European REE Conference

ERES2014: Fist European Rare Earth
Resources Conference

Raw material resources, like minerals and metals, underpin the functioning of the
European and global economy and inescapably govern the quality of life in each
society. In the recent years the term “critical” raw materials has been applied to
further signify raw materials with both economic importance and predicted supply
risk. Among them the Rare Earth Elements (REE), 17 chemically similar metallic
elements, including the 15 lanthanides, scandium and yttrium.
The REEs are used in a wide range of applications, including metallurgy (metal refining
and metal alloying), catalysts in the automotive and the petro‐chemical industry,
colouring of glass/ceramics, phosphors (LEDs, compact fluorescent lamps, flat panel
displays), lasers, rechargeable solid state batteries (Ni‐MH), fiber optics and others.
Additionally, REE are vital components in emerging technologies such as solid state
fuel cells, superconductors, magnetic cooling, hydrogen storage and high performance
permanent magnets. With constantly increasing applications ranging from micro‐
electronics to wind generators and with very few rare earth mines in operation around
the world, the rare earth elements supply and demand is in the global spotlight.
As result, in the last four years there has been both intense political and research
activity in Europe for exploring potential European REE resources and developing
novel processing and recycling technologies. Under the framework of the EURARE:
Development of a sustainable exploitation scheme for Europe’s rare earth ore deposits
project, this scientific conference on European Rare Earth Resources is now being
organized with the hope that it may serve as the ground for exchanging ideas,
networking different groups and furthering academic and applied knowledge on
European Rare Earth Resources.

ERES2014: 1st European Rare Earth Resources Conference|Milos|04‐07/09/2014 1
International Scientific Committee
Prof Ioannis Paspaliaris, Greece (NTUA), Chairman
Prof. Per Kalvig, Denmark (GEUS)
Dr. Nikos Arvanitidis, Finland (GTK)
Prof. Dimitrios Panias, Greece (NTUA)
Prof. Koen Binnemans, Belgium (KU Leuven)
Prof. Bernd Friedrich, Germany (RWTH)
Prof. Xiaosheng Yang, Finland (GTK)
Dr. Dimitrios Eliopoulos, Greece (IGME)
Dr. Georgios Economou, Greece (IGME)
Dr. Martiya Sadeghi, Sweden (SGU)
Prof. Vadim Kessler, Sweden (SLU)
Dr. Kathryn Goodenough, UK (NERC‐BGS)
Dr. Daniel Cassard, France (BRGM)
Dr. Henrik Schiellerup, Norway (NGU)
Dr. Emma Schofield, UK (Johnson Matthey PLC.)
Mr. Carsten Dittrich , Germany (MEAB Chemie Technik GmbH)
Dr. Chris Hall, UK (Less Common Metals)
Dr. Asmo Vartiainen, Finland (Outotec)
Dr. Nicolo Olivieri, Italy (D’Appolonia S.p.A)
Dr. Bertil Grundfelt, Sweden (Kemakta Konsult AB)
Organizing Committee
Dr. Efthymios Balomenos, Greece (NTUA)
Dr. Srecko Stopic, Germany (RWTH)
Dr. Nikolaos Arvanitidis, Sweden (SGU)
Dr. Juha Kajia, Finland (GTK)
Dr. Peter Tom Jones, Belgium (KU Leuven)
Dr. Yiannis Pontikes, Belgium (KU Leuven)
Congress Secretariat
Heliotopos Conferences
Imerovigli, Santorini Island, GR‐84700
Phone: +302286024758 Fax: +302286023672
Email: eres2014@heliotopos.net

EURARE: Development of a sustainable exploitation
scheme for Europe’s rare earth ore deposits

The main goal of the EURARE project is to set the basis for the development of a
European REE industry that will safeguard the uninterrupted supply of REE raw
materials and products to crucial for the EU economy industrial sectors, such as
automotive, electronics, machinery and chemicals, in a sustainable, economically
viable and environmentally friendly way. 23 partners from 10 different European
countries comprise the multi‐actor EURARE consortium.

The EURARE project has received funding from the European Community’s Seventh
Framework Programme ([FP7/2007‐2013]) under grant agreement n°309373. Project
website: www.eurare.eu

RARE³: Research Platform for the Advanced

Recycling and Reuse of Rare Earths
This KU Leuven funded project targets breakthrough recycling processes based on non‐
aqueous technology for the two main applications of rare earths: permanent magnets
and lamp phosphors. This work is part of a more general objective to create fully
integrated, closed‐loop recycling flow sheets. The knowledge platform is steered by an
Industrial User Committee, which includes more than 30 US, EU and Flemish players in
the field of REEs. Moreover, these companies are situated at all different levels of the
REE value chain, from ore extraction, separation, production of consumer goods to
recycling.

The RARE³ project project is funded from the KU Leuven. Project website:
http://kuleuven.rare3.eu
Conference Papers
Session I: Rare Earth Resources ............................................................................... 8
The Rare Earth Crisis And The Critical Materials Institute’s (Cmi’s) Answer ____________________ 9
Karl A. GSCHNEIDNER, Jr. ................................................................................................................. 9
The Role Of Earth Science For The Supply Of Rare Earth Elements __________________________ 12
Yasushi WATANABE, Mihoko Hoshino, Hiroyasu Murakami, Kenzo Sanematsu, Yoshiaki Kon, Yu
Horiuchi ......................................................................................................................................... 12
EXTRACTION AND RECOVERY OF RARE EARTH METALS: CHALLENGES IN PROCESSING __________ 20
Brajendra MISHRA1 , Alexandra ANDERSON2 ................................................................................ 20
China’s Rare Earth Ore Deposits and Beneficiation Techniques ____________________________ 28
Ling Zhi Li1, Xiaosheng Yang2 .......................................................................................................... 28
ECONOMICS OF RARE EARTHS: THE BALANCE PROBLEM _________________________________ 40
Koen BINNEMANS1 ........................................................................................................................ 40
Session II: European Rare Earth Resources (Plenary) ............................................. 52
Unlocking the potential of rare earth resources in Europe ________________________________ 53
Nikolaos ARVANITIDIS1, Kathryn GOODENOUGH2 ......................................................................... 53
THE WORLD‐CLASS REE DEPOSIT TANBREEZ, SOUTH GREENLAND: ITS SIZE AND STRUCTURE _____ 62
Hans K. Schønwandt1, Gregory B. Barnes1 and Thomas Ulrich2 .................................................... 62
Session III‐A: European REE Supply and Demand ................................................... 70
PRIMARY AND SECONDARY SOURCES OF RARE EARTHS IN THE EU‐28: RESULTS OF THE ASTER
PROJECT ________________________________________________________________________ 71
Dominique GUYONNET1, Mariane PLANCHON2, Alain ROLLAT3, Victoire ESCALON2, Stéphane
VAXELAIRE1, Johann TUDURI1 ........................................................................................................ 71
THE ROLE OF RARE EARTHS SUPPLY RISK IN LOW CARBON TECHNOLOGIES INNOVATION _______ 79
Eva BARTEKOVÁ1 ............................................................................................................................ 79
EXPLORING RARE EARTHS SUPPLY CONSTRAINTS FOR A FULL GLOBAL PENETRATION OF
RENEWABLE ENERGY: CASE OF WIND TURBINES AND ELECTRIC VEHICLES ____________________ 94
Komal HABIB1, Henrik WENZEL 2 .................................................................................................... 94
MATERIAL FLOW ANALYSIS FOR IDENTIFYING NEODYMIUM FLOWS AND STOCKS AND RECYCLING
POTENTIALS IN THE EU‐28 ________________________________________________________ 100
Mariane PLANCHON1, Dominique GUYONNET2, Alain ROLLAT3, Victoire ESCALON1, Stéphane
VAXELAIRE2, Johann TUDURI2 ...................................................................................................... 100
Global Pricing Dynamics of the Rare Earths Industry Europe ______________________________ 107
Allan Branch ................................................................................................................................. 107
Session IV: European Framework ......................................................................... 119
EUROPEAN POLICIES ON CRITICAL RAW MATERIALS, INCLUDING REE ______________________ 120
Vasili NICOLETOPOULOS .............................................................................................................. 120
Legislation and best practice in the emerging European rare earth element industry __________ 130
Miranda Keith‐Roach1, Bertil Grundfelt1, Lars Olof Höglund1, Anne Kousa2, Esa Pohjolainen2,
Paloma Magistrati3, Vassiliki Aggelatou4, Nicolò Olivieri5 , Andrea Ferrari5 ............................ 130
A COMPARATIVE ANALYIS OF NATIONAL RAW MATERIALS POLICY APPROACHES ‐ WITH A FOCUS
ON RARE EARTH ELEMENTS IN EUROPE ______________________________________________ 139
Andreas ENDL 1, Gerald BERGER 2................................................................................................ 139
Assessments of boundary conditions and requirements for Rare Earth Underground Mining due to
presence of NORMs ______________________________________________________________ 150
George BARAKOS1, Helmut MISCHO2, Jens GUTZMER3 ............................................................. 150
A FRAMEWORK FOR THE SUSTAINABLE DEVELOPMENT OF RARE EARTH ELEMENTS MINING
PROJECTS ______________________________________________________________________ 158
Sotiris N. KAMENOPOULOS1, Zach AGIOUTANTIS1 ...................................................................... 158
Session III‐B: REE Processing‐ Electrodeposition ................................................... 167
SOLUBILITY OF RARE EARTH OXIDES IN MOLTEN FLUORIDES _____________________________ 168
Xiaoling Guo*, Jilt Sietsma, Yongxiang Yang ................................................................................ 168
ELECTRODEPOSITION OF RARE EARTH METALS IN IONIC LIQUIDS _________________________ 176
E.BOURBOS1a, I.GIANNOPOULOU1, A.KARANTONIS 2, D.PANIAS1b, I.PASPALIARIS1 ..................... 176
ELECTROCHEMICAL RECOVERY OF RARE EARTH ELEMENTS FROM MAGNET SCRAPS‐ A THEORETICAL
ANALYSIS ______________________________________________________________________ 184
V. PRAKASH*, Z.H.I. SUN, J. SIETSMA, Y. YANG ........................................................................... 184
PROPOSAL OF A NEW TYPE OF SULPHIDE ELECTROLYTE FOR RARE EARTH MOLTEN SALT
ELECTROLYSIS __________________________________________________________________ 194
Johannes ÖHL1 ............................................................................................................................ 194
RARE EARTH EXTRACTION FROM NdFeB MAGNET AND RARE EARTH OXIDES USING MOLTEN SALT
PROCESSES _____________________________________________________________________ 201
Aida ABBASALIZADEH1,*, Lidong TENG2, Seshadri SEETHARAMAN3, Jilt SIETSMA1 and Yongxiang
YANG1 .......................................................................................................................................... 201
Session V‐A: REE Processing‐ Mineral Beneficiation / Leaching ........................... 214
ADVANCING KEY TECHNOLOGIES FOR REE MINERAL PROCESSING _________________________ 231
Michael HASCHKE, Horst MÄRTEN, Harald KALKA, Jana NICOLAI ............................................... 231
Unlocking Rare Earth Elements from European apatite‐iron ores __________________________ 241
Bertil I. Pålsson1*, Olof Martinsson2, Christina Wanhainen2 and Andreas Fredriksson3 .............. 241
Application of modern methods for red mud processing to produce rare earth elements. ______ 252
Olga PETRAKOVA1, Gennadiy KLIMENTENOK1, Andrey PANOV1, Sergey GORBACHEV1 ............. 252
HIGH PRESSURE ACID LEACHING OF A NICKEL LATERITE ORE TO EXTRACT SCANDIUM _________ 261
Şerif KAYA1, Yavuz Ali TOPKAYA2 ................................................................................................. 261
LEACHING OF RARE EARTHS FROM BAUXITE RESIDUES USING IMIDAZOLIUM BASED IONIC
LIQUIDS. _______________________________________________________________________ 273
Panagiotis Davris1, Efthymios Balomenos, Dimitris Panias2 and Ioannis Paspaliaris ................... 273
Session VI‐A: REE Processing‐ REE Separation ...................................................... 285
FUNCTIONALISED MIL‐101(Cr) AS NOVEL EUROPIUM ADSORBENT ‐ PROCEEDINGS ___________ 286
Jeroen DE DECKER1, Jeriffa DE CLERCQ2, Els DE CANCK1, Pascal VAN DER VOORT1. ................... 286
MAGNETIC METAL OXIDE BASED ADSORBENT FOR EXTRACTION AND SEPARATION OF REE _____ 294
Gulaim Seisenbaeva1, Peter Svedlindh2,Vadim Kessler 1 ............................................................. 294
ORGANIC‐LIGAND GRAFTED MAGNETIC NANOADSORBENTS FOR EXTRACTION AND SEPARATION OF
REE. __________________________________________________________________________ 302
Elizabeth Polido*, Seda Demirel Topel, Vadim G. Kessler, Gulaim A. Seisenbaeva. .................... 302
APPLYING IONIC LIQUIDS TO RARE‐EARTH SEPARATIONS ________________________________ 311
Kristian LARSSON1, Koen BINNEMANS 1 ....................................................................................... 311
Session V‐B: REE Occurrences in EuropeEURARE IKMS: An Integrated Knowledge
Management System for Rare Earth Element Resources in Europe ...................... 325
Daniel Cassard1, François Tertre1, Guillaume Bertrand1, Frands Schjøth2, Jørgen Tulstrup2, Tjerk
Heijboer2, Jouni Vuollo3 ............................................................................................................... 326
REE DEPOSITS AND OCCURRENCES IN GREENLAND _____________________________________ 336
Kristine THRANE, Per KALVIG & Nynke KEULEN .......................................................................... 336
The Motzfeldt Centre of the Igaliko Nepheline Syenite Complex, South Greenland ‐ A major
resource of REE elements _________________________________________________________ 344
Tapani Tukiainen .......................................................................................................................... 344
ALTERATION OF EUDIALYTE AND IMPLICATIONS FOR THE REE, ZR, AND NB RESOURCES OF THE
LAYERED KAKORTOKITES IN THE ILÍMAUSSAQ INTRUSION, SOUTH WEST GREENLAND _________ 353
Anouk Borst1,2, Tod Waight2, Matthijs Smit2, Henrik Friis3, Troels Nielsen1 ................................ 353
REE‐ MINERALS IN CARBONATITE, ALKALINE AND HYDROTHERMAL ROCKS, NORTHERN AND
CENTRAL FINLAND _______________________________________________________________ 362
Thair Al‐Ani 1* and Olli Sarapää2 ................................................................................................ 362
Rare‐earth elements minerals in carbonatites of the Kola Alkaline Province (northern
Fennoscandia) __________________________________________________________________ 373
Anatoly N. ZAITSEV1,2, Frances WALL 3, Anton R. CHAKHMOURADIAN4 .................................. 373
MINERALOGY AND GEOCHEMISTRY OF STRATEGIC METALS AND REE IN COASTAL HEAVY MINERAL
SANDS FROM NISYROS VOLCANO, GREECE ___________________________________________ 379
Athanasios GODELITSAS1, Ioannis TZIFAS2*, Platon GAMALETSOS1, Giorgos ECONOMOU3, Pavlos
MEGALOVASILIS4 and Paraskevi NOMIKOU1 ............................................................................... 379
Session VI‐B: REE Ore formation .......................................................................... 396
RARE EARTH ELEMENTS IN KARST‐BAUXITES: A NOVEL UNTAPPED EUROPEAN RESOURCE? ____ 397
Éimear DEADY1*, Evangelos MOUCHOS2, Kathryn Goodenough3, Ben Williamson2 and Frances
Wall2 ............................................................................................................................................. 397
THE IMPORTANCE OF TECTONIC SETTING IN ASSESSING EUROPEAN RARE EARTH POTENTIAL ___ 409
Kathryn M GOODENOUGH1,*, Eimear DEADY 2, Frances WALL3, Richard A SHAW2,Paul LUSTY2 . 409
THE PALAEOPROTEROZOIC SKARN‐HOSTED REE MINERALISATIONS OF BASTNÄS‐TYPE: OVERVIEW
AND MINERALOGICAL – GEOLOGICAL CHARACTER _____________________________________ 415
Erik JONSSON1,2*, Karin HÖGDAHL2, Fredrik SAHLSTRÖM2, Per NYSTEN1, Martiya SADEGHI1 .... 415
REE behaviour in hydrothermal alteration zones, Sungun porphyry‐Cu deposit, NW Iran _______ 423
Batoul Taghipour1*, Mohammad Ali Mackizadeh2 ....................................................................... 423
RARE EARTH ELEMENT SIGNATURES IN MANGANESE ORES ASSOCIATED WITH TETHYAN
OPHIOLITES, IRAN; INDICATORS OF PRIMARY ORE FORMING PROCESSES ___________________ 432
Alireza Zarasvandi¹, Mohsen Rezaei1, Martiya Sadeghi2 .............................................................. 432
Session VII: REE Recycling Technologies ............................................................... 444
RECOVERY OF RARE EARTH METALS FROM END‐OF‐LIFE PERMANENT MAGNET SCRAP: AN
OVERVIEW _____________________________________________________________________ 445
Yongxiang YANG .......................................................................................................................... 445
LEACHING OF RARE EARTH ELEMENTS: PAST AND PRESENT ______________________________ 463
Sebastiaan PEELMAN1, Zhi H.I. SUN1, Jilt SIETSMA1, Yongxiang YANG1 ....................................... 463
Extraction of lanthanides from spent polishing agent ___________________________________ 473
Alexander Poscher1, Stefan Luidold2, Holger Schnideritsch3 and Helmut Antrekowitsch4 .......... 473
THE REE POTENTIAL IN “URBAN” ORE DEPOSITS: AN EVALUATION ON CONTENTS AND
PROSPECTING TOOLS FROM ITALIAN MUNICIPAL SOLID WASTE INCINERATORS. _____________ 484
Valerio FUNARI1, Syed Nadeem Hussain BOKHARI2, Thomas MEISEL2, Luigi VIGLIOTTI3, Roberto
BRAGA1 ........................................................................................................................................ 484


Session I: Rare Earth Resources

THE RARE EARTH CRISIS AND THE CRITICAL MATERIALS
INSTITUTE’S (CMI’S) ANSWER

Karl A. GSCHNEIDNER, Jr.
Critical Materials Institue, Ames Laboratory, and
Department of Materials Sciences and Engineering
Iowa State University
Ames, Iowa 50011‐3020 USA
Email: cagey@ameslab.gov
Abstract
The Critical Materials Institute (CMI) of the Ames Laboratory is a U.S. Department of
Energy (DOE) Innovation Hub dealing with energy critical materials. At the present
time CMI is concerned with five rare earth materials (yttrium, neodymium, europium,
terbium and dysprosium), and lithium and tellurium. The Institute has four focus areas
– (1) diversifying supply, (2) developing substitutes, (3) improving reuse and recycling,
and (4) crosscutting research. There are four national DOE laboratories involved (lead
by the Ames Laboratory, together with Idaho National Laboratory, Lawrence Livermore
National Laboratory and Oak Ridge National Laboratory), six industrial organizations
and seven universities.
Extended Abstract

In response to a Department of Energy Funding Opportunity Announcement (FOA) on
critical materials a consortium, called the Critical Materials Institute (CMI), lead by the
Ames Laboratory was informed on January 8, 2013, that they were selected as the
latest DOE Energy Innovation Hub. The actual start date of CMI was June 1, 2013.

The CMI is composed of four focus areas: (1) diversify the global supply chain, (2)
develop substitute materials, (3) enhance recycling, reuse and efficient use of
materials, and (4) crosscutting research. The CMI is led by Alex King (Ames
Laboratory), Director; Rod Eggert (Colorado School of Mines), Deputy Director; and
Karl Gschneidner (Ames Laboratory), Chief Scientist. The CMI consortium is composed
of four national DOE laboratories, six industrial organizations, and seven academic
institutions. The four national laboratories are: the Ames Laboratory, Oak Ridge
National Laboratory, Idaho National Laboratory and Lawrence Livermore National
Laboratory. The six companies are: Advanced Recovery, Inc., Cytec, General Electric
Co., Molycorp, OLI Systems, Inc., and Simbol Materials. The academic institutions are:
Rutgers University, University of California‐Davis, Colorado School of Mines, Brown
University, Iowa State University, Florida Institute of Phosphate Research, and Purdue
University. CMI’s mission is to eliminate materials criticality as an impediment to the
commercialization of clean energy technologies for today and tomorrow.

The CMI addresses seven critical or near‐critical chemical elements: yttrium,
neodymium, europium, terbium, dysprosium, lithium, and tellurium Of these seven
elements only the five rare earth metals are on the European Critical Raw Materials
list. Initially 35 projects were selected for study by applying several criteria: potential
for impact at a key point in a material’s lifecycle in a realistic timeframe; integration of
strengths and capabilities across the Hub; a clear path to deployment, i.e.
commercialization plan in place; an annual evaluation to address continued adherence
to the timeline and the above criteria. As things change projects are expected to be
terminated and new ones started.

Focus area 1 (diversifying supply) has three thrusts: new sources, transformational
processes, and new uses for co‐products. Within the three thrusts there are eight
projects.

Focus area 2 (developing substitutes) has three thrusts: magnets, lighting, and
materials development framework. In this focus area there are nine projects.

Focus area 3 (improving reuse and recycling) has only two thrusts, but nine projects.
The two thrusts are: source separation and concentration, and transformative
extraction and materials production.

Focus area 4 (crosscutting research) consists of three thrusts and nine projects. The
three thrusts are: enabling science; environmental sustainability; and supply chain,
economic analysis and policy.

Within the 35 CMI Projects there are 125 tasks being carried out by about 200
scientists, engineers, and technicians at the various organizations in the USA. This
amounts to about 80 full‐time equivalent technical and 20 full‐time equivalent support
personnel. As a result of recent both internal and external reviews several CMI
projects were combined to improve the interactions between the staff working on
related topic reducing the number of Projects to 30. Furthermore, tellurium was
dropped as one of the seven critical elements because the company who was going to
manufacture CdTe photovoltaic solar cells for energy conversion decided to use other
photovoltaic materials. Also, the funding for several projects were reduced and may
be zeroed out in the next 6 to 12 months depending upon meeting milestones. At least
two new projects are expected to start in the second year, but they have not been
announced.

During the first year of operation the staff of the Institute disclosed 10 inventions for
possible patent protection.

The Critical Materials Institute offers a variety of educational opportunities through
several partners. In particular Colorado School of Mines offers several short courses
such as Recycling Metals from Industrial Waste, and Mineral Processing. While Iowa
State University offers a broad based course on the rare earths, which is available for
distance learning.

The Critical Materials Institute has a membership program in which organizations can
be involved as a Team Member, an Affiliate Member or an Associate Member. Team
members have either research subcontracts from CMI or provide cost sharing funds.
Affiliates (full or observing), through paid membership in the CMI, are informed about
CMI research outcomes, may submit proposals for funding graduate student projects,
and provide input to CMI. Associates are entities who want to access the unique
capabilities and expertise of the CMI on a one‐time basis. Associates, Team and
Affiliate members may provide funding for a specific project via a Cooperative
Research and Development Agreement (CRADA) or a Work for Others (WFO)
agreement; all research agreements are done on a best effort basis and with full cost
recovery.

THE ROLE OF EARTH SCIENCE FOR THE SUPPLY OF RARE
EARTH ELEMENTS
Yasushi WATANABE1, Mihoko Hoshino2, Hiroyasu Murakami3, Kenzo
Sanematsu4, Yoshiaki Kon5, Yu Horiuchi6
1
Department of Earth Resource Sciences, Faculty of International Resource Sciences, Akita
University, 1‐1, Tegata, Akita, 010‐8502 Japan
2, 3, 4, 5
Geological Survey of Japan, AIST, Higashi 1‐1‐1, Tsukuba, 305‐8567 Japan
6
Himeshima Village, Ohita Prefecture, Higashi Kunisaki, Ohita, 872‐1501 Japan
Email1: y‐watanabe@gipc.akita‐u.ac.jp, email2: hoshino‐m@aist.go.jp, email3:
hiroyasu.h‐murakami@aist.go.jp, email4: k‐sanematsu@aist.go.jp, email5: yoshiaki‐
kon@aist.go.jp, email6: kikakusinkou06@vill.himeshima.lg.jp
Abstract
Earth Science can play an important role in various stages of rare earth elements
production, including target defining, geochemical prospecting, evaluation of mineral
resource and reserve, beneficiation, etc. This is because of the limited experience in the
exploration and exploitation of REE by exploration and mining companies outside
China, compared to the other metal commodities. In particular, some of the heavy rare
earth elements demanded by the industry are new targets for the companies. However,
ongoing vast researches on the occurrence, behaviour, concentration and extraction of
REE will be able to ease the criticality of REE in near future.
Introduction
The Behaviour of rare earth elements (REE) in melts and fluids have been widely
applied in earth science as tracer of magmas genesis and evolution, index of oxidation
state of magmas and fluids, geochronology (Sm‐Nd radiometric dating), etc. Despite
the wide recognition of the behaviour of REE in magmas and fluids, it has not been
well understood how REE are transported, fractionated and concentrated to form
economic‐grade rare earth deposits. For example, why only a few carbonatites such
as Mountain Pass and Bayan Obo possess ore‐grade rare earth concentration1, which
ligand (fluorine, chlorine, etc.) is most important for the transportation of REE in
magmas and fluids2, are still debated. Now vast researches on the occurrence,
behaviour, concentration of REE are ongoing, and wide spectrum of knowledge is
accumulating.
Recent development of rare earth products used in high‐technology and/or new
materials (NdFeB magnets, semiconductor, ceramic knives, fuel cell, etc.) has
demanded not only the light rare earth elements (LREE; La‐Eu) but also some of the
heavy rare earth elements (HREE; Gd‐Lu + Y) such as Dy, Tb and Y. Those HREE have
not been well explored in the history, except for xenotime in placer sands along
continental coasts and ion‐adsorption type deposits in southern China. Thus, new
exploration and exploitation projects of HREE meet a lot of challenges. Earth science
can contribute to such projects to overcome difficulties in providing exploration
strategy, knowledge of mineralogy and mineral chemistry, as well as ore mineral
beneficiation. This paper presents some examples that scientific researches
contributed to the REE (especially HREE) exploration projects.
Project targeting
World resource database
Until the progress of many REE exploration projects in recent years, there was no
dataset that show the resource potential of the individual REE such as Nd, Pr, Dy and
Tb that are used for the production of NdFeB magnets. A preliminary attempt was
made in 2006 to evaluate the resource of such specific elements to reveal potential
regions on the basis of the available datasets that compiled whole rare earth resources
of individual deposits in the world3. The result shows that Brazil has the greatest
potential of Nd and Pr, and three regions, Canada, southern Africa and Saudi Arabia,
have good potential of Dy and Tb (Figure 1). Such new data were used for targeting
exploration regions for HREE.

Figure 1: Calculated Nd, Pr, Dy and Tb resources based on the data available before
2006. The amounts of Nd and Pr, and Dy and Tb are combined together, respectively.
Favourable lithology and environment
The present production of HREE is dominated from ion‐adsorption deposits in
southern China. This‐type deposits are formed in weathering profiles of granite and
felsic volcanic rocks. Rare earth elements absorbed on clay minerals formed by
weathering are easily leached with diluted ammonium sulphate solutions. For this
reason, the weathering profiles containing more than 500 ppm ion‐exchangeable REE
with commercially significant sizes (area and thickness) are considered as ores4.
Recent studies show that HREE are concentrated in fractionated ilmenite‐series
granites5, and rare earth elements are present as secondary fluorocarbonates in
addition to refractory apatite and allanite in the host granites of the ion‐adsorption
deposit areas6, 7. Thus, favourable conditions to form HREE‐enriched ion‐adsorption
deposits are 1) tectonically stable and tropical to semi‐tropical climate regions to
preserve thick weathering profiles, and 2) regions of fractionated ilmenite‐series
granitoids, which underwent deuteric alteration to form fluorocarbonates to supply
ion‐exchangeable REE to the profiles. The exploration strategy based on these results
leads to the discovery of ion‐adsorption type mineralization in Southeast Asia such as
Vietnam and Thailand 7, 8.
It has been well known that HREE are typically concentrated in peralkaline rock
complexes, in particular, in the most fractionated parts of the complexes with other
HFSE such as Zr and Nb. Accumulated mineralogical and geochemical knowledge by
scientific researches9,10 immediately leads to the exploration projects in the regions
(northeast Canada, Greenland, and southern Africa) where peralkaline rock complexes
are common. These projects are supported by continuous geochemical and
mineralogical studies at Strange Lake11, Thor Lake12,13, etc., which provide
indispensable information for mineral concentration and REE extraction.
Geochemical prospecting
Recent spread of portable XRF analyser in geochemical prospecting has enabled us to
detect low grade REE mineralization. Comparison of the chemical data obtained by
portable XRF with the results by ICPMS in the laboratory shows that La and Ce is
detectable as low as 150 ppm and Y as low as 30 ppm for the ion‐adsorption clay
samples (Figure 2). These results show that portable XRF analyser is effective in the
survey of ion‐adsorption and other type REE deposits, in particular, HREE exploration,
because Y detection limit is lower than La and Ce. This method was applied to delineate
Y anomalies in a fluorite prospect in South Africa14 (Figure 3).


Figure 2: Comparison of analytical results by portable XRF and ICPMS for the samples
of ion‐adsorption clay in Conception, Chile. Vertical and horizontal axes are the results
by XRF and ICPMS, respectively.

Figure 3: Y concentration (ppm) in termite mounds in the Blockspruit prospect in South
Africa, measured by portable XRF analyser14.
Mineralogy and mineral chemistry
A number of rare earth minerals are present in alkaline rock complexes and
pegmatites, and detailed mineralogical works are indispensable for the recovery of
these minerals. Some of the minerals contain minor amounts of REE and if these
minerals contain important elements (Dy, Tb), they become a target for exploration
and exploitation. It has been known that zirconium silicates are occasionally
concentrated in HREE15, although, these minerals have not been regarded as a
resource source until recently, due to scarcity of geochemical data and lack of the
established beneficiation methods. Recent advance in trace element quantitative
analysis by LA‐ICPMS has made it easier to determine rare earth concentration in such
minerals16, 17.
The Jabal Tawlah deposit in Saudi Arabia is reported to be HREE‐concentrated
microgranite with Nb and Zr in the Proterozoic sedimentary rocks. The granite forms
a small sill‐like body with 330m extension and a maximum thickness of 50m, and 6.4Mt
of ore is estimated18. The microgranite is heavily fractionated and is composed of
quartz, albite and microcline with accessory zircon, Nb oxides and REE minerals. Total
REE content including Y is about 1.2wt% as oxides. Major ore minerals are columbite,
fergusonite, Ca‐Y‐F minerals, and xenotime. A mineralogical study shows that Ca‐Y‐F
minerals are composed of waimirite‐(Y) (YF3) and a Ca‐Y‐F undetermined phase (Figure
4). Waimirite19 is a mineral approved by IMA in December 2013, and nearly half of REE
in the deposit is present in these new and undetermined phases, sharing the other half
with xenotime. The deposit is regarded to be the highest Dy grade deposit in the world.

Figure 4: SEM image of Ca‐Y‐F phase and waimirite‐(Y) (a) and chemical compositions
of these minerals determined by EPMA. The compositions of tveitite‐(Y) from Norway
and Russia are also shown.

One of the target minerals for HREE extraction is eudialyte that contains a few to 10
wt% of REE. Because eudialyte is relatively enriched in HREE and occur abundantly in
peralkaline complexes, several projects aim to produce REE from eudialyte. However,
the REE content in eudialyte is variable even in the same unit of an intrusion and
precise chemical analysis is necessary. An example is the Ilímaussaq peralkaline
complex in Greenland, discovery place of eudialyte, which is divided into lower
kakortokite and upper lujavrite. The REE compositions of eudialyite determined by LA‐
ICPMS show that eudialytes in the kakortokite and lujavrite contain 2.2‐2.3wt% and
4.1‐4.9wt% REE, respectively20.
A whole rock analytical method using LA‐ICPMS newly developed is to use a glass bead
ablation method similar to XRF techniques17. This method has several advantages: 1)
higher sensitivity than that achieved by the XRF method, 2) obviation of erroneous
measurements due to incomplete dissolution of heavy minerals, and 3) simple and
rapid sample preparation procedures for the analysis of both major and trace
elements. The precisions of this method are better than 10% for Y, La, Pr and 20% for
the other REE, and the method can provide REE compositional data of rocks faster than
conventional ICPMS analysis.
Extraction
The extraction techniques of REE from conventional minerals (bastnäsite, monazite,
xenotime) and ion‐adsorption clay were well established21,22. However, to establish
economically feasible REE extraction methods for other unconventional minerals are
challenging.
Apatite
It has been known that apatite contain minor amounts of REE23, and REE by‐production
has been explored in the process of phosphoric acid production. The methods using
hydrochloric or nitric acid can extract more than 80% of REE, but when sulphuric acid
is used, recovery of REE is low because majority of REE is incorporated into gypsum
residue24. Various attempts are being conducted for the extraction of REE from
apatite, including a hydrochloric acid method25.
Zircon
Zircon is a refractory mineral and generally contains less than 1wt% of REE, however,
many varieties of zircon containing a large amount of REE, Th, U, Nb and Ta have been
reported from peralkaline rocks and granites13. A REE extraction experiment indicates
that more than 90% of REE can be leached out from REE‐bearing metamict zircon with
1M‐HCL solution at temperature of 150°C26.
Conclusion
Earth Science can play an important role for the supply of REE, by providing exploration
strategy, target determination, geochemical prospecting, evaluation of resource and
reserve, and beneficiation. This is because the exploration and mining experience is
still immature for the mining sectors outside China. However, abrupt increase of REE
demand, in particular, some specific elements such as Dy and Tb, urged industrialised
countries to secure raw REE materials. Vast scientific researches are ongoing to clarify
the occurrence, behaviour, concentration of REE in rocks, minerals, and solutions,
which will be able to ease the criticality of REE in near future.
References
1. C. Wu, “Bayan Obo controversy: Carbonatites versus iron oxide‐Cu‐Au‐(REE‐U)”, Resource Geol.,
58 (4) 348‐354 (2008).
2. A. Williams‐Jones, “Rare earth element ore genesis‐The great unknown”, in Goldschmidt 2013
Conference Abstracts, p. 2501 (2013).
3. G. J. Orris, R. I. Grauch, “Rare earth element mines, deposits, and occurrences”, USGS Open‐File
Report 02‐189, 78p. (2002).
4. Z. Bao, Z. Zhao, “Geochemistry of mineralization with exchangeable REY in the weathering crusts
of granitic rocks in South China”, Ore Geology Review, 33 (3‐4), 519‐535 (2008).
5. S. Ishihara, H. Murakami, “Fractionated ilmenite‐series granites in southwest Japan: Source
magma for REE‐Sn‐W mineralizations”, Resource Geology, 56 (3), 245‐256 (2006)
6. S. Ishihara, R. Hua, M. Hoshino, H. Murakami, “REE abundance and REE minerals in granitic rocks
in the Nanling Range, Jiangxi Province, Southern China, and generation of the REE‐rich weathered
crust deposits”, Resource Geology, 58 (4), 355‐372 (2008).
7. K. Sanematsu, Y. Kon, A. Imai, K. Watanabe, Y. Watanabe, “Geochemical and mineralogical
characteristics of ion‐adsorption type REE mineralization in Phuket, Thailand”, Mineralium
Deposita, 48 (4), 437‐451 (2013).
8. T. Mentani, T. Ohmura, Y. Watanabe, T. Urabe, “So‐called ion adsorption type” REE deposits found
in weathered crust of ilmenite‐series granite in northern Vietnam”, in the Geological Society of
America Abstracts with Programs, v. 42, n. 5, p. 336, 2010.
9. S. Salvi, A. E. Williams‐Jones, “The role of hydrothermal processes in the granite‐hosted Zr, Y, REE
deposit at Strange Lake, Quebec/Labrador: Evidence from fluid inclusions, Geochimica et
Cosmochimica Acta, 54 (9), 2403‐2418 (1990).
10. R. P. Taylor, P. J. Pollard, “Rare earth element mineralization in peralkaline systems: the T‐zone
REE‐Y‐Be deposit, Thor Lake, Northwest Territories, Canada”, in Rare Earth Minerals, Edited by A.
P. Jones, F. Wall and C. T. Williams. Chapman & Hall, London, UK, 1996.
11. A. Kerr, H. Rafuse, “Rare‐earth element (REE) geochemistry of the Strange Lake deposits:
Implications for resource estimation and metallogenic models”, Current Research, Newfoundland
and Labrador Department of Natural Resources, Geological Survey, Report 12‐1, 39‐60 (2012).
12. E. R. Sheard, A. E. Williams‐Jones, M. Heiligmann, C. Pederson, D. L. Trueman, “Controls on the
concentration of zirconium, niobium, and the rare earth elements in the Thor Lake rare metal
deposit, Northwest Territories, Canada”, Economic Geology, 107 (1), 81‐104 (2012).
13. M. Hoshino, Y. Watanabe, H. Murakami, Y. Kon, M. Tsunematsu, “Formation process of zircon
associated with REE‐fluorocarbonate and niobium minerals in the Nechalacho REE deposit, Thor
Lake, Canada”, Resource Geology, 63 (1), 1‐26 (2013).
14. Y. Horiuchi, T. Ohno, M. Hoshino, K‐C. Shin, H. Murakami, M. Tsunematsu, Y. Watanabe,
“Geochemical prospecting for rare earth elements using termite mound materials”, Mineralium
Deposita, 49 (in press).
15. Z. Johan, V. Johan, “Accessory minerals of the Cínovec (Zinnwald) granite cupola, Czech Republic:
indicators of petrogenetic evolution”, Mineralogy and Petrology, 83 (1‐2), 113‐150 (2005).
16. F. Y. Wu, Y. H. Yang, M. A. W. Marks, Z. C. Liu, Q. Zhou, W. C. Ge, J. S. Yang, Z. F. Zhao, R. H. Mitchell,
G. Markl, “In situ U‐Pb, Sr, Nd and Hf isotopic analysis of eudialyte by LA‐(MC)‐ICP‐MS”, Chemical
Geology, 273, 8‐34 (2010).
17. Y. Kon, H. Murakami, Takagi, T., Y. Watanabe., “The development of whole rock analysis of major
and trace elements in XRF glass beads by fsLA‐ICPMS in GSJ geochemical reference samples”,
Geochemical Journal, 45 (5), 387‐416 (2011).
18. A. R. Drysdall, N. J. Jackson, C. R. Ramsay, C. J. Douch, D. Hackett, “Rare element mineralization
related to Precambrian alkali granites in the Arabian Shield, Economic Geology, 79 (6), 1366‐1377
(1984).
19. O. R. Rigon Minuzzi, J. M. T. Mirras Ferron, A. C. Bastos Meto, V. P. Pereira, “Primeira Notícia da
descoberta de waimirita e atroarita, dois novos minerais na Mina de Pitinga, AM, Brasil”, Pesquisas
em Geociências, 30 (1), 99‐101 (2003).
20. K. Pfaff, T. Krumrei, Marks, M., Wenzel, T., Rudolf, T., Markl, G., “Chemical and physical evolution
of the ‘lower layered sequence’ from the nepheline syenitic layering in peralkaline felsic liquids”,
Lithos, 106 (3‐4), 280‐296 (2008).
21. F. Xie, T. A. Zhang, D. Dreisinger, F. Doyle, “A critical review on solvent extraction of rare earth
from aqueous solutions”, Minerals Engineering, 56 (1), 10‐28 (2014).
22. G. A. Moldoveanu, V. G. Papangelakis, “Recovery of rare earth elements adsorbed on clay
minerals: I/ Desorption mechanism, Hydrometallurgy, 117‐118, 71‐78 (2012).
23. P. M. Ihlen, H. Schiellerup, H. Gautneb, Ø. Skår, “Characterization of apatite resources in Norway
and their REE potential‐A review”, Ore Geology Review, 58, 126‐147 (2014).
24. F. Habashi, “The recovery of the lanthanides from phosphate rock”, J. Chem. Tech. Biotechnol.,
35A, 5‐14 (1985).
25. F. Pereira, E. Bilal, “Phosphoric acid extraction and rare earth recovery from apatites of the
Brazilian phosphatic ores”, Romanian Journal of Mineral Deposits, 86 (2), 49‐52 (2012).
26. M. Hoshino, Y. Watanabe, K. Sanematsu, Y. Kon, M. Shimizu, “Characteristics of zircon suitable for
REE extraction”, International Journal of Mineral Processing, 102‐103, 130‐135 (2012).

EXTRACTION AND RECOVERY OF RARE EARTH METALS:
CHALLENGES IN PROCESSING
Brajendra MISHRA1 , Alexandra ANDERSON2
Department of Metallurgical & Materials Engineering, Colorado School of Mines
1500 Illinois Street, Golden, Colorado, 80401, USA
1
bmishra@mines.edu, 2allieanderson0@gmail.com
Abstract
Rare earths metals, including yttrium and scandium, are being increasingly used in
clean energy technologies, colored phosphors, lasers and high intensity magnets. The
commitment to clean energy technologies by governments and the projected growth
in power and transportation sectors across the globe ensure that the demand for rare
earth metals and compounds would continue to escalate. This demand implies that, to
ensure unhindered technological innovation, it is essential to possess secure supply
chains for rare earth elements. In order to ensure secure rare earth supply and
attenuate supply‐demand imbalance post 2014, it is of utmost importance to look at
opportunities to process intelligently, recycle and reuse Rare Earth Elements from
secondary sources, such as post‐consumer and manufacturing process wastes.

Introduction

Increasingly, the U.S. government, academia, domestic industry, and the public
acknowledge the imperative that we need to conserve energy and natural resources
while exercising judicious stewardship of the environment. The issue of sustainability
is and should be paramount in how we design, manufacture, use, and retire the many
products we consume throughout the world. Inorganic materials are not renewable;
the need exists for the development of technologies to address materials recovery and
recycling. Research supporting materials recovery and recyclability is inherently
multidisciplinary and must respond to the needs of a multiplicity of commercial
stakeholders from throughout the materials supply chain.

Despite growing efforts to recycle metals, we fail to recover half of the domestic post‐
consumer metal scrap reclaimable from retired products, and we continue to rely on
primary metals production to fulfil two thirds of our manufacturing needs. Use of
primary metals, in lieu of scrap, increases global energy consumption as well as the
production of greenhouse gases. In order to augment recycling rates the materials
community needs to upgrade recovery and recycling processing technologies to
maximize the capture of post‐consumer scrap and minimize the quantity of
manufacturing scrap.

Rare earth elements are a group of 17 elements, which include 15 Lanthanides,
Scandium and Yttrium. In spite of what the name suggests these elements are not
“rare”. However, in recent years rare earth elements have become strategically critical
for developed and developing economies around the world which is primarily due to
the shortage of discovered minable resources1. Before 1948, the placer deposits of
Brazil and India were the chief sources of rare earth metals for the rest of the world.
With increasing demand newer supply sources were needed and for a while the
monazite deposits in South Africa played an important role before the production was
dominated by Bastnasite reserves in Mountain Pass and China2.

According to a forecast done by IMCOA, the world rare earth demand is projected to
rise to 200,000 tons by 2014 and the Chinese production is expected to be around
160,000 tons3. In addition, the demand of rare earths in China itself has increased by
380% between 2000 and 2009, which is believed to be the primary reason behind the
export cuts on rare earth4. These developments have made rare earth elements a
strategically important material as evident by the Rare Earths and Critical Material
Revitalization Act of 2010 approved on September 29, 2010 which aims to establish an
R&D program within the DOE to assure long term supply of rare earth materials5. The
US Department of Energy published an analysis of the criticality of selected rare
metals, the most critical elements were identified to be Dysprosium, Neodymium,
Terbium, Yttrium and Europium ‐ what are also known as the heavy rare earth
elements6. Based on the demand and supply position of common rare‐earths, the
prices of common metals like Ce, Nd, Sm, La and Y, went up by 150% to 700% in a short
period of six months between January and August 2010.

Preparation of Rare Earth Oxides

Rare earth oxides can be produced by five commonly used methods: selective
oxidation, selective reduction, fractional crystallization, ion exchange and solvent
extraction.

Selective Oxidation: The rare earth metals that can be separated by selective
oxidation are cerium, praseodymium and terbium. Selective oxidation is effective for
these three rare earth metals because they can be oxidized from trivalent state to a
tetravalent state, which allows for separation. The most commonly used oxidizing
agents used for these reactions are persulfates, permanganate and hypochlorite. Air
or oxygen is also used as an oxidizing agent1. In the selective oxidation of
praseodymium and terbium, Pr (IV) and Tb (IV) will precipitate out of solution as rare
earth oxides and settle because the tetravalent states for praseodymium and terbium
are not stable in an aqueous solution. Whereas, in the selective oxidation of cerium,
the tetravalent cerium must be separated from the rare earth mixture by selective
dissolution of trivalent elements in a dilute acid or complete dissolution of the species
in a concentrated acid followed by precipitation. The tetravalent cerium can be
precipitated from the acidic solution through pH control2.

Selective Reduction: The rare earth metals that can be separated by selective
reduction are samarium, europium and ytterbium. Selective reduction is effective for
these three rare earth metals because they can be reduced to a divalent state. McCoy
developed the commercially used method for selective reduction of Eu. In this
method, Eu (III) is reduced to Eu (II) by zinc in a chlorine solution. The divalent Eu is
precipitated from the solution as a sulfate and is then dissolved in HCl, which causes
the formation of an oxide3. This method does not work for samarium or ytterbium
because zinc does not reduce these metals. Marsh developed the commercially used
method for the selective reduction of ytterbium and samarium. In this method, rare
earth mixtures are placed in an acetate solution. Samarium and europium can then
be separated by reductive extraction into dilute sodium amalgam4.

Fractional Crystallization: Fractional crystallization is a process of separating the
components of a solution on the basis of their different solubilities. In this method,
the solution is evaporated until the least soluble component in solution crystallizes
out. In the fractional crystallization of rare earths, rare earth solutions are mixed with
a salt and then heated. The solutions are then allowed to gradually cool so that rare
earth oxide crystallizes out of solution. This method is usually most suitable for
lanthanides at the lower end of the series4. Ammonium nitrate salts are commonly
used for the separation of praseodymium and neodymium from solutions. Double
magnesium nitrates are commonly used for the separation of samarium, europium,
gadolinium and the ceric group. Bromates and ethyl sulfates are used in the
separation of the yttric group2.

Ion Exchange: Ion exchange reactions are a two‐step process that involves the
interchange of ions between an aqueous solution and an insoluble resin. The first step
is absorption, which removes the metal ions from the solution and the second step is
elution, which is the recovery of metal compounds from the resin bed. This method
suitable for rare earth separation on a small scale because it is a batch process. A
commercially used ion exchange method involves the recovery of rare earth
compounds through the use of amberlite type resins and elution with completing
agents such as citric acid. EDTA and HEDTA are two other commonly used completing
agents because most all rare earth combinations can be successfully separated
through the use of these two agents5.

Solvent Extraction: Solvent extraction is the most economical method for separating
rare earth metals today. In solvent extraction, an aqueous solution is mixed with an
immiscible organic solvent. The metal is then transferred to the organic phase and
there is a phase separation due to the different densities. In the solvent extraction of
rare earth metals, the aqueous solution contains the desired metals to be separated,
which are subsequently transferred to the organic phase. The organic extractants
used in this method can be classified as neutral, acidic, basic or a chelating agent.
Neutral organic extractants generally yield low separation factors, which make them
suitable for the recovery of mixed rare earth oxides. TBP (tributyl phosphate) is the
most commonly used neutral extractant. Basic organic extractants are used to recover
anionic lanthanides from aqueous solutions. Acidic extractants are very powerful and
can be used to separate most rare earth metals. DEPHA is the most commonly used
acidic extractant because it can be used to recover rare earth compounds from many
different types of solutions4.

Methods for producing rare earth metals
The three primary methods practiced for producing pure rare earth metals from
compounds. These three methods are (1) reduction of anhydrous chlorides and
fluorides (2) direct reduction of rare earth oxides, and (3) fused salt electrolysis of rare
earth chlorides or oxide fluoride mixtures6.

Direct Reduction of Rare Earth Oxides: Direct reduction of rare earth oxides process
does not require the pre‐transformation of the rare earth oxides into halides. The
earliest attempts to directly reduce rare earth oxides to metals were made between
1890 and 1912 by Winckler, Matignon, and Hirsch. They used magnesium or calcium
as a reducing agent for the rare earth oxides. Although, this method proved to be
ineffective because the resulting mixture consisted of magnesium oxide or calcium
oxide and a metal, which could not be separated. Several years later, Mahn revisited
the magnesium reduction route. His method was to dip blocks of rare earth oxides in
magnesium that was heated to 1050 °C. Although the magnesium metal attacked the
metal oxide block, the only products obtained were very dilute solutions of rare earth
metals in magnesium2.

In 1953, a major breakthrough was made by Daane for the reduction of volatile rare
earth metals. In this process, rare earth oxides are mixed with lanthanum chips in a
tantalum crucible fitted with a tantalum condenser. The mixture is heated to high
temperature under vacuum so that the metal produced by the lanthanum reduction
sublimes as crystalline condensates on the crucible wall. The metal was then collected
from the crucible wall. The metal contained residual La, O, and H, but these impurities
were removed by subliming the metal once more7. Onstatt also used this reduction –
distillation method, but he used calcium instead of lanthanum as the reducing agent.
This method turned out to be less effective than the lanthanum method because the
metal had to be distilled three to four times in order to remove the majority impurities
and the metal was only around 40%. Achard experimented using carbon as the
reducing agent for the reduction – distillation process, but the distilled product from
this process contained high amounts of oxygen. Achard found that a second
distillation at a higher pressure resulted in purer rare earth metals2.
In 1988, General Motors developed a molten salt process for the direct reduction of
rare earth metals. To perform this process, a rare earth oxide is placed inside a
reaction vessel with calcium chloride and sodium chloride. The sodium chloride is
used to reduce the calcium chloride to calcium metal. The calcium metal then is used
to reduce the rare earth oxide to a rare earth metal. The reaction vessel is then heated
to a temperature above the melting point of the constituents but below the
vaporization temperature of the sodium metal. The molten constituents must be
rapidly stirred in order to keep them in contact with one another. As the reaction
progress, calcium chloride is added to the reaction bath as necessary to main a weight
percent of 70% of the combined weights of both the calcium chloride and the sodium
chloride. The following reaction explains the process:

REnOm + mCaCl2 +2m Na nRE +m CaO +2m NaCl

The “n” and “m” quantities are the number of moles of constituent and the relation of
n and m is determined by the oxidation start of the rare earth metal. When the
reaction is complete and the stirring has stopped, the reduced metal can be recovered
in a clean layer from the bottom of the reaction vessel8.

Reduction of Anhydrous Rare Earth Halides: The second method for the production
of rare earth metals is through the reduction of their halides. This reduction method
requires the pre‐transformation of the rare earth oxides into either chlorides or
fluorides. The rare earth halides are less stable than the rare earth oxides which means
that there are more options available for reducing agents.

Metallothermic Reduction of Rare Earth Chlorides: Trombe and Mahn performed one
of the first metallothermic reductions of rare earth chlorides. They prepared cerium,
neodymium and gadolinium by reduction of their trichlorides with magnesium. This
reduction produced magnesium – rare earth alloy. The magnesium was subsequently
removed through vacuum distillation. A few year after Trombe and Mahn experiment,
Keller and Karl prepared cerium metal by reducing cerium dichloride with calcium in
dolomite lined steel bombs. A Ca – I booster reaction was also used in this process to
achieve a good slag – metal separation. The cerium metal was then vacuum melted
to remove the calcium slag. The metal was obtained in 99% yield2.

Spedding and Daane also developed a technique for producing La, Pr, Nd, and Gd
metals that utilized calcium as a reducing agent. The rare earth oxide and calcium
reaction was performed in tantalum crucible that was heated above the melting
temperature of the metal being reduced. The reduction was also performed in a
vacuum or inert atmosphere. The metal was obtained in 99% yield9.

Lithium and sodium have also been used as reducing agent for rare earth chlorides.
Block and Campbell were able to prepare high purity yttrium by utilizing both elements
as reducing agents. In one technique, the chloride and lithium were placed in a
molybdenum crucible and heated to 850 – 900 °C. The reaction took place in an inert
atmosphere for four to five hours. The excess lithium and lithium chloride was
removed through vacuum distillation. The metal was obtained in 95 – 99% yield2. In
another technique, sodium is distilled into an yttrium chloride crucible. The reaction
took place at 800 – 850 °C for 5 ‐7 hours. The sodium chloride slag was then poured
off and any excess sodium was removed through distillation. Only 61 to 85 % of the
metal could be recovered through this process6.

More recent metallothermic studies have focused on lithium vapor reduction of rare
earth trichlorides. Nolting, Simons and Klingenberg have used lithium vapor reduction
of rare earth trichlorides to prepare yttrium and other rare earth metals that have high
melting temperatures. This process involves reducing yttrium trichloride with lithium
vapor inside a sealed stainless steel chamber. This reduction produces a lithium
chloride slag, which is then distilled away from remaining yttrium crystals. The metal
crystals are consolidated through the arc melting process10.

Metallothermic Reduction of Rare Earth Fluorides: Metallothermic reduction of rare
earth fluorides have also been studied but these processes were developed much later
than processes involving rare earth chlorides and not used as often as rare earth
chloride reduction.

In the early 1950’s, Gray performed experiments in which rare earth fluorides were
reduced with lithium inside a steel bomb. Iodine was added to the reaction to lower
the slag melting point and the mixture was heated to 1220 °C. Cerium, lanthanum and
neodymium were all successfully recovered from this method11.
Soon after Gray’s method was tested, Daane and Spedding developed a very successful
technique for reducing rare earth fluorides. In this process, calcium is used as the
reducing agent. Calcium and the rare earth fluoride are heated in an argon
atmosphere at very temperatures. This process resulted in good metal – slag
separation and the slag could be removed from the metal through vacuum melting.
This method is very similar to what was adopted for industrial practice7.

Reduction through Molten Salt Electrolysis: Molten salt electrolysis is also a viable
option for the production of rare earth metals. Electrolytic methods are typically less
complex and expensive than metallothermic reduction methods. This type of process
is not completely limited by considerations of chemical stability. Instead, electrolytic
methods rely on the theoretical decomposition potential to determine feasibility.

Molten Salt Electrolysis of Rare Earth Chloride Mixtures: Lanthanum, cerium,
praseodymium and didymium (alloy of neodymium) have been produced by fused salt
electrolysis of chlorides on a commercial scale. The maximum operating temperature
of a chloride electrolysis cell is 1100 °C because at higher temperatures the volatility
of the electrolyte constituents, resulting is severe attack of the cell wall material by
rare earth metals. Lanthanum, cerium, praseodymium and didymium have been
successfully recovered through chloride electrolysis because all of these elements
have melting points less than 1100 °C. Rare earth elements with high melting
temperatures, such as gadolinium, dysprosium and yttrium, have been recovered in
limited quantities by fused salt electrolysis of their fluorides rather than their
chlorides12.

The technique for fused salt electrolysis of chlorides that was adopted in industry uses
an iron or carbon cathode and a graphite anode. The cathode is either the cell
container itself or carbon or ions block at the bottom of the cell. The electrolysis takes
place in a ceramic, iron, carbon or graphite lined reactor. The electrolyte contains the
rare earth chloride and NaCl, KCl, BaCl2 or CaCl2. The electrolyte bath is typically
heated between 800 – 900 °C. Lanthanum, cerium and didymium metals of 99% or
greater purity have been produced by this method12.

The current efficiencies of these cells were usually between 45 – 50%. The current
efficiencies in these cells are influenced by various factors. At the cell operating
temperature (typically 900°C), the rare earth metals can react with O2, H2, CO2 or
graphite forming products that have high melting points and low electrical
conductivity, thus decreasing the current efficiency. The presence of oxychlorides can
also decrease the current efficiency. Oxychlorides have high melting points than the
rare earth chlorides and have limited solubility in the electrolyte. Their presence in
the melt increases the viscosity causing current losses due to the lower mobility of the
ions in solutions and high operating temperatures6.

Molten Salt Electrolysis of Rare Earth Oxide‐Fluoride Mixtures: Molten salt
electrolysis of rare earth oxide‐fluoride mixtures is typically used to produce rare earth
metals with high melting temperatures. The first successful technique for producing
metal through oxide‐fluoride electrolysis was developed by Gray in the early 1950’s.
The process utilized a graphite anode and molybdenum cathode. A rare earth oxide
was dissolved in a REF3 – LiF – BaF2 electrolyte and the mixture was externally heated
in a carbon reactor. The reaction also took place in an argon gas atmosphere and an
inert crucible was placed below the cathode to recover the rare earth metal.

During the 1960’s, much of the research and development of oxide‐fluoride
electrolysis was carried out by the Reno Metallurgy Center of the United States Bureau
of Mines. An important feature of the electrolytic cells that were developed at Reno
was that there was layer of frozen bath on the interior surface of the graphite reactor.
This layer was called a skull and kept the rare earth metals from coming in contact with
the graphite crucible. The cells were first used to recovery cerium metal. During
electrolysis, cerium nodules formed on the surface of the skull. After the procedure
was complete and the cell was cooled to room temperature, the skull was taken into
open air and broken up to recover the cerium nodules. This technique was also used
to recover neodymium, praseodymium and didymium. The current efficiencies for
these cells typically ranged from 75‐95%. The electrolyte consisted of the RE2O3, REF,
LiF, and BaF2. Rare earth metals could be produced in 99% or greater purity through
this method12.
Summary
Very small amount of rare‐earth metals is produced outside China. Several methods
have been studied on a laboratory scale as well as some on the pilot level in other
countries. The high thermodynamic stability of rare‐earth oxides as well as halides
makes the metal production challenging. Metallothermic reduction of oxides has
shown better potential for RE production than reduction of halides or processes based
on electroreduction.
References

[1] D. Bauer and R. Lindstrom, “RECOVERY OF CERIUM AND LANTHANUM BY OZONATION OF
LANTHANIDE SOLUTIONS.,” no. May, 1968.
[2] C. Gupta and N. Krishnamurthy, “Extractive metallurgy of rare earths,” Int. Mater. Rev., vol.
37, no. 5, 1992.
[3] N. Mccoy, “The Extraction of Europium from Monazite Residues and the Preparation of Pure
Europium Compunds,” vol. 1145, no. 3, 1936.
[4] M. Albander, “Chemical Extraction, Benefication, Production and Application of the Rare Earth
Metals,” 2012.
[5] F. H. Spedding, “The Separation of Rare Earths by Ion Exchange,” no. 7, 1947.
[6] C. Gupta and N. Krishnamurthy, Extractive metallurgy of rare earths. Boca Raton, Fl: CRC Press,
2005.
[7] A. Daane and F. Spedding, “Preparation of yttrium and some heavy rare earth metals,” J.
Electrochem. Soc., pp. 9–11, 1953.
[8] R. A. Sharma, “Umted States Patent [ 19 ] Date of Patent :,” 1986.
[9] F. H. Spedding, H. A. Wilhelm, W. H. Kelleri, D. H. Aiimann, A. H. Daane, C. C. Hach, and R. P.
Ericson, “Production of Pure Rare Earth Metals,” pp. 553–556.
[10] H. Nolting, C. Simmons, and J. Klingenberg, “Preparation and properties of high purity yttrium
metal,” J. Inorg. …, vol. 14, no. 6, pp. 208–216, 1960.
[11] P. M. . Gray, “Production of Cerium Metal by Electrolytic and Thermal Reduction Processes.”
[12] N. Krishnamurthy and C. Gupta, “Rare Earth Metals and Alloys by Electrolytic Methods,” 2002.


CHINA’S RARE EARTH ORE DEPOSITS AND
BENEFICIATION TECHNIQUES
Ling Zhi Li1, Xiaosheng Yang2

1
China Western Mining Co., Ltd, Xining, Qinghai, 810001 P.R.China
Email: lzli8288@sina.com

2
Mineral Processing Laboratory, Geological Survey of Finland, 83500 Outokumpu, Finland
Email: jason.yang@gtk.fi

Abstract
China possesses the largest resource of rare earth elements (REE) in the world and the
reserve of REE in China takes up over 40% of the total global reserve. China has
dominated the world REE production since 1980s. Bayan Obo REE‐Nb‐Fe deposit
situated in Inner Mongolia is the world’s largest known REE ore deposit with the reserve
of 48 Mt REO and the largest REE production resource. Other two major in production
REE deposits are Mianning REE deposit in Sichuan and Weishan REE deposit in
Shandong. The weathered crust elution –deposited REE ore (ion adsorption REE
deposit), as the world unique REE deposit found in China, is currently the second largest
REE production resource containing high contents of heavy REE. In the paper, the ore
properties of these REE deposits are introduced and the renovation and progress on
the beneficiation techniques including flowsheets, flotation and leaching reagents and
technological indexes on these ores are discussed.

Introduction
China has the largest resource of rare earth elements (REE) in the world. According to
Mineral Commodity Summaries in 2001 the reserves and reserve bases of REE in the
world in 2000 are shown in Table 1. It is indicated that the reserve of REE in China takes
up over 40% of the world reserve. The newest Mineral Commodity Summaries 2014
shows the similar results.

The resources of rare earth elements (REE) in China are mainly divided into four types
of deposits which consist of 95% total REE resources of China. They are the Bayan Obo
REE−Fe−Nb Ore Deposit in Inner Mongolia, the Mianning REE Ore Deposit in Sichuan
Province, the Weishan REE Ore Deposit in Shandong Province and the Weathered
Crust Elution−deposited Rare Earth Ore (The Ion Adsorp on Type Rare Earth Ore)
mainly in Jianxi Province, but also distributed in Guandong, Fujian and Guanxi
Provinces.
Table 1: the reserves and reserve bases of REE in the world in 2000 (REO, 104 t)
Country reserve Reserve % Country reserve Reserve %
base base
China 4300 4800 43 South 39 40 0.4
Africa
Russia 1900 2100 19 Brazil 28 31 0.3
USA 1300 1400 13 Malaysia 3.0 3.5 0
Australia 520 580 5.2 Sri Lanka 1.2 1.3 0
India 110 130 1.1 Other 1704.8 1814.2 17
Canada 94 100 0.9 Total 10000 11000 100

The distributions of REE in the China’s REE ore deposits are shown in Table 2. It is
observed that high percentages of three main light rare earth elements La, Ce and Nd
are found in Bayan Obo REE−Fe−Nb Ore Deposit, Sichuan Mianning REE Ore Deposit
and Shandong Weishan REE Ore Deposit. However, in the Ion Adsorption Type Rare
Earth Ore (especially type B) high contents of Y and other heavy REE are contained.

Table 2: the distributions of REE in the China’s REE ore deposits
Content of Bayan Obo Sichuan Shandong Ion Adsorption Type Rare
REO REE−Fe−Nb Mianning REE Weishan REE Earth Ore
Ore Ore Ore Type A Type B
La2O3 24 − 26 28 − 30 32 31 − 40 2−5
CeO2 50 45 − 50 48 − 50 3−7 1−2
Pr6O11 3−5 5 4 7 − 11 1−2
Nd2O3 16 − 18 12− 14 11 − 12 26 − 35 3−5
Sm2O3 1.5 1.5 − 2 1.0 – 1.8 4−6 2−4
Eu2O3 0.2 0.4 0.1 – 0.2 0.5 0.12
Gd2O3 0.4 0.8 – 1.0 0.3 – 0.5 4 6
(Tb – Lu)2O3 0.2 – 0.3 1 0.4 – 0.6 4–5 15 − 20
Y 2O 3 0.3 0.76 0.1 – 0.3 9 − 11 >60
The Bayan Obo REE−Fe−Nb Ore Deposit
Mineralogy
The Bayan Obo REE‐Nb‐Fe deposit is situated in Inner Mongolia on the northern edge
of the North China Kraton, 135 km northwest of Baotou (110°E, 41°47′N). It is a giant
polymetallic rare earth element (REE)‐Fe‐Nb ore deposit of hydrothermal origin. It was
first discovered as a Fe deposit by Prof. Ding Daoheng in 1927. The estimated total
reserves of Fe (average grade Fe 35 wt%), Nb (average grade of 0.13 wt%) and REE
(average grade REO 6 wt%) are 1.5 Bt, 1 Mt and 48 Mt. It is the world’s largest known
REE ore deposit. The Bayan Obo mining area is shown in Figure 1.


Figure 1: Satellite image of Bayan Obo mining area (modified from google)

The ore composition in the Bayan Obo is very complex, 71 elements and 170 minerals
are found, one element could exist in several or more than ten different minerals,
mineral symbiosis relationship is close and complicated, and dissemination size is fine.
More than 90% rare earth elements in ore are existed in independent minerals, and
about 4% ~ 7% of REE dispersed in iron minerals and fluorite. Total 15 kinds of rare
earth minerals are found but the principal ones are bastnaesite [(Ce,La,Nd)(CO3)F] and
monazite [(Ce,La,Nd)PO4] with the ratio of 7:3 or 6:4, whereas magnetite and hematite
are the dominant Fe‐ore minerals. Meanwhile, several kinds of niobium minerals are
contained in the ore such as columbite (FeNb2O6), aeschynite
((Y,Ca,Fe)(Ti,Nb)2(O,OH)6) and fersmite ((Ca,Ce,Na)(Nb,Ta,Ti)2(O,OH,F)6) etc. The main
gangue minerals are quartz, apatite, dolomite and fluorite etc. These minerals consist
of both primary and secondary metamorphic minerals of the host rocks and the
epigenetic ore and gangue minerals introduced by hydrothermal solutions. The
minerals of the Bayan Obo ore deposit are listed in Table 3.
Most of the Bayan Obo minerals, particularly in the banded ore, are very fine to
extremely fine grained. Most Fe−REE−Nb minerals are closely associated with fine
grain sizes. The grain sizes of REE minerals are in the range of 0.01 – 0.074 mm and the
amount of less than 0.04 mm is accounted for 70% ~ 80%.


Table 3: principal minerals of the Bayan Obo deposit

Beneficiation techniques

Compared with domestic and abroad single bastnaesite ores, the Bayan Obo ores are
much more difficult to upgrade due to its similar physical and chemical properties with
iron minerals and gangue minerals associated closely, a lot of experimental studies by
many research institutes in China on the rational exploitation of the Bayan Obo REE
deposits had been conducted since the late 1950s and more than 20 beneficiation
techniques have been reported. A significant progress on REE minerals processing had
been made in the early 1990s and the mixed and separate REE concentrates with high‐
grade and high recovery were obtained from the iron ores.

The flowsheet of low intensity magnetic separation (LIMS) – high intensity magnetic
separation (HIMS) – flotation (Figure 2) was considered the most successful one to be
used in industrial plants which was developed by Changsha Metallurgical Research
Institute in 1990.
As shown in Figure 2, after crushing the ore is ground to 90 – 95% passing 0.074 mm.
By LIMS (rougher and cleaner) magnetite is recovered in the LIMS concentrate. The
tailings of LIMS is processed through HIMS rougher at the magnetic field strength of
1.4T to recover another Fe mineral hematite and most REE minerals. And hematite
and REE minerals contained in the HIMS rougher concentrate are separated by HIMS
cleaner at the magnetic field strength of 0.6T. Combined LIMS and HIMS concentrates
reported to the Fe reverse flotation to get the final Fe concentrate and the HIMS
cleaner tailings containing most REE minerals is processed by flotation to get REE
concentrates.
At the REE flotation circuit, the feed grade of REO is 9.78 −12% (HIMS cleaner tailings).
The flotation is performed at low alkaline condition (pH9) and the flotation reagents
contained naphthyl hydroxamic acid as the collector of REE minerals, sodium silicate
as the depressant of silicates and J10 as the frother. The feed solids is 35‐45% wt. One
stage of rougher flotation, plus one scavenger flotation and two cleaner flotation
stages produced an REO concentrate at a grade of 55% and a secondary REO
concentrate at a grade of 34% with the combined recovery of 72−75%. However, these
concentrates are the mixture of REE minerals mainly bastnaesite and monazite. The
separation of bastnaesite and monazite has been reached by further flotation using
phthalic acid or Benzoic acid as the collector of bastnaesite and alum as the depressant
of monazite.

Ore
REE Flotation feed
Crushing
REE Flotation Rougher
Grinding
REE Flotation Scavenger
Classification
REE Flotation Cleaner 1
LIMS Rougher
REE Flotation Cleaner 2 Tailings
LIMS Cleaner
HIMS Rougher
1.4 T
HIMS Cleaner REE Conc REE Conc
(REO>60%) (REO>30%)
Fe Reverse Flotation 0.6 T Tailings
REE Flotation feed
Fe Conc Tailings

Figure 2: beneficiation flowsheet of the Bayan Obo deposit ore

Different flotation reagents had been tested. Only a rare earth concentrate with grade
15%~20% and the low recovery could be gotten by using fatty acids, such as oxidized
paraffin and oleic acid, as the collector in rare earth mineral processing in earliest. C5
~ 9 hydroxamic acid by saponification was primarily used for rare earth mineral
flotation from 1970s and 1980s, then the cyclic alkyl hydroxamic acid synthesized by
naphthenic acid was used as collector because the synthesis materials of fatty acids C5
~ 9 in short supply. A miracle result of up to 60% REO flotation concentrate was
obtained by processing previous concentrate of REO 25% ~ 30% from gravity
separation concentrator when the major adjustments of the compatible depressants
were made. In 1976, the plant test was succeeded, and the commercial production
began in 1978. By application of the reagents system, the rare earth concentrate of
REO> 60%, flotation recoveries of 60% to 65% were obtained. Since then, high‐grade
commercial rare earth concentrates were produced.
Because the selectivity of cyclic alkyl hydroxamic acid and alkyl hydroxamic acids was
found poor the collector of H205, as the representative of aromatic hydroxamic acid,
was successfully developed in 1986 by using naphthalene as raw material through
sulfonated ‐ hydrolysis ‐ alkali fusion hydroxylation for preparation of intermediates
naphthol, and then acylation reaction and condensation system with hydroxylamine (‐
NH‐OH), to make naphthyl isobutyl group hydroxamic acid. The rare earth grade and
recovery could significantly be improved by using new generation of rare earth mineral
collector H205 with only water glass and the activator sodium fluoride in the flotation
process was not needed. After the 1990s, a hydroxamic acid with dual activating
group, the No.8 flotation oil, improved from H205, has been used in the Bayan Obo
rare earth plant as a collector.

By the end of 2012, the Bayan Obo rare earth flotation concentrator had been
developed into an annual output of 250,000 t of REE concentrates (contained 50%
REO).
The Sichuan Mianning REE Deposit
Mineralogy

The Sichuan Mianning REE Ore Deposit was found between 1985–1986. It is an alkaline
pegmatite carbonate type rare earth deposit. The industrial reserve is 1×106 t with the
average grade of REO 3.7%. It is rich in light REE and heavy REE as well such as
europium, yttrium etc. The chemical analysis results are shown in Table 4.

Bastnaesite is the main rare earth mineral and chevkinite and parisite are found in the
ore. Other associated minerals are barite (BaSO4), fluorite (CaF2), iron and manganese
minerals and small amount of galena. The ore is divided into granule and powder
types. The granule ore has coarse grain size usually >1 mm. The grain size of
bastnaesite is between 1 to 5 mm. The powder ore is the weathered product of original
ore and takes up about 20% of total ore weight. The grade is about REO 3−7% and the
grain size is 80% passing 325 mesh.
Table 4: Chemical analysis results of Mianning REE ore
Element Wt % Element Wt % Element Wt %
REO 3.70 F 5.50 Na2O 1.39
TFe 1.12 CaO 9.62 MnO 0.73
SiO2 31.00 MgO 1.10 BaO 21.97
Al2O3 4.17 S 5.33 K2O 1.31
FeO 0.43 P 0.24

Three types of flowsheets are industrially used: the gravity concentration flowsheet,
the magnetic – gravity concentration flowsheet and the gravity – flotation flowsheet.
The coexistence of massive and powdered black sludge, Fe‐Mn weathered amorphous
aggregates slime greatly influences the floatability of REE minerals.

Gravity separation
The ore was ground to 62% passing 200 mesh and hydro classified into the four size
fractions. The shaking tables were used to process the fractions separately. Three
different grade bastnaesite concentrates were obtained with the grades of 30%, 50%
and 60%. The overall recovery was 75%.
Magnetic –gravity separation
After grinding the ore was concentrated by LIMS and HIMS separation and the
magnetic concentrate with the grade of 5.64% was obtained. The recovery of magnetic
circuit was 74.2% and yield 42%. And then the magnetic concentrate was classified
into four size fractions and processed separately by shaking tables. The final
concentrate with the grade of REO 52.3% was obtained. The overall recovery of RE was
around 55%.

Gravity separation‐flotation
As shown in Figure 3 the ore was primarily ground to 50% passing 200 mesh and hydro
classified into four size fractions. The classified fractions were concentrated by the
shaking tables separately. The grade of overall gravity concentrate was REO 30% with
the RE recovery of 74.5%. The gravity concentrate was reground to 70% passing 200
mesh for flotation. C5−9 hydroximic acid (H205) and phthalate in the ratio of 1:1,
sodium carbonate, sodium silicate were used as the flotation reagents at pH8‐9. By
one rougher, one cleaner and one scavenger flotation the concentrate with the grade
of REO 50−60% was acquired at the RE recovery of 50−60%.

Ore
Grinding
Hydro classification
Shaking table Shaking table Shaking table Shaking table
Grinding Tailings
C5−9 hydroximic acid (H205) and
phthalate (1:1), Na2CO3,
Rougher Flotation
Na2SiO3, pH 8‐9
Cleaner Flotation Scavenger Flotation
Tailings
Concentrate (REO 50‐60%)

Figure 3: Gravity separation‐flotation process at Mianning REE min
The Shandong Weishan REE Deposit
Mineralogy
The ore deposit was first found in 1958 and the exploration was finished in 1975. The
reserve of RE was about 2.55×106 t and the average geological grade was 3.13%.
It is a quartz−barite−carbonate type RE ore deposit. The main RE minerals are
bastnaesite and parisite and the main associated minerals are barite, calcite, quartz
and fluorite etc. The grain sizes of RE minerals are coarse in the range of 0.04 −0.5 mm.
The chemical compositions are shown in Table 5.

Table 5: Chemical compositions of Shandong Weishan REE Ore
Element Wt % Element Wt % Element Wt %
REO 3.71 F 0.698 Na2O 3.53
TFe 2.81 CaO 1.18 Th 0.002
SiO2 47.92 MgO 1.18 BaO 11.99
Al2O3 22.48 S 2.1 K2O 1.85


The Weishan REE flotation plant was built in 1982. The ore was ground to 65−75%
passing 200 mesh and REE minerals were floated with once rougher, three scavengers
and three cleaners. In 1980s oleic acid and kerosene were used as the collectors of RE
minerals at acidic condition of pH5 using sulfuric acid. After 1991 with the ore grade
dropped to 3‐4%, a specific collector with formula C6H4OHCONHOH was used with the
addition of sodium silicate and frother L101. The flotation was run at the weak alkaline
condition of pH8−8.5. The RE concentrate with the grade of REO >60% at the recovery
of 60−70% was obtained. Meanwhile, a secondary RE concentrate with the grade of
REO 32% at the recovery of 10−15% was acquired. According to market demand, REE
concentrate with the grade of 45~50% REO at the recovery of 80~85% has been
produced and barite has also been recovered by flotation from REE flotation tailings.

The Weathered Crust Elution‐deposited REE Ore (The Ion
Adsorption REE Deposit)

Mineralogy

China’s weathered crust elution−deposit rare earth ore or called ion adsorp on rare
earth ore is the unique REE deposit in the world. It was found in 1969 in Jianxi Province
containing two types of this deposit, light REE type and heavy REE type, and was also
found in other provinces Fujian, Hunan, Guandong and Guanxi. The known reserve of
weathered crust elution−deposit rare earth ore in China is over 1 ×106t (REO) and the
heavy RE reserve in China occupies over 80% of the world total heavy REE reserve.
Now China produces about 1×104 t concentrate (REO >60%) from this ore annually.
The deposit was considered to form by many years weathering of granite and effusive
rocks. It has the characteristics of shallow ore body coverage, soft ore and very fine
grain size. The 80% −90% REE in the state of positive hydrated ions are adsorbed on
the surface of clay minerals such as kaolinite, halloysite and illite. The ore is relatively
low‐grade, generally only 0.05% to 0.5% REO, with high heavy REE.

Extraction Techniques
Because the REE in the ore mainly occur in the state of hydrated ions adsorbing on the
surface of clay minerals they can’t be effectively concentrated by conventionally
physical methods but extracted by the ion−exchange method. The adsorbed REE ions
are not dissolved in water or ethyl alcohol but they are dissolved in the electrolytes
solutions of NaCl, (NH4)2SO4 and NH4Cl by ion–exchange. The percentages of REE in
different phases are shown in Table 6. It is indicated that over 80% REE are ion
exchangeable.

Table 6: REE concentration in different phases (%, mass fraction)
Rare Earth Aqueous Ion Colloidal Mineral
Ore Soluble Exchangeable Sediment
−3
LN 6.15×10 80.62 5.30 13.36
−3
XF 8.14×10 83.58 3.23 12.63
−3
NH 1.47×10 84.94 4.98 9.09

In 1970s as a major extraction method NaCl was used as the leaching solvent to
dissolve REE by ion exchange. But high solvent concentration and long leaching time
were required. And the solvent consumption and processing cost were high.
Meanwhile, the grade of RE concentrate can only reach 70% REO. Since 1980s
(NH4)2SO4 has been used to replace NaCl as the solvent. A high grade (REO 92%) of RE
concentrate has been achieved.

The process is usually run in a cement bath with the volume of about 10−20 m3. When
the bath is filled with the ore to the height of about 1.5 m the solvent of (NH4)2SO4
solution with the concentration of 1−4% is poured through the ore for leaching. The
pregnant solvent is obtained at the bottom of the bath. Oxalic acid is used as the
precipitant to get mid product of oxalic acid−rare earths and further processed by
burning to get final rare earths product (>REO 92%). Ammonium bicarbonate also has
been used as a competitive precipitant. The flowsheet of the extraction technique is
shown in Figure 4.
Ore
Leaching
Residue
RE bearing solution
oxalic acid precipitation (oxide RE: oxalic acid = 1:2
Filtration
Lquid RE oxalic acid salt
Nutralization Dried 105°C
Filtration Burned 850°C
Residue RE Oxide (REO˃92%)

Figure 4: REE extraction technique for weathered crust elution−deposit REE ore
The in‐situ leaching process also called the solution mining was studied from 1980s to
resolve the ecological environmental problems in the exploitation by the bath leaching
technology. The wells are drilled in the natural orebody and the leaching regent is
injected. REE ions were selectively leached. Collected leachate is precipitated by the
solution of oxalic acid or ammonium bicarbonate to obtain rare earth oxide products.
The water is recycled. It was reported that two key technical problems must be solved
during the in‐situ leaching process. One is how to avoid the leaching reagent solution
spreading around the wells to cause the environmental pollution and the other is how
to recover the leachate so as optimize the leaching rate. The technology has been
utilized in Wenfeng, Jiangxi province. More than 200 t of REE are extracted annually at
the REE recovery of 70%.
Conclusions
China possesses the largest resource of rare earth elements (REE) in the world and
has dominated the world REE production since 1980s.The resources of rare REE are
mainly divided into four deposits which consist of 95% total REE resources of China.
They are the Bayan Obo REE−Fe−Nb Ore Deposit, the Sichuan Mianning REE Ore
Deposit, the Shandong Weishan REE Ore Deposit and the Weathered Crust
Elution−deposited Rare Earth Ore.
The Bayan Obo REE‐Nb‐Fe deposit situated in Inner Mongolia on the northern edge of
the North China is a giant polymetallic rare earth element (REE)‐Fe‐Nb ore deposit of
hydrothermal origin. It is the world’s largest known REE ore deposit with REO 48 Mt
with average grade REO 6 wt%. The ore composition in the Bayan Obo was very
complex and the principal REE minerals are bastnaesite and monazite. Most
Fe−REE−Nb minerals are closely associated with fine grain sizes. The flowsheet of low
intensity magnetic separation (LIMS) – high intensity magnetic separation (HIMS) –
flotation was considered the most successful one to be used in industrial plants which
was developed in 1990s.

The Sichuan Mianning REE Ore Deposit was found between 1985 – 1986. It is a alkaline
pegmatite carbonate type rare earth deposit. The industrial reserve is 1×106 t with the
average grade of REO 3.7%. Bastnaesite is the main rare earth mineral. Three
processes have been used industrially, magnetic – gravity separation process, gravity
separation process and gravity separation – flotation process.

The Shandong Weishan REE Deposit was first found in 1958 and the exploration was
finished in 1975. The reserve of RE was about 2.55×106 t and the average geological
grade of 3.13%. It is a quartz−barite−carbonate type RE ore deposit. The main RE
minerals are bastnaesite and parasite. The flotation plant was built in1982.

Chinese weathered crust elution−deposit rare earth ore or called ion adsorp on rare
earth ore is the unique REE deposit in the world. The 80% −90% REE in the state of
positive hydrated ions are adsorbed on the surface of clay minerals which can’t be
effectively concentrated by conventionally physical methods and usually are extracted
by the ion−exchange method using NaCl, (NH4)2SO4 or NH4Cl as leaching solvents. The
process is usually run in a cement bath and the in‐situ leaching process has been also
used.

References
1. U.S. Geological Survey, Mineral Commodity Summaries, February 2014.
2. Ruan Chi and Jun Tian, Comments on Weathered Crust Elution‐deposited Rare Earth Ores, Journal
of The Chinese Rare Earth Society, Vol.25, No.6, Dec.2007. p.641−649.
3. Ruan Chi and Jun Tian, Weathered Crust Elution‐deposited Rare Earth Ores, Nova Science
Publishers, Inc., New York 2008.
4. Yu Yongfu, Development on beneficiation techniques of Chinese REE deposit ores, West China
Exploration Engineering, series No.63, Mar.2000. p.1−4.
5. Jörg Neßler, Giant Bayan Obo REE‐Fe‐Nb Ore Deposit of Inner Mongolia, China, January 2007
6. Jun Ren et al., Selective flotation of bastnaesite from monazite in rare earth concentrates using
potassium alum as depressant, Int. J. Miner. Process. Vol.59, 2000. 237–245.
7. Cao Yongdan, Cao Zhao, Li Jie and Qu Qilong et al, Current Study Situation and Development on
Flotation of rare earth in Baiyunebo Mine, Mining Machinery, Vol.41, No.1,2013, 93‐96.
8. Xiao Shenming, Research on Method of Rare Earth Ore Processing, World Non‐ferrous Metals,
No.5, 2011, 48‐49.
9. Li Yong, Zuo jicheng, Liu Yanhui, Application and Research of Hydroximic Acid as Collectors on
Flotation of Rare Earth Ores, Non‐ferrous Mining and Metallurgy, Vol.23, No.3, 2007, 30‐33.
10. Luo Jiake, Ren Jun, Tang Fangqong, Zhou Gaoyun, Development on Flotation Reagents of Rare
Earth Minerals in China, Journal of the Chinese Rare Earth Society, Vol.20, No.5, 2002, 385‐391.
11. Yu Yongfu, Dressing Technology of REO Ore and Its Development in China, Journal of China
University of Mining & Technology, Vol.30, No.6, 537‐542.
12. Yu Yongfu, Zhou Chaoying, Advance in Baotou’s Rare Earth Ore Beneficiation Technology, Metal
Mine, Series No.281, 1999, 18‐22.
13. Ci Ru‐an, Wang Dianzuo, A new Process of Ion Type REO, Non‐ferrous Mine, No.5, 1990, 47‐51.
14. Personal Letters with Technical Staff of Baotou Iron and Steel Group Corporation, March 2014.

ECONOMICS OF RARE EARTHS: THE BALANCE PROBLEM
Koen BINNEMANS1
1
Department of Chemistry, KU Leuven – University of Leuven, Celestijnenlaan 200F, P.O. Box
2404, B‐3001 Heverlee, Belgium
E‐mail: Koen.Binnemans@chem.kuleuven.be
Abstract
The balance between the demand by the economic markets and the natural abundance
of the REEs in ores is a major problem for manufacturers of these elements. This is the
so‐called Balance Problem (or Balancing Problem). The ideal situation is a perfect
match between the demand and production of REEs, so that there are no surpluses of
any of the REEs. The Balance Problem implicates that the rare earth industry has to
find new applications for REEs that are available in excess, or to search for substitutions
for REEs that have limited availability and that are high in demand. Different solutions
are proposed to solve the Balance Problem: diversification of REE resources, recycling,
substitution, reduced use and new high‐volume applications. It is shown that the issue
of thorium in REE ores is related to the Balance Problem.

Introduction
The global annual production of rare earths or rare‐earth elements (REEs) is typically
expressed in tonnes of rare‐earth oxides (REOs). At present, the estimated yearly total
production of REOs is about 125,000 tonnes, although accurate production numbers
are very difficult, if not impossible, to find. However, these total production numbers
do not tell anything about the availability of the individual REEs. First of all, not the
total amount of REE ores mined is of importance, but rather the total amount of REE
concentrate separated in the different individual REEs. For some applications, such as
mischmetal, no separated REEs are required. Secondly, the REEs are not present in
equal amounts in the REE ores, due to differences in the natural abundances of these
elements. The general trend in the natural abundance of the rare earths is that the
elements become scarcer with increasing atomic number Z; as the abundances
decrease over the lanthanide series, the heavy rare‐earth elements (HREEs) are much
less abundant than the light rare‐earth elements (LREEs).1 Moreover, elements with
an even atomic number Z are more abundant than elements with an odd atomic
number (Oddo‐Harkins rule). For example, cerium (Z = 58) is more abundant with
respect to its neighbours lanthanum (Z = 57) and praseodymium (Z = 59). Cerium is the
dominating REE in LREE‐rich ores (bastnäsite and monazite), while yttrium is the main
REE in HREE‐rich ores (xenotime and ion‐adsorption ores).

The availability of the individual REEs is of concern for applications that require the use
of pure REEs, such as lamp phosphors and permanent magnets. If the REEs that are
required for high‐volume applications have a low natural abundance, the minimum
quantity of REE concentrate that needs to be processed and separated is that quantity
that will produce at least the amount of REEs required for the critical applications. As
a consequence, some REEs will be produced in larger quantities than required by the
REE market and, subsequently, these need to be stockpiled. The balance between the
demand by the economic markets and the natural abundances of the REEs in ores is a
major problem for manufacturers of these elements. This is the so‐called Balance
Problem (or Balancing Problem).2,3 The ideal situation is a perfect match between the
demand and production of REEs, so that there are no surpluses of any of the REEs. This
would result in the lowest market price for any of the REEs, because the production
costs are shared by all the elements. Unfortunately, a market in balance is very difficult
to achieve, because of changes in demand due to technological evolutions in
applications. The result is a sometimes very high demand of a REE that is a minor
constituent in the ore (such as dysprosium), while the demand of the major
constituent (such as yttrium) is much lower. The size of the REE market is different in
value and volume. For instance, lamp phosphors are a very important application of
REEs in terms of value (1/3 of the REE market), but they are a rather modest
application in terms of volume (about 7%). The Balance Problem is related to the
volumes of the individual REEs consumed.

For the REE industry the Balance Problem is a more important issue than the
availability of REE resources, not to say that it is the most important issue facing the
REE industry. This paper gives an in‐depth analysis of the Balance Problem and offers
several possible solutions.

History of the Balance Problem
The balance problem became an issue when purified REEs were started to be used in
applications. Older applications of REEs consumed mixtures of REEs and no separation
into pure single elements was required. A good example is mischmetal, which is an
alloy of LREEs, with the REEs in the same atomic ratios as they occur in the REE ore
minerals. Also for other older applications such as polishing powders or fluid cracking
catalysts no pure REEs have to be used. All REEs that were produced could be
consumed.

The first application of pure REEs was the use of europium as active component of the
red cathodoluminescent phosphors in colour television screens (first YVO4:Eu3+, later
Y2O2S:Eu3+). In the mid‐1960s and the early 1970s, europium was the most critical REE
because it has a low natural abundance and it was high in demand for the production
of these red phosphors. In that period, nearly all of the global supply of europium was
produced from the bastnäsite ore of the Mountain Pass mine in California. In fact, the
Mountain Pass mine was in that period mainly operated for the production of
europium. The bastnäsite of the Mountain Pas mine is relatively rich in europium (0.1
%) compared to monazite (< 0.05%). Nevertheless, the low concentrations of
europium implied that large surpluses of the LREEs were produced and needed to be
stockpiled: to obtain 1 tonne of Eu2O3 from bastnäsite, one had to excavate an amount
of REE ores that contained 300 tonnes of La2O3, 450 tonnes of CeO2, 38 tonnes of
Pr6O11, 118 tonnes of Nd2O3, 7.3 tonnes of Sm2O3, 1.4 tonnes of Gd2O3 and 0.9 tonnes
of Y2O3.4 It should be mentioned that it is possible to selectively remove europium
from mixtures of REEs by reduction of europium(III) to europium(II), for instance by
chemical reduction with zinc amalgam or by electrochemical reduction.

In the 1970s and 1980s, samarium was the most critical REE because of its use in
samarium‐cobalt permanent magnets. At that time, the production of samarium‐
cobalt magnets was limited by the availability of Sm2O3 on the market. At present, the
share of samarium‐cobalt magnets in the permanent magnet market is less than 2%,
and an excess of samarium is being produced. Before 1985, there were no industrial
applications for dysprosium, and neodymium metal was not produced on an industrial
scale. At that time, it could not be predicted that less than 30 years later, neodymium
and dysprosium would be high in demand for neodymium‐iron‐boron magnets. With
respect to holmium, thulium, ytterbium and lutetium no high‐volume applications
currently exist.

Due to the Balance Problem, there has also been a shift in the types of REE ores used.
Until the 1960s, monazite was the main ore (monazite era). However, for a long time
the REEs were just a by‐product of the production of thorium which was used for the
manufacturing of incandescent gas mantles. In the 1960s, bastnäsite became the most
important REE ore, with the Mountain Pass mine in California as the main production
site (Mountain Pass era). Bastnäsite was ideal to cover the increasing needs of
lanthanum and cerium, and as well as to fulfil the demand of europium. Gradually,
monazite was favoured again because of its higher content of HREEs and yttrium. The
steadily increasing demand of yttrium (mainly for its use in lamp phosphors and yttria‐
stabilised zirconia) implied that also the yttrium‐rich phosphate xenotime was
processed as a REE ore. From the end of the 1980s on, the HREE‐rich ion‐adsorption
ores (laterite clays) from southern China became the most important resource for
yttrium and the HREEs.

Current market situation and evolution in the future
At present the LREE market is driven by the demand for neodymium for neodymium‐
iron‐boron (NdFeB) magnets.5 For instance, about 25000 tonnes of neodymium were
required for the production of magnets in 2011. This means that sufficient quantities
of REE ores had to be mined to produce at least 25000 tonnes of neodymium. Since
the natural abundance of neodymium is relatively low in the LREE ores, cerium,
praseodymium and samarium are produced in excess, and have to be stockpiled. The
lanthanum market is in balance, thanks to its use in nickel metal hydride batteries and
optical glasses. A few years ago, it was predicted that there would be a sharp decline
in the use of nickel metal hydride batteries due to the increased use of lithium‐ion
batteries. However, there are still many safety issues related with these lithium‐ion
batteries; overheating of the organic electrolytes can cause fires. For these reasons, it
is likely that nickel metal hydride batteries will continue to be used for several years
to come in high‐power applications such as electric vehicles. The HREEs are produced
in much smaller quantities than the LREEs, so that the Balance Problem is a bigger issue
for the LREE market than for the HREE market. At present, the HREE market is driven
by the demand for dysprosium, which is used to increase to high‐temperature
performance and resistance to demagnetisation (i.e. the intrinsic coercivity) of NdFeB
magnets. About 1600 tonnes of dysprosium were consumed in 2011. The supply
equals the demand for europium, yttrium and erbium. There is a shortage of terbium,
but this problem can still be solved by the use of stockpiles. Gadolinium, holmium,
thulium, ytterbium and lutetium are produced in excess and are stockpiled. This
stockpiling is not a major issue in terms of volumes, but it causes a loss of potential
profit for the REE producers.

The REE market is a rapidly changing market. New applications as well as the sudden
disappearance of well‐established applications could bring the REE market out of
balance. Although the present REE market is driven by the demand for neodymium
and dysprosium, this could rapidly change, as shown by the historical evolutions in the
REE markets. Although there is currently an oversupply of gadolinium, this global
excess could turn into a shortage if magnetic refrigerators would be mass‐produced.
However, it must be admitted that present research activities in the field of magnetic
refrigerators are focusing on the development of REE‐free materials with a large
magnetocaloric effect. Given the fact that kilogramme quantities of gadolinium are
required for the construction of one magnetic refrigerator, it is not realistic to assume
that such devices would be mass‐produced; there is simply not enough gadolinium
available on the world to achieve this goal.

It is very likely that within the next years a dramatic change will occur in the lamp
phosphor market that will cause an imbalance of the HREE market. At present,
europium, terbium and yttrium are considered as three of the five most critical rare
earths (the other two being neodymium and dysprosium) because of their use in the
red lamp phosphor Y2O3:Eu3+ (YOX), the green lamp phosphors LaPO4:Ce3+,Tb3+ (LAP),
(Gd,Mg)B5O12:Ce3+,Tb3+ (CBT), (Ce,Tb)MgAl11O19 (CAT) and the blue phosphor
BaMgAl10O17:Eu2+ (BAM) in fluorescent lamps and compact fluorescent lamps (energy‐
saving lamps).6 Nevertheless, the fluorescent lamps market is rapidly shrinking due to
the increasing success of the light‐emitting diodes (LEDs). Many types of LEDs do not
contain rare earths at all; they produce white light by a combination of red, green and
blue LEDs standing in close proximity to each other, similarly to the red‐green‐blue pixels
in old‐fashioned cathode‐ray tubes. This change from fluorescent lamps to LEDs will
make the conventional lamp phosphors largely obsolete, although they could still find
use in some niche applications. As a consequence, one may expect an oversupply of
europium. The lamp phosphor Y2O3:Eu3+ is by far the most important application of
europium. Smaller quantities of luminescent europium compounds are used in safety
markers, for instance in the luminescent inks of EURO banknotes. Considering that no
other high‐volume applications of europium are close to market introduction,
europium will no longer be a critical element in the years to come. An oversupply of
europium is less of an issue than an oversupply of cerium, since the global production
volumes of europium are much smaller than those of cerium. For terbium, it is much
easier to find alternative applications. Terbium can substitute for dysprosium in NdFeB
permanent magnets (vide infra). Also for yttrium other applications than lamp
phosphors are possible, since yttrium is a very useful element for the preparation of
high‐tech ceramic materials, for instance, yttria‐stabilised zirconia.

How to solve the Balance Problem?
Increasing the overall REE production to meet the highest demand of any REE and to
stockpile the other REEs with lower demand, seems to be an obvious solution for the
Balance Problem. However, this will increase the overall price of the REEs, due to the
extra costs for separating the REE mixtures and stockpiling the REEs produced in excess
of the demand. Adjusting the overall REE production to optimise the REE producer’s
operational margins will create surpluses of some REEs and shortages of other REEs.
Shortages of a minor constituent lead to dramatic price increases of this REE due to its
(very) high price inelasticity. Preferentially, the REE market is driven by the demand
for elements that are very abundant (cerium and lanthanum), since this will create less
problems with stockpiling of the elements that are available in excess. The unbalanced
use of a single REE in a high volume application has to be avoided, especially if this
element has a low natural abundance, because this will cause a serious imbalance of
the market. There are different solutions that can help to alleviate the Balance
Problem. These will be discussed each in some detail. Since the REE markets are rapidly
changing, it is extremely difficult to extrapolate the future demand of individual REEs
on the basis of data for the present demand of these elements. Long‐term predictions
(for 25 or more years) should be interpreted with caution.7

Solution 1: Diversification of REE resources
At present, the global REE supply is produced from a limited number of ore types:
bastnäsite, monazite, xenotime and ion‐adsorption clays. Bastnäsite and monazite are
rich in LREEs, whereas xenotime and ion‐adsorption ores are rich in HREEs and yttrium.
The composition of these ores can show some variation from deposit to deposit, but
these variations are rather limited. Bastnäsite shows a very limited variation in
composition. The search for new REE deposits gives also access to other types of ores
than the conventional ones.8 Examples include eudialite, euxenite, parisite, synchisite,
churchite, gadolinite, fergusonite, loparite and steenstrupine. Phosphate rocks (mainly
apatite) used for the production of phosphoric acid and phosphate fertilisers are an
important potential REE resource, and was already used in the paste for the recovery
of REEs.9 Even though phosphate rocks contain only small REE concentrations, the
excavated volumes are enormous. Rare earths can also be extracted from industrial
waste streams such as bauxite residue (red mud), a by‐product of the aluminium
production and phosphogypsum, a by‐product of the phosphoric acid production from
phosphate rocks.10 Bauxite residue is quite rich in scandium.11 By combining the REE
concentrates obtained from different REE resources, it is possible to compose a mixed
concentrate with a composition that is reflecting much better the needs of the
individual REEs by the market than in the case when only one single or just a few types
of REE ores are used. By combining the REE concentration from different ores, it is also
possible to create a mixed REE feed solution of constant composition. This is beneficial
for the optimisation of solvent extraction processes for the separation of mixtures of
REEs.

Solution 2: Recycling
Recycling and closing the materials loop are very important activities for bringing the
REE markets in balance.12,13 Recycling of neodymium and dysprosium from end‐of‐life
NdFeB magnets implicates that less primary REE ores have to be excavated to meet
the global demand of neodymium and dysprosium. Less mining of REE ores implies less
overproduction of cerium and samarium. Recycling of europium, terbium and yttrium
from lamp phosphors also helps to keep the HREE market into balance.14 Thus, by
recycling REEs that are essential for high‐volume applications, an oversupply of REEs
that are less in demand can be avoided. The REE concentrate obtained by recovery of
REEs from end‐of‐life consumer goods can be used to adjust the composition of the
feed of solvent extraction batteries. Recycling is industrially relevant for the most
valuable applications: permanent magnets, lamp phosphors and NiMH batteries.
However, several challenging remain to be dissolved, especially the recovery of
permanent magnets from shredded waste.15,16 To estimate the possible impact of
recycling on the REE markets, it is of importance to get an idea of the global inventory
of the different REEs in use, of the life cycle of REEs in applications and of the life span
of REE‐containing devices. Recently, several research groups started to address this
issue.7,17-22 Recycling of ceria (from polishing powders and car exhaust catalysts) could
be recommended from a sustainable point of view, but it will not help to solve the
Balance Problem due to the oversupply of cerium on the market.23

Solution 3: Substitution
Substitution includes the replacement of one element by another in a given
application. The different REEs have very different electronic and magnetic properties
so that one REE cannot simply replaced by another REE in an application. For instance,
europium is the only REE that can be used in red lamp phosphors. In fact, europium is
unique for its line‐emission in the red spectral region. In this case, substitution of
europium by another REE is not an option. However, in other cases a more critical REE
can be replaced by a less critical one or, even better, by a non‐critical REE. The NdFeB
magnets are a good example of an application where substitution can play a role.24
Part of the neodymium (up to 25%) can be replaced by praseodymium without
significantly changing the magnetic properties of the material. This is a very good
opportunity to consume a large part of the supply of praseodymium (which is less
abundant than neodymium, but has also less specific applications) and to help bringing
the LREE market in balance. This also means that there is no need to use very high
purity neodymium in NdFeB magnets; didymium (the old name used for the mixture
of neodymium and praseodymium) can be used as well. However, it is not possible to
replace part of the neodymium in NdFeB magnets by samarium. Part of the
dysprosium in NdFeB magnets can be replaced by terbium. This is not often done at
present because terbium is a very critical element due to its use in the green lamp
phosphors. However, it can be anticipated that more terbium will become available
for use in NdFeB magnets if the fluorescent lamps will increasingly be replaced by LEDs
(vide supra).

What about the replacement of neodymium‐iron‐boron magnets by samarium‐cobalt
magnets? As described above, SmCo magnets are in use for a longer time than the
NdFeB magnets and for a long time samarium was the most critical REE element in so
far that the production of SmCo magnets was limited by shortages in supply of
samarium, with very high samarium prices as a result. After 1985, SmCo magnets have
been rapidly replaced by NdFeB magnets. NdFeB magnets have a higher energy
density than SmCo magnets, but, more importantly, NdFeB magnets are much cheaper
than SmCo magnets. More than 70% of the mass of NdFeB magnets consists of cheap
iron, while thethe natural abundance of neodymium is much higher than that of
samarium. Furthermore, the SmCo magnets contain also the rather expensive cobalt.
However, SmCo magnets have some advantages compared to NdFeB magnets: they
are much more resistant to demagnetisation at higher temperatures and they are
much more corrosion resistant (SmCo magnets do not need to be coated by an
anticorrosion layer, in contrast to NdFeB magnets). Therefore, SmCo magnets are used
in high temperature applications and in applications where corrosion can be an issue.
Typical markets for SmCo magnets are the aircraft industry and the military industry.
Still, the market of SmCo magnets is very small compared to the huge market of NdFeB
magnets. As indicated above, the share of SmCo magnets in the permanent magnet
market is less than 2%. Given the criticality of neodymium/dysprosium and the
oversupply of samarium, why cannot more SmCo magnets by produced? First of all,
SmCo magnets are more expensive than NdFeB magnets due to the large quantities of
cobalt they contain. Cobalt is not a cheap element and it has suffered from also its own
supply risks in the past. Cobalt is also used in lithium‐ion‐batteries, in the form of
lithium cobalt oxide (LiCoO2) in cathode materials. Secondly, the safety regulations for
working with samarium in an industrial environment have became more strict.
Samarium has two naturally occurring radioisotopes, which are long‐lived ‐emitters:
147
Sm (t½ = 1.06  1011 y; natural abundance = 14.99%) and 148Sm (t½ = 7  1015 y;
natural abundance = 11.24%). Working with bulk samples such as in SmCo magnets or
with samarium‐containing solutions is not a problem, but working with Sm2O3 powder
(which is a precursor for Sm alloys) is restricted by safety regulations due to the
dangers associated with possible inhalation of powders of ‐emitters.

Substitution can also mean replacement of REEs by non‐REEs, even though this can
lead to products with poor, yet still acceptable, performance. Examples include fluid
cracking catalysts based on zeolites without REEs. NdFeB or SmCo magnets can be
replaced by ferrite or alnico magnets, but these magnets have much lower energy
densities. Therefore, REE magnets are unlikely to be replaced in applications where
miniaturisation is of importance. Major research efforts are directed to the
development of cerium‐based magnetic materials. Until 1975, mischmetal was often
used for nodularising graphite in cast iron and for desulfurising steel, but these
metallurgical applications became largely obsolete due to better production methods
and replacement of mischmetal by calcium or magnesium. Ceria in polishing materials
could be replaced by alumina or iron oxide, but this is at present not a feasible option
since there is an oversupply of cerium. NiMH batteries, containing large quantities of
LREEs and especially of lanthanum, could be replaced by lithium‐ion batteries.
Likewise, the LaNi5 alloy in NiMH batteries could be replaced by an REE‐free titanium‐
iron alloy. Sometimes, substitution can be more drastic by simply replacing one
element by another one, and a totally new technology is introduced on the market,
making the older technology based on REEs obsolete. Until about 10 to 15 years ago,
europium was an important component of red cathodoluminescent phosphors used
in cathode‐ray tubes (CRTs) for colour television screens and computer monitors. The
CRTs screens were replaced by plasma displays (still containing REE phosphors) and
LCD screens (using REEs in the phosphors of the backlight fluorescent lamps), but
modern OLED screens do not contain REEs at all. A similar evolution is taking place by
the replacement of fluorescent lamps by LEDs.

Solution 4: Reduced use
By clever engineering, it is possible to reduce the consumption of critical REEs in a
given application, without compromising the performance of the REE‐containing
devices. The NdFeB magnets, which contain dysprosium to protect the magnetic
material against demagnetisation at higher temperatures, are a good example. By
grain boundary diffusion, the dysprosium is concentrated near the grain boundaries of
a sintered NdFeB magnet. As a consequence the required total dysprosium
concentration can be decreased by more than 50%.

It is important not to waste critical REEs in applications that do not required purified
REEs. In the past, no efforts were made to remove neodymium from the concentrate
that was used to prepare mischmetal, although the presence of neodymium was not
essential for applications of mischmetal. At present, neodymium is too valuable for
magnet applications and is removed prior to mischmetal production.

Solution 5: New high‐volume applications
Compared to the other REEs, lanthanum and especially cerium are very abundant and
cheap. For western REE mining companies, the lanthanum and cerium content of their
ores often has a negative effect on the REE basket value and, subsequently, these
elements are a nuisance for these companies. Therefore, it is important to develop
new high‐volume applications for these elements. Research in this area must be
encouraged. An example of an emerging high‐volume low‐tech application is the use
of chlorides of lanthanum, cerium or mixtures thereof in water purification technology
for removal of dissolved phosphate ions from water. The trivalent REE ions have a high
affinity of phosphate ions, resulting in the formation of poorly soluble and readily
filterable REE phosphates. Cerium is being investigated as active component in redox
flow batteries, for instance the zinc‐cerium redox flow battery. Other applications of
cerium include its use for the destructive total oxidation of toxic organic compounds
and as a redox mediator in organic electrosynthesis.

Thorium
The thorium issue is related to the Balance Problem. Nearly all REE ores contain the
radioactive elements thorium and uranium. Although the uranium content of typical
REE ores is low, the thorium content of some REE ores can be high. For instance,
monazite can contain more than 15 wt% of thorium. Natural thorium is a mono‐
isotopic element: all but some traces of natural thorium consist of the isotope 232Th,
an ‐emitter with a half‐life of 14.05 billion years. Due to the safety regulations
associated with the handling of naturally occurring radioactive materials, the thorium
content of REE ores is a major issue for REE producers. At present, there are no large
scale applications of thorium. Furthermore, thorium is considered as radioactive waste
and its disposal can be very expensive. The thorium content of monazite is the reason
why Rhône‐Poulenc (now Solvay) stopped processing monazite at the REE production
plant of La Rochelle in France in 1994. The thorium content of the REE ore of Bayan
Obo is causing severe environmental issues in Baotou (PR China). As long as thorium
is considered as radioactive waste, it will increase the production costs of the rare
earths. The costs of thorium waste disposal need to be shared by the different REE
applications. However, thorium could be used as a nuclear fuel instead of uranium.
Thorium can be used in conventional light‐water nuclear reactors, although in that
case ThO2 has to be mixed with plutonium oxide. Canadian CANDU reactors can burn
thorium as well. A much better approach is to use specially designed nuclear reactors
such as the molten salt thorium reactor, which do not need the use of plutonium.
Interestingly, the REEs were once by‐products of the thorium production. Thorium was
used at the end of the 19th century and the first half of the 20th century to prepare
incandescent gas mantles (99% ThO2 + 1% CeO2) and the REEs were considered more
or less as a waste product, although there were some applications for cerium and
mischmetal. Contrastingly, at present thorium is considered as a dangerous waste
product. However, the transition to a thorium‐based nuclear fuel cycle could lead to
zero waste valorisation of thorium‐containing REE ores, so that both REEs and thorium
can be considered as valuable resources. A positive side effect is that no special
thorium mines need to be opened, because it can be expected that the quantity of
thorium of processing monazite and other thorium‐containing REE minerals is
sufficient to cover the global demand of thorium.

Conclusions
The Balance Problem is an intrinsic problem of REE markets that is caused by the
imbalance between the demand of individual REEs and their natural abundance in REE
ores. Keeping the REE markets in balance is of strategic importance, not only to secure
the supply of all REEs required for technological and other applications, but also to
avoid dramatic price shocks for critical REEs. Different solutions have been proposed
to solve the Balance Problem: diversification of REE resources, recycling, substitution,
reduced use and new high‐volume applications. There is no silver bullet solution for
this problem. However, by combining different strategies the Balance Problem can be
mitigated. The presence of thorium in most of the REE ores is related to Balance
Problem: thorium is now considered as an expensive radioactive waste, but this waste
could be turned into a resource by using thorium in a thorium‐based nuclear fuel cycle.
At present there is too much focus on the figures with respect to the total REE ore
reserves and/or the total REOs production numbers. It must be realised that most of
the current high‐tech applications of REEs require the use of purified individual REEs
rather than mixtures of REEs. Therefore, It cannot be stressed enough how important
it is to effectively deal with the Balance Problem.

Acknowledgments
The author thanks the KU Leuven for financial support (projects GOA/13/008 and IOF‐
KP RARE3).

References
1. Haxel, G. B.; Hedrick, J. B.; Orris, G. J. Rare Earth Elements—Critical Resources for High
Technology: U.S. Geological Survey Fact Sheet 087‐02, available online at:
http://geopubs.wr.usgs.gov/fact‐sheet/fs087‐02/; 2002.
2. P. Falconnet "The Economics of Rare‐Earths", J. Less‐Common Met., 111 (1‐2), 9‐15, 1985.
3. K. Binnemans, P. T. Jones, K. Van Acker, B. Blanpain, B. Mishra and D. Apelian "Rare‐Earth
Economics: The Balance Problem", JOM, 65 (7), 846‐848, 2013.
4. P. Falconnet "Industrial strategy and ecomomics of rare earths", Proceedings of: First
Workshop on the Basic and Applied Aspects of Rare Earths (Venice, Italy, 26‐27th May, 1988),
19‐31, 1989.
5. Goonan, T. G. Rare earth elements—End use and recyclability: U.S. Geological Survey
Scientific Investigations Report 2011‐5094, 15 p., available at
http://pubs.usgs.gov/sir/2011/5094/; 2011.
6. C. R. Ronda "Phosphors for Lamps and Displays ‐ An Applicational View", J. Alloys Compd.,
225 (1‐2), 534‐538, 1995.
7. E. Alonso, A. M. Sherman, T. J. Wallington, M. P. Everson, F. R. Field, R. Roth and R. E. Kirchain
"Evaluating Rare Earth Element Availability: A Case with Revolutionary Demand from Clean
Technologies", Environ. Sci. Technol., 46 (6), 3406‐3414, 2012.
8. A. R. Chakhmouradian, F. Wall "Rare Earth Elements: Minerals, Mines, Magnets (and More)",
Elements, 8 (5), 333‐340, 2012.
9. F. Habashi "The Recovery of the Lanthanides from Phosphate Rock", J. Chem. Technol. Biot.
A, 35 (1), 5‐14, 1985.
10. K. Binnemans, Y. Pontikes, P. T. Jones, T. Van Gerven and B. Blanpain "Recovery of rare earths
from industrial waste residues: a concise review", Proceedings of the Third International Slag
Valorisation Symposium (19‐20 March 2013, Leuven, Belgium), 191‐205, 2013.
11. M. T. Ochsenkühn‐Petropoulou, K. S. Hatzilyberis, L. N. Mendrinos and C. E. Salmas "Pilot‐
Plant Investigation of the Leaching Process for the Recovery of Scandium from Red Mud",
Ind. Eng. Chem. Res., 41 (23), 5794‐5801, 2002.
12. K. Binnemans, P. T. Jones, B. Blanpain, T. Van Gerven, Y. X. Yang, A. Walton and M. Buchert
"Recycling of rare earths: a critical review", J. Clean. Prod., 51, 1‐22, 2013.
13. Tanaka, M.; Oki, T.; Koyama, K.; Narita, H.; Oishi, T. Recycling of Rare Earths from Scrap. In
Handbook on the Physics and Chemistry of Rare Earths, Vol. 43, Chapter 255, Bünzli, J.‐C. G.,
Pecharsky, V. K., Eds.; Elsevier: Amsterdam, 2013; pp 159‐212.
14. K. Binnemans, P. T. Jones "Perspectives for the recovery of rare earths from end‐of‐life
fluorescent lamps", J. Rare Earths, 32 (3), 195‐200, 2014.
15. J. W. Darcy, H. M. D. Bandara, B. Mishra, B. Blanplain, D. Apelian and M. H. Emmert
"Challenges in Recycling End‐of‐Life Rare Earth Magnets", JOM, 65 (11), 1381‐1382, 2013.
16. H. M. D. Bandara, J. W. Darcy, D. Apelian and M. H. Emmert "Value Analysis of Neodymium
Content in Shredder Feed: Toward Enabling the Feasibility of Rare Earth Magnet Recycling",
Environ. Sci. Technol., 48 (12), 6553‐6560, 2014.
17. X. Y. Du, T. E. Graedel "Global Rare Earth In‐Use Stocks in NdFeB Permanent Magnets", J. Ind.
Ecol., 15 (6), 836‐843, 2011.
18. X. Y. Du, T. E. Graedel "Uncovering the Global Life Cycles of the Rare Earth Elements", Sci.
Rep., 1, Article Number: 145, 2011.
19. X. Y. Du, T. E. Graedel "Global In‐Use Stocks of the Rare Earth Elements: A First Estimate",
Environ. Sci. Technol., 45 (9), 4096‐4101, 2011.
20. X. Y. Du, T. E. Graedel "Uncovering the end uses of the rare earth elements", Sci. Total
Environ., 461, 781‐784, 2013.
21. J. H. Rademaker, R. Kleijn and Y. X. Yang "Recycling as a Strategy against Rare Earth Element
Criticality: A Systemic Evaluation of the Potential Yield of NdFeB Magnet Recycling", Environ.
Sci. Technol., 47 (18), 10129‐10136, 2013.
22. K. Nansai, K. Nakajima, S. Kagawa, Y. Kondo, S. Suh, Y. Shigetomi and Y. Oshita "Global Flows
of Critical Metals Necessary for Low‐Carbon Technologies: The Case of Neodymium, Cobalt,
and Platinum", Environ. Sci. Technol., 48 (3), 1391‐1400, 2014.
23. Biswas, D. I. Potential for recovery of cerium contained in automotive catalytic converters:
U.S. Geological Survey Open‐File Report 2013‐1037, 10 p., available at
http://pubs.usgs.gov/of/2013/1037; 2013.
24. O. Gutfleisch, M. A. Willard, E. Bruck, C. H. Chen, S. G. Sankar and J. P. Liu "Magnetic
Materials and Devices for the 21st Century: Stronger, Lighter, and More Energy Efficient",
Adv. Mater., 23 (7), 821‐842, 2011.


Session II: European Rare Earth

Resources (Plenary)
UNLOCKING THE POTENTIAL OF RARE EARTH
RESOURCES IN EUROPE
Nikolaos ARVANITIDIS1, Kathryn GOODENOUGH2
1
Division of Bedrock and Geochemistry, Geological Survey of Sweden, Box 670, 75128
Uppsala, Sweden
2
British Geological Survey, West Mains Road, Edinburgh EH9 3LA, United Kingdom
Abstract
During the last decade the non‐energy raw materials have become part of the EU’s
new industrial strategy and growth agenda, with increased funding opportunities in
the field of mineral resources. The four‐year (2009–2013) ProMine project has provided
a well documented knowledge base of Europe's non‐energy raw material resource
potential. The ongoing EURARE project is focused on enhancing the value chain
perspectives for European Rare Earth Elements (REE) and the Minerals 4EU project
aims to provide harmonized data on European mineral resources.
This paper summarises ongoing work by the EURARE project (www.eurare.eu). REE
mineral potential belts, including genetic types related to igneous, hydrothermal and
sedimentary processes, have been identified across Europe. Potential REE deposits are
hosted by carbonatites, alkaline igneous complexes and volcanic successions, granitic
pegmatites, iron oxide copper gold deposits, and skarns, as well as secondary deposits
such as laterites, bauxites and heavy mineral sands.
This paper considers issues of the various deposit types with respect to REE‐bearing
minerals, grades, volumes, Heavy (HREE) to Light REE (LREE) ratios, radioactive
element concentrations, genetic aspects, all in a perspective of raising the awareness
for exploration and the potential for exploitation.
Introduction
The European Union (EU) is highly dependent on imports of mineral raw materials that
are crucial for a strong European industrial base, sustainable and competitive growth,
and a thriving society. A specific group of minerals and metals are characterized as
critical raw materials, because they are considered to have significant economic
importance for the EU but have clear risks to their supply (1). All the rare earth
elements (REE) belong within this group because of their importance in electronics and
green technologies, and the dominance of China in their production. For several
metals used in electronics and green technology industries, including the REE and
platinum group elements, the EU completely relies on imports. The resulting annual
shortage is about €11 billion, of which 90% corresponds to metallic minerals,
particularly those of major high‐tech applications. Recycling of metal scrap represents
around 40% to 60% of the input to the EU's metal production, according to industry
estimates.
Rare Earth Elements (REE) is the collective name for 17 chemically similar metallic
elements (the lanthanides, scandium and yttrium) that occur together in a wide range
of minerals and are mined collectively. They are usually divided into the light REE
(LREE) and the heavy REE (HREE); the latter are found in relatively lower
concentrations in the Earth's crust. China today controls about 95 % of the world’s REE
production, making the supply of these elements for European industry highly
vulnerable (2).
At the same time, Europe’s mineral potential is under‐explored, both with regard to
the subsurface (particularly deeper than 150 metres), and the sea‐bed in the EU
member states' exclusive economic zones. Major opportunities for access to raw
materials exist within the EU today, especially for mining at greater depths or in small
deposits. The sea‐bed could also contain valuable raw materials, including REE
deposits, leading to growing world‐wide competition for marine mineral deposits.
More intense and advanced exploration for REE mineral deposits on land and in the
marine environment is therefore needed, and seafloor mineral resources are the
subject of growing European interest.
This paper aims to provide an overview of different REE mineral deposit types in
Europe, and to provide a basis for further research into the mineralizing processes
causing REE enrichments, and controlling the formation of feasible grades and
volumes. It intends also to deliver some preliminary guides for exploration and
exploitation potential.
Available datasets
Mineral occurrence databases exist for most countries, but they typically lack full
information on REE resources, and very rarely include quantitative estimates of the
resource size or how likely it is to go into production. This paper summarises the
information that is available from a number of projects carried out under various EU
funding schemes and commissioned studies (Fig.1), as follows:
 ProMine (http://promine.gtk.fi) has produced a pan‐European primary mineral
occurrence database and anthropogenic concentration (mainly mining wastes)
database, which also contains REE information (3,4). About 270 REE occurrences
were recorded. However this database does not provide information on size of
occurrence or on ratio of LREE to HREE.
 EuroGeoSource (http://www.eurogeosource.eu) has collected mineral resource information, but
only REE occurrences in Greenland are present.

 EuRare (http://www.eurare.eu) aims to set the basis for the development of a European rare earth
element (REE) industry by safeguarding the uninterrupted supply of REE raw materials and products
crucial for sectors of the EU economy. It will define and assess all exploitable REE mineral
resources and REE demand in Europe (Appendix I), and develop an Integrated Knowledge
Management System (IKMS) for EU REE resources, which will provide information on REE and build
up the knowledge to be developed within the frame of the project.

 Minerals4EU (http://www.minerals4eu.eu/) will include the development of an EU Minerals
Knowledge Base Platform that goes along and in line with the Strategic Implementation Plan for
European Innovation Partnership on Raw Materials, and a Minerals Yearbook that will include
resource and reserve data for primary minerals. Publication is anticipated in 2015.

 Production statistic data are compiled on a yearly basis by British Geological Survey
(BGS) staff
(https://www.bgs.ac.uk/mineralsuk/statistics/europeanStatistics.html). The
numbers are based on data that are obtained from the national statistical agencies
or geological surveys within the individual countries. These data are usually easily
obtainable since companies are being taxed according to the actual production,
and therefore obligations exist in most countries to report such information to the
public authorities.
European geology and metallogeny
The geological evolution of Europe extends over some 3.8 billion years of Earth history.
The oldest rocks in Europe are found in the Archaean cratons of Greenland, the UK,
and the Scandinavian countries, whilst the youngest Cenozoic rocks are formed along
the Mediterranean and Atlantic fringes. The rocks of Europe have been affected by a
large number of orogenic events, such as the Svecofenian and Caledonian orogenies
in the north and the Variscan and Alpine ones in central and southern Europe. These
generated various types and grades of deformation and metamorphism, associated
with a range of compositions and volumes of magmatic lithologies emplaced into
widely varying tectonomagmatic settings. Metallogenetic evolution throughout this
period resulted in emplacement of several major mineral belts, such as the
Fennoscandian VMS, IOCG and Iron Oxide deposits, the Iberian Pyrite belt, and the
porphyry Cu and epithermal Au Carpathian‐Balkan belt (3). Years of research have led
to a good understanding of base and precious metal, copper, nickel and iron
occurrences, but the genetic processes related to REE mineralisation are not as well
understood. The importance of understanding tectonic setting in REE exploration is
developed further by Goodenough et al. (this volume).
Primary resources: proposed genetic type classification
The majority of significant primary REE resources in Europe are in alkaline igneous
rocks and carbonatites, and the most important deposits occur in extensional rift‐
related igneous provinces of a range of ages (Goodenough et al., this volume). In
Northern Europe, these igneous provinces are deeply eroded and the major deposits
lie within plutonic silicate and carbonatite igneous rocks, such as the spectacularly
layered peralkaline syenites of the Ilimaussaq intrusion in southwest Greenland
(Kringlerne and Kvanefjeld deposits; 5) or the Norra Kärr syenite in Sweden (6). Also of
importance across Europe, but most particularly in Sweden, are zones of hydrothermal
and pegmatite mineralization. In many cases the source of the REE in these
mineralizing systems is not well understood. Each primary deposit has a different,
commonly very complex, ore mineralogy and many REE‐bearing minerals also have
high contents of U and Th. For this reason, each deposit presents its own challenges in
processing, and a range of factors will affect the economic and environmental case for
the deposit.
In Southern Europe, the rift‐related provinces are at much shallower levels of erosion,
and the surface expression of alkaline magmatism is in alkaline volcanic rocks that do
not contain significant REE concentrations. In these areas, secondary deposits formed
by weathering (bauxites and laterites) and sedimentary processes are more important.
It is likely that primary REE deposits remain to be discovered at depth within these rift
zones.
Classification systems currently used by BGS, USGS and in Australia, as well as modified
versions developed and applied in the ProMine, ASTER (7) and EuroGeoSource
projects, have been evaluated, integrated and adjusted to fit the geological setting,
metallogenetic evolution and mineral belts in Europe, as part of the EURARE project.
Taking into account the INSPIRE Mineral Deposit Group and Mineral Deposit Type code
lists, a simplified genetic classification approach (host rock type, mineral forming
processes and composition, and LREE and HREE concentration and distribution) is
applied, dividing REE mineralisations/deposits into two main categories:
 Primary deposits (Fig.1) linked to magmatic and hydrothermal REE‐mineral forming processes :
o Carbonatites (LREE‐enriched deposits with bastnäsite, allanite, apatite and monazite as
the main minerals e.g. Fen in Norway, Sökli in Finland (8), Sarfartoq and Qaqarssuk (9) in
west Greenland)
o Alkaline‐peralkaline igneous rocks (LREE‐ or HREE‐enriched deposits with apatite,
eudialyte, gadolinite and loparite among the main minerals e.g. Norra Kärr in Sweden (6),
Katajakangas in Finland, Kvanefjeld and Kringlerne in south Greenland). This deposit type
includes the largest resources known in Europe at present.
o Iron Oxide Apatite (IOA) and Iron Oxide Copper Gold (IOCG) deposits (usually LREE‐
enriched with apatite and/or allanite e.g. Kiruna magnetite‐apatite deposits (10) in
Sweden)
o Granitic pegmatites (ranging from LREE‐enriched, containing allanite and monazite, to
HREE‐enriched with gadolinite as main mineral e.g. Ytterby mineralisation in Sweden, Evje‐
Iveland pegmatite district in Norway)
o Hydrothermal/hydrogenetic (mainly LREE‐enriched and varying grades, with bastnäsite,
allanite, monazite, gadolinite and parasite as common minerals e.g. Bastnäs‐Norberg‐Nora
(11) skarn mineralisation zone in Sweden)
o Sedimentary concentrations on land and in the marine (12) environment (e.g. LREE‐
enriched in U‐deposits, fluorite‐barite mineralisations, with disseminations of synchysite
and parisite, phosphates in NW Greece on land, and REE mineralisations associated with
deep sea‐floor polymetallic nodules and cobalt‐rich crusts)
 Secondary deposits (Fig.1) related mainly to sedimentary remobilization and weathering processes
of mainly REE‐bearing igneous rocks:
o Placers, of paleo‐, marine and alluvial affiliation (commonly LREE and U‐Th enriched, with
monazite, xenotime, allanite as common minerals, e.g. Nea Peramos coastal heavy sands
in northern Greece, Klocktorpet and Tåsjö paleoplacer apatite mineralisations in Sweden)
o Residual‐lateritic/bauxitic (usually LREE‐enriched, high‐graded and monazitic e.g. central
Greece monazite‐bastnäsite lateritic/bauxitic mineralisations),
o Ion‐adsorption clays/residual clays (HREE‐enriched, commonly low‐grade high‐volume,
associated with weathered REE‐bearing granites, with REE being adsorbed by clay minerals
such as kaolin and halloysite).

Secondary resources
Secondary resources are those not directly derived from natural REE occurrences, but
rather from material that has already been used by humans. Mining wastes, including
various waste streams such as weakly mineralized waste rocks, processing tailings and
metallurgical residues, could turn out to be potential resources for recovery of
exploitable grades of REE. The abandoned historical mine waste dumps in Bastnäs
district and the stockpiled apatite leftovers in Kiruna, both in Sweden, as well as the
red mud residues of Al smelting in Greece and elsewhere in Europe (Deady et al., this
volume), are prominent examples of raw materials that might be secondary resources
of REE. Recycling of material from the consumer electronics industries represents
another source of the REE (2). However, in spite of the fact that recycling is a vital
component of a resource efficient economy and a long‐term target to secure REE
supply, the EU will still be very much dependent on primary mineral resources
extraction for the foreseeable future.

Exploration data availability
There have been several national‐based geochemical surveys that have carried out
soil, bedrock and stream sediment sampling across Europe. Many EU member states
have produced geochemical maps for varying subsets of the REE on national and
regional scales. However this information, although very valuable in regional‐ scale
REE exploration campaigns, is rather variable and fragmented, and requires much
harmonization before Europe‐wide REE mineralised belts can be recognised. Projects
reporting data and delivering pan‐European geochemical maps include FOREGS (13,)
and GEMAS (http://www.eurogeosurveys.org/projects/gemas/), carried out by the
EuroGeoSurveys Geochemistry Expert Group, the latter based on sampling of
agricultural and grazing land soils (14, 15, and 16).
Geophysical measurements applying airborne and ground magnetic, electromagnetic
and gravimetric methods have been conducted across Europe on national and regional
scales. They may provide valuable data and offer efficient REE exploration tools at a
range of scales, but this is based more on the interpretation of the responses and
signatures received from the wall and host rocks of the mineralisation rather than the
REE mineralisation itself. As yet there is limited information to show how
interpretation of geophysical data may be better used to pinpoint REE mineralization.
Fig.1: Overview of major REE mineralisation types in Europe and Greenland based on
EURARE data and information (17). There is obvious exploration potential and high
prospective interest for primary deposits in Greenland, the Nordic countries and the
British islands, and secondary deposits in mainly NW France, Greece and west Balkans

Preliminary recommendations
Economic aspects
There are some advanced REE projects listed on Table I, providing an overview of
developments taking place in Europe. Economically feasible projects in Europe are
progressing in Sweden (Norra Kärr) and Greenland (Kringlerne and Kvanefjeld) with a
total potential (resources and reserves) of all three together in the order of 30 million
tonnes REO. The trend in Europe is the same as in the rest of the world since 1960s,
showing that REE mining projects are mainly related to primary deposit types
associated with carbonatites and alkaline igneous rocks, compared to prior to 1950
when most bulk REE production was derived from monazite‐bearing placers.
Table I: Major REE advanced projects in Europe and Greenland
Deposit name Country Type Size
41.6 Mt @ 0.57%
Sweden Alkaline igneous
Norra Kärr TREO
4.5 Mt @ 0.6%
Sweden Iron oxide ‐ apatite
Olserum TREO
122 Mt@ 1.4%
Greenland Alkaline igneous
Kvanefjeld TREO
Greenland Carbonatite 5.9 Mt@ 1.8% TREO

Sarfartoq
494 Mt @ 0.07%
Turkey Placer
Aksu Diamas TREO

Exploitation potential
Based on the geological and compositional characteristics of the major REE
mineralisation types in Europe and Greenland, it is obvious that the strongest
exploration potential and highest prospective interest are for primary deposits in
Greenland and the Nordic countries, and secondary deposits in NW France, Greece
and the west Balkans. A range of factors may affect exploitation potential of REE
deposits: these include ratio of more valuable HREE to LREE; overall grade and
tonnage; content of the radioactive elements; environmental considerations;
proximity to transport and infrastructure; and availability of processing routes for the
deposit type and mineralogy. Currently, the larger, high‐grade carbonatites and
alkaline rocks are of highest exploration potential and priority, but processing of these
hard‐rock deposits is energy intensive and costly. For this reason, development of new
exploitation technologies could mean that the lower‐grade but more easily processed
secondary deposits may grow in importance.
More efficient exploration
Mineral exploration is the only way to ensure secure and sustainable supply of raw
materials. Even if 100% recycling efficiency is attained this will never be able to meet
the increasing supply demand for REE. Exploration has for many years been focused
on base metal sulphide and iron oxide mineralised environments. New advanced
technologies related to geochemistry and geophysics were mainly developed to fit to
the geological and metallogenetic characteristics of these systems. There is now a
need to better understand the REE ore forming processes of all the primary and
secondary deposit types, in order to develop the right exploration guides and apply
them to European geological settings, making the discovery of new ore deposits
possible. Carbonatites and alkaline rocks may be of high priority but other types such
as the granitic pegmatite IOA & IOCG ones need to be explored more efficiently.
Examples of required end products might be,
 Distribution of REE (total, HREE, LREE and individual elements) occurrences in Europe (Map)
 Reserves and resources in Europe and in individual countries (Reports, Excel spreadsheets, maps)
 Mineral potential maps for the REE in Europe
Acknowledgements
The research leading to these results has received funding from the European
Community’s Seventh Framework Programme [FP7/2007‐20013] under the grant
agreement no 309373. This publication reflects only the authors’ view, excepting the
Community for any liability.
References
1. REPORT ON CRITICAL RAW MATERIALS FOR THE EU, Report of the Ad hoc Working Group on
defining critical raw materials, DG Enterprise, European Commission European Commission, (2014).
http://ec.europa.eu/enterprise/policies/raw‐materials/files/docs/crm‐report‐on‐critical‐raw‐
materials_en.pdf.
2. UNCTAD‐United Nations Conference on Trade and Development, “Commodities at a glance:
Special issue on rare earths”, No5, United Nations, New York and Geneva (2014).
3. Cassard D., Bertrand G., Maldan F., Gaàl G., Kaija J., Angel J.M., Ballas D., Billa M., Christidis C., Eilu
P., Grazea E., Inverno C., Kauniskangas E., Lintinen P., Maki T., Matos J., Meliani M., Mladenova V.,
Navas J., Niedbal M., Perantonis G., Pyra J., Santana H., Serafimovski T., Serrano J.J., Strengel J.,
Tasev G., Tornos F., Tudor G.,. ProMine pan‐European Mineral Deposit database: a new dataset for
assessing primary mineral resources in Europe. PROMINE Workshop, Nancy (2012).
4. Nano‐particle products from new mineral resources in Europe ‐ ProMine – Final Report Geological
Survey of Finland (2013).
5. Sørensen, L.L., Kalvig, P., Hanghøj, K. Rare earth element potential in Greenland. Reporting the
BMP/GEUS mineral resource assessment workshop. Danmarks og Grønlands Geologiske
Undersøgelse Rapport 2011, 80, 30 p + 1 DVD, (2010).
6. Bluemel, B., Leijd, M, Dunn, C., Hart, C.J.R., Saxon, M., Sadeghi, M. Biogeochemical expression of
rare earth element and zirconiummineralization at Norra Kärr, Southern Sweden. Journal of
Geochemical Exploration, 133, 15–24 (2013).
7. Charles N., Tuduri, J., Guyonnet, D., Melleton, J. & Pouret, O. Rare earth elements in Europe and
Greenland: a geological potential? An overview. High‐tech elements‐deposits and processes, 12th
SGA Biennial Meeting. Proceedings, Volume 4 (2013)
8. Sarapää, O., Ani, AT., Seppo, I., Lauri, L., Sarala, P., Torppa, A., Kontinen, A. Rare earth exploration
potential in Finland. Journal of Geochemical Exploration, 133, 25–41,(2013).
9. Knudsen, C. Petrology, geochemistry and economic geology of the Qaqarssuk carbonatite complex,
southern West Greenland. Monograph Series on Mineral Deposits , 29, 110 pp (1991).
10. Jonsson, E., Troll, V. R., Högdahl, K., Harris, C., Weis, F., Nilsson, K. P. & Skelton, A. Magmatic
origin of giant central Swedish ”Kiruna‐type” apatite‐iron oxide ores. Scientific Reports 3:, 1644,
1‐8. DOI: 10.1038/srep01644, (2013).
11. Andersson, U. B., Holtstam, D. & Broman, C. Additional data on the age and origin of the Bastnäs-
type REE deposits, Sweden. In E. Jonsson et al. (eds.) Mineral deposit research for a high-tech world,
Proceedings of the 12th biennial SGA meeting, 1639-1642 (2013).
12. Kato, Y, Fujinaga, K, Nakamura K, Takaya,Y, Kitamura,K, Ohta J, Toda,R, Nakashima,T & Iwamori,H.
Deep‐sea mud in the Pacific Ocean as a potential resource for REE. Nature Geoscience 4, 535‐539,
(2011).
13. Sadeghi, M., Morris, G.A., Carranza, E.J.M., Ladenberger, A., Andersson, M. Rare earth element
distribution and mineralization in Sweden: An application of principal component analysis to
FOREGS soil geochemistry. Journal of Geochemical Exploration, 133, 160–175, (2013).
14. Sadeghi M., Petrosino P, Ladenberger A, Albanese S, Andersson M, Morris G, Lima A, DeVivo B, and
The GEMAS Project Team. Ce, La and Y concentrations in agricultural and grazing‐land soils of
Europe, Journal of Geochemical Exploration, 133, 202–213 (2013).
15. Demetriades, A. & Reimann, C.,. Mineral deposits of Europe. Chapter 3 In: C. Reimann, M. Birke,
A. Demetriades, P. Filzmoser, & P. O'Connor (Editors), Chemistry of Europe's agricultural soils – Part
B: General background information and further analysis of the GEMAS data set. Geologisches
Jahrbuch (Reihe B 103), Hannover, Fig. 3.8, p.77) (2014).
16. Reimann, C., Flem, B., Fabian, K., Birke, M., Ladenberger, A., Négrel, P., Demetriades, A.,
Hoogewerff, J., The GEMAS Project Team. Lead and lead isotopes in agricultural soils of Europe ‐
The continental perspective. Applied Geochemistry, 27, Fig. 2, p.534, (2012).
17. Sadeghi, M., Jonsson, E., Kalvig, P., Keulen, N., Goodenough, K., Deady, E., Muller, A., Eliopoulos,
D., Cassard, D., Bertrand, G., Vuollo, J., Nysten, P., Bergman, T., Söderhielm, J., Arvanitidis, N., and
Eurare WP1 team. Rare Earth Element Resources in Europe with focus on the Nordic countries‐ 1st
International Symposium on Development of Rare Earths, Baotou, China, (2014).

THE WORLD‐CLASS REE DEPOSIT TANBREEZ, SOUTH
GREENLAND: ITS SIZE AND STRUCTURE
Hans K. Schønwandt1, Gregory B. Barnes1 and Thomas Ulrich2
1
G.B. Barnes & Associates, 47 Labouchere Road, South Perth, W.A. 6151, Australia
2
Department of Geoscience, Aarhus University, Hoegh‐Guldbergsgade 2, 8000 Aarhus,
Denmark
hank@mail.tele.dk, thomas.ulrich@geo.au.dk
Abstract
The Tanbreez ore deposit is a highly fractionated ortho‐magmatic Zr‐Nb‐Ta‐REE
deposit in the southern part of the 1.13Ga old Ilímaussaq intrusive complex in South
Greenland. The commodities are hosted in the zirconosilicate mineral eudialyte,
occurring concentrated in kakortokite at the floor of the exposed intrusion. The
kakortokite sequence is outcropping over an area of 5 x 2.5 km and has a total thickness
of 335 m. A conservative estimate specifies the resource to more than 4 billion tons.
Linear correlations between ZrO2 and individual REE indicate that eudialyte is by far
the main REE bearing mineral in kakortokite. Estimated average grades are 1.75%
ZrO2, 0.18% Nb2O5 and 0.6% total REO, of which heavy REE make up 30% (including
yttrium).
Introduction
Many of the metals used in today’s technology such as REE are enriched in alkaline
and carbonatitic intrusive complexes. The rare earth market is dominated by China.
Therefore there is a concern outside China about supply of these elements. This has
launched an exploration activity for these elements all over the world. Greenland,
which is known for its alkaline igneous provinces, has been the focus for the industries
REE exploration for the last 10‐15 years.
The Ilímaussaq alkaline complex in southern Greenland host several mineral
deposits one of which is Tanbreez, a world class REE deposit. The anagram,
Tanbreez, is based on the major elements (Ta‐Nb‐REE‐Zr) contained in the
deposit. Moreover, there are elevated concentrations of the elements Hf and Y.
Uniquely; all the commodities are contained in the mineral eudialyte, which is a
rock‐forming mineral in the agpaitic nepheline syenite kakortokite.
Rimbal Pty Ltd, the Australian owner of Tanbreez Mining Greenland A/S, holds an
exploration license since 2001 over the deposit and has invested the last 13 years
in understanding and testing the deposit. The previously existing knowledge,
dating as far back as 1880’s, included drilling and mapping programs and
extensive metallurgical test programs carried out by former license holder has
contributed to the present knowledge of the deposit
In this contribution we describe the structure and thickness of the kakortokite (ore
host sequence) and its contact relationships to the rocks surrounding it.
Furthermore, we show the geochemical correlation of REE and Zr in kakortokite,
including an estimate on the overall size of the deposit.
Geological Setting
The Meso‐Proterozoic Gardar Province in South Greenland contains 10 intrusive
complexes that range in composition from alkali granite to nepheline syenites and
gabbroic dikes. These rocks intruded the 1.9 Ga old Julianenhåb granite basement that
is unconformably overlain by sandstones and basalts of the Eriksfjord formation1, 2.
The alkaline to agpaitic Ilímaussaq intrusion has been dated at 1.13 ± 0.05Ga (Sm‐Nd
mineral isochron3) and belongs to the younger Gardar event. The intrusion is 17 km x
8 km in size and at least 3 km thick. The composite intrusion is cut by the Tunulliarfik
Fjord (Fig 1). The geology of the Ilímaussaq intrusive complex has been described as a
succession of three main phases that crystallized at about 3‐4km depth4,5,6 .

Fig. 1A) Geological map of the Ilímaussaq intrusion. B) Satellite image of the southern
part of the intrusion with the exploration license area in green. Also shown is the drill
hole location of the stratigraphic hole in the central part of the kakortokite.

Phase 1 is an outer ring dyke of augite syenite, which partly rims the complex to the
west, east and south and is up to 1 km wide. The augite syenite also overlies the
intrusive complex.
Phase 2 is alkali granite and quartz syenite in the roof zone of the complex.
Phase 3 consists of peralkaline rocks, evolving into the dominant agpaitic nepheline
syenite of the complex. This phase can be subdivided into a:
Roof sequence that crystallizes from the roof downwards, forming the following
succession of rocks downwards: pulaskite, foyaite, sodalite foyaite and naujaite (Fig. 1
naujaite‐foyaite series). The three first mentioned rock types occur as a thin, but
persistent cover of part of the complex on both sides of the Tunulliarfik Fjord (Fig 1a).
The naujaite is by far the thickest unit of the roof sequence with a thickness between
600 m and 800 m7.
Floor sequence solidifying upwards. The floor sequence consists of a rhythmically
layered rock called kakortokite.
Intermediate series that is sandwiched between roof and floor series, consisting of
different types of lujavrites.
The kakortokite contains a significant amount of the mineral eudialyte, and hence, is
enriched in zirconium, rare earth elements, niobium and tantalum. The kakortokite
sequence is regarded as the host of a world‐class, multi‐element deposit.
The structure of the kakortokite body
Kakortokite and lujavrite constitute the lower sequence of the stratified intrusion
where kakortokite is overlain by lujavrite. The exposed thickness of the two rock types
is nearly 1000 m8
The kakortokite has been subdivided into three parts8. From bottom to top:
‐ Lower layered kakortokite (LLK) (approx 209 m thick)
‐ Slightly layered kakortokite (SLK) (approx 35 m thick)
‐ Transitional layered kakortokite (TLK) (approx 40 m thick).
All planar elements are locally parallel and the combined structure recalls a simple
saucer shaped pattern with steep to vertical structures at the periphery, which rapidly
turns into a general dip of 10‐15 degrees for the main part of the kakortokite8. Steeper
dips of the sequence occur just north of Lakseelv, but this is probably linked to drag
associated with the Lakseelv Fault (Fig 1b).
The kakortokite sequence is exposed over an area of 5 x 2.5 km on the south side of
the Kangerdluarsuk Fjord (Fig 1) and ranges from sea level up to about 400 m asl. The
exposed area comprises 95% kakortokite and 5% related rocks, mostly syenite dykes
and sills.
The Lower layered kakortokite and the Transitional Layered Kakortokite show
subhorizontal layered structures, whereas the Slightly Layered Kakortokite sequence
is a virtually unlayered grey and fine‐grained kakortokite.
The spectacular layering in LLK is composed of black, red and white layers, depending
on the amount of specific minerals. The black layers are rich in the mineral
arfvedsonite and contain lesser amounts of eudialyte, alkali‐feldspar and nepheline.
The red layers are composed of eudialyte, with lesser amounts of arfvedsonite, alkali‐
feldspar and nepheline, whereas the white layers are composed of alkali‐feldspar and
nepheline with lesser amounts of arfvedsonite and eudialyte. Locally sodalite occurs
as a significant mineral phase.
Conventionally, the three colored layers together form a unit. The exposed part of the
kakortokite consists of 29 units labelled +1 to +17 and ‐1 to ‐11 above and below unit
0, respectively9.
The layering can be recognized from a distance; however, in some cases the exact
contacts between layers is not always clearly defined close up and in drill core due to
more transitional contacts between the main mineral phases of the layers. There is
some variation in thickness between the layers as well as in texture and grain size,
which helps to identify some horizons. On average, a unit is about 10‐13 m thick. There
are, however, units that are not fully developed; in some cases black/red layers are
very faint or missing. An average relative thickness of the individual layers is 1.5m for
black, 1m for red and 10m for white layers10. In general, the eudialyte content (main
host for Zr, and REE) of the black and white layers is slightly below 10vol%, whereas
the eudialyte content of the red layers is around 30‐40 vol%.
Following is a description of the contact relationships of the kakortokite sequence to
the upper, lower, and marginal boundaries of the intrusion.
Boundaries of the kakortokite body
The upper boundary of the kakortokite sequence is a gradual and concordant
boundary to the overlaying lujavrite sequence.
The lower boundary of the kakortokite is not exposed, which has led to speculations
on the thickness of the kakortokite. Exploration drillings by Tanbreez Mining
Greenland A/S in 2007 in the coastal area of Kangerdluarsuk near the southern
boundary of the Ilímaussaq intrusion (Fig 1b) encountered a black to grey, porphyritic,
fine grained rock with peppered sodalite in the matrix below the kakortokite. The rock
contains 5‐10 vol% subhedral to euhedral, poikilitic biotite grains 1‐2 mm in diameter,
and aggregates of aenigmatite, biotite and arfvedsonite in a groundmass of feldspar,
nepheline and sodalite. Geochemically the rock composition is equivalent to a tephri‐
phonolite. During drill core logging this rock was temporarily call “Black Madonna” and
that name will also be used throughout this paper. Black Madonna was found in 8 drill
holes so far, but none of the drill holes has completely penetrated Black Madonna. The
longest interval drilled in Black Madonna is 220 meter.
There is a reasonable good correlation of the contact between kakortokite and the
massive occurrence of Black Madonna between the drill holes. It is apparent, that the
top of Black Madonna is dipping towards the northwest. Based on the limited data it
seems that Black Madonna has a dip similar to the layering of the kakortokite. This was
corroborated by the depth of the occurrence of Black Madonna in the stratigraphic
hole (DX‐01) drilled in 2010 about 1.5 km east of the drillings in 2007 (Fig 1b). This
further indicates that Black Madonna could lie under the entire kakortokite sequence.
The contact between Black Madonna and the Lower layered kakortokite is well defined
without been sharp. In the contact zone rounded inclusions of Black Madonna often
5‐10 cm long occur in the layered kakortokite and it is not uncommon to encounter
inclusions 50 m or more from the contact. However there is a clear decrease in number
of inclusions away from the contact.
A rim of massive agpaitic nepheline syenite, which consists of the same minerals as
the layered kakortokite sequence, characterizes the marginal borders of the
kakortokite sequence. This led ref. 11 to describe the rocks along the rim as a not
banded kakortokite. The rim has also been referred to as marginal pegmatite by ref. 9,
12. Because the rim is a complex mixture of different eudialyte‐bearing rock types,
including pegmatite, a more appropriate name would be ‘border zone’ of the
kakortokite. The border zone represents a transition between the layered kakortokite
and the augite syenite ring dyke of the Ilímaussaq intrusion (Fig 1a). The border zone
is about 150 m wide at sea level at Kangerluarsuk and decrease to about 50 m at an
altitude of about 450 m. The zone is made up of eudialyte‐bearing nepheline syenite,
which shows a variation in texture from medium to coarse‐grained massive rock The
inner boundary of the border zone against the kakortokite is gradual over about 5‐10
m where the layering becomes more subtle and turns into a massive agpaitic rock. The
outer boundary of the border zone towards the augite syenite is subvertical, sharp and
without a chilled margin. The border zone’s complex intergrowth of texturally
different kakortokite like rocks is cut by a eudialyte poor pegmatite mainly containing
feldspar, nepheline, sodalite, arfvedsonite and aegirine. The proportion of the
pegmatite increases towards the augite syenite and at the same time it largely parallels
the augite syenite. In the outer 10 m of the border zone they occur as a subvertical
swarm of pegmatite locally associated with hydrothermal alteration. The pegmatite is
the youngest member of the border zone and from an economical point of view, down
grades part of the border zone and this part will most likely not be subject to mining.

Size of the kakortokite body
The exceptionally well‐exposed kakortokite in combination with the subhorizontal
structure of the sequence were used by ref. 9 to estimate the tonnage of the lower
layered kakortokite. The calculation was based on detailed mapping of the 29 rhythmic
units in the exposed part of kakortokite. Ref. 13 repeated the estimate based on the
geological map: Southern part of the IlÍmaussaq Complex, South Greenland, 1:20000
published by Grønlands Geologiske Undersøgelse (1988). Ref. 9 and 13 measured
209m and 218 m, respectively for the total thickness of the lower layered kakortokite,
which is within the expected uncertainty of this kind of survey. The volume of
kakortokite was determined using a planimeter to measure the area on the map
between the contours of the different rhythmic units.
Exploration drillings in 2007 showed that the lower contact of the kakortokite and
Black Madonna occurs about 20 m below sea level on the coast of Kangerluarsuk (Fig
1), which means that 20 m should be added to the thickness measured on the surface.
Assuming 214 m as the average for the layered kakortokite based on the previous
studies, the total thickness becomes 234 m.
Subsequently, the stratigraphic borehole DX‐01 (Fig 1b) was drilled in the central part
of the kakortokite body in 2010, and shows a total thickness of the lower layered
kakortokite of 269 m, which is 35 m more than the measured thickness on the surface.
The difference in thickness cannot convincingly be explained by the uncertainty in the
measurements, but is most probably related to structural conditions. This makes the
correlation between numbered layers on the surface and the layering observed in drill
cores difficult, except for correlation over short distances and involving kakortokite
layers with very characteristic textures.
Previous estimates of the tonnage of the lower layered kakortokite west of
Laksetværelv (Fig 1b) yielded 3.7 billion tons9. The estimate was based on the thickness
of the lower layered kakortokite of 209 m, which is 60 m less than the thickness found
in drill hole DX‐01. The new estimate, using the thickness from DX‐01, and using a
linear interpolation from previous estimates would increase the tonnage to 4.7 billion
tons. This number would increase even more if kakortokite of the slightly layered
kakortokite, the transitional layered kakortokite and the border zone would be
included.
Ore Grade
Resource assessments in earlier studies9, 13 show a variation in ore grade of the lower
layered kakortokite between 1.4% ZrO2 to 1.9% ZrO2. The average ore grades based
on assay data from drillings from the area along Kangerluarsuk where the initial mining
is planned is 1.75% ZrO2, 0.18% Nb2O5 and 0.6% total REO including yttrium. The
commodities are all contained in eudialyte, a Na‐rich zirconosilicate mineral. Eudialyte
is by far the most abundant Zr‐bearing mineral in kakortokite, occurring in the black,
white and red layers. The bulk rock data show close linearly correlation between ZrO2
and Nb, Ta and light and heavy REO (Fig 2), which is a clear indication that eudialyte is
the dominant REE‐bearing mineral.
The distribution of the total REO in the ore (kakortokite) shows a quantity of 30.9%
heavy REE (including Y) and 69.1% light REE (Fig 3). Investigations have shown that no
or very little cryptic variation occurs in the minerals of kakortokite10,14 , consequently,
little change in the eudialyte composition is expected in ore and therefore the
magnetic properties of eudialyte would remain the same for the benefit of the planed
magnetic concentration of eudialyte . Importantly, drill core assays show elements like
U and Th have background values (20ppm and 53ppm, respectively), which is an
advantage in processing the ore.

Fig. 2. Assay data showing good correlation between Zr and Dy as a representative of
the REE.

Fig. 3. Proportion of the different REE+Y found in the Tanbreez deposit

Conclusion
The Tanbreez polymetallic Zr‐Nb‐Ta‐REE deposit is hosted in the Ilímaussaq intrusion
in South Greenland. The ore body is made up of kakortokite that is characterized by
rhythmically layered units of layers that are either dominated by feldspar, arfvedsonite
and eudialyte, respectively. The kakortokite sequence is defined by gradual contacts
to the overlying lujavrites and an unexposed (only encountered in drill cores) contact
zone with numerous xenoliths to the underlying Black Madonna unit. The border zone
contact to the augite syenite ring dike is well defined and the outer part of the zone
cut by pegmatoid dikes.
Zr shows perfect correlations with REE in bulk rock analyses, which indicates that the
main ore mineral is eudialyte. The thickness of the lower layered kakortokite sequence
is a about 270m, and the volume of the ore body equals about 4.7 billion tons of ore.
The average ore grade where the initial mining is planned is 1.75% ZrO2, 0.18% Nb2O5
and 0.6% total REO. The proportion of HREE is up to about 30%.

References
1. V. Poulsen, The sandstones of the Precambrian Eriksfjord Formation in South Greenland. Rapport
Grønlands Geologiske Undersøgelse, 2, p. 16 (1964).
2. F. Kalsbeek and P. N. Taylor, Isotopic and chemical variation in granites across a Proterozoic
continental margin‐ the Ketilidian mobile belt of South Greenland. Earth and Planetary Science
Letters, 73, 65‐80, (1985).
3. C.R. Paslick, A. N. Halliday, G. R. Davies, K. Mezger, B.G.J. Upton, Timing of Proterozoic magmatism
in the Gardar Province, southern Greenland. Geological Society of America Bulletin, 105, 272‐78,
(1993).
4. J. Konnerup‐Madsen, J. Rose‐Hansen, E. Larsen, Hydrocarbon gases associated with alkaline
igneous activity: evidence from compositions of fluid inclusions. Rapport Grønlands Geologiske
Undersøgelse, 103, 99‐108, (1981).
5. H. Sørensen and L. M. Larsen, Layering in the Ilímaussaq alkaline intrusion, South Greenland. Origin
of Igneous Layering, Dordrecht, 1‐28 (1987)
6. J.C., Bailey, R., Gwozdz, J. Rose‐Hansen and H. Sørensen, Geochemical overview of the Ilimaussaq
alkaline complex, South Greenland. Geology of Greenland Survey Bulletin 190, 35‐53, (2001)
7. S. Andersen., H. Bohse, and A. Steenfelt,.: A geological section through the southern part of the
Ilimaussaq intrusion. Rapport Grønlands Geologiske Undersøgelser 103, 39‐42. (1981)
8. H. Bohse, H. and S. Andersen, S.: Review of the stratigraphic divisions of the kakortokite and
lujavrite in southern Ilímaussaq. Rapport Grønlands Geologiske Undersøgelse, 103. 130 pp, (1981).
9. H. Bohse, C. K. Brooks, and H. Kunzendorf: Field observations on the kakortokites of the Ilímaussaq
intrusion, South Greenland. Rapport Grønlands Geologiske Undersøgelse ,38. 43 pp., (1971).
10. J. Ferguson: The significance of the kakortokite in the evolution of the Ilímaussaq intrusion, South
Greenland. Bulletin Grønlands Geologiske Undersøgelse . 89, 193 pp., (1970).
11. N. V. Ussing: Geology of the country around Julianehåb, Greenland. Meddeleleser om Grønland
38, 426 pp., (1912).
12. H. Sørensen, H. Bohse, and J. C. Bailey: The origin and mode of emplacement of lujavrites in the
Ilímaussaq alkaline complex, South Greenland. Lithos, 91, 286‐300, (2006).
13. G. B. Barnes: Ilimaussaq (Narsaq) Deposit, Exploration Licence 2001/08. Annual report 2005.
Internal Report (2005)
14. K. Pfaff, T. Krumerei, M. Marks, T. Wenzel, T. Rudolf, G. Markl: Chemical and physical evolution of
the ‘lower layered sequence’ from the nepheline syenitic Ilímaussaq intrusion, South Greenland:
Implications for the origin of magmatic layering in peralkaline felsic liquids. Lithos, 106, 280‐96.
(2008).


Session III‐A: European REE
Supply and Demand

PRIMARY AND SECONDARY SOURCES OF RARE EARTHS
IN THE EU‐28: RESULTS OF THE ASTER PROJECT
Dominique GUYONNET1, Mariane PLANCHON2, Alain ROLLAT3, Victoire
ESCALON2, Stéphane VAXELAIRE1, Johann TUDURI1
1
BRGM ‐ ENAG, 3 av. C. Guillemin, 45060 Orléans (F)
2
BIO by Deloitte, 132 Avenue Charles de Gaulle, 92200 Neuilly sur Seine (F)
3
SOLVAY Group, 24 rue Chef de Baie, 17000 La Rochelle (F)
d.guyonnet@brgm.fr
Abstract
The ASTER project aims at establishing flows and stocks of certain critical rare earth
elements at the scale of the EU‐28, considering both primary and secondary sources.
Material flow analyses were performed taking into account processes along the value
chain (separation, manufacture, use, waste management) and including also
lithospheric (geologic) stocks. While this paper focuses on fluorescent lamp phosphors
(Tb, Eu, Y), a paper by Planchon et al. (this conference) addresses permanent magnets
and batteries (Nd, Pr, Dy). Results suggest that given the magnitude of flows in the EU‐
28, the development of a mining project in Sweden and/or Greenland would contribute
significantly to reducing heavy rare earth; e.g. Tb, criticality.

Introduction
Member state and European policies with respect to the safe supply of mineral raw
materials aim at improving eco‐efficiency1. The objective of eco‐efficiency is to
increase the quantity of services per unit mass of raw material, while decreasing
environmental impacts. One of the pillars of eco‐efficiency is recycling. In order to
efficiently position the development of innovative recycling processes, it is important
to have a clear picture of flows of stocks of mineral raw materials all along the value
chain. The on‐going ASTER project, started in 2012 for a period of three years, aims at
applying material flow analysis (MFA) to selected rare earth elements (REEs) for the
EU‐28. Certain rare earth elements are deemed “critical”, as they combine importance
for strategic sectors of the economy with risks of supply shortage3, 4. The recent JRC
report2 identifies six critical REEs (Dy, Eu, Tb, Y, Pr and Nd) that are precisely the REEs
that were selected for the ASTER project.

Given the importance of recycling, the ASTER project combines an “application”
approach with an “element” approach. Table 1 shows the correspondence between
the two. These three applications were selected because of their importance in terms
of use of the selected REEs, and also because recycling processes for these applications
either already exist or else are currently under development5.

Table 1: Correspondence between applications and REEs addressed in the ASTER
project

Application REEs
Phosphor powders Eu, Tb, U
Permanent magnets Nd, Pr, Dy
NiMH batteries Nd

While REEs such as Nd are used in a variety of other applications, it was considered of
lesser importance to examine, e.g., flows of Nd in ceramics, as there are currently no
realistic prospects of recycling Nd contained in these products. In this paper, we adopt
the following distinction between light (LREE) and heavy (HREE) rare earth elements6:
‐ LREE : La, Ce, Pr, Nd, Sm ;
‐ HREE : Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb, Lu (plus Y).
Methods
A systemic appoach
For the case of fluorescent light phosphors, the system that was analysed is described
schematically in Figure 1. The figure illustrates a certain number of processes with,
starting upstream in the value‐chain: separation (S) of rare earth concentrates,
fabrication (F) of trichromatic phosphor powders, manufacture (M) of fluorescent
lamps (LFL and CFL; resp., linear and compact fluorescent lamps), use (U) of these
lamps in the economy, waste management (WM) of end‐of‐life products and final
elimination (landfill and/or dissipation in the environment). These various processes
are linked by arrows that represent flows from one to another. Differences between
flows into or out of a process imply an addition or subtraction to stock within the
process. The dashed line in Figure 1 represents the limits of the studied system (the
EU‐28). However, an additional “process” appears in Figure 1 indicated as “L”
(lithosphere). The lithosphere represents geological potentialities in the form of
potential REE resources (as opposed to reserves7). For geological and economic
reasons, the limits of this process were taken as continental Europe (including the
Scandinavian shield and the Kola Peninsula) plus Greenland. Therefore this process
straddles the limits of the EU‐28.

It is important to note that the system analysed in this study is an “open” system: there
are inflows and outflows to and from the system all along the value chain. This is in
contrast with most previous REE MFA studies which have addressed global flows and
hence a “closed” system8. In the latter case, flows can be “constrained” by world
mining production, as reported by, e.g., the USGS.

Figure 1: Schematic of the system investigated in the case of fluorescent‐lamp
trichromatic phosphors. LFL = linear fluorescent lamps; CFL = compact fluorescent
lamps

Flows were estimated over a number of years exceeding the expected average lifetime
of the considered products (taken as 6 years for fluorescent lamps), in order to obtain
estimates of in‐use stocks. Year 2010 was selected as the reference year for the final
results of the calculations, due to data availability at the start of the project.
Information sources
Data mining constitutes the main task of any MFA. Sources of information included
statistical custom data (EUROSTAT, Global Trade Atlas, etc.), specialized reports10, 11,
USGS data12, company reports on product sales, etc. With respect to the upstream
portion of the value chain, the same HS codes (listed in Table 2) were used as in the
Öko‐Institut report11. GTA data for Chinese exports based on 8‐digit HS codes that are
specific to individual REEs were also used.

These HS codes represent raw products that contain mixtures of REEs. In order to
obtain data on individual REEs, a disaggregation method was applied13. The method
considers the relative market shares of products containing REEs (e.g., auto catalysts,
magnets, batteries, polishing powders, etc.) and the composition of the REE mixtures
in each product.

Table 2: HS codes used for REE‐containing products in the upstream portion of the
studied system

HS code Description in EUROSTAT
28053010 Intermixtures or interalloys of rare‐earth metals, Scandium and Yttrium
28053090 Rare‐earth metals, Scandium and Yttrium (Excl. intermixtures or interalloys)
28461000 Cerium compounds
28469000 Compounds, inorganic or organic, of rare‐earth metals, of Yttrium or of
Scandium or of mixtures of these metals (Excl. Cerium)

Certain data from EUROSTAT had to be corrected to account for the fact that Austria
no longer reports its REE‐related data after 2008. GTA data for Austria were
disaggregated into imports and exports from intra‐ and extra EU‐28, considering
averages calculated over the pre‐2008 period.

For the evaluation of geological potentialities in the lithosphere, over 350 occurrences
and/or deposits located within the EU‐28 were analysed14 (see also Charles et al., this
conference). The mineralogical characteristics of these deposits and occurrences
provide information regarding potential resources, the reliability of which depends
largely on the state of progress of exploration projects. However, the objective was to
provide orders of magnitude for potential resources (lithospheric stocks), that could
be compared with the magnitude of flows in the EU economy.
Uncertainties and reconciliation
Information collected in the course of a MFA for critical metals is inevitably fraught
with uncertainty. This uncertainty is typically of an “epistemic” (reflecting incomplete
knowledge) rather than a “stochastic” (reflecting random variability) nature.
Uncertainty implies that MFAs do not balance and therefore reconciliation methods
are used15. Such methods generally assume that uncertainty can be represented in the
form of Gaussian probability distributions (see, e.g., the STAN software9).

However, epistemic uncertainty (such as typically ensues from expert‐opinion) is more
faithfully represented by the nested intervals (or fuzzy sets) of possibility theory16.
Therefore, as part of the ASTER project, a methodology was developed for reconciling
MFAs under fuzzy constraints17, 18. The methodology described in17, 18 provides an
alternative to the classical least‐squares minimization approach, which relies on a
purely probabilistic representation of uncertainty, whereas uncertainties in this
context are typically epistemic (reflecting the incomplete/imprecise nature of
available information). With the fuzzy‐constraint reconciliation methodology,
uncertainty relative to flows and/or stocks is represented using fuzzy sets (e.g. a
minimum value, a maximum value and a preferred value). The approach then consists
in finding the maximum level of consistency between input information, while taking
into account the various constraints (mass balances and memberships).

Results
The results presented herein are preliminary as the project is still on‐going. An
important deliverable of the project will be Sankey diagrams for the considered REEs.
These diagrams are described qualitatively below for the case of REEs in fluorescent
light phosphors.

A first step of the disaggregation procedure consisted in converting tons of products
corresponding to the customs data for the HS codes of Table 2, into tons of rare earth
oxide mixtures (REOs). The conversion factors were selected based on data from
USGS12. For the metals and alloys (codes 280530), the conversion factor is 1.2 based
on stoichiometric considerations. For codes 284610 and 284690, conversion factors
were taken as, resp., 0.67 and 0.72.

Application of these conversion factors to historical imports and exports of raw
products containing REEs in the EU‐28 yields Figure 2. The significant decrease in
imports observed in 2009 is interpreted as a consequence of the 2008 financial crisis.

25 000
Imports
20 000
Exports
15 000
Tons REO
10 000
5 000

Figure 1: EU‐28 imports and exports of REOs based on EUROSTAT (and GTA for Austria
post‐2008 correction)

Application of the disaggregation methodology to these data provides information
relative to flows F1 and F2 of Figure 1, which can then be compared with information
provided by SOLVAY, the only company in the EU‐28 to perform separation of REEs for
trichromatic phosphors. Separation is also performed in Estonia by SILMET, but not for
phosphor applications (mainly for Nd). Further to separation, SOLVAY produces
phosphor precursors that are sent to process “F” (Figure 1) to produce phosphors.
Information relative to imports and exports of phosphors is obtained based on HS code
320650 (inorganic products of a kind used as luminophores, whether or not chemically
defined). Expert information was necessary to define the typical composition of these
phosphors, taking into account that the composition is different for imported versus
exported phosphors (in and out of the EU).

Further down the value chain, data was collected for manufactured products
containing trichromatic phosphors. The two HS codes of Table 3 were used.

Table 3: HS codes for fluorescent lamps

HS code Description in EUROSTAT
31501530 Fluorescent hot cathode discharge lamps (excluding ultraviolet lamps, with double ended cap)
31501510 Fluorescent hot cathode discharge lamps, with double ended cap (excluding ultraviolet lamps)

Results of the analysis suggest that the orders of magnitude of Tb, Eu and Y flows in
the EU‐28 related to fluorescent lamp phosphors are resp., 10, 15 and 180 tons. Taking
into account the history of flows into use in the EU‐28 over a period exceeding the
expected average lifetime of the fluorescent lamps, in‐use stocks for Tb, Eu and Y in
2010 are estimated as, resp., 140, 220 and 2300 tons. The annual additions to in‐use
stock in 2010 are estimated to be resp., +8, +13 and +140 tons/yr for these elements.

With respect to flows F13 and F14 in Figure 1 (recycling), there was no recycling in
2010 in the EU‐28, as SOLVAY started its recycling operations (in Saint‐Fons and La
Rochelle, France) in 2012. As an indication, the order of magnitude of Tb recycled each
year is a few tons per year.

These figures can be compared to geological potentialities derived based on the
analysis of three important REE projects: the Norra Kärr project in Sweden and two
exploration projects in Greenland (Kvanefjeld and Kringlerne). The Norra Kärr project
is well advanced and has recently received the authorization to operate from the
Swedish government. We estimate the resource of Tb at this site to be on the order of
2 000 tons. The potentials of Kvanefjeld and especially Kringlerne are far superior
(resp. 10 000 and 110 000 tons Tb), however, the likelihood of these projects entering
into production is much lower, due in particular to environmental constraints and
social acceptability. Nevertheless, these figures suggest that if just one of these
projects enters the production phase (with Norra Kärr as the most likely candidate),
the criticality with respect to certain heavy REEs such as Tb will be significantly
reduced.

Conclusions
Results of the on‐going ASTER project provide estimates of flows and stocks of certain
REEs considered as critical. This paper addresses the results relative to REEs in
fluorescent lamp phosphors, while another paper (Planchon et al., this conference)
addresses Nd and Dy.

The data collected to‐date suggest a somewhat “reassuring” picture with respect to
criticality of Tb, Eu and Y. It is the opinion of the authors of this paper that the crisis
regarding REEs that was experienced in 2011 is essentially over. It is reminded that in
July 2011, peak prices of Dy and Tb exceeded average prices over the period 2002‐
2009 by factors on the order of, resp., 20 and 10, as a result of stricter quotas on
Chinese exports. Several factors have contributed to reducing the pressure on these
substances. On the one hand, countries that import REEs have diversified their sources
of supply of LREEs (e.g., Mount Weld in Australia and Mountain Pass in the U.S.). Also,
there are currently considerable efforts devoted to optimizing the quantities of REEs
used in products, as for example Dy in permanent magnets. Substitution is also an
important issue. Regarding Tb, the development of LED lighting, which use very little
REEs and virtually no Tb, will contribute in the future to reducing the pressure on this
element. In the case of Eu in fact, the development of LEDs could very well lead to a
large excess of supply versus requirements in the relatively near‐future.

Detailed results of the ASTER project are currently in the process of being published.
It is anticipated that the approach of associating both primary and secondary sources
in the systemic analysis of critical metals will be applied to other substances (e.g., In,
Ge, Ga, etc.), as is the case for projects that have already been initiated. It should be
noted however that, as for REEs, data regarding such “small” metals are hard to come
by. On the one hand, there is very little information available regarding reliable
estimates of resources6 for substances that are essentially by‐products of carrier
metals19. Also, considerable effort is still needed to quantify the quantities of these
substances in products, where they may be present in very low concentrations (on the
order of a ppm). Such information is crucial in order to evaluate the economic
feasibility of recycling processes.
Acknowledgments
The material presented herein was developed within the ASTER project, supported by
the French National Research Agency; ANR (project ANR‐11‐ECOT‐002).
References
15. COM. Roadmap to a resource efficient Europe. COM 571, European Commission, Brussels, 2011.
16. Moss, R., Tzimas, E., Willis, P., Arendorf, J., Thompson, P., Chapman, A., Morley, N., Sims, E.,
Bryson, R., Peason, J., Tercero‐Espinoza, L., Sartorius, C., Ostertag, K. Critical metals in the path
towards the decabonisation of the EU energy sector. Assessing rare metals as supply‐chain
bottlenecks in low‐carbon energy technologies. JRC Report EUR 25994 EN. 242 pp., 2013.
17. NRC. Minerals, Critical Minerals, and the U.S. Economy. U.S. National Research Council,
Washington, D.C., National Academies Press, 245pp., 2008.
18. EC. Critical raw materials for the EU. Report of the ad‐hoc Working Group on defining critical raw
materials. European Commission, June 2010.
19. Binnemans, K., Jones, P.T., Blanpain, B., Van Gerven, T., Yang, Y., Walton, A., Buchert, M. Recycling
of rare earths: a critical review. Journal of Cleaner Production, 51, 1‐22., 2013.
20. Wall, F. Rare earth elements. In: “Critical Metals Handbook”, 1st edition (Ed. Gus Gunn), John Wiley
& Sons, Ltd., 2014
21. PERC. Pan‐European standard for reporting of exploration results, mineral resources and reserves.
Geological Service of Belgium, Brussels, Belgium, 2013.
22. Du, X., Graedel, T.E. Uncovering the global life cycles of the rare earth elements. Scientific Reports,
1(145): 1‐4, 2011.
23. Brunner, P.H., Rechberger, H. Practical Handbook of Material Flow Analysis. Lewis Publishers,
2004.
24. ROSKILL. Rare Earths & Yttrium: Market Outlook to 2015. 14th edition. Roskill Information Services
Ltd, London, UK, 2011.
25. Schüler, D., Buchert, M., Liu, R., Dittrich, S., Merz, C. Study on rare earths and their recycling. Öko‐
Institut Report, Darmstadt, Germany, 140 pp., 2011.
26. USGS. Mineral commodity summaries 2013. U.S. Geological Survey, 198 pp., 2013.
27. Goonan, T.G. Rare earth elements – End use and recyclability. U.S. Geological Survey Scientific
Investigations Report 2011‐5094, 2011.
28. Charles, N., Tuduri, J., Guyonnet, D., Melleton, J., Pourret, O. Rare earth elements in Europe and
Greenland: a geological potential? An overview. 12th meeting of the Society of Geology Applied
to Mineral Deposits (SGA), 12‐15 August 2013, Uppsala, Sweden, 2013.
29. Kelly, J.D. Techniques for solving industrial nonlinear data reconciliation problems. Computers and
Chemical Engineering, vol. 28, pp. 2837‐2843, 2004.
30. Dubois, D., Prade H. Possibility Theory: An Approach to Computerized Processing of Uncertainty,
Plenum Press, New York, 1988.
31. Dubois, D., Fargier, H., Ababou, M., Guyonnet, D. A fuzzy constraint‐based approach to data
reconciliation in material flow analysis. International Journal of General Systems (in press), 2014.
32. Dubois, D., Fargier, H., Guyonnet, D. Data reconciliation under fuzzy constraints in material flow
analysis. EUSFLAT‐2013. Eighth conference of the European Society for Fuzzy Logic and
Technology, September 11‐13, Milano, Italy, 2013.
33. Verhoef, E.V., Dijkema, G.P., Reuter, M.A. Process knowledge, system dynamics, and metal
ecology. Journal of Industrial Ecology, Vol 8, 1‐2, pp. 23‐43, 2004.

THE ROLE OF RARE EARTHS SUPPLY RISK IN LOW
CARBON TECHNOLOGIES INNOVATION
Eva BARTEKOVÁ1
1
UNU‐MERIT/Maastricht University, Keizer Karelplein 19, 6211 TC Maastricht, the
Netherlands.
bartekova@merit.unu.edu
Abstract
Considering the quest for alternative energy and transportation modes and their
importance for sustainable growth, this paper examines to what extent the supply risk
of rare earths poses a barrier to the increased adoption of low carbon technologies.
Using secondary data collected on offshore wind turbines and electrically powered
vehicles, the analysis allows determining the actual quantities of rare earths used
within permanent magnets and batteries. The results of this paper disprove the
widespread allegation that availability risk impedes deployment of offshore wind.
Contrary to this, a potential supply shortage would disrupt the further development of
the automotive industry and its electrification.
Introduction
Within the past couple of years, growing concerns over massively skewed world
production and supply of rare earths have emerged in the specific context of clean
energy technologies. On the one hand, the supply risk arises from an institutional
inefficiency of the market. Despite the fact that China is currently the largest rare
earths producer globally with a market share of 89%, and at the same time possesses
single largest reserves, concentration per se is not the main driver of supply disruption.
Instead, it is China’s protectionist policies in light of its national interests of industrial
upgrading which substantially distort the market 1. On the other hand, the increased
adoption of low carbon technologies, driven by institutional settings such as the EU's
Strategic Energy Technology Plan (SET‐Plan), as well as by industrialisation effort of
developing countries 2, require stable supply of rare earths. This growing imbalance
between increased demand and tighter supply implies that while low carbon
technologies are intended to tackle climatic changes, their deployment seems to be
adversely affected by resource scarcity. The question arises then to what extent do
potential supply shortages of rare earths pose barriers to the deployment of latter.

The aim of the present paper is to examine the specific case of offshore wind turbines
and alternative technology vehicles ‐ the two technologies which were identified by
scientific literature as major drivers of change in rare earths demand, and which are
often incorrectly claimed by headlines of newspaper articles as impossible to be
manufactured without rare earths content. In fact, several forecasts by national and
international organisations drawn up recently, acknowledge the source of supply risk
as being the rapid increase in consumption and China’s export policies 3‐7. Along these
lines, 8 construct demand scenarios of rare earths originating from production of wind
turbines and electric vehicles. However, these usage projections underlie many
simplifications and assumptions that do not depict the status quo of technology.
Consequently, 9 attempts to determine quantitative shares for HDDs, mobile phones,
wind turbines and electro mobility in rare earths permanent magnets. While the
author offers a realistic picture on the former two using primary data analysis, his
estimates prove incomplete for the latter two. In fact, he only considers permanent
magnets in electric traction drives, disregarding those present in other parts of a
vehicle (e.g. sunroofs and automatic door locks), what leads to an underestimation of
their respective demands for rare earths. Similarly, batteries are not taken into
consideration. Moreover, figures are based on vehicles’ sales in Germany and the US
only. In this light, the present paper estimates global sales data on both energy and
car maker/model levels globally, what allows drawing more robust conclusions. In
terms of wind turbine technology, at present no analysis exists on the content of rare
earths used within offshore wind.
Methods
The analysis is carried out by first compiling an overview of various topologies of
generators, electric motors and batteries. For this purpose, specialised engineering
literature was analysed. Second, in order to determine the quantities of rare earths
used in offshore wind, all offshore wind farm projects (i.e. deployment phase of a
farm) were screened globally. Data was collected for fully and partially commissioned
projects, as well as for those under construction in 2014, from dedicated online
databases and news portals on offshore wind 10‐15 and subsequently enriched with
information on generator designs from manufacturers' websites. Projects under
construction where turbine types were not publicly announced were omitted.
Similarly, for electrically powered vehicles, data was collected from manufacturers'
websites and specialised online portals 16‐20 for all models in production by end of
2014. Only passenger cars were taken into account. Buses, heavy duty vehicles,
motorcycles and three‐wheelers are not within the scope of this analysis. Also, vehicles
in prototype or pre‐production stages and those not mass produced were excluded,
while discontinued models were kept. Since global sales figures broken down to car
maker/model levels are not publicly available, an estimation was carried out based on
information gathered from various sources 21‐31 and from respective car
manufacturers' websites.
Offshore Wind Turbines
The central priority of the SET‐Plan is to create a low‐carbon economy and wind power
is one of the technologies foreseen to contribute to this transition6. Wind is the fastest
growing renewable energy source, having increased 8‐fold over the past 10 years 32.
But limited space for further development of wind energy onshore, availability of
higher wind speeds as well as opportunities to launch industrial wind projects of
magnitudes parallel to power stations, led to progressive siting of wind turbines
offshore 33. In line with this, recent R&D projects are focusing on challenges primarily
in offshore technology. In terms of drive train innovations, the focus lies on
improvement of reliability of gearbox and on development of direct drive and hybrid
drivetrain technologies. In essence, the main difference between these lies within
their drivetrain designs, which in turn influence the requirement for rare earths. The
detailed discussion on respective topologies can be found in 1 and has also been
summarised in the Table 1.

Table 1: Generator types in wind turbine technologies and their respective
permanent magnet and rare earths contents.

On the one extreme, traditional geared designs – with squirrel cage generators (SCIG)
or with doubly‐fed induction generators (DFIG) 34 – are rare earths free. On the other
extreme, while gearless direct drive machines contain no rare earths if excited by
electromagnets (electric excitation synchronous generator (EESG‐DD)), when designed
with permanent magnets (permanent field excitation generators (PMSG‐DD)) these
turn very rare earths intensive. In fact, some 650 kg/MW of magnets are used in these
machines, with a rare earth content of 250 kg/MW 35, 36. Note that there is a lack of
consensus on permanent magnet content within these generators, opinions ranging
from 500‐625 kg/MW 37, through 600‐800 kg/MW 38, up to 1000 kg/MW reported by
various consultancies. Hybrid drivetrains, containing a single‐stage (PMSG ‐ SG) or a
two‐stage gearbox (PMSG ‐ MG) combined with a medium speed permanent magnet
generator 33, represent a compromise between the two standards. Their substantially
smaller magnets with 160 kg/MW and 80 kg/MW respectively, contain up 10 times
less neodymium, dysprosium, praseodymium and terbium per MW of power produced
compared to PMSG‐DD machines 39, 40. Finally, High Temperature Superconductor
(HTS) technology utilises minute quantities of yttrium, lanthanum and cerium within
its YBCO conductors ‐ the latter two being the most abundant rare oxides 41, 42.

By end of April 2014, a total of 99 offshore wind projects has been operating 2'463
turbines and producing 7.62 GW of offshore wind power globally. Following
observations stand out: in terms of capacity generated, the most widely spread design
is the conventional geared drive train with SCIG, which together with DFIG make up
96% of current offshore market. Permanent magnet driven technology lags
significantly behind covering the remaining market share of 4%, with an equal split of
2% between direct drive and hybrid generator designs. In terms of the distribution of
generator types across projects, 16% of all commissioned projects make use of
permanent market drivetrain technology. This can be explained by the fact that many
of these turbines are currently at prototype stage and are being tested in low numbers
at various test sites. Finally, average power range of generator types is illustrated in
Figure 1 (left). While conventional geared drive train designs have similar average
powers of 2'881 MW for SCIG and of 2'766 MW for DFIG, this is substantially higher
for PMSG topologies with 3'447 MW in aggregate. The average of 4'500 MW for EESG‐
DD is to be interpreted with caution since currently there is only one turbine deployed
offshore worldwide. Having said this, one can conclude that PMSG drivetrains, though
with currently lower market share, tend to be deployed within larger power range
turbines. While the UK is the leader in total installed offshore capacity, followed by
Denmark and Germany, by far the most permanent magnet driven machines are at
present deployed offshore in China.

Accounting for projects under construction in 2014, in addition to those already
commissioned, the capacity generated increases by more than half to 12.72 GW and
number of turbines increases to 3'871. Share of PMSG turbines increases to 16% of
total capacity, with shares of hybrid models increasing to 5% and 6% for PMSG‐SG and
PMSG‐MG respectively, while share of PMSG‐DD increases to 5%. Furthermore, share
of projects relying on permanent magnet driven technology increases to 21%. Figure
1 (right) indicates that the average power range of permanent magnet excited direct
drive machines increases slightly to 3'621 MW and the total capacity produced
increases by 600% ‐ from currently 284 MW to 2’000 MW of produced power once
these projects turn fully operational. At the same time, increase in geared drivetrains
remains below 50%. Market share of electrically excited generator technology remains
unchanged. To summarize, there is acceleration in deploying permanent magnet
drivetrains in the near future, at the expense of traditional geared technologies. This
trend is predominantly driven by Germany which is planning to add 600 MW of power
produced by PMSG turbines offshore, followed by China and the UK with 464 MW and
436 MW respectively, and Belgium with half as much additional power produced at
full commissioning of projects. While European countries will be installing only hybrid
drives, half of added MW in China will be produced by direct drive generator wind
turbines.

Figure 1: Generator designs in global offshore wind projects: commissioned (left);
commissioned and under construction in 2014 (right).

In terms of quantities of rare earths used, assuming contents as indicated in Table 1,
108 tonnes of permanent magnet and 40 tonnes of rare earths have been used for all
currently commissioned offshore wind farm projects so far. For all newly installed
turbines in projects under construction, manufacturing of generators required 430
tonnes of permanent magnets, the rare earths content of which was slightly more than
150 tonnes. On the one hand, this represents an increase in demand for rare earths by
approximately 380%, while on the other this constitutes less than half a percent of
respective neodymium and dysprosium supplies forecasted for 2014. In general,
future expansion of offshore wind is tied to regulatory stability, as well as to existence
of cable routes and high voltage transmission lines linking offshore wind farms to
mainland grid 33. Penetration of PMSG topologies in particular, will depend on
availability risk and pricing volatility of rare earths on the one hand, and competition
from alternative technologies on the other.
Advanced Technology Vehicles
Besides reduction of greenhouse gases' emissions, another main line of argumentation
in favour of advanced technology vehicles is the reduction of importer countries'
dependence on foreign oil. The question remains though, to what extent does the
technology shift from internal combustion engine to advanced technology vehicles
expose us to dependence on rare earths.

Based on the level of hybridisation, modern vehicles can be classified into internal
combustion engine vehicles (ICEV), hybrid electric vehicles and all‐electric vehicles
(AEV) 43. The latter two can be further broken down based on energy sources used into
battery, fuel cell and solar energy models. The latter two technologies are still in early
stages of development and hence are not in the scope of the present analysis. Battery
electric vehicles (BEV) are propelled entirely by battery packs, while full‐hybrid
vehicles (HEV) run on fuel alone, and plug‐in hybrids (PHEV) use the combination of
gasoline and electric motors 43‐45. An overview is provided in Table 2. One thing all
advanced technology vehicles have in common is the electric motor in their
drivetrains. An overview of main motor design drives together with their specifications
and comparison of performance is provided by 46. In short, the most common design
is the permanent magnet motor with currently more than 80% of market share, as
compared to the induction motor with a market share of 11% only 47. Just like in wind
turbines, magnets in electric motors contain neodymium, dysprosium and traces of
praseodymium and terbium. It has been estimated that the electric traction drive
contains some 0.6 kg of rare earths 48.

The case of batteries is somewhat different in that several battery technologies co‐
exist on the market. An overview of secondary battery systems is reported in 43, 49‐53.
Battery technologies currently used in respective vehicle energy models are discussed
in 1 and also summarised within Table 2. In short, besides the two most widespread
designs ‐ nickel metal‐hydride (NiMH) and lithium‐ion (Li‐ion) ‐ other battery
technologies exist on the market: nickel‐cadmium (Ni‐Cd) and lead acid (PbA)
(outdated); zinc‐air (Zn‐air), lithium‐sulphur (Li‐S) and lithium‐iron phosphate (LFP)
(under development); and Zebra and lithium‐metal‐polymer (LiPo) (unproven
viability). However, currently only the NiMH battery contains rare earths, its negative
electrode being made of hydrogen absorbed in metal alloy of the type
MN(Ni,Co,Al,Mn)_5. Such batteries contain up to 10% mischmetal (MN), made up
primarily of lanthanum and to some lesser extent of cerium, neodymium and
praseodymium, as well as of metallic impurities 51, 54‐57. In fact, their large usage of rare
earths ‐ 3.5 kg as estimated by 48 ‐ was one of the reasons why rare earths free lithium
battery technology started to catch up.

Table 2: Vehicle technologies classified by hybridisation and their respective rare
earths contents.

Figure 2 (left) summarises the shares of individual designs adopted by car
manufacturers within their vehicles. NiMH batteries are employed within nearly a
third of currently available advanced technology vehicle models, what makes them the
second most deployed battery technology. However, when looking at the breakdown
of uses across specific vehicle energy models as illustrated in Figure 2 (right), it
becomes obvious that the latter are only used within HEV. The reason behind are
limitations in specific power and energy densities, as well as lack of scalability with
respect to its size. In fact, while HEV are propelled by both ICE and electric motor, in
BEV and to some extent in PHEV it is the battery which powers the electric drivetrain
and hence higher performance of the former is required. This makes NiMH batteries
less suitable to be deployed in the latter two 58, 59.

In terms of rare earths quantities used, 60, 61 claim Toyota Prius to be the most rare
earths intensive consumer product ever made with its 11 kg of rare earths content.
However, other sources are more moderate with their estimates ranging from 2.17 kg
to 2.45 kg of rare earths in Prius II batteries alone 5, 62. Finally, according to estimates
by 48, a hypothetical mid‐size HEV with a nickel metal hydride battery, contains 3.5 kg
and 0.6 kg of rare earths within its battery and electric motor respectively, and another
up to 0.44 kg scattered across over 700 vehicle parts. Note that this estimate decreases
to half when considering rare earths content in basic vehicles in developing countries
on the one hand, and that this quantity is contained in all conventional ICEV, for
example as small magnets contained in sunroof and automatic door locks, or in
catalytic converter 48. Two thirds of this quantity is estimated to be made up of
neodymium and 6% of dysprosium. As negligible as these quantities sound across
single vehicles, for the total of 87.25 million vehicles produced globally in 2013 63,
29'700 tonnes of rare earths elements were used based on the above assumptions.
This quantity when converted into REO content constitutes 31% of global rare earths
supply for 2013. In terms of individual rare earths, approximately 70% and 120% of
forecasted global neodymium and dysprosium supplies for 2014 are consumed.
Considering tighter supply estimated for 2016 64, respective consumptions would
increase to 80% and 210%.

Figure 2: Battery technologies deployed: in advanced technology vehicles (left); across
vehicle energy models (right)

For 2013, estimates on global vehicles’ sales indicate that HEV are dominating the
market with a global share of approximately 86%, while respective shares of PHEV and
BEV are 6% and 8%. Based on data collected on battery technologies and on sales
figures it was possible to determine that NiMH batteries currently have an
approximate market share of 75% of all HEV. This implies that, electric motors and
batteries in electrically powered vehicles require 723 tons and 3'413 tons respectively,
what represents an additional 4% of global rare earths supply in 2013. Based on
individual supplies forecasted for 2014, the shares of neodymium, lanthanum and
dysprosium are 4%, 5% and 14%, respectively. Considering the growing deployment of
electrically powered vehicles projected to reach annual global sales of 6.1 million by
2023 27, these alone would require 18% of current annual rare earth's production. This
is an approximation and only holds when assuming that global production stays the
same, as well as market share of NiMH batteries and of rare earths permanent magnet
motors remains unchanged, and that no minimisation takes place. In general though,
there is a substantial uncertainty in deployment trends of energy models and of their
charging infrastructures. These depend on incentives and regulations by governments,
as well as on fuel prices. Additionally, price decreases envisaged for Li‐ion battery
packs and potential price volatilities and availability risk of rare earths might speed up
the paradigm shift towards the former.
Impact of Rare Earths Price Volatility on Low Carbon
Technologies’ Producers
Though China has been dominating the rare earths market since mid‐1990s and export
and production quotas were in place already in early 2000s, the rare earths
problematic has only become a prominent topic in recent years. One of the reasons
behind are the small quantities of rare earths used in single technologies, as well as
the largely inelastic demand. Moreover, since rare earth prices are rather low relative
to the total cost of the final product, changes in prices did not have substantial impact
on producer costs. However, price hikes like those in 2010‐2011 called innovation of
direct drive turbines in question, despite them being a superior technology on open
sea65‐67. For illustration, assuming a PMSG‐DD with rated power of 2 MW, rotor
diameter 82 m and hub height 80 m, and following the manufacturing cost breakdown
according to components as specified by 33, cost of 80 USD/kg of required quantity of
permanent magnet constitutes 23% of generator costs and 6% of total component
costs. Assuming a variation of prices of permanent magnet up to 420 USD/kg during
the rare earths crisis in 2011 68, this increases the share of permanent magnet costs to
almost 2/3 of the cost of generator and a quarter of cost of all components. This in
turn increases the cost per kW by 50% relative to competing DFIG gearbox technology
with same parameters. Consequently, uncertainty about high and volatile prices and
threat of supply risk of rare earths is inducing producers to adjust the designs of their
drivetrains. For example, within the EU project on developing a 20 MW wind turbine
(INNWIND), while its predecessor (UPWIND) considered conventional drive train
optimisation, the new project is to focus on superconductive and magnetic pseudo
direct drive generator (PDD) which integrates a low ratio magnetic gear thanks to
which the substantial size and cost of conventional direct drive generators are
eliminated 69‐71.

For what concerns raw material costs in advanced technology vehicles, according to
the component cost comparison 47, the cost of a 50 kW permanent magnet electric
motor is by 30% ‐ 200% higher than that of a 50 kW copper rotor induction motor,
depending on the pricing of rare earth magnets. Indeed, substantial price swings and
uncertainty about stable supply of rare earths spurred R&D on induction electric
motors by some car manufacturers 72‐75. Others are developing electric motors with
reduced amounts of dysprosium 76‐78. Also, Hitachi has been developing a prototype
of a motor using amorphous metal for iron cores and ferrite magnets as rotors 79, while
Hybrid Electric Vehicle Technologies has patented switched reluctance motor which
eliminates use of rare earths by ferromagnetic magnet rotor 80. Batteries instead,
constitute the most expensive part of electric vehicles ‐ ranging from a third to over a
half of HEV's component cost. The cost of rare earths used within a NiMH battery,
when converted into rare earths mischmetal composed of 63.0 wt.% La, 17.4 wt.% Ce,
10.7 wt.% Pr, 2.2 wt.% Nd and 6.7% others 81, increased substantially from 21 USD on
30/12/2005 to 121 USD on 18/04/2014. Considering metal prices as of end of June
2011 when rare earths crises peaked, the estimated cost of total rare earth content
increased by 3300% when compared to 2005 year‐end. Nonetheless, during the time
of higher prices producers presumably use up their private stocks so as to mitigate
temporary price hikes. Currently, there seems to be a paradigm shift ongoing from
NiMH to lithium technology, which despite its higher production costs is cheaper in
terms of materials used 82. As an example, Ford has been switching away from NiMH
battery in its earlier Escape model to Li‐ion battery in its Fusion and C‐MAX models.
Finally, the quantities of rare earths contained in regular ICEV remain a serious issue
in the coming years. Nevertheless, it is expected that in the long run these can also be
replaced by other materials resulting from basic research initiatives as discussed in 1.
Conclusions: Separating Myth from Reality
This paper examined the role of rare earths supply risk in the specific context of low
carbon technologies deployment. The results disprove the widespread allegation that
availability risk of rare earths impedes deployment of offshore wind per se. Contrary
to this, the deployment of permanent magnet technology in particular might become
prone to supply risk, depending on its future market share and level of hybridisation
adopted. While its share of total capacity produced is currently on increase, its share
of annual global supply of rare earths is negligible. The automotive market and its
electrification instead, are substantially dependent on availability of rare earths. They
required over a third of the global rare earths production in 2013 and the share is
expected to increase further with growing deployment of both conventional ICEV and
electrically powered vehicles. Therefore a potential supply disruption would impede
further development of the market altogether.
To conclude with, uncertainty about volatile prices and threat of supply shortages
induce manufacturers to shift away from technologies containing rare earths, and
ultimately render innovation in the latter economically non‐viable. In the face of
increasing demand for these, it might follow that manufacturers, and thus all their
innovation activities, relocate to China with view of sourcing cheaper inputs at stable
supply. Consequently, it becomes imperative for policy makers to design policies in
order to avoid getting the world “trapped” in the rare earths dependence. In this light,
concerted action should be directed towards R&D on material substitution within
permanent magnets, on increasing efficiency of rare earths recycling and on
developing sustainable mining and processing techniques. It remains to be seen
whether China’s distortive industrial policies will be abolished following the recent
resolution on the WTO dispute.
References
1. Barteková, E., An Introduction to the Economics of Rare Earths. 2014,
Submitted for Publication as Working Paper, UNU‐MERIT/Maastricht
University.
2. Sampath, P.G., L. Michael, and C. Razo, Technology and Innovation Report
2011. 2011, UNCTAD. p. 1‐85.
3. European Commission, Critical Raw Materials for the EU. 2010, Report of the
Ad‐hoc Working Group on Defining Critical Raw Materials. Brussels: European
Commission. Enterprise and Industry. p. 1‐85.
4. Bauer, D., et al., US Department of Energy Critical Materials Strategy. 2010, US
Department of Energy.
5. Bauer, D., et al., US Department of Energy Critical Materials Strategy. 2011, US
Department of Energy.
6. Moss, R., et al., Critical Metals in Strategic Energy Technologies. 2011, JRC‐
Scientific and Strategic Reports, European Commission Joint Research Centre
Institute for Energy and Transport.
7. Kawamoto, H., Japan's Policies to be Adopted on Rare Metal Resources.
Quarterly Review, 2008. 27: p. 57‐76.
8. Alonso, E., et al., Evaluating Rare Earth Element Availability: A Case with
Revolutionary Demand from Clean Technologies. Environmental Science &
Technology, 2012. 46(6): p. 3406‐3414.
9. Zepf, V., Rare Earth Elements: A New Approach to the Nexus of Supply, Demand
and Use: Exemplified Along the Use of Neodymium in Permanent Magnets.
2013: Springer.
10. 4C Offshore Limited, 4C Offshore Wind Farm Database. 2014.
11. Pierrot, M., The Wind Power ‐ Wind Turbines and Wind Farms Database. 2014.
12. Recharge. 2013; Available from: http://www.rechargenews.com/news/.
13. Wind Power Offshore. 2013; Available from:
http://www.windpoweroffshore.com/.
14. Wind Power Monthly. 2013; Available from:
http://www.windpowermonthly.com/.
15. Offshore Wind. 2013; Available from: http://www.offshorewind.biz/.
16. China Auto Web. 2014; Available from: http://chinaautoweb.com/chinese‐
brands/.
17. Hybrid Autos. 2014; Available from: http://www.hybrid‐autos.info/.
18. Hybrid Cars. 2014; Available from: http://www.hybridcars.com/.
19. Plug‐In Cars. 2014; Available from: http://www.plugincars.com/.
20. US Department of Energy. 2014; Available from:
http://www.fueleconomy.gov/.
21. EV Obsession. World Electrified Vehicle Sales (2013 Report). 2013; Available
from: http://evobsession.com/world‐electrified‐vehicle‐sales‐2013/.
22. evsroll.com. Hybrid Car Statistics. 2013; Available from:
http://www.evsroll.com/Hybrid_Car_Statistics.html.
23. Green Car Congress. Navigant Research Forecasts New EV Global Sales of >
346'000 Units in 2014; 10 Predictions for the Year. 2014; Available from:
http://www.greencarcongress.com/2014/01/20140108‐navigantev.html.
24. Hybrid Cars. Top 10 Best‐Selling Hybrids. 2013; Available from:
http://www.hybridcars.com/best‐selling‐hybrids.
25. Mock, P., European Vehicle Market Statistics ‐ Pocketbook 2013. 2013,
International Council on Clean Transportation.
26. Sehgal, V., White Paper on Global Electric Cars Market. 2013, RNCOS Business
Consultancy Services.
27. Schreffler, R. Analysts Downsize Global PHEV, EV Sales Forecasts. 2013;
Available from: http://wardsauto.com/sales‐amp‐marketing/analysts‐
downsize‐global‐phev‐ev‐sales‐forecasts.
28. Stoddard, H. Global Hybrid, EV Volume to See Incremental Growth by 2016.
2013; Available from: http://wardsauto.com/plants‐amp‐production/global‐
hybrid‐ev‐volume‐see‐incremental‐growth‐2016.
29. Trigg, T. and P. Telleen, Global EV Outlook: Understanding the Electric Vehicle
Landscape to 2020. 2013, International Energy Agency.
30. World Nuclear Association. Electricity and Cars. 2014; Available from:
http://www.world‐nuclear.org/info/Non‐Power‐Nuclear‐
Applications/Transport/Electricity‐and‐Cars/.
31. Japan Automobile Dealers Association. 2014 New Passenger Car Sales Monthly
Rankings. 2013; Available from:
http://www.jada.or.jp/contents/data/ranking.html.
32. Global Wind Energy Council. Global Cumulative Installed Capacity 1996‐2013.
2014; Available from: http://www.gwec.net/wp‐
content/uploads/2012/06/02_glob‐cum‐inst‐wcap‐96‐13.jpg.
33. Hau, E. and H.V. Renouard, Wind turbines: Fundamentals, Technologies,
Application, Economics. 2013: Springer.
34. Kostopoulos, D., H. Polinder, and A. van den Brink. High Temperature
Superconducting Generators for Direct Drive Wind Turbines: A review. 2012.
Presented during EWEA 2012 Conference, Copenhagen, Denmark.
35. Jensen, B.B. and A.B. Abrahamsen. High Temperature Superconducting (HTS)
Technology for Generators. 2011. Presented during 2nd International
Conference on Drivetrain Concepts for Wind Turbines, Bremen, Germany.
36. Lacal Arántegui, R., K. Suomalainen, and T. Corsatea, 2012 JRC Wind Status
Report: Technology, Market and Economic Aspects of Wind Energy in Europe.
2012, JRC‐Scientific and Policy Reports, European Commission Joint Research
Centre Institute for Energy and Transport.
37. Constantinides, S. The Magnetic Material Challenge. 2010. Presented during
ARPA‐E Workshop on Rare Earth and Critical Materials, Arlington, Virginia.
38. Jensen, B.B., N. Mijatovic, and A.B. Abrahamsen, Development of
Superconducting Wind Turbine Generators. Journal of Renewable and
Sustainable Energy, 2013. 5(2): p. 023137.
39. Vestas. Rare Earth Elements from a Life Cycle Assessments Perspective. 2013;
Available from: http://www.vestas.com/en/about/sustainability\#!material‐
use.
40. Wittrup, S. Permanente Magneter Volder Vestas Problemer i Produktionen.
2011; Available from: http://ing.dk/artikel/permanente‐magneter‐volder‐
vestas‐problemer‐i‐produktionen‐123609.
41. Janssen, L.G.J., et al., Scientific Assessment in Support of the Materials
Roadmap Enabling Low Carbon Energy Technologies: Wind Energy. 2012, JRC‐
Scientific and Strategic Reports, European Commission Joint Research Centre
Institute for Energy and Transport.
42. Buchert, M. Rare Earths ‐ A Bottleneck For Future Wind Turbine Technologies?
2011; Available from: http://www.oeko.de/oekodoc/1296/2011‐421‐en.pdf.
43. Tie, S.F. and C.W. Tan, A Review of Energy Sources and Energy Management
System in Electric Vehicles. Renewable and Sustainable Energy Reviews, 2013.
20: p. 82‐102.
44. Çağatay Bayindir, K., M.A. Gözüküçük, and A. Teke, A Comprehensive Overview
of Hybrid Electric Vehicle: Powertrain Configurations, Powertrain Control
Techniques and Electronic Control Units. Energy Conversion and Management,
2011. 52(2): p. 1305‐1313.
45. International Energy Agency ‐ Hybrid and Electric Vehicle Implementing
Agreement. Clean Vehicle Options. 2013; Available from:
http://www.ieahev.org/about‐the‐technologies/.
46. Pellegrino, G., et al., Comparison of Induction and PM Synchronous Motor
Drives for EV Application Including Design Examples. Industry Applications, IEEE
Transactions on, 2012. 48(6): p. 2322‐2332.
47. Burwell, M., J. Goss, and M. Popescu. Performance/Cost Comparison of
Induction‐Motor & Permanent‐Magnet‐Motor in a Hybrid Electric Car.
International Copper Association 2013; Available from:
http://www.coppermotor.com/wp‐content/uploads/2013/08/Techno‐
Frontier‐2013‐MBurwell‐ICA‐EV‐Traction‐Motor‐Comparison‐v1.8‐Eng1.pdf.
48. Alonso, E., et al., An Assessment of the Rare Earth Element Content of
Conventional and Electric Vehicles. SAE International Journal of Materials &
Manufacturing, 2012. 5(2): p. 473‐477.
49. Van den Bossche, P., et al., SUBAT: An Assessment of Sustainable Battery
Technology. Journal of Power Sources, 2006. 162(2): p. 913‐919.
50. Catenacci, M., et al., Going electric: Expert Survey on the Future of Battery
Technologies for Electric Vehicles. Energy Policy, 2013. 61: p. 403‐413.
51. Kopera, J.J., Inside the Nickel Metal Hydride Battery. Cobasys, MI, 2004.
52. Råde, I. and B.A. Andersson, Requirement for Metals of Electric Vehicle
Batteries. Journal of Power Sources, 2001. 93(1): p. 55‐71.
53. Rydh, C.J. and B. Svärd, Impact on Global Metal Flows Arising from the use of
Portable Rechargeable Batteries. The Science of the total environment, 2003.
302(1‐3): p. 167‐184.
54. Constantinides, S., Rare Earth Elements in Transportation. 2013, Presented
during Materials Science & Technology 2013 Conference, Montreal, Canada.
55. Lichtenberg, F., et al., Development of AB 5 Type Hydrogen Storage Alloys with
Low Co Content for Rechargeable Ni‐MH Batteries with Respect to Electric
Vehicle Applications. Journal of Alloys and Compounds, 1997. 253: p. 570‐573.
56. Ovshinsky, S.R., M.A. Fetcenko, and R. J., A Nickel Metal Hydride Battery for
Electric Vehicles. Science, 1993. 260(5105): p. 176‐181.
57. Shukla, A.K., S. Venugopalan, and B. Hariprakash, Nickel‐based Rechargeable
Batteries. Journal of Power Sources, 2001. 100(1): p. 125‐148.
58. Axsen, J., A. Burke, and K.S. Kurani, Batteries for Plug‐in Hybrid Electric Vehicles
(PHEVs): Goals and the State of Technology Circa 2008. 2008, Institute of
Transportation Studies, University of California.
59. Bitsche, O. and G. Gutmann, Systems for Hybrid Cars. Journal of Power Sources,
2004. 127(1): p. 8‐15.
60. Bryce, R., Rare Earth Metals and Their Use in Advanced Batteries. 2010,
Presented during 2010 Sustainable Mobility Seminar, California, the USA.
61. SAE International. Hybrids, EVs Face New Challenge: The Possible Rarity of Rare
Earths. 2010; Available from: http://articles.sae.org/8544/.
62. Schüler, D., et al., Study on Rare Earths and their Recycling, in Öko‐Institut eV
Darmstadt. 2011.
63. Organisation Internationale des Constructeurs d'Automobiles. World Motor
Vehicle Production Statistics ‐ 2013. 2014.
64. Kingsnorth, D.J., The Rare Earths Industry: Undergoing Rejuvenation. 2013,
Presented during Rare Earths 2013 Conference, Shanghai, China.
65. Bills, K.A. and A.F. Sison. Gamesa Wind: Medium Speed Drivetrains Offer Best
Compromise. 2013; Available from: http://www.wind‐
drivetrain.com/MediaCenter.aspx.
66. Böhmeke, G. Samsung Heavy Industries: Gearbox vs Direct Drive. 2013;
Available from: http://www.wind‐drivetrain.com/MediaCenter.aspx.
67. Sörensen, H. Vattenfall Denmark: Balancing Capex and Opex. 2013; Available
from: http://www.wind‐drivetrain.com/MediaCenter.aspx.
68. Polinder, H., Overview of and Trends in Wind Turbine Generator Systems. 2011,
Presented during XEMC‐Darwind Symposium, Eindhoven, The Netherlands.
69. InnWind. WP3 ‐ Electromechanical Conversion. 2014; Available from:
http://www.innwind.eu/Work‐Packages/WP3‐Electromechanical‐Conversion.
70. Magnomatics. Electrical Machine with Fully Integrated Magnetic Gear Offering
Unrivalled Torque Density. 2014; Available from:
http://www.magnomatics.com/Technology/pseudo‐direct‐drive.aspx.
71. InnWind. Innovative Wind Conversion Systems: 10 ‐ 20 MW For Offshore
Applications. 2013; Available from: http://www.innwind.eu/Publications.
72. Tesla Motor. About the Size of a Watermelon, with a Lot More Juice. 2014;
Available from: http://www.teslamotors.com/roadster/technology/motor.
73. Continental Corporation. Renault Electric Vehicles Equipped with Continental
Electric Motor. 2011; Available from: http://www.continental‐
corporation.com/www/pressportal_com_en/themes/press_releases/3_auto
motive_group/powertrain/press_releases/pr_20110913_elektromotor_en.ht
ml.
74. Vignaud, A. and H. Fennel, Efficient Electric Powertrain with Externally Excited
Synchronous Machine without Rare Earths Magnets Using the Example of the
Renault System Solution. 2012, Presented during Vienna Motor Symposium
2012, Austria.
75. Ohnsman, A. Toyota Readying Motors That Don't Use Rare Earths. 2011;
Available from: http://www.bloomberg.com/news/2011‐01‐14/toyota‐
readying‐electric‐motors‐that‐don‐t‐use‐rare‐earths.html.
76. Renault Nissan. When Less is More: Nissan Cuts Rare Earth Element Usage in
Electric Motors by 40 Per Cent. 2012; Available from: http://blog.alliance‐
renault‐nissan.com/blog/when‐less‐more‐nissan‐cuts‐rare‐earth‐element‐
usage‐electric‐motors‐40‐cent.
77. Redall, B. and J. Gordon. Analysis: Search for Rare Earth Substitutes Gathers
Pace. 2012; Available from: http://www.reuters.com/article/2012/06/22/us‐
rareearths‐alternatives‐idUSBRE85L0YB20120622.
78. Syrett, L. Rare Earths to Become Rarer in Electric Vehicles. 2012; Available from:
http://www.indmin.com/Article/3133551/Rare‐earths‐to‐become‐rarer‐in‐
electric‐vehicles.html.
79. Hitachi Ltd. Highly Efficient Industrial 11kW Permanent Magnet Synchronous
Motor without Rare‐earth Metals. 2012; Available from:
http://www.hitachi.com/New/cnews/120411.html.
80. Environment News Service. No Rare Earths in Next Generation Electric Vehicles.
2012.
81. Lee, J.Y., et al., Mm‐Ni Type Hydrogen Storage Alloy for Ni/MH Secondary Cell.
2000, Washington, DC: U.S. Patent and Trademark Office.
82. Toyota. Lithium‐Ion Battery. 2010; Available from:
http://www.toyota.com/esq/vehicles/batteries/lithium‐ion‐battery.html.

EXPLORING RARE EARTHS SUPPLY CONSTRAINTS FOR A
FULL GLOBAL PENETRATION OF RENEWABLE ENERGY:
CASE OF WIND TURBINES AND ELECTRIC VEHICLES
Komal HABIB1, Henrik WENZEL 2
1
Department of Chemical Engineering, Biotechnology and Environmental Technology,
University of Southern Denmark. Niels Bohrs Alle 1, 5230, Odense M, Denmark.
2
Department of Chemical Engineering, Biotechnology and Environmental Technology,
University of Southern Denmark. Niels Bohrs Alle 1, 5230, Odense M, Denmark.
Email1: koh@kbm.sdu.dk, Email2: henrik.wenzel@kbm.sdu.dk
Abstract
The dependency on critical resources like Rare Earth Elements (REEs) has been
pronounced as a potential barrier to a broader implementation of emerging renewable
energy technologies. This study explores the dependency of such technologies
especially wind turbines and electric vehicles along with other background end‐uses on
two key REEs, i.e. neodymium (Nd) and dysprosium (Dy). Our study reveals that a highly
accelerated rate of REEs mining is unavoidable in order to keep up with the pace of
increasing demand from new technologies required in a renewable energy strategy for
meeting the climate change challenge. Recycling does not seem to be in a position to
close the wide gap between future demand and supply by 2050 mainly due to the long
lifetime of key end‐use products.

Introduction
The Rare Earth Elements (REEs) consist of 15 elements from the Lanthanide series plus
two more elements namely scandium and yttrium. This group is the largest chemically
coherent group in the periodic table, and all the elements are found together in
geologic reserves. These elements have similar chemical characteristics, which on one
hand allow them to substitute for each other in their final application but on the other
hand make their separation and refinement to pure metal a very difficult process. REEs
have unique physical and chemical properties which make them highly attractive in
many of today’s hi‐tech applications e.g. permanent magnets having Neodymium (Nd)
and Dysprosium (Dy) are widely used in computers, audio systems, electric and hybrid
vehicles, cell phones, wind turbines, Magnetic Resonance Imaging (MRI) machines,
and others. The performance level provided by these magnets is currently unmatched
with any other available magnets, which makes these elements highly attractive for
the above mentioned technologies in order to get the same high level of performance
1‐6
.

The wind turbines and Electric Vehicles (EVs) are the two key emerging clean energy
technologies, which are considered highly promising in order to reduce our
dependency on fossil fuels as well as greenhouse gas (GHG) emissions. However, the
full global penetration of these clean energy technologies in future may be constrained
by the possible supply restrictions of REEs. REEs are considered as critical resources by
a number of studies, mainly because of the monopoly situation, i.e. what China
currently supplies more than 85% of the REEs production. The remaining is supplied
by Australia, USA, India, Brazil, Malaysia, and other countries. During the recent years,
China has reduced the export of REEs to the rest of the world in order to prioritize the
domestic demand, and to increase the production and export of high value goods using
REEs e.g. permanent magnets, motors, and batteries. Due to the introduction of export
quota by the Chinese Government, the prices of REEs skyrocketed in 2011. This sharp
increase in price led the stakeholders to seek some other solutions e.g. investing in
mines outside China, substituting REEs by some other elements and increasing the
recycling rates 7.

The current study addresses the issue of possible future supply constraints of Nd and
Dy for the emerging clean energy technologies as well as other potential end‐use
sectors. The aim of this study is to explore the risk of potential supply constraints
considering the predicted high demand of Nd and Dy and their forecasted supply by
2050. In order to reach the aim, an ultimate renewable energy scenario is developed
to show the future demand of Nd and Dy, not only for the wind turbines and EVs but
also for other end‐use sectors such as computers, audio systems, industrial motors,
and others. Simultaneously, a historical trend based projected scenario is developed
regarding the primary production of Nd and Dy up to 2050 in order to check if the
demand will outstrip supply. Moreover, different recycling efficiencies for various end‐
use sectors have been assumed to investigate the effect of secondary production on
the demand and supply of Nd and Dy.

Materials and Methods
Demand scenario

In order to reveal the potential supply constraints of REEs for the emerging clean
energy technologies, an ultimate renewable energy scenario is modelled and assessed
in this study. To estimate the demand of Nd and Dy by wind turbines, it is important
to model the global electricity demand by 2050, the share of electricity produced by
wind turbines and the market penetration rate of direct‐drive wind turbines as these
turbines are so far dependent on the strong permanent magnets containing Nd and
Dy. Table 1 shows the required share of wind power as well as the market share of
direct‐drive wind turbines in order to meet the ultimate renewable energy scenario by
2050. Table 2 presents the share of EVs sold by 2050.

Table 1: Overview of the Ultimate renewable energy scenario and the share of wind
power produced by direct‐drive wind turbines up to 2050 7

Scenario Total Share of Share of Wind Power Market
Electricity Renewable wind power demand penetration rate
demand energy (%) (%) 2050 (TWh) of direct‐drive
(TWh) wind turbines up
to 2050 (%)
Ultimate
Renewable 37,656 100 33 12,426 50
Energy

Table 2: Overview of EVs deployment in the Ultimate renewable energy scenario 7

Scenario Passenger Sales of Market Sales of EVs in
vehicles by passenger penetration rate 2050
2050 Vehicle in 2050 of EVs by 2050 (million/year)
(million) (million/year) (%)
Ultimate
Renewable 1800 170 100 170
Energy

For all the other end‐use sectors such as electric bicycles, computers, audio systems,
medical applications etc., an annual growth rate of 4% is considered until 2050.
Furthermore, the amount of Nd and Dy for direct‐drive wind turbines is considered to
be 150 and 14 Kg MW‐1 respectively. For EVs, the amount of Nd and Dy per vehicle is
0.62 Kg of Nd and 0.09 kg of Dy. In this study, the average lifetime is considered to be
20 years for wind turbines, 10 years for EVs and 10 years for all the other background
end‐uses 7.
Supply scenario
In the current study, both primary and secondary supply of Nd and Dy have been
modelled up to 2050, where primary supply refers to the mine production whereas
secondary supply refers to the supply of Nd and Dy from recycling of end‐of‐life
products. The primary supply of Nd and Dy has been projected following the historical
production trend of REEs from 1994 to 2014. The secondary supply has been modelled
assuming 90% recycling rate for wind turbines, 70% for EVs and 40% for other
background end‐uses at the end of their lifetimes (as mentioned earlier) by 2050.
Results and Discussion
Figure 1 presents the demand and supply (primary and secondary) comparison of Nd
and Dy by 2050. Results show that the annual total demand of Nd reaches from almost
16 Gg in 2007 to approximately 242 Gg in the ultimate renewable energy scenario by
2050. It can be noted that wind turbines do not seem to be the major end-user of Nd
compared to other end-use sectors especially EVs by 2050. Same trend can be
visualized in case of Dy, where annual total demand of Dy rises from nearly 1.8 Gg in
2007 to approximately 32 Gg in the ultimate renewable energy scenario by 2050. Here
again, EVs appear to be the major end-users of Dy followed by other end-uses.
As Figure 1 reveals, the Business As Usual (BAU) projected primary supply is not likely
to be enough to support the forecasted demand by 2050. In case of Nd, the gap between
the forecasted future demand and supply is 77% in the ultimate renewable energy
scenario. Moreover, the results clearly show that the linear trend based projection of Nd
mining is not likely to be enough to meet the increasing demand even in short (2020) to
medium term (2030). The same results can be visualized for Dy, where it becomes
evident that the BAU projected primary supply seems to meet only 16% of the projected
demand in the modelled ultimate renewable energy scenario by 2050.
Furthermore, Figure 1 shows the potential role of recycling in lowering the future
demand of Nd and Dy. It becomes clear that recycling seems to reduce the gap between
supply and demand from 77% to 55% for Nd by 2050. For Dy, it becomes evident that
recycling can help in reducing the gap between forecasted demand and supply from
84% to 60% in the modelled scenario by 2050. The reason why recycling is not likely
to meet a considerable share of demand by 2050 is because of longer lifetimes of
products, which delays the recovery of resources from the post-consumer applications.

Figure 1: Overview of the forecasted demand and the projected primary and
secondary supply of Nd (upper) and Dy (lower) in the Ultimate renewable energy
scenario by 2050, where the bars present demand and the slopes present the supply
of Nd and Dy 7.

Conclusion
From the results presented above it can be concluded that the historical trend based
projected primary supply of Nd and Dy is not likely to meet the forecasted high
demand of these resources by 2050. Recycling does not seem to play significant role
in reducing the gap between demand and supply of Nd and Dy by 2050, primarily due
to the long lifetimes of key end‐use products such as wind turbines and electric
vehicles. A highly accelerated rate of REEs mining is unavoidable in order to keep up
with the pace of increasing demand from new technologies required in a renewable
energy strategy for meeting the climate change challenge.

References
34. J. Speirs, Y. Houari, M. Contestabile, R. Gross and B. Gross, ‘’Materials Avaialability: Potential
constraints to the future low carbon economy; Working paper II: Batteries, magnets and
materials’’, UK Energy Research Centre, UK, 2013.
35. British Geological Survey, ‘’Rare Earth Elements’’, UK, 2011. Available at
http://www.bgs.ac.uk/downloads/start.cfm?id=1638
36. G.B. Haxel, J.B.Hedrick and G.J. Orris, ‘’Rare earth elements: critical resources for high technology’’.
US Department of the Interior, US Geological Survey, USA, 2002. Available at
http://pubs.usgs.gov/fs/2002/fs087‐02/fs087‐02.pdf
37. X.Y. Du and T.E. Graedel, ‘’Global Rare Earth In‐Use Stocks in NdFeB Permanent Magnets’’, Journal
of Industrial Ecology, 15:836‐43 (2011).
38. S. Hoenderdaal, L.T. Espinoza, F. Marscheider‐Weidemann and W. Graus, ‘’Can a dysprosium
shortage threaten green energy technologies?’’, Energy, 49:344‐55 (2013).
39. H. Kara, A. Chapman, T. Crichton, P. Willis and W. Morley, ‘’Lanthanide resources and
alternatives’’, Department for Transport and Department for Business, Innovation and Skills. 2010.
Available at
http://www.oakdenehollins.co.uk/pdf/lanthanide_resources_and_alternatives_may_2010.pdf
40. K. Habib and H. Wenzel, ‘’Exploring rare earths supply constraints for the emerging clean energy
technologies and the role of recycling’’, Journal of Cleaner Production, in press. DOI:
10.1016/j.jclepro.2014.04.035.

MATERIAL FLOW ANALYSIS FOR IDENTIFYING
NEODYMIUM FLOWS AND STOCKS AND RECYCLING
POTENTIALS IN THE EU‐28
Mariane PLANCHON1, Dominique GUYONNET2, Alain ROLLAT3, Victoire
ESCALON1, Stéphane VAXELAIRE2, Johann TUDURI2
1
BIO by Deloitte, 132 Avenue Charles de Gaulle, 92200 Neuilly sur Seine (F)
2
BRGM ‐ ENAG, 3 av. C. Guillemin, 45060 Orléans (F)
3
SOLVAY Group, 24 rue Chef de Baie, 17000 La Rochelle (F)
mplanchon@bio.deloitte.fr
Abstract
This paper addresses the material flow analysis (MFA) of Neodymium (Nd) in the EU‐
28. Due to their recycling potential, the Nd applications considered in this study are
permanent magnets and nickel‐metal hydride batteries. Flows and stocks of Nd are
considered all along the value chain within the EU, from the import of concentrates
containing Nd, Nd purification and metal production, manufacture of magnets and
batteries, use of products in the economy, and finally the management of end‐of‐life
products. The synthesis of this wealth of information in the form of Sankey diagrams
provides a holistic vision of the dynamics of Nd within the EU economy. An innovative
aspect of the project compared to existing MFAs is the incorporation of lithospheric
stocks of continental Europe in the analysis.

Introduction
On February 2011, the European Commission adopted a strategy document which sets
out targeted measures to secure and improve access to raw materials for the EU1.
Based on a three‐pillar approach, it aims at improving access to raw materials for
Europe by (1) fair and sustainable supply of raw materials from international markets,
(2) fostering sustainable supply within the EU, (3) boosting resource efficiency and
promoting recycling. In this context, the Commission identified 14 minerals as critical
for the economy 2 based on their risk of supply interruption. Rare Earth elements
(REEs) were among these identified critical raw materials.
A comprehensive data inventory of the material flows in industry and society is
essential to get the solid ground for informed discussion and decision making on the
supply of raw materials. Such data provide information on the quantity and quality of
raw materials in each stage or phase of the flow. This allows more decisions that
provide balanced, secure and sustainable supply throughout the entire material flows.
The on‐going ASTER project, started in 2012 for a period of three years, aims at
applying material flow analysis (MFA) to selected rare earth elements (REEs) for the
EU‐28. Certain rare earth elements are deemed “critical”3, as they combine
importance for strategic sectors of the economy with risks of supply shortage2, 4. The
recent JRC report3 identifies six critical REEs (Dy, Eu, Tb, Y, Pr and Nd) that are precisely
the REEs that were selected for the ASTER project.
While the paper by Guyonnet et al. (this conference) focuses on the REEs in phosphor
lamps (Tb, Eu, Y), this paper addresses REEs contained in permanent magnets and
NiMH batteries, i.e. Nd, Pr and Dy.
Description of the studied system
For the case of permanent magnets and NiMH batteries, the two systems that were
analysed are described schematically in Figures 1 and 2. These figures illustrate a
certain number of processes with, starting upstream in the value‐chain: separation (S)
of rare earth concentrates, fabrication (F) of NdFeB alloy (for permanent magnets) or
LaNiM5 mischmetal alloy (for NiMH batteries), manufacture (M) of permanent
magnets or NiMH batteries, insertion of these products in applications (A) such as cars,
computers, cell phones, wind turbines, etc. – only for permanent magnets, use (U) of
these applications in the economy, waste management (WM) of end‐of‐life products
and final elimination (landfill and/or dissipation in the environment). These various
processes are linked by arrows that represent flows from one to another. Differences
between flows into or out of a process imply an addition or subtraction to stock within
the process. The dashed line in Figure 1 and 2 represents the limits of the studied
system (the EU‐28). An additional “process” appears in Figure 1 and 2 indicated as “L”
(lithosphere). The lithosphere represents geological potentialities in the form of REE
resources (as opposed to reserves5). For geological and economic reasons, the limits
of this process were taken as continental Europe (including the Scandinavian shield
and the Kola Peninsula) plus Greenland.


Figure 1: Schematic illustration of the system investigated in the case of permanent
magnets

Figure 2: Schematic illustration of the system investigated in the case of NiMH
batteries

Flows were calculated for the year 2010, taken as the reference year of the study and
estimated over a number of years exceeding the expected average lifetime of the
considered products (taken as, e.g., 7 years for a NiMH portable battery, 3 years for a
laptop, and 13 years for a car), in order to obtain estimates of in‐use stocks and end‐
of‐life product flows.
Information sources
Information sources are the same as those described in Guyonnet et al. (this
conference). They include Eurostat trade and production data, as well as literature
data such as specialised reports on rare earth elements6, 7, USGS data8, company
reports on product sales, etc. Expert knowledge has been used to fulfill data gaps. For
the evaluation of geological potentialities in the lithosphere, over 350 deposits or
occurrences located within the EU‐28 were analysed9, considering mineralogical
characteristics of these deposits and occurrences.
Some results to‐date
As the ASTER project is still on‐going, the results presented herein are preliminary and
only describe qualitatively the Sankey diagrams elaborated for Nd in permanent
magnets and NiMH batteries.
Flows and stocks of Nd were calculated for each step of the value chain of permanent
magnets and NiMH batteries.
The method to assess the quantities of produced and traded (imported and exported)
rare earth oxide mixtures (REOs) in EU‐28, based on the customs data has been
described by Guyonnet et al. (this conference). With respect to the upstream portion
of the value chain, the same HS codes were used as in the Öko‐Institut report7 (see
Table 1 of Guyonnet al., this conference). GTA data for Chinese exports based on 8‐
digit HS codes were also used. These HS codes represent raw products that contain
mixtures of REEs. In order to obtain data on individual REEs, a disaggregation method
was applied10. The method considers the relative market shares of products containing
REEs (e.g., auto catalysts, magnets, batteries, polishing powders, etc.) and the
composition of the REE mixtures in each product.
Application of this disaggregation methodology to these data provides information
relative to flows F2, F3 and F4 of Figure 1, and F1, F2 and F3 of Figure 2. Expert
judgement used to check the consistency of the calculated flows with the industrial
reality was provided by SOLVAY, a major player of the separation industry in Europe.
In Europe, separation of Nd for magnet applications is performed by SILMET in Estonia.

The Fabrication step involves the production of the NdFeB alloy for permanent
magnets and the LaNiM5 mischmetal for NiMH batteries. The Manufacture step
involves the production of the NdFeB permanent magnets and the NiMH batteries.
Information relative to imports and exports of NdFeB alloy and NdFeB magnets is
obtained based on HS codes 85051100 (Permanent magnets of metal and articles
intended to become permanent magnets after magnetization, excl. Chucks, clamps and
similar holding devices) and 85051990 (Permanent magnets and articles intended to
become permanent magnets after magnetization, of materials other than metal or
agglomerated ferrite). Expert information was necessary to disaggregate the
quantities traded under these codes that regroup both alloys destined to permanent
magnet production and permanent magnets, the market shares of NdFeB magnets
within permanent magnets (in and out of the EU), and the typical composition of these
products.
Imports, production and exports of NiMH batteries were quantified on the basis of
Eurostat data for portable and industrial NiMH batteries (HS code 85078020: Nickel‐
hydrid accumulators; ProdCom code 31402383: Nickel‐hydrid accumulators) and on
literature data for NiMH batteries contained in electric and hybrid vehicles11.

Further down the value chain (Application step and Use step), data was collected for
manufactured products containing permanent magnets and NiMH batteries. The
following applications were considered concerning permanent magnets‐containing
products: electric and non‐electric car, laptop and desktop, cell phone, hard drive,
washing‐machine, refrigerator, air‐conditioner, scanner, fax and printer, shaver and
little electric equipment, camera, CD players, headphones and earphones, RMI, and
wind turbines. Information about production and trade of these products, as well as
Nd content and market shares was obtained in specialised or company reports6, 12, 13,
14, 15, 16, 17, 18
.
Results of the analysis suggest that the orders of magnitude of Nd flows in the EU‐28
related to permanent magnets and NiMH batteries are on the order of, resp., 700 and
120 tons entering the “Use” process. In contrast with the Sankey diagrams for Tb, Eu,
Y (Guyonnet et al., this conference), that are relatively “balanced” all along the value
chain, Sankey diagrams for Nd (and Dy) are very imbalanced due to large quantities in
imports to the “Applications” and “Use” processes. Taking into account the history of
flows into use in the EU‐28 over a period exceeding the expected average lifetimes of
the NiMH batteries and the permanent magnet‐containing products, in‐use stocks for
Nd in 2010 are estimated as, resp., 17,000 and 600 tons. The annual additions to in‐
use stock in 2010 are estimated to be resp., +190, and +50 tons/yr for these elements.
The flows related to end‐of‐life products (F17 in Figure 1 and F11 in Figure 2) are also
assessed using the expected average lifetimes of the NiMH batteries and the
permanent magnet‐containing products. The Nd in the waste flow is estimated as 530
tons due to permanent magnets application and 70 tons from battery applications.
Those values represent an upper bound of Nd in the waste flow as the market shares
or Nd content of the products arriving at end‐of‐life in 2010 are potentially lower (less
Nd in the composition) than the products used in 2010. With respect to flows F12 in
Figure 2 (recycling), there was no recycling of Nd in 2010 in the EU‐28. About 100 tons
of REEs from mischmetal are recycled each year but Nd is not part of these REEs. As
no recycling of Nd was performed in 2010, the Nd contained in the waste flow ends up
in landfills or in dissipative uses such as the cement or steel industry.

These figures can be compared to geological potentialities based on the analysis of
three important REE projects: the Norra Kärr project in Sweden and two exploration
projects in Greenland (Kvanefjeld and Kringlerne). The Norra Kärr project is well
advanced and has recently received the authorisation to operate from the Swedish
government. We estimate the resource of Nd at this site to be on the order of 33 kt.
The potentials of Kvanefjeld and especially Kringlerne are far superior (resp. 750 kt and
3,5 Mt of Nd), however, the likelihood of these projects entering into production is
much lower, due in particular to environmental constraints and social acceptability.
Conclusions
Results of the on‐going ASTER project provide estimates of flows and stocks of certain
REEs in the EU‐28, considered as critical. This paper addresses the results relative to
REEs in permanent magnets and NiMH batteries, while another paper (Guyonnet et
al., this conference) addresses Tb, Eu and Y. Detailed results of the ASTER project are
currently in the process of being published.

Criticality with respect to Nd has been somewhat relaxed in recent years, with the
diversification of supply sources provided by the Mount Weld deposit in Australia and
the Mountain Pass mine in the U.S. The preliminary results of the ASTER project show
that should one of the projects in continental Europe and Greenland enter the
production phase (with Norra Kärr as the most likely candidate), then criticality with
respect to Nd will be significantly decreased.

There is potentially more pressure on Dy. However, current efforts with respect to the
optimisation of Dy use in magnets and to substitution contribute to reducing such
pressure. Also, preliminary estimates of Dy resources for the three projects mentioned
above are resp., 12, 64, 710 kt Dy. This should help tackle the forecasted future
increase of Nd and Dy consumption due to the development of the European wind
turbine farms.

Acknowledgments
The material presented herein was developed within the ASTER project, supported by
the French National Research Agency; ANR (project ANR‐11‐ECOT‐002).
References
41. COM(2011) 25 final
42. DG ENTR (2010) Critical raw materials for the EU Report of the Ad‐hoc Working Group on defining
critical raw materials http://ec.europa.eu/enterprise/policies/raw‐materials/files/docs/report‐b_en.pdf
43. Moss, R., Tzimas, E., Willis, P., Arendorf, J., Thompson, P., Chapman, A., Morley, N., Sims, E.,
Bryson, R., Peason, J., Tercero‐Espinoza, L., Sartorius, C., Ostertag, K. Critical metals in the path
towards the decabonisation of the EU energy sector. Assessing rare metals as supply‐chain
bottlenecks in low‐carbon energy technologies. JRC Report EUR 25994 EN. 242 pp., 2013.
44. NRC. Minerals, Critical Minerals, and the U.S. Economy. U.S. National Research Council,
Washington, D.C., National Academies Press, 245 pp., 2008.
45. PERC. Pan‐European standard for reporting of exploration results, mineral resources and reserves.
Geological Service of Belgium, Brussels, Belgium, 2013.
46. ROSKILL. Rare Earths & Yttrium: Market Outlook to 2015. 14th edition. Roskill Information Services
Ltd, London, UK, 2011.
47. Schüler, D., Buchert, M., Liu, R., Dittrich, S., Merz, C. Study on rare earths and their recycling. Öko‐
Institut Report, Darmstadt, Germany, 140 pp., 2011.
48. USGS. Mineral commodity summaries 2013. U.S. Geological Survey, 198 pp., 2013.
49. Charles, N., Tuduri, J., Guyonnet, D., Melleton, J., Pourret, O. Rare earth elements in Europe and
Greenland: a geological potential? An overview. 12th meeting of the Society of Geology Applied
to Mineral Deposits (SGA), 12‐15 August 2013, Uppsala, Sweden, 2013.
50. Goonan, T.G. Rare earth elements – End use and recyclability. U.S. Geological Survey Scientific
Investigations Report 2011‐5094, 2011
51. EPBA Sustainability report, 2011
52. BIO Intelligence Service, Rhodia (2010) « Potentiel de recyclage des métaux rares »
53. INSEAD (2011) “Rare and Critical Metals as By‐products and the Implications for Future Supply”
54. Öko‐Institut (2012) “Recycling critical raw materials from waste electronic equipment”
55. Westphal L (2013) “Permanent magnets from small waste electrical and electronic equipment”.
Proceedings Sardinia 2013, 14th International Waste Management and Landfill Symposium.
56. W. Benecki, T. Clagett, S. Trout (2010) “Permanent Magnets 2010 – 2020 : A Comprehensive
Overview of the Global Permanent Magnet Industry”
57. Rademaker JH, Kleijn R & Yongxiang Y. (2013) Recycling as a Strategy against Rare Earth Element
Criticality: A Systemic Evaluation of the Potential Yield of NdFeB Magnet Recycling. Environmental
Science & Technology. 47: 10129−10136.
58. Hiroyuki Mikami (2012) Technologies to replace rare earth elements

Global Pricing Dynamics of the Rare Earths Industry
Europe
Allan Branch
Managing Director and CEO, Krucible Metals Ltd, 1/68 Railway Avenue, Railway Estate,
Townsville, Queensland, Australia
Allan.branch@yahoo.com.au
Abstract
This research examines the enigma of a market with critical global demand,
deliberately controlled geopolitical supply, yet prices which when not depressed are
reported as incomprehensible. Scientific interest and funding is dependent on rare
earths having a commercial market, so prices are of crucial importance. The results
show that regardless of how critical rare earths are, only 5% to 10% are used in defence
and aerospace applications, they are almost exclusively used in consumer products,
which implies strict price ceilings for components and raw materials. With China
supplying 87% of product and 60% of export demand coming from Japan, those two
countries essentially define the supply and demand market. Japan primarily
manufactures consumer products. The vast majority of demand therefore sits in what
is called an elastic market, and the solutions manufacturers use when faced with pricey
components in these markets are to find alternative suppliers or to eliminate the need
for the components all together. During the search by manufacturers for alternatives,
prices usually hike temporarily because it takes time to source alternatives or change
production, but then the price settles in a so called shift to a new equilibrium. These
dynamics are evident in the historical rare earths market.
Introduction
This economic research was first described on June and July, 2013 on Investointel.com
with the subtitle, “Am I the only one who thinks rare earth prices will always be low?
1
” The work was updated in June 2014 comparing predictions with subsequent events
and amongst others drew the attention of the USA Critical Materials Institute,
suggesting that it would be relevant to European strategies.
Elementary economics states that when a commodity is in crucial demand and supply
of that commodity is low or restricted, the price goes up. When entering the rare
earths sector in 2012 it was clear to the author that few participants had an answer as
to why rare earth prices were not high, and they lamented the fact in vocal
frustration2.
China has long intended for its essential monopoly of the rare earths elements to allow
control of global rare earth prices3. Yet with most of the world’s rare earth production
in China and their efforts to manipulate prices through rigorous control of supply,
prices continue to drop. China has done everything it can think of, stopping or limiting
production, eavesdrop on producers, eradicate pirating and refusing supply to some
clients. Every now and then there is a bump and everyone gets excited, but the trend
is clear. Rare earths go into energy, space and military sectors, but that accounts for
only about 10% of the demand. With China supplying 87% of product (down from as
much as 95% with the advent of Lynus and Molycorp) and with as much as 80% of
export demand coming from Japan, those two countries, for all intents and purposes,
define the supply and demand market.
Japan mainly manufactures consumer products. Regardless of how important rare
earths are, they are therefore used almost exclusively in cheap electronic
commodities, which have established retail prices and strict ceilings on the cost of
components and raw materials. The vast majority of demand therefore sits in what is
called an elastic market. An elastic market is one in which high commodity prices are
not tolerated. That is because the eventual products are themselves price sensitive,
usually consumer products for mass markets.
The two solutions manufacturers use when faced with pricey materials in these
markets, are to find cheaper suppliers or to eliminate the need for the material all
together. During the search by manufacturers for these alternative solutions, prices
do usually hike for a little while because it takes time to source alternatives or retool
production lines, but then the price settles back down again and there is a so called
shift to a new equilibrium between supply and demand.
Once this erroneous price manipulation fails, then the manipulator generally changes
tack attempting to retain or recapture market share. All of these dynamics are clearly
evident in the rare earths market3,4.
The results of this research suggest that the rare earths market is entering a final
phase, where prices will remain moderate but stable. Business models or projections
based on hoped for or sustained high prices are probably doomed, but those
companies presenting profits on current prices are probably safe and worth
investigating.
Rare Earths, Elastic Markets & Sheep
A global market can sometimes be thought of as a chess game, so when business plans
are based on only one side’s strategies, the surprises from the other side can be
devastating. Since chess is a game between only two players, a less limiting analogy is
a large set of simultaneous chess games, with each player constantly looking at other
players’ moves for inspiration and learning experience. So while China and Japan are
playing it out, so are India and Japan, and so are the USA and Canada and Australia and
Kazakhstan and so on. While China has focused on its game, the rest of the world has
been playing out its games.
The missing factor in much of this, in my opinion, has been the effect of the Global
Financial Crisis caused by the housing bubble. Not because of the crisis but because of
the bubble. The GFC is the most recent, most discussed, and most felt global financial
fiasco of modern times. I like the historical housing prices in Figure 1 because it shows
the trend lines which predict where the prices will return to once the bubble bursts
and the scatter settles down. When the historical rare earth prices chart is examined
and compared, it is clear that this has been a bubble too as seen in Figure 2
Figure 1: Collapse of the Housing Bubble in 2006 precipitating the Global Finance
Crisis.

Figure 2: Comparison of housing bubble of Figure 1 and the rare earth prices of 2011
show a distinct similarity suggesting the high rare earth prices were an anomaly and
not a new price level.
It is possible that some of the rare earths price hike of 2011 was caused directly or
indirectly by the mentality of the GFC. Many independent commentators at the time
like Jack Lifton5 and Constantine Karayannopoulus6 said as much. John Hykawy7 was
prescient very early in the piece directly linking the rare earth and housing bubbles.
Others like Roger Nusbaum8, Nick Kurtis9 and Mark Smith10 thought otherwise opining
it was a supply and demand thing. Some built their businesses on those opinions.
Figure 3 shows the stages of a modern global market bubble11. A bubble accelerates
out of all proportion, bursts, and then settles to where the trend line would naturally
come out. House prices skyrocketed, then collapsed below value, and have started to
stabilize by climbing back to their normal growth curves, based naturally on capital
appreciation and inflationary indices.
Figure 3: Stylised model of a financial or economic bubble. John‐Paul Rodrigue,
Department of Economics & Geography Hofstra University 200811.
The reason that rare earths “will remain moderate and stable” is twofold. Firstly,
historical prices have always been moderate. The very long term pricing chart as
shown in Figure 4 shows that they came into their own as a commodity with the
invention of the transistor, that there are little surges at times of technological
revolution (integrated circuit, microprocessor, internet, ipods), but generally have
followed what one would expect for an Elastic Market.
Figure 4: Long term historical rare earth prices.
Secondly, prices now are probably close to or below the trend line, seen in Figure 5.
There is perhaps an over compensation, so a dampened oscillation back to normal can
be anticipated. That is where I think we are now.
Figure 5: Rare Earth pricing bubble. Bloomberg.
And there is a subtle reason for the oscillation. It is based on the fundamental
difference between “Supply” and “Demand”. Demand can be instantaneous: a need,
a desire, an opportunity. It is a concept. Supply is a real‐world effect, based on physical
and practical factors. It takes time to plan production, build factories, develop
prototypes, make things. This lag causes interesting phenomena, such as the
temporary spike in the first place when manufacturers have no immediate options as
they search for alternatives. Afterwards the supply and demand equilibrium is shifted
since once a producer finds an alternative to an expensive component, they will rarely
go back, even if the price drops. A lesson for China to learn; may have in fact just learnt
it.
Supply and Demand is not a snapshot, it is a process in constant flux, always struggling
to maintain an equilibrium against the “noise”. Afterwards it will be easy to see what
has happened, but in the midst of upheaval, all is speculation.
Bubbles are caused by a mix of avarice and sheep like blind following. What must be
kept in mind is the sheer size of the world which in its modern mode is intolerant of
exploitation. Giant countries forget their power in a market place is affected by the
choices of billions of consumers and a couple hundred chess players.
Rare Earth Prices Will Do Exactly This
Since the advent of the rare earth pricing bubble, which really started to show towards
the end of 2010 ending abruptly at the middle of 2011 and which has been in a spiral
ever since, there has been speculation about the future prices of these critical metals.
Speculation has taken two basic forms: that which expects or hopes the prices will
return to their 2011 highs12,13,14 and that which simply despairs about the continued
decline in prices15. The rare observer played it safe16.
The Author’s article of June 7, 2013, “Rare Earths and Elastic Markets”17, essentially
dispelled the thought that prices would return to the 2011 highs arguing that those
prices were an anomaly. Being a doomsayer is all well and good, but helps no one. If it
is true that prices were a bubble, then what would be really helpful would be some
prediction of where the prices will finally settle, at what price and when. The author is
not the first to see the benefit of this18.
The first problem in this exercise is to define what one actually means by “rare earth
prices.” Even defining “rare earths” is complex enough let alone their prices, given that
they are really a disparate group and it is dangerous to bag them together. In fact there
is no such thing as a rare earth price. There are 17 rare earths, so there are 17 price
trends, and each element can be supplied in several salts, blended or metal form, each
with their own pricing trends. Even a so called basket price is dependent on whose
basket is being discussed.
Analysis showing rare earths are an elastic markets agreed with the general opinion
that at least some rare earths will return to stronger prices eventually. Which ones is
another problem to solve. It is not as simple as just separating the light rare earths
from the heavy rare earths which is the gut reaction19. In fact there are contradictory
definitions of HREE and LREE just to make it harder, and some reports refer to a middle
set which is not officially recognised by the US Geological Survey. And although the
HREE are generally the heavier ones from europium to lutetium they include one of
the lightest; yttrium. Nothing is straight forward.
Technically the definition of a heavy rare earth, (or a light one), is based on electron
configurations on an inner shell affecting its magnetic and orbital radius properties.
Because of the sequence as these electrons fill out, it does follow the atomic number
more or less, but it is more to do with what the chemical and physical properties are
like given the unusual electron configurations. So heavy rare earths do certain things
that light ones do not, such as having greater magnetic characteristics, or different
crystalline structures in hard metal alloys or silicon structures. Through other cosmic
forces, the odd numbered electron configurations make the heavies rarer on earth,
but it is the chemical and physical characteristics of the heavies that make them
important in today’s technology world and their difficulty in separation (not their
prevalence) that makes them expensive.
So relevant factors determining which rare earths will remain at modest prices and
which ones are currently undervalued are such things as type of use (commodity or
not), rarity or abundance, geopolitical importance, ease of preparation or separation,
advent of new processors, changing technology, and substitutability.
The objective here is to work out if some rare earths will soar in price and if so, which
ones, to what price and when. There are different ways to approach this problem, but
the method chosen is based on the belief that the 2011 prices were a bubble and that
bubbles tend to return to earth as an oscillation.

Figure 6: Damped Oscillation.
The mathematics of oscillation can be quite daunting, and using this model requires
estimates of the level of damping. Damping is the rate at which an oscillation settles,
like a pendulum coming to a stop by itself. High damping means it stops quickly and
slight damping means that it takes forever to come to a complete stop. Critical
damping is that unique situation where it comes to a stop smoothly, like a car halting
at a stop light. It is convenient sometimes to refer to critical, half‐critical and quarter‐
critical damping as seen in Figure 6.
It is instructive to compare the half‐critical damped curve in Figure 6 with the recovery
stage (return to the mean) of the generic economic bubble curve in Figure 3.
This suggests that the recovery from a bubble can be modelled approximately by a
damped oscillation. The peak price eventually caves in with everyone following it, until
it has over reacted and everyone brings it back, which is the reason for the reference
to sheep. What can rarely be determined is the level of damping, so the prices might
oscillate for a cycle or two, or if it returns to normal carefully, can be critically or even
over damped.
What causes damping in a physical system is a loss of energy, a result of resistance,
and there is a conceptual version of resistance happening in an economic bubble too.
Oscillations and bubbles start with a push. Eventually there is resistance to the
irrationality of the price when it is too high or too low, and things spring back to
normal. This is a meeting of human psychology and commercial rationalisation, and is
seen in the traditional supply and demand curves. This part of this research will suggest
when the price will settle.
The next thing to determine is the base line, and to discover if it is linear, exponential,
increasing, decreasing, or what. This part of the exercise will suggest at what price the
curve will settle and is akin to the dashed line in Figure 3.
Both of these tasks, “when” and “where”, require raw data and Metal‐Pages and
Gareth Hatch of Technology Metals Research, LLC have been generous enough to
provide the necessary data. An example of one heavy rare earth and one light:
europium and praseodymium respectively were chosen for reasons given below.
Figure 7 shows the Chinese FOB prices for these over the last 5 years. The difference
in price between a heavy and a light is clearly evident in this figure, which uses the
same price axis on the left.
While Eu prices seem to have averaged around $US500 per ton until early 2011 when
the bubble started, and then rose to above $5,500 at its peak in August 2011, they
have been clawing their way back down again to $883 in June 2013 when the analysis
was done and has been kept so that predictions made then can be compared to what
has happened since. Although at lower prices, Pr has done the same thing, averaging
something like $30 per ton before, peaking at $250 also in August 2011 and in July
2013 was at $74 per ton. Pr did seem to make a false earlier start to its bubble, starting
in August 2010.
6000 Monthly rare earth prices
5000 Data Courtesy Technology Metals, LLC
4000
3000
Pr FOB
2000
1000 Eu FOB
0
Mar 2012
Aug 2012
Jan 2013
Jun 2013
Oct 2011
Jun 2008
Jul 2010
Dec 2010
Nov 2008
Apr 2009
Sep 2009
Feb 2010
May 2011
Figure 7: Monthly Rare Earth Prices.
Considering just europium (Eu) to begin with, both linear and an exponential trend
lines seem to fit equally well, so both are shown in Figure 8 with no prediction as to
which trend line is the closer match. Also shown at left in Figure 8 is the curve for the
downward damped trend as the bubble has been bursting.
6000 Eu Trend lines pre bubble EU Eu FOB
5000
4000 Poly. (EU Eu
3000 FOB)
2000 Poly. (EU Eu
FOB)
1000
Linear (EU Eu
0
FOB)
1 7 131925313743495561

Figure 8: Rare Earth Pricing Predictions for Europium.
The reason for choosing europium is to see if non‐commodity rare earth elements
present a different pricing future compared with those used in commodities or
consumer products. Many rare earths have uses in both markets; while europium is a
toxic, reactive, rare element that has uses in a small number of applications including
some that use its nuclear properties. For that reason, it is likely to have a growth
market which is not strictly price sensitive, and therefore may have an exponential
trend line. The oscillation curve in Figure 8 suggests that the prices have bottomed and
should really be around $1,200 ‐ $1,300 per ton today instead of the $880 it currently
is. The maths (not shown) also suggests that if the growth trend is indeed exponential,
the price will bounce back to a respectable $2,000 within about 8 to 12 months. If the
trend is more linear, the data suggests the prices are almost back to normal, but will
continue to decrease for a similar period before returning to around $1,000 per ton in
less than a year. So that was the prediction for europium, that the price is undervalued
and will settle at between $1,200 and $2,000 per ton in about 8 months having
dropped below their value and are about to bounce back.
300 Praseodymium Pricing trends and
250 predictions
200
150 Pr FOB
100 Poly. (Pr FOB)
50 Linear (Pr FOB)
0
Mar 2012
Dec 2012
Sep 2013
Mar 2009
Dec 2009
Sep 2010
Jun 2011
Jun 2008
Figure 9: Rare Earth pricing bubble. Bloomberg.
The same exercise for Praseodymium in Figure 9 suggests that Pr prices were about
right in June 2013, that the prices had returning at a ‘critically damped” rate to what
they should be if the bubble had not occurred. This means that the price of Pr at $74
per ton then has bottomed out. The trend line equation (not shown) suggests the price
of Pr will double about every 4 years from now.
Praseodymium applications are concentrated in the consumer products industries,

glass works, filters, flints, and so on, and as such is more likely to follow a more linear
and tyicial supply and demand character.
Another reason for choosing these two elements is that even though they are classified
as one a heavy and the other a light, they are very close together in atomic number
(59 and 63). So it is the technical difference of electron configuration that is making
the difference in use and price more than the physical characteristics like weight,
melting point or density.
In August 2014 the price of europium is $1,000 per kgm and praseodymium is $150
per kgm (Figure 10) showing that more than a year after the original analysis and
forecast made in June 2013, the prices of these two rare earths is approximately as
predicted (allowing for a linear curve for europium).
Figure 10. Current prices for europium and praseodymium (Metal‐Pages) August
2014
References
1. Allan Branch, Investorintel, “Rare Earths and Elastic Markets,” June 7, 2013. “Rare earths, Elastic
Markets & Sheep,” June 9. 2013. “REE Prices Will do Exactly This,” July 15, 2013.
2. Tim Worstall, Forbes, “There’s something a little strange going on in China’s Production of the
various rare earth oxides and metals. Something I’m not sure that I understand properly,” Oct 27,
2012.
3. Editorial, Australian Newspaper, “As much as 50 per cent of China’s rare‐earths smelting capacity
has been idled in an effort to get prices back off the floor,” Nov 8, 2012.
4. Editorial, Financial Times, “China has very recently changed tack on its rare earths policy amid fears
that its hard line on producers threatens its very dominance of the global market which is being
eroded by growth in alternative supplies …” Nov 22, 2012).
5. Kevin Moore, “Is the Automotive Industry Driving Away from Rare Earths?” Metal‐Pages Rare
Earths North America Conference, New York June 2014.
6. Jack Lifton (private correspondence).
7. Constantine Karayannopoulus, (private correspondence).
8. John Hykawy, (private correspondence).
9. Roger Nusbaum, (private correspondence).
10. Nick Kurtis, (private correspondence).
11. Mark Smith, (private correspondence).
12. Rodrigue, J‐P; Comtois, C.; Slack, B. (2009). The Geography of Transport Systems (2nd ed.).
London: Routledge. ISBN 978‐0‐415‐48324‐7.
13. Tracy Weslosky, Investorintel, “They are back: Rare Earth Market turns around,” May 2012.
14. Alessandro Bruno, Investorintel, “Lynas’s Success in Malaysia could be catalyst for REE rebound,”
May 6 2013.
15. Alessandro Bruno, Investorintel, “The REE turnaround has arrived and this time it’s here to stay,”
July 11, 2013
16. HongPo’, Investorintel, “Forecasts of China’s Rare Earths Exports suggest an irreversible downward
Trend,” April 23 2013
17. Robin Bromby, Investorintel, “Tellurium – the newest critical metal? And an insight into Sri Lanka’s
unique graphite,” July 4, 2013.
18. Steve Mackowski’s, Investorintel, “Crystal Balls and the Rare Earths Future (2017 and 2022),”
September 2012.
19. Jennifer Getsinger, “Rare earth Elements and Rare Metals”, Resource World (August 2009)


Session IV: European Framework
EUROPEAN POLICIES ON CRITICAL RAW MATERIALS,
INCLUDING REE
Vasili NICOLETOPOULOS
Owner of Natural Resources GP, 11 Maiandrou St , Athens 115 28 , Greece
v.nicoletopoulos@naturalresources.gr
Abstract
This paper deals with the criticality of REE and other materials. 'Critical' raw materials
[CRM] have growing economic and technical importance, high risk of supply shortage,
low political‐economic stability of the main suppliers and low substitutability/recycling
rates. EU policies in CRM are studied, from the 2010 EC report to current monitoring of
CRM ‐‐ including the CRM list review and actions by Member States and stakeholders.
Specific EU policies on mining, processing, recycling, or substitution are presented: the
Raw Materials Initiative, the European Innovation Partnership on Raw Materials,
taxation & trade policies, R&D in substitutes/recycling and the EU‐US Free Trade
Agreement. Finally, Chinese CRM policies and the WTO actions on Chinese minerals,
including Mg/REE/W/Mo exports, are analyzed.

History of CRM in the EU

EC Definition of 'Critical Raw Materials’

Securing reliable and undistorted access of certain raw materials is of growing concern
within the EU and across the globe. CRM have a high economic importance to the EU
combined with a high risk associated with their supply1.
In the framework of the Raw Materials Initiative [RMI] on which see below, the first
European CRM analysis, published in 2010, identified 14 CRM from a candidate list of
41 non‐energy, non‐food materials2. The EC formally adopted this list in 2011 and
stated that it would continue to monitor the issue and update every 3 years1. Thus, in
May 2014, the EC published a revised and extended CRM list including 13 of the 14
materials from the previous list, with only tantalum moving out due to a lower supply
risk. Six new materials appeared: borates, chromium, coking coal, magnesite,
phosphate rock and Si metal, bringing the number up to 20 CRM. The other 14 raw
materials are: Sb, Be, Co, fluorspar, Ga, Ge, In, Mg, natural graphite, Nb, platinum
group metals, heavy REE, light REE and W. CRM are highlighted in the pink shaded zone
in Fig. 1 below.

Figure 1: The 2014 Criticality Chart
EU primary supply across all candidate materials is around 9%. CRM supply from EU
sources is even more limited. China is the most influential in terms of global supply.
Other countries have dominant supplies of specific raw materials, eg the USA [Be] and
Brazil [Nb]. For the platinum group metals and borates, supply is more diverse but still
relatively concentrated.

Raw Materials Initiative

RMI aims at providing the framework conditions to establish an integrated strategy to
face the challenges related to the access to non‐energy and non‐agricultural raw
materials and tries to facilitate a sustainable supply of raw materials. Following the
November ‘08 Communication that aimed at improving access to Raw Materials in
Europe on a 3‐pillar approach3: ensuring a level playing field in access to resources in
third countries, fostering sustainable supply of raw materials from European sources
and boosting resource efficiency and promoting recycling, in June 2013, the EC
adopted a Report4 on the implementation of the RMI, providing an overview of the
ongoing initiatives while highlighting the joint interest for the EU and third countries
rich in raw materials to work in partnership. EU member states were requested to
develop national strategies in support of the raw materials industry and of fostering
access to resources.

European Innovation Partnership on Raw Materials EIP]

As an RMI follow‐up, the EC launched in November ‘12 the EIP to reduce the EU's
dependency on imports of raw materials; to promote production and exports both by
improving supply conditions from within and outside the EU, and by providing
resource efficiency and alternatives in supply; and to bring Europe to the forefront in
the raw materials sector, whilst also mitigating the sector’s negative environmental,
social and health impacts, bringing together EU countries, companies, researchers and
NGOs5.
For example, new technologies can help to extract deeper, in more remote areas and
under harsh conditions. Action is also needed to develop CRM substitutes and to
improve recycling of, among others, the 17 kg electric and electronic equipment waste
that each EU citizen produces annually today5.
The current situation calls for targeted innovation and research efforts, breakthrough
technologies and multidisciplinary approaches, as well as demand‐side interventions
eg standards, public procurement, regulatory measures. Such efforts are also essential
for the manufacture of new and innovative products eg batteries for electric cars,
photovoltaic systems and devices for wind turbines5.
2. Other national CRM initiatives
Countries are placed in different supply chain stages. Japan is focusing heavily on
substitution, China on processing and metallurgy, Australia in sustainable mining and
Canada in exploration. S. Korea is investing $300m over 10 years for research into 40
technologies covering refining, smelting, processing, recycling and substitution. Russia
has an active program on materials stockpiles and export restrictions. China has
tightened export quotas ostensibly to secure internal supply. The US legislature has
long being studying7,7a CRM, the country had a stockpile for strategic defense materials
and recently China bought a quantity of MoO for the same purpose6.
In a broader context, concerns over conflict minerals arising from the DR Congo and
neighbouring states lead to materials stewardship schemes and legislation for
responsible sourcing and traceability of raw materials, eg for Co and Au the Extractive
Industries Transparency Initiative and the International Council on Mining and Metals
Materials Stewardship Scheme. Similarly, the US Dodd‐Frank Act requires electronics
companies to disclose their sources of cassiterite, wolframite and tantalum. Similar
regulation is now under consideration in the EU6.

REE Resource Efficiency and Recycling

End‐of‐life recycling rates for all of REE are below 1%, according a recent report by
UNEP. REEs are recycled from pre‐consumer scrap eg from permanent magnets. In
Japan approximately 1/3 of REE for permanent magnets went into new scraps and
then recycled, just like recovering the EU’s CRM, including REE, as per a recent study
that identified the following recovery opportunities6:
 Permanent magnets: hard disk drives are a current opportunity; wind turbines and
hybrid/electric vehicles present a long term potential at the end of their lifetime
 NiMH batteries: from both portable electronics and hybrid/electric vehicles if they can
be collected and identified. Umicore and Rhodia have now announced a jointly‐
developed process to recover REE concentrate that can then be refined into REE
materials.
 Phosphors: Rhodia has developed a new process for the recovery and separation of
REE contained in used fluorescent powder in lighting applications. identified the
following recovery opportunities6:

US: Strategic Stockpiling of CRM

The Strategic Mineral Advisory Council is encouraging the U.S. DoD to move away from
stockpiling strategic/CRM from China to creating and nurturing a U.S.‐based REE
supply chain7. DOD is in the process of identifying REE resources to stockpile. The
Strategic Materials Advisory Council is particularly concerned about the DOD
recommendation that $120.43 million of HREE, which are only produced in China, be
added to U.S. stockpiles. “The root cause of these material shortages is our ongoing
dependence on Chinese suppliers,” said council executive director Jeff Green. “We
urge DOD to move…to the only appropriate and permanent solution—the creation and
nurturing of a U.S.‐based REE supply chain…[the US] must not rely on research projects
and substitutions alone to close the current supply gap [and] must not accept the
status quo of material and technological dependence on China. Instead, [DoD should]
engage the industrial base domestically and from U.S. allies, to achieve sustainable
HREE development for defense and essential civilian requirements.’
Some analysts, however, note that creating a strategic CRM could have also negative
side effects eg insufficient flexibility, impact to the price of the commodity.
3. Foreign Trade Policies on CRM
Materials security and criticality has been of growing interest internationally. Several
countries, including suppliers and users of raw materials have instigated studies and
initiatives to develop national strategies for securing a stable supply of raw materials8.
A number of outreach activities were pursued in international fora such as G20 and
the OECD aiming at fostering a global dialogue on how to achieve a framework to
ensure a sustainable supply of raw materials.
EU trade policy will continue to monitor ongoing and future negotiations of Foreign
Trade Agreements and other bilateral frameworks to ensure that raw materials are
considered and integrated, where appropriate. Import data on the most critical raw
materials for wind energy [magnets in turbines], Nd and Dy, are collected. During
negotiations for the planned Transatlantic Trade and Investment Partnership in July
‘13 the EC said that raw materials are a key priority sector to the EU and that it wants
the US to ban export restrictions on selling raw materials and energy to the EU.

Industrial Minerals Trading
As well as blocking export restrictions, the Initial EU position paper9 says trans‐Atlantic
trade in these products could be promoted by increasing transparency and
predictability of trade deals. This would be “the first and most important step towards
a better [global] governance of trade in raw materials,” which, in turn, would boost
investment, aid production and promoting the functioning and expansion of related
infrastructure10.
Specifically, the EU wants the agreement to “encourage transparency in the process
of licensing”, including the allocation license conditions for trade and investment. It
wants negotiators to consider incorporating rules inspired by the global Extractive
Industry Transparency Initiative [EITI] on declaring payments to governments. This
Initial EU position paper9 stresses that “once exploitation is permitted [...] non‐
discriminatory access for resource exploitation, including for corresponding trade and
investment related opportunities, should be guaranteed by regulatory
commitments.”

US DoD Urges Assessment of REE Supply Risk July 11, '14
The agency in assessing REE supply risks for the US military industry, showing the
strategic importance of these minerals and the will of the US to limit its reliance on
China11.
4. Chinese CRM policies

Chinese Export License Systems: REE etc
China often uses trade‐restrictive measures eg export duties and quotas on key raw
materials, including Sb, bauxite/alumina/aluminium, Co, coke, Cu, fluorspar, In, Pb,
magnesite/magnesium carbonate, Mn, Mo, Ni, RE, Au, Sn, W, wood, yellow
phosphorus and Zn11. Other discriminatory barriers for the REE processing industry
include the unwritten pre‐condition of any company to have their request for a permit
evaluated, blocking new FDI in the sector12.
Since 2000, China applies production caps to regulate REE supply and instated an
export quota system, whereby Chinese companies (including foreign JVs) have to
obtain a licence to export any REE. Licences are attributed separately to domestic
producers/traders and to foreign JVs. As from 2007, China collects an export duty on
RE which depending on the REE varies between 15 and 25%. Since 2009, the export
quotas have been reduced significantly12.
The EU is also concerned about the flexible export tax policies in the Non Ferrous
Metals Stimulus Plan of 2009. The use of flexible export tax rebates clearly aims at
reducing exports of lower value products such as raw materials and this could be
interpreted as a government subsidy for high‐tech products with high added‐value11.

China and the CRM Value Chain

The report “Dominating the World, China and the Rare Earth Industry” by the National
Institute of Advanced Studies, Sept ’13 13 postulates that China achieved its REE
monopoly thanks to US policy oversights, including the permission to buy
Magnequench in 1995 that allowed China to become a global player in the value chain
for REM, alloys and magnets. Some members of Congress opposed the sale, pointing
out that the buyers were majority controlled by the Chinese government. Through
Magnequench, China gained access to the most important portfolio of REE patents in
the world, including REE magnets for cruise missiles and other advanced military
hardware. Coincidentally, ‘such components started to spread worldwide after the
transfer of Magnequench to China’.

Taxation
The Chinese MIIT is considering raising taxes on REE14. A counter‐argument raised by
the industry is that taxes create an uneven playing field for companies operating
legally, and paying the taxes, compared with illegal miners, also creating a big pricing
gap between legal and illegal enterprises.
Chinese resource taxes have a history of creating unfairness, owing in part to the
criteria on which they are implemented. China’s Ministry of Finance and State
Ministration of Taxation raised taxes on rare earth ores in 2011, but the impact of the
taxes differed greatly between provinces, due to the varying ratios of ore‐to‐processed
REE products. In June ’13 the MIIT sought opinions on whether to continue hiking the
resource tax on REE minerals, but could not reach a conclusion, but it seems each
domestic government is adopting different standards and the execution is confusing.
REE Recycling – and Smuggling…

Illegal trading practices are often used in China to circumvent direct control of
important secondary raw material flows, including recycled REE. Preliminary statistics
of June ’14 show illegal smuggling hidden among 67 REE recycling projects across the
country. MIIT urged the government to regulate REE recycling15.

Subsidising REE Environmental and Efficiency Upgrades

China government’s decision to subsidise REE producers pursuing technological
upgrades, shows the commitment of the country to address energy‐efficiency and
environmental issue and to promote value‐added products16.
5. WTO
The first dispute settlement case on several industrial raw materials was initiated in
June ‘09 by the EU, the US and Mexico12. No amicable solution with China was found,
so a Panel was established in Dec ’09 and upheld most of the EU’s claims. China
appealed certain aspects but the Appellate Body upheld all the key claims raised by
the EU, and confirmed the findings made by the Panel, including [a] that China’s export
restrictions on several industrial raw materials are in breach of WTO rules and are not
justified for environmental protection or conservation policy reasons and [b] that
China had committed not to levy export duties. For procedural reasons, the Body could
not rule on additional claims regarding technical aspects of China’s administration of
export quotas and licences and to a minimum export price. Preliminary analysis
indicates that China has taken steps to comply with the ruling but in practical terms
there are still significant barriers, eg minimum export prices seem to still be in place.
In March ‘12, the US, the EU and Japan initiated a second complaint with respect to
China’s restrictions on the export of various forms of REE, W and Mo used in the
production of electronic goods17. These restrictions include export duties, export
quotas, minimum export price requirements, export licensing requirements and
additional requirements and procedures in connection with the administration of the
quantitative restrictions. A panel established with key findings that China imposes
three export restrictions: export duties (taxes); export quotas on the amount of those
materials that can be exported in a given period; limitations on the enterprises
permitted to export the materials ‐‐ all inconsistent with China’s WTO obligations.
On 17 April 2014, China notified of its decision to appeal certain issues. At the same
time, a very interesting suggestion appeared in a China Daily report : ‘Beijing should
try and get around the WTO ruling by offering to remove the export quotas on LREE
but keep ones on medium and HREE, which have smaller production volumes’17.
The Ministry of Commerce says the Chinese position is really just about protecting
resources18 and the environment for sustainable development. But the next sentence
is something more of a stretch: ‘It has no intention of favoring domestic industry
through distorted trade’.
Gao Wei, general manager of Shanghai Huaming Gona Rare Earth New Materials,
welcomed the removal of the current tariff and quota systems that had weakened
Chinese REE competitiveness in terms of pricing and Chinese products are priced
higher than their foreign rivals because of the tariff and quota systems and said that
‘Lifting the quota and reducing tariffs will help domestic producers digest their massive
volumes of REE and bolster the production of medium and HREE’.
The president of the US National Mining Association, welcomed news of the WTO
ruling but pointed out ‘this was no panacea to solving the US’s mineral supply chain
woes. American companies were still 100% reliant on China for supplies of REE and
other minerals [and should] develop local resources and also look downstream, in
Africa, Australia, Canada and elsewhere — after all, it was a processor once upon a
time and could be so again18…So far, many REE projects around the globe have secured
financial backing from Chinese, S. Korean or Japanese interests. Japan is clearly aiming
to end its reliance on China for REE supplies for its downstream industries. Perhaps
rather than joining the long line of exploration companies looking for REE deposits, the
strategy for America would be to provide the downstream money and technology’18.

Chinese Retaliation?

Chinese authorities may deal with the WTO by imposing higher taxes on producers of
REE on the basis of the value of minerals, rather than the volume as is the case now.
This change will result in higher production and, of course, export prices. But China’s
revenge could be longer‐term and designed to cause deeper damage19.
6. What next?
The EC report from on the Implementation of the Raw Materials Initiative’ of June 24,
’134 stated that ‘a wide‐ranging public consultation on the strategy was going well; the
majority of EU member states had responded…along with non‐EU countries, regional
governments and the industrial minerals industry’.
The report confirmed that by 2020 actions would include: setting up innovative pilot
projects, including demonstration plants on exploration, mining and processing;
finding substitutes for raw materials in short supply; creating a network of research
centers on sustainable raw materials management; using EU‐standardized
instruments for surveying resources; creating a 3D geological map; modelling raw
material trends; and undertaking complete lifecycle assessment of EU raw materials
needs.
However, Roskill20 point out some disadvantages of the report, ‘hence it should be
read with caution’:
 “Terms like ‘risk’ and ‘critical’ are emotive,” and the end result could be panic and

‘unnecessary substitution
 Considering “the wider supply chain dynamics” Roskill provides a case study of cobalt
as CRM. Although it is right to highlight the supply risk situation in the DRC, “a volatile
state does not necessarily result in volatile levels of raw material exports.” Indeed,
from 1996 to 2003, when main periods of conflict occurred, mine production
increased at a 28 pct CAGR
 Metals and oil make up 95 pct of DRC’s total export revenue, with mining and
quarrying accounting for ~12 pct of its GDP, hence DRC is unlikely to put policies in
place that would negatively impact those industries.
 DRC is not the sole determinant of Co supply. For instance, 2/3 of Co is produced as a
by‐product of Ni and Cu, meaning that demand for either of those metals has the
ability to impact Co supply.

On the specific issue of recycling, the following comments were made by Zboril21.
Regulations on recycling have not been framed coherently and tend to focus on
individual, isolated aspects of collection and recycling and do not take account of the
market forces at work in the systems and processes. The conflict between market
forces and the existing regulatory framework should be analyzed in detail to achieve
better balanced results. The EU should possibly negotiate emergency terms with the
WTO, setting clear and transparent conditions for export restrictions/duties on wastes
of strategic importance. Recycling should be supported by improving collecting
infrastructure, creating legal certainty and an equal level playing field and by removing
unnecessary administrative burdens. Lastly, all individual elements of the EU Climate
Change Policy should be the result of consistent energy, raw material and industrial
policies.
References
1. Defining 'critical' raw materials, European Commission, Accessed July 15, 2014.
2. “Rare Earths Worldwide: An Industry and Policy Analysis”, Natural Resources GP study, Greece,
2011
3. Raw materials, IMA Europe, Accessed July 15, 2014.
4. Report from the Commission to the European Parliament, the Council, the European Economic and
Social Committee and the Committee of the regions on the Implementation of the Raw Materials
Initiative, June 24, 2014.
5. About the European Innovation Partnership (EIP) on Raw Materials, European Commission,
Accessed July 15, 2014.
6. A. Chapman, et. Al., L. T. Espinoza et. Al., “Study on Critical Raw Materials at EU Level, Critical Raw
Material Profiles”, Oakdene Hollins, Fraunhofer ISI, December 2013.
7. D. Kosich, “Strategic Mineral Advisory Council Calls for US REE Supply Chain”, Mineweb, March 25,
2013.
7a. R. Silberglitt, et al., “Critical Materials Present Danger to U.S. Manufacturing”, RAND, 2013.
8. A. Chapman, et. Al., L. T. Espinoza et. Al., “Study on Critical Raw Materials at EU Level, Final
Report”, Oakdene Hollins, Fraunhofer ISI, December 2013.
9. “EU ‐ US Transatlantic Trade and Investment Partnership, Raw materials and Energy, Initial EU
position paper”, European Commission, July 2013.
10. “REE: Light to Heavy?’”, Natural Resources GP study, 2014.
11. A. Torrisi, “US Department of Defense urges assessment of rare earths supply risk”, Industrial
Minerals Magazine, July 11, 2014.
12. “Raw Materials ‐ export restrictions”, European Commission, Creation Date: 01 Jul 2009, Last
update/check February 24, 2014.
13. N. Mancheri et al., “Dominating the World: China and the Rare Earth Industry”, International
Strategy & Security Studies Programme (ISSSP), National Institute of Advanced Studies, April 2013.
14. A. Li Im, “More Chinese Rare Earth Firms Meet Ministry Conditions”, Industrial Minerals Magazine,
June 19, 2013.
15. A. Torrisi, “China’s Ministry of Industry and Information Technology calls for regulations on rare
earths recycling”, Industrial Minerals Magazine, June 13, 2014.
16. A. Torrisi, “China government to subsidise environmental and efficiency upgrades of rare earths
companies”, Industrial Minerals Magazine, July 10, 2014.
17. China — Measures Related to the Exportation of Rare Earths, Tungsten and Molybdenum, World
Trade Association, April 8, 2014.
18. R. Bromby, “China’s REE Export Policies and America’s Needs‐ is it Time for a Total Rethink”,
Investorintel, November 2013.
19. A.Bruno, “China pondering retaliation to recent US and WTO moves”, InvestorIntel, May 2014.
20. C.McLeod, “New EU Critical Raw Materials Report Has “Disadvantages”: Roskill”, Cobalt Investing
News, June 10, 2014.
21. J. Zboril, “Raw Material Polic”, EESC, Raw Material Conference, Athens, June 20, 2014.
22. V. Nicoletopoulos, “Recent Technology and Policy Developments in Critical Materials CRM, with
REE emphasis”, presented at Federchimica Workshops on Rare Earths; Eco‐Innovation, Milano,
Italy, November 20, 2013.

Legislation and best practice in the emerging European
rare earth element industry
Miranda Keith‐Roach1, Bertil Grundfelt1, Lars Olof Höglund1, Anne
Kousa2, Esa Pohjolainen2, Paloma Magistrati3, Vassiliki Aggelatou4,
Nicolò Olivieri5 , Andrea Ferrari5
1
Kemakta Konsult AB, Sweden
2
Geological Survey of Finland, Finland
3
Fen Minerals A/S, Norway
4
Institute of Geology & Mineral Exploration, Greece
5
D’Appolonia S.p.A, Italy
miranda@kemakta.se, bertil@kemakta.se, anne.kousa@gtk.fi,
esa.pohjolainen@gtk.fi, paloma@bergfald.no, vasaggelatou@igme.gr,
nicolo.olivieri@dappolonia.it, andrea.ferrari@dappolonia.it
Abstract
This paper explores the environmental legislation and best available technique
reference documents in place in the European Union to support the emerging European
rare earth element (REE) industry. One of the issues for REE mining, beneficiation and
processing is the radioactivity that can be associated with the ores, arising from the
presence of elevated U and/or Th and their daughter products. Therefore, workers may
need to be protected from radiation during the exploitation of REE resources, and the
public protected by suitable disposal of the radioactive wastes produced. Potential
regulatory gaps and best practices are identified, based on the experience of countries
with a REE industry.

Introduction
Rare Earth Element (REE) mining, processing and exploitation are large scale industries
that use a wide range of chemical substances and generate significant quantities of
waste. Additionally, the ores contain variable amounts of impurities such as non‐target
toxic metals, fluorine and radionuclides that may be released from the ore during
processing into the product or waste streams, and/or represent safety issues to the
workers. The most significant environmental impact of mining is often on the surface‐
and groundwater quality as seen at Bayan Obo, China1, Mountain Pass, USA2, and a
processing plant in Sichuan, China1. The large volumes of tailings that are generated
during hard rock mining are a source of waterborne contamination. Tailings are also a
major environmental hazard if the tailings impoundment dam fails.

Dust and gaseous emissions are also important vectors for both environmental
contamination and the exposure of workers to toxic substances. For example,
processing at Bayan Obo has led to the release of fluoride, dust and waste gases
containing HF and SO21.

Past REE mining and processing has therefore led to significant environmental impacts
in several non‐EU countries, including Brazil, China, Malaysia and the USA. In addition
to the environmental damage caused, remediation of contaminated sites can be
expensive. For example, the estimated cost of remediating the Mitsubishi REE
processing site in Bukit Merah, Malaysia is $100 million3. It is therefore important to
ensure that the regulatory framework in the EU will support the development of a well
managed REE industry with acceptably low environmental impacts. Since mining and
industrial processing are established industries in the EU, there is a body of relevant
legislation that could be expected to cover the REE industry. Best practice is described
in best available technique reference documents (BREF) for the management of
tailings and waste‐rock in mining activities4 and the non ferrous metals industries5.
However, although generic environmental protection methodologies are included in
these BREF, specific issues relating to REE tailings and processing are not considered
specifically. Therefore, in this paper, we explore the EU legislation and guidelines that
are in place to support this emerging industry and identify potential gaps.

Naturally Occurring Radioactive Materials (NORM)
REE ores often contain sufficient natural radioactivity to be considered naturally
occurring radioactive materials (NORM). The radioactivity arises from U and/or Th and
their daughter products (Figure 1) that are either co‐mineralised with the REE or
present in other minerals within the ore. It is therefore a safety issue at mining sites
and during processing: a significant relationship has been found between the
inhalation of Th‐containing dusts and lung cancer in miners in a 20 year study at Bayan
Obo6; two REE processing plants and a waste facility in Brazil became contaminated
with Th and its daughters7; and workers have received elevated doses in the REE
industry worldwide8.

The separation of the different radionuclides in the decay chain from the REE during
beneficiation and processing occurs at a point determined by the chemistry and
physicochemical associations in the ore. Given that there are 8 elements within the
238
U decay chain (Figure 1) and 8 in the 232Th chain, their separation during processing
is not simple and it is important to ensure that the final REE products are below the
NORM activity level (1 Bq/g of a single radionuclide in the decay chain). Examples of
radionuclide partitioning during REE processing are given by the IAEA8.


Figure 1: Radioactive decay chain of 238U (a = years; d = days, h = hours, m = minutes,
s = seconds)

Analysis of the Regulation of the REE Industry
In this section, EU legislation is compared with legislation in states with an existing REE
industry or experience with NORM, focussing on USA and Western Australia.

Regulatory Framework
Environmental regulation and implementation of the regulations are clearly important
for any potentially polluting industry, but are not without their challenges. The 11 km2
tailings impoundment at Bayan Obo, China has contaminated the surrounding area1,
and provides an extreme example of the environmental consequences of an
unregulated REE industry. Introducing regulation in China has also been difficult
because of the large number of small companies and a significant level of illegal
mining. Forced mergers have consolidated the industry, but the system is still reliant
on a level of self‐regulation. Similarly, although Russia’s legislation in many cases
meets or exceeds commonly accepted international standards, its enforcement has
been uneven9. India has also experienced challenges trying to eliminate illegal mining,
in this case due to easy access to REE‐containing beach sands. Europe’s existing mining
industries mean that legislation is in place and in use. The Mining Waste Directive
(2006/21/EC) is a key piece of legislation, and links with the Water Framework
(2000/60/EC) and Groundwater (2006/118/EC) Directives for the management of
mine water. The Industrial Emissions Directive (IED; 2010/75/EU) applies to the
beneficiation and processing of the REE ores, and also ties in with the Water
Framework Directive and Groundwater Directive. Additionally, processing wastes are
controlled by the Waste Framework Directive (2008/98/EC) and Landfill Directive
(1999/31/EC). The Basic Safety Standards (2013/59/EURATOM) address the
management of NORM.

Penalties are needed for non‐compliance with the regulations and the “polluter pays”
principle is embedded in legislation in the EU (Environmental Liability Directive
(2004/35/EC)), in the USA (Comprehensive Environmental Response, Compensation
and Liability Act (CERCLA)) and Western Australia (Contaminated Sites Act (2003)). The
Contaminated Sites Act (2003) also states the penalties associated with non‐
compliance, including a daily penalty to discourage delay. Environmental quality
standards also support industrial regulation. In the EU, the Water Framework and
Groundwater Directives require threshold values to be set for only a limited list of
substances, with only some relevance to the REE industry. The threshold values set
also differ between member states. In 2011, China brought in specific pollutant
discharge standards for the rare earth industry10.

Environmental Impact Assessment (EIA)
Environmental impact assessment prior to the onset of activities is an important
process for minimising the environmental risks associated with mining. The
Environmental Impact Assessment Directive (2011/92/EU) applies to for EU mine sites
with a surface area greater than 25 hectares. There is also a need for appropriate EIA
if a proposed mine may impact a Natura 2000 site, as defined by the Habitats Directive
(Directive 92/43/EEC) and Birds Directive (2009/147/EC). The definition of when an
EIA is required varies between countries, but EIA is needed in the USA, Russia and
Western Australia. Additionally, China is introducing an environmental risk assessment
system for the REE industry10.

Tailings management legislation
The tailings arising from REE mining are large‐scale wastes with a significant pollution
potential. The Mining Waste Directive (2006/21/EC) requires all EU waste facilities to
have a permit, and the application for this includes a waste management plan for
minimizing environmental impacts. Facilities with a significant accident hazard also
require an emergency plan to be drawn up by the competent authority. Additionally,
the operator must provide a financial guarantee before operations start to provide a
resource for restoring the waste facilities. The operator must maintain the site until
the competent authority approves site restoration and closure, and then the operator
must maintain and monitor the site for as long as the competent authority considers
necessary. This is similar to legislation in Western Australia and the USA. In the USA,
wastes that are “uniquely associated” with mining and processing are excluded from
being regulated as hazardous wastes (Bevill amendment), and these are instead
treated as non hazardous solid wastes. The Contaminated Sites Act (2003) in Western
Australia introduces the full life‐cycle costs principle, which ensures that the cost of
effective waste disposal is included in the cost of the products. Western Australia also
has specific guidelines on safe design and operating standards for tailings storage11,
which play a similar role to the BREF for the management of tailings and waste rock in
mining activities in the EU4. However, Western Australia also provides specific
guidelines for the handling of the NORM mining waste12, which is not the case in the
EU.

NORM legislation
The explicit inclusion of NORM in the latest versions of the IAEA Basic Safety Standards
(2011) and the EU Basic Safety Standards (2013/59/EURATOM) will support more
consistent regulation of NORM. However, the actual regulation of the handling and
disposal of NORM wastes differ between countries and EU member states. The EU
Basic Safety Standards state that “activities in industries processing materials with
naturally occurring radionuclides, or activities related to such processing” are within
its scope, suggesting that REE mining and processing are relevant. This is already the
case in the UK, where REE mining and processing are specifically identified as a NORM
industry that falls under their national regulations. The new BSS also address worker
exposure to radon explicitly, which will be important for both worker exposure and
waste management in the REE industry.

Countries with large scale NORM industries have developed their regulations of NORM
wastes to a higher level than others, often taking lead from the UN, ICRP and IAEA.
Norway has integrated its legislation for activities that involve or may involve
radioactive pollution or radioactive waste management into its 2011 Pollution Control
Act, which controls their permit system13. A repository has also been built for the
disposal of NORM wastes from the oil industry with > 10 Bq g‐1 of 226Ra, 228Ra or 210Po.
The repository is operated by a private company, and the owners are required to have
a fund for closure and remediation. The state has also guaranteed to manage the site
if the company is no longer able.

In Australia, radioactivity is regulated by the Radiation Protection and Control (Ionising
Radiation) Regulations (2000). However, since these regulations define ‘radioactive
ores’ as those with a specific activity greater than 35 Bq g‐1, the mining, processing and
waste disposal of lower activity ores are regulated through state‐level environmental
and mining laws. The Australian Radiation Protection and Nuclear Safety Agency
(ARPANSA) provides The Code of Practice and Safety Guide: Radiation Protection and
Radioactive Waste Management in Mining and Minerals Processing14 to support the
legislation, and this gives prescriptive practice‐specific radiation safety requirements.
Additionally, Western Australia has extensive guidelines for managing NORM in the
mining and processing industry12. A Radiation Management Plan and a Radioactive
Waste Management Plan must be written and approved prior to any work taking place
that involves NORM14. The depth and level of detail of the Radioactive Waste
Management Plan should reflect the amount and activity of the wastes, as well as the
degree of processing, as this can affect the ease with which radionuclides leach. The
dilution of NORM tailings (prior to any chemical treatment that breaks secular
equilibrium) with lower activity waste from the same site is encouraged to reduce the
overall activity concentration to below the regulatory limit. Dilution is also allowed for
materials to be used in other activities such as road building12. Potentially valuable
tailings can be stored in tailings dams if adequate safeguards are in place12.

Management of the mine site
As well as requiring management during operation, mines also require stabilisation
and monitoring during idle periods, which typically occur when the market price of the
commodity falls, and long term post closure management. Failure to stabilise sites
effectively can lead to unauthorised discharges, as seen at the Pea Ridge mine and the
South Maybe Canyon mine site, USA2, both of which have REE‐rich ores. The EU does
not have direct legislation for mining, but uses a variety of legislation including the
Mining Waste Directive (2006/21/EC) discussed above. This influences mining permits
since tailings are often disposed of at the mine site. It is strengthened by Seveso III
(96/82/EC), which addresses operator responsibility to prevent major accidents, with
disused mines as well as disused and operational tailings disposal facilities being within
the scope of the directive. However, given the legislative approach in the EU, aspects
of mine management are not addressed explicitly and, as a result, there are
differences between the regulations in different member states. A different legislative
approach in the USA (Surface Mining and Control Act of 1977) gives more direct control
of the mine, including idle periods and post closure. The waste rock piles, the tailings
ponds, and other mine areas must be stabilized and managed during idle periods, and
restored to its previous condition on closure.

Site management in the case of bankruptcy is another important issue, given the large
sums of money involved in waste management and site remediation. The USA has a
“superfund” (legislated in CERCLA) generated by a tax on chemical and oil companies,
which can pay for the Environmental Protection Agency to clean up hazardous
abandoned sites and respond to short term emergencies. In Western Australia, the
Director of an insolvent company may be liable for the costs of remediation if the
insolvency is linked to avoidance of responsibility for site remediation. Ultimately, if
the director is also bankrupt or not judged to be liable, the state takes responsibility.

There is currently a discussion of a disaster risk‐sharing fund in the EU to cover large
scale industrial accidents (damage exceeding €100 million), funded by a mandatory
insurance premium of a percentage of the annual net sales. A report prepared for the
EC15 raises questions relating to this. The fund could not subsidise operators or it would
be in breach of the polluter pays principle, but if the funds are to be re‐paid, there is
the question of whether liability should be capped. Different industries have different
levels of risk and there are differences in the extent to which environ‐mental damage
is covered by existing private insurance in different member states.

The Non‐Extractive Industry Panel suggests that the fund is unnecessary in the EU,
given the stringency of the Environmental Impact Assessment, Mining Waste, Seveso
III, and Environmental Liability Directives, and the financial controls they define16. They
argue that adherence to the EU law should make the fund unnecessary, and that the
fund may encourage low standards in industry. However, environmental protection
against large scale accidents does need consideration, as does the management of
sites and contaminated land that are no longer under ownership. One issue here is
licensing of subsidiaries, whose financial losses are not covered by the parent
company. Equally, the financial guarantees associated with the Mining Waste Directive
may only cover a relatively small part of the overall costs.

Conclusions
The regulations to support REE mining and processing industries in Europe generally
compare favourably with international standards. The main issues that have been
identified here are:
 several different pieces of environmental legislation control aspects of REE
exploitation in the EU, which emphasizes the need for focused guidance
documents
 the next updates of the relevant BREF documents4,5 should therefore include
REE‐specific information, including the management of NORM
 the current environmental quality standards should be reviewed to encompass
the main hazards associated with REE mining and processing
 there is a lack of direct legislation for mine management, which may lead to
differences between the regulations in different member states
 REE NORM waste management regulation is likely to differ between member
states
 the question of funding for remediation of abandoned sites is relevant to REE
mining
Acknowledgements
Community’s Seventh Framework Programme ([FP7/2007‐2013]) under grant
agreement n°309373. This publication reflects only the authors’ view, exempting the
Community from any liability”. Project web site: www.eurare.eu”
References
1. D. Schüler., M. Buchert., R. Liu., S. Dittrich and C.Merz. Study on rare earths and their recycling.
Final report for the Greens/EFA Group in the European Parliament. Öko Institut, 2011.
http://www.oeko.de/oekodoc/1112/2011‐003‐en.pdf. [accessed 14/05/2014].
2. EPA, Rare Earth Elements: A Review of Production, Processing, Recycling, and Associated
Environmental Issues. EPA/600/R‐12/572, U.S. Environmental Protection Agency, 2012.
3. Bradsher K, Mitsubishi Quietly Cleans Up Its Former Refinery, New York Times, 2011
http://www.nytimes.com/2011/03/09/business/energy‐environment/09rareside.html?_r=0.
[accessed 14/05/2014]
4. EC, Reference document on the best available techniques for management of tailings and
waste‐rock in mining activities. European Commission, Brussels, 2009
5. EC, Reference document on the best available techniques in the non ferrous metals industries.
European Commission, Brussels, 2001
6. X.A Chen., Y.G. Chen, H.J. Xiao, G.D. Feng, Y.H.Deng, Z.L. Feng, L.A. Chen, X.M. Han, Y.J. Yang.,
Z.H Dong and R. Zhen, “20 year follow‐up study on the effects of long‐term exposure to thorium
dust”, Chinese Medical Journal 116, 692‐694 (2003).
7. D.C. Lauria and E.R.R Rochedo, ”The legacy of monazite processing in Brazil”,. Radiation
Protection Dosimetry 114, 546‐550 (2005).
8. International Atomic Energy Agency, Radiation protection and NORM residue management in
the production of rare earths from thorium containing minerals. Safety Report Series No 68.
International Atomic Energy Agency, Vienna, Austria, 2011.
9. J.A. Josefson, Overview Russian Environmental Regulation, King & Spalding, Moscow, Russia,
2012.
http://www.kslaw.com/imageserver/KSPublic/library/publication/RussianEnvironmental.pdf
[accessed 09/10/2013].
10. Chinese Government’s Official Web Portal, Full Text: Situation and Policies of China's Rare Earth
Industry, 2012 http://www.gov.cn/english/2012‐06/20/content_2165802_6.htm [accessed
22/11/2013]
11. Department of Mines and Petroleum, Guidelines on the safe design and operating standards
for tailings storage. ISBN 0 7309 7808 7, Government of Western Australia, 1999.
12. Department of Mines and Petroleum, Radiation safety guidelines,.
http://www.dmp.wa.gov.au/6745.aspx. [accessed 14/05/2014]
13. A. Liland, P. Strand, H. Amundsen, M. Natvig, M. Nilsen, R. Lystad and K.E.Frogg, “Advances in
NORM management in Norway and the application of ICRP’s 2007 recommendations”, Annals
of the ICRP, 332‐342, 2011.
14. ARPANSA, Radiation protection and radioactive waste management in mining and mineral
processing. Radiation Protection Series publication No 9. Australian Radiation Protection and
Nuclear Safety Agency, Australia, 2005
15. BIO Intelligence Service, Study to explore the feasibility of creating a fund to cover
environmental liability and losses occurring from industrial accidents, Final report prepared for
European Commission, DG ENV, 2013.
16. K. Coppenholle, C. Sykes, R. Batier, C. Marlet, C. Hebestreit, G. Merke, W. Lox, K. Gabel, M.
Wyart‐Remy and D. Fincke, Letter of the Non‐Energy Extractive Industry Panel (NEEIP) on the
feasibility of creating a fund to cover environmental liability and losses, 2013. http://www.ima‐
europe.eu/sites/ima‐
europe.eu/files/publications/Letter%20of%20the%20NEEIP%20to%20Mr%20PELLEGRINI%20
on%20the%20feasibility%20of%20creating%20a%20fu....pdf [accessed 14/05/2014]

A COMPARATIVE ANALYIS OF NATIONAL RAW
MATERIALS POLICY APPROACHES ‐ WITH A FOCUS ON
RARE EARTH ELEMENTS IN EUROPE
Andreas ENDL 1, Gerald BERGER 2
1, 2
Institute for Managing Sustainability, Vienna University of Economics and Business,
Welthandelsplatz 1, Building D1, A‐1020 Vienna, Austria
andreas.endl@wu.ac.at, gerald.berger@wu.ac.at
Abstract
The EU is highly dependent on raw material imports due to a combination of factors,
such as price volatility, geo‐political regimes affecting trade, and a global surge in
demand. Consequently, in 2008 the European Commission launched a process for the
development of a policy framework (1, 2) fostering the sustainable supply and use of
raw materials. However, since many regulatory issues are in the competence of EU
Member States, specific and tailor‐made policy responses need to be designed and
implemented on the national level.
Therefore, this paper analyses national policy mechanisms in 4 EU Member States
(Austria, Finland, Greece, and Sweden) that address REE challenges and formulate
policy responses in the context of general raw materials policy. It will shed light on
policy instruments and governance procedures regarding different approaches to
tackle raw materials supply in the context of governance for SD (participation,
reflexivity, and long‐term vision and short term action). Hence, the paper’s remit lies in
explicating the policy governance regime and making a case for best practices with
regard to different sustainable raw materials supply approaches.
Introduction
Overall, the functioning of economies and society’s well‐being largely depends on the
secure supply and use of natural resources and raw materials in particular. However,
with world population projections estimating more than 9 billion in 2050 and rapid
economic growth in newly industrialising countries, raw materials demand continues
to rise strongly (3, 4). Furthermore, per capita raw material consumption is currently
substantially higher in industrialised nations than in less developed countries (5).
Continuing population growth, combined with increasing affluence, does not only put
a strain on resource availability, but also influences the amount of waste and emissions
produced. In the long run, these trends will lead to surpassing planetary boundaries
and thus will also affect the well‐being of people and the environment (6, 7, 8).
Additionally, the rapid development of emerging economies in the Asia‐Pacific region
led to an increased metal ores and industrial minerals consumption by a factor of 8.6
during the period of 1970‐2008 (9). Furthermore, if the world economy continues to
grow following its current development path, i.e. a “business‐as‐usual” scenario,
global resource extraction would significantly grow in the future (10).
European Union security of supply challenges
Threats to secure supply of certain raw materials and the emergence of critical
materials in the EU depend on a series of different political, economic and
technological factors. Among the most crucial factors threatening EU’s secure supply
of certain raw materials is its high dependency on raw material imports. In 2011, for
example, the EU trade balance (5) for fuel and mining products showed an immense
asymmetry between imports and exports (i.e. the EU imported over six times more
fuel and mining materials than it exported).
In this regard, the EU’s import dependency is subject to a range of market distortion
factors, such as price volatility, as well as the interaction (i.e. distortions through in‐
transparency and speculation) between physical and financial commodities markets,
and export restrictions or privileged access (2). Beyond international aspects,
competition on different land use types (recreation, agriculture, built‐up land) might
further restrict the access to European raw material sources (11).
In addition, the physical raw material demand for development and rapid diffusion of
key enabling technologies such as renewable energy technologies in the future will
aggravate these trends (11, 12).
The European Union raw materials policy framework
Besides tackling environmental and social impacts of raw material supply and use, one
of the major challenges in the raw materials policy debate is to achieve a secure supply
base. Raw material shortages are an imminent threat to the EU’s secure supply and
availability.
As a consequence of these above mentioned trends, the EU acknowledged the
importance of raw materials supply for its economic development and society overall.
Thus, in 2008 the EU instigated the Raw Materials Initiative (RMI) to manage responses
to non‐energy and non‐agricultural raw material issues. As outlined by Tiess (13), so far
the European Union did not pursue a common minerals policy.
Since then, the EU has pursued a 3 pillar‐based approach to improving access to raw
materials for Europe (see Figure 1) which got reinforced in a new strategy document
in 2011 (2). This three‐pillar approach encompasses actions on “1. ensuring a level
playing field in access to resources in third countries”; “2. fostering sustainable supply
of raw materials from European sources; and “3. boosting resource efficiency and
promoting recycling.” (1). Another important element of this strategy engages in an
international approach by embedding the EU raw materials policy agenda in the wider
set of its external policies. In this regard, it aims to foster good governance, human
rights, conflict resolution, transparency of activities, and creation of local value added
in developing countries.
Furthermore, recent developments dealing with security of supply issues in EU policy
directly (European Innovation Partnership‐EIP) or indirectly (i.e. through resource
efficiency and waste management policies: Europe 2020 ”Resource Efficiency Flagship
Initiatives“; ”Roadmap to a Resource‐Efficient Europe“) have been put forward. Since
then, these strategies are shaping the European policy framework on raw materials
security of supply and resource efficiency.
Subsequently, this EU level development triggered the design of several raw material
or mineral strategies (henceforth referred to as National Raw Material Strategies –
NRMS) in several EU Member States (Austria, Denmark, Finland, France, Germany,
Greece, the Netherlands, Portugal, Sweden, United Kingdom) to better accommodate
particular national economies raw materials needs and specific circumstances.
A Rationale for a governance for SD approach in raw materials strategies
In that sense, this paper investigates how the European framework for sustainable
supply and use of raw materials outlined in the Raw Materials Initiative (RMI) and the
role of rare earth elements (REE) scarcity is taken up in tailor‐made policy strategies at
the national level in four European Member States (Greece, Finland, Sweden, Austria).
From our perspective, we detect three specific aspects why governance for sustainable
development (SD) plays a major role for successful NRMS:
1. The RMI calls for “defining a National Minerals Policy, to ensure that mineral
resources are exploited in an economically viable way, harmonised with other
national policies, based on sustainable development principles (…)” (2). Following
this proposition for SD within National Minerals Policies, we argue that governance
principles for SD (14) are a fundamental basis for applying effective minerals policy
at the national level.
2. Overall, governance for SD consists of major building blocks, such as stakeholder
participation and reflexivity that are prerequisites for societal self‐steering.
Specifically, with regard to the inclusion of non‐state actors in policy making and
societal steering in general, governance for SD provides a striking rationale: In that
sense, governments usually have certain limitation, for instance, lack of necessary
authority or means to tackle societal challenges, and difficulties to achieve general
legitimacy or support from interest groups (15). Following these limitations,
governance for SD refers to the involvement of and interaction among multiple
societal actors for a deliberative process of collecting and generating knowledge to
successfully tackle societal challenges (14). In the context of sustainable raw
materials management, and, in particular, the extractive sector multi‐actor
approaches (16) guarantee greater commitment and acceptance for state‐designed
intervention (EIP on Raw Materials).
3. In the context of primary extraction and, particularly, challenges tackled in NRMS,
a large number of diverse stakeholders are directly impacted or involved, and a
manifold set of issues are to be addressed. Such stakeholders and issues cover, inter
alia, “not in my backyard” constellations and local communities, mining impacts on
ecosystems, long‐term business investment decisions for prospection and
extraction etc. Consequently, governance for SD aspects, such as inclusive
stakeholder participation during NRMS design and implementation, are crucial to
facilitate legitimacy of policy options and steering.
Following this rationale, we will shed light on policy instruments and governance
procedures during the design and implementation phase with regard to governance
for SD principles (17, 18, 19, 20): participation, reflexivity, and long‐term vision as well as
short term action. Thus, the paper’s remit lies in explicating different approaches to
tackle raw materials supply in the context of governance for SD.
Methodology
The data basis for the analysis was created by a two‐pronged approach: Through
desktop research of the respective policy documents (i.e. NRMS of Austria, Finland,
Sweden and Greece) and qualitative interviews with policy makers of the respective
countries (who carrying major responsibilities for design and implementation of
corresponding strategies1) form the basis of the data collection methodology. The four
strategies in our analysis are:
‐ Austria: “Austrian Mineral Resources Plan“ (2010)
‐ Greece: “National Policy for the Exploitation of Mineral Resources” (2012)
‐ Sweden: “Sweden’s Minerals Strategy” (2012)
‐ Finland: “Finland’s Minerals Strategy” (2010)
By means of these two data collection methods, the authors address the following
research questions:
Firstly, the questions to what extent major building blocks or governance principles for
SD are incorporated into the strategy by means of qualitative interviews with policy
makers. In doing so, the authors apply as analytical criteria for principles of governance
three major buildings blocks to the concept of governance for sustainable
development: i) participation and stakeholder involvement, ii) reflexivity and learning,
and iii) long‐term visioning and short‐term action.
Secondly, the authors investigated the specific role rare earth elements play with
regard to the three‐pillar approach by means of text analysis and interviews with policy
makers.

1
Ministries of interviewed policy makers: Federal Ministry of Science, Research and Economy (AT);
Ministry of Employment and the Economy (FI); Ministry of Environment, Energy and Climate Change
(GR); Ministry of Enterprise, Energy and Communications (SE)
Results
The following paragraphs explicate on the research questions outlined in the
methodology section: In doing so, the first part outlines the policy genealogy and how
specific EU Member State conditions are taken into account, and the role of access to
REE plays. In the second part the authors outlined the degree to which governance
principles for SD (i.e. participation, reflexivity, and long‐term visioning and short‐term
action) are applied during the design and implementation stage of NRMS.
Policy genealogy and driving forces of national raw material strategies
As regards the four represented case study countries, all share the fact that their
present NRMS is the first umbrella strategy encompassing a strategic approach
towards access to raw materials on a national level. Greece, Sweden and Finland
reported that the EU level policy framework – the 2008 RMI – was the common driver
for initiating the development of their NRMS. Only in the case of Austria, in 2001 the
parliament (“Nationalrat”) instigated the development for an overall strategy for more
cross‐cutting as well as coherent policy mechanisms securing access to raw materials.
In that sense, the Austrian Mineral Resources Plan, although still influenced by the
RMI, has been considered as a best practice example by the European Commission.
Apart from the European Commission providing a major impetus for the development
of NRMS, national framework conditions have strongly contributed to their
development and content focus. In particular in the case of Sweden and Finland,
economic factors played a major role. In these two countries, a “mining boom”
through increased investment in mining and exploration activities led to augmented
political awareness on the extractive sector’s activities., Moreover, environmental
challenges and more effective land use planning (avoiding conflicts in use and
guaranteeing access to important mineral deposits) have been vital concerns for the
development of the Austrian and Greek strategies. In Austria, the issue of public
acceptance or conflicts with local communities for on‐going and future mining projects
contributed to development of a NRMS.
Prior to this development, none of the four countries did have a coherent and strategic
approach (i.e. predecessor umbrella strategy) for access to raw materials from
domestic sources. Instead, these countries only had a mix of single stand‐alone
instruments addressing different issues, such as land‐use planning, covering different
materials and different industrial sectors. Only in the case of Greece, mineral policy
has been managed through a tripartite approach of three different laws for raw
materials forming a more coherent framework.
Addressing access to Rare earth elements
This paper explores to what extent NRMS are focusing on access to REE and if they
foresee concrete actions. As regards the four country cases, none of the strategies
makes a clear reference or encompasses actions specifically addressing access to REE.
Nevertheless, in some countries, non‐commodity specific actions include aspects on
REE scarcity and access. In relation to exploration activities Greece and Austria either
already identified respective deposits (Austria) or are currently prospecting for
potential deposits or anomalies (Greece). Sweden is taking measures for a criticality
assessment of certain materials where REE might play a role. Interestingly, in Finland
some companies involved in the implementation of concrete strategy actions consider
extracting REE from old tailings in case world market prices increase.
Despite some more sporadic and unspecific efforts with regard to REE, none of these
countries engage in actions for REE extraction yet. However, increased efforts for
exploration (Greece) and safeguarding access to deposits by restricting other land uses
(Austria) might lead to future extraction projects.
While there is no significant REE production in Europe, but only uncharacterised and
unexplored deposits (21), NRMS of Austria, Finland, Sweden and Greece show no
commodity specific focus on securing REE supply. This to some extent depicts a
perceived mismatch of EU level topical importance of REE supply compared to EU MS
follow up in NRMS. Nonetheless, the RMI’s other two pillars cover issues such as raw
materials diplomacy with exporting countries and strategies increasing recycling rates
of WEEE which highlights a different non‐primary extraction approach towards
securing REE supply. These two pillars are taken up in several EU MS policy strategies
but go beyond the scope of this paper.
Participation and stakeholder involvement in policy design and

implementation
One of the key principles of SD is participation and stakeholder involvement
throughout the whole policy cycle (i.e. ranging from policy design to implementation,
evaluation and revision). In that sense, we asked policy makers about the respective
role stakeholders (state as well as non‐sate actors) played in the design,
implementation and monitoring of the respective NRMS.
In general, depending on the responsibility for raw materials management in the
respective country, the ministry of environment (Greece) or the ministry of economics
(Austria, Finland, Sweden) had the main responsibility for policy design.
In terms of involvement of state actors, the four countries cases basically feature three
different collaboration approaches during the design phase of NRMS: 1) exclusive
development, 2) partially inclusive, and 3) fully inclusive strategy development
process. The three approaches are based on the degree of involvement of i) state
actors (i.e. ministries), and ii) non‐state actors (i.e. industry, public authorities,
academic as well as private research institutions, and civil society organisations such
as environmental NGOs).
Concerning the exclusive development approach, the Greek Ministry of environment
was solely responsible for strategy development, involving only a limited number of
non‐state actors, such as one industry association, one public authority and two
academic institutions. On the other end of the scale, the Swedish NRMS is
characterised by a fully inclusive development approach: Not only have all ministries
been consulted and involved in the design process, but also a broad variety of non‐
state actors covering industry, public authorities, academic and private research as
well as civil society and environmental organisations. Both Austria and Finland are
characterised by a partially inclusive development process, involving two ministries in
the co‐development process and a number of non‐state actors such as industry, public
authorities, academia and private research as well as civil society and environmental
organisations.
In a next step, we explored the extent to which non‐state actors are taken on board
during the implementation of actions of the respective NRMS. For our four country
cases, we again identified three different approaches for strategy implementation and
the role non‐state actors play: 1) “state‐actor‐only implementation”, 2) “shared but
differentiated implementation”, and 3) “fire‐and‐forget implementation”. As regards
“state‐actor‐only implementation”, the Greek NRMS is mainly followed up by state
actors (i.e. ministry of environment supported by other ministries where
responsibilities overlap). Non‐state actors are involved on a rather ad‐hoc and on‐
demand basis in respective consultation procedures or committees during the
implementation process. A similar case can be found concerning the steering of the
Austrian NRMS where the responsibility for implementation is not located on national
ministerial level but rather on the level of federal states.
An implementation approach with a higher degree of non‐state actor involvement
(“shared but differentiated implementation”) is applied in the NRMS of Sweden.
Essentially, the government appoints one authority responsible for implementation of
individual actions. More specifically, the government provides the framework
conditions (financing, goals and targets, time‐frames, monitoring), whereas these
authorities enjoy a considerable amount of freedom and independence during the
implementation process.
Concerning the more active role of non‐state actors, Finland’s NRMS allows for the
largest degree of freedom and non‐state actor involvement during implementation.
Generally, the responsibility for implementing action proposals (outlined in the 2013
action plan) in the NRMS is split among different actors: State‐actor steering is limited
in so far, as the action plan only comprises non‐binding advice on actor’s involvement
and provides opportunities for new actors entering action proposals. Leading actors in
these actions proposals cover a broad variety, such as industry associations or
environmental NGOs. These actors are also participating in the government steering
group which regularly follows up on the development status of strategy
implementation.
Finland and Sweden, when compared to Austria and Greece, feature greater inclusive
stakeholder participation during NRMS action implementation and to a lesser extent
in policy design. Since primary extraction challenges necessitate such multi‐
stakeholder approaches (see for example recent EU policy developments such as the
EIP on raw materials), these policy regimes, on the one hand, might contribute to
increased legitimacy for steering, and, on the other hand, facilitate implementation of
individual actions.
Reflexivity and learning in the implementation process
Another important building block of governance for SD refers to the concept of
reflexivity and learning as an ability to adapt to changing conditions within a given
(governance) system. The following paragraphs highlight the instruments for
monitoring and evaluation applied in NRMS to adapt to either changing framework
conditions or respond to varying degrees of action implementation.
Generally, the four country cases exhibit quite diverse systems, e.g. applying more ad‐
hoc or systemic and ex‐ante in‐built approaches towards institutionalised learning in
NRMS.
Regarding Greece, for example, 9 broad objectives are outlined in the NRMS for which
currently no monitoring or evaluation mechanism (i.e. no indicator or assessment
frameworks) exists for assessing the implementation status. In the case of Austria, the
main criteria for successful implementation will be the translation of the major
instrument (land use planning law) into federal state law. However, no current
monitoring and evaluation system exists to measure its implementation process.
The Swedish NRMS applies a mixed approach of several targets apportioned to 12
actions areas and two horizontal objectives. However, no concrete indicator
framework exists for measuring these targets but instead the responsible authorities
produce a yearly qualitative evaluation report. This report not only informs the
government and all involved actors but also a wider group of stakeholders at a yearly
forum.
On the other end of the scale, Finland developed the most comprehensive system for
policy learning and reflexivity. Besides a set of 12 general as well as specific objectives,
the NRMS features an indicator framework for following up on the measurement of
objectives and produces a qualitative evaluation report.
As regards reflexive learning, concrete evaluation and monitoring approaches
combined with multi‐actor involvement, as in the case of Finland and to a lesser extent
for Sweden, allows for flexibility to address changing conditions and facilitated by the
collective knowledge and understand of plurality of interests and needs.
Engaging in long term visioning and short term actions
This part engages in the process of envisioning long‐term transformation processes,
fundamental to sustainable development, and translating them into short term actions
necessary for applying policy instruments.
Generally, the patterns of specific time frames, actions, and overall implementation
frameworks is quite diverse: While Greece, Austria, and Sweden do not have a
concrete implementation or roadmap, Finland has a fully fledge “sustainable
extractive industry” action plan. In order to better understand the different nature of
implementation framework, we categorised the four country cases into 1) no
implementation framework (Greece, Austria) with “continuous and unspecific”, 2)
“overall specific”, and 3) “differentiated specific” time‐frames and actions for
implementation.
With regard to the first category, we consider a continuous implementation of actions
(no concrete action plan or implementation framework) with no concrete time
schedule for specific actions. While in the case of Greece, a mix of short and long—
term actions are initiated and followed up in a step by step basis, the Austrian NRMS’
major instrument is under gradual negotiation with federals states for changing their
legislation accordingly.
Furthermore, “overall specific” in the case of Sweden refers to a partial
implementation framework which considers an overall time frame (3 years) for short
as well as longer‐term actions with a specific financial framework for that given period.
On the other hand, Finland developed a fully‐fledged policy action plan (2013)
encompassing differentiated and specific implementation time frames across short
and longer term actions.
Considering, on the one hand, for long‐term perspectives while, at the same time,
strategically designing implementation for the short‐term, is especially relevant for
successful societal transformation processes, such as sustainable development in the
extractive sector. In that case, Finland provides a best practice case through a balanced
combination of:
 Long‐term envisioning: such as a set of broad objectives outlined in an umbrella
strategy (i.e. NRMS), and at the same time; and
 Short‐term actions: strategically designing short actions accompanied by specific
time‐frames, implementation pathways, financial framework and multi‐actor
consortia (i.e. policy action plan).
Discussion and conclusions
As outlined in the previous sections, EU MS NRMS demonstrate diverse policy regimes with
regard to 1) the three major building blocks in governance for SD, and 2) the actual
accentuation for securing REE from a primary extraction perspective.
In a nutshell, we want to highlight four major differences and similarities in terms of
approaches to REE supply and policy governance regimes for SD:
 Common among the four selected NRMS is a perceived mismatch of EU level topical
importance of REE supply compared to EU Member State follow up in NRMS in the area of
primary extraction.
 Since primary extraction challenges necessitate multi‐actor approaches, Finland and, to a
minor extent, Sweden support these policy regimes and, thus, contributing to increased
legitimacy for steering and facilitate implementation of individual actions.
 Concrete policy learning approaches, combined with multi‐actor involvement as in the
case of Finland and to a lesser extent for Sweden, allow for flexibility to address changing
conditions facilitated by collective knowledge.
 Both, considering long‐term perspectives and, at the same time, strategically designing
implementation for the short‐term through policy roadmaps and action plans (Finland), is
especially relevant for successful societal transformation processes such as sustainable
development in the extractive sector.
This work is related to the project COBALT. The research leading to these results is associated
with the Community's Seventh Framework Programme (FP7/2007‐2013) under grant
agreement n° 603509 (COBALT, www.cobalt‐fp7.eu/).
References
1. European Commission, “Tackling the Challenges in Commodity Markets and on Raw Materials”, European
Commission, Brussels, COM(2011) 25 final, 2011.
2. European Commission “The raw materials initiative — meeting our critical needs for growth and jobs in
Europe”, European Commission, Brussels, COM(2008) 699 final, 2008.
3. UNEP, “Decoupling natural resource use and environmental impacts from economic growth”, A Report of
the Working Group on Decoupling to the International Resource Panel, M. Fischer‐Kowalski, M. Swilling,
E.U. von Weizsäcker, Y. Ren, Y. Moriguchi, W. Crane, F. Krausmann, N. Eisenmenger, S. Giljum, P. Hennicke,
P. Romero, P. Lankao, , A. Siriban Manalang, S. Sewerin, UNEP, 2011.
4. M.,Dittrich, S. Giljum, S. Lutter and C. Polzin, “Green economies around the world? Implications of resource
use for development and the environment”, SERI, Vienna, 2012.
5. EEA, “Material resources and waste — 2012 update”, EEA, Copenhagen, 2012.
6. J, Rockström, W. Steffen, K. Noone, Å. Persson, F. S. Chapin, E.Lambin, T. M. Lenton, et al. , “Planetary
Boundaries: Exploring the Safe Operating Space for Humanity”, Ecology and Society 14 (2): 32 (2009).
7. UNEP, “Assessing the Environmental Impacts of Consumption and Production: Priority Products and
Materials”, A Report of the Working Group on the Environmental Impacts of Products and Materials to the
ERES2014: 1st European Rare Earth Resources Conference|Milos|04‐07/09/2014

International Panel for Sustainable Resource Management, E. Hertwich,, E. van der Voet, S. Suh, A. Tukker,
M. Huijbregts P. Kazmierczyk, M. Lenzen, J. McNeely, Y. Moriguchi, 2010.
8. EEA, “Environmental pressures from European consumption and production. A study in integrated
environmental and economic analysis”, EEA Report No 2/2013, EEA, Copenhagen 2013.
9. UNEP, “Recent Trends in Material Flows and Resource Productivity in Asia and the Pacific”, UNEP Division
of Early Warning and Assessment, Bangkok, 2013
10. SERI, http://www.materialflows.net/trends/analyses‐1980‐2008/global‐resource‐extraction‐2005‐2030/,
Sustainable Europe Research Institute Vienna, 25.6.2014 (2014).
11. I. Öhrlund, “Science and Technology Options Assessment: Future Metal Demand from Photovoltaic Cells
and Wind Turbines”, European Parliament (2011); http://go.nature.com/VUOs7V
12. European Commission, “Making Raw Materials available for Europe’s Future Wellbeing. Proposal for a
European Innovation Partnership on Raw Materials”, European Commission, Brussels COM(2012) 82 final,
2012.
13. G. Tiess, “Minerals policy in Europe: Some recent developments“, Resource Policy 35(3), 190–198, (2007).
14. J. Meadowcroft, “Who is in Charge here? Governance for Sustainable Development in a Complex World”,
Journal of Environmental Policy and Planning, 9 (3‐4) 299‐314 (2007).
15. A. van Zeilj‐Rozema, R. Cörvers, R. Kemp and P. Martens “Governance for sustainable development: a
framework”, Sustainable Development, 16 (6) 410‐421, (2008).
16. A. Endl and G. Berger, “A stakeholder perspective on sustainable raw materials management”, COBALT
working paper No. 1, www.cobalt‐fp7.eu, May 2014.
17. S. Baker, “In Pursuit of Sustainable Development: A Governance Perspective”, Paper presented at the 8th
International Conference of the European Society for Ecological Economics (ESEE), Ljubljana, 29 June – 2
July, 2009.
18. G. Berger, “Governance for Sustainable Development: Concepts, principles and challenges”, Paper
presented at the Eurofound Expert Meeting: Industrial Relations and Sustainability, Brussels, December 3,
2009.
19. R. Kemp, S. Parto, R. B. Gibson, “Governance for sustainable development: Moving from theory to
practice”, International Journal of Sustainable Development, 8 (1/2) 12–30, (2005).
20. H. Shiroyama, M. Yarime, M. Matsuo, H. Schroeder, R. Scholz, A.E. Ulrich, “knowledge integration and
multi‐actor dimensions in risk management”, Sustainability Science, 7 (1) 45–55, (2012).
21. European Commission, “Report on the critical raw materials for the EU ‐ Report of the Ad hoc Working
Group on defining critical raw materials”, European Commission, Brussels, COM(2011) 25 final, 2014.


Assessments of boundary conditions and requirements for
Rare Earth Underground Mining due to presence of NORMs
George BARAKOS1, Helmut MISCHO2, Jens GUTZMER3
1
Helmholtz Institute Freiberg for Resource Technology, Halsbruecker Strasse 34, 09599 Freiberg,
Germany
2
Institute for Mining and Special Civil Engineering, TU Bergakademie Freiberg, 09599 Freiberg,
Germany
3
Institute for Mineralogy, TU Bergakademie Freiberg, 09599 Freiberg, Germany
g.barakos@hzdr.de, helmut.mischo@mabb.tu‐freiberg.de, j.gutzmer@hzdr.de
Abstract
Rare earths contain NORMs, such as uranium, thorium and their progeny, like radium and
radon. The varying concentrations of NORMs are quite often significant enough to result in
occupational and environmental radiation exposures during the mining, milling and
processing procedures of rare earths and compounds. Ventilation is the primary technique of
controlling ambient concentrations of radon. Fresh air volume flow rates, the distribution of
airflow within the mine and the radon emanation rate are primary factors affecting such
concentrations. In this paper, it is attempted to determine the factors that may result in
radiation risks and evaluate the boundary conditions that will contribute to the restriction or
even elimination of radon progeny, with a goal to use the evaluations in order to build an
overall assessment tool.

Presence of NORMs in Rare Earth Elements
Rare Earth Elements (REEs) in economically exploitable concentrations are mostly found in
minerals such as bastnaesite, monazite, xenotime, gadolinite, fergusonite, samarskite and
allanite. All of these ore minerals contain low to significant levels of thorium and uranium
series nuclides. They are thus typical examples of the Natural Occurring Radioactive Materials
(NORMs). Rare earths themselves contain naturally radioactive isotopes as well, such as 138La,
142
Ce, 144Nd, 145Nd, 147Sm, 148Sm, 149Sm, 152Gd, 156Dy and 176Lu. The uranium content within
rare earth minerals varies from insignificant percentage to 0.8 wt. % while the thorium
content varies from 0.1 to 10 wt. % depending on the mineral and occurrence. Bastnaesite
(Ce, La)CO3F contains 0.1–0.2 wt. % thorium (ThO2) and negligible concentrations of uranium
(U3O8). Monazite (Ce,La,Pr,Nd,Th,Y)PO4 contains 0.2–0.4 wt. % U3O8 and 4.5–9.5 wt. % ThO2.
Xenotime Y(PO4) contains almost equal percentages of uranium and thorium, 0.81 wt. % U3O8
and 0.83 wt. % ThO2 respectively1.


Radioactive Decay Chains of NORMs
NORMs decay with respect to time into many elements called ‘daughters’, a process
associated with the release of radiation. The list of subsequent daughter products is known
as the ‘decay chain’. Radon is a product of the radioactive decay chain of primordial uranium
or thorium (Fig. 1), specifically the isotopes 238U, 235U and 232Th. The isotopes formed from
this decay are, 219Rn (‘actinon’), 220Rn (‘thoron’) and 222Rn (‘radon’). Due to the low abundance
of 235U and the short half‐lives of thoron (55.6 sec) and actinon (3.96 sec), most work
concentrates on 222Rn and its progeny.

Figure 1: Uranium decay chain with the half‐life of progeny (Source: SSVEE).

Radon is unique among the decay products because it is a gas, thus is capable of migrating
from the source location into the atmosphere. It is colorless, odorless, tasteless and
chemically inert with a half‐life of 3.82 days. Radon undergoes 4 rapid decays, forming
airborne progeny, starting with 218Po (‘polonium’) which decays into 214Pb (‘lead’), which
decays into 214Bi (‘bismuth’) and finally into 214Po (‘polonium’). On entering the mine
atmosphere and without sufficient ventilation, this decay chain leads to significant
radioactivity (alpha, beta and gamma radiation) in the mine.
Occupational Exposure Limits and Impacts on Human Health
Radon concentration is measured in units of radioactivity rather than mass because the actual
mass involved is too small2. Common units are; the Becquerel (Bq) for radioactivity and the
Sievert (Sv) for the equivalent dose in the human body, while sometimes in mining the historic
unit Working Level (WL) is also used. Occupational exposure limits were derived to protect

the health of underground miners over a working lifetime. The recommendations issued by
the International Commission on Radiological Protection (ICRP), are that no person shall be
exposed to more than 14 mJh/m3 (≈4 WLMs) annually. The effective dose is 20 mSv/year
averaged over 5 consecutive years or 50 mSv allowed in any 1 year3. The lifetime limit is 400
mSv.

Occupational exposure of underground miners to radon has been related to severe health
impacts decades ago, making it the second leading cause of lung cancer after cigarette
smoking4. It was first identified as a cause of cancer in uranium miners in 19245. But despite
the fact that most studies are focused on its bearing risks, it is not the radon itself that is
actually responsible for the health problems, but rather the short lived radon daughters
(SLRDs) and their decay products, as described above. The radon may be thought of merely a
source for the SLRDs.
Factors Resulting in Radon Spreading in Underground Mining
Radon and progeny emanation can be attributed to their natural occurrence inside the rare
earth element bearing minerals. However, the emission of radon from the undisturbed rock
can be considered as insignificant compared to the amount of radon released after the
rupture of the ore during mining, when the specific surface area of the rock increases. It’s
spreading in the mine atmosphere, combined with the failure to restrict and/or eliminate it,
may be imputed to a series of factors.

It is of utmost importance to evaluate boundary conditions such as dust reduction, the
residence time of the excavated ore in the mine sites, the possible bioleaching of the ore, the
haulage system, the operation of crushing and grinding mills, as well as the use of tailings in
potential backfill mining. Moreover, one of the most important factors responsible for radon
concentration is the lack of sufficient air flow in the mine. Inadequate ventilation may result
in the insufficiently slow removal of radon gas, thus giving it time to accumulate inside the
mine atmosphere (Fig. 2). Another factor is the presence of groundwater in a REE mine. Radon
partially dissolves in water and thus can be transported further away by groundwater flow.
Water‐borne radon will inevitably be outgassed into the air due to de‐pressurization and
constant agitation6. Much of the radon will also decay before it has the opportunity of release.
Assessments for Controlling Radon and Progeny Emanation
The evaluations for maintaining a safe working environment in rare earth mines are basically
the same as in conventional mines; however, the occurrence of significant amounts of
NORMs, that are the focus of the mining operation, imposes special boundary conditions in
order to prevent them from having sufficient time to build up

Figure 2: Ascending of radon and radon daughters’ concentration versus time after drop
down of ventilation (Reinhard Wesely, 1983).

significant quantities of their progeny. These conditions can be established based on those
applied to uranium mining operations. However, it needs to be stressed that, different to
uranium ore deposits, REE deposits will have, in many cases, a Th rather than a U tenor.
Mine planning
Ore dust presents the dominant source of exposure to radon and progeny, thus a proper
mining method selection can reduce the dust production. As in every case, there is no single
appropriate mining method that will efficiently deal with all the radiation issues and risks.
Excavation techniques, crushing and transportation as well as the long time residence of the
broken ore in the mine should be the operations to evaluate so as to keep radon accumulation
below the occupational safety standards.

Common excavation practices can be separated into mechanical excavation and use of
explosives. With respect to dust production during mechanical excavation, several studies
have been made, indicating that the deeper the cut and the larger the chips, the less dust is
produced from the ore removed7. The sharpness of the cutting tools is also a factor, since
worn bits without their carbide tips produce much more dust8. When it comes to drilling, dust
can be controlled with dry collectors at the tip of the drill bit or with water injection through
the drill steel. Foam injection can also be used when excessive water is a problem9. Blasting
is done when no one is expected to enter the affected area. Consequently, blasting should
take place at the end of the shift when all other operations are finished.

During dropping and transporting of the mined ore, dust can also be produced in significant
amounts. Depending on the mining method, the broken ore may fall from several meters
height to the ground or can travel long distances through conveyor belts, that can be of

moderate cost but may be a significant source of dust, or mine railways that usually generate
less dust. Enclosure, drenching and ventilation are the primary techniques to deal with these
dust emission factors. In the case of wetting the ore on conveyor belts, the amount of water
is an issue again for not leaving a sticky mud residue on the belts. Radon‐bearing ore dust can
be accumulated in areas where crushers and mill plants are installed. In REE mining it would
be preferable not to have a crushing plant within the underground mine, yet if inevitable,
strict dust control measures should be implemented.

Adequate wetting can be extremely important for dust control. The majority of the dust
particles are not released to the air, but stay attached to the surface of the broken ore10. The
presence of water, though, may be an additional problem and the ample drenching of the ore
within the limits of necessity is crucial. Furthermore, remote operation of machinery or cabins
in all vehicles for the machine operators, either during mining or mucking of the ore should
be evaluated as a condition that reduces human exposure to radioactivity. Personal
protection devices such as dust masks can be used. Job rotation is also recommended in high
concentration areas.

Some mining methods raise more issues than others. Bioleaching is a cost effective method,
but most of the excavated ore is left underground creating high radon levels that are difficult
to control. Similarly in shrinkage stoping, the residence of the ore on site will result in radon
and progeny accumulation. Among REE mines applying cut and fill methods, those using
tailings as backfill are found to have higher radon levels than those using rocks or cement.
Concrete can be used to reduce radiation from the exposed ore on the back or walls;
therefore it could also be used as part of the backfill mixture. In sublevel stoping the broken
ore is falling from significant height, creating the “piston effect” that leads to generation of
big amounts of ore dust. The sublevel caving method is well suited for a high degree of
automation and remote operations with corresponding high productivity; however this
method may result to high radon progeny concentration levels and should thus possibly be
avoided.
Ventilation
The design of the overall ventilation system in a REE mine is based on the mining operation;
on ventilation‐air‐transit time; on radon emissions from the wall‐rock and broken ore, the
haulage, the tailings backfill and the groundwater. The aim is to lower radon residence times
to 10‐15 minutes, in order to attain a limit of 10‐20% of the theoretical yield of progeny in
the atmosphere11. It is the total air volume flow rate through the mine that determines the
time air takes to travel from the inlet to the production areas and to the outlet of the mine.
The primary applications are main fan and multi‐fan ventilation systems. Main ventilation can
either be forced (blowing), exhaust (suction) or a combination of the two. Multi‐fan
ventilation uses forced, exhaust, main and auxiliary fans, and a parallel multiple fan operation
technique based on energy saving and building partition ventilation12.

With respect to radon confinement, forced ventilation is more effective, even mandatory in
many cases and countries. In forced ventilation the positive pressure on the intake airway not
only prohibits radon from being released from rocks but also blocks it from entering into
airways that channel fresh air into working faces. The effectiveness of the method increases
as the fracturing of the rock is growing. Nevertheless, air leakage due to positive pressure
state can be a significant issue, even more if air doors are not well maintained. Infrastructure
costs are relatively high, as an independent intake shaft has to be built, in addition to a return
airshaft. On the contrary, in a state of negative pressure due to suction ventilation, where
fresh air moves into the mine because the pressure is lower, the release of radon and progeny
is speed up either because of the operations or the wall rock emission.

Development headings are areas of higher risk and their proper ventilation is critical. Auxiliary
ventilation is to be used in such areas11. Continuous air change should be planned for
headings. During production, flow‐through ventilation must be employed to keep the working
staff in clean fresh air. Moreover, a forcing ventilation system should be used to deliver fresh
air to the faces. This air from production areas is, of course, not to be reused to ventilate
other areas. The distribution of pressure is harder to manage when multiple operations are
taking place at the same time at various levels. Ideally, a retreat system should be
implemented with a minimum possible number of working faces ventilated in series,
decreasing the effective surface in intake airways. If this is not possible, then each mining
area could be designed as an independent ventilation block. A split system of ventilation
should be employed, in which fresh intake air is distributed to working blocks with respect to
relative needs. Air from each section is afterwards collected in an isolated return airway
without contaminating other active mining areas. The number of personnel required to work
or travel in return airways should be kept to a minimum.

Beyond production areas, ventilation assessments should be made for all rooms and facilities
in a rare earth underground mine. Primary airways should be kept free from mining activity,
so that relatively high air velocities can be readily maintained. Air speed depends on local
regulations. Mined‐out areas should be kept on the return side of the ventilation system and
be sealed whenever possible. Ramps are likely to be contaminated, for the airflow is not
stable and is likely to change with the development of the mine. Conveyor ways, crushing
facilities and ore passes in general should be ventilated so that exhaust air can be directed to
the return air system quickly. Warehouses, repair shops and laboratories should be positively
ventilated by controlled air volume flow rate. All unventilated areas in the mine must be
sealed and marked, stating the hazard, to prohibit inadvertent entries. Moreover, to prevent
radon flow from goaf areas or old drifts, hermetisation with sealing dams or concrete
platforms in raise drifts should be evaluated. Last to consider is the maintenance of such a
complicated system. Surveys and safety inspections are used to verify if the ventilation
system is in compliance and if it meets all defined objectives. Determination of airflow and

the proof that the minimum limits are kept; evidence that air velocity is within the minimum
and maximum limits; proof that the climate limits are met, are evaluations of primary
importance. Any deficiencies found during an inspection should be acted upon promptly.
Monitoring
The principal objectives of monitoring are to evaluate occupational exposures with respect
to the accepted standards and limitations and to provide data for adequate control. This can
be done by collecting samples in short time periods, primarily in the working faces, where
human exposure is more frequent. Thereby, there will be better detection and evaluation of
the principal sources of exposure, assessment of the effectiveness of the control equipment,
detection of anomalies in the mining operation and prediction of the effect of future
operations on contamination levels.
Personal protection measures
Thorough personal hygiene should be required of all personnel and personal cleanliness
should be mandatory at the end of each work shift. Resting and changing rooms should be
isolated from working areas, and provided with convenient access to washing facilities.
Personal protective equipment such as respirators, dust masks or air helmets should be used
in areas where airborne dust is high or in exceptional circumstances when the ventilation has
dropped down or during a maintenance task for which adequate ventilation is not available.
Gloves should be worn for any direct contact with concentrates. Dosimeters should also be
carried from all personnel.
Groundwater assessments
Confining water from coming in contact with airborne radon can be accomplished by using
pipes for sealing or by applying grout covers ahead of development11 to divert water flow
from the dust production areas. Furthermore, sufficient mine drainage can prevent the
creation of stagnant pools of water contaminated with radon.
Conclusions
The presence of NORMs, especially of radon, is a paramount safety and health issue for the
underground mining industry. This poses challenges for ventilation engineers to minimize the
radiation contamination of the air and thus the human exposure. The most efficient methods
to achieve control are; mechanical dilution ventilation, confinement or suspension of
radiation source, dust reduction and control as well as personal protection measures. The
improvements made so far have reduced radon concentration in underground mines,
however, the work to further improve the evaluations continues. In an ideal underground
rare earth mine radon accumulation is prevented and radiation concentration is kept to very
low levels with respect to safety and cost efficiency. For this reason, a thorough investigation
was made in this paper to evaluate the primary boundary conditions and requirements. What

can be done further on is to use these evaluations in order to build an overall assessment
tool, to adjust and apply special techniques in mining of radon‐emitting deposits.
References
1. P. M. B. Pillai, “Naturally Occurring Radioactive Material (NORM) in the extraction and processing of Rare
Earths”, in Proceedings of the international symposium on NORM, pp 197‐221, Seville, Spain, 2007.
2. D. B. Apel and Z. Hashisho, “Radiation Control”, in SME Mining Engineering Handbook, pp 1625‐1632, 3rd
Edition, Society for Mining, Metallurgy and Exploration Inc., Colorado, U.S.A., 2011.
3. Occupational Radiation Protection in the Mining and processing of raw materials, IAEA Safety standards
series, 2004.
4. H. Frumpkin and J. M. Samet, “Radon”, CA Cancer J. Clin., 51 (6): 337‐344 (2001).
5. D. M. Loring, “A study of radon regulation and pathology as it relates to underground hard rock mining”,
in Proceeding of the 12th U.S./North American Mine Ventilation Symposium, pp 59‐63, Society for Mining,
Metallurgy and Exploration Inc., Nevada, U.S.A., 2008.
6. R. L. Patnaik, V. S. Srivastava, V. N. Jha, A. K. Shukla, R. M. Tripathi and V. D. Puranik, “Radiological Safety
Aspects of Low Grade Uranium Mine in India”, in Mine Ventilation, Ninth International Mine Ventilation
Congress, pp 1013‐1022, Dhanbad, India, 2009.
7. J. Ludlow and R. J. Wilson, “Deep cutting: Key to dust‐free long walling”, in Coal Mining and Processing, 19
(8): 40‐43 (1982).
8. J. A. Organiscak, A. W. Khair and M. Ahmad, “Studies of bit wear and respirable dust generation”, in
Transactions SME, Vol. 298, Society for Mining, Metallurgy and Exploration Inc., Colorado, U.S.A., 1995.
9. S. J. Page, “An evaluation of three wet dust control techniques for face drills”, U.S. Department of Interior,
Bureau of Mines, RI 8596, Pittsburgh, U.S.A., 1982
10. F. N. Kissel, “Handbook for Dust Control in Mining”, National Institute for Occupational Safety and Health,
NIOSH Publications, Pittsburgh, U.S.A., 2003.
11. C. Gherghel and E. De Souza, “Ventilation requirements for uranium mines”, in Proceeding of the 12th
U.S./North American Mine Ventilation Symposium, pp 65‐69, Society for Mining, Metallurgy and
Exploration Inc., Nevada, U.S.A., 2008.
12. P‐H. Hu and X‐J. Li, “Analysis of radon reduction and ventilation systems in uranium mines in China”, in
Journal of Radiological Protection, 32 (3): 289‐300 (2012).
13. M. J. McPherson, “Radiation and radon gas”, in Subsurface ventilation and environmental engineering, pp
457‐487, Chapman & Hall, London, U.K., 1993.
14. M. A. Tuck, “Mine Ventilation”, in SME Mining Engineering Handbook, pp 1577‐1594, 3rd Edition, Society
for Mining, Metallurgy and Exploration Inc., Colorado, U.S.A., 2011.
15. F. C. Bossard, “Ventilation of uranium mines”, in A manual of mine ventilation design practices, Floyd C.
Bossard & Associates Inc., Butte, U.S.A., 1983.


A FRAMEWORK FOR THE SUSTAINABLE DEVELOPMENT OF
RARE EARTH ELEMENTS MINING PROJECTS
Sotiris N. KAMENOPOULOS1, Zach AGIOUTANTIS1
1
School of Mineral Resources Engineering, Technical University of Crete, Greece
snkamenopoulos@gmail.com, zach@mred.tuc.gr
Abstract
The dependence of modern society on critical and/or strategic minerals such as Rare Earth
Elements (REEs) is continuously growing due to the numerous end uses and other factors that
have been recognized formerly. This paper proposes the adoption of a framework that can be
used to guide extraction of REEs under Sustainable Development (SD) principles. The proposed
framework includes fundamental elements that contribute to a holistic sustainable platform
for REEs. It expands on the existing sustainable development framework and proposes: (a) five
basic pillars: economy, society, environment, technology, and (geo)politics, (b) three
controlling/limiting factors: policy, governance, and stakeholders, and (c) a number of output
quantities to be used in decision making which can also be utilized as SD indicators
Introduction
The significance of Rare Earth Elements (REEs) has been recognized in terms of their uses,
their trade, the number of recent global initiatives, and the number of related geopolitical
events/reports1. Some of the findings are the following: China is considered a dominant
player in REEs world production; the US, EU and Japan are the major importers of Chinese
REEs; China’s demand for REEs is increasing; the main end uses of REEs include the energy
and defense sectors; REEs applications may provide low cost efficient energy; since 2010 a
large number of REE‐related initiatives have been proposed within the EU and the US; the
substitution of REEs is rare and/or impossible and/or in preliminary status; the recycling
potential of REEs suffers by a number of constraints; and a significant number of critical
geopolitical events/reports related to REEs have been identified since 2010. In addition, an
effort was made to initiate a roadmap for the sustainable mining of REEs. This effort included
the stream mapping of the REEs production process, the identification of stakeholders, and
the detection of hazards/vulnerabilities of REEs mining2. This effort was initiated since until
that time there was a complete lack of such a roadmap. The substitution of REEs might be a
difficult task, thus turning the problem of REEs substitution into a problem of “minimizing the
amount of REEs (…) as little as will give the desired effect” to the end product3. The demand
for REEs is expected to rise in the future4. In addition, the following demand related facts
should be considered when someone is dealing with the sustainable development of REEs:
(a) the world population is expected to increase5, (b) the global Gross Domestic Product (GDP)
is also expected to increase6, and (c) the largest non‐Chinese REEs mining companies are
facing critical financial problems a few years after they started their operations7. Considering

all of the abovementioned information the complete global lack of a framework for the
Sustainable Development of REE mining projects would seem a paradox. This paper discusses
the adoption of a framework that can be used for guiding the Sustainable Development of
REE mining projects.
Sustainable Development Schematic Models
The definition of Sustainable Development (SD) was established in 1987 by the “Brundtland
Commission”10. In 1992, world leaders presented the principles of sustainable development
at the United Nations Conference on Environment and Development in Rio de Janeiro,
Brazil11. During that conference it was agreed that sustainable development consists of three
elements: economic development, social development and environmental protection.
Twenty years later, during the Rio+20 summit, it was concluded that “minerals and metals
make a major contribution to the world economy and modern societies”12. A “sustainable
path” was described as “one that allows every future generation the option of being as well
off as its predecessors”13. Different schematic models picturing SD have been proposed: as
‘pillars’, as nested circles, or as overlapping circles14 (Figure 1).
Figure 1: Different Description Models of Sustainable Development
Sustainable Development in the Minerals Industry
In today’s world the challenge of sustainable mining is critical indeed. This was recognized in
2002 in the World Summit on Sustainable Development (WSSD), the Johannesburg Summit8.
The Johannesburg summit enhanced “the contribution of mining, minerals and metals to

sustainable development” and called for actions “to address the environmental, economic,
health and social impacts and benefits of mining” through the participation of stakeholders
and encouraged world community to develop sustainable mining practices. In 2012, the
United Nations RIO+20 Conference on Sustainable Development9 acknowledged that “mining
activities should maximize social and economic benefits, as well as effectively address
negative environmental and social impacts. In this regard, we recognize that Governments
need strong capacities to develop, manage and regulate their mining industries, in the interest
of sustainable development”.
Recognizing the need to move forward and develop REE mining projects in a sustainable
manner, a framework for the SD of REEs mining projects is proposed in this paper. This
framework is based on the concept of the “overlapping circles”, where the classic three‐circle
schema is complemented with more sustainability circles as well as a number of
controlling/limiting factors/challenges that interact with or within the circles.
A Framework for the Sustainable Development of Rare Earth

Elements Mining Projects
Today, the minerals sector and in particular the development REE mining projects are facing
several challenges2,15. Any recommended action for the establishment of a sustainable
mining framework should be within the context set off by the “Brundtland Commission”, the
Rio Summit (AGENDA 21), and the Rio+20 recommendations. The proposed framework for
the SD of REE mining projects includes fundamental elements that contribute to a holistic
sustainable platform for REEs including:
 Five components represented as circles: economy, society, environment, technology,
and (geo)politics.
 Three controlling/limiting factors: policy, governance, and stakeholders.
 A number of output quantities to be used in decision making: indicators
The recommended framework is pictured in Figure 2 and has a global application. To better
encapsulate the concept of sustainable path, the “Swiss Cheese” model of accidents was
adopted and adjusted to the scope of this analysis (Figure 3, left). Reason’s “Swiss Cheese”
model has been proved to be a very useful tool in “accident” analyses: every “accident” is a
result of “unsafe acts” created by decision makers and/or latent conditions16, 17.

Figure 2: Sustainable Development Framework for REE Mining Projects
It is now broadly recognized that accidents in multifaceted conditions occur due to multiple
causes that jointly interrelate resulting to the “accident”. If the SD of REE mining project is
considered as a desired multifaceted state that may contain several latent conditions then
the ideal sustainable path should be determined, which produces the most efficient
sustainability level without “accidents”. The ideal sustainable path for the SD of REE mining
projects is presented in Figure 3 (right).
The SD models pictured in Figure 1 have a significant weakness: it is difficult to quantify each
criterion and its contribution to the overall scheme, since the models do not provide a
measurable deviation of each probable sustainable metric entity (i.e. indices, ratios,
indicators, etc.) from the ideal sustainable path. As a result, there is no meaningful
interrelation/interconnection between the circles of SD and fuzziness is created for decision
makers. To overcome this weakness the circles of sustainability should be examined from a
different point of view, i.e., the vertical intersection of the circles should be studied (Figure
4).

Figure 3: Adjusted “Swiss Cheese” and Sustainable Development of REE mining projects (left)
and Ideal Sustainable Path of REEs model (right)
Figure 4: Vertical intersection of the sustainability “overlapping” circles
The advantages of representing SD for REE mining projects are the following:
a. It provides better understanding of the SD interrelated elements.
b. It provides a more practical vision of the SD Path.

c. It provides the ability to measure the deviation of probable metric entities from the
ideal SD Path.
d. It provides the ability to decision makers to avoid latent conditions of sustainability by
reducing the deviations and make better decisions that will be closer to the ideal SD
Path.
e. It clearly provides the “go‐no‐go” option to decision makers of mining projects by
adjusting trade‐offs between the different SD elements/metric entities.
f. It clearly sets up the relationship between stakeholders involved in the REEs SD, policy
and governance.
Application of the generic framework in REEs Sustainable

Development
A common expression in management says that “what gets measured gets done” 18.
Evaluating REEs mining projects from the sustainability point of view is very critical for
decision makers and all stakeholders because it can provide measurable positive or negative
impacts of such projects. The best way to perform such evaluations is by using appropriate
indicators. Thus, the core of the proposed framework (Figure 2) is based on indicators. The
next step would be the development of a decision support system which will incorporate
selected indicators and assist decision makers/stakeholders to better assess the impact of
any REEs project from the sustainability point of view. A sustainability indicator ca be defined
as “a parameter, or a value derived from parameters, which points to, provides information
about, describe the state of a phenomenon/environment/area, with a significance extending
beyond that directly associated with a parameter value” 19. The selection of indicators should
be based on the five proposed pillars of Figure 2. The overall process which details how the
generic framework can be applied in the case of REEs is presented in Figure 5. The following
attributes should be taken into consideration when selecting indicators:
1. The number of indicators is a considerable attribute. A large number of indicators
could be difficult to manage and analyze.
2. The word “development” on itself implicates a dynamic process. Thus, the selected
indicators shall be manageable and recordable through time.
3. Indicators shall be accessible at a reasonable time.
4. Indicators shall be pertinent to the geographical region exercised.
5. Indicators shall be easy to be communicated and reported.

Selected indicators for each of the five pillars are presented as examples:
 PILLAR 1 (Economy), Indicator: “Impact of the REEs project into the national/local
economy” (income per person during operations and after mining finishes)
 PILLAR 2 (Society), Indicator: “Impact of the REEs mining project to prosperity in
national/local society” (number of jobs created and/or level of unemployment during
operations and after mining finishes)

 PILLAR 3 (Environment), Indicator: “Impact of the REEs mining project to
environment” (level of radionuclide activity concentrations during operations and
after mining finishes)
 PILLAR 4 (Technology), Indicator: “Impact of REEs mining project to the local high tech
entrepreneurship” (number of high tech business permits issued by national local
authorities during operations and after mining finishes)
 PILLAR 5 (Geopolitics), Indicator: “Impact of REEs mining project to global secure
supply of REEs” (percentage‐reduction‐of Chinese monopoly in global REEs
production)

Figure 5: The process which describes how the generic framework can be applied in the case
of REEs
Conclusions
In this paper a framework was proposed for the Sustainable Development of REE mining
projects. This framework incorporated five basic “overlapping” circles: economy, society,
environment, technology, and (geo)politics. The proposed framework also includes three
controlling/limiting factors: policy, governance, and stakeholders, and indicators to be used
in decision making. Furthermore, to better encapsulate the concept of sustainable path, the
“Swiss Cheese” model of accidents was adopted. Finally, the “overlapping” circles of SD are
proposed to be examined from their vertical intersection. This new approach provides a
practical vision and better understanding of the SD Path, the quantification of the deviation
from the ideal SD path, the “go‐no‐go” ability to SD decision makers and the ability to avoid
latent SD conditions. Finally, it was explained hoe the generic SD framework can be applied
in the case of REEs.

References
1. S.N. Kamenopoulos, Z. Agioutantis, “Rare Earth Elements: A Review and Analysis of their Multi‐dimensional
Global Importance”, Cuprum Magazine nr. 3 (64), Wroclaw, Poland (2012)
2. S.N. Kamenopoulos, Z. Agioutantis, “Rare Earth Elements: Building a Best Practices Roadmap to Sustainable
Mining”, 6th International Conference on Sustainable Development in the Minerals Industry, 30 June – 3
July 2013, Milos island, Greece (2013)
3. j. Lifton, “Replacing rare earths and ‘other’ technology metals: world technology to take a step
backwards?”, http://investorintel.com/rare‐earth‐intel/replacing‐rare‐earths‐technology‐metals/ (2013)
(assessed on 11 February 2014)
4. massachusetts Institute of technology (MIT), “Mission 2016: Strategic Mineral Management, Rare Earth
Elements: Supply and Demand”,
http://web.mit.edu/12.000/www/m2016/finalwebsite/problems/ree.html (assessed on 12 February
2014)
5. United Nations, “World population projected to reach 9.6 billion by 2050 – UN report”,
http://www.un.org/apps/news/story.asp?NewsID=45165 (2013) (assessed on 11 February 2014)
6. Organization for Economic Cooperation and Development, Environmental Outlook to 2050: Climate
Change Chapter, http://www.oecd.org/env/cc/49082173.pdf (2011) (assessed on 11 February 2014)
7. J.Lifton, “Lynas and Molycorp should go bankrupt”, http://investorintel.com/rare‐earth‐
intel/investorintelreport‐lynas‐molycorp‐go‐bankrupt‐says‐jack‐lifton/ (2014) (assessed on 02 April 2014)
8. United Nations Conference on Sustainable Development, RIO+20 The Future We Want, Decisions by Topic:
Mining, Report A/RES/66/288 (2012) at: http://sustainabledevelopment.un.org/index.php?menu=1259
(assessed on 27 April 2014)
9. United Nations, Report of the World Summit on Sustainable Development Johannesburg, South Africa, 26
August‐ 4 September 2002 at:
http://www.un.org/ga/search/view_doc.asp?symbol=A/CONF.199/20&Lang=E (assessed on 29 March
2014)
10. World Commission on Environment and Development, http://www.un‐documents.net/wced‐ocf.htm
(1987) (assessed on 14 February 2014)
11. United Nations, “AGENDA 21”, United Nations Conference on Environment & Development, Rio de Janeiro,
Brazil, (1992)
12. United Nations Conference on Sustainable Development, “Rio+20: The Future we want”,
http://www.uncsd2012.org/about.html (2012) (assessed on 15 February 2014)
13. National Research Council, Commission on Geosciences, Environment, and Resources, “Assigning
Economic Value to Natural Resources”, http://www.nap.edu/catalog/4844.html (1994) (assessed on 12
February 2014)
14. International Union for Conservation of Nature (IUCN), “The Future of Sustainability Re‐thinking
Environment and Development in the Twenty‐first Century”,
http://cmsdata.iucn.org/downloads/iucn_future_of_sustanability.pdf (2006) (assessed on 14 February
2014)
15. Mining, Mineral and Sustainable Development (MMSD), “Breaking New Ground”, Earthscan Publications,
Ltd http://62.50.73.67/DocRoot/ev8jEJvTiMYd4mJhGGHQ/finalmmsdreport.pdf (2002) (assessed on 10
February 2014)
16. J. Reason, “The contribution of latent human failures to the breakdown of complex systems”, Philosophical
Transactions of the Royal Society (London), series B. 327: 475‐
484.http://sustainabledevelopment.un.org/content/documents/Agenda21.pdf (1990) (assessed on 10
February 2014)
17. European Organization for the Safety of Air Navigation, EUROCONTROL Experimental Centre, “Revisiting
the “Swiss Cheese” model of accidents”, EC Note No. 13/06, http://i3pod.com/wp‐
content/uploads/2011/04/Revisiting‐the‐Swiss‐Cheese‐Modek‐EEC‐note‐2006‐13.pdf (2006) (assessed on
17 February 2014)

18. Dr. R. Batterham, “An Australian perspective on indicators of sustainability for the mineral extraction
industry”, XXIII International Mineral Processing Congress, Workshop: A Review on Indicators of
Sustainability for the Minerals Extraction Industries” Istanbul, Turkey, October 13‐14, 2005
19. Organization for Economic Cooperation and Development (OECD) “OECD Environmental Indicators:
Development Measurement and Use”, 2003 at: http://www.oecd.org/environment/indicators‐modelling‐
outlooks/24993546.pdf (assessed on 20 May 2013)


Session III‐B: REE Processing‐
Electrodeposition

SOLUBILITY OF RARE EARTH OXIDES IN MOLTEN FLUORIDES
Xiaoling Guo*, Jilt Sietsma, Yongxiang Yang
Department of Materials Science and Engineering, Delft University of Technology
Mekelweg 2, 2628 CD Delft, The Netherlands
*X.Guo‐1@tudelft.nl
Abstract
The limited solubility of rare earth oxides (REOs) in fluoride melt has been an obstacle to
preparation of rare earth metals (REMs) by oxide‐fluoride electrolysis. However, very little
work has been performed in this field. This paper presents a comprehensive analysis of the
available data from previous publications. It reveals that the REO solubility increases with
temperature and a linear relationship is obtained between the natural logarithm of the
solubility and the reciprocal of the absolute temperature. The rare earth fluoride (RF3) is
crucial to REO dissolution and higher solubility is expected in melts with higher RF3 content.
The alkali fluoride (AF) can lower the melting points of binary systems and improve their
electrical conductivity. The alkali earth fluoride (AF2) can further lower the melting points and
improve the stability of the melts.

Introduction
Thanks to the unique physical and chemical properties, rare earth elements are widely used
in materials and consumer products, and their essential application in permanent magnets,
lamp phosphors, rechargeable NiMH batteries has built a tight link to our modern life 1.

The first attempt to prepare rare earth metals by electrolysis of a oxide‐fluoride bath was as
early as 1907 2. This was followed by many other investigations owing to the advantages of
the process: low cost, environment‐friendliness and easier handleability. The solubility of rare
earth oxide in fluorides and in‐depth understanding of the dissolution behaviour are both of
fundamental and engineering importance for the design of electrolyte compositions and
process operation. Unlike extensive studies on Al2O3‐cryolite system 3‐5, little work has been
done for REO‐fluoride melts. Even worse, there is inconsistency between results from
different studies. This makes the access to reliable data more difficult.

Table 1 lists the major previous studies on solubility of rare earth oxides in molten fluorides.
In the present paper, the data from these studies are summarised, reconstructed and
analysed systematically for a better understanding of the influence of temperature and melt
components on REO solubility in fluoride melts.


Table 1 Investigations on the solubility of rare earth oxides in fluoride melts

Solubility,
Year REO Melt Temperature, K mol. % Reference
1961 La2O3 LiF(‐BaF2)‐LaF3 1223 0.61 ‐0.78 Porter, 1961 6
LiF‐NdF3‐MF2(M = Mg,
1987 Nd2O3 1273 0.14 ‐ 0.96 Du, 1987 7
Ca, Ba)
1991 Nd2O3 LiF‐BaF2‐NdF3 1073 ‐ 1173 1.7 – 2.6 Wu, 1991 8
1994 Y2O3 LiF‐YF3 998 ‐ 1273 0.26 – 2.1 Reddy, 1994 9
Stefanidaki,
2002 Nd2O3 LiF‐NdF3‐MgF2 1023 ‐ 1173 0.08 – 0.38
2002 10
2008 Nd2O3 LiF‐NdF3 1323 ‐ 1423 1.0 – 2.0 Hu, 2008 11
La2O3, Sm2O3, Pshenichny,
2012 LiF/NaF/KF‐ZrF4 873 ‐ 1073 0.039 – 0.73
Ho2O3 2012 12

Influence of different factors
Temperature
Temperature is an unavoidable factor in studying a thermodynamic property. Figure 1 a
shows the dependence of the REO solubility in fluoride melts on the temperature. It is clear
that the REO solubility increases with the increase in temperature.

The dissolution of REOs into fluoride melts can be simplified as an endothermal reaction

REO(s) REO(dissolved) (0)

The equilibrium moving to the right is favoured at higher temperature and consequently
higher solubility of rare earth oxides is obtained in the melts.

In Figure 1 b, the natural logarithm of the REO solubility in fluoride melts is plotted against
the reciprocal of the absolute temperature for the same data as Figure 1a. It is obvious that
most of the data can be interpreted by a linear relationship
1
lnSREO  a   b (1)
T
where SREO is the REO solubility in mol.% and T the temperature in K.
For example, the solubility of La2O3 in 51LiF‐49ZrF4 can be described using eq. (1) with a =
(5.4  0.2)  103 K and b = 4.5  0.2, with R2 = 0.992, in the temperature range of 873 ‐ 1073
K.


The fitting parameters for the rest of the systems are given in Table 2. Most of the plots have
a R2 with more than two nines, showing good linearity for these data sets. Meanwhile, the
differences among the slopes and intercepts indicate the thermodynamic diversity of these
systems. The parameters for the linear regression equation are exclusive to that specific
system and it is not possible to extend them to an unknown system for solubility estimation.
Nevertheless, these parameters are helpful for a more detailed study on thermodynamic
features of the systems.

2.8 1.2
0.8
2.4 Y2O3 in 80LiF-20YF3
0.4 Y2O3 in 75LiF-25YF3
2.0 Y2O3 in 60LiF-40YF3
0.0 La2O3 in 51LiF-49ZrF4
Sm2O3 in 51LiF-49ZrF4
SREO, mol.%
1.6 -0.4
Ho2O3 in 51LiF-49ZrF4
lnSREO
1.2 -0.8 Nd2O3 in 77LiF-23NdF3

Nd2O3 in 79LiF-6MgF2-15NdF3
-1.2 Nd2O3 in 66LiF-2BaF2-32NdF3
0.8
-1.6 Nd2O3 in 72LiF-2BaF2-26NdF3
Nd2O3 in 77LiF-2BaF2-21NdF3
0.4
-2.0 Nd2O3 in 40LiF-60NdF3
Nd2O3 in 50LiF-50NdF3
0.0 -2.4 Nd2O3 in 58LiF-42NdF3
900 1000 1100 1200 1300 1400 6 7 8 9 10 11 12
T, K 1/T x 104, K-1
a b

Figure 1 Solubility of rare‐earth oxides in fluoride melts as a function of temperature
(reproduced with data from reference 7 ‐ 12)

Rare earth fluoride (RF3) content
Usually, a fluoride melt used to dissolve a rare earth oxide consists of the corresponding rare
earth fluoride, an alkali metal fluoride, and sometimes an alkali earth metal fluoride, e.g.
Nd2O3 in NdF3‐LiF‐BaF2. The RF3 content, thus, is another essential factor to be considered.

Figure 2 shows the solubility of rare‐earth oxides in fluoride melts as a function of rare earth
fluoride content. Most of the published data give the same conclusion that the REO solubility
increases with the RF3 content in the melts except a set of data from Wu’s study 8. As his work
involved a ternary system, this disagreement could be attributed to the influence from BaF2.


Table 2 Linear fitting parameters for data plots in Figure 1 (b)

System R2 a [103 K] b
Y2O3 in 80LiF‐20YF3 0.986 ‐4.1  0.2 2.7  0.2

Y2O3 in 75LiF‐25YF3 0.991 ‐4.4  0.2 3.2  0.2
Y2O3 in 60LiF‐40YF3 0.960 ‐5.2  0.5 4.5  0.4
La2O3 in 51LiF‐49ZrF4 0.993 ‐5.4  0.2 4.5  0.2
Sm2O3 in 51LiF‐49ZrF4 0.980 ‐5.9  0.4 5.1  0.4
Ho2O3 in 51LiF‐49ZrF4 0.969 ‐6.8  0.6 6.2  0.6
Nd2O3 in 77LiF‐23NdF3 0.987 ‐3.8  0.2 1.7  0.2
Nd2O3 in 79LiF‐6MgF2‐15NdF3* ‐‐ ‐4.5213 2.093
Nd2O3 in 66LiF‐2BaF2‐32NdF3 0.999 ‐4.69  0.08 4.96  0.07
Nd2O3 in 72LiF‐2BaF2‐26NdF3 1.000 ‐3.27  0.01 3.63  0.01
Nd2O3 in 77LiF‐2BaF2‐21NdF3 1.000 ‐2.03  0.02 2.48  0.02
Nd2O3 in 40LiF‐60NdF3 0.994 ‐3.3  0.2 3.0  0.1
Nd2O3 in 50LiF‐50NdF3 0.999 ‐4.47  0.08 3.64  0.06
Nd2O3 in 58LiF‐42NdF3 1.000 ‐6.29  0.02 4.75  0.01
* with only two data

Based on the study of Raman spectra of NdF3‐LiF‐Nd2O3 system, Stefanidaki proposed a
reaction between Nd2O3 and the molten NdF3‐LiF 10:

Nd2O3 +NdF 3 +3(x  1)LiF  3NdOFx( x 1)  3(x  1)Li (2)

This reaction shows that LiF acts only as a F‐ ion donor, and that Nd3+ can support the REO
dissolution process. He also pointed out that the most reasonable candidates of Nd‐O‐F
complex seem to be NdOF43‐ and NdOF54‐ among the mononuclear compounds and Nd2OF106‐
and Nd2OF84‐ among the binuclear complexes 10.

Hu described the Nd2O3 dissolution in the melt as a process in which the long‐range order in
3‐
Nd2O3 is destroyed by the interaction between O2‐ and ion complex NdF6 and NdF4‐ to form
NdOF54‐ or Nd2OF106‐ 11. Although the formula of the complexes is not exactly the same as
Stefanidaki’s, his finding confirmed the important role of NdF3 content in Nd2O3 dissolution
into molten LiF‐NdF3. The Nd‐F ion complex increases with the content of NdF3 in the binary
system, which is beneficial to Nd2O3 dissolution 11, 13.

Similar interaction of oxygen ions with RE‐F complexes is expected in other REO‐fluoride
systems, resulting in a positive influence of RF3 content on REO dissolution.

xAF, mol.% (binary system)
0.9 0.8 0.7 0.6 0.5 0.4 0.3
2.8
2.4
2.0
Y2O3 in LiF-YF3 at 1023K

1.6
SREO, mol.%

1.2 Y2O3 in LiF-YF3 at 1273K
Nd2O3 in LiF-NdF3 at 1173K
0.8 Nd2O3 in LiF-2BaF2-NdF3 at 1173K
Nd2O3 in LiF-2BaF2-NdF3 at 1123K
Nd2O3 in LiF-2BaF2-NdF3 at 1073K
0.4 Nd2O3 in LiF-NdF3 at 1323K
0.0
0.1 0.2 0.3 0.4 0.5 0.6 0.7
xREF, mol.%

Figure 2 Solubility of rare‐earth oxides in fluoride melts as a function of rare earth fluoride
content (reproduced with data from reference 8 ‐ 11)

Alkali metal fluoride (AF) content
The main purposes of an alkali metal fluoride addition are to lower the melting point of the
electrolyte and improve the electrical conductivity.

The melting points of rare earth fluorides are all higher than 1400 K. Those of light lanthanide
fluorides, the main elements prepared by molten salt electrolysis, are even as high as 1650 ‐
1766 K (see Table 3). However, the cell operation temperature is usually below 1400 K to
secure smooth production and good quality and yield of final products. The addition of LiF
can significantly lower the melting points of the melts. The eutectic points are 19 ‐ 27 mol. %
RF3 (around 70 wt. %) and around 1000 K (see Table 3). The addition of NaF and KF can lower
the melting points of the binary systems as well 14.

A study from Hu showed that the electrical conductivity of NdF3‐LiF melts increased with
temperature and the LiF content 15‐16. The radius of Li+ ion is much smaller than the other
cations, resulting in lower resistance of its movement and higher mobility in the melt.
Therefore, the conductivity is closely linked to the speed of Li+ ions and the obstruction of
such complexes as NdF63‐ and NdF4‐ to their movement under electric field. It is easy to figure
out that the melts with the higher concentration of Li+ ions and less complex NdF63‐ and NdF4‐
would have higher conductivity.


Table 3 Melting point of RF3 and eutectic temperature of RF3‐LiF binary systems

Fluoride YF3 LaF3 CeF3 PrF3 NdF3 SmF3 GdF3 DyF3
Melting point *, K 1421 1766 1710 1672 1650 1579 1504 1427
Eutectic temperature (RF3‐LiF) **, K 968 1043 1028 1023 1011 971 973 973
* data from reference 17
** data from reference 14

Since higher REO solubility can be obtained in melts with higher RF3 content, the AF content
in a binary melt ( xAF  1  xRF3 ) should have a negative effect (see Figure 2).

Pshenichny compared solubility of lanthanide oxides in fluorozirconate melts and found that
the solubilities decrease in the increasing order of Ln3+ ion radius, i.e. SLa2O3 > SSm2O3 > SHo2O3,
and in the increasing order of alkali metal cation radius in the melts, i.e. SLiF > SNaF > SKF 12 (see
Figure 3). This is consistent with the results from Ambrová 18. The acidic power of a cation Z
was calculated and used to interpret the dissolution behaviour of oxides in the melts 12. The
acidic power of binary systems decreases sequentially as the cationic radius of alkali metal
increases. The decrease in solubility of lanthanide oxides can then be associated with the
decrease in polarizing power of the cation.

0.8
0.6
SREO, mol.%
Ho2O3 in 51LiF-49ZrF4
0.4 Ho2O3 in 50.5NaF-49.5ZrF4
Ho2O3 in 58KF-42ZrF4
La2O3 in 51LiF-49ZrF4
La2O3 in 50.5NaF-49.5ZrF4
0.2 La2O3 in 58KF-42ZrF4
Sm2O3 in 51LiF-49ZrF4
Sm2O3 in 50.5NaF-49.5ZrF4
0.0 Sm2O3 in 58KF-42ZrF4
850 900 950 1000 1050 1100
T, K

Figure 3 Solubility of lanthanide oxides in fluorozirconate melts (reproduced with data from
reference 12)

Alkali earth metal fluoride (AF2) content
The addition of AF2 in the melts can further lower their melting points together with limiting
the evaporation of LiF and lowering the melt viscosity 2.


Wu investigated the melt composition before and after dissolution tests and confirmed the
positive impact of BaF2 on preventing evaporation of LiF due to the formation of complexes
8
.

Figure 4 shows the influence of AF2 addition on the REO solubility in LiF‐MF2‐NdF3 melts (M =
Mg, Ba). In some cases it has a positive effect while sometimes negative, but the variations
are moderate. This suggests that the REO solubility should stay low and stable with AF2
addition. It could be meaningful in some cases. For example, Nd‐Mg alloy prepared by co‐
electrowinning from an LiF‐MgF2‐NdF3 melt containing dissolved Nd2O3 should be possible as
the low solubility of Nd2O3 is not reduced by the MgF2 addition.

2.8
2.4
2.0
SREO, mol.%
1.6 LiF-MgF2-15NdF3 at 1073K

0.4 LiF-MgF2-15NdF3 at 1133K
LiF-MgF2-30NdF3 at 1133K
72LiF-BaF2-NdF3 at 1073K
0.0
0.00 0.02 0.04 0.06 0.08 0.10 0.12
xMF2, mol.%

Figure 4 Solubility of Nd2O3 in LiF‐MF2‐NdF3 melts (M = Mg, Ba) (reproduced with data from
reference 8 and 10)

Conclusions and future work
This paper comprehensively summarises, reconstructs and analyses the available data from
previous publications on the REO solubility in molten fluorides. The REO solubility increases
with temperature and RF3 content. There is a simple linear relationship between the natural
logarithm of the solubility and the reciprocal of the absolute temperature. The alkali metal
fluoride addition, mostly LiF, can lower the melting points of binary systems and improve
their electrical conductivity. The alkali earth metal fluoride, usually as a third additive, can
further lower the melting points and limiting the evaporation of LiF.

As a key factor for optimized electrolysis, study on the REO solubility can help the electrolyte
selection and feed rate control. A more detailed analysis of the solubility data is needed to
find out a method for estimation of the solubility on the basis of thermodynamic calculations,
which would assist selection of electrolyte, aiming for a higher REO solubility without

compromising other electrolyte properties (e.g. viscosity, electrical conductivity). More work
should be done on the structure of the REO‐fluoride system to reveal the mechanism of REO
dissolution. Studies on dissolution kinetics will enable a better feed rate control to lead to a
smoother cell operation.
Acknowledgement
The authors would like to acknowledge the EU FP7 project REEcover (Project ID: 603564) for
financial support of this manuscript preparation.
References
1. 1. K. Binnemans, P. T. Jones, B. Blanpain, T. Van Gerven, Y. Yang, A. Walton M. Buchert, "Recycling of Rare
Earths: A Critical Review", J. Cleaner Prod., 51 (0) 1‐22 (2013).
2. 2. N. Krishnamurthy C. K. Gupta, Extractive Metallurgy of Rare Earths, CRC press, 2004.
3. 3. Y. S. Zhang R. A. Rapp, "Modeling the Dependence of Alumina Solubility on Temperature and Melt
Composition in Cryolite‐Based Melts", Metall. Mater. Trans. B‐Proc. Metall. Mater. Proc. Sci., 35 (3) 509‐
515 (2004).
4. 4. H. M. Kan, N. Zhang X. Y. Wang, "Dissolution Rate Determination of Alumina in Molten Cryolite‐Based
Aluminum Electrolyte", J. Ctr. S. Univ., 19 (4) 897‐902 (2012).
5. 5. T. E. Jentoftsen, O. A. Lorentsen, E. W. Dewing, G. M. Haarberg J. Thonstad, "Solubility of Some
Transition Metal Oxides in Cryolite‐Alumina Melts: Part I. Solubility of FeO, FeAl2O4, NiO, and NiAl2O4",
Metall. Mater. Trans. B‐Proc. Metall. Mater. Proc. Sci., 33 (6) 901‐908 (2002).
6. 6. B. B. E. A. Porter, Determination of Oxide Solubility in Molten Fluorides, U.S. Dept. of the Interior,
Bureau of Mines, Washington, DC, 1961.
7. 7. S. Du, M. Wu, F. Du Y. Liu, "Solubility of Rare Earth Oxides in Alkali and Alkali‐Earth Metal Fluoride
Melts", Chin. Rare Earths, 8 (2) 59‐62 (1987).
8. 8. W. Wu, J. Sun, L. Hai H. Gao, "Nd2O3 Solubility in Fluoride Melt", Chin. Rare Earths, 12 (3) 34‐37 (1991).
9. 9. R. G. Reddy S. G. Kumar, "Solubility and Thermodynamic Properties of Y2O3 in LiF‐YF3 Melts", Metall.
Mater. Trans. B‐Proc. Metall. Mater. Proc. Sci., 25 (1) 91‐96 (1994).
10. 10. E. Stefanidaki, G. M. Photiadis, C. G. Kontoyannis, A. F. Vik T. Ostvold, "Oxide Solubility and Raman
Spectra of NdF3‐LiF‐KF‐MgF2‐Nd2O3 Melts", J. Chem. Soc., Dalton Trans., (11) 2302‐2307 (2002).
11. 11. X. Hu, "Study on Ionic Structure and Its Application of NdF3‐LiF‐Nd2O3 System Melts", Doctoral
dissertation, Northeastern University, 2008.
12. 12. R. N. Pshenichny A. A. Omelchuk, "Interaction of Rare‐Earth Oxides with Binary Molten Mixtures of
Zirconium and Alkali Metal Fluorides", Russ. J. Inorg. Chem., 57 (1) 115‐119 (2012).
13. 13. X. Hu, Z. W. Wang, Z. N. Shi, G. M. Lui, B. L. Gao, X. Z. Cao, X. D. Luo, J. Z. Cui, Y. G. Nie, C. Wang, G. H.
Chen, Y. W. Wu, S. Ma, D. Lu C. Ma "Nd2O3 Solubility in NdF3‐LiF‐Nd2O3 Melts", in Proceedings of 2007 Non‐
Grid‐Connected Wind Power Systems, Edited by W. Gu,M. OprisanL. Sun.American Scholars Press,
Marietta, 2007.
14. 14. R. Thoma, Phase Diagrams of Binary and Ternary Fluoride Systems. In Advances in Molten Salt
Chemistry, J. Braunstein,G. MamantovG. P. Smith, Eds. Springer US: 1975; pp 275‐455.
15. 15. X. Hu, "Electrical Conductivity and Nd Solubility of NdF3‐LiF‐Nd2O3 Melts", Charact. Miner. Met.
Mater., 79‐82 (2007).
16. 16. X. Hu, Z. Wang, B. Gao Z. Shi, "Study on the Electrical Conductivity of NdF3‐LiF‐Nd2O3 System Melts
Determined by Cvcc Technique [J]", J. NE Univ.: Nat. Sci., 29 (9) 1294‐1297 (2008).
17. 17. D. L. Perry, Handbook of Inorganic Compounds, Second Edition, Taylor & Francis, 2011.
18. 18. M. Ambrová, J. Jurišová, V. Danielik J. Gabčová, "On the Solubility of Lanthanum Oxide in Molten Alkali
Fluorides", J. Therm. Calorim., 91 (2) 569‐573 (2008).

ELECTRODEPOSITION OF RARE EARTH METALS IN IONIC
LIQUIDS
E.BOURBOS1a, I.GIANNOPOULOU1, A.KARANTONIS 2, D.PANIAS1b,
I.PASPALIARIS1
1
Department of Mining and Metallurgical Engineering of National Technical university of Athens, 9,
Iroon Polytechniou str. 15780 Zografou Campus, Athens , Greece
2
Department of Chemical Engineering of National Technical university of Athens, 9, Iroon
Polytechniou str. 15780 Zografou Campus, Athens , Greece
(a) ebourbos@metal.ntua.gr, (b) panias@metal.ntua.gr
Abstract
In this paper the available literature on the application of ionic liquids in rare earth metals
electrodeposition, is briefly discussed and the main electrochemical properties of ionic liquids,
in relation with such applications, are presented. In addition, the results of a preliminary
investigation for the suitability of pyrrolidinium‐based ionic liquids on the electrodeposition of
lanthanum are presented.
Introduction
Production of rare earth metals and alloys are conventionally performed by high temperature
molten salts electrolysis, which creates highly corrosive environment and demands high
energy consumption. Recently, the use of organic solvents has been implemented for the
electrodeposition of rare earths; however they were considered inadequate, due to their high
volatility and their inflammability.
Ionic liquids gain increasing attention in the recent years, as electrolytes for the recovery of
metals more electropositive than hydrogen. It concerns for salts that are generally liquid at
ambient conditions, consisting of a bulky organic cation and a smaller organic or inorganic
anion [1]. The variety, the combination, the position and the size of cations and anions adjust
remarkable properties for ionic liquids, such as very low melting point, high chemical and
thermal stability, negligible vapour pressure, ability to dissolve a wide range of organic and
inorganic compounds, important ionic conductivity and broad electrochemical window [2].
These properties of ionic liquids, in combination with the non‐flammability they present,
render them advantageous for the electrodeposition of rare earth metals, in comparison with
the organic solvents currently used [3‐5].
Preliminary studies have proven that ILs are capable solvents and electrolytes of reactive
metals, such as Li, Na, Al, Mg [6‐8]. In this paper the applications of ionic liquids in the
electrodeposition of rare earth metals are reviewed and preliminary results of the
electrodeposition of lanthanum are presented.

Electrodeposition of rare earths in ionic liquids
The electrodeposition of rare earth elements from ionic liquids has been mainly investigated
with four different families of ionic liquids based on their cation. It concerns for imidazolium,
pyrrolidinium, ammonium and the phosphonium‐based ionic liquids with a variety of anions.
Imidazolium‐based ionic liquids
It is the most widely investigated family of ionic liquids for metals electrodeposition.
Especially, the imidazolium chloroaluminate systems have been used successfully for the
electrodeposition of Al and refractory metals such as Ta.
In the case of reduction of rare earth metals from imidazolium‐based ionic liquids, Tsuda et
al. [9‐10] reported the co‐deposition of La with Al from a LaCl3‐saturated Emim‐AlCl3 system
with excess of LiCl and SOCl2 and not in pure Emim‐AlCl3. The lanthanum content of the
resulted Al‐La alloy was 0.05% up to 0.50%, in terms of mass fraction. In addition, the co‐
deposition of La‐Co alloys was studied in the ionic liquid Bmim‐CF3SO3 using copper coils as
cathodes [5]. Moreover, the co‐deposition of Sm and Co in BmimBF4 was reported [5],
through inductive effect and the content of Sm in the final alloy was estimated between 5%
and 40%.
Lin [11] et al in 1993 approached the electrochemical behaviour of Ce(III)/Ce(IV) couple in the
basic aluminium system (AlCl3 – MeEtimCl) and via ultraviolet‐visible spectroscopy it was
stated that Ce(III) and Ce(IV) form complexes such as [CeCl6]x‐, whereas x=3 and 2
respectively. Jagadeeswara et al. [12] researched the electrochemistry of the systems Ce(III),
Sm(III) and Eu(III)/Eu(II) in BmimCl. The application of cyclic voltammetry and
chronopotentiometry at platinum working electrode revealed the reduction of trivalent Ce
and Sm ions to the divalent state at various temperatures and found them to be quasi‐
reversible reactions. In 2006, Yamagata et al. [13] studied the electrochemical behaviour of
redox reactions Sm(III)/Sm(II), Eu(III)/Eu(II) and Yb(III)/Yb(II) in the ionic liquid Emim‐TFSI
through cyclic voltammetry and it was found to be quasi‐reversible or irreversible.
Pyrrolidinium‐based ionic liquids
Glukhov et al. [14] have stated the reduction of Y, Gd and Yb in the ionic liquid BMP‐CF3SO3
on both, platinum and copper electrodes. On the voltammetric curves performed with the Pt
electrodes in the systems Y and Gd with this IL, the observed cathodic peak with a maximum
value of ‐2.6 V (rel. Ag/0.1 M Cl‐) was attributed to the reduction of the trivalent cations to
the metallic state. The electrodeposition of these metals was performed under potentiostatic
or galvanostatic conditions and a tenuous black precipitate was formed on the electrode’s
surface. It was also shown that the reduction of Yb(III) to the metallic form occurs step by
step via Yb(II) formation and the limiting stage of the cathodic process is the adsorption of
the metal cation on the electrode. Legeai et al. [15] have shown that the reduction of La in
OMP‐TFSI is an irreversible reaction and that La electrodeposition in this ionic liquid under

potentiostatic polarisation conditions ‐1.5 V (vs Ag/AgCl), for 120 min, at 298 K, resulted in
the cathodic deposit of a thick La film of 350 nm.
In addition, Yamagata et al. [13] studied by cyclic voltammetry the redox reactions of the
Sm(III)/Sm(II) couple in BMP‐TFSI ionic liquid using a glassy carbon electrode and determined
the redox reactions as quasi‐reversible or irreversible. They also studied the electrochemistry
of Yb(III) in BMP‐TFSI by cyclic voltammetry and concluded to a cathodic and an anodic peak
at around ‐0.95 V (vs Ag/Ag+), which considered attributable to the reduction of Yb(III) and
the oxidation of Yb(II), respectively. The electrochemical behaviour of Eu(III) in BMP‐TFSI was
researched by Rao et al. [16] with glassy carbon and stainless steel electrodes, at various
temperatures. The cyclic voltammograms revealed a quasi‐reversible behaviour of the redox
couple Eu(III)/Eu(II) and an irreversible behaviour of the couple Eu(II)/Eu(0), which presented
a cathodic peak attributed to the reduction of Eu(II) to the metallic state. In the same ionic
liquid (BMP‐TFSI) that contained also chloride anions, Hussey et al. [17] studied the
electrochemical and spectroscopic behaviour of Nd(III) and Pr(III). According to their results,
the electrolytic dissolution of both metals in this ionic liquid’s system produces only the
respective trivalent cations, which can be reduced to the Ln2+(II) state, but the resulting
divalent species exhibit only transient stability, undergoing rapid disproportionation to
Ln3+(III) and Ln0 states.
Ammonium based ionic liquids
Glukhov et al. [14] investigated the reduction of Y, Gd, Yb in the ionic liquid Bu3MeN‐CF3SO3
and concluded that the deposition of the above metals is only possible on a copper substrate
and does not occur on a platinum one. Bhatt et al. [18] reported the reduction of selected
lanthanide cations (La, Sm and Eu) to the zero‐valent state in the ionic liquid Me3NBu‐TFSI.
The lanthanide cations were introduced to the ionic liquid as the TFSI hydrate complexes.
Cyclic voltammograms revealed a cathodic peak at ‐2.4 V vs Fc+/Fc for the lanthanum system
attributed to the reduction to metallic state, whereas two peaks were observed for Sm‐ and
Eu‐ systems, the less negative were associated with (III)/(II) reaction and the more negative
with the reduction to the metallic state.
Phosphonium based ionic liquids
In 2008 Matsumiya et al. [19] studied the electrochemical behaviour of Sm(III) in the
phosphonium based ionic liquids. The diffusion coefficients of Sm(III) were determined via
cyclic voltammetry, chronoamperometry, chronopotentiometry and were estimated to be on
the order of 10‐8 cm2s‐1. It was stated that the trivalent samarium complexes in the
phosphonium based ionic liquids were assumed to present higher ion mobility in respect to
those in the ammonium ionic liquids, due to weaker interaction between the cation and the
anion in phosphonium‐based ILs. In addition, Kondo et al. [20] researched the suitability of
P[225]‐TFSI for the electrodeposition of neodymium. The diffusion coefficient of Nd(III) was
investigated by chronopotentiometry with cylindrically symmetrical diffusion and was
estimated on the order of 10‐11 m2s‐1 at 100 oC. The cyclic voltammograms of Nd(III)/P[225]‐

TFSI containing 0.5 mol dm‐3 NdTFSI3 presented an anodic and a cathodic peak currents that
were attributed to the oxidation and the reduction of Nd(III), respectively. The reduction of
Nd cation was observed at ‐2.5 V vs a platinum wire immersed in the solution used as a quasi‐
reference electrode, compensated for the Fc+/Fc redox couple. The electrodeposition of
neodymium in the specific ionic liquid was performed by potentiostatic electrolysis at ‐3.1 V
vs Fc+/Fc at 150 oC. A black electrodeposit was fine and formed uniformly neodymium particle
and was observed by SEM and identified from the evaluation of EDX.
Moreover, the electrochemical behaviour and electrodeposition of dysprosium (Dy) in the
phosphonium‐based ionic liquid P[225]‐TFSI were investigated by Kurachi et al. [21]. The cyclic
voltammetric measurements resulted in one step reduction of the trivalent dysprosium ion
in phosphonium‐based ionic liquid. On the other hand, no anodic peak ascribed to the
oxidation of dysprosium metal was observed in this electroanalytical study. The diffusion
coefficient and the activation energy for diffusion of the trivalent Dy complex in this IL were
also estimated, using semi‐integral analysis. The diffusion coefficient of Dy(III) was calculated
to be 2.0×10‐12 m2s‐1 at 25 oC, closed to that of the trivalent lanthanoid ion such as Eu(III) and
Sm(III) in phosphonium‐based ionic liquids. In addition, the activation energy for diffusion
was estimated to be 65 kJmol‐1 (0.5 M) and 49 kJmol‐1 (0.075 M).
Chlonine based ionic liquids
In 2012 Ishii et al. [22] investigated the physical and electrochemical properties of choline‐
based ionic liquids and their capacity to reduce iron and neodymium cations. The novel
cationic blended ionic liquids which were constituted by the 2‐
hydroxyethyltrimethylammonium bis(trifluoromethylsulfonyl)imide Ch‐TFSI and P[225]‐TFSI
was applied for the electro‐recovery of neodymium. In this study the starting materials were
voice coil motors for hard disk drivers and the electrodeposits obtained were examined by
SEM/EDX and XPS. According to the results of this study, iron group and rare earth metals can
be recovered separately by electrodeposition in two stages.
Preliminary studies on lanthanum electro‐recovery from

pyrrolidinium‐based ionic liquids.
Experimental work of the specific research focused on evaluating the use of pyrrolidinium‐
based ionic liquids for the electro‐recovery of reactive metals, such as the Rare Earth Metals
are. More precisely, the reduction of lanthanum in the ionic liquid N‐butyl‐N‐
methylpyrrolidinium bis(trifluoromethylsulfonyl)imide (BMP‐TFSI) was studied, at room
temperature. BMP‐TFSI was selected among the different pyrrolidinium‐based ionic liquids,
because it presents electrochemical stability under cathodic polarisation, satisfying ionic
conductivity and it is an air and water stable electrolytic medium.
Materials and apparatus
The ionic liquid BMP‐TFSI was supplied by Solvionic and Lanthanum nitrate hexahydrate was
purchased by Alfa Aesar. The lanthanum nitrate salt was heated at 100 oC under vacuum for

72 h and was dissolved in pure acetone, where subsequently the proper amount of electrolyte
was added to form a solution’s concentration of the order of 0.06 M. The solution was placed
at 60 oC under vacuum for approximately 2h to remove acetone and then at 100 oC under
vacuum for 24 h to remove residual water.
Linear and cyclic voltammetry tests were performed in a three microelectrodes cell (PAR)
connected to a VersaSTAT3 potentiostat (PAR); the obtained experimental data were
analysed with the VersaStudio software (PAR). The working electrode was a platinum disk of
d = 1.98 mm, as a counter electrode a Pt wire immersed directly into the solution was used
and finally, as a reference electrode the redox couple Ag/0,1 M AgNO3 in acetonitrile was
employed. The working electrode was polished with aluminum paste 1 μm on a velvet pad
and electrochemically in 1 M sulfuric acid. Although BMP‐TFSI exhibits hydrophobic
characteristics and it is stable under normal atmospheric conditions, the cyclic voltammetry
and chronoamperometry measurements were performed in a glove box, providing an inert
atmosphere by purging nitrogen gas to minimise oxygen and moisture contamination. The
morphology of electrodeposits was examined by Scanning Electron Microscope
(JEOL6380LV), provided with Energy Dispersive Spectrometry.
Results
The electrochemical window was determined at room temperature by cyclic voltammetry
and the cathodic limit of the electrolyte was determined at ‐3.5 V vs Ag/Ag+. Figure 1 shows
the cyclic voltammogram recorded in the system BMP‐TFSI/La. As it is obvious, a generation
of an intense cathodic peak begins at ‐2.2 V vs Ag/Ag+, attributed to the reduction of trivalent
lanthanum cation (La3+) to the metallic state. The absence of corresponding anodic peak on
the reverse scan allows the assumption that lanthanum reduction is irreversible [15].
Moreover, the cathodic peak at ‐1 V is ascribed to the limited reduction of hydrogen cations,
due to slight moisture [2] that inevitably was present and the narrow shoulder at ‐1,5 V is
attributed to an adsorption La3+ reaction occurring on the electrode, prior to the reduction of
lanthanum [14, 15].

Figure 1: Cyclic voltammograms recorded with scan rate 20mV/s vs Ag/Ag+ on a Pt electrode
at 25 oC: (1) BMP‐TFSI and (2) for the solution of 0.06 M La(NO3)3 in BMP‐TFSI.

Furthermore, the electrodeposition of lanthanum was realised on a copper substrate, under
potentiostatic conditions for 5 h at ‐3.1 V (vs Ag/Ag+) at 25 oC. The electrodeposits were
examined in SEM and EDS analysis (Figure 2) revealed the electrodeposition of lanthanum.
The EDS analysis also revealed the existence of fluorine and sulfur that may occur, due to the
TFSI anion adsorption on the electrode's surface that could explain the absence of anodic
peak.

Figure 2: SEM image depicting the electrodeposits of lanthanum(left),
EDS analysis of the electrodeposits(right)


Conclusions
Ionic liquids are promising electrolytes, suitable for the electrodeposition of rare earth
metals. The research in this area has proven their ability to reduce and to permit the
electrodeposition of drastic metals, without the drawbacks present in the currently used
technology.
The preliminary electrochemical investigation of BMP‐TFSI revealed its potential use for
reduction of lanthanum and appears, due to its physicochemical properties, to be an
appropriate medium for the electrodeposition of rare earth elements.

Acknowledgments
“The research leading to these results has received funding from the European Community’s
Seventh Framework Programme ([FP7/2007‐2013]) under grant agreement n°309373. This
publication reflects only the author’s view, exempting the Community from any liability”.
Project web site: www.eurare.eu”
References
1. H. Ohno “Electrochemical aspects of ionic liquids”, John Wiley & Sons, Inc., Hoboken, New Jersey, (2005).
2. H. Ohno, in: F.Endres, A.P. Abbott, D.R. Mcfarlane (Eds.), Electrodeposition from ionic liquids, Wiley‐VCH
Verlag GmbH & Co., 2008.
3. T.Tsuda, C.L. Hussey “Electrochemical applications of room temperature ionic liquids”, The Electrochemical
Society Interface, Spring 2007, 42‐49.
4. R.E. White, T. Tsuda, C.L. Hussey “Modern aspects of Electrochemistry” No 45, 2009, 63‐174.
5. Y. Cui, Y. Hua, Y. Lin. “Applications of ionic liquids in electrodeposition of rare earths”, J. Chongqing Univ:
Eng Ed [ISSN 1671‐8224], 2010, 9(4): 167‐176.
6. A. Ispas and A. Bund “Electrodeposition from ionic liquids”, The Electrochemical Society Interface, Spring
2014, 47‐51.
7. A.Ispas, B. Adolphi, A.Bund, F.Endres. “On the electrodeposition of tantalum from three different ionic
liquids with the bis(trifluoromethyl sulfonyl) amide anion” J.Phys. Chem. Chem. Phys., 2010, DOI:
10.1039/b922071m.
8. T. Jiang, M.J. Chollier Brym, G. Dubé, A. Lasia, G.M. Brisard. “Electrodeposition of aluminium from ionic
liquids:Part I electrodeposition and surface morphology of aluminium from aluminium chloride (AlCl3)–1‐
ethyl‐3‐methylimidazolium chloride ([EMIm]Cl) ionic liquids”, J. Surface & Coatings Technology 201 (2006)
1–9.
9. T. Tsuda, T. Nohira, Y. Ito.”Electrodeposition of lanthanum in lanthanum chloride saturated AlCl3 ‐1‐ethyl‐
3‐methylimidazolium chloride molten salts” J. Electrochimica Acta ISSN 0013‐4686, 2001,46 (12), 1891‐
1897.
10. T. Tsuda, T. Nohira, Y. Ito. “Nucleation and surface morphology of aluminum‐lanthanum alloy
electrodeposited in a LaCl3‐saturated AlCl3‐EtMeImCl room temperature molten salt” J. Electrochimica
Acta ISSN 0013‐4686, 2002, 47(17), 2817‐2822.
11. F.M. Lin, C.L. Hussey. ”An Electrochemical and spectroscopic study of cerium in the basic aluminum
chloride‐1‐methyl‐ethylimidazolium chloride molten salt” J. of the Electrochemical Society, 1993,140 (11),
3093‐3096.
12. R.C. Jagadeeswara, K.A. Venkatesan, K. Nagarajan et al. “Electrochemical and thermodynamic properties
of europium(III), samarium(III) and cerium(III) in 1‐butyl‐3‐methylimidazolium chloride ionic liquid” J. of
Nuclear Material, 2010, 399(1), 81‐86.

13. M. Yamagata, Y. Katayama, T. Miura. ”Electrochemical behavior of samarium, europium and ytterbium in
hydrophobic room temperature molten salt systems” J. of the Electrochemical Society, 2006, 153(1), E5‐
E9.
14. L.M. Glukhov, A.A. Greish, L.M. Kustov. “Electrodeposition of rare earth metals Y, Gd, Yb in ionic liquids” J.
Russian Journal of Physical Chemistry A, 2010, 84(1), 104‐108.
15. S.Legeai,S. Diliberto, N.Stein et al.” Room‐temperature ionic liquid for lanthanum electrodeposition” J. of
Electrochemistry Communications, 2008, 10(11), 1661‐1664.
16. C.J. Rao, K.A. Venkatesan, K. Nagarajan, et al.”Electrochemical behavior of europium(III) in N‐butyl‐N‐
methylpyrrolidinium bis(trifluoromethylsulfonyl)imide J.Electrochimica Acta, 2009, 54(20), 4718‐4725.
17. C.L.Hussey and L.H.Chou. “An Electrochemical and spectroscopic study of Nd(III) and Pr(III) coordination in
the 1‐butyl‐1methylpyrrolidinium bis(trifluoromethylsulfonyl)imide ionic liquid containing chloride ion” J.
Inorg.Chem. 2014, 53, 5750‐5758.
18. Al. Bhatt, I. May, V.A. Volkovich, et al. “Structural characterisation of a lanthanum bistriflimide complex
La(N(SO2CF3)3(H2O)3 and an investigation of La, Sm and Eu electrochemistry in a room‐temperature ionic
liquid, [Me3NnBu][N(SO2(CF3)2” J.Inorg.Chem. 2005, 44, 4934‐4940.
19. M. Matsumiya, S. Suda, K. Tsunashima et al. “Electrochemical behaviors of multivalent complexes in room
temperature ionic liquids based on quaternary phosphonium cations” J. of Electroanalytical Chemistry,
2008, 622(2), 129‐135.
20. H. Kondo, M. Matsumiya, K. Tsunashima, S. Kodama. “Attempts to electrodeposition of Nd from ionic
liquids at elevated temperatures” J. Electrochimica Acta 66, 2012, 313‐319.
21. A. Kurachi, M. Matsumiya, K. Tsunashima, S. Kodama. “Electrochemical behavior and electrodeposition of
dysprosium in ionic liquids based on phosphonium cations” J. Appl. Electroch., 2012, 42, 961‐968.
22. M. Ishii, M. Matsumiya, S. Kawakami. “Development of recycling process for rare earth magnets by
electrodeposition using ionic liquids media” J. ECS. Transactions, 50(11), 549‐560, 2012.


ELECTROCHEMICAL RECOVERY OF RARE EARTH ELEMENTS
FROM MAGNET SCRAPS‐ A THEORETICAL ANALYSIS
V. PRAKASH*, Z.H.I. SUN, J. SIETSMA, Y. YANG
Department of Materials Science and Engineering, Delft University of Technology Mekelweg 2,
2628CD, Delft, The Netherlands
*P.Venkatesan@tudelft.nl
Abstract
Recovery of rare earth elements (REEs) from secondary resources is of great environmental
and economic importance. In the case of magnet scraps, multi‐step hydrometallurgy and
liquid metal extraction with high energy consumption are usually applied to accomplish
acceptable recovery rate of REEs from the scrap. This paper aims to provide theoretical
framework for using electrochemistry to selectively extract REEs into aqueous solution.
Thermodynamic and electrochemical behaviour of different components in the alloy are
theoretically predicted and correlated with available data from the literature. The role of
microstructure and different elements in dissolution mechanism of the REE magnet is
reviewed. Through understanding various factors involved in dissolution behaviour of the
alloy, an electrochemical process with enhanced selectivity to recover REEs can be
established.
Introduction
Rare earth elements are of great importance in modern industry for producing various
functional materials like permanent magnets, rechargeable batteries, catalysts, lamp
phosphors because of their unique magnetic and electronic properties. Due to the increased
demand and supply shortage, REEs are listed as critical raw materials of strategic importance
in many countries in the world1,2. This has promoted research for recovering REEs from end‐
of‐life (EOL) products and the scrap created during production. Neodymium iron boron
magnet (Nd2Fe14B) is one of the major usages for rare earth elements and has 62%3of the
market share of permanent magnetic material. Apart from neodymium, REEs like dysprosium
and praseodymium are used in varying amounts (in total 25‐30%) for obtaining required
magnetic properties4. Cobalt (0‐10%) is often added to increase the Curie temperature and
other elements like vanadium, oxygen and nitrogen can be added in trace amounts to reduce
corrosion5a,5b.

The complex composition of EOL magnets and magnet scrap makes the recovery of REEs
rather difficult and demands a range of metallurgical and chemical techniques Various
methods have been reported in the past for recovering neodymium and other rare earth

elements from the magnets. Lyman et al.6 used the leaching method and have obtained high
recovery of neodymium (98%) by dissolving the scrap in sulfuric acid followed by using alkali
hydroxides to precipitate double salts of neodymium which can be further treated to obtain
neodymium oxide. Other research groups have also tried selective leaching 7, ultrasound‐
assisted leaching8 and a combination of pre‐roasting the scrap along with selective leaching
and solvent extraction9 to recover the REEs from the scrap. Although the above mentioned
hydrometallurgical methods have achieved appreciable recovery of REEs, they require
multiple steps to enhance leaching selectivity, expensive pre‐treatment techniques and
consumption of large amounts of chemicals. Hence for an aqueous chemical method to be
economically viable for recycling, enhanced selectivity of REEs recovery combined with low
consumption of chemicals and reduced number of steps involved is of paramount
importance.

In this paper, the thermodynamics and the corresponding electrochemical behaviour of
various elements present in waste magnets are discussed and compared during dissolution
into an aqueous solution. With improved understanding of these behaviour, a process to
selectively extract REEs from waste magnet can be established.

Thermodynamic evaluations

A typical composition of commercial magnet and magnet scraps obtained from literature9,10,11
is listed in Table 1.

Table 1: Typical composition of magnet scraps
Elements Nd Dy Pr Fe B Co C N Others
Wt% 23‐25 3.5‐5 0.05‐ 62‐69 1 0‐10 0‐0.14 0‐0.1 1‐2

5
*Others‐ Cu, Al, Ga, Si, Gd, Mo.

The standard Gibbs energy of formation for different rare earth species are calculated from
the literature12,13 and their corresponding Nernst equation are summarized in Table 2. It can
be noticed that the standard electrode potentials (E RE3+/RE = ‐2.2 to ‐2.4 V vs SHE) of the
rare earth elements are very close owing to their similar chemical nature.

0
Table 2: Calculated values of ∆G (kJ/mol) and standard reduction potential at 298 K and the
corresponding Nernst equation. Soluble species concentrations (except H+) is 10−1.0 M

Reaction ∆G0 (KJ/mol) E0 (V) Nernst equation
Nd 3e ⇔ Nd ‐671.6 ‐2.32 E = ‐2.32 + 0.020 log [Nd3+]
Nd 3H2 O⇔Nd OH 3 3H 3e ‐563.3 ‐1.94 E = ‐1.94 +0.059 pH
Nd3 3H2 O⇔Nd OH 3 3H 108.3 3 pH = 19 – log [Nd3+]
Dy 3e ⇔ Dy ‐664.0 ‐2.29 E = ‐2.29 + 0.020 log [Nd3+]
Dy 3H2 O⇔Dy OH 3 3H 3e ‐564.4 ‐1.95 E = ‐1.95+ 0.059 pH
Dy 3H2 O⇔Dy OH 3 3H 98.6 3 pH = 17.3 – log [Dy3+]
Pr 3e ⇔ Pr ‐680.3 ‐2.35 E = ‐2.35 + 0.020 log [Pr3+]
Pr 3H2 O⇔Pr OH 3 3H 3e ‐568.0 ‐1.96 E = ‐1.96 +0.059 pH
Pr 3H2 O⇔Pr OH 3 3H 111.9 3 pH = 19.68 – log [Pr3+]

Their very negative potential values which fall below the hydrogen evolution line indicate a
strong reactivity in aqueous solutions. Experimental observation of corrosion and passivity
behaviour of neodymium in different conditions have also confirmed rigorous hydrogen
evolution14. In particular, comparison of the behaviour of iron and REEs in aqueous solution
at different pH can be of importance. Koyama et al.15 used a Pourbaix diagram (Fig. 1) to see
stability of different species of neodymium and iron in different pH and they selectively
leached out Nd3+ by leaving out iron as Fe2O3 in the solution.

In contrast to selective leaching where pH plays a vital role, using scrap magnets in an
electrochemical cell offers an additional advantage of controlling parameters like potential
and current density to manipulate and selectively dissolve REEs. Fig. 2 shows standard
electrode potentials of different components of scrap magnets vs Ag/AgCl electrode (0.25 V
vs SHE). The other major elements present in the waste magnets in terms of weight
percentage are iron and cobalt. The standard reduction potentials of iron and cobalt are ‐0.44
V and ‐0.28 V vs SHE respectively. EOL magnets also contain copper, as a minor additive or in
the coating along with nickel to prevent corrosion.

Fig. 1: Pourbaix diagram of Fe‐H2O and Nd‐H2O system9

It can be noted that, when magnets are deployed as anodes in a 3‐electrode electrochemical
cell (Ag/AgCl reference electrode), selective dissolution can be accelerated by controlling the
potential and maintaining it in the region between dissolution of rare earths and the region
in which metallic iron, cobalt and other elements are stable. Further anodic polarization of
the scrap magnets can accelerate the dissolution process.

Fig. 2: Standard reduction potential (V vs Ag/AgCl) of different elements in scrap magnet9a

Analyzing elements individually can theoretically form a basis for selective dissolution but
dissolution of individual components from an alloy is governed by their partial
thermodynamic properties. Partial potential of an individual component in an alloy can be
written as 16

ln N (1)
where is the difference between equilibrium electrode potential of the component i in
the alloy and the corresponding potential of the same component in the individual phase. R,
T, F ,Zi, Ni are molar gas constant, temperature, Faraday’s constant, number of electrons
transferred and molar fraction of the component respectively. By definition mole fraction is
less than one and hence the partial potential of a component in the alloy will always be more
positive than the individual component potential. Hence experimental determination of the
extent of positive shift for individual components in the Nd2Fe14B alloy in different electrolytic
environments is vital.

An electrolyte environment which is non‐inhibiting and encouraging free corrosion process is
essential. Passivity is in principle possible in highly alkaline solutions and Song et al.17
observed that NdFeB magnets form passive hydroxide layers in sodium hydroxide and in
oxalic acid. Jingwu et al.18 observed similar type of passive behaviour in phosphoric acid with
low corrosion current. However the alloy freely dissolves in other acids such as sulfuric acid,
nitric acid and hydrochloric acid with high current density and is also susceptible to attack by
chloride and other ions even in the solutions in which they exhibit passivity. Hence for active
dissolution of the magnets it is preferred to have an acidic conditions with strong acids having
high corrosion current.

Microstructure of the magnet and dissolution mechanism

Another crucial aspect of selective dissolution is to understand the correlation between
microstructure of the alloy and the corresponding dissolution mechanism. NdFeB magnets
have a matrix ferromagnetic phase (Φ) of Nd2Fe14B tetragonal compound and is surrounded
by intergranular regions containing a neodymium rich phase (n) and a boron rich phase (η) 19.
The complexity of the intergranular region in the microstructure of the neodymium magnets
depend upon the additional elements such as cobalt, aluminum and gallium . In order to
selectively dissolve rare earths, an understanding of dissolution mechanism in different
environments is thus necessary. Several authors19,20,23 have proposed that neodymium rich

phases (n) corrode preferentially due to the formation of a galvanic couple owing to negative
standard potential of rare earths. This is followed by the boron rich phase dissolution and this
renders the matrix loose finally creating disruption of the matrix phase. Corrosion in Nd‐Fe‐B
alloy is a natural process and can even result in the pulverization of the magnet. The scheme
of step by step dissolution process presented by Schlutz19 et al. is given in Fig. 3.

Fig. 3: Schematic illustration of dissolution process of Nd‐Fe‐B magnets by corrosion19

Mao et al.20 observed the dissolution process of individual phases by synthesizing them
separately and subjecting them to different corrosion tests. The Nd rich phase had the most
negative open circuit potential followed by matrix phase and the boron rich phase, indicating
that the Nd rich phase has the highest electrochemical reactivity. These results further
strengthen the conclusion that selective dissolution of rare earths are feasible and is in fact a
naturally occurring corrosion process.

The composition of different components in the alloy also play a crucial role in dissolution
process. Corrosion studies of magnets with higher weight percent of iron in phthalate buffer
showed reduced current density, indicating formation of passive layers like iron oxide
preventing corrosion21. Aluminum, gallium and copper have reduced the corrosion as they
reduce the strength of galvanic coupling among magnetic phases22 and cobalt has been
observed to improve corrosion resistance as it prevents magnetic pulverisation23. High rate of
dissolution is observed in low carbon and oxygen alloy when the nitrogen content exceeds
0.1 wt%. Oxygen impedes corrosion when it is added between 0.6‐1.2 wt% and the optimum
amount of carbon was found to be 0.1wt% as anything above that facilitates corrosion5a.
Dysprosium and niobium have been said to have formed stable intermetallic phases which
retards corrosion24 but also high amount of dysprosium (16 %) has been deemed as
disadvantageous 25. On the whole, it is important to know the nature of the scrap to
accelerate the dissolution process by various means and by altering the environment.


Electrochemical cell: cathode products and anode slime

Table 2 lists the possible anodic reactions in the electrochemical cell. Besides dissolution of
rare earth metals, co‐dissolution of iron and cobalt by chemical or electrochemical means is
also possible although not desirable.

Rare earths, due to their very negative potential liberate hydrogen on the cathode in aqueous
solutions from an electrochemical process. Due to the same, the rate determining step
becomes mass transport as the charge transfer is relatively fast. The cathodic reactions in
such case are

2H 2e ⇔ H (2)

M ne ⇔ M (3)

2H 0 2e ⇋ H 2 OH (4)

Hydrogen evolution reaction though inevitable, can be kinetically slowed down with proper
choice of a cathode as it will hinder deposition of non‐REEs as well as modify the pH. Schlutz
et al.14 observed water reduction to be the prominent cathodic reaction for pure neodymium
at lower concentrations (≤0.01 M) and at higher concentrations they observed hydrogen
evolution. In addition to just acting as counter electrode, the cathode can also be chosen in
such a way that it allows deposition of co‐dissolved non‐REEs. In order to have a cathode to
allow deposition of non‐REEs, it should a) be inexpensive b) have slow kinetics and low
exchange current density for hydrogen evolution reaction c) favor the deposition of non‐REEs
like iron, nickel or cobalt. Besides if such a deposit from scraps can also be attractive in terms
of commercial applications (such as FeNi alloys) it can be an added benefit.

A simple schematic is illustrated in Fig. 4 for the conceptual framework of the electrochemical
cell described in the paper. Fig. 4 a) shows the scrap magnets placed as anodes in a simple 2‐
electrode electrochemical cell

Fig.. 4 Simple schematic of the proposed electrochemical method for REE recovery

Fig.. 4 b) shows an in‐progress dissolution. Although it is preferred to selectively dissolve only
rare earth ions, chemical dissolution and the complex microstructure of the alloy might also
encourage co‐dissolution of non‐REEs. These co‐dissolved non‐REEs can be further deposited
back on the cathode which results in a partially purified solution enriched in rare earths which
is suitable for further elemental recovery and the rest of undissolved elements can be
collected as anode slime. The proof of principle experiments are currently in preparation.

Conclusion

This paper establishes the theoretical framework required to perform an electrochemical
recycling of end‐of‐life scrap NdFeB magnets. By drawing ideas from different fields such as
electrodeposition, selective dissolution, corrosion and also by projecting thermodynamic
predictions for different rare earths of our interest, it can be observed that not only enhanced
selective recovery of rare earths is possible but also it is, to an extent, a naturally favored
process. The next step of this research will be to build up a 3‐electrode electrochemical cell
which can help establish the proof of concept described in the paper.


A more detailed look into the dissolution mechanism and kinetics of scrap magnets by using
complimentary tools such as electroanalytical techniques and microscopy will be essential.
Once the dissolution mechanism is understood, influence of different parameters such as pH,
temperature, potential difference and electrolyte composition can be analyzed and
optimized. The further goal will be to build up a lab scale cell which can handle few hundreds
of grams of scrap in an efficient manner.

Acknowledgment
The authors wish to acknowledge the EU FP7 Marie Curie project EREAN (Project ID 607411)
for financial support.

References
1. Critical raw materials for the EU, “Report of the ad‐hoc working group on defining critical raw
materials”, (2010).
2. US Department of Energy, “Critical materials strategy”, (2010).
3. Oliver Gutfleisch , Matthew A. Willard , Ekkes Brück , Christina H. Chen , S. G. Sankar , and J. Ping
Liu, “Magnetic Materials and Devices for the 21st Century: Stronger, Lighter, and More Energy
Efficient”, Advanced materials, 23 821–842 (2011).
4. M. Sagawa, S. Fujimura, H. Yamamoto, Y. Matsuura and K. Hiraga, "Permanent Magnet Materials
Based on the Rare‐Earth‐Iron‐Boron Tetragonal Compound”, IEEE Trans. Magnetics 20 1584‐ 1589
(1984).
5. a) A. S. Kim, F. E. Camp and E. J. Dulis, Effect of oxygen, carbon and nitrogen content on corrosion of
Nd‐Fe‐B magnets, IEEE Trans. Magnetics,26 (5) 1936‐1938 (1990) .
b) M. Sagawa, S. Fujimura, H. Yamamoto, Y. Matsuura and K. Hiraga, IEEE Trans. Magnetics 20 (5)
1584‐1589 (1984)
6. J.W Lyman & G.R.Palmer, Report of investigations 9481, US bureau of mines, 1993.
7. Lee, J.‐c., Kim, W.‐b., Jeong, J., Yoon, I.J., J. Korean Inst. Met. Mater. 36, 967 (1998)
8. Tanaka, Y., Zhang, Q.W., Saito, F, “Sonochemical recovery of metals from recording media”, J.
Chem. Eng. Jpn. 35 173‐177 (2002).
9. Mikiya Tanaka, Tatsuya Oki, Kazuya Koyama, Hirokazu Narita and Tetsuo Oishi, “Handbook on the
Physics and Chemistry of Rare Earths”, 43 159‐255 (2013).
9a) Standard Potentials in Aqueous Solution. Allen J. Bard, Roger Parsons, Joseph Jordan
CRC Press, Aug 27, 1985 (Values adjusted to Ag/AgCl reference electrode by researcher)
10. Koji Miura, Masahiro Itoh, Ken‐Ichi Machida, “Extraction and recovery characteristics of Fe element
from Nd–Fe–B sintered magnet powder scrap by carbonylation”, Journal of Alloys and Compounds,
466 (1‐2) 228‐232 (2008).
11. Amjad Ali, “Magnetic properties and corrosion resistance of coated and uncoated rare earth based
permanent mangnets”, PhD thesis , University of Engineering and Technology Lahore, 2011.
12. George K. Schweitzer, Lester L. Pesterfield, “The aqueous chemistry of elements”, Oxford university
press, 2010.

13. Allen J.Bard, Roger Parsons, Josep Jordan, “Standard potentials in aqueous solution”, CRC press,
1985.
14. R. Sueptitz , M. Uhlemann, A. Gebert, L. Schultz, “Corrosion, passivation and breakdown of passivity
of neodymium”, Corrosion Science 52 886‐891 (2010)
15. Koyama, K., Kitajima, A., Tanaka, M., Rare Earths 54 36‐37 (2009)
16. I.K.Markashov, “Anodic Dissolution and Selective Corrosion of Alloys”, Protection of Metals Vol. 38(
2) 139‐145 (2002)
17. Y. W. Song, H. Zhang, H. X. Yang and Z. L. Song, “A comparative study on the corrosion behavior of
NdFeB magnets in different electrolyte solutions”, Materials and Corrosion 2008 59(10) 794‐801
(2008)
18. Zheng Jingwu, Jiang Liqiang, Chen Qiaoling , “ Electrochemical corrosion behaviour of Nd‐Fe‐B
magnet in different acid solutions” , Journal of Rare earths 24 218‐222 (2006)
19. L. Schultz , A.M. El‐Aziz, G. Barkleit, K. Mummert, “Corrosion behaviour of Nd–Fe–B permanent
magnetic alloys “, Materials Science and Engineering, A267 307 (1999)
20. Shoudong Mao, Hengxiu Yang, Zhenlun Song , Jinlong Li, Huagen Ying, Kefei Sun, “Corrosion
behaviour of sintered NdFeB deposited with an aluminium coating “, Corrosion Science 53 1887‐
1894 (2011)
21. M. Rada, J. Lyubina, A. Gebert, O. Gutfleisch, L. Schultz, “Corrosion behavior of Nd–Fe–B/α–Fe
nanocomposite magnets” , Journal of Magnetism and Magnetic Materials 290–291 1251–1254
(2005)
22. A‐A.El Moniem, “Passivity and its breakdown of sintered NdFeB‐based magnets in chloride
containing solution”, Corrosion Science 46 2517–2532 (2004)
23. S. Szymura, H. Bala and G. Pawlowska, Yu. M. Rabinovich, V. V. Sergeev and D. V. Pokrovskii,
“Modification of the magnetic properties and corrosion resistance of Nd‐Fe‐B permanent magnets
with addition of cobalt”, Journal of the Less‐Common Metals, 175 185‐198 (1991)
24. L.Q. Yu, Y.H. Wen, M. Yan, “Effects of Dy and Nbon the magnetic properties and corrosion
resistance of sintered NdFeB”, Journal of Magnetism and Magnetic Materials 283 353–356 (2004)
25. S. Szymura, H. Bala and G. Pawlowska, Yu. M. Rabinovich , “Effect of impurities on the corrosion
behaviour of neodymium”, Journal of Applied Electrochemistry, 23 1017‐1024 (1993)


PROPOSAL OF A NEW TYPE OF SULPHIDE ELECTROLYTE FOR
RARE EARTH MOLTEN SALT ELECTROLYSIS
Johannes ÖHL1
1
Fraunhofer Institute for Manufacturing Engineering and Automation (IPA), Department of Plating
Technology, Nobelstr. 12, D‐70569 Stuttgart, Germany
johannes.oehl@ipa.fraunhofer.de
Abstract
A new sulphide electrolyte is proposed for usage in rare earth molten salt electrolysis in the
recycling of rare earths from magnet scrap. This electrolyte can be used at much lower
temperatures and with much lower ecological impact than currently used chloride electrolytes
in primary production in China. Most of the other properties of the electrolyte have yet to be
determined. Most important aspects are the solubility of involved compounds, formation of
anodic products, current efficiency and purity of rare earth metal product. The new technique
is not only an important aspect in the recycling process of rare earths, but might also be
implemented in primary production.

Introduction
The demand for rare earth metals has been increasing sharply since the year 2000. The most
volatile rare earth products are currently rare earth magnets which contain a high amount of
Neodymium besides smaller amounts of Praseodymium, Dysprosium and Terbium1. Modern
technologies like electric mobility or wind energy are highly depending on a reliable supply
with rare earth magnets2.

The difficulties with rare earth supply started with the rising demand for specific rare earth
metals, like Neodymium for magnets. This situation led to a strong advance in prices for some
rare earth metals while the less demanded metals became cheaper. Due to the composition
of rare earth ores, an increase in production of specific rare earths is not possible, therefore
the price gap became even greater³.

Another problem is the monopoly position of China in rare earth supply with over 90 % of
rare earths coming from China in the last years4. In 2010, Chinese rare earth exports were
restricted because of the demand on China’s own markets5. Rare earth prices, in particular
for the magnet metals, rose sharply in the rest of the world (Fig. 1).

Figure 1: Index of rare earth metal prices used in magnets (Neodymium, Dysprosium,
Terbium), 2003=1006, 7

After this incident, several countries dependent on rare earths started looking for alternative
sources. In the US, Canada and Australia, new mining projects were founded and abandoned
mines were reactivated8. In Europe, no relevant primary production of rare earths is in place,
mostly due to ecological requirements. Therefore, recycling is a promising and highly
discussed source of rare earths in Europe9. Since the beginning of 2013, there was a slight fall
in rare earth prices, but still the high prices and the dependence on China as main supplier
make recycling of rare earth magnets a possible successful alternative10.

One key aspect in both primary winning and recycling of rare earth metals is the reduction of
rare earth salt. For over a hundred years, molten salt electrolysis has been used for the
reduction of nonprecious metals such as rare earths11. Despite its old age, molten salt
electrolysis has not advanced much, mostly due to difficult handling at high temperatures
(>1000°C) and the resulting lack of scientific research12. It is still possible to significantly
improve molten salt electrolysis in terms of energy consumption and processability.

Figure 2: Schematic picture of Neodymium molten salt electrolysis13

State of the Art
Existing molten salt electrolysis techniques are mainly chloride and oxy‐fluoride electrolysis.
The temperature in each case is often set to a higher value than the melting temperature of
the desired rare earth metal, so the metal can be collected in a crucible below the cathode.
These temperatures are between 800°C (Cerium) and up to 1700°C (Yttrium) respectively14.
Alternatively, cathodes for formation of liquid alloys of cathode metal and rare earth or
cathodes for solid precipitation can be used. The furnace chamber has to be flooded with
inert gas, mostly argon, to prevent the reactive rare earth metals from corroding. Heating of
industrial scale electrolysis baths is carried out by direct current, so no external heating
system has to be used.

Chloride electrolysis uses rare earth chlorides in an eutectic mixture of sodium chloride and
potassium chloride or other mixtures of alkaline chlorides. The cathode, either the bath
container or a block electrode, consists of graphite, iron, or inert high‐melting metals such as
molybdenum. Chloride electrolysis is only reasonable below 1100°C, because the solubility of
the product metal in the electrolyte increases significantly and the aggressive medium attacks
the bath set‐up14. The stability of rare earth metal in the electrolyte also influences the
current efficiency which is relatively low at 45‐50 %14. The chloride electrolysis produces
Chlorine gas which is very toxic and corrosive. Despite its downsides, chloride electrolysis is
used in rare earth production in china today.

In oxy‐fluoride electrolysis a mixture of rare earth fluoride and oxide is dissolved in alkaline
metal fluoride. Due to the high corrosiveness of fluoride components and the high
temperatures, the bath may only contain graphite and refractory metal (e. g. molybdenum or
tungsten) components. The better stability of high oxidation states in fluoride media has a
positive effect on the current efficiency which is around 80 %. The anodic product is mainly
carbon dioxide that is formed by the carbon electrode and oxygen from the melt. While not
as harmful as chlorine, carbon dioxide emissions are controversially discussed because of
climate change15. An even greater problem is the evolution of small amounts of extremely
toxic and corrosive fluorine.

Latest trends in molten salt electrolysis research are highly concerned with higher energy
efficiency and less ecological impact through waste products. An improvement in energy
efficiency is reached by higher current efficiency of traditional processes16 and lower process
temperature, e. g. through new electrolyte compositions17.

An interesting technique to reduce anodic exhaust gases is the use of sulphides in molten salt
electrolysis to produce sulphur at the anode. The implementation of sulphides has been
described in basic research, but not yet for rare earths. The production of aluminium from
aluminium sulphide has been discussed18. Additionally, sulphides were used in the form of a

metal sulphide bulk material on the cathode. Tungsten and molybdenum have been won by
this process on a laboratory scale19, 20.

Characteristics of molten salts
Several important parameters of the electrolyte salt in molten salt electrolysis can be
determined. The most obvious property is the melting point of the electrolyte, either of a
single salt or a (eutectic) mixture. The melting point fixates the minimal possible temperature
for the electrolysis. Normally, a temperature of about 100 K higher is necessary to reach a
sufficient thermal movement of the melt and a sufficient diffusion of ions21.

The electrolyte has a great contribution on the anodic product of electrolysis. It is dependent
on available anions that can be oxidized on the anode. The anion can either come from the
rare earth salt in the melt or the electrolyte salt. If more than one anion is used in the melt,
the reacting one can be determined by the standard potential of the related redox pair. The
anion with the highest potential is least likely to get reduced. Unfortunately, the electrode
material and the deposition mechanism may affect the realistic potential of deposition,
causing deposition at higher or lower potentials than the standard potential. These effects
can be used advantageous for the process.

Solubility of rare earth salts and rare earth metals is also important. For the rare earth salt, a
high solubility in the electrolyte is desirable. A greater amount of rare earth ion can be
provided in the melt, resulting in a lower impact of diffusion on the electrodeposition22.

New Idea: Sulphide Electrolytes
As mentioned above, sulphides have already been used as a bulk electrode to win
Molybdenum, Tungsten and Sulphur. In this new concept, the sulphides are used as the main
electrolyte, not a sulphide salt of the target metal, mostly because of the poor availability of
rare earth sulphides and the hard soft acids bases theory (HSAB). Due to the necessary
electrochemical stability, sulphide compounds of alkaline or alkaline earth metals are
suitable. Unfortunately, those compounds have a melting point of at least 840°C (K2S)
therefore the use of lower melting polysulphides is planned (melting point K2Sx: 471°C).24

In the electrolysis process, the sulphide components will be oxidized to sulphur on a graphite
anode. For the cathode, the use of the target rare earth metal is optimal for an easy post‐
processing. Also other electrode materials will be considered, like steel, nickel‐base alloy or
ceramic23.


An important aspect in sulphide electrolysis is the collection of the won sulphur. With its
boiling temperature at 444.6°C24, the sulphur is expected to be gaseous at process conditions.
The most reasonable solution would be to guide the sulphur through a heated pipe and
condense it in a predefined separate container. Because of the difficulties in furnace
geometry, this installation cannot be realised in short‐term, so there will only be a hood at
the anode for preventing the sulphur from distributing in the furnace chamber.

Experiments will be carried out in a retort furnace with a specially designed and insulated
electrode feed‐through. The conditions in the furnace chamber are temperatures up to
1100°C and inert atmosphere with nitrogen or argon gas. At first, only neodymium as the
most important rare earth compound in magnets will be considered. Typical other rare earth
metals in magnets are Praseodymium, Dysprosium and Terbium.

Figure 3: Proposed set‐up for sulphide electrolysis of Neodymium

With the electrochemically active salt in the anodic reaction being the sulphide electrolyte, it
is expected that several rare earth salts with more stable anions than sulphide can be
processed in this kind of molten sulphide electrolysis. In rare earth processing, mainly oxides,
chlorides, and fluorides occur which are all electrochemically more stable than sulphide.
Additionally, rare earth sulphides are processible, although those salts do currently not play
a role in rare earth treatment. A submitted project funded by BMBF will be dealing with the
winning of Neodymium sulphides for recycling purposes. The plurality of usable Neodymium
would be a great advantage of the sulphide electrolysis.

Future Plans
The next steps in validation of sulphide electrolytes for molten salt electrolysis have to be
experimental testing of the postulated ideas. At first, we need to examine the basic properties

of alkali polysulfides concerning molten salt electrolysis, in particular melting temperature
and general behaviour like vapour pressure and reactivity under electrolysis conditions.

As a key aspect, the solubility of affected chemicals, rare earth oxides, chlorides, and
fluorides, will be quantified with high temperature potentiometric methods. As mentioned
before, the solubility of Rare Earth salt has to be as high as possible for a sufficient supply
with Rare Earth cations in the melt and a good current efficiency. On the other hand, the
solubility of deposited metal in the melt has to be very low. If dissolved, the metal can migrate
through the cell to the anode, get oxidized and start all over, significantly decreasing current
efficiency. Additionally, the further treatment of the metal is much more difficult if it is
dissolved.

Another very important aspect is electrochemical testing. Via cyclovoltammetry in a three
electrode set‐up, multiple reaction properties can be determined. At first, the potentials of
the electrochemical reactions can be recognized at the points of highest current. Additionally,
information about the reduction mechanism may be obtained, for example intermediate
oxidation states of the products (RE3+  RE2+  RE0) or surface reactions on the electrode13.
Another important outcome of electrochemical analysis is the theoretical yield of metal from
the integration of the Faraday current over time.

The quality of electrodeposited rare earth metal is controlled by x‐ray diffraction (XRD) and
x‐ray fluorescence analysis (RFA). In electrolysis, it is always possible to co‐deposit other
metals from the melt, even if their standard potential is significantly lower than the one of
the desired metal. Additionally, the mass of won rare earth metal is important for comparison
with theoretical yield for determination of current efficiency.

For validation of a possible industrial usage of our technique, scale‐up is essential. With the
current equipment, electrolysis with up to 10 l of electrolyte can be performed.

Conclusions
Sulphide electrolysis is considered a promising technique because of the possible low process
temperature, a lack of exhaust gas, and lower voltage compared to state of the art oxide or
chloride electrolysis. The technique can be used both in rare earth winning both from primary
and secondary sources. Similar reactions have already been verified in laboratory scale in
other fields.

A great effort in verifying the technique has still to be done, on laboratory scale as well as in
scale‐up. If the experiments lead to favourable results, sulphide electrolysis will become a
serious alternative in electro‐winning of non‐precious metals.


References
1. X. Du, T. E. Graedel, “Global Rare Earth In‐Use Stocks in NdFeB Permanent Magnets”, J. Ind. Ecol. 15 (6)
836‐843 (2011).
2. J. H. Rademaker, R. Kleijn, Y. Yang, “Recycling as aStrategy against Rare Earth Element Criticality: A Systemic
Evaluation of the Potential Yield of NdFeB Magnet Recycling”, Environ. Sci. Technol., 47 (18) 10129‐10136
(2013).
3. K. Binnemans, P. T. Jones, K. Acker, B. Blanpain, B. Mishra, D. Apelian, “Rare Earth Economics: The Balance
Problem”, JOM, 65 (7) 846‐848 (2013).
4. L. Levkowitz, “China’s Rare Earth Industry and its Role in the International Market”, U.S.‐China Economic
and Security Review Commission, Washington D. C., 2010.
5. K. Binnemans, P. T. Jones, B. Blanpain, T. van Gerven, Y. Yang, A. Waltron, M. Buchert, “Recycling of rare
earths: a critical review”, J. Clean. Prod., 51 1‐22 (2013).
6. http://www.metal‐pages.com/metalprices/rareearths/, accessed 14.05.2014
7. http://www.wirtschaftsdienst.eu/archiv/jahr/2011/3/2540/, accessed 20.05.2014
8. L. Meinert, “Mineral Resources Supply & Information with a Focus on Rare Earth Elements”, at US‐EU
Workshop on Informed policy‐making through improved mineral raw materials data, U.S. Geological Survey
(USGS), Brussels, 2012.
9. R. Fechner, S. Legler, T. Engelmann, „Germany – A source of Raw Materials“, TechProtect GmbH, Böblingen,
2012.
10. J. Liu, “Rare Earth Magnet Recycling”, at Rare Earth Elements Workshop, Colorado, 2012.
11. H. Zhu, “Rare Earth Metal Production by Molten Salt Electrolysis”, in Encyclopedia of Applied
Electrochemistry, Edited by R. F. Savinell, Springer‐Verlag, Berlin, 2013.
12. H. Numata, H. Akatsuka, H. Matsuura, “Optimization of a Molten Salt Electrolytic Bath Geometry for Rare
Earth Metal Recovery using a Finite Element Method”, Z. Naturforsch. A, 68a 48‐58 (2013).
13. Wikimedia commons, http://commons.wikimedia.org, accessed 07.05.2014.
14. N. Krishnamurthy, C. K. Gupta, “Rare Earth Metals and Alloys by Electrolytic Methods”, Min. Pro. Ext. Rev.,
22 477‐507 (2001).
15. R. Adounka, A. Bouvier, A. Fölzer, “Schmelzflusselektrolyse von Seltenerd‐Metallen”, in Proceedings of the
33rd metallurgical seminar, Society for Mining, Metallurgy, Resources and Environmental Technology
(GDMB), Clausthal‐Zellerfeld, 1997.
16. D. K. Sahoo, H. Singh, N. Krishnamurthy, “Current efficiency in electro‐winning of lanthanum and cerium
metals from molten chloride electrolytes”, Rare Met., 32 (3) 305‐311 (2013).
17. L. Viera, A. H. Whitehead, B. Gollas, “Mechanistic Studies of Zinc Electrodeposition from Deep Eutectic
Electrolytes”, J. Electrochem. Soc., 161 (1) D7‐D13 (2014).
18. N. Q. Minh, R. O. Loutfy, N.‐P. Yao, „A preliminary study of the electrolysis of Aluminium Sulfide in Molten
Salts”, Argonne National Laboratory, Springfield, 1983.
19. G. Li, D. Wang, X. Jin, G. Z. Chen, “Electrolysis of solid MoS2 in molten CaCl2 for Mo extraction without CO2
emission”, Electrochemistry Communications, 9 1951‐1957 (2007).
20. T. Wang, H. Gao, X. Jin, H. Chen, J. Peng, G. Z. Chen, “Electrolysis of solid metal sulphide to metal and
sulphur in molten NaCl‐KCl”, Electrochemistry Communications, 13 (12) 1492‐1495 (2011).
21. L. C. Dewan, C. Simon, P. A. Madden, L. W. Hobbs, M. Salanne, “Molecular dynamics simulation of the
thermodynamic and transport properties of the molten salt fast reactor fuel LiF‐ThF4”, J. Nuc. Mat., 434
322‐327 (2013).
22. Y. Castrillejo, M. R. Bermejo, E. Barrado, A. M. Martínez, P. Díaz Arocas, “Solubilization of rare earth oxides
in the eutectic LiCl‐KCl mixture at 450°C and in the equimolar CaCl2‐NaCl melt at 550°C”, J. Electroanal.
Chem., 545 141‐157 (2003)
23. E. Kvalheim, G. M. Haarberg, A. M. Martinez, H. M. Jahren, “Inert Anodes for Oxygen Evolution in Molten
Salts”, ECS Transactions, 16 (49) 367‐374 (2009).
24. W. M. Haynes, CRC handbook of chemistry and physics, 94th Edition, CRC Press, Boca Raton, 2013.

RARE EARTH EXTRACTION FROM NdFeB MAGNET AND RARE
EARTH OXIDES USING MOLTEN SALT PROCESSES
Aida ABBASALIZADEH1,*, Lidong TENG2, Seshadri SEETHARAMAN3, Jilt
SIETSMA1 and Yongxiang YANG1
1
Department of Materials Science and Engineering, Delft University of Technology, Delft, Netherlands;
2
Department of Materials Science and Engineering, Royal Institute of technology, Stockholm, Sweden;
3
Salt Extraction AB, Sweden

abbasalizadeh@tudelft.nl, lidong@kth.se, raman@kth.se, J.Sietsma@tudelft.nl,
Y.Yang@tudelft.nl
Abstract
In the present paper, the feasibility of the recovery of neodymium and dysprosium from spent
Nd‐Fe‐B magnets (~6 wt% Dy) was investigated by using molten salt processes. The salt bath
consisted of eutectic composition of NaCl‐KCl‐LiCl mixture. In order to enable an efficient
dissolution of metal in the molten salt phase, AlCl3 was used as a chlorinating agent. Iron‐free
electrodeposition could be carried out successfully. EDS analysis of the electrodeposit revealed
that co‐deposition of the dysprosium occurs along with neodymium at the cathode. The
process shows that this method is well‐suited for recovering rare earth metals from magnetic
scrap containing these metals.
Furthermore the set‐up design for recovery of neodymium and dysprosium from their oxides,
with regards to the previous studies on the neodymium magnets, was investigated. Stability
of different fluorides and chlorides salts was studied by means of thermodynamic calculations.
AlF3 based molten salt systems were studied in detail as the solvent in the rare earth oxide
electrochemical extraction of Rare Earth Elements with Al.
Introduction
International concerns have been raised on the supply shortages of rare earth elements since
China, the largest producer of rare earths, has reduced the export of these elements recently,
while, on the other hand, global demand has increased over last years.
Neodymium magnets are used in different application such as hard disk drives, voice coil
motors, magnetic resonance imaging sources because of their superior magnetic properties.
Dysprosium is often used as additive element in neodymium magnets in order to keep the
magnetic properties at high temperatures. An effective recovery method for Dy and Nd is
needed, since almost all magnets are disposed after being used1.
Hydro‐electrochemistry is not a feasible method for the recovery of rare earths since these
elements have highly negative electrode potential and also they react with water and oxygen.
Therefore, molten salt solutions are selected for electrowinning of rare earths2. In this work

we have used molten salt chloride in order to electro‐reduce the rare earth elements from
their magnet.
The other important issue which has been discussed in the recent years among the
researchers is the recovery of rare earths from rare earth oxides (REO). Different methods
were used in order to remove the oxygen from rare earth oxides: direct electrochemical
deoxidation process3, solid state electrotransport (SSE)4‐6, oxyhalide formation7 and calcium‐
halide deoxydation8 which was further combined with electrolysis in order to balance the CaO
activity in the molten salt3. Yet the strong affinity of rare earth metals to oxygen8, 9 has made
it difficult to industrialize any of these methods except molten salt electrolysis.
In this paper we will discuss the possibility of using aluminium chloride as the Nd (and Dy)
chlorination agent for NdFeB magnets, and using aluminium fluoride as a strong fluxing agent
in the molten fluoride to react with rare earth oxide and form rare earth fluoride which can
be further subjected to electrolysis under the applied voltage.
Thermodynamic considerations
NdFeB magnets containing Dy

Efficient dissolution of metal in the molten salts depends on the choice of fluxing agent.
Earlier works on the extraction of Fe,Cr and Nd from industrial electric arc furnace slag (EAF),
chromite ore10 and spent neodymium magnets11 respectively, have proven that AlCl3 can act
as a powerful chlorinating agent. The reaction between the magnets and AlCl3 leads to the
metal chloride formation. The formed metal chloride will be subjected to the electrolysis and
will be reduced on the cathode.
The standard Gibbs energy for different metal chlorides formation (metals present in the
magnets), using AlCl3 as the chlorinating agent, was calculated by using the FactSage software
(FactSage 6.3). The results show that the Nd and Dy trichlorides are more stable than AlCl3,
while the formations of FeCl3, FeCl2 and BCl3 are not favoured. The Gibbs energies of the
corresponding reactions are listed in table 1.

Table 1: Gibbs energy values of the reaction of AlCl3 with the different metals in the system.
ΔGo[kJ/mol] ΔHo[kJ/mol] ΔGo[kJ/mol] ΔHo[kJ/mol]
Chlorination Reactions
at 1073 K at 1073 K at 298 K at 298 K
AlCl3(salt)+Dy(s)=DyCl3+Al(liq.) -206.2 -353.5 -291.2 -293.9
AlCl3(salt)+Nd(s)=NdCl3+Al(liq.) -247.8 -368 -331.4 -330.2
AlCl3(salt)+Fe(s)=FeCl3 +Al(liq.) 296.7 337.7 296.4 306.6
AlCl3(salt)+1.5Fe(s)=1.5FeCl2+Al(liq.) 207 165.1 176.8 193.2
AlCl3(salt)+B(s)=BCl3 +Al(liq.) 180.2 195.3 242.4 303

From these results it can be concluded that Nd and Dy react with AlCl3 and as a result, NdCl3
and DyCl3 are formed in the salt bath. It should be noted that these calculations are based on

the pure substances at their standard state at 1073 K, whereas the activity of Nd and Dy in
the magnet as well as the activity of the chlorides would change after the dissolution in the
salt bath. The formed rare earth chloride will be decomposed according to the reaction (1):

3 (1)
The decomposition voltage of the different metal chlorides and alkali chlorides were
calculated using FactSage software and the results are presented in Figure 1.

Figure 1: Calculated decomposition voltage as a function of temperature for the chlorides
formed in LiCl–KCl–NaCl–AlCl3 molten salt bath.

An overpotential of 0.8 V was suggested for the decomposition of NdCl3 in the earlier work11.
Considering that Nd and Dy possess similar properties and also based on the decomposition
voltage of DyCl3 shown in Figure 1 (2.6 V at 800⁰C), a voltage of 3.4 V was applied for the
electrodecomposition of DyCl3.

Electrochemical reduction of rare earth oxides
For the electrochemical reduction of REOs, the first step is the electrolyte selection. The
reduction potential of the solvent should be more negative than the reduction potential of
the rare earth oxides, meaning that the solvent has to be more stable than the solute. In order
to compare the stability of different chloride and fluoride solvents, thermodynamic studies
were performed on the solvent components as well as the oxides. The Gibbs energy
formations of the rare earth oxides and the most common molten salts, calculated using
FactSage, are listed in table 2.


Table 2: Gibbs energy values of the decomposition of different salts and oxides in the
system
Reaction ΔG(kJ/mol) at 850º C Decomposition potential(V)
Nd2O3 (s)= 2Nd(s)+3/2 O2(g) 1491.7 -2.5
Dy2O3(s)= 2Dy(s)+ 3/2 O2(g) 1534.9 -2.6
NdF3 (s)=Nd(s) + 1.5 F2(g) 1392.6 -4.8
DyF3(s)=Dy(s) + 1.5 F2(g) 1414.5 -4.8
CaF2 =Ca + F2 1036.8 -5.3
LiF = Li + 0.5F2 504.6 -5.2
NaF= Na + 0.5 F2 468.3 -4.8
KF= K + 0.5 F2 459.3 -4.7
CaCl2= Ca + Cl2 625.8 -3.2
LiCl= Li + 0.5 Cl2 322.8 -3.3

From the thermodynamic results, it can be seen that between CaF2, LiF, KF and NaF at 850°
C, the most stable fluoride is CaF2. Hence the relative stability of these metal fluorides is:
CaF2>LiF>NaF>KF. Comparing the decomposition voltage of rare earth fluorides and the alkali
fluorides and chlorides, it can be seen that only CaF2 and LiF are more stable than NdF3 and
DyF3. In other words, in the case of using NaF, KF, CaCl2 and LiCl, we can expect that Na, K, Ca
and Li (respectively) will be reduced on the cathode before the reduction of rare earths.
Therefore, for the electrochemical deposition of rare earth oxides, the most suitable solvents
among fluorides are CaF2 and LiF. Considering the high melting point of calcium fluoride
(1418°C), the eutectic composition of LiF‐CaF2 (79‐21 mol%, shown in Figure 2), is a suitable
option to be used as the molten salt solvent for the electrochemical reduction of the rare
earth oxides.


Figure 2. LiF‐CaF2 phase diagram

12
The experimental results of Hamel et al. support the thermodynamic results in the present
study. These authors have measured the standard potential of different fluorides solvent in
order to find the most suitable solvent for reduction of Nd. Their results show that in LiF‐NaF
and LiF‐KF systems, no electrochemical reduction of Nd is observed.

Comparing the relative advantages of AlF3 and AlCl3 for electrochemical reduction of Nd and
Dy from Nd magnets, trials with AlCl3 indicated the loss of some of the aluminium chloride
added to the vapour phase from the molten chloride bath before getting dissolved in the
molten salt due to the high vapour pressure11. AlF3 on the other hand, was found in the
present work as a suitable reduction agent candidate for the electrolysis of rare earth oxides
in molten fluorides. From the calculated Gibbs energy of the reaction,

2 2 Δ 850°C 177.7 (2)

it is seen that aluminium fluoride can react with rare earth oxides forming rare earth fluorides,
which can be further reduced at the cathode. In‐situ formation of REF3 is very important since
the solubility of REOs is very low in the molten fluorides. Stefanidaki et al.13 have studied the
oxide solubility and Raman spectra of Nd2O3 in the alkali fluorides. They have reported that
the NdF63‐ anion is the dominant complex in the eutectic NdF3‐LiF melt system and when
Nd2O3 is added, NdOF54‐ complex might form in the melt. They have found that the solubility
of Nd2O3 varies from 0.15‐0.38 mol% when the NdF3 concentration changes from 15‐30 mol%
at 900°C. This is in support of the earlier results that MyOx oxide solubility in electrolyte is
enhanced with the MFx salt13. Another important advantage of in‐situ formation of REF3 as

the result of reaction of REO with aluminium fluoride is that the formation of rare earth
oxyfluoride might be prohibited.

Contradictory results have been reported on the electrochemical reduction of rare earth
oxyfluorides in the molten salts, which shows that further investigation is needed. According
to Taxil et al.14, Ln fluorides in the presence of the metal oxides will form rare earth
oxyfluoride which is an insoluble product. Stefanidaki et al.15 have shown that the neodymium
oxyfluoride is not reduced to neodymium metal. In the voltammetric characterization of LiF‐
NdF3‐Nd2O3 system, they have observed the same voltammogram as the one for LiF‐NdF3
system. They have concluded that neodymium is reduced on the tungsten cathode by
electroreduction of neodymium fluorides (present in the form of [NdF6]3‐ ), while oxygen is
generated on the glassy carbon anode by oxidation of neodymium oxyfluorides (present in
the form of [NdOF5]4‐), producing CO and CO2 gasses15. They believed that electrochemical
production of neodymium in an oxyfluoride melt is possible at low voltage electrolysis, in
which fluoro‐carbon compounds are not formed. Thudum et al.16 have shown that
neodymium in LiF‐CaF2‐NdF3‐Nd2O3 and LiF‐CaF2‐LaF3‐Nd2O3 systems can be reduced both
from neodymium oxyfluorides and neodymium fluoride ions, depending on the neodymium
oxyfluorides to neodymium fluoride ions (OF/F) molar ratio. At low OF/F, [NdF6]3‐ are reduced
to neodymium. Meanwhile above a critical Nd2O3 concentration, [NdOF5]4‐ are cathodically
active ions and are reduced on the cathode. Kaneto et al.17 have suggested that in the
oxyfluoride system, oxygen is generated on the anode, while fluorine can be produced at the
anode at higher cell voltage.
As it was discussed earlier, in‐situ formation of AlF3 might reduce the problem of low solubility
of REOs in molten fluorides.
Experiments for electrochemical reduction of Dy and Nd from NdFeB
magnets containing Dy

The ternary eutectic composition of LiCl‐KCl‐NaCl salt (55 mol‐% LiCl, 35 mol‐% KCl and 10
mol‐% NaCl) was dried at 473 K for at least 24 hours. High purity aluminium chloride (99%)
was added as the fluxing agent. Neodymium magnets containing approximately 6 wt%
dysprosium were crushed into small particles and were used as Dy source. According to the
results from scanning electron microscopy (SEM) equipped with an energy dispersive
spectroscopy probe (EDS) analysis, chemical composition of the magnet was indicated as Fe14
Nd1.4 Dy0.6 (boron, being a light element could not be detected in EDS analysis).
Experiments on reduction of Nd and Dy from scrap magnets in molten chlorides were
performed at temperature 1073 K for 6 hours. The flux/neodymium and flux/salt ratio were
chosen as 2 (molar fraction) and 20 (wt%) respectively. Table 3 shows the amounts of
different components used in the experiments.


Table 3. Amounts of different components used in the electrolysis experiment (wt%)
Composition AlCl3 Nd2Fe14B NaCl KCl LiCl
(wt%) 3.14 9.18 1.59 7.48 6.63

Salt mixture, flux and magnets were heated up to 1073 K in an alumina crucible in a vertical
furnace. Inert atmosphere containing argon gas was used which was dehydrated by passing
through silica gel. Graphite rods were chosen as anode and cathode in view of their additional
advantage as oxygen getters. The electrolysis was started by dipping the electrodes into the
salt bath and a constant voltage of 3.4 V was applied by a DC power supply (HP, Hewlett,
6632A) based on the decomposition voltage of the DyCl3 at 1073 K. Figure 3 shows the
schematic diagram of the set‐up. After 6 hours electrolysis, the crucible was cooled down
under the argon gas. The deposited layer on the graphite cathode was separated and washed
with distilled water in order to dissolve the salts. After removing the salts, the deposited
powder was dried and prepared for the analysis. To investigate the morphology and
composition of the deposited product, SEM/EDS analysis was carried out.

(a) (b)
Figure 3. (a) Schematic diagram of the set-up (b) Image taken from graphite cathode after
electrolysis at 1073 K for 6 hours in argon atmosphere.
Results
The microstructures of the cathode samples were analyzed by SEM and are presented in
Figure 4. The phase with bright contrast in this image, indicated by A was confirmed to be a
metallic phase consisting mostly of dysprosium and neodymium. The phase with dark
contrast indicated by B is aluminium oxide phase. The composition of the metallic phase on
cathode samples, analyzed by EDS is presented In Figure 5. The data from EDS point analysis
show presence of neodymium and dysprosium in the deposited product which is dominant

in comparison to other elements. It should be noted that oxygen detection with EDS is not
reliable.

Figure 4. SEM image of cathode deposition after electrolysis of magnet, using AlCl3 as flux, at
V=3.4 V and T=1073 K during 6 hours. A is the metallic phase (Dy‐Nd), B is the oxide phase .
cps/eV
14
12
10
Cl Al
Dy O Dy Cl Nd Dy
6
Nd
0
2 4 6 8 10
keV

Figure 5. EDS pattern of Dy‐Nd deposit on graphite cathode in LiCl‐KCl‐NaCl molten salt at
1073 K.
The formation of Dy‐Nd metallic phase on the cathode was also investigated by EDS mapping
analysis of the sample, shown in Figure 6. The intensity of the colour in the image related to
neodymium is very low, however it can be seen that dysprosium and neodymium are
distributed in the same areas, confirming the formation of metallic phase, since these regions
are poor in oxygen. It can be also seen that aluminium and oxygen are distributed in the same
areas which shows the formation of aluminium oxide phase. Dysprosium deposition along
with neodymium was confirmed using molten salt electro‐deposition method.


Figure 6. Mapping images of Dy‐Nd‐Al‐O deposit on graphite electrode in LiCl‐KCl‐NaCl
molten salt at 1073 K
It was not possible to detect boron in the EDS analysis due to it being a light element. In the
earlier studies on the electrochemical reduction of neodymium from Nd magnet scraps18, the
results from wavelength dispersive spectroscopy (WDS) showed that boron remains in the
bulk salt bath, and the intensity of the boron peak in the cathode sample was lower than the
detection limit, shown in figure 7.

Figure 7. Intensity scan over the boron peak position in the salt bath sample and cathode
sample from the earlier study 18.

Discussion
The eutectic composition of LiCl‐NaCl‐KCl ternary mixture was used as electrolyte for the
electrochemical decomposition of NdFeB magnets containing dysprosium. Aluminium
chloride was used as the fluxing agent. It is proven in different studies19,20 that aluminium
chlorides exist in the form of AlCl4‐ ionic species in the alkali chlorides solvents, which results
in the formation of a pseudo–binary solution.

The results from SEM/EDS show the presence of Nd‐Dy metallic phase in the deposited
material. This is explained by the proximity of their electrode potentials. It can be concluded
that neodymium and dysprosium have been dissolved in the alkali chloride melt, forming
RECl3, In fact, the results from Raman spectrometry have confirmed that neodymium(III)
exists as NdCl63‐ complex with octahedral symmetry in the molten alkali chlorides21.Thus, in
the chloride melt, the rare earth elements dissolve according to22:

6 (3)

Considering the cathodic and anodic electrochemical reactions, these can be represented as:
Anodic reaction 3 3 (4)
Cathodic reactions 3
3 (5)

RE3+ is the most stable state of the rare earth metals in the molten salts23. It has been reported
that most of the rare earth elements (La, Ce, Pr, Y) have a single decomposition signal in the
molten chlorides, however Nd reduction occurs in two steps22. The formation of divalent rare
earth metal ions in the chlorides melt is most likely to be one of the reasons for decreasing
the current efficiency, which would be caused by two‐step reduction of the rare earth metals:

(6)
2 (7)

Contradictory results have been reported on the electrochemical mechanism of neodymium
and dysprosium reduction in molten salts. It has been reported that the reduction process of
NdCl3 to Nd metal in pure chloride24, LiF–CaCl2 melts25, LiF15, LiF–CaF226 and LiCl‐KCl27 is a one‐
step mechanism. However, De Córdoba et al.28 and Masset et al.29 have confirmed that the
reduction of NdCl3 takes place in two steps.
Electrode material is one of the factors that can influence the electrochemical reduction
behaviour of the rare earths in molten salts. Castrillejo et al.30 have observed that their cyclic
voltammogram results exhibit different behaviours of dysprosium on the W and Al

electrodes. They have observed that, on the W electrode, Dy is reduced in two steps, viz.
Dy(III) → Dy(II) → Dy(0), while on the Al electrode the electrochemical reaction would be15 :

3 3 (8)
In order to get a clarification of the electrochemical mechanism of rare earth reduction in the
molten salts, further studies are required.
The study of the electrochemical reduction of Nd and Dy from neodymium magnets
containing dysprosium is based on a new process line for the electrolytic recovery of rare
earths31, which can be shown as the flow chart in Figure 8.

Figure 8. Schematic diagram of the salt extraction process
Conclusion

Feasibility of the neodymium and dysprosium extraction from Nd2Fe14B magnets containing
6% dysprosium was investigated using the molten salt method. In the present approach it
was shown that neodymium and dysprosium recovery from magnetic scrap enables a direct
separation of these metals from iron, eliminating the oxide or halide conversion steps.
Simplicity of this method due to the single step recovery of the RE metals from the magnet
scrap, makes this process attractive from the industrial point of view. Compared to
hydrometallurgical approaches, this process has advantage of being more environmental‐
friendly since the salt bath can be reutilized without contaminating the environment.

Further, based on thermodynamic calculations, it was shown that strong tendency for AlF3
towards reaction with Nd2O3 and Dy2O3 would enhance the in‐situ formation of rare earth
fluorides in the salt bath and thereby increase the solubility of REOs in the molten fluorides.
Hence AlF3 can act as a powerful flux agent in the fluorides melt for recovery of rare earth

metals from their oxides. The use of AlF3 will be experimentally explored in the forthcoming
studies.
Acknowledgements
Part of this work that is related to rare earth extraction from scrap magnet was carried out in
Royal Institute of Technology (KTH) in Sweden and was supported by Swedish Foundation for
Strategic Environmental Research (MISTRA) through Swedish Steel Producers Association
(Jernkontoret). We are also grateful to the EU‐FP7 project of “EREAN” (Project ID 607411) for
the financial support on the reduction of rare earth from rare earth oxides.

References

1. A. Kurachi, M. Matsumiya, K. Tsunashima and S. Kodama, “Electrochemical behaviour and
electrodeposition of dysprosium in ionic liquids based on phosphonium cations”, J Appl Electrochem, 42,
961–968 (2012).
2. J. Lodermeyer, M. Multerer, M. Zistler, S. Jordan, H.J. Gores, W. Kipferl, E. Diaconu, M. Sperl, and G.
Bayreuther, “Electroplating of Dysprosium, Electrochemical Investigations, and Study of Magnetic
Properties”, Journal of The Electrochemical Society, 153 (4) C242‐C248 (2006).
3. K. Hirota, T.H. Okabe, F. Saito, Y. Waseda and K.T. Jacob, “Electrochemical deoxidation of RE–O (RE=Gd,
Tb, Dy, Er) solid solutions”, Journal of Alloys and Compounds, 282, 101–108 (1999)
4. D. Fort, “The ultra purification of rare earth metals: gadolinium and neodymium”, Journal of the Less‐
Common Metals, 134, 27 – 44 (1987)
5. D. Fort, V.K. Pecharsky, K.A. Gschneidner, “Solid state electrotransport purification of dysprosium“, Journal
of Alloys and Compounds, 226, 190‐196 (1995).
6. R. G. Jordan and D. W. Jones, “The purification of the rare earth metals: ii. Solid state electrotransport
processing of terbium” R. G. Jordan and D. W. Jones. Journal of the Less‐Common Metals, 42, 101 – 110
(1975).
7. J. D. Corbett, J. D. Smith and E. Garcia, “The selective removal of oxygen from rare earth and thorium metals
through oxyhalide formation”, Journal of the Less‐Common Metals, 115, 343 – 355 (1986).
8. T.H. Okabe, K. Hirota, E. Kasai, F. Saito, Y. Waseda and K.T. Jacob, “Thermodynamic properties of oxygen
in RE–O (RE=Gd, Tb, Dy, Er) solid solutions”, Journal of Alloys and Compounds 279 184–191 (1998).
9. O. N. Carlson, R. R. Lichtenberg and J. C. Warner, “Solid solubilities of oxygen, carbon and nitrogen in
yttrium”, Journal of the Less‐Common Metals, 35, 275‐284 (1974).
10. X. Ge, O. Grinder, and S. Seetharaman, “The salt extraction process: a novel route for metal extraction Part
I – Cr, Fe recovery from EAF slags and low grade chromite ores”, Mineral Processing and Extractive
Metallurgy, 119(1) 27‐32 (2010)
11. A. Abbasalizadeh, S. Seetharaman, L. Teng, S. Sridhar, O. Grinder, Y. Izumi, and M.Barati, “Highlights of the
salt extraction process”, The Journal of The Minerals, Metals & Materials Society, 65(11) 1552‐58 (2013)
12. C. Hamel, P. Chamelot, P. Taxil, “Neodymium(III) cathodic processes in molten fluorides”, Electrochimica
Acta, 49, 4467–4476 (2004).
13. E. Stefanidaki, G. M. Photiadis, Ch. G. Kontoyannis, A. F. Vik and T. Østvold, “Oxide solubility and Raman
spectra of NdF3–LiF–KF–MgF2–Nd2O3 melts”, J. Chem. Soc., Dalton Trans., 2302–2307(2002)
14. P. Taxil, L. Massot , C. Nourry, M. Gibilaro, P. Chamelot, L. Cassayre, “Lanthanides extraction processes in
molten fluoride media: Application to nuclear spent fuel reprocessing”, Journal of Fluorine Chemistry, 130,
94–101 (2009).

15. E. Stefanidaki, C. Hasiotis and C. Kontoyannis, “Electrodeposition of neodymium from LiF–NdF3–Nd2O3
melts”, Electrochimica Acta, 46, 2665–2670 (2001).
16. R. Thudum, A. Srivastava, S. Nandi, A. Nagaraj and R. Shekhar, “Molten salt electrolysis of neodymium:
electrolyte selection and deposition mechanism”, short communication.
17. A. Kaneko, Y. Yamamoto and C. Okada, “Electrochemistry of rare earth fluoride molten salts”, Journal of
Alloys and Compounds, 193, 44‐46 (1993).
18. A. Abbasalizadeh, , L. Teng, s.. Sridhar, S. Seetharaman, “Neodymium extraction using salt extraction
process”, submitted to journal of Mineral Processing and Extractive Metallurgy.
19. H. Lichum, and R. A. Osteryoung, 1981. “Chemical and electrochemical studies in room temperature
aluminium halide containing melts in Ionic Liquids”, Plenum Press, Edited by D. Inman and D. G. Lovering,
New York (1981).
20. G. Charlot, and B. Tremillon, “Les reactions chimiques dons les solvants et les sels fondus”,Paris: Gauthier–
Villars (1963).
21. Y. A. Barbanel, V. V., Kolin, V.P. Kotlin, and A. A. Lumpov, “Coordination chemistry of actinides in molten
salts”, J. Radioanal. Nucl. Chem, 143, 167‐179 (1990).
22. Y. Castrillejo, M.R. Bermejo, E. Barrado, A.M. Martı´nez and P. Dı´az Arocas, “Solubilization of rare earth
oxides in the eutectic LiCl‐KCl mixture at 450° C and in the equimolar CaCl2‐NaCl melt at 550°C”, Journal of
Electroanalytical Chemistry, 545, 141‐157(2003).
23. H. Zhu, “Rare Earth Metal Production by Molten Salt Electrolysis”, SpringerReference. May 2013.
24. l. Wu, H. Zhu, Y. Sato, T. Yamamura, and K. Sugimoto, In Proceedings of the Ninth International Symposium
on Molten Salts, Edited by C.L. Hussey, D.S. Newman, G. Mamantov and Y. Ito, The Electrochem. Soc Proc.
Series: NJ, Pennington (1994)
25. C. Hamel, P. Chamelot, and P. Taxil, “Neodymium(III) cathodic processes in molten fluorides”,
Electrochimica Acta, 49, 4467–4476 (2004).
26. C. Nourry, L. Massot, P. Chamelot and P. Taxil, P. “Electrochemical reduction of Gd(III) and Nd(III) on
reactive cathode material in molten fluoride media”, J. Appl. Electrochem, 39, 927–933 (2009).
27. J.Serp, M. Allibert, A. L. Terrier, R. Malmbeck, M. Ougier, J. Rebizant, and J. P. Glatz, 2005.
“Electroseparation of Actinides from Lanthanides on Solid Aluminum Electrode in LiCl‐KCl Eutectic Melts”,
J. Electrochem. Soc., 152(3) C167–C172 (2005).
28. G. De Córdoba, A. Laplace, O. Conocar, J. Lacquement, and C. Caravaca, “Determination of the activity
coefficient of neodymium in liquid aluminium by potentiometric methods”, Electrochemical Acta, 54, 280–
288 (2008).
29. P. Masset, R. J. M. Konings, R. Malmbeck, J. Serp, and J. P. Glatz, “Thermochemical properties of
lanthanides (Ln = La, Nd) and actinides (An = U, Np, Pu, Am) in the molten LiCl–KCl eutectic”, Journal of
Nuclear Materials, 344, 173–179 (2005).
30. Y. Castrillejo, M.R. Bermejo, A.I. Barrado, R. Pardo, E. Barrado and A.M. Mart´ınez, “Electrochemical
behaviour of dysprosium in the eutectic LiCl–KCl at W and Al electrodes”, Electrochimica Acta, 50, 2047–
2057 (2005).
31. S. Seetharaman, and O. Grinder, US patent application nr. 12/991128, ref. no.: 12057 (2010)


Session V‐A: REE Processing‐ Mineral
Beneficiation / Leaching


RECOVERY OF RARE EARTH ELEMENTS FROM CLAY
MINERALS

Vladimiros G. PAPANGELAKIS1 and Georgiana MOLDOVEANU2
1,2 Department of Chemical Engineering and Applied Chemistry, University of Toronto
200 College Street, Toronto, Canada, M5S 3E5

vladimiros.papangelakis@utoronto.ca, georgiana.moldoveanu@utoronto.ca

Abstract
Continuous development of advanced technologies has created increasing demand for rare
earths elements (REE), with global emphasis on identifying new alternate sources to ensure
adequate supply. Clay deposits containing physically adsorbed lanthanides are
substantially lower grade than other types of REE mineral resources; however, their
abundance in sub‐tropical regions, existence as surface layers requiring low mining cost,
and their high cation exchange capacity make them economically‐important sources of rare
earths. At the University of Toronto we have conducted a systematic study of clay minerals
from various locations. It was found that REE are easily recovered via an ion‐exchange
mechanism during leaching with monovalent salt solutions under ambient conditions,
based on a 3:1 stoichiometric ratio between the trivalent lanthanides and the exchange
monovalent cation. The present paper reports on the most important findings of this
investigation.

I. Background
Rare earths, also known as lanthanides, are a series of 17 elements with unique
properties that make them essential for the hi‐tech industry because of their various uses
in high strength permanent magnets, lasers, automotive catalytic converters, fiber
optics/superconductors, and electronic devices. They are grouped depending on the
atomic number, in “light” rare earth elements (REE)—La, Ce, Pr, Nd, and “middle & heavy”
REE—Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb, Lu, Y. Because of the ongoing development of
new advanced technologies, there is an ever‐increasing demand for lanthanides in the
international markets, with emphasis on identifying new resources to ensure adequate
supply for present and future use. World production of REE is dominated by China, the
United States and Australia; however, in terms of reserves and resources, China
dominates the world potential with reserves estimated to be around 50% of the total
while completely leading and controlling the global production at ~ 90%.1 For the last 3
decades, R&D in the field of REE in most of the Western world has slowed down due to
the import of these elements from China. Consequently, the development of specialized
extraction, refining and processing technologies, including equipment and training of

engineering expertise were allowed to lapse, creating thus a dependence on Chinese
supplies. Starting with 2005 China – the undisputed leader in both REE innovation and
trade, has been restricting yearly export quotas for REE ores in order to have enough
resources for its own industries and to gain control over the global market, especially for
the heavy and scarcer rare earths.2 Consequently, the last decade has brought a renewed
concerted global drive towards REE research and development, led by USA and Japan,
with the dual scope of finding new resources and improving processing/extraction
technologies, as summarized by Adachi, Imanaka and Tamura.3

REE are part of many various rock‐forming minerals, but the most commercially
significant sources, as reviewed by Kanazawa and Kamitani,4 fall into the following
categories:
1) Bastnasite, (REE, Ce)(CO3)F, is a magma‐derived fluorocarbonate mineral containing
65‐75 %wt. light rare earth oxides (REO) and accounts for more than 80% of global REO
production. The two major sources in the world for lanthanides are bastnasite deposits
at Mountain Pass, California (U.S.A) – devoted solely to REE production, and Bayun‐Obo,
Inner Mongolia (China) – mined primarily for iron ore and REE as by‐product.
2) Monazite, (REE)PO4 is a light REE phosphate containing 55‐65 %wt. REO, associated
with granites and beach sands in Australia, Brazil and India. Until about 1965 monazite
was the main REE source; since then, the use of monazite has been considerably reduced
due to radioactivity caused by thorium and radium.
3) Xenotime (Y,REE)PO4 is an yttrium‐rich phosphate containing 25‐60 %wt. Y2O3 and
other heavy REE. It is mainly recovered as by‐product of mining for titanium, zirconium
and tin in Malaysia, Indonesia and Thailand.
4) Weathered Crust Elution‐deposited Rare Earth Ores (common name: Ion‐adsorption
Clays) are aluminosilicate minerals (e.g. kaolinite, illite, and smectite) containing 0.05‐
0.3% %wt. REEs are physically adsorbed at sites of permanent negative charge. The ion‐
adsorption clay deposits are the result of in‐situ lateritic weathering of rare‐earth rich
host rocks (granitic or igneous), which lead over geological times to the formation of
aluminosilicate clays. These very fine mineral particles have the capability of adsorbing
lanthanide ions released/dissolved during weathering. Sub‐tropical climates present
ideal conditions for this lateritic process to occur. The best example of this formation
process exists in Southern China (latitudes 24‐26°N), where many of such deposits are
known to exist, as described by Bao and Zhao.5 In spite of their low grade, ion‐adsorption
clays account for ~ 35% of the China’s total REE production, according to Yang et al.6

Carbonate and phosphate sources, in spite of being high grade, are associated with
elevated recovery costs due to difficulty in mining, separation, beneficiation and need of
aggressive conditions to dissolve the REE. For example, bastnasite is generally leached
with concentrated H2SO4 or HCl, whereas monazite/xenotime concentrates need to be
baked either in 98% H2SO4 or 70% NaOH to render REE soluble. Secondary REE sources

include apatite and the so‐called “hard‐rock” deposits such as zircon, titanate, niobate,
allanite, eudialyte, gadolinite, but these contain less than 30% REE and require even
harsher conditions for breaking down the mineral matrix (e.g. caustic bake followed by
acid leaching); the processing of these ores is mainly directed towards extraction of
niobium, tantalum and zirconium.7

Although ion‐adsorption clay deposits are substantially lower grade than other types of
lanthanide sources, the lower grade is largely offset by the easier mining and processing,
costs, and the very low content of radioactive elements (normally associated with
yttrium). These deposits are mined by open‐pit methods and no ore beneficiation is
required. A simple leach using monovalent sulphate or chloride salt solutions at ambient
temperature can produce a high‐grade REO product.8,9,10 Because of their abundance in
surface layers in nature, ease of mining and processing, these clays warrant a detailed
study as important sources of rare earths.

II. Formation of Weathered Crust Elution‐deposited Rare

Earth Ores (Ion‐adsorption Clays)
The ion adsorption clay deposits of REE were first discovered in 1969 in the Jinangxi
province (southern China) and declared a novel type of exogenous rare earth ore. Since
then, more deposits were discovered and mined all throughout the South of China, as
indicated by Figure 1.

Figure 1: Locations of REE‐bearing ion‐adsorption clay ores in China5

The formation of this type of ore is due to physical, chemical and biological (microbial)
weathering of REE‐rich granitic and volcanic rocks under warm, humid, slightly acidic

conditions in subtropical zones. According to Bao and Zhao,5 the weathering crusts are
up to 30 m deep and divided into four layers: (A) An upper humic layer of quartz, organic
matter and soil: 0–2 m thick, with very low/nil REE content; (B) a strongly weathered
layer enriched in REE: 5–10 m thick with kaolinite, halloysite, quartz and mica; (C) a
semi‐weathered layer: 3–5m thick with kaolinite and sericite; (D) a weakly weathered
bottom layer with the same mineral compositions as the host rock. 80‐90% of the
adsorbed REE report to the strongly weathered layer (B) whereas less than 15% are
found in the semi‐weathered later (C). Depending obviously on the nature of the original
host rocks, the general components of the weathered ores are kaolinite, halloysite and
mica, with a typical composition (as %wt.) of ~70% SiO2, 15% Al2O3, 3‐5% K2O, 2‐3%
Fe2O3 and less than 0.5% of CaO, MgO and other elements.8 The metallogenic mechanism
of formation for the weathered REE deposits, as proposed by Peng11 and later expanded
by Hendrick12 is presented in Figure 2.

Considering the geological and climate conditions for the formation of REE‐bearing
weathered ores, there is no reason to limit the occurrence of this type of deposits within
Chinese borders. While at the present time China is the only country to actively pursue
and develop this type of resource to commercially produce REE, recent geological
surveys (summarized by Chen13) have led to the discovery and investigation of similar
ion adsorption clay deposits in South America14 and Africa,15 located in the same sub‐
tropical weathering areas as depicted in Figure 3.

More clay
minerals
Completely
weathered
layer Migration and
REE ions
ore‐forming Ion‐adsorption
Upper‐crust deposits
Granites
REE ions
Half‐
weathered
layer
Decomposition Less clay
Weatherin
and alteration minerals

Figure 2: Metallogenic mechanism of formation for REE‐
rich adsorption‐type ores (adapted from Peng11 and Hendrick12)


Figure 3: Area (shaded) favourable for ion adsorption clays formation

III. Nature of Rare Earth Elements in Ion‐adsorption Clays

Clay minerals are part of the phyllosilicate class, containing layered structures of shared
octahedral aluminum and tetrahedral silicon sheets; water molecules and hydrated
cations can move in and out of the interlayer spaces. Very often, isomorphous
substitution of one cation with another (of similar size but with lesser charge, e.g. Al3+ for
Si4+ or Mg2+ for Al3+) within crystal structures leads to a charge imbalance in silicate clays,
which accounts for the permanent negative charge on clay particles, thus the ability of
clays to attract cations to the surface. Amphoteric ‐OH groups at the surface/edge of clays
(i.e., silanol and aluminol) may also contribute to surface charge (pH‐dependent
reversible charge).16

According to Bradbury and Baeyens17 as well as Piasecki and Sverjensky,18 for acidic and
near‐neutral conditions (pH < 6.5‐6.8), most of the surface‐adsorbed lanthanides occur
as simple or hydrated cations such as “clay‐REE” or “clay‐REE(H2O)n” species derived
from straightforward cation‐exchange reversible reactions at the permanent negative
charge sites on the clays (physisorption); for pH > 7 the prevalent form is the hydrolyzed
“clay‐O‐REE2+” species derived from permanent complexation reactions at the
amphoteric surface hydroxyl groups (chemisorption) or soluble carbonate/bicarbonate
complexes.8


Due to various weathering conditions (i.e. nature of host rocks, water and soil pH,
temperature, pressure, redox conditions) there are three main categories of REE present
in the ion‐adsorption clays, as described by Chi and Tian8:
1. Colloid phase: REE deposited as insoluble oxides or hydroxides or as part of colloidal polymeric
organometallic compounds. These species have low occurrence in ores at the slightly acidic natural
conditions and can be recovered only by acid leach.
2. Exchangeable phase: REE occur as soluble free cations/hydrated cations or part of positively‐
charged complexes in solution adsorbed species on clays. These species account for 60‐90% of
total content of rare earths in ores and can be recovered by ion‐exchange leaching with
monovalent salts.
3. Mineral phase: REE part of solid fine particles with same mineral matrix as the host rocks (REE
part of the crystal lattice). This phase usually accounts for the balance from the ion‐exchangeable
phase towards the TREE content and can be recovered only by aggressive conditions (alkaline
bake and acid leach).

The vast majority of the ion‐adsorption ores present the “negative cerium anomaly”,
meaning that, contrary to the majority of lanthanide elements which are usually
physically adsorbed as trivalent ions, Ce3+ can be easily oxidized by atmospheric oxygen
(O2) to Ce4+,19 and precipitates as cerianite, CeO2. Consequently, the formation of the
mineral cerianite facilitates a natural separation of Ce from the other adsorbed trivalent
lanthanides, as described by Bao and Zhao,5 and makes it impossible to be recovered by
ion‐exchange leaching.

Depending on the nature of the original host rocks, other metals will get dissolved and
carried downstream during the weathering, decomposition and alteration processes.
The main impurities associated with the ion‐adsorption ores are usually Al, Mg, Ca, Mn,
Zn and Fe. While most base metals occur as part of the mixed mineral phase and do not
leach out during the mild ion‐exchange REE leaching conditions, Al especially and to a
lesser extent Ca and Mg constitute the major impurities physically adsorbed on clays that
are liable to get desorbed during the process along with the lanthanides.8,14

IV. Overview of Leaching Technologies for the Ion‐Adsorption

Clays
As previously described, the ion‐adsorption clays contain anywhere between 0.05 to 0.3
wt.% rare earths, out of which 60‐90% occur as physically adsorbed species recoverable
by simple ion‐exchange leaching. In the typical procedure, the ores are leached with
concentrated inorganic salt solutions of monovalent cations. During leaching, the
physisorbed REE are relatively easily and selectively desorbed and substituted on the
substrate by the monovalent ions and transfer into solution as soluble sulphates or

chlorides, following a 3:1 stoichiometry (Eqn. 1). Solubilized REE are usually selectively
precipitated with oxalic acid to form oxalates (Eqn. 2) that are subsequently converted
to REO via roasting at 900 °C according to Eqn. 3. Finally, the mixed REO are separated
into individual REE by dissolution in HCl and fractional solvent extraction.

2 Clay‐REE + 3M2SO4  2 Clay‐M3 + REE2(SO4)3 (1)
REE2(SO4)3 + 3 H2C2O4 + 10 H2O  REE2(C2O4)3·10H2O + H2SO4 (2)
REE2(C2O4)3·10H2O  REE2O3 + 3CO + 3CO2 + 10H2O (3)

Various research conducted on the desorption of REE from clays via ion‐exchange
leaching8,9,10 indicated that, regardless of the initial content, not all REE reached similar
extraction levels. Coppin et al.20 reported that the amount of trivalent lanthanide ions
adsorbed on smectite and kaolinite was inversely proportional to the ionic radii and
pointed to a fractionation during selective sorption of lanthanides, with heavy elements
(i.e., higher atomic number: Tb to Lu) being adsorbed stronger that the light ones (i.e, La
to Gd). They related this behavior to the “lanthanide contraction” in the ionic radii going
from light to heavy REE. Based on these observations, it was inferred that desorption
must exhibit a similar trend, with heavy REE being more difficult to extract, probably
according the trend

Sc > La > Ce > Pr > Nd > Sm > Eu > Gd > Tb > Dy > Ho > Y > Er > Tm > Yb > Lu

Ever since the discovery of the weathered crust elution‐deposited rare earth ores in the
Jingxi province about 45 years ago, China was at the forefront of research and
development of this kind of unique deposits, applying the ion‐exchange leaching
procedure for the extraction of lanthanides via three successive generations of
technology, as summarized by Chi and al.21 and presented below.

The first generation leaching technology –batch leaching with NaCl

In the early 1970s, the ores were processed by opencast mining, sieved and leached with
~1M NaCl in barrels, followed by oxalic acid precipitation. The main disadvantages of this
initial approach were small scale, low yields, high lixiviant concentration needed and
poor product quality (<70% TREO content due to Na oxalate coprecipitation), which
greatly surpassed the advantages of extremely low costs and fast processing times. By
mid‐1970s the procedure was changed to bath leaching in concrete pools in order to
increase production; however, the main disadvantage of low product purity remained
and, because of the largely unregulated and illegal/clandestine mining and extraction
practices, the environmental impact was devastating, including severe loss of vegetation
and biodiversity, soil erosion and water contamination (both streams and phreatic).


The second‐generation leaching technology – batch and heap leaching with
(NH4)2SO4
In the early 1980s, 1M NaCl was replaced by ~ 0.3M (NH4)2SO4 as lixiviant for batch
leaching, which required less reagent consumption due to increased desorption
capabilities of NH4+ as compared to Na+ and led to improved final product purity (> 92%
TREO content). The procedure was so successful that it became the primary leaching
method for the next approximately 20 years and largely contributed to the intense
development of REE ion‐adsorption research. However, the second‐generation batch
leaching technology led to environmental impact as well, due to mining‐related
deforestation and discharge of tailings and was ultimately replaced in early 1990s by the
heap leaching procedure.

In the traditional heap leaching procedure, the soil pile (1.5‐5 m high) is built on a flat
impermeable (leak‐proof) layer 5‐20 cm thick inside a cofferdam ~ 50 cm high as to
prevent solution overflow. The lixiviant is injected into the top of the pile at a solid to
liquid (S:L) ratio of ~0.25:1 and accumulates at the bottom in the collecting ditch.
Washing is performed with clean water at a S:L ratio of ~0.6:1; depending on the size of
the ore heap, leaching time ranges from 100 to 320 hours and REE extraction can reach
up to 90%. This procedure is very well suited for the processing of very low‐grade ores.


The third‐generation leaching technology – in‐situ leaching with
(NH4)2SO4
The intense and largely unregulated use of successful batch and heap leaching with
ammonium sulphate for ~ 2 decades led to severe and long‐lasting environmental,
ecological and health damages in southern China; as an example based on Yang’s
paper,6 by 2010 the REE mining in Guangdong region alone has left ~302
abandoned mines, 191 million tons tailings and 153 km2 of destroyed forests.

In June 2011, in an effort to regulate the industry and deal with the environmental
effects, the Chinese government enforced a ban on surface mining and batch/heap
leaching while implementing mandatory in‐situ leaching technology for the
processing of the ion‐adsorption clays, as being more advantageous in terms of
surface vegetation clearing and soil disturbance. The basic principle of in‐situ
leaching (also called “solution mining”) is injection of leaching solution directly
into the natural orebody and retrieval of the pregnant solutions for further
processing. Leaching holes with a depth of 1.5‐3 m and diameter of ~0.8 m are
drilled 2‐3 m apart, for up to 100 m, the lixiviant (~0.3 M (NH4)2SO4) is injected at
high pressure, flows through the pores of the orebody and the loaded leach
solution is pumped above‐ground through the recovery wells; the whole process
(including injection of water for washing) can take up to 400 days and ultimately
85‐90% of REE are extracted.
The in‐situ leaching technique is also currently applied in China for the recovery
of residual REE from the tailings of older batch and heap leaching operations.

The implementation of in‐situ leaching requires comprehensive geological
surveys ‐ specific to each site ‐ in order to determine the hydrogeological structure
of the area, ore characteristics, grade, orientation and the surrounding rock
infiltration properties. The procedure can only be applied to an orebody with
suitable permeability and placed over solid bedrock without fissures. Failure to
conduct diligent geological surveys may result in serious environmental
degradation such as underground water contamination, mine collapse, landslides
and severe loss of REE recovery.

V. Evaluation of Leaching Potential of Various Ion‐

adsorption Clays at University of Toronto
As new ion‐adsorption REE deposits are being explored and discovered in the rest
of the world, research on REE extraction from clays has expanded outside of China
as well. For the last five years, the University of Toronto has conducted systematic
in‐depth studies on the leaching chemistry and optimum conditions for REE
extraction from clay samples obtained from various geographical locations.
Overall, it was determined that, under atmospheric conditions, the leaching power
of monovalent ions for REE extraction depends on the hydration energy of the
exchange ion, following the order Cs+>NH4+>Na+>Li+, in both sulphate and
chloride.9 Batch leaching studies also revealed that the ion exchange process
achieved equilibrium in as little as 5 minutes, regardless of the experimental
conditions; ambient temperatures and moderately acidic pH values (4‐5)
represent optimum conditions for maximum REE recovery.10

Based on those previous studies, a benchmark leaching procedure was established
and research has been conducted at the University of Toronto in order to compare
leaching characteristics among different clay samples obtained from Africa, Asia
and South America and evaluate lanthanide extraction. The final aim is to develop
a fully contained optimized process for field implementation that minimizes the
impact to environment by providing options for efficient reagent use, maximized
extraction and recycle/regeneration of the lixiviant.22

Experimental
Solid Characterization
The REE contents of the clay samples were determined by acid digestion at 220°C,
using the Ethos EZ microwave digestion system, followed by inductively coupled
plasma optical emission spectrometry of the solution (Agilent 720 ICP‐OES).
Batch Leaching Tests

The leach solutions were prepared using ACS reagent grade ammonium sulphate
and deionized water. The benchmarked procedure for leaching is: 0.5M (NH4)2SO4
(natural pH ~ 5.2), ambient conditions, liquid to solid (L:S) ratio of 2:1 (vol./mass),
30 minutes total time. The slurry was agitated via magnetic stirring then the
mother liquor was separated by vacuum filtration. The filter cake was washed by
deionized water, and the wash water was collected separately for analysis. The
resultant solutions were diluted with 5% (vol.) nitric acid and analysed by ICP‐
OES to calculate the REE extractions.

Table 1 shows the REE content (as %wt.) of the various clay samples tested while
Table 2 shows the relative distribution of individual REE in the clays.

Table 1: REE content of ion‐adsorption clays from different geographical origins

South
Origin Africa Asia
America
REE
A1 A2 A3 A4 A5 B1 B2 C1
(wt%)
Y 0.029 0.014 0.018 0.012 0.010 0.157 0.047 0.120
La 0.175 0.029 0.179 0.046 0.025 0.007 0.098 0.045
Ce 0.026 0.017 0.022 0.045 0.028 0.006 0.020 0.012
Pr 0.028 0.007 0.027 0.012 0.004 0.003 0.019 0.010
Nd 0.100 0.023 0.088 0.026 0.016 0.012 0.069 0.029
Sm 0.017 0.004 0.017 0.006 0.004 0.004 0.018 0.006
1.76E‐ 4.76E‐ 6.36E‐
Eu 0.001 0.001 0.001 0.001 0.002
04 04 05
Gd 0.011 0.004 0.009 0.003 0.003 0.015 0.013 0.010
1.06E‐
Tb 0.002 0.001 0.001 0.001 0.004 0.002 0.006
04
Dy 0.006 0.002 0.002 0.002 0.001 0.026 0.010 0.022
4.60E‐ 3.33E‐ 1.43E‐
Ho 0.001 0.001 0.005 0.002 0.007
04 04 04
Er 0.002 0.001 0.015 0.012 0.001 0.038 0.025 0.021
2.00E‐ 8.69E‐ 6.85E‐ 2.32E‐
Tm 0.001 0.004 0.002 0.005
04 05 05 06
Yb 0.002 0.001 0.001 0.001 0.000 0.016 0.003 0.026
2.66E‐ 2.66E‐ 2.53E‐ 1.71E‐ 2.40E‐ 4.95E‐
Lu 0.002 0.005
04 04 04 04 04 04
TREE 0.399 0.108 0.380 0.165 0.095 0.300 0.330 0.326

Despite total rare earth element (TREE) content being consistent with grades
expected for the ion‐adsorption clays, no specific pattern of preferential REE
accumulation and distribution was observed, except that all ores seem to be rich
in La and Y and Nd; although some similarities in terms of relative composition are
observed within clays originating from the same geographical areas (e.g. A1
through A5), there is no consistent trend. This is probably due to wide variations
in composition of original host rocks and incongruent dissolution‐adsorption
processes during the weathering processes.


Table 2: Relative distribution of REE within clay samples

Origin Africa Asia South
America
REE A1 A2 A3 A4 A5 B1 B2 C1
(%REE)
Y 7.2 12.9 4.6 7.0 10.7 52.5 14.4 36.9
La 43.8 26.8 47.0 27.6 26.7 2.2 29.7 14.0
Ce 6.4 16.1 5.7 27.4 30.0 2.1 6.0 3.6
Pr 7.0 6.1 7.2 7.4 4.2 0.9 5.8 3.0
Nd 24.9 21.6 23.2 15.5 16.7 3.9 20.8 8.9
Sm 4.4 4.1 4.4 3.4 3.7 1.3 5.6 1.8
Eu 0.2 0.8 0.2 0.1 0.5 0.0 0.3 0.7
Gd 2.8 3.9 2.4 2.1 2.9 5.2 4.0 3.0
Tb 0.4 0.6 0.3 0.3 0.1 1.3 0.6 1.7
Dy 1.4 2.1 0.6 1.4 1.5 8.7 3.0 6.9
Ho 0.2 0.7 0.1 0.2 0.2 1.8 0.6 2.0
Er 0.6 1.2 4.0 7.0 1.4 12.8 7.6 6.4
Tm 0.1 0.1 0.0 0.0 0.6 1.2 0.5 1.5
Yb 0.4 1.0 0.2 0.5 0.4 5.3 1.0 8.0
Lu 0.1 0.2 0.1 0.1 0.3 0.7 0.1 1.6
TREE 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0

The clay samples listed in Table 1 were leached using the benchmarked procedure
described above to investigate the terminal REE extraction levels (shown in Table
3) and TREE leaching kinetics, respectively (presented in Figure 4).

From data in Table 3 it can be observed that all the minerals investigated are the
ion‐adsorption type, i.e., the lanthanides are physically adsorbed and can be easily
recovered via a simple ion‐exchange leaching procedure, as described by
Moldoveanu and Papangelakis.9,10 The extraction levels vary between 40 to 80%,
consistent with the predicted exchangeable REE percentage, as described by Chi
and Tian8. Again, there is no consistent trend regarding REE desorption within
clays of similar geographical origin; this could be explained in terms of:

(1) The ores contain various mixtures of clay materials with different cation‐adsorption
capacities, i.e., clays with lower cation exchange capacity (CEC) such as kaolinite will
adsorb less ions than the ones with larger CEC such as smectites, hence less is available for
desorption (possible observed for clays A4 and A5).
(2) The overall exchangeable fraction of REE is low, regardless of the high REE content, the
balance being the matrix‐bound lanthanides (possible observed for clays C1).

As a general common trend for all clays, cerium does not exhibit large extraction
due to its presence mostly as CeO2 while some of the heavy REE also show poor
extraction probably because of the stronger adsorption, as described by Coppin et
al.20

Table 3. Final REE extraction from ion‐adsorption clays from different
geographical origins

Origin Africa Asia South
America
%E A1 A2 A3 A4 A5 B1 B2 C1
Y 77.3 72.6 77.5 67.1 57.2 80.5 79.9 70.8
La 83.6 76.9 73.3 66.4 61.2 85.0 80.5 82.9
Ce 0.0 16.9 12.7 6.8 11.6 0.0 0.0 35.6
Pr 75.1 65.0 84.7 74.1 35.1 62.9 69.4 72.8
Nd 80.8 74.1 83.7 68.0 46.0 86.3 87.9 83.2
Sm 90.6 85.2 86.8 66.3 96.4 43.1 81.2 76.9
Eu 62.1 79.0 62.0 43.8 84.3 61.3 83.0 36.2
Gd 82.6 63.0 79.8 74.8 62.3 90.1 84.9 61.6
Tb 84.1 62.8 23.7 19.2 37.3 91.5 90.1 35.6
Dy 80.9 76.3 90.4 74.2 52.2 87.5 79.7 61.5
Ho 75.5 0.0 97.7 96.9 97.5 82.9 77.2 51.8
Er 86.8 73.5 80.9 63.0 44.7 75.2 91.5 53.0
Tm 53.4 79.4 11.2 23.1 7.0 57.9 26.8 47.0
Yb 73.2 64.0 9.4 16.0 25.4 77.9 73.2 61.9
Lu 52.3 34.9 18.9 31.7 12.2 78.5 67.5 51.5
TREE 76.6 64.0 72.8 58.9 42.6 80.3 82.1 68.7

In terms of extraction kinetics, all materials investigated showed a common trend
of fast REE desorption which is the typical behaviour of the ion‐adsorption
minerals. While each clay sample possesses a slightly different REE content and
REE extraction end point, it can be seen that extractions reach the terminal levels
in a very similar fashion, typically reaching plateau in less than 15 minutes of batch
leaching time. The authors demonstrated10 that desorption kinetics are very fast
independently of leaching conditions such as temperature, pH, and agitation,
which influence only terminal extraction levels.

Figure 4: Kinetic response of ion‐adsorption clays from various geographical
origins to benchmark leaching conditions (0.5M (NH4)2SO4, 60 min leaching
under ambient conditions)

Conclusions
Past and ongoing research conducted at University of Toronto demonstrated the
existence and commercial potential of ion‐adsorption clay deposits outside China
and established a unified benchmark procedure for REE leaching from these types
of minerals. It was found that, in spite of variations in clay origin and REE content,
all REE consistently reached peak extraction levels under ambient conditions with
fast kinetics. However, the terminal extractions were generally clay and element‐
specific.
Based on those studies, an optimized process ready for field implementation is
currently being developed, that highly minimizes the impact to environment by
providing options for efficient reagent use, maximized extraction and
recycle/regeneration of the lixiviant.

Acknowledgements
The authors would like to thank Molycorp Inc. and Tantalus Rare Earth A.G for
providing clay samples.

References
1. United States Geological Survey 2011 (USGS), “Mineral Commodity Summaries: Rare Earths”,
http://minerals.usgs.gov/minerals/pubs/commodity/rare_earths/mcs‐2011‐raree.pdf,
accessed 05/21/14.
2. C. Hurst, “China’s Rare Earth Elements Industry: What Can the West Learn?” Institute for the
Analysis of Global Security, 2010.
http://fmso.leavenworth.army.mil/documents/rareearth.pdf, accessed 05/21/14.
3. G. Adachi, N. Imanaka, S. Tamura, “Research Trends in Rare Earths: A Preliminary Analysis”,
Journal of Rare Earths, 28(6) 843‐846 2010.
4. Y. Kanazawa, M. Kamitani, “Rare Earth Minerals and Resources in the World”, Journal of Alloys
and Compounds, 408 1339‐1343 2006.
5. Z. Bao, Z. Zhao, “Geochemistry of Mineralization with Exchangeable REY in the Weathering
Crusts of Granitic Rocks in South China”, Ore Geology Reviews, 33 519‐535 (2008).
6. J. J. Yang, A. Lin, X. L. Li, Y. Wu, W. Zhou and Z. Chen, “China’s Ion‐adsorption Rare Earth
Resources, Mining Consequences and Preservation”, Environmental Development, 8 131‐136
(2013).
7. C. K. Gupta, N. Krishnamurti, Extractive Metallurgy of Rare Earths, CRC Press, New York, 2005.
8. R. Chi, J. Tian, Weathered Crust Elution‐deposited Rare Earth Ores, Nova Science Publishers, NY,
2008.
9. G. Moldoveanu, V. Papangelakis, “Recovery of Rare Earth Elements Adsorbed on Clay Minerals:
I. Desorption Mechanism”, Hydrometallurgy, 117‐118 71‐78 2012.
10. G. Moldoveanu, V. Papangelakis, “Recovery of Rare Earth Elements Adsorbed on Clay Minerals:
II. Leaching with Ammonium Sulphate”, Hydrometallurgy, 131‐132 158‐166 2013.
11. S. Peng, “Geological Characteristics and the Prospecting Criteria of the Granite‐weathering
Crust Ion‐adsorption Type REE Deposits in Nanling Area, South China”, in Materials Science
Forum 33‐42, Trans Tech Publications, Switzerland, 1991.
12. J. B. Hendrick, “The Global Rare Earth Cycle”, Journal of Alloys and Compounds, 225 609‐618
1995.
13. Z. Chen, “Global Rare Earth Resources and Scenarios for Future Rare Earth Industry”, Journal
of Rare Earths, 29(1) 1‐6 2011.
14. A. Rocha, D. Schissel, A. Sprecher, P. de Tarso and J. Goode, “Process Development for the Serra
Verde Weathered Crust Elution‐deposited Rare Earth Deposit in Brazil”, in Rare Earth Elements
– Proceedings of the 52nd Conference of Metallurgists (COM 2013), Edited by I. London, J. Goode,
G. Moldoveanu and M. Rayat. Metallurgical Society of the Canadian Institute of Mining,
Metallurgy and Petroleum (MetSoc‐CIM), Montreal, Canada, 2013.
15. TRE Project, Tantalus Rare Earth AG, http://www.tre‐ag.com/operations/tre‐project,
accessed 05/21/14.
16. A. Meunier, Clays, Springer, Berlin, 2005.
17. M. H. Bradbury, B. Baeyens, “Sorption of Eu on Na‐ and Ca‐montmorillonites: Experimental
Investigations and Modeling with Cation Exchange and Surface Complexation”, Geochimica and
Cosmochimica Acta, 66 3964‐3979 2002.
18. W. Piasecky, D. A. Sverjensky, “Speciation of Adsorbed Yttrium and Rare Earth Elements on
Oxide Surfaces”, Geochimica and Cosmochimica Acta, 72 2325‐2334 2008.
19. A. J. Bard, R. Parsons, R. Jordan, Standard Potentials in Aqueous Solutions, Marcel Dekker, New
York, 1985.
20. F. Coppin, G. Berger, A. Bauer, S. Castet and M. Loubet, “Sorption of Lanthanides on Smectite
and Kaolinite”, Chemical Geology, 182 57‐68 2002.
21. R. Chi, J. Tian, X. Luo, Z. Xu and Z. He, “Basic Research on the Weathered Crust Elution‐deposited
Rare Earth Ores”, in Rare Earth Elements – Proceedings of the 52nd Conference of Metallurgists
(COM 2013), Edited by I. London, J. Goode, G. Moldoveanu and M. Rayat. Metallurgical Society
of the Canadian Institute of Mining, Metallurgy and Petroleum (MetSoc‐CIM), Montreal,
Canada, 2013.
22. F. Cheuk, G. Moldoveanu, S. Abendschein, E. Krause and V. Papangelakis, “Extraction of Rare
Earth elements from Various Ion‐adsorption Clays”, paper accepted at the 53rd Conference of
Metallurgists (COM 2014) to be held in Vancouver, Canada, Sept.28‐Oct 1, 2014.

ADVANCING KEY TECHNOLOGIES FOR REE MINERAL
PROCESSING
Michael HASCHKE, Horst MÄRTEN, Harald KALKA, Jana NICOLAI
Umwelt‐ und Ingenieurtechnik GmbH Dresden (UIT), Germany
m.haschke@uit‐gmbh.de
Abstract
UIT Dresden has scrutinised several recent investigations on optimisation potential for
alternative extraction and processing of technologically critical, high‐demand REEs
from some primary and secondary REE deposits: (1) Ion‐adsorption clays as current
global main source of heavy REE (e.g. Dysprosium, Terbium), (2) Monazite ore as main
source of light‐middle REE (e.g. Neodymium) but also NORM (e.g. Thorium, Uranium,
Radium), and (3) Mine tailings as alternative source of REE (mainly Cerium). Extraction
and processing of REE from ion‐adsorption clays stands out as being easily accessible
with relatively low capitalisation and operational costs, low NORM content. Extraction
from ion‐adsorption clays may have high remediation costs depending on extraction
method. Monazite processing is considered high‐risk due to locally high Thorium
content with challenges for sustainable disposal. REE extraction from tailings through
heap‐leaching shows potential for low‐cost recovery, but many tailings do not contain
the high‐value heavy REE required to make such REE recovery economical.
Introduction
Recent commodity studies indicate heavy Rare Earth Elements (REE) as most critical
mineral resource throughout the next decades as these resources are among the main
sources for innovative ‘green’ technologies for energy generation and e‐mobility.
However, current R&D projects place little effort on research and development of the
key issues related to (heavy) REE mining and extraction which include (1) Development
of more economic, alternative and environmentally, socially and economically
sustainable metallurgical REE extraction technologies, and (2) Removal (and re‐
utilization) of naturally occurring radioactive material (NORM) from the REE mineral
processing stream. The exploration and mineral processing industry is facing these
challenges regardless of previous, current or future market prices of technology
metals, unless substitute materials enter the market on an industry scale.

In response to the expected shortage of technology metals in high‐tech industry
countries, numerous (non‐China) REE deposits are currently being developed
throughout the world (status May 2014: 57 reporting code‐compliant REE projects;
TechMetalResearch, May 2014), buts many projects seem to be lingering at, or do not
pass the scoping or prefeasibility stage as their capital and operating costs are high.

Western world REE projects generally possess higher capitalisation costs for
metallurgical plant setup and operational costs for hydrometallurgical REE extraction,
concentration, and separation of rare earth oxides. Numerous feasibility studies have
shown that large scale commercial operations are required; such large scale rare earth
production is currently lacking in Europe. The quest for economic (lower cost)
extraction, processing and separation of REE is, and will remain, among the most
important challenges for any REE project regardless of REE market price fluctuations.

A second challenge is economic recovery. Many REE deposits throughout the world
show appreciable REE content, but very few deposits contain the desired high value
and technologically critical heavy REEs content (e.g. Dysprosium, Terbium, Europium),
which are required to make REE mining and processing economical. Without the high‐
value REE content, metallurgical REE processing facilities are required to process
unrealistically large amounts of low‐moderate value REE ore for economic REE
recovery. In addition, the metallurgical processes required to isolate individual REE
metal species are complex and every REE deposit requires a different processing chain,
or at least some modification and optimisation within the metallurgical processing
chain for optimal recovery.

A third challenge is NORM removal from REE ore. All minerals with economically
relevant REE content (bastnaesite, monazite, steenstrupine, xenotime, ± eudialyte, ±
allanite) contain naturally occurring radioactive material (NORM); some monazite ore
contains up to 12 wt.% of Thorium (e.g. Steenkampskraal, South Africa). Processing of
ore with economic REE quantities tends to accumulate substantial NORM content and
standard metallurgical laboratories in Europe are generally not equipped or licenced
for large‐stream NORM material separation. Outside of Europe, high‐volume REE
processing facilities for NORM include the Lynas Advanced Materials Plant (LAMP) in
Gebeng industrial estate in Kuantan, Malaysia, for processing of monazite ore from
Mt. Veld from Western Australia, and in India there is a state run operation which
treats Monazite concentrates which are high in NORMs.

A fourth challenge, but also an opportunity, is REE extraction from mining tailings.
Mining tailings are pre‐processed, typically silt‐size materials which can be
enriched in REE depending on the former main resource; for instance former
uranium mining tailings tend to show REE enrichments. This study discusses a
project in Queensland, Australia, where former uranium mine tailings show some
economic light REE‐enrichment where one main challenge lies in optimized heap‐
leaching design with sufficient light REE recovery.

UIT Dresden has developed several conceptual approaches for potential low‐
moderate cost potential projects with optimised extraction schemes from primary and
secondary REE deposits: (1) Ion‐adsorption clays as current main source of heavy REE
(e.g. Dysprosium, Terbium), (2) Monazite ore as main light‐middle REE (e.g.
Neodymium) yet also NORM (e.g. Thorium, Uranium, Radium) bearing ore, and (3)
Mine tailings as alternative source of light REE (e.g. Cerium). By discussing type
examples for REE deposits, we present preliminary results from some prefeasibility
and feasibility studies, and outline selected research opportunities in REE extraction
and NORM separation technology. The results find direct application in our own
affiliated mining operations and consulting projects worldwide. The ion‐adsorption
clay project is in direct response to official requests by Govt. for development of
sustainable leaching technologies.
REE Recovery from Ion‐Adsorption Clays
Ion‐adsorption clays are currently the world’s main source of heavy REE which contain
some of the most critical technology metals as shown by the United States Geological
Survey (USGS), the British Geological Survey (BGS), and the European Commission. For
instance, the main current supply of Dysprosium (Dy) are ion‐adsorption clays in the
Xunwu and Longnan province, southern China, followed by xenotime minerals at
Dubbo Zirconia (Australia) and Kvanefjeld (Greenland). Based on current forecasts, the
Dy supply cannot meet short‐ and/or long‐term demand (until 2050); not even when
including Dy from recycling products.

REE enriched ion‐adsorption clays are known from the equatorial belt around the
globe (tropical high‐humidity climate with intense chemical weathering), notably Mt.
Weld Australia, Araxa Brazil, Kangangunde Malawi and the Tantalus project
Madagascar, but currently all economically important REE‐rich ion‐adsorption clays
deposits are situated in southern China where ion‐adsorption clay mining for more
than two decades led to the China monopole in heavy REE production. REE‐rich ion‐
adsorption clays typically consist of saprolite and kaolinitic clays (Xunwu and Longnan)
containing microcrystals of bastnaesite, synchisite and allanite adsorbing the RE3+ ions
with typical concentrations of 0.05‐0.2% REO, out of which 2% tend to consist of
Dysprosium. If weathering occurred above REE‐enriched alkaline granitoids they may
show enrichments in heavy REE as the soluble light REE are being mobilised during
weathering processes.

At Longnan, China, large parts of the heavy REE production from ion‐adsorption clays
are produced by in‐situ leaching and surface/mountain‐top mining by tank/heap
leaching with ammonium sulfate and Na‐chloride. REE in‐situ leaching produces large
environmental issues as the leaching solutions are difficult to control. With increased
environmental controls and operating standards the operating costs are expected to
increase significantly.

The leaching chemistry in in‐situ recovery (ISR) operations revolves around acid
dissolution of matrix minerals and/or complexation of the REE into solution (especially
in ion‐adsorption clays). The latter can be achieved with various organic molecules that
can be generated through bacteria in or ex situ. Common complexing reagents used in
the context of REE processing are citric and lactic acid. Inorganic ligands have also been
explored, such as NaCl or (NH4)2SO4, with good results. Table 1 shows a comparison of
REE concentrations in ion‐adsorption clay profiles and bastnaesite, xenotime and
monazite.

Table 1: Rare Earths content of major source minerals in China (in % total REO; sources:
Roskill 2011, USGS 2011).

Rare Ion‐Adsorption Clays
Bastnaesite Xenotime Monazite
Earth
Bayan Obo Guangdong Xunwu Longnan Guangdong
Oxide
Light
La2O3 23.0 1.2 42.0 1.8 23.0
CeO2 50.0 3.0 2.3 0.4 42.7
Pr6O11 6.2 0.6 8.8 0.7 4.1
Nd2O3 18.5 3.5 30.8 3.0 17.0
Medium
Sm2O3 0.8 2.2 3.8 2.8 3.0
Eu2O3 0.2 0.2 0.5 0.1 0.1
Gd2O3 0.7 5.0 2.9 6.9 2.0
Heavy
Tb4O7 0.1 1.2 Trace 1.3 0.7
Dy2O3 0.1 9.1 Trace 6.7 0.8
Y2O3 Trace 59.3 8.0 65.0 2.4
Total 99.6 85.3 99.1 88.7 95.8

Technical issues with ion‐adsorption clay mining by in‐situ leaching include the
conditioning of the lateritic soil profile regarding permeability. At Longnan, REE
concentration in sample location tends to be highest at shallow depths of 3‐7m.
Standard ISR operation require saturated host rock conditions plus a critical hydraulic
head above the ISR pump (typically >45m hydraulic head), which rules out standard
ISR mining technology for recovery of shallow REE mineralization from ion‐adsorption
clays but requires a modified approach. UIT Dresden is currently conducting a
research initiative (supported by the German Ministry of Economics and Research
BMBF) on extraction of ion‐adsorption clays from Madagascar (Tantalus project).
The goal of this project is to develop an optimized extraction technology for heavy
REE recovery from ion‐adsorption clays through in‐situ/heap leaching technology.
The focus of this R&D project includes:

(1) the geotechnical improvement of permeability of ion‐adsorption clays in
laterite and saprolite,
(2) the development of alternative bio‐hydrometallurgical REE extraction
technologies for selective and sustainable REE extraction,
(3) process simulation of extraction and separation of adsorptive‐bonded REE
as base for optimisation of REE extraction by using thermodynamic and
experimental data.

One goal of this project between UIT Dresden, the Helmholtz Centers of Resource
Technology and Resource Ecology, GUB Engineering AG, IPK Gatersleben and
GMBU Society for Support of Medical‐, Bio‐ and Environmental Technologies, is a
re‐assessment of and optimisation of existing hydrometallurgical extraction and
separation methods with development of new biotechnological extraction
methods regarding minimised processing costs and environmental footprint.

REE Recovery from Monazite Ore
Monazite, a REE‐bearing phosphate, is a typical ingredient of placer deposits and
known for its light‐medium REE content (mainly Lanthanum, Cerium and Neodymium),
but can also contain heavy REEs (particularly Yttrium; Ni et al., 1995). Monazite is a
major source of Neodymium which ranks among the high‐demand and critical
technology elements for magnets used in wind generators; Neodymium is among the
key ingredients for wind‐power generators as state‐subsidised form of ‘green’
alternative energy in Germany and other states in Europe.

However, monazite is also an important carrier of NORM. Typical monazite generally
contains 4.5–9.5% Thorium as ThO2, and 0.2–0.4% Uranium as U3O8, depending on the
region of mineral origin (Table 2). When decay products of either are present in the
minerals, the radioactivity levels may be significant as monazite is concentrated in the
processing stream. Typical Indian plant processings are in the order of 3,000‐4,000
tons/year of monazite.

Table 2: Typical radionuclide content of some REE mineral ores in % of
composition (Source: IAEA NORM V Proceedings 2008).
Total Rare Earth
U3O8 ThO2
Oxides
Monazite, Australia 61.33 0.34 6.55
Monazite, India 59.68 0.37 9.58
Monazite, Malaysia 59.65 0.24 5.90
Monazite, Thailand 60.20 0.44 5.76
Monazite, Rep. of Korea 60.20 0.45 5.76
Monazite, DPRK 42.65 0.18 4.57

Table 3 shows a comparison of radioactivity resulting from processing of different ore
minerals, with specific reference to monazite concentrate.

Table 3: Radioactivity in mineral sands (Source: IAEA 2003 Tech Report 419).
Thorium Uranium

ppm Bq/g ppm Bq/g
Ore 5 ‐ 70 0.04 – 0.60 3 ‐ 10 0.07 – 0.25
Heavy mineral concentrate 80 ‐ 800 0.60 – 6.6 <10 ‐ 70 <0.25 – 1.70
Ilmenite 50 ‐ 500 0.40 – 4.1 <10 ‐ 30 <0.25 – 0.75
Rutile <50 ‐ 350 <0.40 – 2.9 <10 ‐ 20 <0.25 – 0.50
Zircon 150 ‐ 300 1.2 – 2.5 150 ‐ 300 3.7 – 7.4
Monazite concentrate 10,000 – 55,000 80 – 450 500 – 2,500 12 – 60
Processing tailings
200 – 6,000 1.5 – 50 10 – 1,000 0.25 – 25
(incl. monazite)
Bastnaesite 58 ‐ 74 ‐‐‐ 0.11 – 0.20 0.64 – 0.94
Xenotime, Malaysia 54.00 0.81 0.83 26.20
Gadolinite 32 ‐ 46 ‐‐‐ <2 ‐‐‐

When processing monazite, the Thorium hydroxide cake produced during the
extraction of REEs from monazite contains the U fraction as well as the unreacted
monazite. The quantity generated amounts to 250kg per 1ton of monazite processed.

Solvent extraction with tributyl phosphate (TBP) from an aqueous 8 N nitric acid
solution of Thorium and mixed REE permits the recovery of Thorium, Uranium, Cerium,
and Cerium‐free REE from monazite. Other commercially significant processes involve
precipitation of Thorium pyrophosphate or basic salts from the leach liquor and
subsequent recovery of the REE in solution as double sulfates, fluorides, or hydroxides
or even selective solubilisation of Thorium in the ore treatment stage itself.

The sulfuric acid process (acid baking)is used at the Lynas Advanced Materials Plant in
Malaysia. Here concentrated sulphuric acid is contacted with monazite concentrated
at 650 deg C. In another industrial process (caustic cracking), fine ground monazite is
generally attacked with a 60‐70% sodium hydroxide solution at 140‐150°C. Previously,
grinding and caustic digestion were performed as one step which enabled ~50%
savings in caustic soda consumption. This yielded a clean separation by digesting the
ore with caustic soda at 170°C under a pressure of several atmospheres. The mixed
RE‐Thorium hydroxide cake is processed for REE and Thorium recovery by a variety of
methods.

The radioactive equilibrium of the Th‐232 chain is broken at the acid extraction stage.
Thereafter the buildup and decay of activity at different stages follow separate routes.
The activity inventory at various stages is marked by disequilibria and the activity at
each stage has to be accounted for by taking the material balance, fractionation and
buildup/decay. Hazards from Thorium may result from both external and internal
sources. External hazards are due to high energy beta and gamma rays, while internal
hazards are mainly due to alpha emitting radionuclides deposited inside the body.

In Europe, mineral processing facilities are required to account for accumulation of
Thorium during extraction and processing, but NORM processing facilities are lacking.
However, UIT Dresden is currently in the process of expanding its laboratory facilities
to include hydrometallurgical processing of NORM material from mineral concentrate
consistent with IAEA regulations. IAEA provides regularly updated exemption levels. A
material is exempted without further consideration if the effective dose expected to
be incurred by a member of the public is of the order of 10 µSv/yr, or less. Where bulk
quantities of radionuclides are involved, IAEA makes provision for exemption using a
dose criterion commensurate with natural background levels, that is about 1 mSv/yr.
This criterion is satisfied if the activity concentration does not exceed 1 Bq/g for
uranium and thorium series radionuclides, and 10 Bq/g for 40K (IAEA Safety Guide Level
No. RS‐G‐1.7, 2004; Chambers et al., 2010).

The activity concentration criteria of 1 Bq/g for Uranium and Thorium series
radionuclides and 10 Bq/g for 40K are generally accepted as the most practicable way
forward in determining which materials should be considered for regulation in terms
of the requirements for practices. A dose criterion of the order of 1 mSv per year is
now commonplace as a de facto standard for exemption of NORM from the regulatory
requirements for practices. For a plant processing of 3,000‐4,000 t/year of monazite,
the individual annual exposures for the past 25 years were in the range of 3‐9 mSv.
The representative exposure for REE production are 0.64 man mSv (standard deviation
0.19 man mSv) per ton of REE concentrate produced by chemical processing of
monazite.
REE Recovery from Mining Tailings
UIT Dresden has developed initial heap leaching concepts for REE recovery and seeks
to pursue a feasibility study for leaching REE from former Uranium mining tailings.
Mary Kathleen is a former Uranium mine (excavated between the 1950’s and 1980’s)
with relatively high REE content, and the only commercial example of a REE‐U skarn
deposit. While Uranium (present as uraninite in fine disseminations enclosed in
allanite at about 35% of the ore at outcrop, usually replacing garnet) was mined and
processed, the REE were not recovered and consequently significant REE
concentrations are present (documented) in the unprocessed waste rocks and tailings
(milling residues). The West Tip waste rock samples at Mary Kathleen (La up to 7,310
ppm, Ce up to 11,600 ppm, U up to 916 ppm) show two times the concentration of Ce,
La and U compared to Stockpile and Crusher pile samples; this could be a priority site
for initial reprocessing. The sampled mineralogy comprises mainly garnet,
clinopyroxene, allanite, chlorite, calcite, pyrite, and locally apatite. An overview of REE
content in Mary Kathleen mining tailings is shown in Table 4.

Up until now there were no plans to recover the 4.4% total Rare Earths oxides (TREO)
in the existing tailings dam. About 9.2 million tons of ore were milled with ore grades
of 0.10‐0.15% U3O8 (source: ASX Announcement Universal Resources 2007) so that it
is assumed that there are approximately 7 million tons of ore tailings at the site in a
purpose‐built dam structure with about 3% TREO mainly in allanite and stillwellite as
economically relevant REE minerals, making it one of the largest light REE deposits
(mainly La, Ce) in Australia. Although tailings reprocessing could be economically
attractive for mining projects in Europe, the lack of NORM processing facilities
prevents this project from being economically feasible.

.

Table 4: Calculation of REE content (in %REO and value) in Mary Kathleen tailings with price comparison (as of April 22, 2013) relative to other major REE deposits. From a value basis,
the Mary Kathleen tailings appear to be at the least-value end of the REO market due to their paucity of HREE where the current value (and scarcity) lies. A Mary Kathleen tailings play
would be a mostly Cerium/Lanthanum commodity play which would only be interesting if there would be a very inexpensive method of extracting REO from the tailings so that one would
not have to compete purely on price.

22‐Apr‐13 Mary Kathleen Mt Weld Mountain Pass Nolans Bore Kvanefjeld Baiyunebo Thor Lake Basal Zone
Price (USD) % Value % Value % Value % % Value % Value % Value
La oxide La2O3 9.6 33.5 3.22 25.5% 2.45 33.0% 3.17 19.1% 1.84 27.5% 2.64 27.1% 2.60 0.3% 0.03
Ce oxide CeO2 9.6 51.5 4.94 46.7% 4.48 49.0% 4.70 48.7% 4.68 42.0% 4.03 49.9% 4.79 0.4% 0.04
Nd oxide Nd2O3 72 9.1 6.55 18.5% 13.32 13.0% 9.36 20.6% 14.82 12.9% 9.29 15.4% 11.09 15.8% 11.38
Pr oxide 80 4 3.20 5.3% 4.26 4.0% 3.20 5.9% 4.74 4.2% 3.36 5.2% 4.12 1.7% 1.36
Sm oxide Sm2O3 25 0.006 0.00 2.3% 0.57 5.0% 1.25 2.3% 0.58 1.6% 0.40 1.2% 0.29 10.4% 2.60
Dy oxide Dy2O3 630 0.05 0.32 0.1% 0.78 0.00 0.3% 2.02 1.1% 6.93 0.3% 1.89 9.8% 61.74
Eu oxide Eu2O3 1250 0.05 0.63 0.4% 5.54 0.1% 1.25 0.4% 4.88 0.1% 1.25 0.2% 2.38 1.6% 20.00
Tb oxide Tb4O7 1100 0.02 0.22 0.0% 0.08 0.1% 0.88 0.2% 2.20 0.00 1.8% 19.80
Y oxide 1.4% 0.00 7.7% 0.00 0.00 29.0% 0.00
Total 19.07 31.47 22.93 34.42 30.10 27.15 116.95


References
32. ASX Announcement Universal Resources 2007
33. Chambers, D.B., Monabbati, M., Garva, A., Feasby, G., 2010, Human exposure to radioactivity from
mining and industrial residues, Session 2, NORM VI Proceedings, IAEA, Vienna.
34. Hawkins, B.W., 1975, Mary Kathleen Uranium Deposit, in: Knight, C.L. (Ed.) Economic Geology of
Australia and Papua New Guinea, 1. Metals. Australasian Institute of Mining and Metallurgy
(AUSIMM), Melbourne, p. 398‐406.
35. IAEA Techreport No. 419, 2003, Extent of Environmental Contamination by Naturally Occurring
Radioactive Material (NORM) and Technological Options for Mitigation.
36. IAEA, 2004, The long term stabilization of uranium mill tailings, Final report of a co‐ordinated
research project, IAEA‐TECDOC 1403.
37. IAEA, NORM V, 2008, Proceedings of an International Symposium, Seville, Spain, STI/PUB/1326.
38. Ni, Y. and Hughes, J.M., 1995, Crystal chemistry of the monazite and xenotime structures,
American Mineralogist, vol. 80, p. 21‐26.
39. Roskill Information Services Ltd., 2011, Rare Earths & Yttrium: Market Outlook to 2015, 14.
40. U.S. Dept. of Energy – Critical Materials Strategy 2011, 189 p.
41. www.techmetalsresearch.com; accessed May 2014


Unlocking Rare Earth Elements from European apatite‐
iron ores
Bertil I. Pålsson1*, Olof Martinsson2, Christina Wanhainen2 and Andreas
Fredriksson3
1
MiMeR – Minerals and Metallurgical Research Laboratory, Luleå University of Technology,
Sweden
2
Geosciences and Environmental Engineering, Luleå University of Technology, Sweden
3
LKAB Mining Division, Kiruna, Sweden
bertil.palsson@ltu.se, olof.martinsson@ltu.se, christina.wanhainen@ltu.se,
andreas.frediksson@lkab.com
Abstract
Rare Earth Elements are known to occur within apatite‐magnetite ores of the Kiruna
type. Previously it was assumed that the REE was associated to the apatite part of the
ore. It is now shown that the REE follows the apatite to a much less degree, and that
the REE are contained in the minerals monazite, allanite and some unknown REE‐phase
still to be identified. Monazite occurs as inclusions in the apatite as well as free
particles. Allanite is to some degree in mixed grains with magnetite but also in free
particles. Monazite mainly reports to the apatite concentrate while allanite largely
goes to the tailings. Laboratory flotation also shows some preferential concentration
of heavy REE over light REE to the apatite concentrate. A new EU project, REEcover,
will characterise the REE phases in this type of iron ore and among the different process
streams within the mineral beneficiation chain.
Introduction
Deposits of apatite iron ores are the major source for iron in the European Union and
are exploited by LKAB in several mines in the Kiruna area in northern Sweden, which
is an important mining province dominated by Fe and Cu deposits. The combined
production from two large underground mines and one open pit was for 2013 25.5 M
tonnes of pellets and fines and a total ore production of 2275 M tonnes from 11 mines
during the last 127 years. These magnetite deposits with/without haematite have
mineralogical and textural features that are different to most other types of iron
deposits in the world. Most characteristic is the presence of apatite as an important
gangue mineral, which contributes to high phosphorus content of the ores. Many
apatite iron ores contain 1 to 5 % P on average and apatite has been produced as a by‐
product at Malmberget during the First and Second World War but also in 1952‐1953.
In the Kiruna area, apatite has been produced from the Rektorn deposit in 1942 to
1946, and 1985‐1988 about 0.4 Mt apatite concentrates were produced by flotation

from Kiirunavaara ore1. Elevated contents of rare earth elements (REE) in apatite from
the apatite iron ores in Northern Sweden was early noticed and investigated in some
detail in the 1960’s2. The samples then consisted of gangue minerals from different
ores, flotation concentrates of apatite and handpicked apatite grains. These results
demonstrated a strong enrichment of the LREE in apatite with total REE+Y contents of
1545 to 8855 ppm and that Monazite occurred as inclusions in Apatite2. In a recent
study3, the chemistry and occurrence of REE‐minerals in the Kiirunavaara deposit have
been further investigated by optical microscopy, scanning electron microscope (SEM)
and electron micro probe analysis (EMPA). The samples included apatite concentrates
obtained from tailings and ore from drill cores. This study confirmed previous results
but showed that allanite and monazite may be common as larger grains in the ore and
act as important carrier of the light rare earth elements.

TEXTURE AND CHEMISTRY OF REE‐BEARING MINERALS
Three main REE‐bearing minerals have been identified by optical microscopy, SEM and
EMP analysis and analysed for their chemical composition by electron micro probe3.
Rarely grains of thorite with moderate enrichment of heavy rare earth elements were
found.
Apatite
Apatite Ca5(PO4)3(OH,F,Cl) occurs mainly as disseminated grains in the ore or form
band, schliren or veinlets. Disseminated apatite occurs interstitial to magnetite as
subhedral and equidimensional or prismatic grains up to 0.5 mm large but mostly in
the size of 0.05‐0.1 mm. In banded type of D‐ore, apatite often form prismatic grains
with the long axis oriented parallel with the banding. Apatite forming patches‐
stringers‐schliren may by slightly coarser in grain size (up to 1mm) and occur together
with varying amounts of biotite, magnetite and calcite. Locally there occur skeletal ore,
which is coarse grained with apatite and/or carbonate occurring interstitial to tabular
magnetite crystals.

The apatite is a fluorapatite with 3.75 % F and 0.06 % Cl on average. There is no
significant variation in composition for apatite in the tailings and samples of ore. Total
content of REE is not possible to obtain by EMPA, due to values below detection limits
for most rare earth elements. Highest values were obtained for Ce, while La, Nd and Y
were only detected in some apatite grains. Total contents for rare earth element
oxides (REO) including only La, Ce and Nd are in the range of 0.04 to 0.91 with an
average of 0.25 percent by weight (Table 1). Slightly higher REO contents in apatite has
been reported in earlier studies4,5 (0.07 to 1.57 and averages of 0.38 to 0.85 percent
by weight) from spot analyses on individual grains of apatite by EMPA and LA‐ICP‐MS

while complete REO‐analyses of concentrates of apatite generated values from 0.45
to 1.04 percent by weight2,6. All the reported data demonstrates a strong dominance
for the light rare earth elements.
Monazite
Monazite (Ce,La,Nd,Th)PO4 occurs texturally in two different forms within the apatite
crystals. The monazite inclusions are mainly forming rods oriented parallel to the c‐
axis of the apatite grains but may also be tabular with a more random orientation
(Figure 1a). In most cases, inclusions are most abundant in the inner part of apatite
grains while the rim often lacks inclusions (Figure 1b). The rod‐shaped inclusions have
a length of up to 80 μm but are most commonly 10‐30 μm in size. The tabular shaped
monazite inclusions are mostly 5‐20 μm in size. The highest amount of monazite
inclusions is found in larger apatite grains within veins and schliren, while
disseminated more fine grained apatite in D‐ore might lack visible inclusions.
Inclusions of monazite are also rare in samples containing higher contents of biotite.
Besides occurring as inclusions in apatite, monazite is also locally found as up to
150 μm large single grains together with magnetite and gangue minerals.

Total content of REO in monazite (La, Ce, Pr, Nd) varies between 68.87 and 71.87 with
an average of 70.0 percent by weight (Table 1). Earlier reported chemical data4 for
monazite from Kiirunavaara is similar in contents with REO (La‐Dy) varying from 68.11
to 68.92 percent by weight.
Allanite
In general, allanite Ca(REE,Ca)Al2(Fe+II,Fe+III)(SiO4)(Si2O7)O(OH) is rare in samples
containing apatite with abundant inclusions of monazite but is rather common in
samples containing biotite. Allanite occurs partly as single grains in association with
apatite but also strongly intergrown with magnetite in up to 1 mm large aggregates
(Figure 1c). Euhedral grains of allanite may be up to 300 μm in size but most commonly
allanite occur in fine grained and anhedral to subhedral aggregates together with
apatite, biotite and locally chlorite (Figure 1d). The REO content (La, Ce, Pr, Nd) of
allanite varies from approx. 15 to 25 with an average of 22.5 percent by weight and
with dominance for the LREE (Table 1). Others reported4 values of approx. 25.5 % REO
(La, Ce, Pr, Nd, Sm, Gd, Dy) for allanite from Kiirunavaara.


a b
c d

Figure 1. Micrographs of apatite, monazite and allanite from the Kiirunavaara deposit3
a) photomicrograph of apatite grain with rod shaped and tabular inclusions of
monazite, crossed nicoles b) photomicrograph of apatite (grey) with monazite
inclusions mainly in the inner part of grains, crossed nicoles c) SEM backscatter
micrograph of allanite (medium gray) intergrown with magnetite (light gray) and
apatite (dark grey) d) photomicrograph of apatite (ap) with some biotite (bi), chlorite
(ch), and aggregates of allanite (al).

MATERIALS AND METHODS
This part is an update on earlier reported flotation results7. The materials received at
LTU were: General sample consisting of sub‐samples from a large number of drill‐holes
in the tailings pond, 5.85 % P2O5 and ≈1250 mg/kg REE; Coarse, low‐phosphorus
sample, 3.8 % P2O5; Fine high‐phosphorus sample, 6.6 % P2O5. Both the general sample
and the coarse product have wide particle size distributions, cf. Figure 2. A visual
inspection of the samples gave that this was due to pebble chips, size 5‐10 mm, from
the pebble mills. These chips must be removed prior to any flotation. They may be
removed by screening/wet classification or by grinding. A slight grinding is preferable
since it will also fresh the mineral surfaces for flotation.


Table 1. Monazite, allanite and apatite from Kiruna tailings (% by weight)
Mass‐ Monazite, S.D. Allanite, S.D. Apatite, S.D.
% n=11 n=14 n=39
SiO2 0.18 0.11 32.66 1.51 0.07 0.09
TiO2 0.01 0.01 0.13 0.09 0.01 0.02
Al2O3 0.03 0.06 15.58 2.00 0.16 0.60
FeO 0.21 0.15 15.09 1.20 0.12 0.07
MnO 0.03 0.03 0.34 0.16 0.02 0.02
MgO 0.00 0.01 0.32 0.10 0.01 0.01
CaO 0.13 0.07 11.97 1.76 54.55 0.53
Na2O 0.00 0.00 0.00 0.00 0.07 0.06
BaO 0.03 0.04 0.01 0.02 0.02 0.03
La2O3 14.816 1.879 6.989 3.006 0.020 0.042
Ce2O3 37.840 0.857 11.442 1.826 0.200 0.160
Nd2O3 11.951 1.106 2.446 1.106 0.027 0.050
Pr2O3 5.147 0.316 1.271 0.391 ‐
Y2O3 0.280 0.088 0.319 0.250 0.014 0.017
ThO2 0.471 0.282 0.062 0.056 0.017 0.017
UO2 0.024 0.031 0.000 0.000 0.000 0.002
P2O5 28.61 1.01 0.12 0.05 42.34 0.43
F 0.65 0.05 0.24 0.05 3.86 0.37
Cl ‐ ‐ 0.03 0.02
SUM 70.034 22.467 0.252
REO

At LTU, the general sample was sieved into fractions from 20 to 600 μm with sieve
ratio 2, and the sieve fractions were sent for chemical analysis. The results did show
that there was hardly any P2O5 in the fractions >300 μm, but approx. 30 % by weight
<20 μm. The REE do mostly follow the P2O5 distribution, but do not have the same
peak between 38‐250 μm, and there is more REE in the fractions >300 μm. To further
investigate, the correlation between P2O5 (apatite) and the REE, the mass and element
distributions for all sieve fractions were subjected to Principal Component Analysis
(PCA). A loadings (variable) plot (Figure 3) in the directions PC2 and PC3 shows that
REE is correlated, but not totally, with the P2O5 distribution. In contrast, the correlation
to Ca is strong and this implies that REE to a large extent substitutes for Ca. The
conclusion is that REE in the Kiruna tailings pond is, besides in apatite, also present in
some other Ca mineral, and that might be allanite or another REE‐containing mineral.

Figure 2. Particle size distributions for samples and after grinding7

Based on a pre‐study, the grinding time to freshen the mineral surface was locked at
10 minutes in a stainless steel rod mill (to not introduce colloidal iron). All flotation
tests were run with a talc pre‐float, to lower the Mg load, and an apatite rougher
flotation followed by three cleaner stages. The tests were run with process water from
Kiruna. This was sent to LTU in a 1 m³ tank and used in grinding, dilution after grinding,
in flotation cells and as rinsing water in flotation.

A full factorial statistical test plan with three parameters was used: pH adjusted by
NaOH solution, 9 – 11; Reagent (Collector) – Atrac 1563, 130 – 260 g/ton; Depressant
– water glass, 500 – 1000 g/ton.

RESULTS
Flotation tests
The tests were evaluated with grade‐recovery, selectivity curves, and difference
indices for the main analyses. Product balances were calculated for all analyses,
including individual REE, total REE, HREE and LREE. An example is given in Table 2 and
Figures 4a‐b. They show that it seems possible to achieve >30 % P2O5 in the
concentrate with approx. 70 % recovery and fair selectivity towards MgO, SiO2 and

Fe2O3. The selectivity diagram (Fig. 4b) also shows that REE is discharged in the rougher
and the first cleaner stage. Heavy REE (Y, Eu to Lu) follows apatite better, but are also
lost in the same way.

Figure 3. Multivariate loadings (variable) plot over mass and element distributions in
the general sample7

Table 2. Products balance for a good test7; pH, collector, depressant – all low
Product Mass Assays (%)
Notation Name (g) (%) SiO2 Al2O3 CaO Fe2O3K2O MgO MnO Na2O P2O5 TiO2
1 Talc conc 42.9 4.3 41.90 5.00 10.80 13.20 1.31 12.80 0.18 1.58 3.96 0.61
mp 1 Feed to flotation 965.8 95.7 32.66 6.03 12.77 25.99 1.62 5.21 0.17 2.08 6.24 0.73
2C Apatite conc 119.6 11.9 2.04 0.23 51.10 1.77 0.14 1.45 0.31 0.16 32.10 0.07
2CR Cl-middl 3 25.0 2.5 11.60 1.61 39.10 6.18 0.54 4.36 0.35 0.83 19.90 0.28
2B Clean conc 2 144.6 14.3 3.69 0.47 49.03 2.53 0.21 1.95 0.32 0.27 29.99 0.11
2BR Cl-middl 2 35.7 3.5 24.40 4.01 25.70 11.20 1.01 6.83 0.28 1.39 11.30 0.40
2A Clean conc 1 180.3 17.9 7.79 1.17 44.41 4.25 0.37 2.92 0.31 0.49 26.29 0.17
2AR Cl-middl 1 152.3 15.1 36.60 6.81 12.00 18.40 1.80 7.71 0.19 2.24 4.01 0.36
2 Rogher conc 332.6 33.0 20.98 3.75 29.57 10.73 1.02 5.11 0.26 1.29 16.09 0.25
mp 2 Tail 633.2 62.8 38.80 7.22 3.94 34.00 1.94 5.26 0.12 2.49 1.06 0.99
Calc.Feed 1008.7 100.0 33.06 5.98 12.68 25.44 1.61 5.53 0.17 2.06 6.14 0.73

Product Mass Distributions (%)
Notation Name (g) (%) SiO2 Al2O3 CaO Fe2O3K2O MgO MnO Na2O P2O5 TiO2
1 Talc conc 42.9 4.3 5.4 3.6 3.6 2.2 3.5 9.8 4.6 3.3 2.7 3.6
mp 1 Feed to flotation 965.8 95.7 94.6 96.4 96.4 97.8 96.5 90.2 95.4 96.7 97.3 96.4
2C Apatite conc 119.6 11.9 0.7 0.5 47.8 0.8 1.1 3.1 22.1 0.9 62.0 1.1
2CR Cl-middl 3 25.0 2.5 0.9 0.7 7.6 0.6 0.8 2.0 5.2 1.0 8.0 0.9
2B Clean conc 2 144.6 14.3 1.6 1.1 55.4 1.4 1.9 5.1 27.3 1.9 70.0 2.1
2BR Cl-middl 2 35.7 3.5 2.6 2.4 7.2 1.6 2.2 4.4 5.9 2.4 6.5 2.0
2A Clean conc 1 180.3 17.9 4.2 3.5 62.6 3.0 4.1 9.4 33.1 4.3 76.6 4.1
2AR Cl-middl 1 152.3 15.1 16.7 17.2 14.3 10.9 16.9 21.0 17.3 16.4 9.9 7.4
2 Rogher conc 332.6 33.0 20.9 20.7 76.9 13.9 21.0 30.5 50.5 20.7 86.4 11.4
mp 2 Tail 633.2 62.8 73.7 75.8 19.5 83.9 75.6 59.7 45.0 76.0 10.8 85.0
Calc.Feed 1008.7 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0

Figure 4. a) grade‐recovery and b) selectivity diagrams for a good test7
Response surfaces
To get a better view of how the parameters influence the responses, contour plots
were made for recoveries of P2O5, REE and HREE, cf. Figures. 5a‐5c. The overall view is
that the best recoveries are for pH 9, high collector dosage and low water glass
addition. Note that in all cases there is a local maximum for pH 11, high collector and
high water glass. Therefore, the experimental space is 3‐dimensional with a saddle
point at the centre of the test series. It is interesting to note that said recoveries show
the same general pattern. This indicates that the REE and HREE are either locked with
apatite or in free‐floating minerals with similar flotation properties as apatite.

a
b
c

Figure 5. Contour plots for a) recovery of P2O5, b) recovery of REE and c) recovery of
HREE to apatite concentrate

DISCUSSIONS
The flotation results show that the best conditions for both apatite and REE winning
are pH 9, high collector dosage (260 g/t) and low depressant addition (500 g/t). For
this condition, laboratory scale flotation gives an apatite concentrate with 30 % P2O5
at 70 % recovery.

REE follows apatite in the cleaning, but not totally. The recovery of total REE to apatite
concentrate is approx. half of that for apatite. For the heavy REE, it is slightly better.
The final grade of REE in the flotation concentrate is fairly constant and related to the

P2O5 grade in the apatite concentrate. This implies that REE is not totally in apatite,
and that there must be at least one unknown REE phase present. In later investigations
it has been seen that a shorter grinding time leads to slightly higher REE recoveries,
implying that a not so fine grinding gives better recoveries of liberated REE minerals.

When a full‐scale recovering‐dredging operation of the tailings pond is running during
the winter, the feed to the concentrator will be just above freezing. Unfortunately,
preliminary tests indicate unacceptable losses of apatite and REE when the flotation is
run below 10 °C with the present reagent mix.

It appears that a large proportion of the REE is lost already in the rougher flotation and
the first cleaner flotation. If this loss is caused by REE present in allanite, a test of the
magnetic properties of the tailings might give some indication of the possibility to use
high‐gradient magnetic separation and/or amine based flotation to increase the total
REE recovery.

FUTURE WORK
In the attempts to recover some of Europe´s most critical REEs from existing mine
waste, it is fundamental to understand the occurrence of all REE containing minerals
in the tailings and adapt flotation and other techniques to fit these specific minerals.
This is one of the objectives in the recently started European project REEcover, where
several European countries collaborate on finding new, and more efficient, techniques
for characterization, analysis, extraction, and metal production of REEs.

As can be seen in the data presented in this paper, there is a need improving the
accuracy of the characterization of REE‐bearing minerals, and also on improving the
detection of elements of very low concentrations, e.g. the HREE´s. To establish in detail
the distribution and character of REE‐bearing minerals with the purpose to develop
the right extraction technique, quantitative analysis using Inca software will be
performed on the Kiirunavaara material, as well as LA ICP MS analysis to be able to
detect the absolute concentrations of REE´s in different mineral phases.

In a recent mineralogical survey by LKAB on an apatite concentrate, the mineral grades
according to Table 3 were found. The mineral analyses come from Table 1. The
calculated combined REO contents in mg/kg are compared with chemical analyses
(ICP‐MS) from flotation tests7. The two flotation tests gave nearly the same
concentrate quality. The agreement between calculated and analysed contents is not
so good, but what stands out is that chemically analysed Y grades are approx. five
times higher than theoretically calculated. This indicates that the missing REE phase
carries a relatively high Y content, and possibly other HREE’s.

Table 3. Comparison of REE‐content in apatite concentrate
Concentrate Analysed grades in minerals
Mineral Grade (%) %SumREO %La2O3 %Ce2O3 %Pr2O3 %Nd2O3 %Y2O3
Apatite 70.65 0.252 0.020 0.200 0.027 0.014
Monazite 0.32 70.034 14.816 37.840 5.147 11.951 0.280
Allanite 0.40 22.467 6.989 11.442 1.271 2.446 0.319
Calculated concentrate grade contribution (mg/kg)

SumREO La2O3 Ce2O3 Pr2O3 Nd2O3 Y2O3
1 780 141 1 415 0 189 100
2 241 474 1 211 165 382 9
899 280 458 51 98 13
Combined REO 4 920 895 3 084 216 669 122
N101:Test REE 3 550 608 1 290 164 603 472

N105:Test REE 3 484 617 1 250 161 587 486
Conversion REE/REO 0.8000 0.8527 0.8537 0.8545 0.8574 0.7874

Test equivalent bulk chemistry REO 4 396 718 1 488 190 694 608
Difference (Bulk-EMPA) eq REO -524 -177 -1 596 -25 25 487
REFERENCES
1. S. Bergman, L. Kübler and O. Martinsson, “Description of regional geological and geophysical
maps of northern Norrbotten county”, Sveriges Geologiska Undersökning, Ba 56, 110 p (2001).
2. T. Parak, “Rare Earths in the apatite Iron ores of Lappland together with some data about the
Sr, Th and U content of these ores”, Economic Geology, 68:210‐221 (1973).
3. O. Martinsson, E. Öberg, and A. Fredriksson, “Apatite for extraction ‐ Mineralogy of apatite and
REE in the Kiirunavaara Fe‐deposit”, in XXVI International Mineral Processing Congress ‐ IMPC
2012, New Delhi, India, September 24‐28, 2012. ISBN: 81‐901714‐3‐7.
4. D.E. Harlov, U.B. Andersson, H.‐J. Förster, J.‐O. Nyström, P. Dulski and C. Broman, ”Apatite‐
monazite relations in the Kiirunavaara magnetite‐apatite ore, northern Sweden”, Chemical
Geology, 191:47‐72 (2002).
5. Å. Edfelt, M.P. Smith, R.N. Armstrong and O. Martinsson, “Apatite chemistry – applications for
characterizing apatite‐iron and IOCG deposits”, in The Tjårrojåkka apatite‐iron and Cu (‐Au)
deposit, Northern Sweden, Edited by Å Edfelt, Doctoral Thesis 2007:17, Luleå University of
Technology, 2007.
6. R. Frietsch and J.‐A. Perdahl, “Rare earth elements in apatite and magnetite in Kiruna‐type iron
ores and some other ore types”, Ore Geology Reviews, 9:489‐510 (1995).
7. B.I. Pålsson, and A. Fredriksson, “Apatite for extraction. II. Flotation of apatite and rare earth
elements from old tailings dumps”, in XXVI International Mineral Processing Congress ‐ IMPC
2012, New Delhi, India, September 24‐28, 2012. ISBN: 81‐901714‐3‐7.


Application of modern methods for red mud processing
to produce rare earth elements.
Olga PETRAKOVA1, Gennadiy KLIMENTENOK1, Andrey PANOV1, Sergey
GORBACHEV1
1
RUSAL ETC Ltd., 4 Klara Tzetkin Yekaterinburg, Russia
Olga.Petrakova@rusal.com
Abstract
Nowadays red muds of the Ural refineries, wherein practically the whole content of
scandium and the rare‐earth elements content of bauxites is concentrated, are
considered as one of the most promising raw sources of scandium in Russia. The paper
contains a review and analysis of the methods of REE and scandium extraction from
red mud known in the art to choose the most favorable one from ecological and
economic point of view among methods/group of methods of REE and scandium
extraction from this type of industrial wastes.

Introduction
Successful development of advanced engineering and technological industries is
primarily determined by the use of new materials and products based on these
materials. The use of scandium as an alloying additive in producing alloys makes it
possible to produce materials with improved performance characteristics, that is, high
strength, ductility, conductivity, etc. Another promising area of scandium application
is the production of high temperature ceramics, phosphors, dielectrics, catalysts,
refractories, etc. [1].
Currently, the red mud from Ural alumina refineries, in which almost all scandium and
rare‐earth metals contained in bauxite are concentrated, is regarded as one of the
most promising sources of raw scandium in Russia. Most scandium and lanthanum
contained in red mud is bound with iron and titanium minerals that undergo no
changes in alumina production [2]. In addition to the high metal content (more than
100 g/t), red mud is completely prepared for further metallurgical processing, unlike
other mineral formations. RUSAL's alumina refineries (UAZ and BAZ) annually stockpile
about 2 million tonnes of red mud, which contains more than 200 tonnes of scandium.
Notwithstanding the fact that scientists from around the world have been studying the
problem of integrated red mud processing over the last 50‐60 years, the least studied
area, in our opinion, is that associated with the recovery of rare‐earth elements and
scandium from red mud. According to the results of a patent search conducted by
Australian scientists from CSIRO, from 1964 to 2008, 135 patents were granted for

inventions in the field of red mud application in metallurgy, representing 17% of the
total number of patents for red mud application (the patent analysis is based on the
ISI Derwent Innovations Index data). A total of 10 patents, 2 and 8, respectively, were
granted for the rare‐earth metal and scandium recovery processes [3].
The purpose of this article is to summarise the information on the currently available
methods of rare‐earth metal and scandium recovery from red mud, as well as to select
the most acceptable (from environmental and economic perspectives) method / group
of methods for recovering rare‐earth metals and scandium from these types of man‐
made materials.
The technologies for rare‐earth metal recovery from red mud can be divided into two
large groups: technologies based on the solid phase rare‐earth metal pre‐
concentration methods followed by hydrometallurgical recovery of valuable
components, and technologies for rare‐earth metal digestion from the total volume of
red mud. The first group of technologies includes such methods of rare‐earth metal
and scandium concentration as gravitational processing, magnetic separation,
reduction roasting and reduction melting.

Technologies for rare‐earth metal recovery from red mud using
rare‐earth metal pre‐concentration methods
There are a number of studies by Soviet and Russian research teams related to the
application of a combination of methods of classification, gravity concentration and
magnetic separation to produce rare‐earth metal and scandium concentrate. The red
mud slurry was separated by density into sink and float. The sink (heavy fraction) was
then subjected to magnetic separation with a magnetic field intensity of 40‐160 kA/m
to produce magnetic concentrate containing rare‐earth elements, scandium and
yttrium, which can be recovered by any of the existing methods. The magnetic
concentrate produced was characterised by a low content of impurities, the recovery
rate of rare‐earth metal, scandium and yttrium was 50‐60%, whereas there occurred
no significant concentration of these elements as compared to their content in the
initial red mud (no more than 1.5 times) [4].
The authors of the study [5] proposed a method for producing scandium‐containing
magnetic concentrate from red mud using the method of red mud slurry classification
by particle size of 40‐60 microns and further magnetic separation of fine impurities,
with pH = 1.5‐4.0 and magnetic field intensity – 30‐80 kA/m. The magnetic concentrate
yield was 4‐7 % wt, the concentration of scandium in the concentrate was increased
to 0.032‐0.04 % wt compared to the scandium concentration in the initial red mud
(0.01 % wt) [5]. There is also a method of scandium concentrate production using the
combined methods of classification and magnetic separation, which consists in
separating fine‐grain fraction of red mud through classification, vibration cavitation

processing and further magnetic separation into magnetic and non‐magnetic product.
Under this method, the magnetic product was subjected to additional classification to
produce iron‐containing and scandium‐containing concentrates with 0.040% Sc2O3.
The concentrate yield was 5% of the initial red mud volume; the throughout recovery
of scandium into concentrate – 19% [6]. The application of the methods above for pre‐
concentrating rare earth metals and scandium makes it possible to reduce the total
volume of the solid phase targeted for digestion of valuable components (normally,
acids). This results in reduced operating costs by minimisation of waste subject to
disposal (acidic effluent, leaching cake, etc) as well as due to the process flow diagram
simplicity. Notwithstanding the relatively low operating costs, these methods are
characterised by a relatively low throughout recovery of scandium and rare‐earth
metals into the final product, as scandium recovery into the concentrate is less than
20%, and some part of scandium is lost in the process of sandium consentrate
preparation into commercial scandium and rare‐earth metal oxides. That is why their
cost effectiveness is still questionable. It should be noted that the red mud particle
size distribution has a strong influence on the technical and economic performance
indicators of the magnetic separation method used in rare‐earth metal concentrate
production, in particular, for red mud resulting from processing bauxite from the
Sredne‐Timansky bauxite mine at UAZ, the average particle size is less than 10 microns,
which precludes achieving the performance indicators specified in the studies [4‐6].
The application of the pyrometallurgical methods of red mud processing for pre‐
concentration of the valuable components is based on the reduction bake out and
reduction melting processes to produce iron and REEcontaining dross, which are
separated using the magnetic separation method. Subsequently, the dross is sent for
digestion of rare elements, i.e. rare‐earth metals and scandium, the digestion agent
being a 20‐30% solution of H2SO4 [7‐10]. The application of this group of methods
results in a sharp increase in capital and operating costs (high energy consumption).
Hence, the process of rare‐earth metal and scandium recovery is economically
inefficient.

Technologies for rare‐earth metal digesting from the total

volume of red mud
The second group of technologies for producing rare‐earth metals and scandium from
red mud includes a set of methods for digesting valuable components from the total
volume of red mud, followed by further recovery from the solutions. This group of
technologies can be divided into several subgroups depending on the type of digesting
agents used for scandium recovery into solution: acid and carbonaceous digestion. The
method of scandium carbonaceous digestion from red mud is based on the ability of
scandium to dissolve and form anionic complexes, like (Sc(CO3)n)(2n‐3)‐ (n≥2) over the

excessive hydrogen carbonates and carbonates of alkali elements [11]. The
carbonaceous process of scandium recovery into solution, followed by its separation
by hydrolysis or by sorption, is the most acceptable taking into account the specificity
of the alumina production. At the same time, this process ensures a relatively low rate
of scandium recovery into solution and does not allow simultaneous recovery of the
rare‐earth elements, the content of which in red mud exceeds 1 kg/t [12‐15]. Thus, the
problem of producing the combination of valuable elements (Sc, Ti, Zr, REM, etc.)
contained in red mud in sufficiently high amounts can be solved using the acid
digestion method only.
Based on the results of analysing literature data on the research in the given field
conducted here and abroad, a conclusion was drawn regarding the most suitable
digesting agents for recovery of rare‐earth metals and scandium from red mud. Some
of the most popular acid reagents for rare‐earth metal and scandium recovery are
sulfuric, hydrochloric and nitric acids. The bulk of studies are related to the use of
diluted and concentrated sulfuric acid for digesting rare‐earth metals and scandium
both from red mud and magnetic concentrate, as well as dross, after reduction
melting. The holders of the patent [16] propose a method for producing scandium‐
containing concentrate from red mud that includes the following stages: 1) digestion
of scandium from red mud with 10.0‐13.5% sulfuric acid with a preceding vibration
cavitation treatment of slurry, duration – 15‐35 minutes, process temperature – 30‐
95С, 2) scandium sorption from sulfuric solutions on phosphoric ion exchangers, 3)
wash down of the ion exchanger with a 1.0N HCl solution in order to remove titanium
and zirconium impurities, 4) desorption of scandium with carbonaceous solution, and
5) sedimentation of scandium concentrate using sodium caprate. The relatively low
acid concentration makes it possible to selectively recover scandium into solution,
leaving most of the iron in the solid phase. The method makes it possible to achieve a
70‐75.5% yield of scandium into the concentrate with its concentration in the final
product – 22.5‐25.0%. In the application [17], it is also proposed to treat red mud with
a low concentration sulfuric acid (less than 10%) instead of using ineffective
hydrochloric acid; scandium recovery into solution was 50%, yttrium recovery – 70%.
No data on recovery of other rare‐earth metals into solution are given in the patent,
no methods of scandium and yttrium recovery from solutions are specified. Russian
scientists studied the process of red mud treatment by the methods used in processing
low‐quality uranium ores, that is: the method of rare‐earth metal and scandium
digestion with mineral acids, along with the ion exchange separation of radioactive
and valuable components. Using the method of ion exchange in slurry with sulfuric
acid, these scientists demonstrated the possibility of recovering titanium, scandium,
uranium and thorium from the red mud at Nikolaev Alumina Refinery (Ukraine), where
bauxites from Brazil, Guyana, Australia, Guinea and India are processed [18]. The
advantages of using sulfuric acid for digesting processes over other reagents include
but are not limited to low cost, the high level of knowledge related to the chemistry of

sulfate compounds, and the ease of the sulfate ion disposal by its binding into slightly
soluble calcium sulphate.
Another popular reagent from the perspective of the research on rare‐earth metal and
scandium digestion from red mud is hydrochloric acid. Some researchers also suggest
the successive use of hydrochloric acid and sulfuric acid solutions for scandium
recovery from red mud [19]. To recover yttrium from red mud, there exists a widely
known method for red mud treatment with a 5.7% solution of hydrochloric acid at a
temperature of 85‐95С, where 70% of yttrium is recovered into the solution. This
method implies a single‐time or double circulation of the digesting solution in order to
increase yttrium content in the solution up to 0.16 g/l. The resulting solution is
subjected to further treatment using known methods for producing yttrium and its
components [20]. Chinese researchers from the Taiwan University of Technology
proposed a method for recovering scandium from modified red mud. This method
includes the following stages: hydrochloric digesting of scandium, scandium recovery
through extraction using composite extractant, and scandium concentrates
sedimentation with oxalic acid [21, 22]. These same researchers suggest producing
rare‐earth metal concentrate by direct hydrolysis deposition from hydrochloric red
mud digesting solution already silicon free [23].
Canadian researchers from Orbite Aluminae proposed a method of red mud complex
processing to produce iron‐bearing product, alumina and rare‐earth metal
concentrate [24]. This method includes the following stages: 1) Classification of
aluminium‐bearing material (50‐80 micron fraction will be separated), 2) Treatment of
the material with hydrochloric acid (with a concentration of 15‐45%) making it
possible, at a certain temperature and under certain pressure, to dissolve aluminium
and other elements, such as iron, magnesium and other metals, including rare‐earth
elements and/or rare metals. Silicon dioxide and titanium remain completely in the
solid phase; 3) Separation of the chloride solution from the solid residue; 4) Saturation
of the acidic solution with dry high‐concentration hydrogen chloride gas, which results
in aluminum chloride hexahydrate crystallisation (sedimentation) with minimum
concentration of other impurities; 5) Calcination of aluminium chloride
hexahydrate at high temperature and production of alumina. The high‐concentration
hydrogen chloride released during calcination is passed through water to achieve even
higher concentration, the final solution can be used at the digestion stage (recycling);
6) Ferrous chloride remaining in the solution is concentrated and hydrolysed at low
temperature to produce Fe2O3 (a type of hematite), the acid being recovered during
the process; 7) Treatment of the solution enriched with rare‐earth elements and/or
rare metals after hematite removal. Rare‐earth elements and rare metals in the form
of chloride are recovered from the solution using known methods such as the ultimate
REM concentrate (e.g., liquid solvent extraction). To extract rare‐earth metals,
scandium and some other elements, it is suggested to use oxy‐phenyl phosphoric acid
and toluene, HCl is suggested to be used for re‐extraction [24]. The main disadvantage

of the proposed technology is the complexity of the process of muriatic acid
recirculation, which requires serious measures to be taken to create an airtight system
for catching hydrogen chloride gas, as well as high capital expenses to ensure the
corrosion resistance of the equipment in contact with gaseous HCl. Similar drawbacks
hold true for the method of rare‐earth metal recovery from red mud using SO2
dissolved in water [25]. At the same time, the use of sulfurous acid makes it possible
to selectively recover rare‐earth metals into solution, leaving iron in the solid phase.
The use of alkyl carboxylic acids as a digesting agent also makes it possible to recover
selectively rare‐earth elements into solution without extracting any significant
amounts of iron into solution [26,27]. These technologies were trial tested on a pilot
plant using sintering red mud from the Boksitogorsk Alumina Refinery. Particularly, the
use of octanoic acid or the mix of synthetic fatty acids makes it possible to recover up
to 30% of scandium and 9% of La into solution. The resulting organic phase containing
rare‐earth metals, aluminium and calcium was washed down with sulfuric acid
solution (pH 3.0) in order to recover these elements into the aqueous phase, calcium
precipitated from the solution as insoluble sulfate (gypsum), whereas rare‐earth
metals precipitated from the solution by way of its treatment with a 10% NaOH
solution. Scandium was recovered from the organic phase after removing rare‐earth
metals, Al and Ca from it using sulfuric acid solution (pH 1.5) [26]. The same
researchers suggested a method of rare‐earth metal and scandium recovery from red
mud involving the digestion of aluminium, calcium and rare‐earth metals with formic
or acetic acids. It was proven that the recovery rate of the target products depends
significantly on the digestion time and temperature, i.e. the higher the temperature,
the faster the maximum recovery rate of the following elements can be achieved: Al
50 %, Сa 85 %, Sc 63 %, Yb 20 %. The digestion solution was treated with sulfuric acid
until the equilibrium value of pH 1.5 was reached in the aqueous phase, calcium in the
form gypsum precipitated, formic acid was distilled from the filtered solution, the
distillation residue was then treated with sodium carbonate solution, this resulted in
rare‐earth metals precipitating in the form of hydroxides, whereas aluminium and
scandium remained in the solution as carbonates [27]. The distinctive feature of the
last two methods of rare‐earth metal recovery into solution using alkyl carboxylic acids
is the necessity to use quite expensive reagents and formation of a large amount of
gypsum at the acid regeneration stage.
Application of nitric acid for rare earth metals (REM) recovery from red mud was
tested at the pilot unit, which makes it possible to feed 5 kg of red slurry at one time.
The pilot unit was designed and manufactured in the NTUA Chemical Engineering
Department laboratory (Athens, Greece) [28]. The designed technology includes
selective recovery of scandium and REM to solution by red slurry digestion with diluted
nitric acid solution; recovery of scandium and lanthanides using ion‐exchange
technique, and removal of diluted iron from their solution; further recovery of
scandium by liquid extraction from eluate using di‐(2‐ethylhexyl) phosphoric acid, and

directing it to purification and concentration [29, 30]. The rate of scandium, yttrium
and lanthanides recovery to solution reached 50‐75 % depending on the element. Iron
dissolution was insignificant. The acid capacity of a digesting agent was selected so
that the scandium content could be increased in the solution at minimum dissolution
of iron (< 4 %). The technology involves a process for nitric acid recycling. The limiting
stage for the whole technology is digestion of scandium and REM from red mud.
Therefore the pilot unit was manufactured in the NTUA laboratory.
The article [28] presents results of studies related to the following parameters’
influence on digestion process: type of agitation, solid‐to‐liquid ratio, concentration
of digesting agent, number of digestion stages, preliminary treatment of red slurry
with concentrated acids for maximum scandium recovery at minimum dissolution of
iron. The optimum conditions were selected for the scandium digestion process at a
pilot production scale. The highest scandium concentrations in the solution were
reached at the solid‐to‐liquid ratio increase, and media pH about 0. If pH increases,
scandium concentration in the solution will reduce. The maximum rate of scandium
recovery in the solution is 50 %, yttrium – 55 %. Behaviour of other REM during
scandium recovery from red mud was not studied in this article.
Unfortunately, the authors did not consider the issues related to recycling red mud,
and the environmental impact of the technology. Treatment of large volumes of red
mud with nitric acid involves the need to recover the nitrate ions excess absorbed by
red mud. As a result, large volumes of flushing water are produced. It is difficult to
neutralise such flushing water. As opposed to sulphate ions, nitrate ions cannot form
slightly soluble salts, and consequently, they cannot be removed using less expensive
techniques. At the same time, maximum allowable concentration for nitrate ions (45
mg/l) is much lower than for sulphate ions (500 mg/l).
Having reviewed various technologies for REM and scandium recovery from red mud,
we believe that, in terms of economic efficiency and advanced high‐tech solutions for
rare earth elements recovery and concentration, the most attractive option is the
technology involving all stages of sulphuric‐acid digestion of scandium and REM from
red mud, scandium and REM sorption, and concentration of scandium and REM from
the eluting solution.

Conclusion
At present there are many open issues related to selecting the most suitable
technology for scandium and REM recovery from red mud of alumina production. In
particular, most scientific studies lack information on specific REM behavior during red
mud processing. These studies are mainly focused on scandium and yttrium recovery
and related to the scientist’s assuming that all the rest of the elements are recovered
with the same volume as yttrium. However, this fact is not proven by the results of the
work [25, 26, 29, 30]. Besides, discussion of studies related to direct recovery of

scandium and REM from red mud does not cover the red mud recycling and
environmental issues.
At the same time, it is obvious that digestion of scandium and REM from red mud
should involve acid solutions with concentration making it possible to reach the
highest REM recovery rate at minimum iron dissolution. Since digestion solutions
contain insignificant volumes of scandium and REM, the most suitable technique for
concentration is ion‐exchange adsorption, which makes it possible to recover
selectively REM and scandium using different types of ionites, or varying desorption
conditions.

References
1. Scandium in bauxites and clays. / V.N. Lavrenchuk, A.V. Stryapkov, E.N. Kokovin; GUP SO Kamensk‐
Uralsk Printshop, 2004. – 291 pages.
2. Behaviour of scandium and lanthanum during alumina and bauxites production / V.A.Derevyankin,
T.P. Porotnikova, E.K. Kocherova, and others // Izvestiya (bulletin) of higher educational
institutions. Non‐ferrous metallurgy. 1981. No 4, P 86‐89.
3. C. Klauber, M. Gräfe, G. Power. Bauxite residue issues: II. Options for residue utilisation//
Hydrometallurgy.‐ June 2011.‐ V. 108; Issues 1‐2 ‐ P. 11‐32.
4. Patent 2147623 Russia, C22B59/00, C22B7/00, B03C1/00. Method of rare earth metals, scandium
and yttrium recovery from red muds of alumina production / S.L. Orlov, I.Yu. Entelis, B.N. Smirnov
– No 99120766/02; applied October 6, 1999; published April 20, 2000, Bulletin No 13‐2003.
5. Patent 1715874 А1, С 22 В 59/00//С 01 А 7/02. Processing technique for red muds from alumina
production / A.N. Kaluzhskiy, V.S. Anashkin, G.N. Klimentenok, and others. ‐ No 4120629; applied
September 19, 1986; published February 29, 1992, Bulletin No 8.
6. Application 2012104844/02 Russia, C01F7/02, C22B59/00, B03C1/00, B03B7/00. Processing
technique for red mud from alumina production / V.S. Anashkin, A.N. Bukharov, G.L. Girshina and
others ‐ Applied February 10, 2012; Published April 27, 2013.
7. L. Piga, F. Pochetti, L. Stoppa. Recovering metals from red mud generated during alumina
production // J. Met. – 1993. – V. 45 (11). – P. 54–59.
8. Crnko, V. G. Logomerac. Treatment of dissolved red mud slag by solvent extraction: technological
experiments // TRAVAUX (ICSOBA). 1976 ‐ No 13. ‐ P. 513‐518
9. Sargic V., Logomerac V.G. Leaching and extraction in the complex processing of red mud // Travaux
ICSOBA. – Zagreb. – 1974. № 11. – Р. 71‐78
10. E. Ercag, R. Apak. Furna Furnace Smelting and Extractive Metallurgy of Red Mud: Recovery of TiO
and Pig Iron // J. Chem. T echnol. Biotechnol.‐ 1997. ‐№. 70. Р. 241‐246.
11. Scandium ions condition in water solutions / L.N. Komissarova, ZhNKh, ‐ 1980. Vol. 2, Issue 1, P
143‐151.
12. Study of red mud treatment and recovery of concentrate scandium and other products: Report on
R & D / Sverdlovsk UPI. ‐ 1986. – P 49.
13. Patent 2247788 Russia. Method for scandium production from red mud / S.P. Yatsenko, N.A.
Sabirzyanov, L.A. Pasechnik, and others; Published March 10, 2005, Bulletin No 7.
14. S.P. Yatsenko, L.A. Pasechnik, I.N. Pyagay. Production of concentrates and pure scandium oxide
from dump mud using fuel gas from alumina production. / IV International Conference of the
Russian Chemical Society named after D.I. Mendeleev ‘Chemical technology and biotechnology of
new materials and products’ ‐ Moscow, 2012. ‐ P 341‐343.

15. A.B. Kozyrev, A.V. Nechaev, A.S. Sibilev, G.N. Klimentenok, O.V. Petrakova. Sorption recovery of
scandium from low‐concentrate raw materials ‐ red mud. / The 2nd Russian conference with
international participation , ‘New approaches in chemical engineering of minerals. Application of
extraction and sorption’ ‐ St. Petersburg 2013, P 145‐146.
16. Patent 2484164 Russia, С22B59/00, C22B3/08, C22B3/24. Method for scandium concentrate
production from red mud / V.S. Anashkin, A.N. Bukharov, G.L. Girshina and others; No
2011153732/02, Applied 27/12/2011; Published June 10, 2013.
17. Patent 2140998 Russia, C22B7/00, C22B59/00. Method for red mud processing / O.D. Linnikov, S.P.
Yatsenko, N.A. Sabirzyanov No 98122283/02. Applied 07.12.1998; Published November 10, 1999.
18. Smirnov, D.I., Molchanova, T.V., 1997. The investigation of sulphuric acid sorption recovery of
scandiumand uranium from the red mud of alumina production. Hydrometallurgy 45 (3), 249–259.
19. Patent 2040587 Russia, C22B59/00. Method for scandium recovery from red mud of alumina
production / V.N. Diev, S.P. Yatsenko, V.S. Anashkin and others.‐ No 93006354/02. Applied
February 3, 1993; Published July 25, 1995.
20. Patent 2057196 Russia, C22B59/00. Method for yttrium recovery from red mud of alumina
production / V.N. Diev, N.A. Sabirzyanov, V.S. Anashkin and others.‐ No 93006354/02. Applied
February 3, 1993; Published July 25, 1995.
21. Patent CN 102061392, C22B3/04, C22B3/40, C 22B59/00, C22B7/00. Method for scandium
recovery from modified red mud by using composite extractant / Aixiu Li, Keqin Wang, Hao Wang,
Yingbin Bai, Xiaoyun Zhang, Haixia Deng, Guohai Zhu, Qingzhen Gong, Meilin Zhang, Publication
Date 18 May 2011 (in Chinese).
22. Patent CN 102071317 (A). C22B3/40, C22B59/00. Composite extracting agent for scandium
recovery from red mud hydrochloric acid leachate/ Aixiu Li, Keqin Wang, Liping Cui, Jin Li,
Publication Date 25 May 2011 (in Chinese).
23. Patent CN 102127642 (A). C22B59/00, C22B7/00. Method for enriсhing rare earth elements from
red mud/ Aixiu Li, Keqin Wang, Fenglin Zhang, Yingbin Bai, Xiaoyun Zhang, Haixia Deng et. al.
Publication Date 20 June 2011 (in Chinese).
24. Patent WO 2013/104059 A1. International Publication Processes for treating red mud/ Boudreault
R., Fournier J., Primeau D., Labrecque‐Gilbert M., Date 18.07.2013
25. Patent 5030424 US/ Sato Recovery of rare earth elements from bayer process red mud/ G.Fulford,
G. Lever, T., Publication Date 9 July 1991.
26. Patent 2034066 Russia, C22B21/00, C22B26/20, C22B59/00, C01F7/02. Method for aluminum,
calcium and rare earth metals recovery from red mud of alumina production / P.V. Komarov; V.K.
Molotilkin; M.S. Polyakov; A.Yu. Shilnikov – No 92010983/02; Appl. December 1, 1992; Publ.
30.04.1995, Bulletin 24‐2000.
27. Patent 2048556 Russia, C22B21/00, C22B26/20, C22B59/00, C01F7/02. Method for aluminum,
calcium and rare earth metals recovery from red mud of alumina production / P.V. Komarov; M.S.
Polyakov; A.Yu. Shilnikov – No 92010984/02; Appl. December 1, 1992; Publ. November 20, 1995.
28. M. Ochsenkuhn‐Petropoulou, K. S. Hatzilyberis, L.N. Mendrinos, C. E. Salmas. Pilot‐Plant
Investigation of the Leaching Process for the Recovery of Scandium from Red Mud // Ind. Eng.
Chem. Res. – 2002/ ‐ V. 41/ ‐ P. 5794‐5801.
29. Ochsenkuhn‐Petropoulou,M., Lyberopulu, T., Parissakis, G. Selective separation and determination
of scandium from yttrium and lanthanides in red mud by a combined ion exchange/solvent
extraction method// Anal. Chim. Acta. – 1995. ‐ V. 315 (1–2). P. 231–237.
30. Ochsenkuhn‐Petropoulou, M., Lyberopulu, T., Ochsenkuhn, K., Parissakis, G. Recovery of
lanthanides and yttrium from red mud by selective leaching// Anal. Chim. Acta. ‐1996. – V. 319 (1–
2). P. 249–254.


HIGH PRESSURE ACID LEACHING OF A NICKEL LATERITE
ORE TO EXTRACT SCANDIUM
Şerif KAYA1, Yavuz Ali TOPKAYA2
1,2 Department of Metallurgical and Materials Engineering, Middle East Technical University
(METU), 06800, Ankara, Turkey
kserif@metu.edu.tr, topkaya@metu.edu.tr
Abstract
The purpose of this experimental study on laboratory scale was to investigate the high
pressure acid leaching (HPAL) behavior of the rare earth element scandium from a
lateritic Ni‐Co ore. In this work; the basic HPAL process parameters (sulfuric acid
concentration, leaching temperature, leaching duration, and feed size) were studied
and the optimum process conditions were determined to be as 260 kg sulfuric acid/ton
of dry ore, 255 °C leaching temperature, 1 hour of leaching duration and 100% ‐850
μm feed size. Under these process conditions; 80.6% of scandium, 85.7% of nickel and
86.2% of cobalt could be extracted into the pregnant leach solution. In order to see the
effect of finer feed size and longer leaching duration together, the ore was further
ground to 100% ‐75 μm and leached for 360 minutes. It was seen that even under these
extreme leaching conditions; only 84.7% of scandium, 88.7% of nickel and 90.6% of
cobalt could be extracted. Thus; the present investigation has shown that the presence
of chemically resistant minerals such as hematite and some clay minerals in the lateritic
ore could be the possible reason for this incomplete dissolution behavior.

Introduction
Lateritic (oxide) and sulfide type nickel ores are the primary nickel and cobalt
resources. Although most of the nickel has been supplied from the sulfide based
resources since 1950’s, the trend in nickel production changed towards the more
abundant laterite ores which constitute for about 70% of the world nickel resources.
To confirm these remarks in the light of the historical data, it is seen that a small
portion of (less than 10%) nickel was produced from lateritic sources in 1950’s.
However, this ratio has increased up to ~42% in 2003, and rose up to ~50% in 2012.
Moreover, it is forecasted that around 55‐60% of nickel will be derived from lateritic
sources in 2020’s [1‐3]. In addition to nickel and cobalt, the presence of scandium has
been detected in these resources and it seems that the importance of lateritic nickel
ores will be more prominent in the near future. Since the supply of scandium is very
scarce throughout the world, its field of application is limited for the time being and it
is mainly used in the production of Al‐Sc alloys, lighting, electronics, laser research and

solid oxide fuel cells (SOFC’s). However, it is believed that new application areas will
emerge parallel to the increase in its supply.

Scandium is known to be present in many ore deposits and it was stated that the ores
can be considered as worthy of exploitation if the scandium content ranges between
20 and 50 g/t [4]. Due to being widely dispersed, it is obtained during processing of
various ores or tailings and residues from previously exploited mines as a by‐product.
In particular, it is obtained during processing of uranium, tin, tantalum, titanium,
tungsten and rare‐earth element sources and the major producer countries are China,
Kazakhstan, Russia and Ukraine [5]. Due to its lack of affinity for the common ore
forming anions, scandium is rarely concentrated in the Earth’s crust and generally
found in more than 100 minerals in trace amounts in the form of solid solutions [5].
Due to the similarities in ionic radius, the most common substitution mechanism is the
replacement of Fe3+ and Al3+ with Sc3+. Scandium in laterites is distributed widely
among many minerals but may be especially associated with goethite, clay minerals or
manganese oxides such as nickel‐rich asbolane, in which it substitutes for Fe3+ and Al3+.
Because scandium is distributed in much the same manner as nickel in lateritic ore
minerals, it can be assumed that the leaching methods that solubilize nickel will also
solubilize scandium. Some of the companies and research organizations working on
leaching and recovery of scandium projects are: Jervois Mining Limited, EMC Metals
Corp., Metallica Minerals, Sumitomo Metals Mining Company, Platina Resources,
Pacific Metals Company and CSIRO.

Therefore; the aim of this study was to investigate the leaching behavior of rare earth
metal scandium together with nickel and cobalt during high pressure sulfuric acid
leaching (HPAL) of a lateritic ore. During this investigation, the basic process
parameters (acid/ore ratio, leaching temperature, particle size, and leaching time)
were studied and important findings have been presented in this paper.

Experimental
Ore Sampling and Materials Characterization
The sample was obtained from about 5000 tons of stockpile from Eskişehir/Yunusemre
lateritic ore body located in Turkey. This representative sample was further splitted up
to suitable fractions by coning and quartering method. Initially the bulk and solid
density measurements were performed. ‘Quantachrome Ultrapycnometer 1000’
helium gas pycnometer was used for its solid density determination. Multiple methods
were used in chemical analyses during the study. Inductively Coupled Plasma (ICP)
analyses were done by ‘Perkin Elmer DRC‐II ICP‐MS’ and ‘Perkin Elmer Optima 4300
DV ICP‐OES’ devices. X‐Ray Fluorescence (XRF) analyses were conducted using a ‘Niton
X‐Met 820’ analyzer, and finally Atomic Absorption Spectroscopy (AAS) analyses were

performed by using ‘Perkin Elmer Analyst 400 AA’ analyzer. For mineralogical
characterization, X‐Ray Diffraction (XRD), Scanning Electron Microscopy (SEM), and
Differential Thermal Analysis (DTA)‐Thermo Gravimetric Analysis (TGA) methods were
used. XRD measurements were done by a ‘Rigaku Ultima‐IV’ model X‐Ray
Diffractometer with a Cu‐Kα X‐ray tube working under 40 kV and 40 mA. Finally, the
presence of minerals was verified with an ‘FEI Nova Nano 430’ model scanning
electron microscope.

Experimental Set‐up and Procedure
High pressure sulfuric acid leaching experiments were performed in a 2 L, Parr‐4532
titanium (grade‐4) autoclave. To observe the leaching behavior; a slurry (150 g dried
ore and 350 cc deionized water) of 30% solid content (excluding acid) were digested
with 96‐98 wt. % reagent grade sulfuric acid with a stirring speed of 400 rpm in all
experiments. Because there was no direct acid injection system in the autoclave,
sulfuric acid was added initially together with the ore slurry and the reactor was
allowed to heat up to the desired leaching temperature in 40‐45 minutes. The start of
the reaction was defined as zero time when the reactor temperature reached to the
set‐point; subsequently, the leaching duration was recorded. After the completion of
the reaction, the system was allowed to cool down to room temperature via water
circulation system inside the reaction chamber. After digestion, the slurry was vacuum
filtered in a Buchner funnel using a Whatman grade‐4 filter paper for solid/liquid
separation. Leach residue remaining on the filter paper was washed well with slightly
acidic (pH 2) deionized water to clean the residue from the remaining pregnant leach
solution fully without the precipitation of a portion of ions at higher pH. After washing,
the leach residue was dried overnight at 105 °C and ground for chemical analyses.
Meanwhile; the filtrate of the slurry was stored for chemical analyses, free acid
determination, and oxidation‐reduction potential (ORP) measurement. In order to
measure the free acid remaining after digestion, the pregnant leach solution was
titrated with 0.2 M sodium hydroxide (NaOH) solution. During titration, 280 g/L of
reagent grade di‐potassium oxalate monohydrate solution was used as a chelating
agent to eliminate the interfering effect of some ions. Then, the residual acid was
calculated from the amount of NaOH consumed via titration. For the oxidation–
reduction potential measurements, a Pt‐Ag/AgCl electrode (saturated with KCl) was
used and the measured values were reported according to the Pt‐Ag/AgCl reference
electrode.


Results and discussion
Characterization Results
The as‐received ore was found to contain about 14% moisture with the bulk and solid
densities of 1.08 g/cm3 and 3.51 g/cm3, respectively. Chemical analysis of the
representative lateritic ore sample is presented in Table 1.

Table 1 Chemical analysis of the lateritic ore

Constituent Fe Ni Co Sc (g/t) Al2O3 MnO MgO
Content (wt. %) 46.2 1.54 0.093 106 3.61 1.04 0.71
Constituent Cr2O3 SiO2 CaO Na2O K2O As S
Content (wt. %) 2.25 14.56 0.98 0.02 0.02 0.12 0.02

Mineralogical analysis of the sample was performed by XRD, SEM and DTA/TGA
methods. XRD results given in Figure 1 showed that hematite (Fe2O3), quartz (SiO2),
and goethite (FeOOH) were the dominant mineral phases present in the ore. The
presence of the goethite phase was also verified by DTA/TGA which exhibited a distinct
endothermic peak between 260‐340 °C according to the following dehydroxylation
reaction 2FeOOH → Fe2O3 + H2O [6]. Additionally, minor amounts of nimite
((Ni,Mg,Fe++)5Al(Si3Al)O10(OH)8) and kaolinite (Al2Si2O5(OH)4 phases were also
detected. Since, the peaks of these phases were small and dominated by the peaks of
previously mentioned major ones; they were verified by separating clay and non‐clay
fractions and analyzing by XRD and SEM in detail.


Figure 1 XRD Patterns of the laterite ore sample and its HPAL leach residue
The SEM results also provided information regarding the other elements having no
distinct mineral phases in the sample. According to the SEM results; nickel, chromium
and manganese were found to be distributed substitutionally within the crystal
structures of the iron oxides and the clay minerals (kaolinite, nimite). Similarly, cobalt
was mainly found to be present in the crystal structure of clay minerals. However,
there was no evidence for the distribution of scandium in the SEM results due to its
relatively lower concentration in the sample. In order to interpret the distribution of
scandium; the ore sample was separated into different size fractions (‐850 +710, ‐710
+500, ‐500 +355, ‐355 +250, ‐250 +180, ‐180 +125, ‐125 +90, ‐90 +63, ‐63 +45 and ‐45
µm) by wet screening and each fraction was analyzed chemically. It was seen that
there was a strong correlation in the distribution of scandium, iron and aluminum,
which gave an indication about the presence of scandium within the crystal structures
of the iron and clay minerals.

Extraction Results
Reliability of the experiments was controlled by repeating four analogous digestion
tests with the same process parameters. The extraction results obtained from these
tests are reported in Table 2.

Table 2 Reliability of HPAL experiments

Element HPAL‐I HPAL‐II HPAL‐III HPAL‐IV Mean Std. Dev.
Ni Ext.(%) 86.2 84.9 86.1 85.5 85.7 0.5
Co Ext. (%) 86.7 86.9 85.6 85.7 86.2 0.6
Sc Ext.(%) 81.5 79.9 80.9 80.1 80.6 0.6

The experiments given in Table 2 were performed with the following process
parameters; 255 °C leaching temperature with 0.260 acid/ore ratio for 60 minutes of
leaching duration and ‐850 μm particle size. It is obvious from Table 2 that the nickel,
cobalt and scandium extractions were sufficiently close to relying on the results of the
experiments. Thus, the given parameters in Table 3 were tested, respectively.


Table 3 Studied process parameters during HPAL experiments

Experiments Fixed Parameters Studied Parameters
Sulfuric Acid/Ore 255 °C, 60 min., ‐850 μm 0.16, 0.18, 0.20, 0.22, 0.24, 0.26, 0.27,
Ratio (w/w) 0.285, 0.30, 0.325, 0.35
Leaching Temp. (°C) 60 min., ‐850 μm, 0.260 245, 255, 265, 275
acid/ore ratio
Leaching Duration 255 °C, ‐850 μm, 0.260 0, 30, 60, 90, 120, 360
(min.) acid/ore ratio
Particle Size (μm) 255 °C, 60 min., 0.260 ‐75, ‐425, ‐850
acid/ore ratio

Effect of Sulfuric Acid/Ore Ratio
Initial the acid requirement was calculated according to the practical data of the
operating plants and previously published studies [7, 8]. The first test was conducted
with 160 kg of sulfuric acid per ton of dry ore and the highest amount of acid loading
experiment was done with 350 kg of sulfuric acid. The effect of increasing acid load on
the metal extractions can be seen in Figure 2.

Figure 2 Effect of acid load on nickel, cobalt and scandium extractions


According to the data plotted in Figure 2, there was no significant change in the
extraction of nickel. However; the cobalt extractions appeared to be adversely
affected beyond 300 kg of acid per ton dry ore. In terms of scandium, the acid additions
beyond 260 kg of acid per ton of dry ore led to a drastic decrease in the extraction of
scandium. In order to understand this anomaly, the leach residues of the performed
experiments were characterized and it was observed that a new iron sulfate hydrate
phase (Fe2O3.2(SO3).H2O) was formed due to excessive acid load as shown in Figure 3.

Figure 3 XRD Patterns of leach residues w.r.t increasing acid concentration

Similar precipitates were also observed at Moa Bay plant in the form of yellow balls at
the point of acid injection in the first autoclave where zones of high acidity existed.
Literature also supports this finding by stating that solids with sulfur contents similar
to those of basic iron sulfate or jarosite form at free acid concentrations above a
‘break‐point’, while a sulfur‐deficient material (hematite) forms at acid concentrations
below this break‐point [9]. Since Sc3+ and Fe3+ have similar ionic radii, the available
findings of the experiments indicated the possible precipitation of scandium together
with iron sulfate hydrate due to higher acid loading. The previous mineralogical studies
also indicated that Sc3+ substitutes for Fe3+ in the jarosite type structures due to similar
reason [10]. Thus, the following HPAL tests were conducted with 260 kg of sulfuric acid
per ton of dry lateritic ore.


Effect of Temperature
The positive effect of higher leaching temperatures during HPAL (in a limited
temperature range) on nickel and cobalt extractions was reported in literature in the
past. Generally, the leaching temperatures higher than 200 °C are selected due to
higher nickel and cobalt extractions together with the dissolution and reprecipitation
of iron, aluminum and some chromium at higher temperatures. These precipitation
reactions provide cleaner pregnant leach solution for further processing and also
decrease reagent consumption by regenerating the consumed acid at these
temperatures. However, due to technical and economic reasons, higher leaching
temperatures are limited up to around 275 °C. Moreover; the experimental studies
reported the negative effect of increasing temperature beyond 270 °C due to co‐
precipitation of nickel in the newly formed precipitates such as ((Mg,Ni)SO4.H2O) [7].
Therefore; the temperature variation experiments were limited to the temperature
range of 245‐275 °C. The results of these experiments are given in Figure 4;

Figure 4 Effect of leaching temperature on nickel, cobalt and scandium extractions

According to the results plotted in Figure 4, it was observed that cobalt was unaffected
from the temperature change. However, there was ~4% increase in the extraction of
nickel with increasing leaching temperature. In terms of scandium, leaching
temperatures higher than 255 °C led to a drastic decrease in the scandium extraction
into the leach solution. To understand this unexpected observation, the XRD patterns
of the experiments given in Figure 5 were investigated and it was seen that the
formation of a new phase came into picture again as experienced in the acid load
experiments.

Figure 5 XRD Patterns of leach residues w.r.t temperature

It was stated before by other workers in this field that leach temperature, acid/ore
ratio and leaching duration have a significant effect on the sulfate content of the leach
residue [11]. In particular, sulfur losses (i.e. solids sulfate content) are greater for
higher leach temperatures and higher acid/ore ratios. The increase in sulfur content
of the leach residue with temperature suggests the formation of a hydrolysis product
with high sulfur content such as basic iron sulfates. To see this effect, the sulfur
content of the leach residues were analyzed and it was seen that sulfur contents
increased from 1.13% to 1.15%, then to 1.31% and finally to 1.70% with 10 °C increases
in temperature from 245 °C to 275 °C, respectively. Thus, the remaining HPAL tests
were conducted at 255 °C to determine the effect of the other parameters.

Effect of Particle Size and Leaching Duration
In the above mentioned work, the acid concentration and temperature were
determined to be 260 kg acid/ton of dry ore and 255 °C within 60 minutes of digestion,
respectively. However, the obtained extraction results were unsatisfactory since only
85.7% nickel, 86.2% cobalt and 80.6% scandium could be extracted into the pregnant
leach solution with these parameters. Thus; in order to increase the metal extractions,
finer particle size and higher leaching durations were tested. In industrial practice, the
pressure leaching is generally conducted between 60‐120 minutes. In order to study a
broader range, the pressure leaching experiments were conducted between 0‐360
minutes. According to the results presented in Figure 6 (left), it seems that the leaching
was almost complete within the first 120 minutes of reaction and there was no effect
of further leaching durations on the extraction of the desired metals.

Figure 6 Effect of leaching duration on nickel, cobalt and scandium extractions (left),
Effect of particle size on nickel, cobalt and scandium extractions (right)

Since longer leaching durations did not give satisfactory result beyond 120 minutes,
higher extraction efficiencies were tried to be obtained by finer particle size.
Therefore; the ore used in the previous experiments (100% ‐850 µm) was further
ground to 100% ‐425 µm and 100% ‐75 µm, respectively. As a result of grinding, only
2‐3% increase was observed in the extraction yields of desired metals within 60
minutes as presented in Figure 6 (right). However; this increase was observed to be
unsatisfactory due to over 10% un‐reacted nickel, cobalt and scandium in the leach
residue. Thus, to investigate an extreme case, the leaching duration of the finely
ground ore (100% ‐75µm) was increased further from 60 minutes to 360 minutes. With
these parameters ~3% increase in the nickel extraction and ~4% increases in the
extractions of cobalt and scandium were observed.

In summary; even with the extreme leaching parameters (‐75 µm particle size and 6
hours of leaching duration), approximately 10‐15% of nickel, cobalt and scandium
could not be extracted into the leach solution. To understand the possible reasons of
these low extraction results, the XRD findings of the original sample and its leach
residue obtained under the HPAL conditions of 260 kg sulfuric acid/ton of dry ore, 255
°C leaching temperature, 1 hour of leaching duration and 100% ‐850 μm particle size
as given in Figure 1 were compared. According to Figure 1, it can be seen that almost
all of goethite was leached and no remaining goethite phase was present in the leach
residue. However, hematite and clay minerals were still present in the leach residue,
which may be due to their incomplete dissolution during leaching. To support this
finding, there are some reported studies referring to the difficulty in leaching hematite
mineral [12, 13]. Therefore; undissolved primary hematite remaining in the leach
residue may be one of the possible reasons of low metal extractions. In terms of clay
minerals, it is known that these minerals are quite soluble during sulfuric acid leaching.
However, in a previous study it was reported that kaolinite and clinochlore (chlorite

group clay member), similar to nimite, exhibited resistance to pressure acid leaching
[14] and remained in the leach residue. According to the literature data, the dissolved
iron species start to precipitate by ex‐situ precipitation in solution on the primary
hematite surfaces. This precipitation reaction was reported to begin within the first
several minutes of leaching [11, 15, 16]. Thus, the precipitation of iron as hematite on
the primary hematite particles could be one of the possible reasons of the un‐reacted
hematite mineral observed in the leach residue.

Conclusions
To test the extraction behavior of nickel, cobalt together with the rare earth metal
scandium in a high pressure acid leaching process, the lateritic ore was digested by
varying the following leach parameters: acid load, leaching temperature, leaching
duration and particle size. In acid concentration experiments, it was seen that the
excessive acid load beyond 260 kg/ton of dry ore led to the formation of iron sulfate
hydrate precipitate and parallel to the formation of this precipitate, the scandium
extractions decreased considerably. In terms of leaching temperature experiments,
there was also a sign of sulfate formation and a decrease in the extraction values of
scandium at higher leaching temperatures. The connection between the iron sulfate
hydrate formation and decreasing scandium extractions was predicted to be due to
simultaneous precipitation of scandium together with iron at higher acid loads and
higher leaching temperatures due to similar ionic radii. Thus, the acid load and
leaching temperature were selected to be 260 kg/ton dry ore and 255 °C, respectively.
The leaching duration and particle size experiments indicated that 10‐15% of nickel,
cobalt and scandium could not be extracted into the leach solution even under the
extreme leaching conditions (‐75 µm particle size and 6 hours of leaching duration) at
255 °C and with 260 kg acid/ton dry ore. Thus, the possible cause of this behavior was
investigated and it was decided to be due to incomplete dissolution of refractory
minerals predicted to be as primary hematite and clay minerals.

Acknowledgements
The authors would like to express their special thanks to META Nikel Kobalt A.Ş. both
for supplying the limonitic ore sample of Eskişehir/Yunusemre and for supporting the
chemical analyses by XRF, AAS and ICP methods. In addition, the contributions of the
METU Central Laboratory in ICP/MS/OES, DTA/TGA, and XRD analyses, the METU
Metallurgical and Materials Engineering Department in SEM analyses and the SGS
Ankara Geochem Lab. in ICP/OES analyses are gratefully acknowledged. The
suggestions and support of MEAB Metallextraktion Company is also appreciated.

Finally, the authors are thankful to TÜBİTAK for providing a financial grant to support
this study.

References

1. Jolly, P., Comparison of various methods for calculating cash costs‐focus on nickel industry.
Proceedings of ALTA 2012 Nickel/Cobalt/Copper‐Processing of Laterites. 2012. Perth, Australia:
ALTA Metallurgical Services.
2. Sudol, S., The thunder from down under: Everything you wanted to know about laterites but
we were afraid to ask. Canadian Mining Journal, 2005. 126 (5): p. 8‐12.
3. Dalvi, A.D., W.G. Bacon, and R.C. Osborne, The past and the future of nickel laterites. PDAC
2004 International Convention, Trade Show & Investors Exchange. 2004. Ontario, Canada: Inco
Limited.
4. Xu, S. and S. Li, Review of the extractive metallurgy of scandium in China. Hydrometallurgy,
1996. 42 (3): p. 337‐343.
5. Gambogi, J., USGS Mineral commodity summaries 2013. 2013, United States Geological Survey
(USGS). p. 141‐142.
6. Földvári, M., Handbook of thermogravimetric system of minerals and its use in geological
practice. Occasional Papers of the Geological Institute of Hungary. Vol. Vol. 213. 2011,
Budapest.
7. Whittington, B.I. and D. Muir, Pressure acid leaching of nickel laterites: A Review. Mineral
Processing and Extractive Metallurgy Review, 2000. 21(6): p. 527‐600.
8. Seçen, B., Pressure Leaching of Sivrihisar‐Yunusemre Nickel Laterites, Metallurgical and
Materials Engineering Department. 2011, Middle East Technical University M.Sc. Thesis:
Ankara. p. 119.
9. Whittington, B.I., R.G. McDonald, J.A. Johnson and D.M. Muir, Pressure acid leaching of arid‐
region nickel laterite ore: Part I: Effect of water quality. Hydrometallurgy, 2003. 70 (1‐3): p. 31‐
46.
10. Dutrizac, J.E. and T.T. Chen, The behaviour of scandium, yttrium and uranium during jarosite
precipitation. Hydrometallurgy, 2009. 98 (1‐2): p. 128‐135.
11. Chou, E.C., P.B. Queneau, and R.S. Rickard, Sulphuric acid pressure leaching of nickeliferous
limonites. Metallurgical Transactions B, 1977. 8B (4): p. 547‐554.
12. Liu, K., Q. Chen, and H. Hu, Comparative leaching of minerals by sulphuric acid in a Chinese
ferruginous nickel laterite ore. Hydrometallurgy, 2009. 98 (3‐4): p. 281‐186.
13. Kaya, Ş. and Y.A. Topkaya, High pressure acid leaching of a refractory lateritic nickel ore.
Minerals Engineering, 2011. 24 (11): p. 1188‐1197.
14. Chen, G., Pressure acid leaching of nickel laterite, Department of Metals and Materials
Engineering. 1998, University of British Columbia., MSc Thesis.
15. Papangelakis, V.G., D. Georgiou, and D.H. Rubisov, Control of iron during the sulphuric acid
pressure leaching of limonitic laterites. Second International Symposium on Iron Control in
Hydrometallurgy, CIM 1996: Ontario, Canada. p. 263‐274.
16. Georgiou, D. and V.G. Papangelakis, Sulphuric acid pressure leaching of a limonitic laterite:
Chemistry and kinetics. Hydrometallurgy, 1998. 49 (1‐2): p. 23‐46.


LEACHING OF RARE EARTHS FROM BAUXITE RESIDUES

USING IMIDAZOLIUM BASED IONIC LIQUIDS.
Panagiotis Davris1, Efthymios Balomenos, Dimitris Panias2 and Ioannis
Paspaliaris
National Technical University of Athens School of Mining and Metallurgical Engineering,
Laboratory of Metallurgy
9, Iroon Polytechniou str. 15780 Zografos Campous – Athens, Greece
E‐mail address: 1panosdavris@metal.ntua.gr, 2panias@metal.ntua.gr
Abstract
Ionic liquids are versatile solvents that can be used as alternatives in the field of
hydrometallurgy. At this study, a brief review of ionic liquids utilization in metal oxides
and mineral leaching will be addressed followed by investigations for direct leaching of
rare earth metals from synthetic REE oxides and metallurgical by‐products. In
particular, leaching of pure oxides of Lanthanum, Neodymium, Dysprosium and Yttrium
and also Bauxite Residues (a solid residue generated from Bayer process in the primary
aluminum industry), is investigated with pure Ionic liquid 1‐ethyl ‐3methylimidazolium
hydrogen sulfate (EmimHSO4/C2mimHSO4). Synthetic rare earth oxides totally dissolve
in neat ionic liquid and high recoveries of rare earths and base metals were observed
from bauxite residues.

Introduction
Ionic liquids (ILs) are ionic pairs which are liquid at ambient or low temperatures. Their
properties include non‐flammability, low vapor pressure, thermal stability, high
conductivity and wide electrochemical window. Ionic liquids have been identified as
alternatives to conventional solvents and are capable of enhancing the yield and
selectivity of reactions [1]. Furthermore ILs are nonvolatile solvents in contrast with
organic solvents and thus are safer to use in above ambient temperatures. They
consist of an organic cation and an inorganic anion of different sizes (Fig. 1).

Figure 5 Chemical structure of 1‐alkyl‐3‐methyl‐imidazolium ionic liquid
Their properties play an important role as solvents for a number of applications like
catalytic reactions [2‐3] separation process [4‐5] and electrochemistry [6]. IL’s can be used
as design solvents or the so‐called task specific ionic liquids, due to their ability to

adjust their physicochemical properties in a variety of applications. Thus hydrophobic
or hydrophilic behavior can be monitored with modifying the number or the length of
alkyl chains in the cation [4] and also their miscibility with water by using different
anions [2]. Second generation ionic liquids that consist of 1‐ethyl/butyl‐3methyl‐
imidazolium cations with discrete inorganic anions are growing in interest for
exploitation in mineral processing.
In this paper a review of the use of ionic liquids in mineral processing will be addressed
and also the study of applying ionic liquid 1‐ethyl‐3‐methyl‐imidazolium hydrogen
sulfate (EmimHSO4) as leaching medium for rare earth oxides will be presented.

Dissolution of metal oxides using ionic liquids
The first reported studies of leaching metal oxides in ionic liquids concerned the
dissolution of UO3 and V2O5 in imidazolium chloroaluminate melt [7‐8]. The solubility of
UO3 was found to be 1.5‐2.5 10‐2 mol/dm3 and the main species in solution was
[UO2Cl4]2‐. Vanadium oxide was found to be very soluble in the basic melt, 0.15 g in 1
g, but also soluble in the neutral melt. Task specific ionic liquids are ionic liquids
designed for specific purposes such as betaine
[9]
bis(trifluoromethylsulfonyl)imide (HbetTf2N), which have been used for the
dissolution of the rare earth oxides; imidazolium, pyridinium, pyrrolidinium,
piperidinium, morpholinium, and quaternary ammonium
bis(trifluoromethylsulfonyl)imide salts were also functionalized with a carboxyl group
for the selective dissolution of metal oxides and hydroxides [10]. Oxides such as rare
earth oxides, UO3, PbO, ZnO, CdO, HgO, CuO, Ag2O, NiO, PdO as well as several
hydroxides were soluble in these ionic liquids whereas Co oxides, Cr2O3, FeO, Fe3O4
found to be insoluble in the same experimental conditions.
Deep eutectic solvents of choline chloride mixtures with hydrogen bond donors such
as carboxylic acids, amines and alcohols have been studied for dissolving metal oxides
[11‐15]
. The solubilities of ZnO, CuO, and Fe3O4 have been determined in 3 deep eutectic
solvents between choline chloride and carboxylic acids at 500C [12]. Fe3O4 is more
soluble in oxalic acid/ChCl system whereas 20 times less soluble in phenylpropionic
acid mixture while CuO showed the opposite behavior, suggesting that deep eutectic
solvents can be applied for selectively leaching of metals. In addition the solubility of
17 common metal oxides have been determined in ionic liquids of choline chloride
mixtures with urea, malonic acid, and ethylene glycol [13]. At present the behavior of
rare earth oxides in such deep eutectic solvents have not been studied.
The dissolution of Ca, Ni, Mn, Co, Cu, and Zn oxides in hydrophobic ionic liquid trihexyl
(tetradecyl) phosphonium chloride (Cyphos IL 101) saturated with a concentrated
aqueous hydrochloric acid solution have been also investigated [16]. Here the IL works
as in solvent extraction, as the dissolved oxides can be in subsequent step recovered
in an aqueous phase through selective stripping.

A summary of ionic liquids reported for metal oxide dissolution is shown table 1.


Table 4 Dissolving metal oxides in ionic liquids
Refer‐
Ionic liquid/leaching agent Metal oxides studied
ence
EmimCl/AlCl3
Uranium‐ Vanadium 7,8
DMPIC/AlCl3
HbetTf2N, [C6Hbet]Tf2N, [C4Hbet]Tf2N,
Sc, Y, La, Pr, Nd,Sm, Eu, Gd,
[HbetmIm]Tf2N, [HbetPy]Tf2N,
Tb, Dy, Ho, Er,Tm,Yb, Lu, U,
[EtHbetmMor]Tf2N,[HbetmMor]Tf2N, 9,10
Pb, Zn, Cd, Hg, Cu, Ag, Ni, Pd,
[HbetmPip]Tf2N, ([HbetmPyr]‐
Mn, Co Cr, Fe, Al and Si
Tf2N),
[Deep eutectic solvents]
Ti, V, Cr, Mn, Fe, Co, Al, Si,
Choline chloride/urea /ethylene 13
Ca, Cu, Ni, Pb and Zn
glycol/malonic acid
(Cyphos IL 101‐ saturated with
Ca, Ni, Mn, Co, Cu, and Zn 16
concentrated hydrochloric acid
Emim(X‐),
Ag,Cu, Ni and Zn 17
X‐= (Cl‐), (SCN−), (DCA−), (OAc−), (Tf2N−)

Ore and wastes processing with ionic liquids
Few studies have dealt also with processing ores and wastes with ILs for leaching
valuable metal oxides.
Electric arc furnace dust: Choline chloride with urea has been applied for processing
electric arc furnace dust [14], [15]. EAF dust consist mainly of metal oxides. Urea mixture
had been chosen due to the low solubility of iron oxide towards high solubility of ZnO
and CuO in it. Leaching of EAF with 2:1 molar ratio ChCl/urea conducted at 600C for
48h. ZnO, Cu2O and PbO can be selectively leached and subsequently recovered from
the liquid through cathodic electro‐deposition. A pilot plant with a capacity of 5kg dust
batches has already been built.
Brass ash: BmimHSO4 has been investigated for leaching of industrial brass ash in order
to recover copper and zinc. Parameters such as dissolution time, IL concentration and
oxidizing agents have been studied [18]. 99% dissolution efficiency has been achieved
for Zn and 24.82% for Cu after 4h of leaching at 700C with 50% v/v BmimHSO4 in water.
To increase the extraction ratio of Cu oxidizing agents were added to the leaching
solution. With H2O2 addition up to 82% of Cu was leached at 700C after 5h.
Sulphidic ores containing gold and silver: The first reported study of BmimHSO4
utilization in metal leaching concerned the extraction of Au and Ag from a complex
gold bearing ore [19]. The leach solution was prepared by dissolving thiourea in
BmimHSO4 in the presence of Fe2(SO4)3. Leaching was carried out at 500C resulting in
85% Au recovery and 60% Ag recovery. The remaining base metals copper, lead, zinc
and iron had low extraction yields. Furthermore recycling of BmimHSO4 has been

studied by stripping the leach liquor with activated carbon. Altering IL organic cation
while maintaining the HSO4‐ anion revealed that as the alkyl chain of the alkyl‐methyl‐
imidazolium hydrogen sulphate increases from butyl to octyl, the extraction of gold
decreases. This may be attributed to the corresponding rise of the viscosity of the IL
solution. Except from the cation alteration, different anions have been studied keeping
the butyl chain of the cation constant. BmimCl, BmimBF4, BmimN(CN)2 and
BmimCH3SO3 were studied for gold extraction. Significant recoveries occur for the
BmimCH3SO3 and BmimHSO4 for gold, silver and base metals. Moderate recoveries
occur for BmimBF4 and low recoveries for BmimN(CN)2 and BmimCl. These
observations suggest that the anion in the ionic liquid plays an active role during the
dissolution process. Furthermore leaching of gold and silver was studied in the
presence of (peroxomonosulfate HSO5‐) or Fe(III) oxidants, and complexing agents
such as thiourea, Cl‐, Br‐ and I‐ using BmimHSO4 and BmimCl as solvent medium [21].
Over 85% of gold was leached and significant amount of silver recovered in the
presence of HSO5−/ thiourea at 25–50 °C in both BmimHSO4 and BmimCl. HSO5−/
thiourea system had the same leaching efficiency as the Fe(III) thiourea system.
Recoveries of gold and silver were significantly higher than in an aqueous medium and
negligible recoveries of base metals obtained in ionic liquids.
Chalcopyrite: Chalcopyrite leaching was studied with 1:1 water: Bmim(BF4) volume
ratio in the presence of Fe(BF4)3 [22}. 90% of copper extraction was achieved after 8 h
at 100 0C. BmimHSO4 and its aqueous solution were used for leaching chalcopyrite [23]
concentrate in temperature range from 50 to 900C. Copper extraction increased from
52% to 88% as ionic liquid concentration in0creased from 10% to 100% at 700C for 24h.
1‐alkyl‐3‐methyl‐imidazolium ionic liquids were studied as solvent medium either in
neat or in aqueous mixtures, in the presence of thiourea and iron (III) as an oxidant [20].
Initially the leaching behavior of pure pyrite was studied in pure BmimHSO4, leading to
poor results. Chalcopyrite ore concentrate behaved differently during its leaching in
the same medium. 85% recovery of copper and less than 8% iron dissolution were
achieved after leaching of chalcopyrite at 600C with 20%w/v in water. The metal
extraction increases with temperature. Moreover increasing the concentration of
BmimHSO4 also lead to higher extraction yields for copper. In comparison to leaching
the same chalcopyrite ore concentrate with an aqueous 1M H2SO4 solution, the
abovementioned BmimHSO4 solution achieves 62% higher Cu leaching efficiency.
A summary of the research works that have studied the use of ionic liquids as solvents
in mineral processing can be seen in Table 2.


Table 5 Ionic liquids in mineral processing
Target Referenc
Ionic liquid/leaching agent Minerals
Metals e
[Deep eutectic solvents] Choline Electric arc furnace Zn, Cu,
14,15
chloride:urea dust (oxidic form) Pb
BmimHSO4/H2O2 or potassium Industrial brass ash
Zn, Cu 18
peroxymonosulfate (oxidic form)
sulphidic ores
BmimHSO4/thiourea and Fe2(SO4)3 19
containing Au,Ag
sulphidic ores
Bmim+X‐ , (X−= Cl−, CH3SO3−,
containing Au,Ag /
N(CN)2−, BF4‐ ,HSO4−)/thiourea and 20
pyrite, sphalerite
Fe2(SO4)3
chalcopyrite Au , Ag
BmimHSO4 or BmimCl
/peroxomonosulfate (HSO5‐) or
sulphidic ores
iron (III), and complexing agents 21
containing Au,Ag
thiourea, chloride, bromide and
iodine
Fe(BF4)3 in water:Bmim(BF4) at 1:1 22
chalcopyrite Cu
BmimHSO4 23

Leaching rare earth oxides using ionic liquids
This study will present preliminary results for utilizing ILs in leaching Rare Earth oxides
from both synthetic oxides and REE containing metallurgical wastes.
Synthetic rare earth oxides: Preliminary experiments investigated the dissolution
behavior of synthetic rare earth oxides in neat IL EmimHSO4. Ionic liquid EmimHSO4
with purity 95% was obtained from Sigma Aldrich and was used without any further
purification. The oxides studied were yttrium oxide, neodymium oxide, dysprosium
oxide and lanthanum oxide which were obtained from Alfa Aesar with 99.9% purity.

The experiments were conducted by introducing 1g of dried rare earth oxide powder
into 10ml of pure ionic liquid (10% w/v pulp density). Temperatures were varied from
90 to 2100 C, where EmimHSO4 remains stable and liquid. The end point of the leaching
experiments was determined by optical observation (fig. 2), at the point where a clear
solution was obtained. The hot solution was then filtered using fiber glass filters and
the resulting filter paper washed with deionized water. Atomic absorption
spectroscopy was used to measure the metal content in the pregnant solution and
thus to determine the total dissolution of the REE oxide. The time needed to achieve
the complete dissolution of different rare earth oxides under different temperature is
shown in figure 3.

Figure 6 Neodymium (left) and Lanthanum (right) oxide dissolving in neat EmimHSO4
at elevated temperatures. The solution becomes clear in less than 5 minutes at 210
0
C.

Figure 7 Time needed for total dissolution of 1 g of various rare earth oxides in 10 ml
of neat EmimHSO4 at different temperatures.
At elevated temperature rare earth oxides dissolve in very short times. At lower
temperature lanthanum and yttrium oxide didn’t dissolve and their solution formed a
highly viscous gel‐like phase. The trend here shows that dissolution is highly effected
with temperature, which can be related to the significant decrease in the viscosity of
EmimHSO4, shown in table 3, leading to better wettability of the solids and faster
dissolution rates.


Table 6 Viscosity of neat EmimHSO4 at different temperatures [24].
Temperature 20 0C 100 0C 150 0C 200 0C
Viscosity (mPa*s) 1354 55 13 4

Bauxite residue (red mud): Bauxite Residue, also known as red mud, is the major solid
waste generated during the primary alumina production with the Bayer process. For
each tone of alumina produced, almost 1 tone of residue is generated. This amounts
to approximately 120 million tons of bauxite residues that are produced worldwide
each year. Significant research effort takes place for the effective utilization of Bauxite
Residues [25‐27], while it’s potential as a source for REE production has also been
proposed [28‐30].
In the present study bauxite residues provided by ALSA S.A in Greece were used.
Chemical analysis of the sample was performed after total dissolution with a fusion
method and rare earth elements were measured using inductively coupled plasma
mass spectroscopy (ICP‐MS) while base metals with atomic absorption spectroscopy
(AAS). Calcium and sodium were determined with XRF method. The results are
presented in table 4.
Table 7 Chemical analysis of ALSA‘s Bauxite residues sample
Loss of
Red mud oxides Fe2O3 Al2O3 CaO TiO2 SiO2 Na2O REO
ignition
Chemical
43% 25% 9% 5% 5% 2% 0.14% 10%
analysis %wt.
Rare earths in Bauxite Residues (mg/Kg)
Ce 512 Pr 30
La 193 Sm 26
Sc 135 Dy 24
Nd 129 Gd 21
Y 114 Yb 13
Er 13

Scandium oxide is 0.02% wt. in the red mud analyzed, a concentration considered to
be exploitable [30]. Overall the REE concentration in the ALSA red mud is in the same
levels of concentrations reported in previous studies [31].
Minerals that have been identified by XRD in red mud are mainly hematite [Fe2O3],
gibbsite [Al2O3*3H2O], diaspore [Al2O3*H2O], cancrinite [Na6Ca2(Al6Si6O24)(CO3)2],
katoite [Ca3Al2(SiO4)(OH)8] and rutile [TiO2]. Measurements on grain size distribution
indicate an average grain size of 12 μm.
Direct leaching of red mud with EmimHSO4 at different pulp densities was contacted
at elevated temperatures. After 24h of leaching at 190oC, filtration at high
temperatures and washing the resulting filter cake with deionized water the filtrate

was analyzed in ICP‐MS and AAS. Experiments at 5% and 2.5% pulp density (w/v) were
carried out. The metal recoveries are shown in figure 4.

Figure 8 Recovery of rare earths and base metals from red mud after leaching with
neat EmimHSO4 at 190 0C for 24h.
At these leaching conditions almost total dissolution of iron and titanium was observed
while the residue consisted mainly of aluminum and silicon. EmimHSO4 showed no
particular leaching selectivity between rare earths and base metals. The average
recovery yield of rare earths increased with decreasing pulp density, reaching a 60‐
70% at the 2.5% pulp density. The resulting concentration of the pregnant leach
solution is given in table 5.

Table 5 Concentration of metals in EmimHSO4 after leaching bauxite residue (2.5%
w/v pulp density at 1900C, 24h).
Elements % recovery ~mg/L Elements % recovery ~mg/L
solvent solvent
La 100 5 Al 36 1205
Nd 70 2 Fe 100 7650
Y 61 2 Ca 39 627
Ce 70 9 Na 86 318
Sc 68 2 Ti 100 700

Kinetic experiments:
Keeping temperature constant at 1900C and pulp density at 2,5%(w/v), kinetic
experiments were conducted with neat EmimHSO4 for ½, 2,6 ,8, and 24h. The metal
recoveries are shown in figure 5.

Figure 9 Recoveries of selected metals versus time after leaching with EmimHSO4 at
190 0C for ½, 2, 6, 8, 24h.
After 2 hours retention time high recoveries for Fe and Ti were achieved. Iron is totally
recovered within the first 2 hours of leaching. Titanium is dissolved fast during the first
two hours, reaching a recovery of 80% and then its leaching rate is decelerated
necessitating 24 hours retention time in order to achieve 100% recovery. Scandium (as
a proxy for all REE) reaches also almost it’s maximum recovery in 2 hours and
prolonging the leaching time does not seem to have a significant effect as in the case
of Ti leaching. Aluminum follows a very slow dissolution process and its recovery
doesn’t exceed 35% after 24 hours
retention time.

Effect of altering the alkyl chain in
1‐alkyl‐3methyl‐imidazolium
hydrogen sulfate:

Altering the alkyl chain in
imidazolium cation by using the
ionic liquids 1‐butyl‐3methyl‐
imidazolium hydrogen sulfate
(BmimHSO4 >95% purity), and 1‐
methylimidazolium hydrogen
sulfate (HmimHSO4 >95% purity)
resulted to small differences on
metal recoveries from Bauxite
residues as seen in figure 6.
Separating the metals in three different groups, it can be concluded that base
Figure 10 Metal recoveries with different elements reach their maximum
cations in imidazolium hydrogen sulfate ionic recovery in EmimHSO4 while Na
liquid (1900C 24h, 2.5% pulp density). and Ca follow an opposite trend.

Rare earth elements such as La, Nd and Ce present their lowest recoveries in
BmimHSO4 whereas Sc recovery is practically unaffected by the alteration of cations.
Overall small changes occur with the variation of the IL cation.
Preliminary results with task specific ILs: Despite the high recoveries of rare earths
achieved with the imidazolium ‐ hydrogen sulfate ILs, the pregnant solutions produced
contain large amounts of dissolved Fe and Ti, making the extraction of the REE from
such solutions impractical. Thus future investigations should be focused to exploit
ionic liquids for selective dissolution between critical metals and iron as the latter is
the most abundant in red mud. As an indication for this potential, in figure 7 the
results of a preliminary leaching test of Bauxite Residue with the task specific IL betaine
bis(trifluoromethylsulfonyl)imide (HbetTf2N) is shown. The results indicate that the
specific IL favours the dissolution of Ca, Na, REE and other minor elements while the
iron of the Bauxite Residue remains practically undissolved.

Figure 7 Direct leaching of red mud with 50% v/v HbetTf2N in water at 60 C with
10%pulp density for 24h.
Conclusions
In the present study direct dissolution of synthetic rare earth oxides and red mud has
been made with ionic liquid EmimHSO4. Preliminary experiments with pure ionic liquid
EmimHSO4 and synthetic rare earth oxides prove that dissolution process is feasible
and occur at shorter times as temperature rises. Neat EmimHSO4 provides the acidity
needed for the metal oxides to dissolve and form metal complexes as other studies
with imidazolium ionic liquids proposed [17]. Bauxite residue also dissolve in
EmimHSO4. Small dissolution of silicon and aluminum occur and high recoveries of
rare earth elements such as yttrium, lanthanum, cerium, neodymium and scandium
have been obtained in the range of 60‐70%. Alteration of the cation in imidazolium
hydrogen sulfate ionic liquid has little effect in the metals recovery. Preliminary results
show that utilizing different ILs, such as the HbetTf2N may even achieve selective
leaching of the REEs. Thus from these studies and literature investigations Ionic liquids
seem to be promising alternative solvents for primary and secondary raw materials
processing.

Acknowledgments
“The research leading to these results has received funding from the European
agreement n°309373. This publication reflects only the author’s view, exempting the
Community from any liability”. Project web site: www.eurare.eu”
References
1. R. SETHI AND T. WELTON, in Ionic Liquids Industrial Applications to Green Chemistry, ed. R. D.
Rogers and K. R. Seddon, Washington, DC,2002.
2. WELTON, T., 2004. Ionic liquids in catalysis, Coord. Chem. Rev., 248, 2459‐2477.
3. SHELDON, R., 2001. Catalytic reactions in ionic liquids, Chem. Commun., 23, 2399‐2407.
4. VISSER, A.E., ROGERS, R.D., 2003. Room‐temperature ionic liquids: new solvents for f‐element
separations and associated solution chemistry, J. Solid State Chem., 171, 109‐113.
5. HUDDLESTON J.G.,. WILLAUER H.D, SWATLOWSKI R.P., VISSER A., ROBERTS D.W., 1998. Room
temperature ionic liquids as novel media for ‘clean’ liquid–liquid extraction, Chem. Commun.,
1765‐1766.
6. HIROYUKI OHNO Electrochemical Aspects of Ionic Liquids Department of Biotechnology, Tokyo
University of Agriculture and Technology.
7. S. DAI, Y. S. SHIN, L. M. TOTH AND C. E. BARNES, Comparative UV‐Vis Studies of Uranyl Chloride
Complex in Two Basic Ambient‐Temperature Melt Systems: The Observation of Spectral and
Thermodynamic Variations Induced via Hydrogen Bonding Inorg. Chem., 1997, 36, 4900.
8. BELL R C, CASTLEMAN A W, THORN D L. Vanadium oxide complexes in room‐temperature
chloroaluminate molten salts [J]. Inorg Chem, 1999, 38: 5709−5715
9. NOCKEMANN P, THIJS B, PITTOIS S, THOEN J, GLORIEUX C, VAN HECKE K, VAN MEERVELT L,
KIRCHNER B, BINNEMANS K. Task‐specific ionic liquid for solubilizing metal oxides [J]. J Phys Chem
B, 2006, 110: 20978−20992.
10. NOCKEMANN P, THIJS B, TATJANA N. Carboxyl‐functionalized task‐specific ionic liquids for
solubilizing metal oxides. [J]. Inorg Chem, 2008, 47(21): 9987−9999.
11. ABBOTT AP, CAPPER G, DAVIES DL ET AL (2003) Novel solvent properties of choline chloride/ urea
mixtures. Chem Commun 2003:70–77
12. ABBOTT AP, BOOTHBY D, CAPPER G ET AL (2004) Deep eutectic solvents formed between choline
chloride and carboxylic acids: versatile alternatives to ionic liquids. J Am Chem Soc 126:9142–9147
13. Abbott AP, Capper G, Davies DL et al (2006) Solubility of metal oxides in deep eutectic solvents
based on choline chloride. J Chem Eng 51:1280–1282
14. ABBOTT AP, CAPPER G, DAVIES DL ET AL (2006) Processing metal oxides using ionic liquids. Trans
Inst Min Met C 115:115–119
15. ABBOTT AP, FRISCH G, HARTLEY J ET AL (2011) Processing of metals and metal oxides using ionic
liquids. Green Chem 13:471–481
16. SIL WELLENS, TOM VANDER HOOGERSTRAETE , CLAUDIA MÖLLER , BEN THIJS , JAN LUYTEN, KOEN
BINNEMANS ., Dissolution of metal oxides in an acid‐saturated ionic liquid solution and
investigation of the back‐extraction behaviour to the aqueous phase. Hydrometallurgy 144–145
(2014) 27–33
17. SIL WELLENS, NEIL R. BROOKS, BEN THIJS, LUC VAN MEERVELT AND KOEN BINNEMANS., Carbene
formation upon reactive dissolution of metal oxides in imidazolium ionic liquids Dalton Trans.,
2014, 43, 3443
18. AYFER KILICARSLAN, MUHLIS NEZIHI SARIDEDE, SRECKO STOPIC, AND BERND FRIEDRICH, (2014)
Use of ionic liquid in leaching process of brass wastes for copper and zinc Int. J. Miner. Metall.
Mater., Vol. 21, No. 2,

19. WHITEHEAD J A, LAWRENCE G A, MCCLUSKEY A. Green leaching: recyclable and selective leaching
of gold‐bearing ore in an ionic liquid [J]. Green Chem, 2004, 6: 313−315.
20. WHITEHEAD J A, ZHANG J, PEREIRA N, McCLUSKEY A, LAWRANCE G A. Application of 1‐alkyl‐3‐
methyl‐ imidazolium ionic liquids in the oxidative leaching of sulphidic copper, gold and silver ores
[J]. Hydrometallurgy, 2007, 88: 109−120.J.A.
21. WHITEHEAD, J. ZHANG, A. MCCLUSKEY, G.A. LAWRANCE Comparative leaching of a sulfidic gold
ore in ionic liquid and aqueous acid with thiourea and halides using Fe(III) or HSO5− oxidant
Hydrometallurgy 98 (2009) 276–280
22. McCLUSKEY A, LAWRANCE G A, LEITCH S K, OWEN M P, HAMILTON I C. Ionic liquids industrial
applications for green chemistry [M]. ROGERS R D, SEDDON K R. Washington D C: American
Chemical Society, 2002, 818: 199−212.
23. DONG Tie‐guang, HUA Yi‐xin, ZHANG Qi‐bo, ZHOU Dan‐gui. Leaching of chalcopyrite with Brønsted
acidic ionic liquid at ambient pressure and low temperature [J]. Hydrometallurgy, 2009, 99: 33−38
24. N. POULIMENOU, I. GIANNOPOULOU, D. PANIAS. Preliminary investigation of ionic liquids
utilization in primary aluminum production, Proceedings of the International Conference on
Mining, Material and Metallurgical Engineering. Prague, Czech Republic, August 11‐12, 2014
25. C. KLAUBER, M. GRÄFE, G. POWER. Bauxite residue issues: II. Options for residue utilization.
Hydrometallurgy 108 (2011) 11–32
26. EFTHYMIOS BALOMENOS, IOANNA GIANNOPOULOU, DIMITRIOS GEROGIORGIS, DIMITRIOS
PANIAS, IOANNIS PASPALIARIS Resource‐Efficient and Economically Viable Pyrometallurgical
Processing of Industrial Ferrous By‐products. Waste Biomass Valor DOI 10.1007/s12649‐013‐9280‐
5
27. R. K. PARAMGURU P. C. RATH V. N. MISRA Trends in red mud utilization ^ a review Mineral
Processing & Extractive Metall. Rev., 26: 1_29, 2005
28. MARIA TH. OCHSENKU1 HN‐PETROPOULOU, KONSTANTINOS S. HATZILYBERIS,
29. LEONIDAS N. MENDRINOS AND CONSTANTINOS E. SALMAS. Pilot‐Plant Investigation of the
Leaching Process for the Recovery of Scandium from Red Mud. Ind. Eng. Chem. Res. 2002, 41,
5794‐5801
30. M. OCHSENKIIHN‐PETROPULU , TH. LYBEROPULU , K.M. OCHSENKIIHN, G. PARISSAKIS. Recovery of
lanthanides and yttrium from red mud by selective leaching. Analytica Chimica Acta 319 (1996)
249‐254
31. WEIWEI WANG, YOKO PRANOLO, CHU YONG CHENG. Metallurgical processes for scandium
recovery from various resources: A review. Hydrometallurgy 108 (2011) 100–108
32. OCHSENKUHN‐PETROPULU, M., LYBEROPULU, TH., PARISSAKIS, G., 1994. Direct determination of
lanthanides, yttrium and scandium in bauxites and red mud from alumina production. Anal. Chim.
Acta 296, 305–313


Session VI‐A: REE Processing‐ REE
Separation


FUNCTIONALISED MIL‐101(Cr) AS NOVEL EUROPIUM

ADSORBENT
Jeroen DE DECKER1, Jeriffa DE CLERCQ2, Els DE CANCK1, Pascal VAN DER
VOORT1.
1
Department of Inorganic and Physical Chemistry, Center for Ordered Materials,
Organometallics and Catalysis (COMOC), Ghent University, Krijgslaan 281‐S3, 9000 Ghent,
Belgium.
2
Department of Industrial Technology and Construction, Industrial Catalysis and Adsorption
Technology (INCAT), Ghent University, Valentin Vaerwyckweg 1, 9000 Ghent, Belgium.
Abstract
In this contribution we report the synthesis of a novel MOF‐based adsorbent for
europium. MIL‐101(Cr) is functionalised in a stepwise manner to create a material
bearing carbamoylmethylphosphine oxide (CMPO) ligands, which are selective towards
lanthanide complexation. Analysis showed that the adsorbent contains ~0.7 mmol of
CMPO ligands per gram. Preliminary adsorption capacity tests have been performed
resulting in a capacity of ~17 mg Eu/g adsorbent.
Introduction
The rare earth elements (REEs), a group of valuable metals, are gaining an ever‐
increasing importance to our modern society. Mainly consisting of the fifteen
lanthanides, these rare earth metals are found in a variety of high‐tech applications.
Especially in the field of clean energy, many of the lanthanides (e.g., neodymium,
dysprosium, europium,…) play a critical role in the design and operation of appliances
and machinery such as wind turbines, hybrid and electrical vehicles and lighting
technology1. With China currently controlling about 97% of the rare earth production,
only a fraction of the mined metals reaches the international market as export
product2. Opposite to this limited supply stands a rapidly increasing demand that is
unlikely to decelerate. Maintaining the rare earth supply will be crucial to a
technologically sustainable future. One important way to achieve this goal is to
anticipate on rare earth recycling. In this contribution we allude to one specific part in
the complex chain of recycling processes, namely the recovery of rare earth metals
from dilute aqueous streams.
Many metal recycling processes (and also primary mining processes, for that matter)
end up with acidic solutions in which various metals are present in their dissolved ionic
form. The current state‐of‐the‐art (rare earth) recovery techniques consist of
sequential liquid‐liquid extractions or precipitation processes to selectively separate
the valuable metals from any unwanted species3. While these techniques have proven
their worth in terms of selectivity, their recuperation yield is not 100 %. Typically, these
techniques generate aqueous waste streams containing several hundreds of
milligrams of rare earth metals per litre. In the past decade it was not necessary, nor

was it economically feasible to recover these small amounts of rare earths. But with
their increasing scarceness, the incentive has become far greater to expand industrial
recovery to these streams as well.
An efficient and cost‐effective way to remove and/or recover specific species from
dilute solutions is adsorption. Being a heterogeneous process, adsorption has the
advantage of easy separation and reuse. An adsorbent for rare earth metals, however,
must meet several requirements. First of all, the adsorbent needs to interact
selectively with the rare earth cations. This implies the embedding of specific
functionalities into the adsorbent support, which neglect unwanted cations (transition
metals, alkali metals,…) as much as possible, if not completely. Another important
property is adsorbent stability. Not only does the support material have to perfectly
survive the acidic aqueous conditions during the adsorption and regeneration, also the
linkage between support and the immobilised functionality requires the proper
stability to prevent leaching and to promote reuse of the adsorbent.
In this study, metal‐organic‐frameworks (MOFs) are introduced as a novel support in
the field of liquid adsorption. More specifically the chromium based MIL‐101 was
functionalised and applied as selective adsorbent for europium. MIL‐101(Cr) is a type
of mesoporous cage‐MOF, introduced in 2005 by Ferey et al.4. Its three‐dimensional
framework is made up of inorganic chromium‐oxide clusters connected via
terephthalate linkers, forming cages of super tetrahedra (Figure 1). The presence of
the aromatic linkers makes post‐functionalization of the MOF particularly easy. The
MIL‐101(Cr) is amongst the most stable mesoporous MOFs to date, possessing
excellent resistance to acidic aqueous media5,6. Furthermore, it is a highly porous MOF,
where the mesoporous cages permit the incorporation of spacious molecules.

Figure 1: (a) MIL‐101(Cr) super tetrahedron. (b) Small super cage. (c) Large super
cage. (d) MIL‐101 MTN zeotype structure.
All of the lanthanides form trivalent cations and as a general rule, these cations tend
to be fairly hard Lewis acids as a result of their high oxidation states. Therefore, when
designing a ligand field to selectively bind the lanthanide cations, it is advisable to
incorporate hard Lewis bases into the ligand structure (cfr. the HSAB concept). In
addition, an interesting phenomenon arises with these cations, in which the binding
of one class of ligands to the cation enhances the affinity of another class of ligands.
This results in a very selective comlexing behaviour7‐9. For the rare earth cations, the
most important types of synergistic ligands are the amide carbonyl group and

phosphine oxide based ligands10. A very efficient system based on the aforementioned
ligand combination is the so‐called CMPO ligand (carbamoylmethylphosphine oxide,
Figure 2). These ligands are known to be very selective towards lanthanide
complexation, while having little to no affinity for ordinary transition‐ or alkali
metals11.

Figure 2: Carbamoylmethylphosphine oxide (CMPO) ligand structure.
By heterogenising these kinds of ligands, one can create a selective adsorbent for rare
earth elements. In this on‐going research, we aim to develop a selective adsorbent
based on the MIL‐101 as solid support. This is performed by a step‐wise build‐up of
the ligand onto the MOF support. The materials are tested in the adsorption of
europium from aqueous solutions. In these preliminary tests, we investigate the
europium retention capacity of these developed adsorbents.
Experimental
Chemicals
Europium(III) oxide (99.9 %) was purchased from Alfa Aesar. Europium standard
solution (1,000 mg/l Eu3+ in dil. nitric acid) was purchased from VWR Chemicals. All
other chemicals were purchased from Sigma Aldrich.
Synthesis of MIL‐101(Cr)
MIL‐101(Cr) was synthesized based on an adapted recipe from Jiang et al.12. In a typical
experiment, terephthalic acid (0.665 g, 4 mmol) and Cr(NO3)3∙9H2O (1.608 g, 4 mmol)
were added to deionised water (20 ml). The resulting suspension was added to a
Teflon‐lined autoclave and subjected to 210 °C during 8 hours under autogenous
pressure (2 hours warm‐up). After cooling down to room temperature, the mixture
was filtered and the solid green product was collected and washed thoroughly with
dimethylformamide (DMF) and water in order to purify the material by removing any
unreacted reagents. Samples were vacuum dried at 120 °C for 24 hours.
Characterization
Nitrogen gas sorption experiments were conducted at 77 K with a Micromeritics Tristar
3000. Samples were vacuum dried at 120°C prior to analysis. The specific surface area
was calculated using the Langmuir method. DRIFTS‐spectra were measured on a
Nicolet 6700 FTIR spectrometer (Thermo‐Scientific) with an MCT detector. Analyses
were performed at 120 °C under vacuum. X‐ray fluorescence (XRF) spectroscopy
measurements were performed on an energy‐dispersive Rigaku NexCG spectrometer.
Elemental analysis (CHNS) was performed on a Flash 2000 (Thermo‐Scientific). Powder
X‐ray diffraction analyses were performed on an ARL X’tra diffractometer (Thermo‐
Scientific). Flame Atomic Emission spectroscopy (AES) was performed on a Varian

SpectrAA 220FS, using 4 bar acetylene and 0,7 bar nitrous oxide. Adsorption tests were
performed using a thermostatic shaking device (Infors HT Multitron standard, Analis,
25 °C, 220 rpm, 24 hours). In a typical adsorption test, 50 mg of adsorbent was added
to 50 ml of aqueous 100 ppm Eu3+‐solution (pH 4 with HCl) in glass vials (w/ plastic
lids). The tests were performed in triplicate. The europium concentration before and
after the tests was analysed by means of Flame‐AES.
Build‐up of the CMPO‐ligand onto the MIL‐101
Figure 3 gives an overview of the functionalization process of the MIL‐101(Cr). In the
first step, the material was chloromethylated to provide a proper anchoring point for
further functionalization. In this procedure13 the sample (1 g) was suspended in 70 mL
of nitromethane. Subsequently 1.9 g of AlCl3.6H2O and 0.4 g of methoxyacetyl chloride
were added. The mixture was then stirred for 5 hours at 100 °C, after which the solid
material was filtered off and washed thoroughly with nitromethane, water and
acetone. The sample was then vacuum dried at 120 °C for 24 hours. The material was
analysed for its chlorine content prior and post chloromethylation with XRF and was
found to contain ~5 mmol chloromethyl‐groups per gram material.
The chloromethylated MOF (1 g) was then reacted with 1,4‐diaminobutane (2.2 g) in
70 mL of DMF under inert atmosphere, by letting the mixture stir for 24 hours at 80 °C
(Step 2). The material was filtered off, thoroughly washed with DMF and acetone, and
vacuum dried at 120°C prior to elemental analysis. The sample contained approx. 5.5
wt.% additional nitrogen after this functionalization step, corresponding to 2 mmol
diaminobutane per gram. Current research is focussing on increasing this loading.
The third step involves the coupling of the carboxylic end‐groups of
diethylphosphonoacetic acid with the primary amines of the functionalised MOF,
forming the amide part of the CMPO‐ligand. A total of 500 mg of amine‐functionalised
MOF is suspended in 25 mL of anhydrous DMF. In a separate round‐bottom flask, 0.6
g of diethylphosphonoacetic acid (3 mmol) is dissolved in anhydrous DMF at room
temperature (inert atmosphere), after which an equimolar amount of
carbonyldiimidazole (CDI, 0.48 g) is added and the mixture is moderately stirred10.

Figure 3: Stepwise build‐up of the CMPO‐ligand on MIL‐101(Cr). Initial MOF: MIL‐101,
chloromethylated: ‐Cl, aminated: ‐NH2, full ligand on MOF: MIL‐101‐LIG.

CDI is a known coupling agent for peptide synthesis, requiring water‐free conditions14.
The reaction of CDI with the carboxylic end‐group can be visually monitored because
the mechanism leads to formation of CO2. The mixture is stirred until the evolution of
CO2 ceases (about 15 minutes). Afterwards both mixtures are combined and stirred
for another 12 hours at room temperature. After filtration and washing with DMF,
water, and acetone the sample is vacuum dried at 120 °C.
The following characterization analyses were performed on each stage of the
(un)functionalised material: (1)nitrogen sorption measurements to determine the
specific surface area of the material, which indicates the anchoring of species (weight‐
based result) and indirectly confirms the (in)stability of the material in the respective
reaction conditions (shape of the isotherm); (2)DRIFTS‐analyses to confirm whether
the proper groups are anchored onto the material; (3)XRF‐measurements to estimate
the loading of the final ligand (via phosphorous analysis); (4) XRD measurements to
further confirm the stability of the material in each step. The results are depicted in
figures 4 to 6.

Figure 4: Nitrogen adsorption isotherm of the unfunctionalised MIL‐101 (•),
aminated MIL‐101 (‐‐) and finalised ligand MIL‐101 (―).

Figure 5: DRIFTS spectra of each step in the functionalization. Inset: zoom on the
aliphatic stretch region.


Figure 6: Powder XRD patterns of the (un)functionalised materials.

From the preservation of the XRD pattern and shape of nitrogen adsorption isotherm,
it can be concluded that the MIL‐101 perfectly maintains its structure throughout the
different functionalization steps. Starting out at SLangmuir: 3000 m²/g, the MIL‐101
gradually decreases in specific surface area upon each functionalization step due to
anchored species that fill up the cages and increase the weight of the material. In the
DRIFTS spectra aliphatic CH stretches appear (2850 – 3000 cm‐1, inset), confirming the
presence of diaminobutane in the aminated sample and the final adsorbent. An
indication of amine groups is visible near 1550 and 1650 cm‐1, corresponding to
primary‐NH2 scissoring. These do not appear as isolated peaks but partially overlap
with vibrations originating from the MIL‐101 structure. In the spectrum of the finalised
adsorbent, vibrations are visible at 1040 cm‐1 and in the region between 1200 and 1260
cm‐1. These indicate the P‐OR ester and the phosphine oxides or phosphonates
vibrations, respectively. Presence of other characteristic peaks could not be proven
unambiguously due to overlap with the MIL‐101 structural vibrations and possible low
loading.
XRF‐analyses showed a phosphorous‐loading of 0.7 mmol/g, which indicates that not
all primary amines have been functionalised with the phosphonoacetic acid group.
Current research focusses on increasing the yield of this solid‐phase peptidization.
The CMPO functionalised MIL‐101 was tested in the adsorption of europium. The
unfunctionalised MOF as well as the aminated one were tested in the same conditions
and were found to be inactive for the adsorption of Eu3+. The results of the final
adsorbent are presented in table 1.

Table 1: Adsorption capacity of the CMPO functionalized MOF.
Material Functional group qe (mg/g) qe (mmol Eu/g)
MIL‐101‐Full Ligand ‐CMPO 17 (±2) 0.11 (±0.01)


The amount of europium adsorbed is calculated using the following expression:
.
were qe is the equilibrium adsorption capacity of europium (mg/g), C0 and Ce the initial
and equilibrium Eu3+ concentration (mg/L) respectively, V the solution volume (mL)
and m the adsorbent mass (mg). A total of 17 mg Eu3+/g (0.11 mmol/g) is adsorbed
with 0.7 mmol/g of CMPO ligands. This corresponds to six or seven ligands for every
Eu3+. The exact manner of coordination or the number of ligands coordinating to one
europium, is not yet determined.
Conclusion
The MIL‐101(Cr) is successfully functionalised and applied for the first time as a rare
earth metal adsorbent in acidic aqueous solution. The solid is post‐modified via three
subsequent steps which resulted in a material with 0.70 mmol per gram of
carbamoylmethylphosphine oxide ligands. The CMPO‐functionalised MOF was used in
preliminary experiments as a europium adsorbent and was able to adsorb ~17 mg
Eu3+/g. More in‐depth europium adsorption experiments will be performed including
regeneration of the CMPO functionalised MOF adsorbent.

References
1. Alonso, E., Sherman, A. M., Wallington, T. J., Everson, M. P., Field, F. R., Roth, R.,Kirchain, R. E.
Environ Sci Technol 2012, (46), 4684.
2. Jordens, A., Cheng, Y. P.,Waters, K. E. Miner Eng 2013, (41), 97.
3. Kara, H. C., A; Crichton, T; Willis, P; Morley, N Lanthanide Resources and Alternatives, Oakdene
Hollins, (2010).
4. Ferey, G. Science 2005, (310), 1119.
5. Ferreira, R. B., Scheetz, P. M.,Formiga, A. L. B. Rsc Adv 2013, (3), 10181.
6. Bernt, S., Guillerm, V., Serre, C.,Stock, N. Chem Commun 2011, (47), 2838.
7. Horwitz, E. P., Chiarizia, R.,Dietz, M. L. React Funct Polym 1997, (33), 25.
8. Reddy, M. L. P., Bharathi, J. R. B., Peter, S.,Ramamohan, T. R. Talanta 1999, (50), 79.
9. Masuda, Y., Zhang, Y. W., Yan, C. H.,Li, B. G. Talanta 1998, (46), 203.
10. Fryxell, G. E., Wu, H., Lin, Y. H., Shaw, W. J., Birnbaum, J. C., Linehan, J. C., Nie, Z. M., Kemner,
K.,Kelly, S. J Mater Chem 2004, (14), 3356.
11. Lambert, B. J., V; Shivanyuk, A; Matthews, S; Tunayar, A; Baaden, M; Wipff, G; Bohmer, V;
Desreux, J.F Inorg Chem 2000, (39), 2033
12. Jiang, D. M., Burrows, A. D.,Edler, K. J. Crystengcomm 2011, (13), 6916.
13. Goesten, M. G., Gupta, K. B. S. S., Ramos‐Fernandez, E. V., Khajavi, H., Gascon, J.,Kapteijn, F.
Crystengcomm 2012, (14), 4109.
14. Podlech, J. Synthesis of Peptides and Peptidomimetics (Methods of Organic Chemistry, Houben‐
Weyl); 4 ed. Stuttgart, (2001); Vol. E22a.


MAGNETIC METAL OXIDE BASED ADSORBENT FOR

EXTRACTION AND SEPARATION OF REE
Gulaim Seisenbaeva1, Peter Svedlindh2,Vadim Kessler 1
1 Department of Chemistry and Biotechnology, Biocenter, Swedish University of Agricultural
Sciences, Box 7015, Almas allé 5, Ultuna, 750 07 Uppsala, Sweden
2Solid State Physics, Department of Engineering, Uppsala University, Box 534, SE‐751 21
Uppsala, Sweden
Gulaim.Seisenbaeva@slu.se; Vadim.Kessler@slu.se
Abstract
Mesoporous transition metal oxides with high surface area (over 200 m2/g) can
efficiently be produced by a thermohydrolysis approach from solid metal alkoxides
precursors. There materials can be made magnetic via insertion of iron oxide
nanoparticles. Protection of materials against acid leaching is achieved by deposition
of a thin layer of TiO2. The produced materials are easily modified by grafting of
phosphonate ligands leading to creation of hybrid adsorbents conserving the
morphology and surface characteristics of the applied oxide matrices. Aminopropyl
phosphonic and imino(dimethylphosphonic) acids were used to achieve a proof‐of‐
concept. The obtained adsorbents revealed good adsorption capacity towards REE
cations (0.18, 0.21, 0.23 and 0.24 mmol/g for Y, Nd, La and Dy respectively). The
obtained materials can be used for water purification and for recycling of REE.

Introduction
Mesoporous transition metal oxides constitute an attractive class of materials due to
a broad spectrum of applications, ranging from heterogeneous catalysts [1], as
matrices for catalyst deposition [2], electrode materials for metal ion batteries [3],
dye‐sensitized solar cells [4], and functional adsorbents for removal of radioactive [5]
and heavy metal pollutants [6] from ground and waste waters. To be attractive for all
these potential uses, the metal oxides should possess a high active surface area,
considerable chemical stability to different types of leaching and dissolution, and in
many cases be essentially pure from residual organics that may deteriorate the surface
characteristics required for alkali metal ion insertion and heavy metal adsorption.
Most often such materials are produced using either large organic or macromolecular
templates [7] and therefore need to undergo long term solvothermal treatment to
improve their crystallinity. Alternatively, these materials can be synthesized in the
absence of a template under rather extreme solvothermal conditions but still require
calcination for removal of the organic residues [8].

Sol‐gel synthesis using metal alkoxides as precursors is potentially a very promising
route for production of mesoporous oxides. Depending on the nature of the alkoxide
groups and introduced hetero ligands, it is possible to influence the size of the particles
and their crystallinity [9]. The latter is improved by use of increased temperature and
with more efficient removal of the ligands. Thus highly crystalline single nanoparticles
of titania have been produced by immersion of solid titanium alkoxide modified with
highly hydrophilic amino acid ligands into boiling water [10]. In a recent study we have
shown that by using solid metal alkoxides derived from highly volatile alcohols (i.e.
MeOH and EtOH), with reaction in boiling water, it is possible to produce materials
with high crystallinity and open mesoporosity, thereby exploiting the evaporation of
ligands as a tool to prevent collapse of the pores [11, 12].
A large surface charge with a dominating total negative charge in a broad pH interval
[13], together with enhanced chemical affinity to highly charged anions makes these
materials efficient as adsorbents in immobilization of a broad spectrum of solute
species. The adsorbates include heavy metal cations such as uranium and thorium [14],
various toxic anions, for example, arsenate [AsO43] [15] and phosphate or
phosphonate [16] anchoring moieties. The adsorption processes rely on the molecular
details of the interactions on the surfaces, and nanomaterials with their very large
surface‐to‐volume ratio are especially attractive adsorbents. In the case of TiO2, very
large surface‐to‐volume ratios are available in the form of nanotubes or nanorods that
can be produced in high yields by acidification and subsequent exfoliation of layered
alkali metal titanates [17] or simply by electrospinning [18]. An interesting alternative
for a facile and cost‐efficient production of porous titania (anatase) nanorods is the
so‐called rapid hydrothermal synthesis. It makes use of nanorods of a metal‐organic
precursor titanium methoxide, Ti(OCH3)4 as a starting material [11]. The inexpensive
and efficient surfactant‐free synthesis of nano titania make related materials
attractive alternatives to mesoporous silica adsorbents.
Titania‐based adsorbents can be surface modified for potential uses in various
applications. Its surfaces can be modified by carboxylates, phosphates and
phosphonates moieties carrying a wide range of organic functionalities. In addition,
such modifications allow TiO2 nanoparticles to be stabilized in both aqueous [18] and
hydrocarbon [19] media. They have also been converted into hybrid organic‐inorganic
materials with principally changed hydrophilicity/hydrophobicity and functionality,
applied, for example, in the sorption of CO2 from gas mixtures [20, 21], in separation
of hydrophobic dyes from diluted solutions [22] and in extraction of heavy metal
cations [23]. Among those a special attention has recently been focused on Rare Earth
Elements (REE). REE, and especially neodymium (Nd) and dysprosium (Dy) are required
for production of magnetic materials used both in electronics [24] and novel energy
production technologies utilizing renewable energy sources [25], and are available as
components in ores abundant in the Northern Europe, Iceland, Greenland, USA and
Canada. Their production today is, however, almost exclusively concentrated to China,

where they are present in more easily exploitable forms and where the environmental
hazards associated with their extraction and separation (in the first hand, production
of huge volumes of acidic wastes and the use of huge volumes of organic solvents) are
not treated with the same care. Also the adsorption of actinide elements accumulating
in the nuclear waste and displaying solution behavior analogous to REE became a topic
of increased attention in connection with recent disaster at the Fukushima nuclear
plant [26]. An effort has been reported recently in creation of hybrid mesoporous silica
based adsorbents bearing alkyl phosphonate function for selective adsorption of REE
[27].
In the present work we discuss results from studies of hybrid adsorbents produced
from mesoporous metal oxides, in the first hand, titania nanorods. They were derived
by a precursor‐driven approach [11] via straightforward immobilization of
aminoethylphosphonic acid AEPA (NH2(CH2)2PO3H2) and imino‐bisphosphonic acid
IMPA (NH(CH2PO3H2)2) on the surface of nanoparticles. As shown, this approach
permits a denser grafting on the titania surface than other approaches [16], and thus
these sorbents have a potentially much higher adsorption capacity.

Synthesis of the oxide material was carried out using a unified procedure by rapid
immersion of the precursor powder into boiling water with subsequent refluxing for
30 min. All oxide materials produced are mesoporous and demonstrate nitrogen
sorption isotherms of type IV according to IUPAC classification (see Fig. 1).
16
14
12
Volume of gas sorbed (a.u.)
10
8
a
6 b
4 c
2 d
0
e
0.0 0.2 0.4 0.6 0.8 1.0
Relative Pressure p/p0

Figure 1. Nitrogen sorption isotherms for mesoporous oxides obtained by
thermohydrolytic treatment of Ti(OMe)4 (a), Zr(OiPr)4(iPrOH) (b), Nb(OMe)5 (c),
Zr(OEt)4 (d) and Ta(OMe)5 (e).
They possess impressively high active surface areas and pore volume fractions (see
Tab. 1) with a sharp pore size distribution and well defined maximum in the meso
scale.

Table1 Textural characteristics of metal oxides produced by thermohydrolysis
Metal oxide[a] Precursor Active surface Mean pore Pore volume,
area, m2/g[b] size, nm[c] cm3/g[d]
TiO2 Ti(OMe)4 278 5.1 0.43
ZrO2 Zr(OEt)4 257 3.2 0.20
ZrO2 Zr(OiPr)4(iPrOH) 321 5.1 0.41
Nb2O5 Nb(OMe)5 410 2.9 0.29
Ta2O5 Ta(OMe)5 98 3.8 0.09
[a] Data after initial drying at 120C for 3h. [b] Calculated using BET method. [c] and
[d] calculated using BJH method on the desorption branch and at p/p0=0,99
respectively

Magnetic properties were introduced via insertion of magnetite nanoparticles. The
Fe3O4 was obtained in nano form by co‐precipitation and used in aqueous dispersion
as solution for thermohydrolysis of the alkoxides precursors. The freshly produced
material was covered by thin layer of TiO2 using CaptiGelTM aqueous titania to insure
stability in acidic media. The produced material displayed strong magnetic properties
(Fig. 2).

Figure 2. Magnetization versus field at T=10K (uncoloured) and T = 300 K (coloured)
for titania‐iron oxide adsorbent
.
Grafting of functional phosphonate ligands was achieved at room temperature by
stirring the matrices in a ligand solution in 48 h. The ceramic yields for the adsorbents
were for both ligands ~ 70%, and the inorganic‐organic hybrid materials were
separated by straightforward sedimentation overnight. The deviation from a
quantitative yield was ascribed to a partial stabilization in solution of small and non‐
precipitating nanoparticle aggregates, splitting from the body of nanorods on
mechanical treatment (stirring in 72 h), by the adsorbed amino phosphonic ligands.

The adsorption of amino phosphonate ligands has already been reported to stabilize
dispersions of small non‐aggregated metal oxide nanoparticles [16]. The effect was
more pronounced for AEPA than for IMPA, as it has been observed previously for other
oxide systems [16].
An insight into surface coordination of the amino phosphonic acids onto TiO2 is
provided by their FTIR spectra. As reported in literature [28], AEPA can form molecular
zwitter‐ionic structure in the spectrum of which the vibrational bands relating as to
the bending NH3+ modes of the initial acid are located at 1482 сm‐1 [δsNH3] and 1643
сm‐1 [δasNH3]. During modification of the surface of titanium dioxide by AEPA the
phosphonate groups are involved into complex formation and possibly deprotonаtion,
leading for the TiO2‐AEPA sample to the shift of the δsNH3 adsorption band to 1510
cm‐1. The FTIR spectra demonstrate thus that the ligands are definitely located on the
surface after grafting and that at least AEPA is present there in zwitter‐ionic form,
indicating possibility that this acid can be connected to the TiO2 surface with release
of only one proton.
Sorption kinetics of Dy3+ cations from aqueous solutions were investigated for both
TiO2‐АЕPA and TiO2‐IMPA (see Fig. 3)

Figure 3. Kinetics of Dy3+ ions sorption by samples TiO2‐IMPA and TiO2‐AEPA

The dynamic equilibrium in REE adsorption was achieved for both TiO2‐АЕPA and TiO2‐
IMPA within 72 h. Such considerable delay in reaching the equilibrium is usually a
feature of mass transport limitations within the materials. These limitations appear
logical due to the ink‐bottle shaped pores that were indicated in the N2 desorption
characteristics in these adsorbents. The Dy3+ adsorption by TiO2‐IMPA was faster than
on TiO2‐АЕPA, and was completed to 95% within 24 h.
The sorption isotherms obtained for Y3+, La3+, Nd3+ and Dy3+ cations (see Fig. 4) belong
to the Langmuir type indicating similar sorption mechanisms in all cases. All isotherms
had slight kinks in the equilibrium concentration region of 0.05‐0.2 mmol/l. These
kinks could speculatively indicate potential conformational changes for the ligands in
the surface layer on TiO2.

Figure 4. Sorption isotherms of Nd3+ and Dy3+ on TiO2, functionalized with
aminophosphonic acids.

The obtained adsorbents show thus relatively modest values of static sorption capacity
(SSС) and compared with the best silica‐based mesoporous nanoadsorbents
constituting the state‐of‐the‐art in the field [29, 30]. However, our results are
impressively better than those from analogous experiments with hybrid organic‐
inorganic adsorbents based on ZrTi‐0.33 mesoporous adsorbent functionalized with,
in particular, amino tris(methylphosphonic) acid: the TiO2‐AEPA shows about 10 times
and TiO2‐IMPA up to 50 times higher adsorption capacity [27]. Much lower price of
mesoporous titania prepared by the utilized approach makes the produced material
potentially interesting for remediation applications. The better efficiency in ligand
involvement in adsorption revealed by the ТіО2‐ІМРА material provides it with 3‐4
times higher capacity compared with ТіО2‐АЕРА.
Conclusions
A simple and efficient methodology for preparation of hybrid magnetic adsorbents
based through grafting of amino phosphonate ligands on the surface of mesoporous
titanium dioxide was developed. The ligand loading varies for different ligands in the
range 0.17‐0.21 mmol/g. The pore volume and active surface area undergo minor
decrease as a result of this transformation.
By comparing the data from several experimental methods it appeared as the ligands
are bound to the ТіО2 surface through loss of one proton per ligand. The adsorption of
REE, Y3+, La3+, Nd3+ and Dy3+, cations from weakly acidic solutions was associated with
a considerable decrease in the pH, which indicating an ion exchange mechanism for
this process. The uptake of REE varied in the range 0.18‐0.24 mmol/g corresponding
to M:L = 1:1 surface complexes being formed. This value is comparable with capacity
of organic ion‐exchange resins.
The produced adsorbents are of interest in separation of REE for water purification
and for recycling processes.


References
1. M. Misono, “Perovskite and Heteropoly Catalysts” in Heterogeneous Catalysis of Mixed Oxides:
Elsevir, Oxford, UK, pp. 25‐35, 157‐159, 2013.
2. a) S. Pany, B. Naik, S. Martha, K. Parida, ACS Appl. Mat. Interfaces, 6 839‐846 (2014); b) J. L. Gole,
S. Ozdemir, Chemphyschem, 11 2573‐2581 (2010); c) S. J. Han, T. K. Yu, J. Park, B. Koo, J. Joo, T.
Hyeon, S. Hong, J. Im, J. Phys. Chem., 108 8091‐8095 (2004); d) D. A. Svintsitskiy, L. S. Kibis, A. I.
Stadnichenko, V. I. Zaikovskii, S. V. Koshcheev, A. I. Boronin, Kinetics Catal., 54 497‐504 (2013); e)
J. C. Rooke, T. Barakat, S. Siffert, B. L. Su, Catal. Today, 192 183‐188 (2012).
3. a) S. G. Lee, H. Deng, J. Hu, L. H. Zhou, H. L. Liu, Int. J. Electrochem. Sci., 8 2204‐2219 (2013); b) N.
A. Kyeremateng, C. Lebouin, P. Knauth, T. Djenizian, Electrochim. Acta, 88 814‐820 (2013); c) X. Lu,
Z. L Jian, Z. Fang, L. Gu, Y. S. Hu, W. Chen, Z. X. Wang, L. Q. Chen, Energy Env, Sci., 4 2638‐2644
(2011); d) L. Hernan, J. Morales, L. Sanches, J. L. Tirado, Electrochim. Acta, 39 2665‐2671 (1994); e)
L. Kavan, A. Attia, F. Lenzmann, S. H. Elder, M. Gratzel, J. Electrochem. Soc., 147 2897‐2902 (2000).
4. a) B. X. Lei, L. L. Zeng, P. Zhang, H. K. Qiao, Z. F. Sun, J. Power Sources, 253 269‐275 (2014); b) C. C.
Chen, W. D. Jheng, C. K. Lin, Ceramics Int., 40 3221‐3226 (2014); c) K. H. Park, E. M. Jin, H. B. Gu, S.
D. Yoon, E. M. Han, J. J. Yun, Appl Phys. Lett., 97 Art. Nr: 023302 (2010); d) B. Y. Jeong, E. H. Jung,
Met. Mater. Int., 19 617‐622 (2013); e) M. Stefik, F. J. Heiligtag, M. Niederberger, M. Gratzel, ACS
Nano, 7 8981‐8989 (2013); f) J. Kim, J. Kim, J. Nanosci. Nanotechn., 11 7335‐7338 (2011); g) S. Yun,
L. Wang, W. Guo, T. L. Ma, Electrochem. Com., 24 69‐73 (2012); h) M. C. Kao, , H. Z. Chen, S. L.
Young, Japan J. Appl. Phys., 52 Art. Nr. 01AD04 (2013).
5. a) D. J. Yang, Z. F. Zheng, H. Y. Zhu, H. W. Liu, X. P. Gao, Adv. Mater, 20 2777‐2781 (2008); b) D. J.
Yang , Z. F. Zheng , H. Y. Zhu , H. W. Liu , X. P. Gao , Adv. Mater., 20 2777‐2781 (2008).
6. a) J. M. Herrmann, Catal. Today, 53 115‐129 (1998); b) R. Upadhyay, S. Kant; R. Navneet, S. Susanta,
RSC Adv., 4 3823‐3851 (2012).
7. D.H. Chen, L. Cao, T.L. Hanley, R.A. Caruso, Adv. Funct. Mater., 22 1966‐1971 (2012).
8. A. Pucci, M.G. Willinger, F. Liu, X. B. Zeng, V. Rebuttini, G. Clavel, X. Bai, G. Ungar, N. Pinna, ACS
Nano, 6 4382 (2012).
9. a) R. Pazik, R. Tekoriute, S. Håkansson, R. Wiglusz, W. Strek, G. A. Siesenbaeva, Y. K. Gun’ko, V. G.
Kessler, Chem. Eur. J., 15 6820‐6826 (2009); b) G. A. Seisenbaeva, V. G. Kessler, Nanoscale, DOI:
10.1039/C3NR06336D (2014).
10. a) G. Goutailler, C. Guillard, S. Daniele, L.G. Hubert‐Pfalzgraf, J. Mater. Chem., 13 342‐346 (2003);
b) R. Rahal, S. Daniele, L.G. Hubert‐Pfalzgraf, V. Guyot‐Ferréol, J. F. Tranchant, Eur. J. Inorg. Chem.,
980‐987 (2008).
11. G.A. Seisenbaeva, J.M. Nedelec, G. Daniel., V.G. Kessler, J. Mater. Chem., 22 20374‐20380 (2012).
12. G. A. Seisenbaeva, G. Daniel, V. G. Kessler,J‐M. Nedelec, Chem Europ J. (2014) accepted.
13. T. Preocanin and N. Kallay, “Point of Zero Charge and Surface Charge Density of TiO2 in Aqueous
Electrolyte Solution as Obtained by Potentiometric Mass Titration”, Croat. Chem. Acta, 79 95‐106
(2006).
14. M. Bonato, K.V. Ragnarsdottir, G.C. Allen, “Removal of Uranium(VI), Lead(II) at the Surface of TiO2
Nanotubes Studied by X‐Ray Photoelectron Spectroscopy”, Water Air Soil Pollut. 223 3845‐3857
(2012).
15. Y.C. Tang, C.N. Wu, X.H. Huang, H.P. Zhang, H.Q. Yu, X. Li, Y. Peng, “Arsenic(III) removal from low‐
arsenic water by adsorption with amorphous mesoporous TiO2”, Desalin. Water Treatm. 49 359‐
367 (2012).
16. R. Pazik, R. Andersson, L. Kepinski, V.G. Kessler, J.M. Nedelec, G.A. Seisenbaeva, “Surface
Functionalization of the Metal Oxide Nanoparticles with Biologically Active Molecules Containing
Phosphonate Moieties. Case Study of BaTiO3”, J. Phys. Chem. C 115 9850‐9860 (2011).

17. Y.V. Kolen'ko, K.A. Kovnir, A.I. Gavrilov, A.V. Garshev, J. Frantti, O.I. Lebedev, B.R. Churagulov, G.
Van Tendeloo, M. Yoshimura, ”Hydrothermal Synthesis and Characterization of Nanorods of
Various Titanates and Titanium Dioxide” , J. Phys. Chem. B 110 4030‐4038 (2006).
18. J. Choi, A. Ide, Y.B. Truong, I.L. Kyratzis, R.A. Caruso, “High surface area mesoporous titanium–
zirconium oxide nanofibrous web: a heavy metal ion adsorbent”, J. Mater. Chem. A 1 5847‐5853
(2013).
19. T. Arita, K. Moriya, T. Yoshimura, K. Minami, T. Naka, T. Adschiri, “Dispersion of Phosphonic Acids
Surface‐Modified Titania Nanocrystals in Various Organic Solvents”, Ind. Eng. Chem. Res. 49 9815–
9821 (2010).
20. F.J. Song, Y.X. Zhao, H.L. Ding, Y. Cao, J. Ding, Y.F. Bu, Q. Zhong, “Capture of carbon dioxide by
amine‐loaded as synthesized TiO2 nanotubes”, Environ. Tech. 34(11) 1405–1410 (2013).
21. F.J. Song, Y.X. Zhao, H.L. Ding, Y. Cao, J. Ding, Y.F. Bu, Q. Zhong, “Capture of carbon dioxide from
flue gases by amine‐functionalized TiO2 nanotubes”, Appl. Surf. Sci. 268 124‐128 (2013).
22. J.J. Li, J.T. Feng, W. Yan, “Synthesis of polypyrrole‐modified TiO2 composite adsorbent and its
adsorption performance on acid Red G”, J. Appl. Polymer Sci. 3231‐3239 (2013).
23. S. Abbasizadeh, A.R. Keshtkar, M.A. Mousavian, “Preparation of a novel electrospun polyvinyl
alcohol/titanium oxide nanofiber adsorbent modified with mercapto groups for uranium(VI) and
thorium(IV) removal from aqueous solution”, Chem. Eng. J. 220 161‐171 (2013).
24. X.Q. Cao, “Application of Rare Earths in Thermal Barrier Coating Materials”, J. Mater. Sci. Tech.
2315‐35 (2007).
25. A.J. Hurd, R.L. Kelley, R.G. Eggert, M.H. Lee, “Energy‐critical elements for sustainable
development”, MRS Bull. 37 405‐410 (2012).
26. S. Schneider, C. Walther, S. Bister, V. Schauer, M. Christl, H.A. Synal, K. Shozugawa, G. Steinhauser,
“Plutonium release from Fukushima Daiichi fosters the need for more detailed investigations”, Sci.
Rep., 3 2988 (2013).
27. I.V. Melnyk, V.P. Goncharyk, N.V. Stolyarchuk, L.I. Kozhara, A.S. Lunochkina, B. Alonso, Y.L. Zub,
“Dy3+ sorption from water solutions by mesoporous silicas functionnalized by phosphonic acid
groups”, J. Porous. Mater. 19 579‐585 (2012).
28. M.L. Roldán, A.E. Ledesma, A.B. Raschi, M.V. Castillo, E.Romano, S.A. Brandán, “A new
experimental and theoretical investigation on the structures of aminoethyl phosphonic acid in
aqueous medium based on the vibrational spectra and DFT calculations”, J. Molecul. Struct. 1041
73‐81 (2013).
29. M.R. Awual, T. Kobayashi, Y. Miyazaki, R. Motokawa, H. Shiwaku, S. Suzuki, Y. Okamoto, T. Yaita,
“Selective lanthanide sorption and mechanism using novel hybrid Lewis base (N‐methyl‐N‐phenyl‐
1,10‐phenanthroline‐2‐carboxamide) ligand modified adsorbent”, J. Hazardous Mater. 252‐253
313‐320 (2013).
30. M.R. Awual, T. Kobayashi, H. Shiwaku, Y. Miyazaki, R. Motokawa, S. Suzuki, Y. Okamoto, T. Yaita,
“Evaluation of lanthanide sorption and their coordination mechanism by EXAFS measurement
using novel hybrid adsorbent”, Chem. Eng. J. 225 558–566 (2013).
31. C.S. Griffith, M. De Los Reyes, N. Scales, J.V. Hanna, V. Luca, “Hybrid Inorganic−Organic Adsorbents
Part 1: Synthesis and Characterization of Mesoporous Zirconium Titanate Frameworks Containing
Coordinating Organic Functionalities”, ACS Appl. Mater. Interfaces, 2 3436‐3446 (2010).


ORGANIC‐LIGAND GRAFTED MAGNETIC

NANOADSORBENTS FOR EXTRACTION AND SEPARATION
OF REE.
Elizabeth Polido*, Seda Demirel Topel, Vadim G. Kessler, Gulaim A.
Seisenbaeva.
Department of Chemistry and Biotechnology, Biocenter, SLU, Box 7015, 75007 Uppsala,
Sweden.
* elizabeth.polido.legaria@slu.se

Abstract
Magnetic hybrid nanoadsorbents functionalized with different organosilane derivates
were synthesized and their affinity to Rare Earth Elements (REE) cations in solution was
tested. Magnetite Fe3O4 nanoparticles (MNP) were chosen as cores because of their
good magnetic properties and stability. The MNP were synthesized and coated with
SiO2 resulting in highly stable core‐shell Fe3O4@SiO2 nanoparticles. Three different
organic chelators were synthesized and attached to the surface of core‐shell stuctures
and to the surface of SiO2 nanoparticles, which were used as prototypes. Ligand
grafting was assesed by FTIR, TGA and 1H and 29Si Solid State NMR. Dy3+ and Nd3+
uptake capacity was tested in both static and dynamic conditions and studied by SEM‐
EDS and complexometric titrations respectively.

Scheme 1. Design of the magnetic nanosorbent materials


Introduction
Lanthanide based materials have gained a lot of importance due to their wide range
of applications, such as high field strength magnets1, sensors2, electrooptical devices3,
catalysts4 etc. This increasing demand for lanthanides resulted in an increased
production of lanthanide through mining. Therefore, separation and purification of
lanthanides from ores urges development and optimization of new technologies.
Among them, one can find high affinity sorbents for sequestration of the lanthanides
from diluted industrial waste streams or the removal/sorption and extraction from
mine leachates. Furthermore, sorbent materials should be economical and reusable in
order to be feasible for industrial application as well as environmentally friendly in
order to reduce the environmental burden after use5.
For this reasons, a special interest was set to magnetic iron oxide nanoparticles for
facilitated removal of adsorbed content. Iron oxide nanoparticles have many
advantages due to their high magnetism and good stability as well as less toxicity when
compared with their metal magnetic counterparts like iron and cobalt6, which is very
important specially for environmental applications. Despite their good stability, it is
necessary to coat their surface with an organic or inorganic shell in order to protect
them from chemical degradation or agglomeration7. Among different types of coating
materials, silica coating is one of the most advantageous methods due to its stability
in aqueous solution. Silica coating allows protection of the magnetic cores and
prevents the direct contact of the magnetic core with external agents or structures
linked to the surface, enhances their biocompatibility, hydrophilicity and dielectric
properties8. Furthermore, it also facilitates surface modification due to the silanol
groups (‐SiOH) on the surface of silica9.
Heavy metal adsorption and ligand conjugated magnetic separation have been
extensively studied in water treatment and environmental cleanup10, but there are not
so many studies regarding the removal of rare earth ions (RE3+) from industrial waste
water or mining leachates with iron oxide nanoparticles. Binnemans and co‐workers11
have studied uptake capacity of all lanthanides serie (La3+, Pr3+, Nd3+, Sm3+, Gd3+, Tb3+,
Dy3+, Ho3+, Er3+, Yb3+, Lu3+) in aqueous solution using EDTA functionalized magnetic
(Fe3O4) and non‐magnetic (SiO2 and TiO2) nanoparticles.
In the present work, highly stable core‐shell Fe3O4@SiO2 nanoparticles were
synthesized and surface modified with organic chelates. Pure SiO2 nanoparticles were
also surface modified with the same organic ligands and used as prototypes.
Adsorption behaviour of all the materials towards Dy3+, Nd3+ and La3+ was tested by
SEM‐EDS, complexometric titrations and luminescent measurementes in the case of
silica prototypes. Selectivity with Nd/Dy and La/Dy binary rare earth mixtures was also
investigated by SEM‐EDS.


Optimization and effects of reaction parameters for Fe3O4 nanoparticles and
core‐shell Fe3O4@SiO2 nanoparticles
The most simple and efficient method for synthesis of Fe3O4 nanoparticles is the
coprecipitation of iron salts12,13, which is performed under non‐oxidizing environment
using a stoichiometric ratio of 2:1 Fe3+/Fe2+ in aqueous solution followed by the
addition of NH4OH solution.
The as‐synthesized Fe3O4 nanoparticles were succesfully coated by silica using Stöber
method14, which is based on the alkaline hydrolisis of tetraethyl ortosilicate (TEOS).
Two different synthetic procedures were carried out (reaction parameters are shown
in Table 1) leading to nanoparticles with different size and physical properties.
Table I. Reaction parameters for Fe3O4@SiO2 nanoparticles synthesis
No Fe3O4 H2O Ethanol NH4OH TEOS Hours Particle
(mmol) (mL) (mL) (mmol) (mmol) (h.) Size
(nm)
I 0.4 ‐ 40 270 11.8 2 800
II 0.4 32 mL 160 53.4 7.0 20 100
For our purpose, it is very important to achieve small particles since they lead to
materials with higher surface area allowing the binding of higher amounts of organic
ligands, which facilitates the coordination with rare earths in solution. Therefore, the
second synthetic pathway, previously reported on the literature7b,15, is more suitable
since it leads to nanoparticles with an average diameter of 100 nm (±15nm) according
to TEM results, which also confirms the effectiveness of Stöber method for silica
coating showing a silica shell of approximately 25 nm. (Figure 1).

A B
SiO
Fe3

C D


Figure 1. HR‐TEM images of Fe3O4@SiO2 nanoparticles (scale bar is 100, 20,5 200 nm for A, B,
C and D respectively)

For industrial purposes, it is very important that the core‐shell nanoparticles are stable
under acidic conditions in order to resist the process of separation from ores or
aqueous solution. Stability for these particles was tested with a solution of HNO3 0,1M
containing KSCN 0,5%.
The Fe3O4@SiO2 nanoparticles synthesized under synthetic pathway II are very stable
under time. Even after 5 months, there was no sign of iron leaching in the solution. In
order to be able to compare, images from nanoparticles synthesized according
synthetic pathway I are also shown (Figure 2), with a molar ratio TEOS/Fe3O4 of 0.125,
in which iron did leach from the solution.

Figure 2. Leaching test images of silica covered Fe3O4 nanoparticles

Adsorption of lanthanides (Dy3+ and Nd3+) on ligand grafted SiO2 nanoparticles
Determination of lanthanides uptake of the produced ligand grafted nanoparticles was
carried out by complexometric titration given the known concentration of initial
lanthanides in solution which is put in contact with the sorbent materials. EDTA
complexates metals in a 1:1 ratio, therefore a complexometric titration can be carried
out with the remaining solution after sorption, and the amount of lanthanides uptaken
can be calculated by substraction, since the initial amount is known and the remaining
amount is determined. To 50 mg of organic ligand grafted SiO2 nanoparticles (L2 and
L3 were tested), a calculated amount of lanthanide salt solutions (equivalent to the
moles of ligand grafted on the surface, calculated from TGA analysis) was added.
NaNO3 was added to facilitate sorption and the mixtures were stayed in static

conditions for different times (2, 8, 24 and 48hours). After that time the samples were
centrifuged and washed one time and the remaining solutions were titrated with
EDTA.
Kinetic curves obtained by these reverse complexometric titrations show a quick
achievement of the lanthanides adsorption equilibria (Figure 2). A higher affinity to
Nd3+ ions can be observed for Ligand‐2 grafted SiO2 nanoparticles whereas ligand‐3
grafted SiO2 nanoparticles have higher affinity to Dy3+ ions. Uptake capacity is, in the
case of ligand‐3 grafted SiO2 nanoparticles, very reasonable for static sorption
processes, resulting in an efficient material comparable to ion exchange resins16,17
(Table 2).

Figure 2. Kinetic curves of Dy3+ and Nd3+ static sorption process on hybrid silica nanoparticles
Table 2. Lanthanides uptake capacity of hybrid silica nanoparticles in static sorption
conditions
Time Dy3+ adsorbed (mmol Dy3+/g Nd3+ adsorbed (mmol Nd3+/g
(hours) sample) sample)
SiO2@L2 SiO2@L3 SiO2@L2 SiO2@L3
2 0,013 0,186 0,048 0,121
8 0,029 0,187 0,030 0,149
24 0,029 0,198 0,049 0,156
48 0,043 0,233 0,051 0,167


Adsorption of lanthanides (Dy3+, Nd3+ and La3+) on surface modified iron oxide
nanoparticles
Same complexometric determination with Na4EDTA was carried out for ligand grafted
iron oxide nanoparticles. Ligand 3 was used in this study because it showed to be the
most suitable for adsorption.
Kinetic curves show a quick achievement of the lanthanides adsorption equilibria
(Figure 3), reaching in the three cases more than 65% of the maximum uptake in 2
hours. Uptake capacity is, for the three lanthanide ions very reasonable and effective
for static sorption processes, resulting in an efficient material for adsorption (Table 3).
The best uptake capacity was achieved for Dy3+, with a maximum uptake of 0,25
mmol/g (equivalent to 40 mg Dy3+/g), whereas maximum uptake was 0,23 mmol/g for
Nd3+ (equivalent to 33,6 mg Nd3+/g) and 0,2
mmol/g for La3+ (equivalent to 27,8 mg La3+/g).

Table 3. Lanthanides uptake capacity of modified iron oxide nanoparticles in static
sorption conditions
3+
Time Dy ads. Nd3+ ads. La3+ ads
(mmol Dy3+ /g) (mmol Nd3+ /g) (mmol La3+ /g)
0 0 0 0
2 0,17 0,21 0,18
8 0,18 0,20 0,14
24 0,23 0,23 0,19
48 0,25 0,23 0,20


Figure 3. Kinetic curves of Dy3+, Nd3+ and La3+ static sorption process on modified iron oxide
nanoparticles
Lanthanides selectivity studies on ligand grafted SiO2 nanoparticles and
surface modified iron oxide nanoparticles
Selectivity on binary rare‐earth mixtures was tested on the ligand grafted SiO2
nanoparticles and the surface modified iron oxide nanoparticles. Nd/Dy and La/Dy
mixtures were used in a 1:1 molar ratio of lanthanide: ligand. Total uptake was
determined by complexometric titrations carried out as previously explained and
lanthanides selectivity was studied by SEM‐EDS analysis.
Tables 5‐6 show the total uptake capacity and SEM‐EDS surface characterization for all
the materials tested. It is necessary to remark that EDS analysis is very influenced by
the heterogeneity and roughness of the surface sample, but it is helpful to give a
qualitative estimation of the selectivity, showing a quite clear selectivity towards Dy3+.
Table 5. Lanthanides mixtures total uptake capacity of different materials in dynamic
sorption conditions
Sample Lanthanides Uptake capacity
(mmol Ln3+/mg)
Fe3O4@SiO2@L3 Nd‐Dy 0,127
Fe3O4@SiO2@L3 La‐Dy 0,026
SiO2@L3 Nd‐Dy 0,242
SiO2@L3 La‐Dy 0,275


Table 6. EDS analysis of sorbent materials with binary rare earths mixtures
Sample %Nd %La %Dy Dy/Nd Dy/La
molar molar molar molar molar
ratio ratio
Fe3O4@SiO2@L3 Nd‐Dy neutral pH 1,017061 ‐ 0,763981 0,751166 ‐
Fe3O4@SiO2@L3 La‐Dy neutral pH ‐ 0,214887 0,439505 ‐ 2,045283
SiO2@L3 Nd‐Dy neutral pH 0,96066 2,362555 2,459304
SiO2@L3 La‐Dy neutral pH ‐ 1,164047 3,49001 ‐ 2,998169

Conclusions
New magnetic nanosorbents with high stability under acidic conditions (over 5
months) were produced and their lanthanides uptake capacity was tested in neutral
pH conditions, showing in many cases capacities comparable with ion exchange resins.
Selectivity with binary rare earths mixtures was also tested and showed effective
selectivity towards Dy3+.
References
1. Peng Zhang, Li Zhang, Chao Wang, Shufang Xue, Shuang‐Yan Lin, and Jinkui Tang, “Equatorially
Coordinated Lanthanide Single Ion Magnets”, J. Am. Chem. Soc., 136, 4484‐4487(2014).
2. Parker D., Bretonniere Y. “Luminescent lanthanide complexes as sensors and imaging probes”,
Molecular Imaging, 49, 123‐146 (2005).
3. Jin T. Inoue S., Machida K., Adachi G., “Photovoltaic Cell Characteristics of Hybrid Silicon
Devices with Lanthanide Complex Phosphor‐Coating Film”, J. Electrochem. Soc., 144(11):
4054‐4058 (1997)
4. Shibasaki M., Yoshikawa N. “Lanthanide Complexes in Multifunctional Asymmetric Catalysis”,
Chem. Rev., 102, 2187‐2209 (2002).
5. Smith E.H. “Uptake of heavy metals in batch systems by a recycled iron‐bearing material”,
Water Research, 30, 2424‐2434 (1996)
6. a) Soenan S.J., Cuyper M.D., Smedt S. C., Braeckmans K. “Investigating the toxic effects of iron
oxide nanoparticles”, Methods in enzymology, 509, 195‐224 (2012). b) Wilkinson K., Ekstrand‐
Hammarstr B., Ahlinder L., Guldevall K., Pazik R., Kepinski L., Kvashnina K. O., Butorin S. M.,
Brismar H., Onfelt B., Osterlund L., Seisenbaeva G.A., Kessler V. “Visualization of custom‐
tailored iron oxide nanoparticles chemistry, uptake and toxicity”. Nanoscale, 4 (78), 7383‐7393
(2012)
7. a) LKu A‐H., Slabas EL., Schuth F., “Magnetic nanoparticles: synthesis, protection,
functionalization and application”, Angew Chem. Int. Ed., 46, 1222‐1244 (2007).b) Pogorilyi R.
P., Melnyk I. V., Zub Y. L., Carlson S., Daniel G., Svedlindh P., Seisenbaeva G. A., Kessler V. G.,
"New product from old reaction: uniform magnetite nanoparticles from iron‐mediated
synthesis of alkali iodides and their protection from leaching in acidic media", RSC Adv. 4 (43)
22606 – 22612 (2014),
8. a) Stjenrndahl M., Andersson M., Hall H.E., Pajerowski D.M. Meisel M.W., Duran R.S.
“Superparamagnetic Fe3O4/SiO2 nanocomposites: enabling the tuning of both the iron oxide
load and the size of the nanoparticles”, Langmuir, 24, 3532‐3536 (2008). b) Im S.H., Herricks
T., Lee Y.T., Xia Y. “Synthesis and characterization of monodisperse silica colloids loaded with
superparamagnetic iron oxide nanoparticles”. Chem. Phys. Lett., 401, 19‐23 (2005). C) Lu A.H.,

Salabas E.L. Schuth F., “Magnetic nanoparticles: synthesis, protection, functionalization and
application”. Angew. Chem. Int. Ed., 46, 1222‐1244 (2007).
9. Stober W., Fink A., Bohn E., “Controlled growth of monodisperse silica spheres in the micron
size range.” Journal of Colloid and Interface Science, 26, 62 (1968).
10. Xu P., Zeng M.G., Huang D.L., Feng C.L., Hu S., Zhao M.H., Lai C., Wei Z., Huang C., Xie G.X., Liu
Z. F., “Use of iron oxide nanomaterials in wastewater treatment: A review.” Sci. of the Total
Envir., 424, 1‐10 (2012).
11. Dupont D., Brullot W., Bloemen M., Verbiest T., Binnemans K., “Selective uptake of rare earths
from aqueous solutions by EDTA‐functionalized magnetic and non‐magnetic nanoparticles.”,
Appl. Mat. & Inter. (2014)
12. Basak, S.; Chen, D.‐R.; Biswas, P., “Electrospray of ionic precursor solutions to synthesize iron
oxide nanoparticles: Modified scaling law.” Chem. Eng. Sci., 62, 1263 (2007).
13. Martinez‐Mera, I.; Espinosa, M. E.; Perez‐Hernandez, R.; Arenas‐ Alatorre., “Synthesis of
magnetite (Fe3O4) nanoparticles without surfactants at room temperature.” J. Mater. Lett.
61, 4447–4451 (2007).
14. Lee, S.‐J.; Jeong, J.‐R.; Shin, S.‐C.; Kim, J.‐C.; Kim, J.‐D. “Preparation and characterization of
porous ultrafine Fe2O3 particles.” J. Magn. Magn. Mater., 282, 147 (2004).
15. Abbas M., Rao B.P., Islam M.N., Naga S.M., Takahashi M., Kim C., “Highly stable‐silica
encapsulating magnetite nanoparticles (Fe3O4/SiO2) synthesized using single surfactantless‐
polyolprocess”, Ceramics Int., 40, 1379‐1385 (2014).
16. El‐Sofany, E.A., “Removal of lanthanum and gadolinium from nitrate medium using Aliquat‐336
impregnated onto Amberlite XAD‐4.”, J. Hazard. Mater., 153, 948‐954 (2008)
17. Texier, A.C.; Andres, Y.; Le Coirec, P., “Selective Biosorption of Lanthanide (La, Eu, Yb) Ions by
Pseudomonas aeruginosa.”, Env. Sci. Tech.,33, 489‐495 (1999).


APPLYING IONIC LIQUIDS TO RARE‐EARTH SEPARATIONS
Kristian LARSSON1, Koen BINNEMANS 1
1
Department of Chemistry, KU Leuven, Celestijnenlaan 200F B‐3001 Heverlee, Belgium;
E‐mail: kristian.larsson@chem.kuleuven.be
Abstract
A way to improve the sustainability of solvent extraction is to replace the organic
diluents used in current separation processes by ionic liquids. The desirable properties
of ionic liquids include a very low vapour pressure, which avoids air pollution and
reduces the risk of fire. In order to successfully apply ionic liquid systems to rare‐earth
separation processes these systems need to be economic, simple and give high
separation factors. For this purpose, undiluted ionic liquids with quaternary
ammonium and phosphonium cations with nitrate anions have been investigated for
the extraction of rare earths from aqueous nitrate solutions using EDTA as an aqueous
complexant. At higher rare‐earth concentration levels, the distribution ratios are
loading‐dependent making the extraction dependent on the composition of rare
earths. Therefore the composition was modelled on the European deposit of Norra Kärr
(Sweden).
Introduction
The use of ionic liquids for the separation of mixtures of rare earths through solvent
extraction is a developing field with high potential because of both the intrinsically
beneficial properties of ionic liquids and the suitable extraction properties of some
ionic liquids.1‐3 Ionic liquids are per definition solvents consisting solely of ions and are
typically made up of salts with organic components and are liquid at useful
temperatures. The properties of ionic liquids generally include a very low vapour
pressure. This can make incorporating them into solvent extraction processes a part
of the transition towards greener processes since this reduces flammability and air
pollution.1 Ionic liquids can however only be considered green provided that they have
a low persistency and a low toxicity. A part of this is the avoiding the use of fluorinated
ionic liquids, due to their persistence in nature.4

Two reasonable starting points when using undiluted hydrophobic ionic liquids,
without functional groups specific to rare‐earth extraction, are choosing a suitable
counter‐ion that will create extractable complexes with the rare earths to be extracted
and choosing an ionic liquid with properties suitable for solvent extraction. For the
counter‐ions, the most obvious choices are sulphate, chloride or nitrate, since the
corresponding acids (H2SO4, HNO3, HCl) can be used to dissolve rare‐earth‐containing

starting materials. Sulphuric acid is somewhat unsuitable due to the low solubility of
some rare‐earth sulphate salts in solutions with high sulphate concentrations.
Hydrochloric acid is a potential choice since rare‐earth chloride complexes can be
extracted using chloride ionic liquids, though only at very high chloride concentrations,
although suitable solvating extractant extractants can be used to extract from chloride
systems.6‐8 The chloride‐chloride (chloride in both phases) ionic liquid systems, such
as tricaprylmethylammonium chloride for extraction of rare earths from chloride
aqueous solutions, are unsuitable due to low distribution ratios.

Nitrates have historically been used for rare‐earth separation systems since nitrates
form highly extractable complexes with rare‐earth ions. The use of
tricaprylmethylammonium nitrate in rare‐earth separation has strong precedents in
the literature, with a number of investigations and patents, but these extractants are
usually used diluted in molecular diluents and not as pure ionic liquids.9–17 Two
commercially available non‐fluorinated hydrophobic ionic liquids suitable for solvent
extraction were therefore chosen for this study: tricaprylmethylammonium chloride
(Aliquat 336) and trihexyl(tetradecyl)phosphonium chloride (Cyphos IL 101), both of
which can have their chloride anions easily replaced by nitrate ions. These ionic liquids
can also be used as diluents for other ionic liquids or classical extractants due to their
physical properties, including a comparatively (to other ionic liquids) low viscosity
when equilibrated with a water phase.

Gaudernack et al. described the use of quaternary ammonium nitrate dissolved in
organic diluents to extract light rare earths from an aqueous nitrate solution, followed
by extraction of the heavy rare earths using a quaternary ammonium thiocyanate
dissolved in organic solvents for the primary purpose of separation yttrium from the
other rare earths.16 This process has an optimum at a nitrate concentration in the
range of 5–7 M, with 0.5 M or less of rare‐earth metals, using a nitrate salt (e.g. sodium
or ammonium nitrate). The given acidity is stated as a range of pH values between 0.5
and 5, where extractability is higher at higher pH and separation increased at lower
pH. Modification of the separation factors by aqueous complexants such as diethylene
triamine pentaacetic acid (DTPA), followed by stripping by ammonium chloride and
subsequent precipitation by oxalic acid when using quaternary ammonium nitrates
dissolved in organic diluents to selectively extract lighter lanthanides was described by
Bauer et al.12. Trimble et al. described the use of an amine extractant to extract light
lanthanide followed by an acidic extractant extraction to remove heavy lanthanides,
with the purpose of attaining pure yttrium by stripping it using an acidic scrub.17
Morais et al. investigated the use of tricaprylmethylammonium nitrate dissolved in
kerosene to separate gadolinium and europium in a sodium nitrate matrix with a
separation factor of 2.0 that was independent of the nitrate concentration.10 Lu et al.
investigated the separation of praseodymium from neodymium using 0.6 M

tricaprylmethylammonium nitrate in an aromatic diluent with DTPA used as an
aqueous complexant.11

In this paper, we report on the separation of rare earths by extraction with
tricaprylmethylammonium nitrate or trihexyl(tetradecyl)phosphonium nitrate from an
aqueous nitrate feed solution containing EDTA as complexant. The significant
progress, compared to some previous studies, proposed in this work is to proceed
towards greener processes for the separation of rare earths. This is achieved by use of
nitrate ionic liquids in their pure form instead of diluted in, for example, kerosene in
order to gain the benefit of ionic liquids and thereby having an insignificant vapour
pressure. The fact that the concentration of the extractant is very high since the ionic
liquid phase acts itself as the extractant also gives high distribution ratios. The growth
in the number of products utilising rare‐earth metals in their construction has led to a
supply risk of rare earths which in turn has lead to the desire to develop geographically
diverse resources of rare earths.18 This study therefore uses a European rare‐earth
source as a model for the distribution of the aqueous solutions used for extraction
studies.

The aqueous complexing agent ethylenediaminetetraacetic acid (EDTA) was used to
increase separation factors. EDTA was chosen because of its favourable properties
since it forms more stable complexes with heavy rare earths than with light rare
earths. This increases the separation factors, since the extractant is a quaternary
ammonium or phosphonium salt that extracts light rare earths preferentially.
Experimental
Tricaprylmethylammonium chloride (Aliquat 336) is a mixed quaternary ammonium
salt containing mainly trioctylmethylammonium and tridecylmethylammonium
chloride (Sigma‐Aldrich, 88.2‐90.6%). Trihexyl(tetradecyl)phosphonium chloride
(Cyphos IL 101) is a phosphonium‐based ionic liquid (Cytec Industries, 97.7%). To
prepare the nitrate form of the ionic liquids, they were pre‐equilibrated three times
for 1 hour or more with a 2.5 M KNO3 solution to exchange the chloride ions by nitrate
ions. After three equilibrations, the chloride levels in the ionic liquid phase were under
the instrumental detection limits (TXRF).

Hydrated rare‐earth nitrate salts were obtained from Aldrich (Y 99.9%, Tb 99.9%, Sm
99.9%), Alfa Aesar (La 99.99%, Nd 99.9%, Dy 99.9%, Ho 99.9%, Yb 99.9%), Chempur
(Pr, 99.9%) and Acros (Gd 99.9%). The following rare earth oxides were used:
yttrium(III) oxide (Rhodia,99.99%), lanthanum(III) oxide (Aldrich, 99.99%),
praseodymium(III,IV) oxide (Janssen chemicals, 99.9%), neodymium(III) oxide (Rhodia,
99.9%), samarium(III) oxide (Rare Earth Products Ltd, 99.9%), europium(III)
oxide(Inframat Advanced Materials, 99.99%), gadolinium(III) oxide (Ventron LFA

products, 99.9%), terbium(III,IV) oxide (Rare Earth Products Ltd, 99.99%),

dysprosium(III) oxide (Rare Earth Products Ltd, 99.99%), holmium(III) oxide (Rare earth
products Ltd, 99.99%), erbium(III) oxide (Acros Organics, 99.99%), thulium(III) oxide
(GFS Chemicals, 99.9%), ytterbium(III) oxide (Rhodia, 99.9%), lutetium(III) oxide (Rare
Earth Products Ltd, 99.9%). Ethylenediaminetetraacetic acid (Acros Organics, 99%) and
potassium nitrate (Chem‐lab, >99%). Sodium nitrate (Chem‐Lab, >99%), ammonium
nitrate (Chem‐Lab, >99%). All dilutions were made using pure water (MilliQ, Millipore,
>18 MΩ/cm). A Picofox S2 (Bruker) total reflection X‐ray fluorescence (TXRF)
spectrometer was used to determine the metal content in solutions.

Batch solvent extraction experiments were performed using approximately 1 mL of
each phase, unless other phase ratios were tested, in which case the ionic liquid
volume was increased. Samples were shaken at constant temperature (30 °C) using a
TMS‐200 turbo thermo shaker (Hangzhou Allsheng Instruments Co., Ltd) for 30 min, or
longer when stated. The pH measurements were performed using a S220
SevenCompact pH/Ion meter (Mettler‐Toledo) and a slimtrode electrode (Hamilton)
and calibrated using pH 1.0 (Merck), 4.0 (Chem‐Lab), 7.0 (VWR) and 10 (Fischer)
buffers.

The distribution ratio (D) of a single metal is defined (equation 1) as the ratio of the
total concentration in the ionic liquid phase ([M]IL) by the total concentration in the
aqueous phase ([M]aq) after extraction and phase separation. The phase ratios are
defined as the volume ionic liquid divided by the volume aqueous phase.

D = [M]IL / [M]aq (1)

The basic assumption regarding the starting material was that an oxide concentrate
has been formed from a starting ore material, thereby attaining an up‐concentration
and the possibility of choosing the chemistry of the separation process. The
concentrate can in principle be in different forms such as oxide, carbonate, hydroxide,
etc. The starting material for this work was based on a HREE‐rich concentrates, such
as the European deposit Norra Kärr in Sweden, with an average composition given in
Table 1.19 This initial process development, described in this paper, did not include the
common impurities of the deposits ore body (zirconium, manganese, iron, etc.), It is
also assumed that cerium is removed by a separate process step prior to (or in
combination with) the re‐dissolution of the oxide material. The optimal conditions for
the separation process are dependent on the starting material and on the chosen
concentrations for the re‐dissolution of the oxide material.

High concentration nitrate aqueous phases with and without EDTA were prepared by
dissolving in nitric acid a mixture of rare‐earth oxides with a composition distribution

mimicking that of the Norra Kärr deposit (Table 1, column A). A 6 M NaNO3 matrix was
spiked with a few representative rare‐earth nitrate salts in order to study the nitrate
extraction systems at low loading (Table 1, column B).

Table 1: Column A: composition of the concentrated rare‐earth oxide solution in HNO3.
Columm B: concentrations of the solution spiked with rare earths for low
concentration, high matrix experiments (6 M NaNO3).

Metal A (M) B (M)

La 0.23 0.011
Pr 0.08 0.012
Nd 0.23 0.012
Sm 0.012
Eu 0.002
Gd 0.068 0.0081
Tb 0.02 0.0054
Dy 0.084 0.0071
Ho 0.02 0.0096
Er 0.029
Tm 0.011
Yb 0.045 0.015
Lu 0.011
Y 1.3 0.0095

Extraction with nitrate‐based ionic liquids
The distribution ratios in Table 2 indicate that, if the nitrate concentration is
sufficiently high, nitrate pre‐equilibrated Cyphos IL 101 is suitable for extracting all rare
earths together. Tricaprylmethylammonium nitrate is better suited for group
separations than trihexyl(tetradecyl)phosphonium nitrate since the separation factors
between light rare earths (e.g. La, Pr, Nd) and heavy rare earths (e.g. Yb) and yttrium
is greater for tricaprylmethylammonium nitrate. There is also a significant difference
in the mid‐range elements (Gd, Tb, Dy, Ho) that could be exploited to separate these
elements.


Table 2: Distribution ratios in tricaprylmethylammonium nitrate (A) and

trihexyl(tetradecyl)phosphonium nitrate (B) using an aqueous phase of 6 M NaNO3
spiked with rare‐earth nitrate salts. 30 min of mixing at 30 °C.

Metal A B
La >300 430 ± 140
Pr >250 230 ± 29
Nd 110 ± 13 200 ± 7.4
Gd 15 ± 1.0 74 ± 6.6
Tb 18 ± 1.2 120 ± 32
Dy 16 ± 1.6 86 ± 13
Ho 14 ± 1.5 72 ± 1.9
Yb 3.7 ± 0.7 18 ± 0.8
Y 4.7 ± 0.8 22 ± 1.1

However, the distribution ratios of ytterbium and yttrium are too high to be used for
separations at the 6 M nitrate concentration. Therefore the distribution ratio was
investigated as a function of the nitrate concentration between 3 M and 5 M nitrate
by diluting the spiked 6 M sodium nitrate solution with pure water before contacting
it with tricaprylmethylammonium nitrate (Figure 1). The separation factors remained
constant over the range of nitrate concentrations while only the distribution ratios
varied. The separation factors versus lanthanum for the heavy rare earths (Yb and Y)
are approximately 200 while for the medium rare earths (Gd, Tb, Dy and Ho) the
separation factors vary between 50 and 70.


Figure 1: Distribution ratios for tricaprylmethylammonium nitrate versus

concentration of sodium nitrate (3–5 M) at 1 h of mixing, 30 °C, pH approximately 2.

The separation factors coupled with the relevant distribution ratios are sufficient for a
process to separate the light rare earths from the heavy and medium ones. However,
there will be significant co‐extraction of both medium and heavy rare earths. To
increase the separation factors, an additional component is needed. An aqueous phase
complexant can be used to increase the separation factors and reduce co‐extraction.
A suitable aqueous complexant is EDTA, which has a varying complex strength across
the lanthanide series with stronger complexes for the heavier rare earths. This
interacts well with the nitrated ionic liquid system since the light rare earths are to be
extracted and the heavy remain in the aqueous phase, which is strengthened by
complexing with EDTA. Additionally DTPA, citric acid, lactic acid and malonic acid were
investigated, but were found to have undesirable properties at high loading, primarily
precipitation or too low solubility of the acid.

A more acidic medium was tested as well (6 M nitric acid), but only very poor rare‐
earth extraction was observed (D < 0.1). Acidic and non‐acidified chloride systems
were investigated in Cyphos IL 101 and Aliquat 336, using 4 M sodium chloride and 8
M hydrochloric acid; however, only poor (D < 0.1) or no extraction at all was observed.

Aqueous complexant assisted separation with tricaprylmethylammonium
nitrate
To evaluate the influence of pH on a system with the aqueous complexant EDTA, the
pH of a highly loaded system with a high EDTA concentration was varied between 1.5
and 3.75 (Figure 2). The distribution ratios were lower at low pH. This shows that the
pH has an impact on the distribution ratios; however, the separation factors remain
fairly constant, as shown in Figure 3. This means that the following investigations into
the effect of lanthanide metal loading and EDTA concentration should have the same
pH but that small variations will have only a small effect on separation factors.


Figure 2: Distribution ratios at high metal concentrations (1.9‐2 M) and high EDTA
concentrations (0.75‐0.8M) for tricaprylmethylammonium nitrate. The variations are
due to the pH modifications. The concentration of the aqueous feed solution is given
in Table 2. Only 5 elements are shown for clarity.

Figure 3: Separation factors at high loading (>93% of 2 M) and EDTA concentration
(0.75‐0.8M), the variation is due to 6M NH3 additions to modify pH, in a 6 M nitrate
matrix (ammonium nitrate). The concentration of the aqueous feed solution is given
in Table 2. Only 5 elements are shown for clarity.

A high rare‐earth metal loading (and thereby low additional added nitrate matrix) is
desirable as is having as low an EDTA concentration as possible while maintaining a
high distribution ratio of elements to be extracted together with high separation
factors. Increasing the EDTA concentration increases the binding strength of the

aqueous phase for especially the heavy rare earths and this increases the separation
factors, however, all distribution ratios are lowered too much at high EDTA
concentrations. Based on the experiments with the 2 M mixed rare‐earth solution (6
M nitrate), a series of dilutions down to 0.5 M of total metal content were investigated
with varying EDTA concentrations utilizing an ammonium nitrate matrix to have a
constant 6 M nitrate in the solutions. The pH in the solutions was kept at
approximately pH 2. The effect of EDTA concentration on the 0.5 M rare‐earth system
is shown in Figures 4 and 5. The distribution ratios decrease strongly with the
increasing EDTA concentration. All distribution ratios go below 1 already at 0.55 M
EDTA. Up to 0.4 M EDTA the distribution ratio of lanthanum remains fairly high and
the separation factors versus yttrium and heavy lanthanides such as dysprosium and
beyond are approximately 10 000 or even higher.

Figure 4: Distribution ratios at 0.5 M initial metal concentration in a 6 M nitrate matrix
(4.5 M ammonium nitrate added). The composition of the aqueous feed solution is
given in Table 2. Only 5 elements are shown for clarity.


Figure 5: Separation factor versus lanthanum at 0.5 M initial metal concentration in a
6 M nitrate matrix (4.5 M ammonium nitrate added). The concentration of the
aqueous feed solution is given in Table 2. Only 5 elements are shown for clarity.

The effect of EDTA concentration on a medium metal loaded system at 1 M rare‐earth
concentration is shown in Figures 6 and 7. The system is similar to the 0.5 M system
with the differences that the distribution ratios are lower and the reduction with
increasing EDTA concentration is lower. The separation factors are also somewhat
lower with the highest being with the highest tested EDTA concentration with
separation factors versus lanthanum for heavy rare earths about a factor of 10 lower
at 1000. The separation factors are very dependent on the variation of the rare‐earth
concentration at a constant, high, EDTA (0.8 M) concentration (Figure 8).


Figure 6: Distribution ratios at 1 M aqueous initial total metal loading in a 6 M nitrate
matrix (3 M ammonium nitrate added). The concentration of the aqueous feed
solution is given in Table 2. Only 5 elements are shown for clarity.

Figure 7: Separation factor versus lanthanum at 1 M initial metal loading in a 6 M
nitrate matrix (3 M ammonium nitrate). The concentration of the aqueous feed
solution is given in Table 2. Only 5 elements are shown for clarity. Separation factors
versus EDTA concentration (M).


Figure 8: Separation factors versus total initial lanthanide concentration (M) with a
constant EDTA concentration of 0.8 M, in a 6 M nitrate matrix (ammonium nitrate
added). The concentration of the aqueous feed solution is given in Table 2. Only 5
elements are shown for clarity.

The rare‐earth metal concentrations in the aqueous phase below 1 M at 0.8 M EDTA
concentration have very high separation factors, but suffer from too low distribution
ratios. The effect on the separation factors versus lanthanum is large, where the
middle lanthanides have a separation factor of approximately 10000 at low initial
metal concentration (0.75 M). However, the distribution ratio of neodymium has
become somewhat too low since it will require many steps to completely extract the
neodymium. Raising the nitrate concentration additionally can counteract this, for
example, an ammonium nitrate matrix can be used to attain a nitrate concentration
above 11 M without precipitation at room temperature.

The experimental results of using the desired high total nitrate concentration (11 M)
with a reasonable initial total metal content (1 M after pH adjustments) combined with
an EDTA concentration (0.2 M) optimized for an initial separation of material above
dysprosium in the lanthanide series are shown in Table 3. Praseodymium and lutetium
concentrations in aqueous (Pr) and organic (Lu) phase were below detection limits.


Table 3: Distribution ratios in tricaprylmethylammonium nitrate and using an aqueous
phase of 11 M nitrate with a 1 M rare‐earth initial metal concentration with a 0.2 M
EDTA concentration. 2 h of mixing at 50 °C, phase ratio 2 (O:A).

Metal Distribution ratio Separation factor (La)

La 520 1
Pr ‐ ‐
Nd 42 13
Sm 6.3 83
Eu 4.7 110
Gd 4.1 130
Tb 2.7 190
Dy 1.1 480
Ho 0.5 1100
Er 0.2 3500
Tm 0.01 35000
Yb <0.01 82000
Lu ‐ ‐
Y 0.4 1200

Conclusions
A system for liquid‐liquid extraction based on the ionic liquid
tricaprylmethylammonium nitrate, has been investigated for rare‐earth separations in
pure form using trace‐level and high concentration of rare earth in nitrate aqueous
solutions with and without EDTA. It was shown that high distribution ratios and
separation factors can be achieved using this system and that the distribution ratio
decreases with increasing atomic number across the lanthanide series. The separation
factors were independent of the nitrate concentration and were improved by using
EDTA as an aqueous phase complexant to increase the separation factor for the
heavier rare earths. The system behaves favourably when the nitrate concentration is
maximized due to the need to increase distribution ratios.
Acknowledgements
agreement n°309373. This publication reflects only the author’s view, exempting the
Community from any liability.

References
1. M. Dietz, “Ionic liquids as extraction solvents: Where do we stand?”, Sep. Sci. Technol., 41 (10)
2047‐2063 (2006).
2. Z. Kolarik, “Ionic Liquids: How Far Do they Extend the Potential of Solvent Extraction of f‐
Elements?”, Solvent Extr. Ion Exch., 31 (1) 24‐60 (2013).
3. Y. Liu, J. Chen and D. Li, “Application and Perspective of Ionic Liquids on Rare Earths Green
Separation”, Sep. Sci. Technol., 47 (2) 223‐232 (2012).
4. S. Wellens, B. Thijs and K. Binnemans, “An environmentally friendlier approach to
hydrometallurgy: highly selective separation of cobalt from nickel by solvent extraction with
undiluted phosphonium ionic liquids”, Green Chem., 14 (6) 1657‐1665 (2012).
5. M. J. Earle and K. R. Seddon, “Ionic liquids, Green solvents for the future”, Pure Appl. Chem., 72 (7)
1391‐1398 (2000).
6. A. I. Mikhailichenko, E. G. Goryacheva, N. P. Sokolova, N. M. Aksenova, L. V. Vdovina and A. P.
Emelyanov, “Extraction Of Rare‐Earth Elements From Chloride Solutions By Salts Of Quaternary
Ammonium Bases”, Sov. Radiochem., 26 (1) 25‐29 (1984).
7. P. K. Khopkar and J. N. Mathur, “Extraction Of Trivalent Actinides And Lanthanides By Tertiary And
Quaternary Amines From Concentrated Chloride Solutions”, J. Inorg. Nucl. Chem. 43 1035‐1040
(1981).
8. Kristian Larsson, Christian Ekberg and Arvid Ødegaard‐Jensen, “Using Cyanex 923 for selective
extraction in a high concentration chloride medium on nickel metal hydride battery waste”,
Hydrometallurgy, 129‐130 35‐42 (2012).
9. M. Černá, E. Volaufová and V. Rod, ”Extraction of light rare earth elements by amines at high
inorganic nitrate concentration”, Hydrometallurgy, 28 (1992).
10. C. A. Morais and V. S. Ciminelli, “Selection of solvent extraction reagent for the separation of
europium(III) and gadolinium(III)”, Miner. Eng., 20 747–752 (2007).
11. D. Lu, J. Horng and Y. Hoh, “The separation of neodymium by quaternary amine from didymium
nitrate solution”, J. Less‐Common Met., 149 219–224 (1989).
12. D. J. Bauer and R. E. Lindstrom, “Differential extraction of rare‐earth elements in quaternary
ammonium compound‐chelating agent systems”, Bureau of Mines Report of Investigations 7524,
(1971).
13. I. Komasawa, K. Hisada and M. Miyamura, “Extraction and separation of rare‐earth elements by
tri‐n‐octylmethylammonium nitrate”, J. Chem. Eng. Jpn., 23 (1990).
14. J. Preston, Hydrometallurgy, “The recovery of rare earth oxides from a phosphoric acid by‐
product. Part 3. The separation of the middle and light rare earth fractions and the preparation of
pure europium oxide”, 42 151–167 (1996).
15. D. J. Bauer and R. E. Lindstrom, “Selective extraction and separation of lanthanides with a
quaternary ammonium compound”, US patent, 3323857 (1967).
16. B. Gaudernack, G. Hannestad and I. Hundere, “Process for separation of yttrium from
lanthanides”, US patent, 3821352 (1974).
17. C. Trimble and D. Strott, “Yttrium purification process”, US patent, 3640678 (1972).
18. European Commission, “Critical raw materials for the EU”, Report of the Ad‐hoc Working Group
on defining critical raw materials, (2010).
19. G. C. Reed, “NI 43‐101 Technical Report, Norra Kärr REE ‐ Zirconium Deposit, Gränna, Sweden”,
Pincock, Allen & Holt, (2011)


Session V‐B: REE Occurrences in
Europe

EURARE IKMS: An Integrated Knowledge Management
System for Rare Earth Element Resources in Europe
Daniel Cassard1, François Tertre1, Guillaume Bertrand1, Frands Schjøth2,
Jørgen Tulstrup2, Tjerk Heijboer2, Jouni Vuollo3
1
Bureau de Recherches Géologiques et Minières (BRGM), Orléans, France
2
Geological Survey of Denmark and Greenland (GEUS), Copenhagen, Denmark
3
Geological Survey of Finland (GTK), Rovaniemi, Finland
Corresponding author: D. Cassard (d.cassard@brgm.fr)
Abstract
The EU‐FP7 project EURARE 'Development of a sustainable exploitation scheme for
Europe's Rare Earth ore deposits' will develop an operational data management
distributed system based on high‐level interoperability standards. Using advances
made in EU‐FP7 former projects such as OneGeology‐Europe, ProMine and
EuroGeoSource, ongoing projects like Minerals4EU in terms of database structure,
harvesting systems, web services, metadata management, and integration of non‐
structured information, and InGeoCloudS project in terms of cloud computing, EURARE
will contribute to implement the standards of a European geoscientific data
infrastructure defined in the EU‐FP7 EGDI‐Scope project. The objective is to develop an
Integrated Knowledge Management System (IKMS) allowing to easily combine rare
earth elements (REE) information related to primary and secondary mineral resources
and to provide end‐users with all the available information from primary sources to
new processing techniques and to waste streams. The IKMS will thus be designed as a
fully‐fledged extension of the EU‐MKDP (EU‐Minerals Knowledge Data Platform)
currently under development within Minerals4EU and which aims to become the future
European Mineral Resources (s.l.) data infrastructure. This assures an effective and
sustainable system designed for facilitating data updates and maintenance and
allowing a full access to information related to the whole REE resources value chain.

1 – Introduction
Data related to raw materials, either metallic, industrial or construction materials,
primary or from wastes (mining and industrial) are, most of the time, available in
Europe. However, they are often scattered amongst a variety of institutions, including
governmental agencies, universities, NGOs and industries. These data are often stored
in databases with their own design/architecture and vocabulary, making any attempt
of merging – even partially ‐ in view of a compilation difficult and time consuming. The
problems regarding availability, quality, organization, accessibility and sharing of data

are common to a large number of policies and are experienced across the various
levels of public authorities. Solving these problems requires measures that address
exchange, sharing, access and use of interoperable spatial data and services both at
national and European levels. This is the aim of the INSPIRE Directive1, but its
implementation in the Member states is just starting and achieving those objectives
remains a major challenge.

The development of the IKMS (Integrated Knowledge Management System) aims to
give a simplified, user‐friendly and efficient access to all available and new data related
to REE (Rare Earth Elements) from national geological surveys, scientific institutes and
universities, relevant industries and professional organizations, as well as from former
European projects such as ProMine2,3 (information on both mineral deposits and
anthropogenic concentrations resulting from mining and downstream activities) and
EuroGeoSource4,5 (information on energy and mineral resources, extraction locations,
production, reserves). The system is also designed to accommodate and manage semi‐
and non‐structured data (e.g., syntheses and statistics in the form of graph charts,
time‐series related to reserves and resources, secondary resources and exploration,
European REE market survey and raw material demand, REE resource exploitation
technologies including ore beneficiation, extraction technologies, end product
development and waste management practices…).

The role of the IKMS is to provide the end‐user with a seamless access to the whole
value chain from REE deposit exploration, mining and extraction of ore, beneficiation
and extraction technologies to treatment of end‐of‐life products and the generation
of "new" materials, with the ability to combine all spatial and non‐spatial pertinent
information in a single reference system.

2 – IKMS general architecture
The principles of the IKMS architecture follow the Implementing rules of the European
INSPIRE Directive1 to make data and services interoperable across Europe over a
distributed infrastructure.

The open architecture (figure 1) defines various components connected together to
build the IKMS, taking into account state‐of‐the‐art developments to enhance the
performance, stability, sustainability and user friendliness of the system. The
connections between the components are specified by standardized interfaces.

The main components of the IKMS architecture are the followings:
 Data on primary and secondary resources from databases provided by geological
surveys and former European projects such as ProMine and EuroGeoSource.

These data are delivered through INSPIRE compliant web services and
according to INSPIRE (plus extensions) data model6 and ERML v.2 data
model7,8,9,10,11.
 A system which gives access to available pertinent layers (e.g. geological,
geochemical, geophysical, geographical, land use maps…) provided as
interoperable services that are registered in a metadata catalogue (accessible via
an INSPIRE discovery service).
 New data and information delivered during the project by other work packages
(statistics analyses, analyses of REE supply and demand in the EU, stocks,
flows, including trends in what products are put on the market and their
composition, analyses of the composition of waste products and wastes arising
from pre-processing and their geographic location in EU) in various formats will
be accessible through the portal. This applies also to any 'open data' of interest
for the project. This semi- or non-structured information will be enriched by the
provider with a metadata record (e.g., Dublin Core type, see below), then
processed and indexed, first extracting relevant information from the documents
(e.g., named-entity recognition on atomic elements such as names of locations,
expressions of time, thematic specific terms…), then classifying this
information according to three facets (spatial, temporal and thematic) and
creating indexes for each facet. The way to display the list of selected documents
will take into account these three facets.
 Besides a map viewer, the IKMS also includes a search engine for all data, layers
and documents delivered by the project and all external pertinent documents.
3 – IKMS detailed architecture
In order to increase the value of the work done in former EU‐FP7 project, and also to
minimize new developments, it was agreed to reuse the Central Database of the
EuroGeoSource project in EURARE. To achieve this goal, there was a need to both
update the structure of the database to be ERML 2.0 compliant, and add the support
of wastes and chemical analyses. There was also a need to update the database to use
the final code lists (or vocabularies) coming from INSPIRE MR specifications which
were themselves improved and validated at an international level by the International
Union of Geological Sciences (IUGS), Commission for the Management and Application
of Geoscience Information (CGI) and the Geoscience Terminology Working Group
(GTWG). Finally, the EuroGeoSource cookbook used to help the partners to serve their
data has been updated accordingly to the new structure of the database.

3.1 ‐ A professional architecture
The IKMS adopted a professional architecture with a harvesting system and a diffusion
system (figure 2). The system is divided into three parts:
 The national level from which ERML 2.0 compliant WFS’ (Web Feature
Service) deliver data to the harvesting system either from (already)
EuroGeoSource structured databases or directly from other data sources in each
country;
 The central harvesting system which regularly reads data from the national
WFS’ and stores it in the Harvesting Database. Data is subsequently being
delivered to the central diffusion system through a database synchronization
mechanism. The Harvesting Database is optimized towards reading data from
the national level and delivering this data to the diffusion system;
 The central diffusion system which is updated with data from the harvesting
system at regular intervals and makes this data available to users through the
IKMS web portal. The Diffusion Database has specific optimization to offer the
better experience to the user for the delivery of the data and the computation of
on-demand services.

The Information Factory is a software component able to process data (indexation,
filtering, descriptive/exploratory statistics computation, automatic report
generation…).

3.2 ‐ The national level provider system
Each participating survey will have a National Provider Database which will have the
same architecture as the Central Harvesting Database. This database is a PostGreSQL
database, which build upon the schema that was used in the EuroGeoSource project,
extended by the requirements from EURARE (including mining wastes and
geochemical analyses). This Provider Database will contain the data of a survey that
was mapped from that survey’s own database using the INSPIRE vocabulary defined in
the code lists and data types. This mapping can be done using an ETL (Extraction‐
Transformation‐Loading) tool such as GeoKettle. The data from this database will then
be transformed to an INSPIRE compliant web feature service (WFS) using the
DeeGree3 framework and toolstack.
3.3 – The harvesting system
The harvesting procedure is based on an updated EuroGeoSource harvesting
procedure. The harvesting from the National Provider/project WFS’ is done with an
ETL tool run in a scheduled job. Aggregation and other manipulation of data in the

harvesting system, if needed, will be performed by procedures within the PostGreSQL‐
database.

The possibility to set up an incremental harvesting system, which would avoid a total
refreshment of harvested data, is still under consideration. The problem here is not of
a technical nature but depends on the fact that some data may not have a time stamp.
3.4 – The Harvesting Database and the INSPIRE MR data model
The Harvesting Database is built by extending the existing EuroGeoSource Central
Database with new tables. Additions and changes to the INSPIRE MR data model
include:
 the Mining Waste extension from INSPIRE, and the modification of the Earth
Material feature used in EuroGeoSource in order to include other materials;
 parts of INSPIRE Draft Guidelines for the use of Observational Measurements
and Guidelines for storing geochemical measurements of Earth Material
specimens;
 the Mined Material part of the Mine extension from INSPIRE v.3;
 the Supergene Processes from the Occurrence extension from INSPIRE v.3.

The delivery of data from the harvesting system to the diffusion system is carried out
by sending SQL update scripts.
3.5 ‐ Vocabularies
Vocabularies used in EURARE are based on INSPIRE MR code lists. These code lists,
since their initial publication in the Data Specification on Mineral Resources – Draft
Guidelines6 have been submitted to the IUGS/CGI/GTWG (cf. supra) for validation at
the international level and have thus been reworked and improved. During this
process, it was ensured that all the needs of the EURARE project regarding REE were
taken into account. This led to the improvement – before their final acceptance ‐ of
several code lists such as the Commodity Code Value, the Deposit Group and the
Deposit Type code lists.
3.6 – The Diffusion Database
The Diffusion Database will be a copy of the Harvesting Database, which will be
specialized for data delivery and services that carry out computations based on the
stored data. This database will be kept updated with the Harvesting Database using
database synchronization. This synchronization will rely on SQL scripts forwarded from
the Harvesting Database to the Diffusion Database. The specialization of this database
will aim to optimize the response time of the services proposed to the users.


The services on top of the Diffusion Database are:
 Simple map visualizations using WMS (Web Map Services).
 EURARE’s dedicated services: these services are described in a dedicated
paragraph below.
 Data download: in some case (and depending on the access rights to the data)
the data will be available for download so that they can be processed in other
applications (e.g., desktop GIS…).
 Search facilities: these facilities will allow the user to search in the whole
database and documents.
3.7 – The non‐structured data
Beside the Diffusion Database, an additional interface will allow experts to feed the
knowledge base with non‐structured data. These non‐structured data can be of
different types (reports, synthesis notes, thesis…). This interface will allow the expert
to add a document (the document will be part of the IKMS and retrievable within the
IKMS) or a link to an existing document in some place accessible via the Internet (the
document will stay on its original server; the end‐user will have to follow a link to
retrieve it). To be able to integrate it into the knowledge base, the expert will have to
include metadata to this document. These metadata will be based on Dublin Core
(http://dublincore.org/) with a specific extension for EURARE (at least quality of the
data). Dublin Core Metadata Element Set has 15 elements covering the most
important properties to describe a document (title, creator, subject, coverage –
temporal or geographical…). These metadata will then be used in the search facilities
to retrieve the documents.
3.8 – The search facilities
The search facilities will be based on a search engine indexing the Central Diffusion
Database, the documents corpus (non‐structured data) and some external databases
(if any). A user interface will allow the end‐users to retrieve data from the whole IKMS
using a simple input (Google‐like search, a simple sentence will search in the whole
IKMS), or using some specialized interfaces (search specific concepts using their main
attributes). Almost all the indexed concepts will have a geographic and temporal
extents (coming from INSPIRE MR/ERML for the Diffusion Database, coming from
Dublin Core metadata for the documents). These extents will allow the user to retrieve
the most accurate response for his search.
3.9 – EURARE's dedicated services
In order to answer the end‐users most common requests, some specific calculation
services will be set up for EURARE on top of the Central Diffusion Database. As it is
difficult to include LREE and HREE in the Commodity Code Value code list, a service can

be developed which will classify each deposit as a HREE or a LREE deposit, taking into
account the main commodity, the second one, etc. This would, for instance, allow
plotting deposits on a map with distinct symbols (HREE or LREE), which could be useful
for end‐users as both groups of REE have different industrial applications. It would also
be possible to do this from the geological or metallurgical point of view, as both
classifications are not similar. Similarly, a service could also calculate the ratio
LREE/HREE, provided that available data is accurate and numerous enough. A service
can also be set up for plotting simultaneously known REE deposits (having REE among
the main commodities) and deposits containing REE (REE present, but (totally)
subordinate). For the latter ones, the selection can be made based on the presence of
one or several REE containing mineral(s) in the paragenesis and/or in the mineralogy.
This could possibly help to better define provinces or districts.

Finally, both harvesting and diffusion systems will be self‐contained systems that can
easily be moved to other installations/platforms. More particularly, the project will
investigate how platforms like InGeoCloudS could be exploited for the system. This
work will be realized in close collaboration with the EGDI‐Scope and InGeoCloudS
projects.

4 – Conclusion
The present work does not only represent an advance in terms of implementation of
a distributed architecture and management of structured, semi‐ and non‐structured
data related to REE. It also contributes to the implementation of a European Mineral
Resources Data Infrastructure and to bring concrete answers to several
recommendations made by the European Commission. The IKMS contributes to the
improvement of the standardization of mineral resources data across participating
organizations, as suggested in the WP3 of the EC Communication “Making raw
materials available for Europe’s future well‐being proposal for a European innovation
partnership on raw materials” (COM(2012) 82 final) and to the improvement of their
availability. It also contributes to implementing land‐use planning best practices as
suggested in the relevant report by Enterprise and Industry Directorate General
(http://ec.europa.eu/enterprise/policies/raw‐materials/files/best‐practices/sust‐full‐
report_en.pdf). The IKMS outputs will be appropriate for use at a variety of scales,
from local/regional land use planning up to the pan‐European level. It allows better
communication with the general public and specific stakeholders within the mineral
sector and facilitates more balanced and well‐informed debates and decisions related
to the raw materials (and notably critical ones) non‐energy extractive industry. Finally,
it contributes to the enhancement of minerals role and related industrial technologies
for competitive growth as part of the Europe 2020 Strategy.


Figures and captions

Figure 1: Simplified architecture of the EURARE IKMS (Integrated Knowledge
Management System). The system is designed to accommodate both structured
data from existing databases (data produced by geological surveys or external
providers) and semi‐ or non‐structured information produced by other work
packages.


Figure 2: The IKMS detailed architecture showing the Harvesting Database, the
synchronization process with the Diffusion Database, the Information factory and
the indexation process of both structured and semi‐ and non‐structured data. Note
the possible use of Excel‐type portrayals for feeding project databases.

References
1. INSPIRE Directive (2007). http://inspire.jrc.ec.europa.eu/ ; http://eur‐
lex.europa.eu/LexUriServ/LexUriServ.do?uri=OJ:L:2007:108:0001:0014:EN:PDF.
2. Serrano J.‐J., Kauniskangas E., Cassard D. (2010). ProMine: architecture, portal and
web services to provide a European mineral resources information system, 4th
INSPIRE Conference. Krakow, 22‐25 June.
3. Cassard D., Bertrand G., Billa M., Serrano J.‐J., Tourlière B., Angel, J.‐M., Gaál G †.
(submitted). ProMine mineral databases: new tools to assess primary and
secondary mineral resources in Europe. ProMine Special Volume. Springer Editor.
4. Gruijters S. and Fruijtier S. (2011). Using cloud computing to increase the usability
of distributed Web GIS Systems. eChallenges e‐2011 Conference Proceedings, 26‐
28 October 2011, Florence, Italy. Paul Cunningham & Miriam Cunningham (Eds.),

IIMC International Information Management Corporation, 2011, ISBN: 978‐1‐

905824‐27‐4, 8 p.
5. EuroGeoSource (2013). EuroGeoSource homepage http://eurogeosource.eu/. In
the menu EuroGeoSource Deliverables: 1) Publications and presentations:
http://eurogeosource.eu/deliverables/publications‐and‐presentations; 2) Public
Reports: http://eurogeosource.eu/deliverables/public‐reports
6. INSPIRE Thematic Working Group Mineral Resources (2013). D2.8.III.21 Data
Specification on Mineral Resources – Technical Guidelines. 156 p. European
Commission Joint research Center Publisher.
http://inspire.ec.europa.eu/documents/Data_Specifications/INSPIRE_DataSpecif
ication_MR_v3.0.pdf
7. EarthResourceML (2013).
http://www.earthresourceml.org/earthresourceml/2.0/doc/ERML_HTML_Docu
mentation/
8. Vuollo J., Cassard D., Simons B., Seymon A. (2010). The Earth resource data
exchange model (EarthResourceML) – a tool for delivering INSPIRE and ProMine
mineral resource data, 4th INSPIRE Conference. Krakow, 22‐25 June.
9. Vuollo J., Simons B., Laxton J., Cassard D., Seymon A. (2012). EarthResourceML
v.2.0 – an upgrade of the CGI‐IUGS Earth Resource Data Model due to INSPIRE
Data Specification. 34th International Geological Congress, Brisbane, August 2012,
Abstracts on CD‐ROM.
10. Cassard D., Serrano J.‐J., Vuollo J. (2012). The use of the standard exchange
EarthResourceML in the ProMine project. 34th International Geological Congress,
Brisbane, August 2012, Abstracts on CD‐ROM.
11. Raymond O., Vuollo J., Richard S. and CGI Interoperability Working Group (2012).
Global Geoscience Data Transfer Standards ‐ GeoSciML and EarthResourceML.
eResearch Australasia Conference, Brisbane, Australia.
http://conference.eresearch.edu.au/eres2012/posters‐2012/


REE DEPOSITS AND OCCURRENCES IN GREENLAND
Kristine THRANE, Per KALVIG & Nynke KEULEN
Department of Petrology and Economic Geology, Geological Survey of Denmark and
Greenland, Øster Voldgade 10, Copenhagen
kt@geus.dk, pka@geus.dk, ntk@geus.dk
Abstract
The Greenland REE potential is considered excellent, consisting of three giant deposits
and several smaller deposits in addition to favourable geological settings for
undiscovered REE‐occurrences. In South Greenland, the main REE occurrences are
related to the Mesoproterozoic alkaline intrusions of the Gardar Province, including
three world‐class deposits: (i) Kvanefjeld and (ii) Kringlerne both hosted by the
Ilímaussaq intrusion, and (iii) the Motzfeldt Sø deposit associated with the Igaliko
Complex. In West Greenland, the REE occurrences are mainly related to carbonatite
intrusions e.g. the Sarfartoq, Qaqarssuk and Tikiussaq intrusions, but some smaller
occurrences also exist at Niaqornakavsak, Attu and Nassuttooq. In the Caledonian fold
belt of East Greenland, occurrences are related to quartz veins in Bjørnedal, as well as
a palaeoplacer deposit at Milne Land.

Introduction
Greenland comprises several large REE deposits in various geological settings. The
largest deposits are hosted in peralkaline intrusions related to the Gardar Province in
South Greenland, including the deposits around Kvanefjeld, Kringlerne, and Motzfeldt
Sø. Additionally, three carbonatite deposits are known from the West coast, and
quartz veins in Bjørnedal and one palaeoplacer deposit (Milne Land) are located in East
Greenland (Figure 1). An overview of the REE‐deposit types and main REE‐minerals are
listed in Table 1. Additionally, Greenland holds potential for undiscovered REE deposits
related to carbonatite magmatism, alkaline intrusions, pegmatite settings, IOCG
mineralising systems, and palaeoplacer environments1,2.

South Greenland
The Mesoproterozoic Gardar province in South Greenland is an intracratonic rift
province of sandstones, and a variety of volcanic and plutonic igneous rocks of alkaline
and peralkaline affinity. Of the intrusions in the province, the most important
economic REE deposit include the Ilímaussaq intrusion (1160 Ma), and the Igaliko
Nepheline Syenite Complex (1273 Ma), hosting the Kvanefjeld and Kringlerne REE

deposits, and the Motzfeldt Sø REE deposit, respectively. In addition, the Gardar
province also encompasses the minor REE carbonatite occurrence at Grønnedal‐Ika.
The Ilímaussaq intrusion is largely emplaced during block subsidence and formed by
three pulses, of which the third pulse formed a layered series of nepheline syenite,
enriched in REE, U, Nb, Ta, Zr, Zn and F. The Ilímaussaq intrusion covers an area of 8
km x 17 km and hosts two world class REE deposits: (i) Kvanefjeld, an intermediate
series sandwiched between roof and floor series, which consist mainly of lujavrites,
and (ii) Kringlerne, the bottom cumulates of kakortokite.
The Kvanefjeld multi‐element deposit is an accumulation of igneous rocks from the
central part of the Ilímaussaq intrusion. The bulk of the REE and the U is associated
with lujavrite rocks containing disseminated steenstrupine, lovozerite and an
unnamed mineral species. Greenland Minerals and Energy Ltd. (GME) has been the
license‐holder since 2007, and reports3 the indicated JORC compliant resources of
Kvanefjeld to be 437 Mt ore grading 1.1 % TREO, hence containing 4.8 Mt TREO (0.18
Mt HREO), 263 Mlbs U3O8, and 1 Mt Zn. In addition, they have located and assessed
two other lujavrite bodies in the vicinity, the Sørensen Zone and Zone 3. However, the
proposed mining project is focused on the Kvanefjeld deposit. The deposit will be
exploited as an open pit mine, with ore beneficiation at an adjacent plant, extracting
sphalerite by flotation followed by hydrometallurgical leaching of steenstrupine and
recovery of the uranium and REE. GME anticipate annual production to be in the range
of 3 Mt ore resulting in 33,000 ton TREO (equivalent to 3,400 ton Nd2O3 and 300 ton
Dy2O3), 1.8 Mlbs U3O8 and 6,636 ton Zn. The feasibility report is in progress.
The Kringlerne multi‐element deposit is hosted in agpaitic kakortokite, and consists of
a total of 29 exposed cyclic layers with a total thickness of about 200 m, made up by
arfvedsonite dominated kakortokite, eudialyte dominated kakortokite, and a
nepheline‐feldspar dominated kakortokite. Eudialyte, the main economic mineral, is
enriched in REE‐Zr‐Nb‐Ta. TANBREEZ is the current license‐holder, and they report4
the inferred JORC‐compliant resource to be at least 4,300 Mt grading 0.65 % TREO, 0.2
% Nb2O3 and 1.8 % Zr2O5 equaling 28 Mt TREO. Their proposed mining project involves
an open pit mine near the fjord, hauling the ore to a nearby beneficiation plant, where
they will produce three products: (i) eudialyte concentrate (REE, Nb, Zr); (ii) feldspar
concentrate, and (iii) arfvedsonite concentrate, all to be shipped for further processing
and/or use outside Greenland. Planned annual production is set at5 500,000 ton ore,
equal to 3,250 ton TREO (equivalent to 400 ton Nd2O3 and c. 90 ton Dy2O3) and 9,000
ton Zr2O5. The feasibility report is in progress.
The Motzfeldt Sø REE deposit is part of the Motzfeldt Centre, which in turn is one of
the intrusions of the Igaliko Nepheline Syenite Complex. It contains extensive U‐Nb‐
Ta‐Zr‐REE mineralisation6,7. Significant Ta‐Nb‐enriched zones relate to altered syenite,
minor pegmatite and diorite dykes and high grade REE intersections are related to
pegmatites at depth. Presently, RAM Resources Ltd. has licensed the area and reports8

a resource estimate of 340 Mt grading 0.26 % TREO, 0.19 % Nb2O3, 0.012 % Ta2O3 and
0.46 % ZrO2. A scoping study is in progress.
West Greenland
West Greenland possesses a number of carbonatite associated REE deposits intruded
between Neoproterozoic and Jurassic. The most important of those, in terms of REE
resources, are Sarfartoq, Qaqarssuk, and Tikiussaq. These carbonatites are located in
southern West Greenland intruded into the Archaean basement of the North Atlantic
Craton. In addition, West Greenland hosts minor REE occurrences at Niaqornakavsak,
Attu and Nassuttooq.

The Sarfartoq carbonatite intrusion is located on a Precambrian thrust zone and
intruded 560 ± 13 Ma ago9,10, during the opening of the Iapetus Ocean. The host rock
consists of orthogneiss with schlieren of amphibolite. The complex has an ellipsoidal
shape and covers about 90 km², of which 15 km² are intrusive carbonatites. The core
of the complex consists of rauhaugite with schlieren of sövite and beforsite dykes,
surrounded by a Na‐type fenite zone with aegerine‐ and pyrochlore‐bearing rocks. The
carbonatite contains REE‐bearing minerals in veinlets of dolomite and REE‐carbonatite
and in shear zones with thorium, uranium, K‐feldspar alteration and limonitisation11.
The most common REE‐minerals are synchysite and zhonghuacerite, which are
relatively enriched in LREE12, most notably Ce, La and Nd. Furthermore, ancylite,
burbankite and Sr‐REE barite were observed13. Hydrothermal activity caused an
enrichment of niobium in the carbonate veins and shear zones, which have been
drilled in the past. Hudson Resources Ltd. currently holds the license of the area and
reports resource estimates that include indicated NI 43‐101 compliant resources of 5.9
Mt averaging 1.8 % TREO and an additional inferred resource of 2.5 Mt averaging 1.6
% TREO, based on a 1 % cut‐off grade12,14.
The Qaqarssuk carbonatite (Qeqertaasaq) complex consists of carbonatite ring‐dykes
that are elliptical to semi‐rectangular in shape and intruded into the Archaean
basement at 165.7 ± 1.9 Ma9,10. The carbonatite covers an area of 15 km² and is
surrounded by fenitised basement. The ring‐dykes consist of sövite, olivine‐sövite, and
dolomite carbonatite. They are cut by late‐stage sövite veins, REE‐carbonatite veins,
ferrocarbonatite and lamprophyre dykes15. The sövites contain increasing amounts of
silicate minerals and inclusions of fenitised basement away from the centre of the
complex. The fenites are composed of albite, alkali amphibole, and alkali pyroxene16.
The major REE minerals in the complex are ancylite, burbankite, huanghoite and
qaqarssukite. NunaMinerals A/S holds the license for the area and report17 that
surface sampling and trenching have yielded up to 13.2 % TREO within the late‐stage
carbo‐hydrothermal REE veins, and that drilling demonstrated that the REE veins are
up to 4.5 % TREO. Metallurgical flow sheets are currently being defined.

The Late Jurassic Tikiusaaq carbonatite complex was discovered in 200518. The
presence of the carbonatite was predicted by a study of regional stream sediment
geochemistry and aeromagnetic data. The complex consists of a central intrusive
carbonatite surrounded by a fenite zone with carbonatite and aillikite dykes, and the
main REE‐mineral is bastnaesite. The exposed carbonate sheets cover 2 km x 3 km,
and the alteration zone extends up to 14 km in diameter. Remote sensing data suggest
that a massive carbonatite is hidden below the glacial terraces. Both magmas have
common origin within carbonated upper mantle source19. The 165‐150 Ma carbonatite
intruded during the opening of the Labrador Sea. Apart from REEs, the carbonatite
complex hosts P, Nb and Ta. The highest grade REEs are associated with a thorium
anomaly. NunaMinerals A/S holds the license for the area and reports up to 9.6 % TREO
in a carbonate float17.
The Niaqornakavsak REE deposit was discovered in 2007 by Avannaa Resources who
still holds the license of the area. The mineralisation is hosted in Early Proterozoic
metasediments and metavolcanic of the Karrat Group, situated in the Karrat Isfjord
region. The mineralisation is a lithological distinct horizon of banded carbonates within
an amphibolite unit. The horizon strikes 1.5 km and dips 32 degrees. The mineralised
body is estimated to contain tens of millions tons of ore. Drill intersections indicate
that TREO+Y typically varies between 0.8 and 1.5 wt. % with some layers containing up
to 1.9 wt. % with an average of 1.03 wt. %. The REE´s are mainly hosted by bastnaesite,
monazite, allanite and other REE silicates20. The Niaqornakavsak bulk samples indicate
an average TREO+Y of 1.36 wt. % of which the average HREO+Y content is 13.06 %. A
metallurgical analysis and flow sheet development is in progress.

East‐Greenland
In the Caledonian fold belt of East Greenland, a couple of occurrences exist in quartz
veins in Bjørnedal, in a palaeoplacer deposit at Milne Land and in an alkaline intrusion
in the Gardiner Complex.

Nb‐REE‐bearing types of mineralisation occur in the Kap Simpson complex in Bjørnedal
on Traill Ø. The mineralisation occurs in course‐grained quartz veins intruding
Jurassic/Cretaceous shales, siltstone and sandstones. The veins are up to 30 m long
and 15 cm wide and contain minor amounts of Nb and REE‐bearing minerals
columbite, euxenite, samarskite, fergusonite, monazite and bastnaesite. Selected
samples contain up to 3.2 % Nb and 3 % REE, but the average content is not reported21.

In East‐Greenland, the Jurassic placer deposit on Milne Land is forming another source
for REEs. The deposit, placed within the Mesozoic basins in central East Greenland,
formed during the opening of the North Atlantic Ocean. The placer consists of heavy
mineral‐bearing arkosic sandstones and breccias that are rich in zircon, monazite, Ti‐

minerals and garnet. The units forms the lowermost 20 m of the Charcot Bugt
Formation, which rests unconformably on Mesoproterozoic migmatic granite. The
best outcrops occur on Hill 800, where these sedimentary rocks are 40‐50 m thick and
are more than 500 m in diameter9,22. CGRG Ltd. currently has the license for the area
and is exploring for Mo‐Zr‐REE‐Ti.
Conclusion
The Greenland REE potential is considered excellent including several known deposits,
as well as favourable geological settings for undiscovered REE‐occurrences such as
alkaline intrusions, pegmatite settings, carbonatites and IOCG mineralising systems.
Currently, two giant deposits that have reached advanced stage of development, plan
to enter operation before 2018.
Figure 1: Known REE deposits/occurrences in Greenland

Table 1: Deposit types and REE‐minerals of the Greenland REE‐occurrences
Occurrence REE‐ Main REE Chemical formula
deposit minerals
type
Attu Granitic Allanite (Ce,Ca,Y,La) 2(Al,Fe3+)3(SiO4) 3(OH)
Bjørnedal Granitic Euxenite (Y,Ca,Ce,U,Th)(Nb,Ta,Ti) 2O6
Samarskite (YFe3+Fe2+U,Th,Ca) 2(Nb,Ta) 2O8
Fergusonite REENbO4
Monazite (Ce,La)PO4
Bastnaesite (Ce,La,Y)CO3F
Gardiner Alkaline Perovskite CaTiO3
Apatite Ca5(PO4)3(F,Cl,OH)
Grønnedal‐ Carbonat Bastnaesite (Ce,La,Y)CO3F
Ika e
Kringlerne Alkaline Eudialyte Na15Ca6(Fe,Mn)3Zr3SiO(O,OH,H2O)3(Si3
O9)2
(Si9O27)2(OH,Cl)2
Kvanefjeld Alkaline Steenstrupin Na14REE6(Mn,Fe)3(Th,U,Zr)(Si6O18)2(PO4
e )7∙3H2O
Lovozerite Na2CaZrREE(H4Si6O18)H2O
Milne Land Palaeopl Monazite (Ce,La)PO4
acer
Motzfeldt Sø Alkaline Pyrochlore, (Na,Ca)2Nb2O6(OH,F)
Bastnaesite, (Ce,La,Y)CO3F
Monazite (Ce,La)PO4
Eudialyte Na15Ca6(Fe,Mn)3Zr3SiO(O,OH,H2O)3(Si3
O9)2
(Si9O27)2(OH,Cl)2
Nassuttooq Granitic Monazite (Ce,La)PO4
Niaqornakavs Carbonat Bastnaesite (Ce,La,Y)CO3F
ak ite? Monazite (Ce,La)PO4
Allanite (Ce,Ca,Y,La) 2(Al,Fe3+)3(SiO4) 3(OH)
Qaqarssuk Carbonat Ancylite Sr(Ce,La)(CO3) 2(OH)H2O
ite Burbankite (Na,Ca) 3(Sr,Ba,Ce)(CO3)2
Huanghioite BaCe(CO3)2F
Qaqarssukite Ba(Ce,REE)(CO3)2F
Sarfartoq Carbonat Synchysite Ca(Ce,La,Nd,Y)(CO3)2F
ite Zhonghuacirit Ba2(Ce,La,Nd)(CO3)3F
e

Tikiusaaq Carbonat Monazite (Ce,La)PO4

ite

References

1. Sørensen, L.L., Kalvig, P. & Hanghøj, K. 2011: Rare earth element potential in Greenland.
Reporting the BMP/GEUS mineral resource assessment workshop 29 November – 1
December 2010. Danmarks og Grønlands Geologiske Undersøgelse Rapport 2011/80, 30 p + 1
DVD.
2. Sørensen, L.L. & Kalvig, P. 2011:Geology and Ore. 20/ 2011. GEUS; 12 pp.
3. http://www.ggg.gl/
4. TANBREEZ presentation, Greenland day, Perth 2013.
5. http://www.tanbreez.com/
6. Armour-Brown, A., Tukiainen, T. & Wallin, B. 1982: The South Greenland uranium exploration
programme. Final report, 95 pp. Unpublished report, Grønlands Geologiske Undersøgelse.
7. Tukiainen. T. 2014: The Motzfeldt Center of the Igaliko Nepheline Syenite Complex, South
Greenland – A major resource of – REE elements. This volume.
8. http://www.ramresources.com.au/
9. Larsen, L.M., Rex. D.C. & Secher. K., 1983: The age of carbonatites, kimberlitic and
lamprophyres from southern West Greenland: recurrent alkaline magmatism during 2500
million years. Lithos 16, 215-221.
10. Secher, K., Heaman, L.M., Nielsen, T.F.D., Jensen, S.M., Schjøth F. & Creaser, R.A. 2009:
Timing of kimberlite, carbonatite, and ultramafic lamprophyre emplacement in the alkaline
province located 64-67N in southern West Greenland. Lithos 112, 400-406.
11. Secher, K. & Larsen, L. M. 1980: Geology and mineralogy of the Sarfartôq carbonatite
complex, southern West Greenland. Lithos 13, 199-212.
12. http://www.hudsonresources.ca/
13. Keulen, N. & Secher, K. 2014: REE-mineralisation in the Sarfartoq Carbonatite, Greenland.
This volume.
14. Hudson commences 2012 work program in Greenland.
15. Knudsen, C. & Buchardt, B. 1991: Carbon and oxygen isotope composition of carbonates
from the Qaqarssuk Carbonatite Complex, southern West Greenland. Chemical Geology 86,
263-274.
16. Knudsen, C. 1991: Petrology, geochemistry and economic geology of the Qaqarssuk
carbonatite complex, southern West Greenland. Monograph Series on Mineral Deposits 29,
110 pp
17. http://www.nunaminerals.com/
18. Steenfelt, A., Hollis, J.A. & Secher, K. 2006: The Tikiusaaq carbonatite: a new Mesozoic
intrusive complex in southern West Greenland. Geological Survey of Denmark and Greenland
Bulletin 10, 41–44.
19. Tappe, S., Steenfelt, A., Heaman, L.M. & Simonetti, A. 2009: The newly discovered Jurassic
Tikiusaaq carbonatite-aillikite occurrence, West Greenland, and some remarks on carbonatite-
kimberlite relationships. Lithos 112, 385-399.
20. http://www.avanaa.com/
21. Harpøth, O., Pedersen, J.L., Schønwandt, H.K. & Thomassen, B. 1986: The mineral occurrences
of central East Greenland. Meddelelser om Grønland, Geoscience 17.

22. Schatzlmaier, P., Schöllnberger, W. & Thomassen, B. 1972: Untersuchung des Vorkommens
an Zirkon und seltenen Erden auf Kote 800 Milneland. Internal report, Nordisk Mineselskab
A/S, 91 pp GEUSDodex # 20669.


The Motzfeldt Centre of the Igaliko Nepheline Syenite

Complex, South Greenland ‐ A major resource of REE
elements
Tapani Tukiainen
Geological Survey of Denmark and Greenland (GEUS)
Department of Petrology and Economic Geology
Øster Voldgade 10
DK‐1350 Copenhagen
Email TT@GEUS.DK
Abstract
The Motzfeldt Centre (1,273 ± 6 Ma) is one of four major alkaline centres belonging to
the Igaliko complex, part of the Mid‐Proterozoic Gardar province of South Greenland.
The Centre covers an area of approximately 150 km2 with excellent 3‐dimensional
exposure. The Motzfeldt Centre is made up of multiple intrusions of syenite. The
syenites were emplaced as two main igneous phases into the Proterozoic Julianehåb
batholith and the unconformably overlying Gardar supracrustal rocks. The apparent
intrusion mechanism was a combination of block stoping, ring fracture and partial ring
dyke formation. The oldest igneous phase, the Motzfeldt Sø Formation, underwent an
extreme in situ differentiation, which resulted in the formation of peralkaline residues
rich in volatile and incompatible elements. The peralkaline residues gave rise to a
complex of late peralkaline sheets of microsyenite and pegmatite, and pervasive
hydrothermal alteration with associated Th‐U‐Nb‐Ta‐Zr‐REE mineralization which
increases in the intensity towards the margins, and especially towards the roof of the
intrusions.
The highest enrichment of REE and Y and the relative enrichment of HREE are confined
to the peralkaline sheet complex in the roof zone of the intrusion. The volumes of the
peralkaline microsyenite with 0.6 – 1.1% TREO is huge and can be estimated to a
minimum of 80 million tons.

Introduction

High‐field‐strength‐elements (HFSE) such as Zr, Hf, Nb, Ta, Y, REE, Th and U are often
enriched in alkaline rocks where a substantial volume of the world’s strategic reserves

Motzfeldt
Centre
Ilímaussaq
intrusion

Figure 1: Geological sketch map of South Greenland1. The location of the Igaliko
Complex, Motzfeldt Centre and Ilímaussaq intrusion is highlighted.

of these elements are found. The mid‐Proterozoic Gardar province in South Greenland
(Fig. 1) representing the products of repeated rift‐related alkaline magmatism9,10 in the
Proterozoic Ketilidian (1,900 ‐1,700 Ma) and Archaean rocks9,10 during the period ~
1,300 – 1,1406,7,8 Ma. There are approximately ten major intrusive complexes of
alkaline rocks of Gardar‐age in southern Greenland. The intense enrichment of the
HFSE‐elements in the agpaitic intrusive units of Ilímaussaq intrusion has been known
and evaluated since 197115. The extensive mineralisation of the HFSE‐elements in the
Motzfeldt Centre of the Igaliko Complex was discovered in late 1970’s4 and Centre has
been a target for geological mapping and exploration for Nb and Ta 13,14.

Geology of the Motzfeldt Centre
The Motzfeldt Centre (Fig. 2) appears to be a classic example of central‐type alkaline
complexes which developed which developed through the successive emplacement of
syenite guided by combination of ring fractures and block subsidence.

The syenite units of the main igneous phase originally comprised a circular body about
18 km in diameter having a shape of “bell jar”. The contact relationships between the
syenites and the syenites/basement rocks in indicate that the emplacement of the
syenites took place passively via permissive stoping.
The Motzfeldt Sø Formation (MSF) is the oldest unit of the centre and now occupies
its outer part. The MSF has been divided into two major concentric zones: namely
inner Nepheline syenite and outer altered syenite. No evidence of these zones as
separate intrusions has been found. The MSF hosts a conspicuous sheet complex of
pegmatite/aplite and peralkaline lujavrite and peralkaline microsyenite.
The Flinks Dal Formation (FDF) occupies the core of the MRS and includes three major
intrusions of phonolitic composition.
The present exposure of syenites is an intricate combination of faulting an topography.
The Centre was affected by two major sets of vertical or nearly vertical faults, one
striking NE‐SW (older) and another approximately E‐W (younger). The fault sets display
both vertical and horizontal component.
The younger E‐W striking faults are characterised by movements in sinistral sense. The
Flinks Dal Fault (Fig. 2) traverses the whole centre with a horizontal component of
about 6 km.
The present altitude of the Eriksfjord Formation/basement unconformity provides a
means to evaluate the magnitude of the vertical movements along the faults. The
Flinks Dal Fault has the most dramatic downthrow of a minimum of 600 metres to the
north.
The Motzfeldt Sø Formation – mineralisation of HFSE‐elements
The texturally and mineralogically highly variable units of the Motzfeldt Sø Formation
(Fig. 3) contain exceptionally high concentrations of Th, U, Nb, Ta, Zr and REE and
volatile components such as F and H2O. Within the inferred roof zone (fig. 4) the
diversity of rock types is most extreme with the highest enrichment of HFSE.
The rock types of the MSF, their mineralogical and chemical characteristics manifest
and extreme internal differentiation of the phonolitic (?) magma. Along with the
progressive crystal fractionation the incompatible elements and volatile components
were increasingly concentrated at the top of the chamber. The build‐up of the volatiles
must have affected the physical and chemical properties of magma/fluid. The
peralkaline magma was able to dissolve large amounts of volatiles12 and there may
have been a gradual transition into a hydrothermal fluid. The formation of the
peralkaline sheet complex could be related to a formation of such a fluid phase.
The MSF and the peralkaline sheets complex are characterised by the pervasive
hydrothermal alteration associated with the subsolidus evolution of the MSF. The
textural relationships of the minerals indicate that the migration involved a continuous
readjustment (precipitation/leaching) of the mineral phases with the proceeding

hydrothermal activity. The proceeding hydrothermal activity was accompanied by the
relative and absolute enrichment the HFSE.

Geologfjeld North East

Motzfeldt
Aries Prospect
RAM Resources Ltd
South East
Motzfeldt

Figure 2: Geological map of Motzfeldt alkaline centre (after Bradshaw2; Jones3;
Tukiainen & al.4). The great numbers of dykes, predominantly alkali trachytes of the
late‐Gardar dyke swarm are omitted for the sake of clarity. The localities referred in
the text are highlighted:

The extent and the intensity of the Th‐U‐Nb‐Ta‐Zr‐REE mineralisation are well outlined
by the high resolution airborne gamma spectrometer survey4.

The mineralogy of the HFSE in the MSF is complex. The REE are dominantly hosted by
REE–carbonates, pyrochlore, monazite and eudialyte (table 1.)

The peralkaline sheet complex was sampled in selected mineralised localities as
vertical profiles traversing from 20 to 200 metres of ground with a sampling density of
reconnaissance character (table 2.).


Formation

Figure 3: Schematic cross section of the Motzfeldt Sø Formation, explanation of
colours same as in Fig. 2

Reserve estimates
The peralkaline sheet complex

The peralkaline sheet complex of the Motzfeldt Centre constitutes a huge reserve of
Zr in the class 1‐2 % ZrO2. The extent and the intensity of the Th‐U‐Nb‐Ta‐Zr‐REE
mineralisation is well outlined by the high resolution airborne gamma spectrometer
survey. Within the inferred roof‐zone, the diversity of rock types is most extreme
and roof zone hosts the highest concentrations of HFSE and REE. Large portions of
the roof‐zone are preserved in North and South East Motzfeldt containing 0.6 ‐ 1.5
% TREO covering 4 and 6 km2 of ground, respectively. Assuming that only the
topmost 50 meters of peralkaline sheet complex is of interest, the rock volumes in
this category are c. 1250 Mt containing 0.6 ‐ 1.5 % TREO corresponding to 75 –
187 million tons of contained REO, respectively.


Eriksfjord Formation
Sheet complex of peralkaline
microsyenite and lujavrite

Figure 4: The peralkaline sheet complex of lujavrite and microsyenite in North
Motzfeldt. The water table of the lake is 162 metres a.s.l., the mountain top is 1500
metres a.s.l. The MSF roof zone below the Eriksfjord Formation indicated with red
colour.

The MSF altered syenite

The enrichments of pyrochlore in the MSF altered syenite have been assessed for their
economic potential for Ta and Nb. The Aries Prospect of the Ram Resources Limited
(Fig. 1) with an exploration potential of 200 – 500 Mt @1800‐2200 ppm Nb205 , 130‐
160 ppm Ta2O5 and 3000‐5000 ppm TREO contained both in Pyrochlore (4‐6 % TREO)
and REE‐carbonate (70 % TREO)
The extraction of Ta and Nb from pyrochlore may also yield significant quantities U,
Th, Zr and LREE.


Table 1: Dominant minerals for Nb‐Ta‐REE‐Zr‐Y‐U in the MSF (Tukiainen11, 1988)
Peralkaline sheet complex

Mineral Nb205 Ta205 REE2O3 ZrO2 Y2O3 UO2 ThO2
Pyrochlore 54.78 1.42 16.1 1.04 ‐ 6.88 0.13 Altered syenite
Columbite 68.13 0.8 ‐ ‐ ‐ ‐ ‐ Altered syenite
Bastnaesite 70.0 Altered syenite
Zircon 46.3 2.53 Altered syenite
Thorite 0.7 5.53 50.58 Altered syenite
Monazite 63.5 3.3 Altered syenite
Eudialyte 2.47 ‐ 5.68 12.21 ‐ ‐ ‐ Lujavrite,
unaltered

Altered Motzfeldt Sø Syenite
Pyrochlore 47.39 6.04 6.04 1.03 3.48 0.34 Altered syenite
Zircon 65.05 2.0 0.1 0.37 Altered syenite
Monazite 64.2 4.2 Altered syenite

Table 2: Content of the HFSE‐elements in selected mineralised localities of the
peralkaline sheet complex of the Motzfeldt Sø Formation

Locality (Map 2) ∑Ce203,La2O3, Nd2O3 Y2O3% U ppm Th ppm ZrO2 % Nb2O5 % Ta2O5%
Geologfjeld 0.1 – 0.3 % 0.06 ‐ 0.1 % 100 – 500 100 – 600 0.6 – 1.5 0.2 – 0.6 % 0.01 – 0.03
Max. 0.9 % Max. 2 % (Max. 1 %) %

North East 0.4– 1.1 % 0.02‐0.05 % 100 – 200 200‐ 300 0.9 – 1.5 0.2 – 0.5 % 0.005 –
Motzfeldt Max. 2 % Max. 1400 Max. 1900 Max. 2 % (Max. 1 %) 0.01%

South East 0.2 – 0.4 % 0.012‐ 200 ‐ 866 300‐4936 1.0 – 1.3 0.2 – 0.3 % 0.01 – 0.03
Mozfeldt Max. 1 % 0.12% Max.9000 Max. 2 % Max. 1 % %


Conclusions
The magmatic differentiation of the Motzfeldt Sø formation resulted in the formation
of peralkaline rocks with the exceptional enrichment of HFSE (Th‐U‐Nb‐Ta‐Zr, Y‐REE)
towards the margins and especially towards the roof of the intrusion.
The preserved inferred topmost roof‐zone of the intrusion in North East and South
East Motzfeldt is likely to host a reserve of c. 1250 Mt with the TREO content varying
from 0.6 ‐ 1.5 % TREO.

References
1. Garde, A.A., Hamilton, M.A., Chadwick, B., Grocott, J. and McCaffrey, K.J.W., The Ketilidian origin
of South Greenland: Geochronology and techtonics, magmatism and fore‐arc accretion during
Palaeoproterozoic oblique convergence. Canadian Journal of Earth Science, 39, 765‐93. (2002)
2. Bradshaw, C., A petrographic, structural and geochemical study of the alkaline igneous rocks of
the Motzfeldt Centre, South Greenland. Unpublished PhD thesis, University of Durham. (1988)
3. Jones, A.P., The petrology and structure of the Motzfeldt centre, Igaliko, South Greenland.
Unpublished PhD thesis, University of Durham. (1980)
4. Tukiainen, T., Bradshaw, C. and Emeleus, C.H., Geological and radiometric mapping of the
Motzfeldt Centre of the Igaliko Complex, South Greenland. Rapport Grønlands Geologiske
Undersøgelse, 102, 78‐83. (1984)
5. McCreath, J.A., Finch, A.A., Simonsen. S.L., Donaldson, C.H., Armour‐Brown, A., Independent ages
of magmatic and hydrothermal activity in alkaline igneous rocks: The Motzfeldt Centre, Gardar
Province, South Greenland. Contrib Mineral Petrol 163, 967‐982. (2012)
6. Blaxland, A.B., Vanbreemen, O., Emeleus, C.H., and Anderson, J.G. (1978) Age and origin of major
syenite centers in Gardar province of South Greenland ‐ Rb‐Sr studies. Geological Society of
America Bulletin, 89, 231‐44. (1978)
7. Upton, B.G.J. and Emeleus, C.H. Mid‐Proterozoic alkaline magmatism in Southern Greenland: The
Gardar Province. In J.G. Fitton, and B.G.J. Upton, Eds. Alkaline Igneous Rocks, Special Publication,
30, p. 449‐71. Geological Society of London. (1987)
8. Upton, B.G.J., Emeleus, C.H., Heaman, L.M., Goodenough, K.M. and Finch, A.A., Magmatism of the
Mid‐Proterozoic Gardar Province, South Greenland: Chronology, petrogenesis and geological
setting. Lithos, 68, 43‐65. (2003)
9. Allart, J.H., Ketilidian mobile belt in South Greenland. In A. Escher, and W.S. Watt, Eds. Geology of
Greenland, p. 121‐51. Geological Survey of Greenland, Copenhagen. (1976)
10. Garde, A.A., Hamilton, M.A., Chadwick, B., Grocott, J. and McCaffrey, K.J.W., The Ketilidian origin
of South Greenland: Geochronology and techtonics, magmatism and fore‐arc accretion during
Palaeoproterozoic oblique convergence. Canadian Journal of Earth Science, 39, 765‐93. (2002)
11. Tukiainen, T., Niobium‐tantalum mineralisation in the Motzfeldt centre of the Igaliko nepheline
syenite complex, South Greenland. In J. Boissonnas, and P. Omenetto, Eds. Mineral deposits within
the European Community, p. 230‐46. Springer‐Verlag, Berlin. (1988)
12. Kogarko, L.N., (1974) Role of volatiles. In H. Søresnsen, Ed. The Alkaline Rocks. Wiley, London. p.
474‐87. (1974)
13. Thomassen, B., The Motzfeldt 87 Project, Final Report. Open file Series 88/1, Grønlands Geologiske
Undersøgelse. (1988)
14. Armour‐Brown.A., (2001) Tantalum exploration: Review of previous exploration, results ofbenefication
studies and recommendations for future work, Angus & Ross plc report 1. (2001)

15. Sørensen, H., Rose‐Hansen, J., Nielsen, B.L., Løvborg, L., Sørensen, E. & Lundgaard, T., The uranium
deposit at Kvanefjeld, Ilímaussaq intrusion, South Greenland. Geology, reserves and beneficiation.
Rapp. Grønlands geol. Unders. 60, 54 pp. (1974).


ALTERATION OF EUDIALYTE AND IMPLICATIONS FOR THE
REE, ZR, AND NB RESOURCES OF THE LAYERED
KAKORTOKITES IN THE ILÍMAUSSAQ INTRUSION, SOUTH
WEST GREENLAND
Anouk Borst1,2, Tod Waight2, Matthijs Smit2, Henrik Friis3, Troels Nielsen1
1
Department of Petrology and Economic Geology, Geological Survey of Denmark and
Greenland, Øster Voldgade 10, 1350, Copenhagen, Denmark
2
Department of Geosciences and Natural Resource Management (Geology Section),
University of Copenhagen, Øster Voldgade 10, 1350 , Copenhagen, Denmark
3
Natural History Museum, University of Oslo, P.O. 1172 Blindern, 0318 Oslo, Norway
amb@geus.dk
Abstract
The layered kakortokites in the southern part of the Ilímaussaq Intrusion are of great
economic interest due to their high concentrations of REE, Zr, Nb and Ta. The
prospective metals are largely contained in eudialyte, a complex sodium‐zirconosilicate
and one of the major cumulus phases. Eudialyte‐group minerals are easily extracted
from the host rock through magnetic separation, and contain 12 wt% ZrO2, 2 % TRE2O3
and 1% Nb2O5 on average. Petrographic investigations show that a large fraction of
eudialyte at Ilímaussaq is replaced by complex aggregates of secondary minerals as a
result of interaction with late‐stage magmatic, presumably Na‐ and F‐ rich aqueous
fluids. The alteration effectively fractionates the major components into the secondary
minerals, producing separate Zr‐, Nb and REE‐ phases, leading to an increased
complexity of the mineralisation and potential ore recovery.

Introduction
The Ilímaussaq Complex in Southern Greenland hosts some of the most evolved
alkaline rock types in the world and is the type‐locality of agpaitic nepheline syenites.
The term agpaitic exclusively refers to peralkaline rocks with a molar (Na+K)/Al ratio
greater than 1.2, containing complex zircono‐ and titanosilicates, such as eudialyte‐
and rinkite‐group minerals, instead of common high field strength element‐bearing
(HFSE) phases like zircon, titanite and ilmenite1,2. Eudialyte represents a prime REE, Nb
and Zr resource provided by these rock types. The mineral can be readily separated
from other minerals (alkali feldspar, arfvedsonite and feldspathoids) by simple
magnetic separation and is highly soluble in acids, providing a potential low‐cost ore
material. Other occurrences of agpaitic rocks are the Khibina and Lovozero complexes
in the Kola Peninsula, Mont Saint‐Hilaire in Canada, the Tamazeght Complex in

Morocco and Norra Kärr in Sweden. All are characterized by extreme enrichment in
alkali metals, halogens (F, Cl, Br) and HFSE (e.g. Zr, Ti, Y, Nb, Ta and REE), and many
host exploitable resources of these elements3.

The Ilímaussaq Complex has the potential for two types of world‐class multi‐element
magmatic ore deposits; 1) the Kringlerne deposit, studied here, which occurs in
cumulates forming the lower part of the intrusion4, and 2) the Kvanefjeld deposit in
the most evolved lujavrites exposed in the Northern part of the intrusion5. The
Kringlerne deposit comprises a ~500 m thick series of rhythmically layered amphibole‐
, eudialyte‐ and feldspar‐ rich nepheline syenites (black, red and white kakortokites,
respectively), which is currently being explored as a potential REE, Zr, Nb and Ta
economic deposit. The resource estimates are 4.3 Bt with an average grade of 1.8 %
ZrO2, 0.2 % Nb2O5 and 0.5 % TREO. The enormous tonnages and relatively high
proportion of heavy relative to light rare earth elements (HREE/LREE) in eudialyte
(approximately 1:3) compared to currently exploited REE deposits, make the deposit
especially attractive3,6. Proposed on‐site processing involve coarse crushing followed
by magnetic separation, producing concentrates of 1) eudialyte, 2) arfvedsonite‐
aegirine and 3) a mixture of felsic minerals.

Field observations combined with petrographic and electron‐microprobe (EPMA)
analyses show that a significant amount of eudialyte in the kakortokites has been
altered forming a suite of secondary minerals including catapleiite (Na‐Zr silicate),
aegirine, pectolite, analcime, alkali‐feldspar, nacareniobsite‐(Ce) (REE‐Nb silicate),
britholite (REE‐phosphosilicate), monazite (REE‐phosphate) and pyrochlore (formulas
provided in Table 1), previously described in detail by Karup‐Møller et al.7,8. The bulk
composition of pseudomorphosed eudialyte appears to be preserved, but with major
components, notably Zr, REE and Nb, redistributed among the newly formed minerals.
The goals of this study are to improve the general knowledge of the relative timing and
processes of crystallisation and alteration, the de facto distribution of elements of
economic interest and possible implications for beneficiation of the ore.

Table 1: Selection of relevant minerals mentioned in text
Mineral name Formula
aegirine NaFe(Si2O6)
arfvedsonite Na3(Fe,Mg)4FeSi8O22(OH)2
analcime NaAlSi2O6H2O
pectolite NaCa2Si3O8(OH)
eudialyte Na15Ca6(Fe,Mn,Na)3(Zr,Ti)3(Si,Nb)2Si24O72(O,OH,H2O)3(Cl,OH)2
catapleiite (Na,Ca)2ZrSi3O9.2H20
zircon (Zr,REE)SiO4
apatite Ca5(PO4)3(F,Cl,OH)
britholite‐(Ce) (Ca,Ce)5(SiO4,PO4)3(OH,F)

Ca‐rich A1 HCa3REE6(SiO4)6(F□)
Ca‐poor A1 (Fe,Mn,Ca)1.5REE6Si6FO22
monazite (Ce,La,Nd,Th)PO4
mosandrite – rinkite group (Na,Ca)3(Ca,Ce)4(Ti,Nb,Zr) (Si2 O7)2 (O,F)4
nacareniobsite‐(Ce) Na3Ca3(Ce,La)Nb(Si2 O7)2 O F3
pyrochlore (Na,Ca)2Nb2O6(OH,F)
fergusonite‐(Ce) (Ce,La,Y)NbO4
fersmite (Ca,Ce,Na)(Nb,Ta,Ti)2(O,OH,F)6

Eudialyte Group Minerals
Eudialyte group minerals (EGM), the most prominent index minerals for agpaitic rocks,
encompass a large compositional range and are inferred to crystallise at various stages
in the magmatic history9. Eudialyte group minerals have a complex structure, with a
general structural formula of N15‐16M(1)6M(2)3Z3M(3)M(4)[Si24O72]OH2‐6X2, where N, M
and Z refer to different crystallographic sites incorporating variable amounts of Na, Ca,
Fe, Mn, Sr, REE, Nb, Si, Zr and Ti as well as volatile components (H2O, Cl, F, OH)10,11. In
the kakortokites EGM are generally Fe‐rich, approaching eudialyte sensu stricto
compositions, with a general formula of
Na15Ca6(Fe,Mn)3Zr3Si(Si25O72)(O,OH,H2O)3(Cl,OH)2, with the major part of the REE and
Nb situated in the M1 and M3 site, respectively12. Owing to their great compositional
variability and sensitivity to hydrothermal re‐equilibration, EGM typically record late‐
stage magmatic and hydrothermal processes in agpaitic systems13‐15. These processes
could result in partial or complete decomposition of the eudialyte into complex
aggregates of secondary minerals8,14,15.

Analytical Methods
Eudialyte and its decomposition assemblages were analysed in approximately 30
samples, including black, red and white kakortokites from layers ‐11 to + 164, using
optical microscopy, back‐scatter electron imagery (BSE) and wavelength dispersive
electron microprobe analyses using a JEOL JXA‐8200 at University of Copenhagen. All
analyses were run at an acceleration voltage of 15 kV, beam current 15 nA and a beam
diameter of 5 or 20 µm, using a set‐up similar to that described by Karup‐Møller et al.
(2010, 2013). Counting times range from 10‐60 seconds, depending on the abundance
and volatility (Na) of the respective elements. Background counting times were half of
the peak counting times.

Petrographic observations
Primary eudialyte description

The eudialyte‐group minerals form well‐developed idiomorphic crystals with an

average grain size of 0.5 ‐ 1 mm. In all three varieties of kakortokite (black, red and
white) EGM are part of the early crystallising assemblage4,12. In black kakortokites EGM
occur between, or are fully enclosed by arfvedsonite crystals. In white kakortokite
eudialyte crystals are generally interlocked between gravitationally aligned laths of
feldspar and grains of nepheline. Intact EGM exhibit complex sector‐zoning, fine
oscillatory zoning (sub‐µm) as well as concentric core to rim zonation (Fig. 1a). Hour‐
glass sector zoning is present in most crystals and represents a crystallographically
controlled feature, presumably caused by preferential incorporation of certain
elements on a specific set of growth surfaces16. Sector‐zoning is associated with minor
differences in Nb and Ce (LREE) contents (< 0.2 wt%) between light and dark sectors.
Rims showing uniformly high or low backscatter‐index typically occur along specific
faces of primary sector‐zoned eudialyte crystals (Fig. 1a). These rims are distinctly
enriched or depleted in REE and presumably reflect changes in the REE‐budget of the
evolving (interstitial) melt in the final stages of crystallisation.
Alteration paragenesis
A significant proportion of the cumulus eudialyte was subjected to late‐stage
replacement reactions. First‐stage alteration occurs as irregular replacement or
symplectite formation along margins and cracks and is associated with crystallisation
of fine‐grained anhedral crystals of the Na‐zirconosilicate catapleiite (mineral formulas
provided in Table 1), as well as µm‐sized Ce‐ and Nb‐rich phases. Upon further
alteration an increasing volume of eudialyte is consumed until the grain is completely
replaced by secondary phases (Fig. 1b,c,d). Relics of eudialyte in some pseudomorphs
retained their original composition, although they are locally associated with minor
loss of Na and Cl. In pseudomorphs representing fully decomposed eudialyte, several
assemblages of secondary minerals are distinguished. Most assemblages are
dominated by catapleiite as the main replacing Zr‐phase (35 wt% ZrO2). Catapleiite
compositions range from a nearly pure Na‐end‐member to being Ca‐rich (up to 0.6
apfu Ca). Aegirine is the main Fe‐bearing phase in nearly all pseudomorphs.
Nacareniobsite‐(Ce) is found in most assemblages and hosts most of the Nb and REE
contents originally present in the eudialyte (Fig. 1b). Both aegirine and nacareniobsite‐
(Ce) have near‐ideal end‐member compositions. This is in contrast to their magmatic
counterparts, which show significant solid‐solution trends in the aegirine‐augite17 and
rinkite‐mosandrite‐nacareniobsite‐(Ce) series18.


Figure 1: Back‐scatter electron (BSE) images of eudialyte and pseudomorphs after
eudialyte; (A) Euhedral eudialyte crystal (Layer 4 Black) with 8‐fold sector zoning, µm‐
scale oscillatory zoning and multiple concentric rims; (B) Common eudialyte
pseudomorph assemblage (Layer 1 red), with catapleiite (Cat), interstitial albite (Ab),
prismatic nacareniobsite‐(Ce) (Ncr), acicular Ca‐poor A1‐silicate (A1, see text) and
aegirine (Aeg). Pseudomorph surrounded by arfvedsonite (Arf) and exsolved feldspar
(Ab and Ksp); (C) Part of pseudomorph (Layer 3 white) with monazite (Mnz), apatite
(Ap), fluorite (Fl), catapleiite, aegirine and albite; (D) Part of pseudomorph (Layer 2
white) with unknown Ba‐REE‐Nb‐phase (Brn), catapleiite, interstitial aegirine,
analcime (Anl) and nacareniobsite‐(Ce) between the catapleiite plates.
Analcime, albite and/or potassium feldspar are found in most assemblages, filling the
interstitial space between catapleiite and other phases (Fig. 1b,c,d). Pectolite is also
common and is often found intergrown with aegirine. Rare earth elements are
concentrated into a wide range of minerals, many uncommon, and generally present
as µm‐sized crystals8. The most commonly observed mineral is a Ca‐rich or Ca‐poor
REE‐silicate containing up to 65% TREO. These REE‐minerals were first described by
Karup‐Møller et al. (2010), as new REE‐phases with a presumed apatite structure

(possibly P‐poor members of the britholite‐group) and were referred to as Ca‐rich and
Ca‐poor A1, respectively. The REE‐minerals occur as clusters of acicular crystals
randomly distributed in the pseudomorphs (Fig. 1b). More P‐rich varieties,
approaching compositions of britholite‐(Ce) (following a REE3+ + Si4+  Ca2+ + P5+
coupled substitution scheme), are also present in several decomposition assemblages.
A Ba‐REE silicate (8 wt% BaO, 2% Nb2O5, 55% REO (La+Ce+Nd) and 22 % SiO2) is found
in six samples. The mineral is texturally indistinguishable from the more common REE‐
silicates described above and forms similar aggregates of acicular crystals (Fig. 1d).
Several REE‐phases typically occur together in a single pseudomorph. At higher P
contents, pseudomorphs contain monazite as the main REE‐bearing phase, associated
with secondary apatite and fluorite. Monazite‐bearing assemblages have only been
observed in two samples of the present study, both in white kakortokites (Fig. 1c).
Because magmatic eudialyte contains virtually no P and only ppm‐level F and Ba, these
components are presumably introduced by the alteration‐mediating fluids. The
absence of Cl‐bearing phases in the pseudomorphic assemblage suggests a significant
net loss in Cl from these domains, which is consistent with the findings of Karup‐Møller
et al. (2010). Niobium mainly partitions into nacareniobsite‐(Ce), although some
pseudomorphs contain additional Nb‐phases, i.e. pyrochlore, fersmite or fergusonite
as small (<5µm) dispersed grains7,8,14. The alteration of eudialyte is associated with
zeolitisation (primarily analcime) of feldspars and feldspathoids and replacement of
magmatic arfvedsonite by secondary aegirine along the margins of the
pseudomorphosed eudialyte grains (Fig. 1b).

Extent of alteration
The alteration of eudialyte ranges from partial replacement along the margin and
cracks of EGM crystal to complete pseudomorphic replacement. In hand specimen,
altered EGM are easily recognised by their pale orange‐grey colour, which is distinctly
different from the glassy pink original eudialyte. Eudialyte pseudomorphs cluster in
zones of intense alteration, but completely decomposed grains may equally be in
direct contact with fully preserved crystals. No petrographic link is observed between
the distribution of the pseudomorphically replaced eudialyte and the proximity to
cracks or veins that could have provided a pathway for percolating fluids. It is therefore
suggested that the alteration occurred as a result of reactions with late‐stage
magmatic fluids present at crystal boundaries and in interstitial spaces at the final
stages of solidification. The varying assemblages (monazite vs REE‐silicate A1) could
reflect variations in the composition of the evolving interstitial magmatic fluids, which
may be linked to the local mineralogy e.g. the presence of magmatic apatite, fluorite
or sodalite. The extent of eudialyte alteration is independent of lithology and is
extremely variable throughout the kakortokites, complicating precise evaluation of the
degree and mode of alteration and net loss or gain of elements. Current observations

suggest that up to 20% of the primary eudialyte is altered, i.e., 20% or more of the REE,
Nb and Zr potential is contained in minerals not currently targeted for processing.

Discussion
Extensive auto‐metasomatism and pseudomophic replacement of primary magmatic
minerals due to a continuous transition from melt to hydrothermal fluid in the final
stages of solidification is a common feature of agpaitic complexes19. Late magmatic
fluids are dominated by H2O‐CO2 or CH4 and generally enriched in Na, Cl and F, allowing
them to complexate and remobilise HFSE and REE19,20. Remobilisation of incompatible
elements by such fluids can both up and down‐grade the resource potential, and is
hence an important aspect in the paragenesis of alkaline rocks.
This study clearly demonstrates the wide‐spread decomposition of eudialyte into
complex aggregates of secondary minerals. These aggregates are invariably dominated
by catapleiite‐group minerals, suggesting alteration by Na‐rich, hydrous fluids. Zircon‐
bearing assemblages have not been described from the current sample set and are
apparently only related to severely altered kakortokites and pegmatitic rock types
within the complex7,8.
Late‐stage alteration of eudialyte has also been described from North Qoroq,
Greenland15; Pilanesberg, South Africa14 and Sushina, India21, indicating a large variety
in possible eudialyte replacement products depending on both the primary
composition of the eudialyte and the physico‐chemical parameters of the
metasomatising fluids. An overview of reported eudialyte decomposition assemblages
is provided by Mitchell and Chakrabarty (2012). In all cases, the bulk compositions of
the replacing phases reflect the initial composition of the primary eudialyte, and
distinct fractionation of the HFSE and REE into the secondary phases is observed. In
the case for Kringlerne, we infer that the auto‐metasomatic alteration of eudialyte is
not associated with large‐scale remobilisation of the elements of interest and does not
affect the overall ore grade of the deposit. Irrespectively, the increased complexity of
the secondary REE‐, Nb, and Zr‐ paragenesis and major reduction in grain size (< 10
µm) will have a significant impact on the recovery of the ore metals.
An important step in optimizing ore recovery is to include the secondary REE, Nb and
Zr assemblages in the magnetic separation process. It is currently unknown how the
alteration products of eudialyte are distributed during magnetic separation and
whether they are concentrated in the eudialyte separate (i.e. the REE‐Nb‐Zr ore) or
the other separates. Two aspects are thus worth considering in future processing tests
of eudialyte concentrates: 1) the magnetic susceptibility of the secondary ore minerals
and 2) the dissolution kinetics of the secondary phases in relation to the acid treatment
procedures used for eudialyte. An example provided by Zakharov et al. (2011)
indicated that low Zr recovery from Lovozero eudialyte concentrates was related to
the presence of acid resistant Na‐Zr‐phases, which formed as a consequence of late‐
stage magmatic alteration of eudialyte. Similar complications may be expected for Zr,

REE and Nb recovery from Ilímaussaq eudialyte concentrates with secondary phases
that are insoluble during eudialyte dissolution procedures. Subsequently, acid
treatment should be optimised for digestion of secondary ore bearing phases in
addition to eudialyte.

Conclusions
The kakortokites at the Ilímaussaq Intrusion provide a world‐class REE, Zr and Nb
deposit which is unique in its size, exposure and mineralogy. Partial decomposition of
the primary ore mineral eudialyte through interaction with late‐stage magmatic fluids
formed complex secondary REE‐, Nb‐ and Zr‐ phases, such as catapleiite,
nacareniobsite‐(Ce), britholite‐(Ce), monazite and fergusonite‐(Ce). The alteration led
to increased dissemination and fractionation of the prospective metals in the host
rock, while leaving the overall ore grade intact. Further studies on the mineralogical
properties and extraction techniques of the secondary REE‐, Nb‐ and Zr‐ minerals will
be of great value for optimal beneficiation of the kakortokites, with potential
applications to other agpaitic ore deposits.
References

1. Sørensen, H., “The agpaitic rocks; an overview”, Mineralogical Magazine, 1997. 61(4): p. 485‐498.
2. Marks, M.A.W., et al., “The Mineralogical Diversity of Alkaline Igneous Rocks: Critical Factors for
the Transition from Miaskitic to Agpaitic Phase Assemblages”, J. Pet., 2011. 52(3): p 439‐455.
3. Sørensen, H., “Agpaitic nepheline syenites: a potential source of rare elements”, App. Geochem.,
1992. 7(5): p. 417‐427.
4. Bohse, H., C.K. Brooks, and H. Kunzendorf, “Field observations on the kakortokites of the
Ilimaussaq Intrusion, South Greenland, including mapping and analyses by portable X‐ray
fluorescence equipment for zirconium and niobium”, Rapport GGU, 1971. 38: p 43.
5. Sørensen, H., “Brief introduction to the geology of the Ilimaussaq alkaline complex, South
Greenland, and its exploration history [Kvanefjeld]”, Geo. Greenland Surv. Bul., 2001. 190: p 7‐23.
6. Bailey, J.C., et al., “Geochemical overview of the Ilimaussaq alkaline complex, South Greenland”,
Geo. Greenland Surv. Bul., 2001. 190: p 20.
7. Karup‐Møller, S. and J. Rose‐Hansen, “New data on eudialyte decomposition minerals from
kakortokites and associated pegmatites of the Ilimaussaq complex, South Greenland”, Bul. Geol.
Soc. Denmark, 2013. 61: p. 47‐70.
8. Karup‐Møller, S., J. Rose‐Hansen, and H. Sorensen, “Eudialyte decomposition minerals with new
hitherto undescribed phases from the Ilimaussaq complex, South Greenland”, Bul. Geol. Soc.
Denmark, 2010. 58: p. 75‐88.
9. Schilling, J., et al., “The compositional variability of eudialyte‐group minerals”, Min. Mag., 2011.
75(1): p. 87‐115.
10. Johnsen, O. and J.D. Grice, “The crystal chemistry of the eudialyte group”, Can.Min., 1999. 37: p.
865‐891.
11. Pfaff, K., et al., “A fast and easy‐to‐use approach to cation site assignment for eudialyte group
minerals”, Neues Jahrbuch fur Mineralogie, 2010. 187(Abhandlungen): p. 69‐81.
12. Pfaff, K., et al., “Chemical and physical evolution of the ‘lower layered sequence’ from the
nepheline syenitic Ilímaussaq intrusion, South Greenland: Implications for the origin of magmatic
layering in peralkaline felsic liquids”, Lithos, 2008. 106(3–4): p. 280‐296.

13. Schilling, J., et al., “Reconstruction of magmatic to subsolidus processes in an agpaitic system using
eudialyte textures and composition: a case study from Tamazeght, Morocco” Can. Min., 2009.
47(2): p. 351‐365.
14. Mitchell, R.H. and R.P. Liferovich, “Subsolidus deuteric/hydrothermal alteration of eudialyte in
lujavrite from the Pilansberg alkaline complex, South Africa”, Lithos, 2006. 91(1–4): p. 352‐372.
15. Coulson, I.M., Post‐magmatic alteration in eudialyte from the North Qoroq centre, South
Greenland. Min. Mag., 1997. 61(1): p. 99‐109.
16. Watson, E.B. and Y. Liang, “A simple model for sector zoning in slowly grown crystals: Implications
for growth rate and lattice diffusion, with emphasis on accessory minerals in crustal rocks”, Am.
Min. 1995. 80: p. 1179‐1187.
17. Larsen, L.M., “Clinopyroxenes and coexisting mafic minerals from alkaline Ilimaussaq Intrusion,
South Greenland”, J. Pet. , 1976. 17(2): p. 258‐290.
18. Rønsbo, J.G., et al., “Rinkite–nacareniobsite‐(Ce) solid solution series and hainite from the
Ilímaussaq alkaline complex: occurrence and compositional variation” Bul. Geol. Soc. Denmark,
2014. 62: p. 1‐15.
19. Salvi, S. and A.E. Williams‐Jones, “Alkaline Granite‐Syenite Deposits, in Rare‐Element
Geochemistry and Mineral Deposits”, R.L. Linnen and I.M. Samson, Editors. 2005, GAC Short
Course Notes: Geol. Ass. Can.. p. 315‐341.
20. Salvi, S., et al., “Hydrothermal mobilization of high field strength elements in Alkaline Igneous
Systems: Evidence from the Tamazeght complex (Morocco)”, Econ. Geol. 2000. 95(3): p. 559‐575.
21. Mitchell, R.H. and A. Chakrabarty, “Paragenesis and decomposition assemblage of a Mn‐rich
eudialyte from the Sushina peralkaline nepheline syenite gneiss, Paschim Banga, India”, Lithos,
2012. 152(0): p. 218‐226.
22. Zakharov, V.I., et al., “Causes of insufficient recovery of zirconium during acidic processing of
lovozero eudialyte concentrate”, Russ. J. Non‐Ferrous Metals, 2011. 52(5): p. 423‐428.


REE‐ MINERALS IN CARBONATITE, ALKALINE AND

HYDROTHERMAL ROCKS, NORTHERN AND CENTRAL
FINLAND
Thair Al‐Ani 1* and Olli Sarapää2
1 Geological Survey of Finland, P.O. Box 96, FI‐02151 Espoo, Finland
2 Geological Survey of Finland, P.O. Box 77, FI‐96101 Rovaniemi, Finland
Email: thair.alani@gtk.fi, olli.sarapaa@gtk.fi

Abstract

REE‐rich minerals were identified and analyzed by electron microprobe from different
targets located in the northern and central Finland. Both primary and hydrothermal
minerals were found namely: phosphates (monazite‐Ce), fluor‐carbonates
(bastnaesite‐Ce), hydrated carbonates (ancylite‐Ce), hydrated aluminium silicates,
(allanite), oxides (fergusonite) and U‐Pb rich minerals. Sokli Jammi‐ Kaulus carbonatite
veins are enriched in LREE, P, F, Sr and Ba hosting in ancylite, bastnaesite, apatite and
monazite.
Allanite‐(Ce) and fergusonite (Y) are abundant in alkaline gneiss of the Katajakangas
REE‐occurrence. The Korsnäs Pb‐REE deposit includes apatite with monazite inclusions,
calcio‐ancylite and bastnasite. The Mäkärä‐ Vaulo REE‐prospect in arkosic gneisses is
dominated by monazite, allanite and xenotime. Albitites at Enontekiö contain
bastnaesite, monazite, allanite, xenotime and U‐rich minerals includes davidite,
masuyite and sayrite.
The Honkilehto Au‐Co‐S‐mineralization at Kuusamo is characterized by U‐rich minerals
with bastnaesite and allanite. The results obtained provide vital insights into the
mineralizing processes associated with REE‐prospects in northern and central Finland.

Introduction

Economic REE deposits are not known in Finland. However, REEs were extracted in
the 1960’s century as a by‐product in the fertilizer production from the apatite
concentrates of the Kola Peninsula and the Korsnäs Pb mine in western Finland (1). In
addition, small Nb‐ REE deposit, related with alkaline gneissic granite, from Otanmäki,
Central Finland has been reported (2). During 2009‐2012, the REE potential of Finland
was studied by Geological Survey of Finland and potential source areas were identified
according to the existing data and preliminary field studies (Fig. 1). The Fennoscandian
Shield hosts several P‐REE deposits in the Devonian Kola alkaline province in NW Russia

and Finland (3, 4). The Sokli carbonatite complex (total area 20 km2) is part of the
province (5, 6) and recent geochemical studies have confirmed that the area hosts high
REE potential. This study has been focused on the fenite aureole and associated late‐
stage cross‐cutting carbonatite dykes that seem to have the highest potential for REE
mineralisation in the Sokli area (7). The REE‐rich accessory minerals bastnaesite,
monazite, allanite, fergusonite and xenotime play a key role in the storage and mobility
of geochemically important trace elements (LREE, Y Th and U) and are relatively
common accessory minerals in carbonatite veins and fenites.

Palkiskuru
#
*
MäkäräVaulo
#
*#
*
Legend Sokli
#
# Deposit
#
* Mineralisation
Caledonian tectonic units
Mainly Neo- and Mesoproterozoic sediments
Mesoproterozoic rapakivi granites
Paleoproterozoic igneous rocks
Honkilehto
Paleoproterozoic schists #
*
Archean rocks
Katajakangas
#
Korsnäs
#

Figure 1: Location map of the studied areas.


In this study, backscattered electron (BSE) imaging and electron microscope analysis
(EDS) are used to document various mineralogical characteristics and paragenesis of
various types of REE minerals from different localities in Finland. The principal REE‐
minerals in studied rocks are phosphates (monazite‐Ce), carbonates (bastnaesite‐Ce),
hydrated carbonates (ancylite‐Ce), hydrate aluminum silicates (allanite), and oxides
(fergusonite). Other REEs such as britholite, thorite, carbocernaite and columbite have
been reported from some of the studied samples.

The research was circled around three objectives: (1) to study the REE, Y, Th and U
composition of the major minerals; (2) to learn the nature, abundance, composition,
grain‐size distribution, textural relationships and associations of REE, Y, Th, U‐rich
accessories; (3) to determine the relative contributions of accessory minerals to REE,
Y, Th and U.

Samples and analytic methods

The samples were collected from drill cores and outcrops of 6 different targets (Table
1). More than 50 polished thin sections were prepared from representative specimens
of REE‐mineral assemblages and examined under a polarizing microscope. The mineral
compositions were analyzed by scanning electron microscope (SEM) JEOL JSM 5900
LV, at the Research laboratory of the Geological Survey of Finland (GTK), Espoo.
Mineral chemistry was investigated by standard electron microprobe techniques on
polished sections, using CAMECASX100/LKP at the Research laboratory, GTK, Espoo.
The accelerating voltage was 15keV with beam current of 20nA. In order to detect
fluorine in the analysis, spot size was kept 5 μm, whereas 1 μm was used where
fluorine was not desirable.

Mineralogy and mineral chemistry

More than twenty REE‐bearing minerals have been identified in the studied samples
from different target areas in Central and North Finland. The mineralogy and
paragenesis of accessory REE‐rich minerals and the associated minerals are
summarized in Table (1). Representative EPMA analyses of the REE‐bearing minerals
are presented in Table (2). It is obvious Ce predominates over both La and all other
REE in most of the studied samples.

Jammi and Kaulus carbonatite veins in Sokli complex


The Sokli carbonatite (ca. 360‐380 Ma) in northeastern Finland is part of the Kola
alkaline province and hosts an unexploited phosphate deposit enriched in Nb, Ta, Zr,
REE and U (4,8). The carbonatite intrusion consists of a magmatic carbonatite core,
which is surrounded by a metacarbonatite and a wide fenite aureole, altogether about
9 km in diameter. The late‐stage carbonatite vein dykes in the central fracture zone
and in the fenite zone have a high potential for REE mineralisation (8, 9). The REE
minerals that occur in Jammi and Kaulus carbonatite veins are almost entirely LREE
ancylite‐(Ce), calcioancylite‐(Ce), monazite, allanite and bastnaesite (Ce). Ancylite‐(Ce)
is the most common in carbonatite veins and occurs as coarse grained phenocrysts
with an average diameter of 300 μm. Calcioancylite (Ce) is commonly associated with
barite, strontianite and pyrite (Fig. 2a).

Table 1: The main REE‐minerals in studied samples.

Locality Rock type REE‐mineral phases
Sokli Jammi Carbonatite F‐apatite, Sr‐apatite, monazite, bastnaesite,
and Kaulus veins ancylite, strontianite, baryte
Katajakangas Alkaline‐gneiss Allanite, monazite, ancylite, bastnäsite, parasite,
fergusonite, euxenite columbite‐tantalite and
pyrochlore
Korsnäs Carbonatite Apatite, monazite, carbocernaite, calcio‐ancylite,
bastnaesite, and barytocalcite.
Mäkärä Arkosic gneiss Monazite, allanite and xenotime
Enontekiö Albite‐ Bastnaesite, monazite, allanite, xenotime,
Palkiskuru carbonate rock davidite, masuyite and /or sayrite.
Honkilehto Carbonate‐ Bastnaesite, allanite, davidite, U‐Pb minerals, U‐Si
mica schist minerals

Monazite (Ce) occurs most commonly in the form of microcrystalline, sporadic,
isolated equidimensional crystals and associated mainly with apatite. The crystal habit
of bastnaesite and allanite in the studied carbonatites appears to be acicular or needle‐
shaped forming either in radial accumulations or intricate cross‐cutting grids within a
variety of minerals such as albite and dolomite(Fig.2b). Apatite in late carbonatite
veins contains significant amounts of REE2O3, which indicates that the rare earth
elements could be a by‐product in the future phosphate production at Sokli.

Katajakangas

The Katajakangas Nb‐REE deposit is located within alkaline gneissic granite at
Otanmäki, central Finland. The Nb‐REE mineralisation consists of a few metres wide,
narrow lenses or layers in sheared quartz‐feldspar gneiss with riebeckite and alkaline

pyroxene. These narrow mineralised zones contain high concentrations of Nb, Zr, Y, Th
and lanthanides, with an estimated Nb‐YREE resource of 0.46 Mt with 2.4% RE2O3,
0.31% Y2O3 and 0.76% Nb205. The Zr and Th contents in the occurrence range from
0.7% to 1.5% and 0.1% to 0.2%, respectively (10). Drill core samples have relatively
high HREE contents compared with samples from carbonatites (7).
New mineralogical data, which is obtained in this study, indicate that REE are present
in allanite, monazite, ancylite (Ce), bastnäsite (Ce), parasite (Ce), whereas, Y is present
in, fergusonite (Y), euxenite (Y) and yttroclumbite. Nb is present in columbite and
fergusonite, whereas, U‐Th occurs in uranpyrochlore (betafite) and in yttrilite (Y, TH).
Electron microprobe (EMPA) data indicates that ferugsonite‐(Y) host Nb and forms
irregulargrains that vary in size from 100µm to more than 400 µm. Fergusonite is not
chemically homogeneous, many crystals display growth zoning with bright domains,
(Fig. 2c). The bright domains show 2.5% U and high contents of Y (23%) and Nb (38%)
as fergusonite (Y) mineral phase (Table 2). Allanite occurs as subhedral to anhedral
crystals and exhibits oscillatory zoning (Fig. 2c). Zoning profiles of allanite (Ce) show
increase in Ce2O3, La2O3 and Nd2O3, whereas decrease in CaO, FeO and Al2O3 from
core to rim (Table 2). Allanite crystals also show acicular crystals and aggregates of
radiating individual crystals (Fig. 2d).

Korsnäs Pb‐REE deposit

The area is dominated by metasedimentary rocks such as mica gneiss calc‐silicate
gneiss and marble, which are recently interpreted as magmatic carbonatite,
composing of coarse grained calcite, feldspar, diopside, REE‐bearing apatite, monazite,
barite, britholite and ancylite. Two groups of sample from the Korsnäs are
represented. The first group is calcite‐marble or carbonatite characterized by
enrichment in volatile‐bearing REE‐carbonates, ancylite‐(Ce) and bastnäsite‐(Ce), with
traces of monazite. Analyses of multigrain calico‐ancylite grains are rich in LREE as Ce
(~45%), La (~25%) and Nd (~12%). whereas bastnäsite contains 21–75 wt% REE2O3,
1.3‐7.5 F and 20‐77% CaO.
Back‐scattered electron (BSE) imaging of multiple grains of calcio‐ancylite revealed the
presence of Ca‐ancylit grains disseminated through calcite (Fig. 3a). Ancylite varies
considerably in form. Most crystals are subhedral, some are euhedral and in few cases
form fine aggregates along bastnaesite. The second group of samples marked by
abundant crystals of apatite with exsolution‐induced domains of monazite grains, and
the monazite contains 49 wt% REE2O3, P2O5 28 wt% and CaO 16 wt%. Back‐scattered
electron (BSE) imaging of studied samples show that monazite and britholite inclusions
occur both randomly and in large clusters of monazite grains included within apatite
grains (Fig. 3b).


Table 2: Electron microprobe analyses of REE‐rich minerals from studied areas.

Mineral Allanite Ancyli Bastnäs Ceri Davidit Euxenit Fergusoni Monaz
chemist Katajakan te ite te e e te ite
ery gas Kaulu Mäkärä Tan Enonte Honkile Katajakan Jammi
s a kiö hto gas
Belt
SiO2 24.1 6.5 5.8 2.1
Al2O3 7.5 1.7
FeO 6.6 1.1 17 1 1.5
TiO2 46.2 1.3
MgO 2.9
CaO 10.4 1.8 4.2 3.2 5.1
P2O5 2.5 0.3 1.1 22.3
SrO 21.9 4
BaO 0.5
PbO 1.5
Cr2O3 11.3
V2O3 3.1
Nb2O5 37.6 39.6
Ta2O5 1 0.2
ThO2 3.1 0.7 0.1 2 6.1
UO2 5.6 4.6 2
Y2O3 0.7 1.5 19.5 23.9
La2O3 13.6 14.4 16.5 3.1 15.5
Ce2O3 24.9 30.6 35.1 76.8 2.8 33
Nd2O3 9 8.6 10.5 1.4 13.5
Pr2O3 2.5
Gd2O3 0.7
SmO 0.8
F 0.6 6.2 1.4 0.9
Total 99.12 74.4 77.0 91 92.1 93 91 99.5


Figure 2: BSE images of REE‐bearing and associated minerals in studied rocks (a)
Clusters of barite, strontianite (white) and ancylite(grey) within calcite, (b) acicular
crystals growth of bastnaesite within apatite and Sr‐calcite, (c) irregular fergusonite (Y)
grains, the bright core marks enrichment of U and/orTh, (d) acicular crystals of allanite
within dolomite.

Mäkärä‐Vaulo

Selection of the Mäkärä and Vaulo areas as REE‐Au‐target in the Tana Belt was based
on LREE‐ and HREE‐anomalies in regional till and bedrock geochemistry and Au
intersections in previous drill holes. Earlier studies indicate that the hydrothermal
quartz‐haematite‐pyrite veins in Mäkärä are generally narrow (1 mm ‐ 2 m) and
formed as tensional fractures (11). The anomalous REE contents in the Tana belt can
be explained by the occurrence of bastnaesite, cerite, allanite and fine xenotime grains
in arkosic gneisses. The REE composition of the minerals is characterized by
enrichment in Ce, La and to a lesser extent Nd. The total REE content decreases from
75 wt % in cerite (Ce), 60% in bastnaesite to 25 wt % in allanite.

SEM –BSE imaging of arkosic gneiss core samples from Mäkärä and Vaulo reveals that
xenotime occurs as scattered grains within Ti‐bearing minerals and as overgrowths on
zircon. Allanite as the main REE‐bearing mineral in the studied samples, occurs as
spheroidal aggregates or as fillings vugs or cavities wiz and albite, and the chlorite and
bastnaesite replacing the allanite grains (Fig. 3c). Electron microprobe analyses EPMA
along the opened fractures in some spheroidal aggregates, have shown needle‐like
bastnaesite crystals and fine apatite grains filling these spaces (Fig. 3d).

Figure 3: BSE images of REE‐bearing and associated minerals in studied rocks. (a) Ca‐
ancylite grains disseminated through calcite, (b) discrete and coarse crystal of apatite
with exsolution‐induced domains of monazite clustering, (c) vug‐filling texture;
chlorite altered to allanite (in the rim) and later to bastnaesite in the center(light grey),
(d) euhedral to subhedral bastnaesite grains filling the vugs and fractures within
quartz.


Enontekiö

The Enontekiö area is characterised by variable rock types; albitites, syenite and
carbonatite veins. In this study four samples were selected for chemical and
mineralogical analyses. The carbonatite vein sampled contains 2.8% P2O5, 0.45% REE
and 256 ppm Nb and characterized by high content of REE minerals as allanite,
ancylite, bastnaesite and xenotime. The albitites samples characterized by high
content of U‐Pb minerals such as davidite‐ (Ce), mausuyite and sayrite as well as high
content of apatite, allanite and monazite. The allanite presents as fine‐grained
aggregates associated with monazite, apatite and zircon and exhibits a well‐developed
chemical zonation with Fe‐ and REE poor core and an Fe‐ and REE‐rich margin. Despite
the good abundance of allanite in the Enontekiö samples (<1%) it represents an
important mineral phase in this locality since it contains more than 40% LREE (Ce, La,
Nd). Monazite is, after allanite, the most important LREE carrier in albitites rocks of
Enontekiö, where it appears as isolated minute crystals with a diameter usually smaller
than 50—60 µm. Multigrain analyses show that the monazite‐(Ce) is rich in LREE as Ce
(~38%), La (~20%), Nd (~10%), CaO (1.5%), P2O5 (28%) and F (<1%). Bastnäsite is found
in most studied samples of Enontekiö rocks, it crystallised around davidite or occurs as
vein filling (Fig). EPMA analysis of bastnasite grains indicate that F more than 3%, and
the predominance of LREE (> 65%), with Ce2O3 (35‐37%), La2O3 (14‐17%) and Nd2O3
(11.5‐13.5%). Davidite as most common U‐Pb minerals in Enontekiö area occurs in
association with bastnaesite and allanite. Davidite is zoned, as shown by bastnaesite
inclusions found near the rim of one grain, and a darker central region in the same
grain (Fig. 4a). Reconnaissance EPMA traverses show that this zoning, at least in part,
reflects the uneven distribution of elements Fe, Y, Si and U. Uranium, Cr, Ti and Pb
tend to be relatively concentrated toward the central (more altered) portion of grains
at the expense of Pb. Darker areas thus have maximum U, Pb and minimum Ti (Fig.
4b).

Honkilehto
The Kuusamo Schist Belt is an Au‐Co‐Cu‐U occurrence with no resource estimate
available. It is mainly hosted by albitised, biotitised and sulphidised sericite quartzite
(12, 13). The studied samples characterized by occurrences of REE‐minerals
(bastnaesite, allanite) and U‐Pb (davidite) with high content of Cr‐magnetite and
apatite.
Bastnaesite crystals are envelope‐shaped around the U‐Pb minerals or as isolated fine
grains around the U‐minerals (Fig. 4c). In Honkilehto, two phases of U‐rich minerals
recognized; firstly U‐Pb rich minerals (richtetite) contain high uranium content 75% U,
20 Pb and 5% Y, Fe , secondly the U‐Si rich mineral (bijvoetite) contain less UO2 (~65)
with Si(~15%). Intergrowth of these two minerals exhibit cauliflower form in most

studied U‐Pb and U‐Si phases (Fig. 4d). All of these U‐minerals are associated mostly
with bastnaesite and allanite.

Conclusions

Detailed mineralogical characteristic revealed three distinct types of REE‐
mineralization as phosphates, carbonates and silicates in studied areas. Mineralogical
and mineral chemical analysis demonstrate that hydrothermal processes were
responsible for the REE mineralization in the studied rocks and confirms that such
processes are predominant in the formation of REE minerals in carbonatites, calc‐
silicate rocks and albitites. During late‐stage processes apatite and carbonate minerals
have been replaced by various assemblages of REE–Sr–Ba minerals in carbonatites.

The results obtained provide a useful reference basis for possible future feasibility
studies, as well as important mineralogical insights into the varied rock types and
mineralizing processes associated with the occurrence of elevated REE in northern
Finland.

Figure 3: BSE images of REE‐bearing and associated minerals in studied rocks. (a)
Bastnaesite growth in the rim of davidite or filling vugs within albite, (b) zoning in

davidite grain related to the concentration of (Pb‐U) versus Ti, (c, d) U‐Pb minerals
(bright) surrounded by bastnäsite (grey).

References

1. N. Lounamaa, “Utvinning av lantanider”, Kemisk Tidskrift, (5) 40‐42 (1972).
2. R. Hugg, “Katajakangas. Geologinen malmiarvio”, Report OU 12/85, Rautaruukki Oy
Malminetsintä, 1985.
3. A. Arzamastsev, V. Yakovenchuk, Y. Pakhomovsky, G. Ivanyuk, The Khibina and Lovozero alkaline
massifs: Geology and unique mineralization, 33 IGC excursion No 47, July 22 – August 2, 2008.
4. M. Korsakova, S. Krasotkin, V. Stromov, M. Iljina, L. Lauri, P. Nilsson, “Metallogenic areas in
Russian part of the Fennoscandian shield”, Edited by P. Eilu, Mineral deposits and metallogeny of
Fennoscandia, Geological Survey of Finland, Special Paper (53) 343–395 (2012).
5. H. Vartiainen, “The petrography, mineralogy and petrochemistry of the Sokli carbonatite massif,
northern Finland”, Geological Survey of Finland, Bulletin (313) 126 p (1980).
6. H.E. O'Brien, P. Peltonen, H. Vartiainen, H., 2005, “Kimberlites, carbonatites, and alkaline rocks”,
in Precambrian geology of Finland: key to the evolution of the Fennoscandian Shield (2005),
Edited by M. Lehtinen, P.N. Nurmi, O.T. Rämö. Precambrian Kimberlites, carbonatites, and
alkaline rocks. Developments in Precambrian Geology, Elsevier, Amsterdam, (14) 605‐644 (2005).
7. O. Sarapää, T. Al Ani, S. Lahti, L. Lauri, P. Sarala, A. Torppa, “ Kontinen, Rare earth exploration
potential in Finland”, Journal of Geochemical Exploration (133) 25‐41 (2013).
8. H. Vartiainen, “Sokli carbonatite complex, northern Finland”, Res Terrae. Ser., A 20, University of
Oulu, 8‐24 (2001).
9. T. Al Ani and O. Sarapää, “Geochemistry and mineral phases of REE in Jammi carbonatite veins
and fenites, southern end of the Sokli complex, NE Finland”, Geochemistry: Exploration,
Environment, Analysis 13 217‐224 (2013).
10. Hugg and V. Heiskanen “Otanmäen alueen niobi‐lantaniditutkimukset”, Raportti OU 28/85,
Rautaruukki Oy Malminetsintä, 1986.
11. P. Eilu, and H. Pankka, “Fingold ‐ a public database on gold deposits in Finland”, Version 1.1
[Electronic resource]. Geological Survey of Finland, Espoo. Optical disc (CD‐ROM), 2009.
12. O.Sarapää, and P. Sarala, “Rare earth element and gold exploration in glaciated terrain: example
from the Mäkärä area, northern Finlan d”, Geochemistry: Exploration, Environment, Analysis 13
131–143(2013).
13. H. Pankka, “Geology and mineralogy of Au‐Co‐U deposits in the Proterozoic Kuusamo volcano
sedimentary belt, northeastern Finland”, A dissertation, Geology, Michigan Technological
University 1992.


Rare‐earth elements minerals in carbonatites of the Kola
Alkaline Province (northern Fennoscandia)
Anatoly N. ZAITSEV1,2, Frances WALL 3, Anton R. CHAKHMOURADIAN4
1
Department of Mineralogy, St. Petersburg State University, St. Petersburg, Russia
2
Department of Earth Sciences, Natural History Museum, Cromwell Road, London, UK
3
Camborne School of Mines, University of Exeter, Cornwall, UK
4
Department of Geological Sciences, University of Manitoba, Winnipeg, Canada
a.zaitsev@spbu.ru, f.wall@exeter.ac.uk, Anton.Chakhmouradian@umanitoba.ca
Abstract
Various rare earth elements minerals occur in late‐stage carbonatites within the Kola
Alkaline Province. The carbonatites are mineralogically diverse rocks and contain
calcite, dolomite, magnesite, siderite and rhodochrosite as rock‐forming minerals. REE‐
minerals are present as accessory, minor and rock‐forming minerals related to two
distinct mineral assemblages: primary (magmatic or crystallized from
carbohydrothermal solution – e.g. burbankite) and secondary (metasomatic – e.g.
ancylite). The REE‐minerals tend to be enriched in light REE, with the Y‐rich mckelveyite‐
group minerals as rare exceptions. Stable and radiogenic isotopes indicate a deep
mantle source for C, O, Sr and Nd in primary minerals and it is thought that the original
formation of REE rich carbonatites was a result of multi‐stage fractional crystallization
of silicate‐carbonate melts.

Introduction
Rare earth‐rich carbonatites are here defined as carbonatites containing levels of rare
earth elements (REE) as oxides of 1 wt.% and higher in which REE minerals often attain
rock‐forming concentrations. In a recent review of compositional and mineralogical
data for Kola carbonatites (including Khibina, Kovdor, Vuoriyarvi and Turiy Mys)
(Zaitsev et al., 2014) established a threshold for the content of REE at 2,000–3,000
ppm, above which REE‐minerals are likely to be present. Such rocks are potential
economic sources of REE and although none of the examples cited in this paper have
been mined for REE, several have previously been exploration targets.

Where successive generations of carbonatite can be defined, all varieties of REE‐rich
carbonatites are always observed to be late‐stage rocks. Many occurrences are
assumed to be products of crystallization from (carbo‐)hydrothermal fluids (e.g. Fen,
Norway; Seblyavr, Kola Alkaline Province) and often all REE‐minerals in carbonatite are

grouped in this category. However, magmatic origins have been thought possible, and
indeed likely for other occurrences (Mountain Pass, USA).
Kola carbonatites
The Kola Alkaline Province contains twenty two complexes (380‐360 Ma) consisting of
various ultrabasic and alkaline rocks, they are, in order of their formation: olivinites,
clinopyroxenites, melilititic rocks (turjaites, uncompahgrites, okaites) and foidolites
(melteigites, ijolites, urtites). Nepheline syenites (foyaites, khibinites, luyavrites) are
principle rock types at two giant Khibina and Lovozero complexes (Fig. 1).
The carbonatites are the youngest intrusive phase in Kola ultrabasic‐alkaline
complexes and they form multi‐stage complex intrusions containing several varieties
of carbonatities. They are mineralogically diverse rocks and contain calcite, dolomite,
ankerite, kutnahorite, magnesite, siderite and rhodochrosite as major rock‐forming
minerals. Other typical minerals are represented by forsterite, diopside, phlogopite,
tetraferriphlogopite, magnesioarfvedsonite, richterite, apatite and magnetite (Wall &
Zaitsev, 2004).
Calcite carbonatites are the early‐stage carbonatites (in terminology of Kapustin
(1980)) in Khibina, Vuoriyarvi, Seblyavr and Sallanlatvi, while late‐stage carbonatites
are quite different in terms of mineralogy in these localities. Dolomite, and rarely
ankerite carbonatites are known from Vuoriyarvi, Seblyavr and Sallanlatvi, magnesite
and siderite carbonatites are present at Sallanlatvi, and Mn‐rich carbonatites
(including kutnahorite and rhodochrosite varieties) occur at Khibina.

The Kola carbonatites are characterised by the diverse mineralogy of their accessory
minerals, and some of these can attain major or minor minerals status in late‐stage
rocks. They contain assemblages of various primary and subsolidus minerals of REE
(e.g. burbankite and carbocernaite in Khibina and Vuoriyarvi, cordylite and
kukharenkoite in Khibina and ancylite in all carbonatites) and Nb and Zr (e.g.
pyrochlore, uranpyrochlore and bariopyrochlore at Khibina; lueshite and pyrochlore
at Sallanlatvi.

The REE‐mineralisation in the Kola carbonatites is highly variable ranging from tens of
ppm in REE‐poor Turiy Mys carbonatites to several wt.% in REE‐rich Khibina, Vuoriyarvi
and Sallanlatvi carbonatites. All carbonatites are enriched in light REE compared with
heavy REE with (La/Lu)CN ratio between 9.3 and 1230 (Zaitsev et al., 2014).

REE‐minerals
A total of twenty‐five REE‐minerals are known from the Kola carbonatites (Wall &
Zaitsev, 2004). Mostly, they are carbonates, including unhydrous sodium‐rich
burbankite and carbocernaite, hydrous ancylite, Ca‐ and Ba‐fluocarbonates synchysite,
bastnäsite and cordylite; in addition to oxides (loparite), silicates (cerite) and
phosphates (monazite) (Table 1).

Burbankite is a typical early‐crystallised mineral in Khibina and Vuoriyarvi late‐sage
carbonatites, of which the latter also contains calcioburbankite. These minerals form
crystals up to 7 cm long and 4 cm diameter, eudhedral hexagonal crystals; also they
occur as 10‐250 µm drop‐like inclusions hosted by calcite or dolomite‐ankerite or
mineralized along boundaries of calcite/dolomite. The minerals have low birefrigence
and optical data resembling apatite. Burbankite and calcioburbankite contain up to
21.7 and 18.5 wt.% REE2O3 respectively in Khibina and Vuoriyarvi.

Burbankite and calcioburbankite often show variable degrees of alteration and several
assemblages of secondary minerals have been established. Rather simple
assemblages, ancylite+strontianite+baryte and synchysite+strontianite+baryte, are
typical for the Khibina carbonatites but the pseudomorphs after burbankite‐
carbocernaite in Vuoriyarvi are very complex with carbocernaite, ancylite, alstonite,
olekmenskite, baryte, monazite and Sr‐enriched calcite (Fig. 1).

Neither burbankite nor calcioburbankite have been found in the Seblyavr and
Sallanlatvi carbonatites but a few samples contain “pseudomorph‐like” polycrystalline
assemblages of ancylite, strontianite and baryte that suggest the possible former
presence of burbankite.

Ancylite is another common REE mineral for the Kola carbonatites. Typically it forms
veinlets and thin veins within burbankite‐calcioburbankite, and also full
pseudomorphs (with strontianite and other minerals) after these minerals. Also,
ancylite forms monomineralic veins cross‐cutting carbonatites and occurs as euhedral
crystals in rock cavities, containing up to 46.5 wt.% REE2O3.

Examples of rare REE minerals are the Ba fluocarbonates (cordylite and kukharenkoite)
in Khibina carbonatites and Y‐rich hydrous carbonates (mckelveyite, donnayite and
ewaldite) in Khibina and Vuoriyarvi carbonatites. These minerals are the latest in the
crystallisation sequence of the rare earth minerals. While cordylite and kukharenkoite

are well studied minerals, the mckelveyite group minerals are not well characterised,
particularly their crystal structure.

Origin of REE carbonatites
Studies of Kola REE‐rich carbonatites (Wall & Zaitsev, 2004) have showen that there
are two groups of rare earth minerals. The first consists of primary minerals crystallised
directly from a magma or carbohydrothermal solution; examples are burbankite,
calcioburbankite, synchysite, bastnäsite (magmatic minerals), ancylite, mckelveyite
(hydrothermal minerals). The second group consists of secondary minerals that were
formed during replacement of early‐formed REE minerals; ancylite and synchysite are
the best examples of these.

Stable and radiogenic isotopes for REE minerals from the Khibina and Vuoriyarvi
carbonatites indicate a deep mantle source for C, O, Sr and Nd in primary minerals, but
show significant changes for stable isotopes and no changes for radiogenic isotopes in
secondary minerals (Zaitsev et al., 2002). Formation of REE rich carbonatites was
probably a result of multi‐stage fractional crystallization of silicate‐carbonate melts
(Chakhmouradian & Zaitsev, 2012).

Figure 1: Simplified map of the Kola province showing the distribution of ultrabasic‐
alkaline‐carbonatite complexes, kimberlites, lamproites and lamprophyres (after Bell
& Rukhlov, 2004 in Wall & Zaitsev, 2004).


Figure 2: Backscattered electron image showing relic of burbankite in platy crystal of
synchysite (centre of image), strontianite (gray) and baryte (white).

Table 1: REE minerals in Kola carbonatites

Mineral Formula
Burbankite (Na,Ca)3(Sr,Ba,REE)3(CO3)5
Calcioburbankite (Na,Ca)3(Ca,Sr,Ba,REE)3(CO3)5
Carbocernaite (Na,Ca)(Sr,REE,Ba)(CO3)2
Ancylite SrREE(CO3)2(OH)xH2O
Synchysite CaREE(CO3)2F
Bastnäsite REE(CO3)F
Cordylite (Na,Ca)Ba(REE,Sr)2(CO3)4F
Kukharenkoite Ba2REE(CO3)3F
Mckelveyite Ba3NaCaY(CO3)6x3H2O
Monazite REE(PO4)
Loparite (REE,Na,Ca)(Ti,Nb)O3
Cerite (REE,Ca)9(Fe,Al)(SiO4)6(SiO3OH)(OH)


References
1. Chakhmouradian A.R. and Zaitsev, A.N. (2012): Rare earth mineralization in igneous rocks: sources
and processes. Elements, 8, 347–353.
2. Kapustin Yu.L. (1980): Mineralogy of Carbonatites. Amerind Publishing, New Dehli.
3. Wall, F. and Zaitsev, A.N. (Editors) (2004): Phoscorites and carbonatites from mantle to mine: the
key example of the Kola Alkaline Province. Mineralogical Society Series, 10. Mineralogical Society,
London.
4. Zaitsev, A.N., Demény, A., Sindern S., and Wall, F. (2002): Burbankite group minerals and their
alteration in rare earth carbonatites — source of elements and fluids (evidence from C–O and Sr–
Nd isotopic data). Lithos, 62, 15–33.
5. Zaitsev, A.N., Williams, C., Jeffries, T.E.; Strekopytov, S., Moutte, J., Ivashchenkova, O.V., Spratt, J.,
Petrov, S.V., Wall, F. Seltmann R. and Borozdin, A.P. (2014): Rare earth elements in phoscorites
and carbonatites of the Devonian Kola Alkaline Province, Russia: Examples from Kovdor, Khibina,
Vuoriyarvi and Turiy Mys complexes, Ore Geology Reviews, 61, 204–225 .


MINERALOGY AND GEOCHEMISTRY OF STRATEGIC

METALS AND REE IN COASTAL HEAVY MINERAL SANDS
FROM NISYROS VOLCANO, GREECE
Athanasios GODELITSAS1, Ioannis TZIFAS2*, Platon GAMALETSOS1,
Giorgos ECONOMOU3, Pavlos MEGALOVASILIS4 and Paraskevi
NOMIKOU1
1
University of Athens, School of Science, 15784 Zographou, Greece
2
Aristotle University, Department of Chemistry, 54124 Thessaloniki, Greece
3
Institute of Geology and Mineral Exploration, Olympic Village, 13677 Acharnae, Greece
4
Technological Educational Institute of the Ionian Islands, Zakynthos, GR, 29100, Greece

Abstract

The northern coast of Nisyros volcano, belonging to the Hellenic Volcanic Arc (HVA),
hosts black sands with heavy minerals, which are herein investigated for the first time
in the literature. This also the first study concerning heavy sands from the entire HVA.
UCC‐normalized bulk spidergrams indicated that the sands are enriched in V, HFSE (Nb,
Ta, Zr, Hf), Co, Sn and P, while they are depleted in LILE and particular incompatible
elements. The study by powder‐XRD and SEM‐EDS showed abundant feldspars, apatite,
Fe‐spinels, tourmaline, pyroxenes (with apatite inclusions), and ilmenite (with apatite
and sulfide inclusions) together with zircons. The heavy mineral sands are also
relatively enriched in REE (exhibiting a remarkable Eu negative anomaly) but the total
content can be considered interestingly low (ΣREE+Y = 240 ppm). The studied volcanic
weathered material was also compared to “fresh” volcanic material, that is a clayely
material (alunite, anhydrite, opal‐CT, quartz, kaolinite) currently deposited into Nisyros
caldera main hydrothermal crater. This is enriched in volatile components (L.O.I. 33.7
wt.% compared to 1.4 wt.% of sands), S (14.28 wt.% compared to 0.10 wt.% of sands)
and other chalcophile elements (Se, Bi, Hg, As, Pb), as well as in P and Ba. Thus, P is an
element enriched in both weathered and “fresh” Nisyros volcanic materials, whereas
there is a great contrast in Fe, Ti, HFSE and chalcophile elements. Moreover, the sands
show different geochemical patterns compared to the volcanic rocks of the island
which are particularly enriched in LILE. It is concluded that the weathering contributed
to coastal accumulation of specific heavy minerals and strategic metals, including V
(1920 ppm) and Nb (245 ppm), while REE are low compared to coastal heavy mineral
sands of Greece and the globe. The same geochemical trends stand when comparing
Nisyros sands with various volcanic rocks from the whole HVA.


Key words: Heavy minerals; strategic metals; rare earths; coastal sands; volcano;
Nisyros

Introduction

The growing demand of strategic metals (e.g. Nb, Ta), actinides (U, Th) and lanthanides
(REE) from the markets enhances geological exploration projects in order to find
commodities which exhibit high exploitation and metallurgical extraction potential.
Although the geological sources of the above elements exhibit a great variety, many
studies have focused on placer deposits. Beach sands are mineral accumulations
typically formed through weathering and erosion of either igneous or metamorphic
rocks. Significant interest for further study show natural high magnetic black sands
which are enriched in so‐called heavy minerals. Their crystals are sufficiently resistant
to weathering processes and their accumulation in coastal areas owns to the gravity
segregation. The phases occurring in heavy mineral sands, in mixture with common
tectosilicates and phyllosilicates, are Fe‐ & Cr‐spinels, Ti‐oxide minerals, phosphate
minerals (apatite, monazite, xenotime, rhabdophane), and specific heavy silicate
minerals such as zircon and allanite (epidote group). The above natural materials are
rich, by nature, in strategic metals, actinides, and lanthanides. Heavy mineral sand
formations are mainly found in ancient or modern ocean and sea beaches and
occasionally within continental land masses 1‐3. However, sands commonly found in
coasts of volcanic islands also show an economic importance, depending on the
extraction ability of metals hosted in the crystal lattice of “accessory minerals” such as
rutile, zircon and ilmenite 4, 1, 5. On the other hand, heavy mineral sands show an
increasing environmental importance due to the presence of actinides (U, Th, and their
decay products) that could pose a radiological risk 6, 7.
Previous studies in Greece, concerned heavy mineral sands derived from
plutonic rocks (i.e. granites), particularly in the long coast (240 km) west of Kavala city,
in Touzla Cape near Thessaloniki city, and also in Sithonia peninsula coasts in Chalikidiki
8‐12
. In the frame of the present study we present, for the first time in the literature,
mineralogical and geochemical data for coastal heavy mineral sands occurring in
southern Greece, and particularly in Nisyros Island which is located in the Hellenic
Volcanic Arc (HVA) at eastern Mediterranean Sea. More specifically we demonstrate
the strategic metal (HFSE elements), REE and actinide elements potential of these
sands. The coasts of the Mediterranean Sea are rich in heavy mineral sands of volcanic
origin due to several volcanoes still active nowadays 13. Thus, this paper could be

considered as a model study for sands from the whole HVA and from relevant volcanic
areas in Mediterranean coasts.

Geology settings

The Hellenic Volcanic Arc (HVA) is a young 5 Ma‐to‐present volcanic arc that has
developed in the pre‐Alpine to Quaternary continental crust of the Hellenic
Subduction System (HSS). Its development is related to the northward subduction of
the last remnant of the oceanic crust of the African plate beneath the southern edge
of the active margin of the European plate, resulting in the formation of numerous
Plio‐Quaternary volcanic centers, namely Crommyonia (Soussaki), Methana‐Aegina‐
Poros, Milos‐Komolos‐Polyegos, Santorini (Thira)‐Kolumbo and Nisyros‐Yali‐Kos 14, 15.
Nisyros volcano is composed of Quaternary volcanic rocks with alternating lava flows,
pyroclastic layers and lava domes erupted during successive periods of activity 16‐ 19.
The evolution of this volcano is associated to a formation of a big caldera in the middle
of the island, with significant post‐volcanic activity. In fact, Nisyros constitues a
stratovolcano which has an average diameter of 8 km (based on sealevel niveau) with
a caldera (4 km in diameter) formed during the explosive eruption of a large pumice.
The oldest age of the volcanic rocks is about 160 Ka and the most recent phreatic
eruption took place in 1867 AD. The caldera floor consists of several hydrothrmal
craters, with the largest (Stefanos) having a a depth of 27 m and diameter of 330 m.
Hydrothermal venting from the craters continues to this day with the last significant
hydrothermal eruptions occurring in 1871‐1888 AD 20. Present‐day activity is
documented by fumarolic degassing of mainly H2S, CO2, H2O, H2 and CH4. The pre‐
caldera period is characterised by a succession of four lavas and four pyroclastic flows
that formed the stratovolcano18. This succession was followed by the rhyolites of Nikia
in the southern part of Nisyros and by pumice deposits. Post‐caldera magmatic activity
21
has produced a series of rhyodacitic domes (Profitis Ilias Synthem) that fill the north‐
western section of the caldera floor, as well as the south‐western flank of the island.
Geothermal drillholes22 have been studied in order to give information about the
basement beneath Nisyros. The boreholes showed intersected units of diorite,
limestone, and marble. In addition skarns have been detected and are formed by
contact metamorphism and hydrothermal alteration of the limestones. Carbonate‐rich
rock‐types are also found as xenoliths in some pyroclastic deposits on the island 17, 23‐

25
. The volcanic stratigraphy of Nisyros has been described by many authors 17, 24, 26, 27.
The earliest magmatism at Nisyros is also related to submarine flows that deposited
porphyritic pillow lavas of basaltic‐andesite composition that crop out on the north‐
western shore, as a result of active uplift of this section of the island 28, 29. There is
evidence of erosion and reworking of these volcanic rocks, indicating exposure above
sea level for part of the depositional history 30. It should be mentioned that the black
heavy mineral sands, occurring in the northern coast of the Island, have never been
investigated by the aforementioned authors.

Material and methods

The sand samples (NISBS) have been collected from the northern coast of Nisyros
Island (Fig. 1). The material appears as fine grained and it is black in the colour due to
the high content in Fe‐spinels and other dark heavy minerals. Furthermore, another
clayely sample (NISC) has been collected from the recent deposits of the interior of
the main hydrothermal crater of the Nisyros volcano caldera (Fig. 1). The second
sample was considered to be studied for comparison reasons, as “fresh” volcanic
material, in contrast to the weathered (re‐processed) volcanic coastal sands.
The bulk chemical composition, concerning major and trace elements, was
determined by Inductively Coupled Plasma Optical Emission & Mass
Spectrometry/ICP‐OES &‐MS (PerkinElmer Sciex Elan 9000) after LiBO2/LiB4O7 fusion
and HNO3‐digestion of 0.2 g of sample. Loss on ignition (LOI), as well as total C and S,
were also measured using standard methods. The mineralogical characterization was
performed by optical microscopy, powder X‐ray Diffraction/XRD (Siemens D5005 –
now Bruker AXS– diffractometer), and Scanning Electron Microscopy/SEM–EDS (Jeol
JSM‐5600 equipped with Oxford EDS) in both free sand grains and polished‐sections.
The fine clay (2 ‐ 0.2 μm) and ultrafine (nano) clay (< 0.2 μm) fractions of the materials,
separated using standard wet‐chemical and centrifugation procedures31, were also
investigated with regard to raw, glycolated and heated samples (550 oC).


Figure 1: Sampling areas concerning black sands (NISBS) in the northern coast of
Nisyros Island (a,b) and clayely material (NISC) in the floor of the main hydrothermal
crater of the caldera (c,d).


The UCC‐normalized spidergrams, based on bulk chemical analyses, indicated that the
studied coastal heavy mineral sands from Nisyros volcano (NISBS) are enriched in V,
High‐Field Strength Elements (HFSE, i.e. Nb, Ta, Zr, Hf), Co, Sn and P, while they are
depleted in Large Ion Lithophilic Elements (LILE) and particular incompatible elements
(Fig. 2 upper). Thus, V concentration has been recorded to be 1920 ppm whereas Nb
concentration 245 ppm. Moreover, concerning the REE content of the sands,
chondrite‐normalized patterns (Fig. 2 lower) showed that they are enriched mainly in
LREE (ΣLREE = 193 ppm) rather than HREE (ΣHREE = 20 ppm), but the total content can
be considered interestingly low (ΣREE+Y = 240 ppm) in respect to heavy mineral sands.
Europium negative anomalies are characteristic in all REE patterns including those
corresponding to various rocks of the HVA. Chondrite‐normalized patterns in previous

study for Nisyros volcanic rocks showed LREE‐enrichment, with similar slight negative
Eu anomalies only apparent in the andesitic samples. Alternatively, the different
degrees of LREE and HREE enrichments may be also be due to magma composition
variations32. On the other hand, the studied “fresh” volcanic material (NISC), currently
deposited into Nisyros caldera main hydrothermal crater, is enriched in volatile
components and other chalcophile elements (Se, Bi, Hg, As, Pb), as well as in P and Ba.
According to the same spidergrams, the material is depleted in LILE as well. The REE
content, compared to the volcanic weathered material (i.e. the sands), is remarkably
lower (ΣREE+Y = 96 ppm), while Eu anomaly also shows a slightly different trend.
Besides, the L.O.I. of NISC is extremely high (33.7 wt.% compared to 1.4 wt.% of sands)
accompanied by a high S content (14.28 wt.% compared to 0.10 wt.% of sands). That
implies abundant hydrated, hydroxylated, and sulfate mineral phases. It is therefore
worthy to note that the heavy mineral sands show different geochemical patterns
compared to the caldera material. Only P is an exception due to the fact that it is an
element that occurs in elevated concentration in both weathered/re‐processed and
“fresh” Nisyros volcanic materials, whereas there is a significant contrast in Fe, Ti, HFSE
and chalcophile elements.


NISBS
NISC
1000
KIMOLOS (Godelitsas et
al_2010)
Kos TsoukalasPhD_2008
Nisyros
Volcanics_Lithos2012
Santorini
J.Vol.Geoth.Res2006
Santorini J.Petrol.2011
100 Santorini Lithos2009
Sample/Chondrite
Aegina
Methana
Poros
Santorini
J.Vol.Geoth.Res1987
Milos
10 Nisyros J.Vol.Geoth.Res1987
Kos J.Vol.Geoth.Res1987
Nisyros Pre Caldera
Lithos2005
Nisyros Post Caldera
Lithos2005
Yali Lithos2005
IosLithos2005
1
Nisyros. Bull Volcanol.1995
La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu

Figure 2: UCC‐normalized spidergrams (upper) and chondrite‐normalized REE
patterns (lower) for Nisyros black heavy mineral sands (NISBS) and hydrothermal
crater material (NISC), compared to volcanic rocks from Nisyros and the entire HVA.

According to a Eu/Eu* ‐ Ce/Ce* discrimination diagram, recently proposed for
sedimentary geological materials33 , the caldera crater “fresh” material is closer to the
massive sulfides field, while the sands, along with the HVA rocks, are closer to the
average volcanics (Fig 3 upper). Additionally, the sands seem to be part of sediments‐
partial field using a (Ce/Yb)NASC ‐ (La/Sm)NASC discrimination diagram (Fig. 3 lower), for
the reason that the material is weathered and re‐processed.


Figure 3: REE discrimination diagrams for sedimentary geological materials (Leybourne
& Johannsen, 2008), as applied to NISBS, NISC, and rocks from Nisyros and the entire
HVA.

Figure 4: SEM‐EDS data indicating strategic metal‐, REE‐ and actinide‐bearing heavy
minerals in the studied Nisyros sands.

The powder‐XRD and SEM‐EDS study of the sands showed common
tectosilicates (feldspars and particularly plagioclase), together with abundant
phosphates (apatite), Fe‐spinels (magnetite, ulvospinel), tourmaline, clino‐ and ortho‐
pyroxenes (with apatite inclusions), and ilmenite (with apatite and sulfide inclusions)
as well as zircons (Fig. 4). Tourmaline crystals appear as elongated trigonal prisms, with
smaller, second order prism faces on their corners. Tourmaline had also been
mentioned in rhyolite dykes, most likely Pliocene, from Ikaria Island in the Aegean

back‐arc region34. Pyrite and barite have also been identified in the sands. It is notable
that micro‐inclusions of sulfides, together with apatite needles, are occasionally
included in ilmenite fragments. Sulfides and sulfates may be significant scavengers of
various trace elements related to bulk geochemical anomalies. It should also be
mentioned that the studied Nisyros heavy mineral sands are rather coarse, in
microscale (Fig. 4), and they do not contain a measurable clay (fine clay and nanoclay)
fraction.

On the other hand, the XRD patterns of the raw clayely material from the
hydrothermal crater of the caldera, showed that the main minerals are alunite and
anhydrite, in accordance to bulk chemical analyses indicated high L.O.I. and S, as well
as opal‐CT and quartz. Besides, careful SEM‐EDS investigation revealed distinct barite
microcrystals with a characteristic platy morphology. Additionally, the XRD of raw,
glycolated, and heated (550 oC) clay fractions indicated a significant amount of alunite
and kaolinite‐group minerals31. Concerning the caldera floor crater and the southern
caldera wall, surface expressions of an active hydrothermal system include
hydrothermal deposits, fumaroles, mud pools and boiling water pools 35, 36. The
caldera model 36 is consisted of a magmatic body, at unknown depth, which supplies
heat and fluids to the hydrothermal system, a deep boiling aquifer (situated at more
than 900 m below sea level) and finally a shallow reservoir at variable temperatures
fed by a mixture of vapor separated by the deep aquifer and meteoric water. Thermal
springs and fumaroles represent the surface manifestations of the deep hydrothermal
system of Nisyros, which has been directly attained by the geothermal wells Nisyros 1
and Nisyros 237. Previous studies have been mentioned a decrease in sulfate content
that could be a result of sulfate minerals precipitation or reduction to sulfide. Alunite
which is abundant also in the clay faction of the material is an evidence of precipitation
of sulfates while the presence of pyrite could represent an example of sulfide
reduction. That aspect has been pointed out by several authors 38, 39, 20. Geothermal
wells both situated at the northeastern periphery of the craters were investigated
about the existed hydrothermal minerals considering the water‐rock interaction40.
Thus, the propylitic zone extends from 950 to 1547 m and it is marked by abundance
of well‐crystallized epidote, adularia, albite, quartz, pyrite, chlorite, and sericite‐
muscovite. Less abundant minerals such as anhydrite, stilpnomelane, wairakite,
garnet, tremolite, and pyroxene, also exist. Moreover, previous mineralogical
investigations indicated in the caldera material minerals such as halotrichite,
alunogen, voltaite, gypsum, sulfur,and tamarugite41, besides to alunite, anhydrite,

opal‐CT and quartz. The volatiles and chalcophile elements, which are enriched in the
NISC material, may be mainly associated to alunite. The alunite supergroup minerals
are defined as having the general formula DG3(TX4)2X’6, where D is normally a large,
mono‐ to tetravalent, ion in 12‐fold coordination (e.g. K, Na, Ca, Pb, REE) or partial
vacancy, G is usually a trivalent cation and minor divalent cations (e.g. Fe, Al), T is a
penta‐ & hexavalent cation and minor Si4+, and X/X’ is O, (OH), minor F and possibly
H2O 42. Thus, TX4 are characteristic alunite anions, usually SO4, PO4, or even AsO4. As a
consequence, certain positive geochemical anomalies of the caldera material, may be
related to elements hosted in alunite T sites (P, As, and perhaps Se), whereas other
elements, such as Bi, Pb and Hg, might substitute in D sites.
The heavy mineral sands from Nisyros volcano represent a weathered and re‐
processed material which originates mainly from the Quaternary acidic volcanic rocks,
lavas, pyroclastic layers and domes. Nisyros rocks belong to the calc‐alkaline series and
their composition ranges from basaltic andesites to rhyolites. According to previous
studies 17 the volcanic rocks of Nisyros exhibit important mineralogical similarities in
relation to the other volcanic centres of the HVA (Milos, Santorini, Methana) and
therefore all above are included in the calc‐alkaline trend and range from basaltic
andesite to silicic rhyolite. The minerals identified in the sands fit to the geological
background of Nisyros, taking into account all previous studies regarding petrology of
the Island’s volcanics. According to mineralogical investigation, euhedral to subhedral
crystals of plagioclases (with characteristic zoning) are abundant in Nisyros rocks23.
Moreover, that is a mineral phase which often shows abrupt compositional changes
reflecting significant changes in chemical potentials that are commonly attributed to
magma mixing, ascent, and degassing 43. Clinopyroxene, which is also abundant in the
sands, has been mentioned as a major mineral phase of basaltic andesitic and dacitic
rocks23. However, orthopyroxene with hypersthenic composition is additionally
present in the sands. In contrast, olivine and hornblende, which are common as
phenocrysts in the parent volcanic rocks (lower andesitic lava flows, pyroclasts, post‐
caldera domes and their enclaves, dacitic and rhyolitic volcanic rocks), are not present
in the sands. This is evidence that the above mineral phases were most likely dissolved
during prolonged erosion and weathering processes, resulted in formation of heavy
mineral sands. Fe‐spinels (magnetite) and ilmenite are commonly present as
microphenocrysts in the parent rocks, although the post‐caldera dacitic‐ rhyolitic
domes lack ilmenite. The black sands contain high content of magnetite and ilmenite,
which is not unusual due to the fact that they are rather resistant to weathering
processes. Hydrothermal pyrite has also been reported as inclusion in Fe‐Ti oxides.

Among the accessory mineral phases, apatite is ubiquitous and is generally included in
plagioclase and pyroxene, whereas zircon is only found in the most evolved rocks 44, 25.
The positive geochemical anomalies, derived from the UCC‐normalized
spidergrams with regard to V, HFSE (Nb, Ta, Zr, Hf), Co, Sn and P, are consistent to the
mineralogical composition. The intense presence of ilmenite affects the elevated
content of Nb and Ta due to the fact that this metallic mineral may host such high‐
valence elements in its structure. Ti‐oxides are likely to be the most important
reservoirs of Nb and Ta in the mantle45, 46, although, if titanite is present, it will strongly
affect the Nb and Ta partitioning. Remarkably high Nb2O5 content (0.28 wt.%) has been
detected in the magnetic fractions of ilmenite from the beach placer deposit of
Chavara, southwest India47. The positive anomalies of tetravalent Zr and Hf are
justified from the zircons inclusions, which are abundant as inclusions in the pyroxene
and ilmenite crystals. According to the literature they are present in all geological units
of the Nisyros and are most likely part of mineral assemblages in rhyodacites to
rhyolites. The main difference of zircons concerns their concentrations in some trace
elements, including Ti, Hf, U, Sc, and Nb. The typically higher Hf, U, Sc, and Nb contents
reflect the slightly more evolved melt composition from which these zircons are
crystallized 48. Also, the Fe‐spinels could host in their lattice the observed elevated
amounts of V and Co. Furthermore, the P anomaly is obviously associated to apatite,
derived directly from the parent volcanic rocks. Apatite supergroup concerns phases
with the general chemical formula IXM12VIIM23(IVTO4)3X (Z=2), where many cations
could substitute in M1, M2, T, or X site. Thus, V, which is enriched in the studied
samples, except its predominant presence into Fe‐spinels, may also be present (in the
form of [V5+O4]3‐) into the apatite structure49. Additionally, Co, and also Sn, may be
associated to sulfide mineral phases. Hydrothermal pyrite contains significant
amounts of trace elements (e.g. As, Pb, Sb, Bi, Cu, Co, Ni, Zn, Au, Ag, Se, Te), which can
be specifically incorporated into nanoparticles50. Most of the nanoparticles are
crystalline and are excluded from the pyrite matrix or precipitated from the
hydrothermal solution. Significant recrystallization is taking place under the conditions
of increasing temperature and therefore the migration of the trace metals via diffusion
through the pyrite structure and melt formation have been observed by many
researchers 51, 52. Recent studies have shown that many of the trace elements originally
thought to be present in the structure of pyrite, are actually present as micro‐
inclusions 53, 50 abundant in early forms of pyrite. Whereas Ni , Co, and Se are
commonly present in the pyrite structure, As, Cu, Zn, Pb, Bi, Sb, Tl, Mo, Ag, Cd, Mn, Hg,

and Te may be either in the structure or as sulfide micro‐inclusions, and Ti, V, U, Ba,
Sn, W, and Cr, are in variably present within micro‐inclusions of matrix material.
Trace‐element variation patterns in previous studies for Nisyros volcanics
feature the characteristic subduction‐related signature of relative enrichment in LILE
(i.e. K, Cs, Rb, Ba, Pb), U, Th and Sr, and depletion in Nb and Ti 32. However, the studied
heavy mineral sands are depleted in LILE, which clearly indicates that the above
elements are preferentially dissolved and removed upon prolonged weathering
processes. Kanafià Synthem which represents the oldest volcanic unit of Nisyros
volcano showed increased incompatible element (e.g., Zr, Nb, Ba, Rb, Ce)
concentrations with increasing SiO2 or decreasing MgO, as it has also been observed
in previous geochemical studies of Nisyros rocks23, 25, 54, 55. Lower Pumice and Upper
Pumice glasses of Nisyros are characterized as high‐K rhyolites56 and their chemical
trends indicate elevated values of incompatible element of primitive mantle. Besides,
negative anomalies at Nb, Ta and also at Ba, Sr and Eu, are consistent with the arc
setting of Nisyros and with feldspar fractionation, respectively57. Empirical and
computational studies have demonstrated that partitioning of tri‐ tetra‐ and
pentavalent cations in clinopyroxene increases with increasing alumina content, due
to enhanced coupled substitution of highly‐charged elements on M sites with
tetrahedrally‐coordinated Al 58, 59. Consequently, clinopyroxene is also a mineral which
should play a crucial role in the accumulation of trace elements in the studied sands.
It is very crucial to mention that previous studies concerning coastal heavy
mineral sands from northern Greece showed elevated concentration of REE and
natural actinide elements 8‐12. In the case of Nisyros sands the REE content is
comparably very low, but, on the other hand, the HFSE contents are much higher. It is
evident that the type of the igneous parent rocks affects the mineralogical and
chemical composition. Thus, it is concluded that in both cases acidic to intermediate
igneous rocks are involved (Kavala pluton, Monopigado granite, Sithonia granite, and
Nisyros volcanics), but it seems that volcanic rocks provide with less REE+Y (and
actinides) and more strategic metals (namely V, Nb, Ta and Co). This conclusion,
beyond the scientific interest, could be especially useful for future exploration and
exploitation projects in Greece and the entire Mediterranean and southern Europe.


Conclusions

Coastal heavy mineral black sands derived, by weathering, from acidic volcanic rocks
of Nisyros island, located in the Hellenic Volcanic Arc / HVA (southern Greece, eastern
Mediterranean), differ significantly to similar sands derived from relevant plutonic
rocks occurring in northern Greece (Kavala, Thessaloniki, Chalkidiki). The mineral
constituents of the studied sands are feldspars, apatite, Fe‐spinels, tourmaline,
pyroxenes (with apatite inclusions), and ilmenite (with apatite and sulfide inclusions)
together with zircons. The Nisyros sands are enriched in specific strategic metals (e.g.
V = 1920 ppm and Nb = 245 ppm) and exhibit unusually low REE content (ΣREE+Y = 240
ppm). This weathered/re‐processed material is also different, in terms of mineralogy
and geochemistry, to Nisyros and HVA volcanic rocks, as well as to “fresh” volcanic
clayely material (alunite, anhydrite, opal‐CT, quartz, kaolinite) currently deposited into
Nisyros caldera main hydrothermal crater. The later one is enriched in volatile
components and other chalcophile elements, whereas according to Eu/Eu* ‐ Ce/Ce*
discrimination diagram it is closer to the massive sulfides field. The present paper is
the first, in the literature, study of heavy mineral sands from the entire HVA, and could
contribute to future exploration and exploitation of coastal sands in Greece and the
entire Mediterranean and southern Europe.

References

1. Kuzvart, M. 1984. Industrial minerals and rocks. Developments in Economic Geology 18.
Elsevier, Amsterdam, 454 pp.
2. Patro, B.C., Sahu, B.K. and Guru, S. 1989. A study on heavy mineral sands of Gopalpur beach,
Orissa. Journal of the Geological Society of India, 33, 243-252.
3. Rattigan, J.H. and Stitt, P.H. 1990. Heavy mineral sands. In: Geological Aspects of the
Discovery of some Important Mineral Deposits in Australia (K.R. Glasson and J.H. Rattigan,
eds), pp. 369-378. Australasian Institute of Mining & Metallurgy Monograph No. 17. AIMM
Australasian Institute of Mining & Metallurgy, Victoria, Australia.
4. Lefond, J.S. 1975. Industrial Minerals and Rocks. American Institute of Mining, Metallurgical
and Petroleum Engineers, Maryland, 1360 pp.
5. Vanecek, M. 1994. Mineral Deposits of the World. Elsevier, Amsterdam, 519 pp.
6. Eisenbud, M. 1987. Environmental Radioactivity from Natural, Industrial and Military Sources.
Academic Press, San Diego, 475 pp.
7. Faure, G., 1992. Principles and Applications of Inorganic Geochemistry. Macmillan, New York,
626 pp.
8. Perissokratis C., Moorby, S.A. Angelopoylos, I., Cronan, P.S.., Papavassilioy, C., Konispoliatis,
N., Sakellariadou, F., Mitropoulso, D. (1988) Mineral Concetrations in the recent sediments off

eastern Macedonia Northen Greece: Geological and Geochemical cosiderations Min. Dep. Eur.
Com. Pp. 530-552.
9. Filippidis A., Misaelidis P., Clouvas A., Godelitsas A., Barbayiannis N., Anousis I., (1997)
Mineral chemical and radiological investigation of a black sand at Touzla Cape, near
Thessaloniki, Greece, Environmental Geochemistry and Health (1997), 19, pp 83-88
10. Pergamalis F., Karageorgiou D. E., Koukoulis A (2001): Contribution of gamma-radioactivity
in the detection of the depositis potantial of Ti, REE, Th, U in the coastal area of Nea Peramos-
Loutra Eleytheron, S. Kavala. Bull. of Geol. Soc. of Greece.
11. Papadopoulos A., Christofides G., Koroneos A. , Stoulos S. 2013. Natural radioactivity
distribution and gamma radiation exposure of beach sands from Sithonia Peninsula Cent. Eur.
J. Geosci., 1-14, Proofreading version
12. Papadopoulos A., Christofides G. , Pe-Piper G., Koroneos A., Papadopoulou L. (2014)
Geochemistry of beach sands from Sithonia Peninsula (Chalkidiki, Northern Greece) Chemie
der Erde, accepted manuscript.
13. Braccini S., Pellegrinelli O., Krämer K. (2013) Mössbauer, X-Ray and Magnetic Studies of
Black Sand from the Italian Mediterranean Sea. World Journal of Nuclear Science and
Technology 3, 91-95.
14. Papanikolaou, D., 1993. Geotectonic evolution of the Aegean. Bull. Geol. Soc. Greece 27, 33–
48.
15. Royden, L. H., Papanikolaou, D., 2011. Slab segmentation and late Cenozoic disruption of the
Hellenic arc. Geochem. Geophys. Geosyst. 12.
16. Georgalas, G., 1958. Nisyros island. Nisyriaka Chronika 21, 22, 23, Athens
17. Di Paola, G.M., 1974. Volcanology and petrology of Nisyros (Dodecanese, Greece). Bulletin of
Volcanology 38, 944–987.
18. Papanikolaou, D.J., Lekkas, E.L., Sakelariou, D.T., 1991. Geological structure and evolution of
Nisyros volcano. Bulletin of the Geological Society, Greece 25 (1), 405– 419.
19. Vougioukalakis, G.E., 1998. Volcanic stratigraphy and evolution of Nisyros Island. Bulletin of
the Geological Society of Greece 27, 239–258.
20. Brombach, T., Caliro, S., Chiodini, G., Fiebig, J., Hunziker, J.C., Raco, B., 2003. Geochemical
evidence for mixing of magmatic fluids with seawater, Nisyros hydrothermal system, Greece.
Bulletin of Volcanology 65, 505–516
21. Zellmer, G.F., Turner, S.P., 2007. Arc dacite genesis pathways: evidence from mafic enclaves
and their hosts in Aegean lavas. Lithos 95, 346–362.
22. Marini, L., Principe, C., Chiodini, G., Cioni, R., Fytikas, M., Marinelli, G., 1993. Hydrothermal
eruptions of Nisyros (Dodecanese, Greece)—past events and present hazard. Journal of
Volcanology and Geothermal Research 56, 71–94.
23. Francalanci, L., Varekamp, J.C., Vougioukalakis, G., Defant, M.J., Innocenti, F., Manetti, P.,
1995. Crystal retention, fractionation and crustal assimilation in a convecting magma chamber,
Nisyros Volcano, Greece. Bulletin of Volcanology 56, 601–620.
24. Limburg, E.M., Varekamp, J.C., 1991. Young pumice deposits on Nisyros, Greece. Bulletin of
Volcanology 54, 68–77.
25. Seymour St., K.S., Vlassopoulos, D., 1992. Magma mixing at Nisyros Volcano, as inferred from
incompatible trace-element systematics. Journal of Volcanology and Geothermal Research 50,
273–299.
26. Volentik, A., Principe, C., Vanderkluysen, L., Hunziker, J.C., 2005a. Stratigraphy of Nisyros
Volcano (Greece). In: Hunziker, J.C., Marini, L. (Eds.), The Geology, Geochemistry and
Evolution of Nisyros Volcano (Greece). Implications for the volcanic hazards: Mémoires de

Géologie (Lausanne), 44, pp. 26–66.

27. Volentik, A., Vanderkluysen, L., Principe, C., 2002. Stratigraphy of the caldera walls of Nisyros
volcano, Greece. Eclogae Geologicae Helvetiae 95, 223–235.
28. Nomikou, P., Papanikolaou, D., 2011. Extension of active fault zones on Nisyros volcano across
the Yali-Nisyros Channel based on onshore and offshore data. Marine Geophysical Research
32, 181–192
29. Tibalbi, A., Pasquarè, F.A., Papanikolaou, D., Nomikou, P., 2008. Tectonics of Nisyros Island,
Greece, by field and offshore data, and analogue modeling. Journal of Structural Geology 30,
1489–1506
30. Volentik, A., Principe, C., Vanderkluysen, L., Hunziker, J.C., 2005b. Explanatory notes on the
“Geological Map of Nisyros Volcano (Greece)”. In: Hunziker, J.C., Marini, L. (Eds.), The
Geology, Geochemistry and Evolution of Nisyros Volcano (Greece). Implications for the
Volcanic Hazards: Mémoires de Géologie (Lausanne), 44, pp. 7–25.
31. Moore, D.M. & Reynolds, R.C. Jr. (1997): X-Ray diffraction and the identification and analysis
of clay minerals.Oxford University press.
32. Spandler C., Martin L.H.J., Pettke T. (2012) Carbonate assimilation during magma evolution at
Nisyros (Greece), South Aegean Arc: Evidence from clinopyroxenite xenoliths, Lithos 146–147
18–33.
33. Leybourne, M.I., Johannesson, K.H., 2008. Rare earth elements (REE) and yttrium inbstream
waters, stream sediments, and Fe–Mn oxyhydroxides: fractionation, speciation, and controls
over REE + Y patterns in the surface environment. Geochim. Cosmochim. Acta 72, 5962–5983.
34. Baltatzis E., Kostopoulos D., Godelitsas A., Zachariadis P., Papanikolaou D. Pliocene
tourmaline rhyolite dykes from Ikaria Island in the Aegean back-arc region: geodynamic
implications Geodinamica Acta 2214 (2009) 189- 199.
35. Chiodini, G., Brombach, T., Caliro, S., Cardellini, C., Marini, L., Dietrich, V., 2002.m
Geochemical indicators of possible ongoing volcanic unrest at Nisyros Island (Greece).
Geophys. Res. Lett. 29 (16). doi:10.1029/2001GL014355.
36. Caliro, S., Chiodini, G., Galluzzo, D., Granirei, D., La Rocca, M., Sacorotti, G., Ventura, G.,
2005. Recent activity of Nisyros volcano (Greece) inferred from structural, geochemical and
seismological data. Bull. Volcanol. 67, 358–369.
37. Dotsika, E. Poutoukis, D., . Michelot J.L, Raco, B.Natural tracers for identifying the origin of
the thermal fluids emerging along the Aegean Volcanic arc (Greece): Evidence of Arc-Type
Magmatic Water (ATMW) participation Journal of Volcanology and Geothermal Research 179
(2009) 19–32
38. Chiodini, G., Cioni, R., Leonis, C., Marini, L., Raco, B., 1993. Fluid geochemistry of Nisyros
island, Dodecanese, Greece. Journal of Volcanology and Geothermal Research 56, 95– 112.
39. Kavouridis, T., Kuris, D., Leonis, C., Liberopoulou, V., Leontiadis, J., Panichi, C., La Ruffa,
G., Caprai, A., 1999. Isotope and chemical studies for a geothermal assessment of the island of
Nisyros (Greece). Geothermics 28, 219–239.
40. Ambrosio M., Doveri M. , Fagioli M. T., Marini L., Principe C., Raco B., Water–rock
interaction in the magmatic-hydrothermal system of Nisyros Island (Greece), 2010. Journal of
Volcanology and Geothermal Research 192, 57–68.
41. Markopoulos A., Stamatakis, M., Marantos I., 1988. Crystallisation of new-age minerals in the
caldera material of Nisyros. Bull. Geol. Soc. Greece XX/2, 393-404.
42. Bayliss P., Kolitsch U. Nichele H., Pring A., 2010. Alunite supergroup: recommended
nomenclature Mineralogical Magazine, 74(5), 919–927.
43. Sonnag I., Altherr R., Ludwig T., 2006. Li, Be, and B in plagioclase phenocrysts from Nisyros,

Greece Goldschmidt Conference Abstracts

44. Wyers, G.P., Barton, M., 1989. Polybaric evolution of calc-alkaline magmas from Nisyros,
southeastern Hellenic Arc, Greece. Journal of Petrology 30, 1–37.
45. Dalou C, Koga KT, Hammouda T, and Poitrasson F (2009) Trace element partitioning between
carbonatitic melts and mantle transition zone minerals: Implications for the source of
carbonatites. Geochimica et Cosmochimica Acta 73: 239–255
46. Tiepolo M, Vannucci R, Oberti R, Foley S, Bottazzi P, and Zanetti A (2000) Nb and Ta
incorporation and fractionation in titanian pargasite and kaersutite: Crystal-chemical constraints
and implications for natural systems. Earth and Planetary Science Letters 176: 185–201.
47. Nair A. G., Babu S.D.S., Damodaran K.T., . Shankar R, Prabhu C.N., 2009. Weathering of
ilmenite from Chavara deposit and its comparison with Manavalakurichi placer ilmenite,
southwestern India. Journal of Asian Earth Sciences 34(2):115-122.
DOI: 10.1016/j.jseaes.2008.03.005.
48. Bachmann, O., Deering, C.D., Ruprecht, J.S., Huber, C., Skopelitis, A., Schnyder, C., 2011.
Evolution of silicic magmas in the Kos–Nisyros volcanic centre, Greece: a petrological cycle
associated with caldera collapse. Contributions to Mineralogy and Petrology 163, 151–166.
49. Pasero,M., Kampf, A.R., Ferraris, C., Pekov, I.V., Rakova, J.,White, T., 2010. Nomenclature of
apatite supergroup minerals. Eur. J. Mineral. 22, 163–179.
50. Deditius, A.P., Utsunomiya, S., Reich, M., Kesler, S.E., Ewing, R.C., Hough, R., Walshe,J.,
2011. Trace metal nanoparticles in pyrite.Ore Geol. Rev. 42, 32–46.
51. Reich, M., Becker, U., 2006. First-principles calculations of the thermodynamic mixing
properties of arsenic incorporation into pyrite and marcasite. Chemical Geology 225, 278–290
52. Tomkins, A.G., Pattison, D.R.M., Frost, B.R., 2007. On the initiation of the metamorphic sulfide
anatexis. Journal of Petrology 48, 511–535
53. Large,R.R., Danyushevsky,L., Hollit,C., Maslennikov,V., Meffre,S., Gilbert,S., Bull,S.
,Scott,R., Emsbo,P., Thomas,H., Singh,B., Foster,J.,2009. Gold and trace element zonation in
pyrite using a laser imaging technique: Implications for the timing of gold inorogenic and carlin
– style sediment-hosted deposits .Econ.Geol.104,635–668.
54. Buettner, A., Kleinhanns, I.C., Rufer, D., Hunziker, J.C., Villa, I.M., 2005. Magma generation
at the easternmost section of the Hellenic Arc; Hf, Nd, Pb and Sr isotope geochemistry of
Nisyros and Yali volcanoes (Greece). Lithos 83, 29–46.
55. Vanderkluysen, L., Volentik, A., Hernandez, J., Hunziker, J.C., Bussy, F., Principe, C., 2005.
The petrology and geochemistry of lavas and tephras of Nisyros Volcano (Greece). In: Hunziker,
J.C., Marini, L. (Eds.), The Geology, Geochemistry and Evolution of Nisyros Volcano (Greece).
Implications for the volcanic hazards: Mémoires de Géologie (Lausanne), 44, pp. 79–99.
56. Le Maitre, R.W., et al., 1989. A classification of igneous rocks and glossary of terms:
recommendations of the International Union of Geological Sciences Subcommission on the
Systematics of Igneous Rocks. Blackwell Scientific Publications, Oxford, U.K.
57. Tomlinson E.L . Kinvig H.S, . Smith V.C , BlundyJ.D. , GottsmannJ, Müller W. , Menzies M.A.
The Upper and Lower Nisyros Pumices: Revisions to the Mediterranean tephrostratigraphic
record based on micron-beam glass geochemistry Journal of Volcanology and Geothermal
Research 243–244 (2012) 69–80
58. Wood, B.J., Blundy, J.D., 2001. The effect of cation charge on crystal-melt partitioning of trace
elements. Earth and Planetary Science Letters 188, 59–71.
59. Francis, D., Minarik, W., 2008. Aluminum-dependent trace element partitioning in clinopyroxene.
Contributions to Mineralogy and Petrology 156, 439–451.

Session VI‐B: REE Ore formation


RARE EARTH ELEMENTS IN KARST‐BAUXITES: A NOVEL

UNTAPPED EUROPEAN RESOURCE?
Éimear DEADY1*, Evangelos MOUCHOS2, Kathryn Goodenough3, Ben
Williamson2 and Frances Wall2
1
British Geological Survey, Kingsley Dunham Centre, Nicker Hill, Keyworth, Nottingham, NG12
5GG, UK
2
Camborne School of Mines, University of Exeter, Penryn, Cornwall, TR10 9FE, UK
3
British Geological Survey, Murchison House, West Mains Road, Edinburgh EH9 3LA, UK
*eimear@bgs.ac.uk
Abstract
Karst‐bauxite deposits form as a result of the accumulation of residual clay minerals in
depressions on a karst limestone surface, and their subsequent lateritic weathering.
Rare earth elements (REE) become concentrated in the bauxite deposits due to
crystallisation of authigenic REE‐bearing minerals, accumulation of residual phases
and the adsorption of ions on clays and other mineral surfaces. All the REE are
concentrated in the red mud waste generated by alumina production from bauxite
through the Bayer process. Red muds thus contain on average 900 ppm REE compared
with typical values of <100 ppm to ~500 ppm REE in the bauxites. Extraction of REE
from red mud has been shown to be feasible although it is challenging due to the
heterogeneous spatial distribution of REE in the bauxites and the need for development
of appropriate processing methods. With annual European production of bauxite
estimated to be approximately 3.5 million tonnes per annum, resulting in
approximately 140 000 tonnes of red mud, understanding the REE resource potential
of bauxites is integral to the assessment of European REE resources.

Introduction
Research into the rare earth element (REE) content of bauxites has a long history, with
authigenic rare earth minerals first identified in the San Giovanni Rotondo deposit in
Italy1. Further research by Maksimović and Pantó2,3,4,5,6,7,8 in particular led to the
discovery of authigenic REE in multiple deposits in the Balkans and Greece (for
example at Marmara, Greece, Vlasenica, Bosnia and Herzegovina and Štitovo,
Montenegro). Research into the potential of bauxites as a source of REE, Sc and Y has
been undertaken by many researchers over the past forty years (e.g. 9,10,11,12,13,14,15).
This paper reviews this previous work to consider the REE resource potential of
bauxites in Europe.


Bauxite is formed from the intense lateritic weathering of residual clays, which
accumulate in topographic lows on continental surfaces. The two main classes of
bauxites are: 1) those that form on aluminosilicate lithologies; and 2) those that form
on limestone, known as karst bauxites16, and which are the focus of this paper. Bauxite
is the primary source for aluminium (Al) production globally and is relatively common
in Europe, with deposits known from the majority of Mediterranean countries and
intermittent exploitation occurring over many decades (Figure 1).

8
5
7
2 6
1
3
4

Figure 1: Map of southern Europe showing selected bauxite occurrences as mentioned in the text:
Olmedo, Sardinia, Italy (1), San Giovanni Rotondo, Italy (2); Parnassos‐Ghiona, Greece (3), Marmara,
Greece (4); Vlasenica, Bosnia and Herzegovina (5); Štitovo pit, Nikšić, Montenegro (6); Grebnik, Kosovo
(7); Nagyharsany, Transdanubian Central Range, Hungary (8).

Demand for REE is increasing, particularly for use in green energy technologies, e.g.
wind turbines etc. As China currently dominates global production17, there is
increasing pressure to identify alternative European sources of these economically
important metals. The growing demand for REE has recently been emphasised by
numerous authors18,19,20, and while recycling may contribute to meeting future
demand, the separation of individual REE from recycled materials is very challenging
with only 1 per cent of REE currently recycled from end‐of‐life products19 (and
references therein). There are currently no mines producing REE in Europe and
although this is likely to change in the future, an alternative, readily available, source
of REE would be attractive.

Red muds produced from alumina processing represents a potentially important
concentration of REE as it has been shown that all of the REE pass through the alumina
extraction process into the waste, and the total REE values are enriched by a factor of
two from the original bauxite ore11,13. Currently red mud is viewed as waste as it does
not have any other purpose, although there have been attempts to find alternative
uses for the material21 (and references therein). .

Bauxite
Karst bauxite deposits are formed by accumulation of residual clays, derived from the
weathering of aluminosilicate minerals22, in depressions on carbonate
palaeotopography. The contents of the depressions are bauxitised through deep
lateritic weathering and desilicification of ferrilitic soils23. The conditions under which
bauxitisation occurs are such that silica is incongruently dissolved from minerals such
as feldspar and kaolin leaving an aluminium‐rich residue24. Bardossy16 distinguishes
several subtypes: Kazachstonian type, Timan type, Ariège type, Tula type and
Mediterranean type. These classifications are based on the composition of the
deposits and the mode of karstification.

Figure 2: Schematic diagram showing the typical classic Mediterranean type karst
bauxite (after Bardossy 198216).

Mediterranean type karst bauxite deposits formed on both the European and Adriatic
Mesozoic carbonate shelves in the Neotethys realm during the Mesozoic to Early
Cenozoic25,26. Bauxitisation in this area occurred from the surface downwards leading
to the ultimate loss of SiO2 and Fe2O3 and accumulation of Al2O3 with depth23. This
paper focuses on Mediterranean type karst bauxite.
Rare Earth Elements in Bauxite
Where REE are available to the weathering system, they can be adsorbed onto the
surfaces of clay residues that form the bauxite deposit and are then concentrated with
depth by bauxitisation27,28. With more intense bauxitisation there is a correlation with
depth between both increased REE content and decreased silica content. Meteoric
water can preferentially leach away silica produced from the dissolution of feldspar
and kaolinite leaving an Al‐rich residue29. However, the process of REE concentration
is dependent on multiple factors and is still poorly understood.

Authigenic rare earth element‐bearing minerals
Where sufficient REE are available within the bauxite, bauxitisation has occurred in
situ and where there has been no later transport or erosion of the deposit, authigenic
REE‐bearing minerals can form. REE are mobilised from the clay as waters percolate
through the system during bauxitisation. As REE‐bearing fluids descend through the
deposit and reach the carbonate they become physically trapped allowing authigenic
minerals to form where REE content is sufficiently high 27,28; the most common is
hydroxylbastnäsite‐(Nd) (Nd(CO3)(OH))14,30. In order for REE minerals to develop, the
initial concentration of REE adsorbed onto the clay must be greater than about 1000
ppm14. If the initial concentration does not reach this threshold REE minerals are
unlikely to form. Increases in REE content of up to almost six‐fold have been recorded
as a result of bauxitisation (e.g. at Nikšić, Montenegro and Grebnik, Kosovo31 (and
references therein)). Where the bauxites have been disturbed during formation REE
can be lost from the system, such as in the karst‐bauxite deposits in the Transdanubian
Central Range in Hungary32. In Hungary reported REE concentrations vary greatly,
ranging from ~81–~10 000 ppm. This is due to the detrital nature of some REE host
minerals such as monazites sourced from the Mecsec mountains in the case of the
Nagyharsany deposit, rather than authigenic mineral growth 9.
Conditions of mineral formation
Key parameters controlling the concentration of REE in bauxite include intensity of
leaching by meteoric waters, pH, Fe content and the availability of ligands such as
fluorine (F).
Water flow: Leaching of the bauxite by meteoric water can readily remobilise adsorbed
REE14. The weathered limestone below the bauxite facilitates drainage and thus clay
layers do not form within the bauxite, so the downward mobility of REE is not
restricted33,26.
pH: The alkaline nature of the underlying limestone creates a high pH that results in
increased stability of REE‐carbonate complexes (e.g.34,35). Stability increases with
atomic number, allowing HREE to be preferentially retained in solution as carbonate
ions25. The distribution of REE along this ‘pH barrier’ is heterogeneous and the REE can
be concentrated in lenses or in micropores, space fillings and microveins14.
Fe content: Fe‐rich goethite has a significant scavenging effect, preferentially fixing
mobile REE, in particular the LREE, within residual Fe‐rich layers36. Positive correlations
exist between Fe2O3 and ∑REE, particularly where Fe oxy‐hydroxides occur as
ferrigenous ooids (e.g. Olmedo Bauxite, Italy25).
F depletion: Bauxite is typically depleted in F; this depletion is a critical factor in
controlling the type of minerals that form. F can be substituted by ‐OH groups to form
hydroxylbastnäsites such as hydroxylbastnäsite‐(Nd), which has been described from

Nikšić in Montenegro. Hydroxylbastnäsites are the most common authigenic REE
minerals found in bauxite6.

(m)
0 Nd
red bauxite
Sample point
6
10 100 600 3000 15000
ppm
Figure 3: Typical Nd distribution through the Stitovo bauxite, Montenegro (after
Maksimovic and Panto 1995 (figure10.1 chapter 10 p262)14). The enrichment of Nd at
the base allows for the formation of authigenic REE minerals such as Nd‐goyazite,
bastnäsite and hydroxylbastnäsite‐(Nd)37,28.

Ce does not have the same per descensum correlation that is observed for the other
REE. This is due to the oxidation of Ce2+ to Ce4+ in the strongly oxidising environment
created during bauxitisation28,38,39. This can result in the skewed distribution of Ce
within bauxite profiles with Ce enrichment occurring in the upper parts of the bauxite
deposit whilst the rest of the REE have been transported downwards.
Rare Earth Elements in Red Mud
To separate alumina from bauxite, a processing route known as the Bayer process is
employed. Bauxite is digested in a hot sodium hydroxide solution to convert the
alumina to aluminium hydroxide, which dissolves in the hydroxide liquor. The other
components of bauxite do not dissolve. The solution is clarified by filtering off the solid
impurities, which forms the so‐called bauxite residue or red mud. The actual
composition of red mud depends on the type of bauxite, the mining location and the
process parameters of the Bayer Process40. REE content in red mud has been shown
to double as a result of the Bayer Process11 and this is consistent in both Jamaican and
Greek deposits10. Mineralogical studies have shown that in the Jamaican red mud
samples the REE are dispersed as ion‐adsorbed elements or as an unknown phase
associated with iron oxides41. The resource potential of red muds in Europe (including
Turkey) is significant with approximately 3.5 million tonnes (Mt) of bauxite ore
extracted in 201217, which resulted in the production of an estimated 140 000 tonnes
(t) of red mud waste.

Rare Earth Elements in Greek Bauxites
The karst‐bauxite deposits of Greece are among the world’s most important sources
of bauxite. Exploitable deposits of bauxites exist mainly in the regions of Mt Parnassus,
Mt Giona and Mt Helikon (central Greece) and reserves are approximately 600 000 t –
the 9th largest bauxite reserves globally42. Previous research on these deposits includes
the quantification of REE and Sc in bauxites and red mud waste43,11, and development
of separation techniques for Sc and REE from the red mud12.
Parnassos‐Ghiona bauxite
The Parnassos‐Ghiona geotectonic zone is characterised by nearly continuous
sedimentation of epicontinental reef‐like carbonates from the Upper Triassic to the
Upper Cretaceous. Bauxites occur in three main horizons (B1–B3) (Figure 4).
Geochemical analysis of samples from this region has been conducted in this study and
the results are presented below in Figure 5 and Table 1.

Eocene flysch overlain by
Quaternary conglomerates

Upper Cretaceous limestone

B3 horizon

Lower Cretaceous limestone

B2 horizon

Upper Jurassic limestone

B1 Horizon

Lower Jurassic limestone

Triassic dolomitic limestone

Figure 4: Stratigraphic column showing the three bauxite horizons (after Laskou and
Economou‐Eliopoulos, 2013)50

1000
B1
100 B2
REE/Chondrite
B3
MB
RM
10 RM N
1
La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Y
Figure 5: Plot of average published REE values for each of the three main horizons (B1–
B3) and red mud (RM); and preliminary data based on 18 samples taken from three
dehydrated red mud waste tips at the Aluminium S.A. site (RM N). Data for mixed
bauxite ore (MB) that enters the plant is shown. All values are chondrite normalised
using McDonough and Sun 199644. [Data from33,43,11,45,46,47,48,49,50].

Tsirambides and Filippidis53 report ΣREE values in the bauxites and lateritic bauxites of
central Greece as 3275 to 6378 ppm. Although these values are significant, these
concentrations are unlikely to be reflective of all Greek bauxite deposits. High values
are likely to result from the sampling of localised areas with REE enrichment along the
footwall limestone, and are not representative of average REE concentrations in the
bauxite. Chondrite‐normalised data for average REE values from the three different
bauxite horizons of the Parnassos‐Ghiona zone are shown in Figure 5. The combined
REE values for bauxite horizons B2 and B3 (Figure 4) were estimated using samples of
the mixed throughput ore from the plant at Aluminium S.A., Greece (sample MB). Data
compiled for this study indicates an average ΣREE+Y value of ~500 ppm for the bauxite
and ~900 ppm for the red mud samples (Table 1). REE concentrations in the collated
data indicate variation across the three main bauxite horizons, with a mixture of these
horizons entering the processing plant (MB). The two‐fold increase in REE in the red
muds compared with the original bauxites is consistent with the literature e.g.11.


B1 B2 B3 MB RM RM N

(ppm) (ppm) (ppm) (ppm) (ppm) (ppm)
Y 44.4 159.1 48.03 55.44 93.11 85.1
La 54.5 258.35 64.08 80.2 150.59 122.19
Ce 178 427.53 195 259.59 421.62 394.92
Pr 11.5 42.95 7.85 13.8 25.62 29.02
Nd 62.33 149.23 60.39 59.86 120.9 107.92
Sm 8.4 57.2 7.36 13.81 29.09 21.4
Eu 1.95 12.99 1.87 2.87 4.97 4.59
Gd 7.2 61 7.1 12.53 23.27 17.86
Tb 1.1 8.66 1.82 1.3 3.17
Dy 6.9 46.88 8.2 7 14.44 18.51
Ho 1.4 9.16 1.9 2.15 4.33 3.8
Er 4.1 23.46 5.74 8.07 17.2 11.24
Tm 0.5 3.27 0.99 1.32 1.89
Yb 4.3 19.01 6.14 7.4 15.6 12.33
Lu 0.7 2.73 0.94 1.38 2.42 1.99
ΣREE+Y 387.28 1281.52 417.41 526.72 923.16 835.93
Table 1: REE values compiled from the literature for B1–3, RM and MB samples. New
data is presented as RM N. Samples were analysed for ∑REE+Y content in the
geochemical laboratories of the Camborne School of Mines.

The variation in REE concentration between specific horizons, particularly the
enrichment observed in B2 is likely to be a function of sampling an enriched zone,
which is not necessarily reflective of the entire horizon. The vertical, and probable
lateral, variation in REE concentrations illustrate the need for a robust sampling
strategy involving the collection of more carefully characterised samples.
Bauxite as a potential resource
Production of REE in the form of RE oxides (REO) is currently limited to six countries:
Russia, USA, China, Brazil, Malaysia and Australia17. With total production of bauxite
from Europe (+Turkey) of ~3.5 Mt17, this results in annual production of ~140 000 t of
waste red mud, with an average homogenised REE content of ~900 ppm, based on
ΣREE+Y from sample RM in Table 1. If these quantities of red muds were processed
and REE extracted, it would result in total REE production of ~1300 t annually. In a
best case scenario, with REE enriched bauxite having an average ΣREE concentration
of ~2000 ppm14, this would result in REE production of a ~2800 t annually. This is based
on REE extracted from the Bayer Process would likely be in the form of oxalates or
carbonates, which have a relatively low market value as additional costs are incurred
in separating the individual REEs. In Greece, Aluminium S.A. is the largest consumer

of Greek bauxite, processing over 1.5 Mt annually and producing 700 000 t of red mud
waste annually51,52. Deposition of dehydrated waste mud started in 2009. Accordingly
the total volume accumulated is approximately 3.5 Mt. Assuming an average ΣREE of
1000 ppm, if processed this could potentially produce ~3500 t of contained REE. It is
notable that the data also show a relatively high content of HREE relative to LREE,
which may have economic implications. However, there are challenges associated with
the exploitation of red mud waste, specifically the difficulty in controlling REE
concentrations. REE concentration varies significantly between deposits and within
individual bauxite horizons and this directly affects the REE concentration of the
resulting red muds.
Discussion
Fully understanding the mineralogy and natural processes of REE enrichment in
bauxites is vital to assessing the REE resource potential of bauxites and red muds in
Europe. Great variability in both the mineral concentration and composition between
and along bauxite horizons is documented14. This means acquiring red muds with
consistent REE concentrations to feed a processing plant would be challenging.
However, there are occurrences which are more enriched in particular REE, such as
the Montenegro deposits at Niksic, which host hydroxylbastnäsite‐(Nd) as the
predominant REE mineral. Improved understanding of the heterogeneity of REE
distribution within the bauxite could allow for the targeting of REE enriched portions
of the bauxite. It is likely that once the supply of REE becomes more geographically
diverse there will be greater emphasis on acquiring metals from sources with the
lowest environmental impact19. Existing stockpiles of red muds could represent a more
environmentally sustainable source of REE for Europe as additional mining is not
necessary and current “stockpiles” of red mud would be viewed as a resource rather
than a waste product with the associated disposal issues.
Although the quantities of REE contained in red muds appear to be low (average ~900
ppm) when compared with primary deposits of carbonatite (e.g. Mountain Pass and
Mount Weld have grades of about 80 000 ppm REE19), in a scenario based upon a REE
grade of 2000 ppm (~2800 t contained REE) red muds could account for about 10 per
cent of EU demand for REE. EU imports of rare earths in 2011 (REE+Y+Sc metals and
compounds) totalled ~26 500 t54.
More significantly, it may contribute to improving security of supply for some REE to
the European market. European production of REE in high demand would ensure a
secure supply source and remove some of the risks associated with the importation of
REE from geopolitically unstable countries.
The economic viability of REE production from European red muds needs careful
assessment due to the investment required in mineral processing and extraction
technology. Further research focused on improved characterisation of red mud

stockpiles, selective mining of REE enriched bauxites, and the development of efficient
REE recovery techniques from red muds is required to fully assess these potentially
important resources.
Acknowledgments
ED and KG publish with the permission of the Executive Director, British Geological
Survey (NERC). BGS © NERC 2014. All rights reserved. The authors would like to thank
Delphi‐Distomon S.A. and Aluminium S.A. for providing all bauxite and red mud
samples. The authors would also like to thank Mr Paul Lusty for his constructive
criticism of the manuscript.
References
6. G. Bárdossy and G. Pantó, “Trace mineral and element investigation on bauxites by electron
probe”. 3rd International Congress, ICSOBA, (International Committee for Study of Bauxite,
Alumina & Aluminium), Nice, France pp. 47‐53 1973.
7. Z. Maksimović and G. Pantó, “Minerals of the rare‐earth elements in karstic bauxites: synchysite‐
(Nd), a new mineral from Grebnik deposit”. 4th International Congress, ICSOBA, (International
Committee for Study of Bauxite, Alumina & Aluminium), Athens, Greece pp. 540‐552 1978.
8. Z. Maksimović and G. Pantó, “Bastnäsite‐(La) and monazite‐(Nd): A new variety of monazite from
the Marmara bauxite deposit (Greece)”. Bull. Acad. Serbe Sci. Arts, 20, 35‐42 (1980).
9. Z. Maksimović and G. Pantó, “Synchysite‐(Nd) from Grebnik bauxite deposit (Yugoslavia)”. Acta
Geologica Academiae Scientiarum Hungaricae, 24(2‐4), 217‐22 (1981).
10. Z. Maksimović and G. Pantó, “Mineralogy of yttrium and lanthanide elements in karstic bauxite
deposits”. Travaux ICSOBA (International Committee for Study of Bauxite, Alumina & Aluminium),
18, 191‐200 1983.
11. Z. Maksimović and G. Pantó, “Hydroxyl‐bastnaesite‐(Nd), a new mineral from Montenegro,
Yugoslavia”. Mineralogical Magazine, 49, 717‐720 (1985a).
12. Z. Maksimović and G. Pantó, “Neodymian goyazite in the bauxite deposit of Vlasenica
(Yugoslavia)”. Tschermaks Mineralogische und Petrographische Mitteilungen, 34(2): 159 (1985b).
13. Z. Maksimović and G. Pantó, “The occurrence and genesis of the hydroxylbastnäsites from
Montenegro, Yugoslavia”. Bull. Acad. Serbe Sci. Arts Cl. Sci. Math. Natur., 27, 15‐20 (1987).
14. G. Bardossy, G Panto, G. Varhegyi, “Rare metals in Hungarian bauxites and conditions of their
utilization”. Travaux ICSOBA (International Committee for Study of Bauxite, Alumina &
Aluminium), 13, 221‐231 1976.
15. A. Wagh and W. Pinnock, “Occurrence of scandium and rare earth elements in Jamaican bauxite
waste” Economic Geology Vol. 82, pp. 757‐761 1987.
16. M. Ochsenkühn‐Petropoulou, T. Lyberopoulou, and G. Parissakis, “Direct determination of
lanthanides, yttium and scandium in bauxites and red mud from alumina production”, Analytica
Chimica Acta, 296, 305‐313 (1994)
17. M. Ochsenkühn‐Petropoulou, T. Lyberopoulou, and G. Parissakis, “ Selective separation and
determination of scandium from yttrium and lanthanides in red mud by a combined ion
exchange/solvent extraction method” Analytica Chimica Acta, 315, 231‐237 (1995).
18. M. Ochsenkühn‐Petropoulou, T. Lyberopoulou, K. M. Ochsenkühn and G. Parissakis, “Recovery of
lanthanides and yttrium from red mid by selective leaching”, Analytica Chimica Acta, 319, 249‐254
(1996).

19. Z. Maksimović and G. Pantó, “Authigenic rare earth minerals in karst‐bauxites and karstic nickel
deposits”. In: A.P. Jones, F. Wall and C.T. Williams, Rare earth minerals, chemistry, origin and
ore deposits, Chapter 10, 257‐279 (1996).
20. L. V. Tsakanika, M. Ochsenkühn‐Petropoulou, L. N. Mendrinos, “Investigation of the separation of
scandium and rare earth elements from red mud by use of reversed‐phase HPLC”, Annals
Bioanalytical Chemistry, 379, 796–802 (2004).
21. G. Bárdossy, “Karst Bauxites, Bauxite Deposits on Carbonate Rocks”. Elsevier, 444pp 1982.
22. T. Brown, N. Idoine, E. Raycraft, R. Shaw, E. Deady, J. Rippingale, T. Bide, C. Wrighton, J. Rodley,
“World Mineral Production 2008‐12” British Geological Survey, Keyworth, Nottingham, 2014.
23. A. Bloodworth, “Track flows to manage technology‐metal supply”, Nature, 505, 19‐20, (2014).
24. F. Wall, “Rare earth elements”. In: G. Gunn, Critical Metals Handbook, Chapter 13, 312‐339 (2014).
25. G. P. Hatch, “Dynamics in the global market for rare earths”, Elements, 8(5), 341‐346 (2012).
26. C. Klauber, M. Gräfe, G. Power, “Bauxite residue issues: II. options for residue utilization”,
Hydrometallurgy 108, 11–32, (2011).
27. P. Freyssinet, C.R.M. Butt, R. C. Morris and P. Piantone, “Ore forming process related to lateritic
weathering”, Economic Geology 100th Anniversary Volume, pp. 681‐722, (2005).
28. N. Gow and G. Lozej, “Bauxite”, Geosciences Canada 20 (1) (1993).
29. W. Bland and D. Rolls, “Weathering: an introduction to the basic principles”, Arnold, 271pp. 1998.
30. P. Mameli, G. Mongelli, G. Oggiano, E. Dinelli, “Geological, geochemical and mineralogical features
of some bauxite deposits from Nurra (Western Sardinia, Italy): insights on conditions of formation
and parental affinity”, Int J Earth Sci (Geol Rundsch) 96, 887–902, (2007).
31. I. Valeton, “Element concentration and formation of ore deposits by weathering”, Catena 21 99‐
129 (1994).
32. Z. Maksimović, “Genesis of some Mediterranean karstic bauxite deposits”. Travaux ICSOBA,
(International Committee for Study of Bauxite, Alumina & Aluminium), Zagreb, Croatia, 13, 1‐14
(1976).
33. Z. Maksimović, E. Roaldset, “Lanthanide elements in some Mediterranean karstic bauxite
deposits”, Travaux ICSOBA, (International Committee for Study of Bauxite, Alumina & Aluminium),
Zagreb, Croatia, 13, 199‐220 (1976).
34. L. Robb, “Introduction to ore‐forming processes”, Blackwell Publishing, 2005.
35. G. Mongelli, “Ce‐anomalies in the textural components of Upper Cretaceous karst bauxites from
the Apulian carbonate platform (southern Italy)”, Chemical Geology 140, 69–79, (1997).
36. Z. Maksimović and G. Pantó, “Contribution to the geochemistry of the rare earth elements in the
karst‐bauxite deposits of Yugoslavia and Greece”, Geoderma, 51, 93‐109, (1991).
37. Z. Maksimović, A. Mindszenty, G. Pantó, “Contribution to the geochemistry of Hungarian karst
bauxites and the allochtony/autochtony”, Acta Geol. Hung., 34 (4), 317‐334 (1991).
38. I. Valeton, M. Bierman, R. Reche, F.F. Rosenberg, “Genesis of nickel laterites and bauxites in Greece
during the Jurassic and the Cretaceous and their relation to ultrabasic rocks”. Ore Geol Rev 2 359–
404 (1987).
39. K.H., Johannesson, W.B., Lyons, K.J., Stetzenbach, R.H., Byrne,”The solubility control of rare earth
elements in natural terrestrial waters and the significance of PO43– and CO32– in limiting dissolved
rare earth concentrations: a review of recent information”. Aquat Geochem 1, 157–173(1995).
40. K.H. Johannesson, K.J., Stetzenbach, V.F. Hodge, W. B. Lyons, “Rare earth element complexation
behavior in circumneutral pH groundwaters: assessing the role of carbonate and phosphate ions”,
Earth Planet Science Letters 139, 305–319 (1996).
41. R.A.,Künhel, “The role of cationic and anionic scavengers in laterites”, Chemical Geololgy 60,3 1–
40, (1987).
42. Z. Maksimović, “Trace elements in some Yugoslav bauxite deposits and their significance”. In: Proc.
IV Jugosl. Symp. on exploration of bauxites, Herceg Novi, 29‐33, 1976.

43. S. Koicki, A. Koicki, Z. Maksimovic, Neutron activation analysis of lanthanides in Yugoslav
bauxites. Glas ‐ Srpska Akademija Nauka i Umetnosti, Odeljenje Prirodno‐Matematickih Nauka, 46,
37‐48 (1980).
44. Z. Maksimović, N. Skarpelis, G. Pantó, “Mineralogy and geochemistry of the rare earth elements in
the karstic nickel deposit of Lokris area, Greece”, Acta Geol. Hungarica, 36 (3), 331‐342, (1993).
45. A. R. Hind, S. K. Bhargava, Stephen C. Grocott, "The surface chemistry of Bayer process solids: a
review", Colloids and Surfaces A : Physicochem. Eng. Aspects, 146, 359–374, (1999).
46. A.N. Mariano and A. Mariano Jr., “Rare earth mining and exploration in North America”, Elements,
8(5), 369‐376 (2012).
47. United States Geological Survey, “Bauxite and Alumina”, U.S. Geological Survey, Mineral
Commodity Summaries, February 2014.
48. M. Ochsenkühn‐Petropulou, K. Ochsenkühn and J. Luck, “Comparison of Inductively Coupled
Plasma Mass Spectrometry with Instrumental Neutron Activation Analysis for the determination
of Rare Earth Elements in Greek bauxites”. Spectrochim. Acta 46B, 51‐65 (1991).
49. W.F. MacDonough and S‐s. Sun, “The composition of the Earth”, Chemical Geology, 120, 223‐253,
(1995).
50. K. M. Ochsenkühn, M. Ochsenkühn‐Petropoulou and G. Parissakis, “Activation analysis of bauxitic
materials by epithermal irradiation, Journal of Radioanalytical and Nuclear Chemistry”, 190 (1), 75‐
79. (1995).
51. M. Laskou and G. Andreou, “Rare earth element distribution and REE‐minerals from the Parnassos‐
Ghiona bauxite deposits, Greece”, In: Eliopoulos D. et al. (Eds), “Mineral Exploration and
Sustainable Development”, 7th Biennial SGA Meeting, Athens, Millpress, Rotterdam, 89‐92 (2003).
52. M. Laskou, “Pyrite‐rich bauxites from the Parnassos‐Ghiona zone, Greece”. In: MAO et al. (Eds),
8th SGA Meeting, “Mineral Deposits Research Meeting the Global Challenge” Beijing, 1007‐1010,
(2005).
53. M. Laskou and M. Economou‐Eliopoulos, “The role of micro‐organisms on the mineralogical and
geochemical characteristics of the Parnassos‐Ghiona bauxite deposits, Greece”. J. Geochem.
Exploration, 93, 67–77, (2007).
54. M. Laskou, M. Economou‐Eliopoulos, I. Mitsis, “Bauxite ore as an energy source for bacteria driving
iron‐leaching and bio‐mineralization”. Hellenic Journal of Geosciences, 45, 163‐174, (2011).
55. M. Laskou, and M. Economou‐Eliopoulos, “Bio‐mineralization and potential biogeochemical
processes in bauxite deposits: genetic and ore quality significance”, Mineralogy and Petrology, 107
(4), 171‐186, (2013).
56. A. Tsirambides and A. Filippidis, Metallic mineral resources of Greece. Central European Journal of
Geosciences, 4(4), 641‐650, (2012).
57. Ch. Anagnostou, “Bauxite resource exploitation in Greece vs. Sustainability”. Bulletin of the Geol.
Soc. Of Greece. Proc 12th international congress, Patras, 2010.
58. A. Tsirambides and A. Filippidis, “Exploration key to growing Greek industry”, Industrial Minerals
44–47, (2012).
59. T. Brown, S. Hobbs, A. Mills, E. Petavratzi, R. Shaw, T. Bide, “European Minerals Statistics 2007‐11”
British Geological Survey, Keyworth, Nottingham, 2013.


THE IMPORTANCE OF TECTONIC SETTING IN ASSESSING
EUROPEAN RARE EARTH POTENTIAL
Kathryn M GOODENOUGH1,*, Eimear DEADY 2, Frances WALL3, Richard A
SHAW2,Paul LUSTY2
1
British Geological Survey, West Mains Road, Edinburgh, EH9 3LA, UK
2
British Geological Survey, Nicker Hill, Keyworth, Nottingham, NG12 5GG, UK
3
Camborne School of Mines, University of Exeter, Penryn, Cornwall, TR10 9FE, UK

* kmgo@bgs.ac.uk
Abstract
Rare earth element (REE) resources are commonly found associated with alkaline
igneous complexes or carbonatites, or as secondary deposits derived from igneous
rocks. Globally, many REE deposits occur around the margins of Archaean cratons,
most in continental rift zones. Europe contains many such rift zones, which are
generally younger in the south. Many of these rifts are intracontinental, whereas
others are associated with the opening of oceans such as the Atlantic.
All these rift systems have the potential to host REE resources, but whereas the older
provinces of northern Europe are deeply exposed, exposures in southern Europe are
largely at the supracrustal level. This paper considers how an understanding of the
tectonic setting of Europe’s REE resources is vital to guide future exploration.
Introduction
Although the rare earth elements (REE) can be concentrated by geological processes
in a range of tectonic settings, the most significant concentrations are typically
associated with alkaline‐peralkaline silicate rocks and carbonatites formed in
extensional intracontinental rifts1. Such rifts are found in all the continents, dating
from the Archaean to the present day. However, the ultimate source of alkaline and
carbonatitic magmatism within the rifts remains controversial: one hypothesis states
that the emplacement of these magmas is largely controlled by structures in the
underlying lithosphere2, whilst the contrasting view invokes a role for mantle plumes3.

This controversy is of the utmost importance for REE exploration. If mantle plumes are
the most important control, REE enrichments could potentially be found anywhere –
even in ocean islands. Alternatively, if the lithosphere is the most important control, it
is perhaps more likely that REE enrichments would be focused in areas with long and
complex lithospheric histories. This paper will assess alkaline magmatism across
Europe in this context, and consider the implications for REE resources.

Intracontinental rifts in Europe
Europe has a long and complex geological history, from the Archaean cratons of
Greenland and northern Scandinavia, to volcanic suites around the Mediterranean
that are only a few years old. This history comprises several cycles of continental
break‐up and collision (Wilson Cycles) and thus includes many intracontinental rifts
(Figure 1) that represent potential REE metallogenetic belts.

Archaean alkaline silicate and carbonatite magmatism is rare across the globe4. In
Europe, the main examples date from c. 2700 Ma and include the Skjoldungen alkaline
province, Greenland and the Siilinjärvi carbonatite in Finland5. Although these are
enriched in the REE, as is typical for alkaline magmas, they are considered unlikely to
host economic REE mineralisation; Siilinjärvi is currently mined for phosphate.

Figure 1: Map of Europe showing the approximate extent of some of the major alkaline
provinces mentioned in the text. Yellow stars indicate important REE deposits. 1:
Sarfartoq; 2: Qeqertaasaq; 3: Kvanefjeld, Kringlerne and Motzfeldt; 4: Fen; 5: Norra
Kärr; 6: Alnö; 7: Sökli.
By contrast, some of Europe’s most important REE mineralisation is associated with
rifts of Proterozoic age. The Gardar Province of south‐west Greenland was emplaced
along the boundary between the North Atlantic Craton and a Palaeoproterozoic

mobile belt to the south. The province comprises two periods of rift‐related
magmatism, one at c. 1280 Ma and a second at 1180‐1140 Ma6. Both episodes are
characterised by extensive dykes and emplacement of alkaline to peralkaline plutonic
complexes. The Motzfeldt centre of the Igaliko complex, and the Ilimaussaq complex,
both comprise peralkaline syenites with zones that have been enriched in REE by
magmatic and/or hydrothermal processes. The Ilimaussaq complex contains two
major REE deposits, Kvanefjeld and Kringlerne. Other syenitic and carbonatitic
complexes within the Gardar Province are the focus of current REE exploration.

In Fennoscandia, the Mesoproterozoic was a time of complex accretionary tectonics,
with collisional zones separated spatially and temporally by localised rifting and
magmatism. Episodic rift‐related magmatism occurred within Fennoscandia
throughout much of the Mesoproterozoic7, 8, although distinct rift‐related igneous
provinces have not been identified. Examples of alkaline intrusions formed at this time
include the Norra Kärr and Almunge syenites in southern Sweden, of which the former
is recognised as an important HREE‐enriched deposit.

Although the Neoproterozoic and early Palaeozoic were marked by rifting in Europe as
the supercontinent Rodinia broke up, there is relatively little evidence of alkaline
magmatism at this time. The exceptions are zones of carbonatite and lamprophyre
intrusion in Greenland and Scandinavia. In west Greenland, one such zone is dated at
555‐605 Ma9 and includes the Sarfartoq carbonatite which is currently being explored
for REE. In Scandinavia, the Fen carbonatite complex in Norway and the Alnø alkaline‐
carbonatite complex in Sweden have both been dated at c. 580 Ma10 and are known
to contain zones of REE enrichment. Rift‐related magmas were also emplaced
elsewhere in Europe at this time, but appear to have been strongly affected by crustal
assimilation, such that alkaline compositions are rarely seen.

In the aftermath of the Caledonian and Variscan orogenies, rifting and alkaline
magmatism developed in many areas across northern Europe. The most dramatic of
these rifts is the Kola Alkaline Province, which was emplaced into the Fennoscandian
Shield during the Devonian. Notably, this area had been the site of repeated alkaline
magmatism during the Precambrian11. The province comprises a number of alkaline
ultramafic, syenite and carbonatite complexes which extend across the Kola Peninsula
in Russia and into Finland. The most notable of these within Finland (and therefore in
the area reviewed by this paper) is the Sökli phoscorite‐carbonatite complex 12, which
has been extensively explored for phosphate.

Rifting and alkaline magmatism also developed through the Carboniferous and
Permian across the UK, southern Scandinavia (particularly the Oslo Rift), and into
northern Germany13. In the Oslo Rift, exposures include both volcanic successions and

alkaline plutons, some having notable enrichments in the REE. The UK Permo‐
Carboniferous magmatism in the Midland Valley of Scotland and Northern England is
at a higher erosional level, with volcanics, dykes and sills but no plutonic rocks
exposed.

The late Mesozoic and Cenozoic represent a major period of rifting in Europe, with
alkaline magmas being emplaced in many areas. In western Europe, Mesozoic to
Cenozoic rifting associated with the formation of the Atlantic Ocean generated large
areas of flood basalt with associated mafic to ultramafic layered complexes that do
not have elevated REE concentrations. However, some alkaline and carbonatitic
magmatism was generated. The Jurassic Qeqertaasaq and Tikiussaq carbonatites at
the margin of the North Atlantic Craton in West Greenland are currently being
explored for REE. In east Greenland, the Gardiner Complex is a Cenozoic alkaline
ultramafic, syenite and carbonatite intrusion which contains some REE‐enriched
zones. Plutons of similar age also occur in the UK, but are typically granitic in
composition and show only minor REE enrichment. Further south, late Cretaceous (94‐
72 Ma14) alkaline magmatism in the Iberian peninsula produced numerous syenitic
intrusions, forming the Iberian Alkaline Province.

Following collision during the Alpine orogeny and subduction around the
Mediterranean, Cenozoic rifts and associated alkaline magmatism developed across
much of central and southern Europe15. North of the Alpine collision zone, the Massif
Central in France, the Rhenish Massif and Rhine Graben in Germany, and the
Bohemian massif all include alkaline volcanic rocks, including some differentiated
compositions. Carbonatites are rare but those that have been identified are of interest
for their REE contents (e.g. the Storkwitz project in Germany). Alkaline volcanism also
occurs in many areas around the margins of the Mediterranean, particularly in the
Roman province of Italy and on Sicily and Sardinia, and in the Pannonian Basin and
Anatolian rifts further east. The current level of erosion means that in these areas, the
central complexes that might contain significant primary REE resources are still
hundreds of metres to kilometres below the surface.
Magmatism in intracontinental rifts and REE enrichments
Intracontinental rifting is a key part of the Wilson Cycle, and as such has occurred at
many times in the geological history of Europe. The majority of Europe’s largest REE
resources are associated with mantle‐derived alkaline magmatism in these
intracontinental rifts.

The original source of magmas in these rifts is in many cases the subject of debate;
some authors suggest derivation from mantle plumes, some from the sub‐continental
lithospheric mantle, and some from the convecting asthenosphere. Whilst the

contribution from mantle plumes typically remains uncertain, it is likely that most
continental rifts include magmas sourced both in the lithospheric mantle and in the
deeper asthenospheric mantle. However, it is notable that alkaline magmatism
develops repeatedly in certain areas, often where major lithospheric structures
provide pathways for magma to rise up into the crust. In these areas, the lithospheric
mantle keels are also likely to have been enriched in a range of elements through
repeated metasomatism associated with fluids rising from subducting slabs.

Figure 2: Schematic diagram illustrating development of lithospheric complexity. The
left‐hand side shows how subduction at the margins of a craton with thick lithospheric
keel will introduce magmas and fluids, carrying a range of elements, into the overlying
crust and lithospheric mantle. The right‐hand side shows how subsequent extension
of this complex lithosphere may produce alkaline magmatism derived from
asthenospheric and metasomatised lithospheric sources.

In Europe, the most important such areas include the margins of the North Atlantic
Craton in Greenland and the margins of the Fennoscandian Shield in Scandinavia. It is
these areas that contain the most significant currently identified European REE
deposits. However, repeated orogenic cycles since the Precambrian have developed
the same level of lithospheric complexity in southern and central Europe, providing a
basis for the Cenozoic alkaline magmatism in these areas. In these younger provinces,
the plutonic counterparts to the surface volcanics are rarely exposed.

To summarise, the principal REE metallogenetic belts in Europe are those in which
alkaline magmas have been emplaced, during extension, into zones of complex
continental lithosphere that offer both a range of magma sources and structural

pathways for magma to rise into the crust. REE are typically concentrated in plutonic
rocks and thus near‐surface resources are found in the more deeply eroded alkaline
provinces of northern Europe. However, it is very likely that comparable primary REE
resources remain to be discovered at depth within southern Europe’s younger rift
provinces.

Acknowledgements
This paper is published with the permission of the Executive Director, British Geological
Survey (NERC). BGS © NERC 2014. All rights reserved. Research presented in this paper
has been funded by the FP7 EURARE project.
References
1. A. R. Chakhmouradian and A. N. Zaitsev, "Rare Earth Mineralization in Igneous
Rocks: Sources and Processes", Elements, 8 347‐353 (2012)
2. A. R. Woolley and D. K. Bailey, "The crucial role of lithospheric structure in the generation and
release of carbonatites: geological evidence", Mineralogical Magazine, 76 259‐270 (2012)
3. R. E. Ernst and K. Bell, "Large igneous provinces (LIPs) and carbonatites", Mineralogy and Petrology,
98 55‐76 (2010)
4. J. Blichert‐Toft, N. T. Arndt and J. N. Ludden, "Precambrian alkaline magmatism", Lithos, 37 97‐111
(1996)
5. D. R. Zozulya, T. B. Bayanova and P. N. Serov, "Age and Isotopic Geochemical Characteristics of
Archaean Carbonatites and Alkaline Rocks of the Baltic Shield", Doklady Earth Sciences, 415A 874‐
879 (2007)
6. B. G. J. Upton, C. H. Emeleus, L. M. Heaman, K. M. Goodenough and A. A. Finch, "Magmatism of
the mid‐Proterozoic Gardar Province, South Greenland: chronology, petrogenesis and geological
setting", Lithos, 68 43‐65 (2003)
7. S. V. Bogdanova, B. Bingen, R. Gorbatschev, T. N. Kheraskova, V. I. Kozlov, V. N. Puchkov and Y. A.
Volozh, "The East European Craton (Baltica) before and during the assembly of Rodinia",
Precambrian Research, 160 23‐45 (2008)
8. K.‐I. Ahall and J. Connelly, "Intermittent 1.53–1.13 Ga magmatism in western Baltica; age
constraints and correlations within a postulated supercontinent", Precambrian Research, 92 1‐20
(1998)
9. K. Secher, L. M. Heaman, T. F. D. Nielsen, S. M. Jensen, F. Schjoth and R. A. Creaser, "Timing of
kimberlite, carbonatite, and ultramafic lamprophyre emplacement in the alkaline province located
64°–67° N in southern West Greenland", Lithos, 112 400‐406 (2009)
10. J. G. Meert, H. J. Walderhaug, T. H. Torsvik and B. W. H. Hendriks, "Age and paleomagnetic
signature of the Alnø carbonatite complex (NE Sweden): Additional controversy for the
Neoproterozoic paleoposition of Baltica", Precambrian Research, 154 159‐174 (2007)
11. H. Downes, E. Balaganskaya, A. Beard, R. Liferovich and D. Demaiffe, "Petrogenetic processes in
the ultramafic, alkaline and carbonatitic magmatism in the Kola Alkaline Province: A review",
Lithos, 85 48‐75 (2005)
12. M. J. Lee, J. I. Lee, D. Garcia, J. Moutte, C. T. Williams, F. Wall and Y. Kim, "Pyrochlore chemistry
from the Sokli phoscorite‐carbonatite complex, Finland: Implications for the genesis of phoscorite
and carbonatite association", Geochemical Journal, 40 1‐13 (2006)
13. R. Miranda, V. Valdares, P. Terrinha, J. Mata, M. Azevedo, M. Gaspar, J. C. Kullberg and C. Ribeiro,
"Age constraints on the Late Cretaceous alkaline magmatism on the West Iberian Margin",
Cretaceous Research, 30 575‐586 (2009)

THE PALAEOPROTEROZOIC SKARN‐HOSTED REE
MINERALISATIONS OF BASTNÄS‐TYPE: OVERVIEW AND
MINERALOGICAL – GEOLOGICAL CHARACTER
Erik JONSSON1,2*, Karin HÖGDAHL2, Fredrik SAHLSTRÖM2, Per NYSTEN1,
Martiya SADEGHI1
1
Geological Survey of Sweden, Department of Mineral Resources, Box 670, SE‐75128 Uppsala, Sweden;
2
CEMPEG, Department of Earth Sciences, Uppsala University, SE‐75236 Uppsala, Sweden
* erik.jonsson@sgu.se

Abstract
The Bastnäs‐type rare earth element (REE) deposits are located in the
Palaeoproterozoic Bergslagen ore province in the Swedish part of the Fennoscandian
Shield. These deposits comprise magnetite‐skarn‐hosted REE silicate mineralisations
that occur along a discontinuous belt, c. 100 km in length, hosted by metamorphosed
c. 1.90‐1.87 Ga Svecofennian volcanosedimentary rocks. The major REE‐minerals are
LREE‐enriched silicates, e.g. cerite‐(Ce), and carbonates, e.g. bastnäsite‐(Ce). Available
evidence suggest that the Bastnäs‐type deposits formed through replacement
reactions between c. 1.9 Ga carbonate units and hydrothermal, magmatic‐dominated
fluids, most likely related to volcanic‐subvolcanic activity coeval with the younger parts
of the host rock sequence. Despite their known relatively small sizes, several are studied
by exploration companies.

Introduction
The increasing global interest in rare earth elements (REE) and their deposits has
prompted both new exploration as well as renewed interest in known occurrences.
Overall, the Nordic part of the Fennoscandian Shield is considered one of the areas
with highest potential for exploration in Europe today, not least for iron, base and
noble metals. In addition, several projects for rare and critical metals including the REE
are presently in different stages of development. Advanced REE projects include the
syenite‐hosted Norra Kärr deposit in southern Sweden1. Yet, the most classic of REE‐
rich mineralisations in the Shield are the skarn‐hosted deposits of Bastnäs‐type, in the
Palaeoproterozoic Bergslagen ore province of south central Sweden (Fig. 1). The
element cerium was discovered in cerite‐(Ce) from the Bastnäs mines2, and a
succession of early studies led to the discovery of additional new elements and
minerals. Subsequently, a number of similar deposits were identified in the west
central part of the province, and brought together under the classification “Bastnäs‐
type deposits” 3. Active during the 1800s, the mines of the Bastnäs field were most

likely the first hard‐rock deposits ever mined explicitly for REE extraction. Still today,
the abundant mine dumps exhibit mineralised material, typically with significant REE
Table 1: Whole rock REE concentrations in mine dump samples, Bastnäs‐type deposits.
Locality Y La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu REEtot
ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm %
Bastnäs 151 29400 44400 4001 13600 1500 173 836 71 262 34 53 4 19 2 9
Johanna 3633 16305 29275 2851 10000 2102 158 1687 197 750 131 297 37 179 24 7
Malmkärra 7780 54300 136500 17500 67900 11600 855 6900 686 2450 305 493 42 172 18 31
Rödbergsgruvorna 1200 17100 27300 2800 10500 1300 78 845 63 210 27 42 4 15 2 6
Södra Hackspik 5080 47100 71900 7600 28300 3600 340 2100 196 812 114 220 19 86 9 17
Östanmossa 3970 28300 53700 6100 24600 3700 171 2300 201 709 97 171 14 61 7 12
Östra Gyttorp 1994 25874 63800 6652 10000 4659 180 2716 282 815 91 178 24 131 18 12
Concentrations of Ce, La, Nd, Pr >10000 ppm, Gd, Pr, Sm >1000 ppm were analysed by ICP‐AES . All other elements were analysed by ICP‐MS at ALS, Vancouver, Canada.
contents (Table 1.).
Over time, the Bastnäs‐type deposits have been debated both as to their formation
and relations to host rocks and other types of mineralisation. Their genesis, together
with that of the associated skarn iron ores and the extensive host‐rock alteration, was
originally interpreted to be related to large‐scale, so‐called magnesia metasomatic
processes, generated by granitoids emplaced at the waning stage of regional
metamorphism3,4,5. Presently hydrothermal scenarios are favoured, involving
reactions between magmatic‐hydrothermal fluids and pre‐existing limestone
layers6,7,8,9.
REE mineral names are mostly given with recommended Levinson modifiers10 (Table
2). Where detailed mineral chemical information is lacking, this is not applied.
Occurrences of what is referred to as “orthite” in older literature, and where sufficient
data is not available, are represented here by “allanite”, as “orthites” reported from
this area have been shown to represent an array of different species.
Table 2: REE‐minerals in the Bastnäs‐type deposits.

Mineral General formula
Allanite‐(Ce) Ca(Ce,LREE)Fe2+Al2[SiO4][Si2O7]O(OH)
Bastnäsite‐(Ce) (Ce,LREE)(CO3)F
Bastnäsite‐(La) (La,LREE)(CO3)F
Cerianite‐(Ce) (Ce,LREE)O2
Cerite‐(Ce) (Ce,LREE,Ca)9(Mg,Fe)(SiO4)6(SiO3OH)(OH)3
Dissakisite‐(Ce) Ca(Ce,LREE)MgAl2[SiO4][Si2O7]O(OH)
Dollaseite‐(Ce) Ca(Ce,LREE)Mg2Al[SiO4][Si2O7]F(OH)
Ferriallanite‐(Ce) Ca(Ce,LREE)Fe3+AlFe2+[SiO4][Si2O7]O(OH)
Fluocerite‐(Ce) (Ce,LREE)F3
Fluocerite‐(La) (La,LREE)F3
Fluorbritholite‐(Ce) Ca2(Ce,LREE)3[SiO4]3F
Gadolinite‐(Ce) (Ce,LREE,Y)2FeBe2Si2O10
Gadolinite‐(Y) (Y,REE,)2FeBe2Si2O10
Håleniusite‐(La) (La,LREE)OF
Lanthanite‐(Ce) (Ce,LREE)(CO3)3F . 8H2O
Monazite‐(Ce) (Ce,LREE)(PO4)
Parisite‐(Ce) Ca(Ce,LREE)2(CO3)3F2
Percleveite‐(Ce) (Ce,LREE)2Si2O7
Synchysite‐(Ce) Ca(Ce,LREE)(CO3)2F
Törnebohmite‐(Ce) (Ce,LREE)2Al[SiO4]2(OH)
Västmanlandite‐(Ce) (Ce,LREE)3Ca(Mg,Fe)2Al2Si5O19(OH)2(F,OH)

Geological setting
The REE‐line7 is located in the western part of the Palaeoproterozoic Bergslagen ore
province in the c. 1.9‐1.8 Ga Svecokarelian orogen (Fig. 1). Archaean and older
Palaeoproterozoic rocks outline the northeastern boundary of the orogen, and to the
south and southwest it is succeded by the 1.85‐1.65 Ga Transscandinavian Igneous
Belt.
During the Svecokarelian orogeny the Palaeoproterozoic rocks in the Bergslagen
province formed in a back‐arc setting inboard of an active continental margin11. The
oldest, 1.90‐1.87 Ga rocks, belong to a volcanosedimentary succession with coeval
intrusive rocks. Polyphase deformation and LP‐HT greenschist to amphibolite facies
metamorphic occurred between c. 1.85 and 1.80 Ga. A suite of younger intrusive rocks
was
Figure 1: Geological overview map of the

Bergslagen ore province, south central Sweden.
Inset map shows the Bergslagen province and its
position in the Fennoscandian shield (left).
Geological map showing the major depsosits of the
REE‐line, the linear occurences of Bastnäs‐type
deposits in west central Bergslagen (right)7,20.


emplaced after the main deformation stage and are dominated by granites and
associated pegmatites. The succession of metavolcanic rocks is estimated to be c. 8
km thick, and are mainly of rhyolitic to dacitic compositions, with subordinate
occurrences of mafic rocks. The felsic metavolcanic rocks, typically with intercalated
marbles and associated skarns, host most of the base metal and iron oxide
mineralisations in the province4,12, including the Bastnäs‐type deposits.
The REE‐line (Fig. 1) stretches about 100 kms in a NE‐SW direction with a moderately
to steeply dipping foliation, and is interpreted to be located in the core of an
antiform13. It hosts a number of groupings of REE‐rich iron oxide skarn deposits with
associated BIF occurrences, as well as similar mineralisations without elevated REE
contents. The mineralisations follow marble horisons interlayered with the
metavolcanic rocks. Near the skarn deposits these rocks were strongly hydrothermally
altered, and subsequently transformed during regional metamorphism into cordierite
and/or andalusite‐bearing mica schists or quartzites14,15. Meso‐scale, ductile folds
have been recorded along the REE‐line7 and the mineral assemblages in the deposits
are recrystallised, indicating that they were formed at an early stage of the orogenic
evolution. Mining of the iron oxide deposits in this area was active from the Middle
Table 3: Observed REE minerals in the bastnäs‐type deposits.
e
d it ?
lan ite
an nd te
äs tm owla ilicatesilica
v r s ‐
l it e
p
roue i te dite e e to e to REE‐ ‐REE
o te g n t u
ite ite hm la o og log C n u l ‐ W
it e e ite te th te ni ite sit ite e
it e ä s nit kis sei bri eri lla lin iu an zit te ve ys bo an pid al na wn ow
an stn rite ria sa lla or oc ria do len th na risi rcle ch rne stm E‐e ‐an ‐a ko kn
Locality All Ba Ce Ce Dis Do Flu Flu Fer Ga Hå Lan Mo Pa Pe Syn Tö Vä RE Fe Mg Un Un
Bastnäs x x x x x x x x x x x x x x x x
Danielsgruvan (Knutsbo area) x x x
Haggruvan x
Högfors x x x x x x x x
Johanna x x x x x
Lerklockan x x
Malmkärra x x x x x x x x x x x
Myrbacksfältet x x
Rödbergsgruvorna x x x x x
Södra Hackspik x x x x x x x
Storgruvan x x x
Stripåsen x x
Stålklockan x ?
Åsgruvan x x
Östanmossa x x x x x x x x x x
Östra Gyttorp x x x x
Based on recent analytical work and 14,3,9,1,7,16
Ages and into the 1980s.

The Bastnäs‐type deposits
The iron oxide skarn‐associated REE silicate mineralisations occur within a specific
sequence of the altered metavolcanic rocks, and they have been separated into
subgroup 1 and 2 based on their location and their mineralogy and geochemistry9.
Subtype 1 includes more iron‐rich silicates and mainly LREE enrichment and subtype
2, exhibiting more magnesium and fluorine‐rich silicates and enrichment in both LREE
and Y+HREE. Their complex and diverse REE mineralogy and the variability between
deposits is shown in Tables 2 and 3.

The Nora area
Two occurrences are known from the Nora area (Fig. 1). The predominant REE‐
assemblage at Rödbergsgruvorna consists of cerite‐(Ce), allanite‐Ce) – ferriallanite‐
(Ce), bastnäsite‐(Ce), and a västmanlandite‐like mineral, in clinoamphibole skarn
associated with magnetite and minor sulphides9,16. REE mineralisation at the Östra
Gyttorp mine occurs in the form of allanite‐(Ce)‐dominated lenses associated with
magnetite ore, hosted by felsic metavolcanic rocks6,16,17. During mining, approximately
100 tons of “allanite”‐dominated rare earth ore was produced and sold during the late
1800s, beside the iron ore12. Minor occurrences of “allanite” as well as other REE
minerals have been noted in the greater Nora area, in addition to the deposits of
Bastnäs‐type.

The Riddarhyttan area
The Riddarhyttan area (Fig. 1) includes the most well‐known deposit of this type, in
the form of the eponymous Bastnäs mines. However, REE mineralisation in this area
occurs also at several other locations within the same, partly altered, metavolcanic
unit14,15. The Nya Bastnäsfältet is composed of two parallel ore horisons, featuring
skarn‐hosted (polymetallic) ores with locally abundant REE minerals, and banded iron
formations, respectively (Fig. 2). The former are magnetite‐dominated, and occur as
pods in mainly clinoamphobole‐dominated skarn, wholly or partly replacing marble. In
the western part of the Bastnäs field unreacted portions of the marble bed are
present.
Rich REE mineralisation was encountered at Ceritgruvan (the cerite mine; Fig. 2). The
REE deposits here are of characteristic magnetite‐skarn type, but with locally
significant Cu‐Co‐Ni‐Bi‐Mo sulphides as well as minor Te, Hg‐bearing minerals18 and
gold‐silver alloys. At Bastnäs the most characteristic REE mineralisation type comprises
either only “allanite” (ferriallanite‐(Ce)), or an assemblage of cerite‐(Ce) and
ferriallanite‐(Ce), with variable amounts of bastnäsite‐(Ce) and törnebohmite‐(Ce). It
is stated that the overall REE ore production from the Cerite mine was c. 4500 tons, of
which cerite alone accounted for several hundreds of tons14,19. Later reworking of the
old dumps at Nya Bastnäs yielded significant additional amounts (B. Högrelius, pers.
comm.).
Storgruvan is another REE mineralisation comprising cerite associated with “allanite”
and törnebohmite where c. 5 tons of REE‐mineralised material were separated and
sold14. The Högfors field to the NE features REE mineralisation within a banded iron
formation3,7, with sparse bands dominated by cerite‐(Ce), a västmanlandite‐type
mineral, ferriallanite‐(Ce) and gadolinite‐(Ce) as well as additional minor minerals that
occur as fine‐grained, recrystallised assemblages in folded bands or layers7.
Outside of this major, ore‐bearing unit, REE mineralisation was also encountered at
Lerklockan, Stålklockan, Myrbacksfältet and others. Additional REE‐mineralised zones

at depth in the greater Bastnäs area were noted during exploratory drilling for gold
and base metals in the 1980s15.

Figure 2: Detailed geological map showing the relations in the Bastnäs ore field,
Riddarhyttan. Modified after Geijer14.

The Norberg area
The major occurrences of Bastnäs‐type REE mineralisation in the Norberg area (Fig. 1)
are the Östanmossa, Malmkärra, Johanna, Åsgruvan and Södra Hackspik mines, all of
them magnetite‐dominated, magnesium‐rich skarn iron ore deposits. The relative
abundance of fluorine in the Norberg area deposits (subtype 2) is manifested by the
abundance of humite‐group minerals, including norbergite. These occur intimately
with REE mineralisation, such as at Johanna, where norbergite is present as subhedral
crystals included in REE silicates. The presence of humite‐group minerals is also
indicated by the common occurrence of “ophicalcite”; a carbonate rock containing
pseudomorphs of humite‐minerals.
The Malmkärra mines are magnetite skarn deposits occurring within a narrow layer of
marble. Here, REE mineralisation in the form of fluorbritholite‐(Ce), västmanlandite‐
(Ce) as well as REE‐carbonates9,20 occurs in a skarn zone in that marble at the contact
towards cordierite schists. In the Johanna mine magnetite mineralisation is associated
with tremolitic skarn with “ophicalcite” pods21, and associated REE mineralisation in
the form of dollaseite‐(Ce) associated with sparse cerite‐(Ce), REE‐carbonates and
gadolinite‐group minerals. At Åsgruvan, another magnetite skarn deposit in marble,
carries sparse REE mineralisation, mainly as “allanite” 21. The mines at Östanmossa

were operated on a magnetite mineralisation similar to that at Åsgruvan21, but feature
more abundant REE mineralisation as dollaseite‐(Ce) and fluorbritholite‐(Ce),
characteristically associated with norbergite9,16,21. The Södra Hackspik mine is another
deposit in fluorine‐rich skarn, featuring locally REE‐rich assemblages, with cerite,
törnebohmite, dissakisite‐(Ce), dollaseite‐(Ce), bastnäsite‐(Ce and La), and
fluorbritholite‐(Ce)3,9,16. In addition, REE enrichment, mainly hosted by “allanite” and
REE‐enriched epidote has been noted in a number of other iron and base metal
mineralisations, such as Stripåsen.
The deposits forming the northeasternmost continuation of the REE‐line are not
represented in Fig. 1, as they are located significantly farther to the east, yet decidedly
forms an offset part of this structure. At Knutsbo, Danielsgruvan and Haggruvan have
been noted for their REE contents, with e.g. "allanite" and törnebohmite described
from a talc‐bearing amphibole skarn3.

Discussion and conclusions
The Bastnäs‐type deposits represent a style of mineralisation without directly similar
deposits occurring outside of the Bergslagen province. All available evidence suggest
that the Bastnäs‐type deposits formed as a consequence of submarine volcanic to
subvolcanic, magmatic‐dominated hydrothermal activity at around c. 1.9 Ga. The iron
ores and the associated skarn as well as REE mineralisation were generated through
reactions between pre‐existing marble beds within the volcanosedimentary sequence,
and hydrothermal fluids, most likely in a sub‐seafloor position. The differences in
mineralogy between the deposits may originate both from local fluid (and mineral)
evolution, the degree of fluid‐rock interaction9, as well as in variations in the local ore‐
forming systems, including volcanic‐hydrothermal facies.
Despite the modest known size of individual Bastnäs‐type deposits, their high grades
and extended distribution along the REE‐line suggest relevant exploration potential,
not least when considering the presence of other metals. A potential problem for the
beneficiation of these deposits is their mineralogical complexity. Yet, in most cases, a
few REE silicates (e.g., cerite‐(Ce), dollaseite‐(Ce), ferriallanite‐(Ce), fluorbritholite‐
(Ce)) tend to dominate, which may improve their chances of utilisation.

References
1. E. Jonsson, "The Norra Kärr REE-Zr project and the birthplace of the light REEs", in Excursion
guidebook SWE3, SWE6 & SWE7, Society for Geology Applied to Mineral Deposits (SGA), 36-51
(2013).
2. W. Hisinger and J. J. Berzelius, Cerium, en ny metall, funnen i Bastnäs Tungsten, från Riddarhyttan
i Westmanland (in Swedish), Henrik A. Nordström förlag, Stockholm 1-24 (1804).
3. P. Geijer, "The geological significance of the cerium mineral occurrences of the Bastnäs type in
central Sweden", Arkiv för Mineralogi och Geologi, 3 99-105 (1961).

4. P. Geijer and N. H. Magnusson, "De mellansvenska järnmalmernas geologi" (in Swedish), SGU
Ser. Ca, 35 1-654 (1944).
5. N. H. Magnusson, "The origin of the iron ores in central Sweden and the history of their alterations",
SGU Ser. C, 643 1-364 (1970).
6. F. Sahlström, E. Jonsson, K. Högdahl and C. Harris, "Magnetite oxygen isotope constraints on the
genesis of Bastnäs-type REE mineralisation in Bergslagen, Sweden", in 31st Nordic Geological
Winter Meeting, abstract volume, Edited by M. D. Johnson, 60 (2014).
7. E. Jonsson and K. Högdahl, "New evidence for the timing of formation of Bastnäs-type REE
mineralisation in Bergslagen, Sweden", in E. Jonsson et al. (eds.) in Proceedings of the 12th biennial
SGA meeting (SGA 2013), Edited by E. Jonsson, Mineral deposit research for a high-tech world,
Uppsala, Sweden, 1724-27 (2013).
8. U. B. Andersson, D. Holtstam and C. Broman, "Additional data on the age and origin of the Bastnäs-
type REE deposits, Sweden", in Proceedings of the 12th biennial SGA meeting (SGA 2013), Edited
by E. Jonsson, Mineral deposit research for a high-tech world, Uppsala, Sweden, 1639-42 (2013).
9. D. Holtstam, and U. B. Andersson, "The REE minerals of the Bastnäs-type deposits, south-central
Sweden", Canadian Mineralogist, 45 1073-1114 (2007).
10. P. Bayliss and A.A. Levinson, "A system of nomenclature for rare-earth mineral species: revision
and extension", American Mineralogist, 73 422-3 (1988).
11. R. L. Allen, I. Lundström, M. Ripa, A. Simeonov and H. Christofferson, "Facies analysis of a 1.9
Ga, continental margin, back-arc, felsic caldera province with diverse Zn-Pb-Ag-(Cu-Au) sulfide
and Fe oxide deposits, Bergslagen region, Sweden", Economic Geology, 91 979-1008 (1996).
12. F. R. Tegengren, "Sveriges ädlare malmer och bergverk" (in Swedish), SGU Ser. Ca, 17 1-406 +
plates (1924).
13. A. Strömberg, "Berggrundsgeologiska kartan 12 F Ludvika SO (bed rock map)", SGU Ser. Af, 128
(1980).
14. P. Geijer and H. Carlborg, "Riddarhytte malmfält i Skinnskattebergs socken, Västmanlands län"(in
Swedish), Kungliga Kommerskollegium, Beskrivningar över mineralfyndigheter, 1 1-343 (1923).
15. P. Ihre and S. Sädbom, "Riddarhytte malmfält. Prospekteringsarbeten 1986" (in Swedish), Sveriges
Geologiska AB, Unpublished prospecting report Division prospektering PRAP 86541 1-261 (1986).
16. F. Sahlström, "Stable isotope systematics of skarn-hosted REE-silicates-magnetite mineralisations
in central Bergslagen, Sweden" Examensarbete vid Institutionen för geovetenskaper 279 1-82
(2014).
17. G. Nordenström, "Fynd af allanit (cerin) vid Gyttorp i Nora Bergslag" (in Swedish), Geologiska
Föreningens i Stockholm Förhandlingar, 12 540-4 (1890).
18. D. Holtstam, and B. Ensterö, "Does the Bastnäs REE deposit in central Sweden belong to the Fe
oxide-Cu-U-Au-REE class of ores?", in 25th Nordic Geological Winter Meeting, Reykjavík,
Abstract volume, 83 (2002).
19. P. Geijer, "The cerium minerals of Bastnäs at Riddarhyttan", SGU Ser. C, 304 1-24 (1921).
20. U. B. Andersson (Editor), "The Bastnäs-type REE mineralisations in north-western Bergslagen,
Sweden", SGU Rapporter and Meddelanden, 119 1-34 (2004).
21. P. Geijer, "Norbergs berggrund och malmfyndigheter" (in Swedish), SGU Ser. Ca, 24 1-162 (1936).


REE behaviour in hydrothermal alteration zones, Sungun
porphyry‐Cu deposit, NW Iran
Batoul Taghipour1*, Mohammad Ali Mackizadeh2
*1- Department of Earth sciences, faculty of sciences, Shiraz University, Shiraz, Iran,
2- Department of Geology, Isfahan University, Isfahan, Iran
taghipour@shirazu.ac.ir, ma_mackizadeh@yahoo.com
Abstract
Sungun porphyry‐Cu deposit is located in northwest Iran and is a part of Urumieh‐
Dokhtar magmatic arc. Porphyry type mineralization is related to the Miocene
monzonite‐quartz monzonite and diorite‐granodiorite intrusive rocks. There are four
distinguished alteration zones in the Sungun deposit which nearly envelope the ore
body. This research is about the behaviour of REE in the potassic, phyllic and argillic
zones with regard to indicated geochemical patterns and economic potential.
Geochemical characteristics of trace elements show many similarities between
alteration zones. There is an obvious LREE enrichment and HREE depletion in the three
alteration zones. Eu anomaly in the potassic zone is weakly negative, and in the phyllic
is nearly positive, but in the argillic zone there is no Eu anomaly shown up. The lower
ratio of Ce/Ce* in the potassic and the respectively higher ratio in the phyllic and argillic
zones reveal prevailing reduced and oxidized conditions. Similar physicochemical
conditions such as pH, suitable complexes in fluids, and presence of secondary minerals
for REE uptake, have contributed to the evolution of the three alteration zones.

Introduction
Sungun‐Ahar metallogenic region is part of Alpian‐Himalayan orogenic belt. Urumieh‐
Dokhtar volcano‐plutonic belt is extended from NW to SE Iran having a length of 1600
Km and a width of 20 Km. This belt includes volcanic, pyroclastic and intrusive rocks
with calc‐alkaline type magmatic affinity. It hosts many Cu‐Mo porphyry deposits i.e.
Sungun.
Many studies have focused on mobility, fractionation and solubility of REEs in
hydrothermal systems (Fulignati et al. 1999). According to Rollinson (1993) during low
grade metamorphism, weathering and hydrothermal alteration REEs are relatively
immobile, except in the cases of high fluid/rock ratios (Michard, 1989). REE mobility
can be favoured by pH, fO2, P‐T conditions, and availability of SO4, PO4, F, CO3 and
OH‐ ligands (Rolland et al., 2003). In this work the REE behaviour in relation to different
hydrothermal alteration zones has been investigated. Also a better understanding of
the physico‐chemical conditions during hydrothermal alteration is considered.

Geological setting

Sungun porphyry copper deposit is located in the central Iranian magmatic arc. The
porphyry host rock is a quartz monzonite stock which has been intruded by NNW‐SSE
and NNE‐SSW orientated andesitic dikes. The Sungun porphyry is Oligo‐Miocene in
age, and was emplaced as a sub‐volcanic complex into upper Cretaceous carbonate
rocks (Calagari, 2004, Mehrpatou, 1993) (Fig. 1). The hydrothermal alteration zones
are distinguished to potassic, phyllic, argillic, and propylitic ones (Calagari, 2003,
Hezarkhani, 2006, Mehrpartou and Tarkian, 1999). A series of late Tertiary Quaternary
sub volcanic rocks including, latite, quartz latite‐andesite were intruded into
sedimentary rocks (Calagari, 2004).
Petrography

On the basis of petrographic studies, alteration is widely developed in the area. Quartz
and sericite are the predominant mineral phases. Potassic alteration zone is taken
place in the upper parts of the stock but makes the deepest alteration zone related to
the porphyry‐Cu mineralisation. Associations of K‐feldspar, primary biotite, secondary
biotite, sericite, plagioclase, quartz, anhydrite, sphene, muscovite, actinolite and
zircon are characteristic minerals of this zone, along with common features of
porphyry and granular textures are common (Fig. 2a, b).
Biotite and amphibole were replaced by muscovite as pseudomorph. Rare jarosite in
this zone is produced by meteoric oxidation of pyrite (Fig. 2c, d). Jarosite itself is partly
altered to goethite (Pirajno, 2009).
Phyllic alteration replaces the potassic zone, by K‐feldspar being altered to sericite
and clay minerals, and plagioclase altered to sericite. The mineral assemblage
includes: plagioclase, sericite, quartz, orthoclase, pyrite, zircon and apatite. The
intergrowth texture of mica minerals can be seen as lepidoblastic texture. In the argillic
alteration zone, metasomatism is very intense, and primary minerals are altered to

clay minerals. Quartz is the only relict mineral in this zone. Propylitic zone is
characterized by chlorite, epidote, calcite and albite.
Petrology
Geochemical data of intrusive and altered rocks are shown in the table 1. Using SiO2/
(Na2O+K2O) diagram (Middlemost, 1994), Sungun intrusive rocks are diorite,
granodiorite and quartz monzonite in composition.
On the basis of TAS diagram (Irvin and Baragar, 1971) the magmatic series shows sub‐
alkaline affinity, whereas when using AFM diagram (Irvin and Baragar, 1971), the
magmatic nature is calc‐alkaline. Regarding the Y/Nb discrimination diagram the
tectonic setting of the area (Pearce, et al. 1984) is determined to be dominated by
volcanic arc granitoids.
Geochemistry
Behaviour of major and trace element during the alteration processed were
determined by using enrichment‐depletion diagram. The mean SiO2 values are 58.82%,
63.91%, 70.71% and 67.38% in unaltered rock, and the potassic, phyllic, and argillic
zones respectively. The highest SiO2 content belong to phyllic and the lowest one to
potassic altered rocks. K2O and MgO contents tend decreasing from the potassic to
phyllic zones (Fig. 3).
Arsenic is an indicator for mineralization and has a magmatic origin. This element
could be associated with Cu and Mo. In the potassic zone Cu value ranges between
5160 to 1400 ppm. Also there is a positive correlation between Mo‐As and Cu‐ As (Fig.
4‐a, b). Molybdenum content in the potassic and phyllic zones is higher than in the
argillic one. Cu and Ag show negative correlation (Fig. 4‐c). There is positive correlation
between Cu/Ag ratios in all of alteration zones (Fig. 4‐d).

Trace and rare earth geochemistry

REE patterns of unaltered and hydrothermal altered rocks were normalized to the
chondrite value (Fig. 5), aiming at to assign the possible fractionation process, during
hydrothermal alteration: (La/Sm)n, (La/Yb)n and (Tb/Yb)n ratios of the altered rocks
were compared to unaltered ones (Fig. 6).
The enrichment of light REE (LREE) relative to heavy REE (HREE) in the alteration zones
was shown by (La/Yb) n ratio (Fig. 6). ∑REE in the potassic (∑REE= 219.9), phyllic
(∑REE= 186.41) and argillic (∑REE=200.5) are higher than in the unaltered rocks (∑REE=
132.89).
The ∑REE in the potassic (219.9) is more than in the phyllic zone (186.41). Decreasing
of REE from phyllic to potassic zones is attributed to increasing of water/rock ratio,
and decreasing pH (Pirajno, 2009).
The Eu/Eu* ratio of unaltered and altered rocks, is lower than 1, which may depend
on the weak Eu anomaly. The negative Eu anomaly is due to plagioclase alteration and
also the low fO2 and high CO2/H2O.
Eu can be concentrated with oxide and sulfates phases during hydrothermal process.
Due to the presence of sulfide phases, like as pyrite, in the phyllic zone, higher Eu/Eu*
ratios would be expected. The Eu/Eu* value of 0,89, in the argillic zone is very similar
to that of unaltered rocks (0.94), implying that Eu is immobile during the alteration
process . Using (La/Sm) n for LREE/HREE fractionation, (La/Sm)n ratio in the argillic,
potassic and phyllic zones are 8.44, 8.21, and 7.29 respectively. The potassic, phyllic
and unaltered rock samples show (Tb/Yb)n 2.45, 3.03 and 1.34 respectively, whereas
this ratio is 8.28 in the argillic zone. The (La/Yb)n is 69.95 in the potassic zone and 62.45
in the phyllic one which are higher than corresponding value in unaltered rock samples
(20.96) (Fig. 6D). It is there by concluded that there is an enrichment of LREE compared
to HREE in the potassic and phyllic zones. The high enrichment of LREE relative to HREE
in the argillic zone is probably due to clay minerals present. The Ce/Ce* value of 0.96
in the phyllic and 96.5 in argillic zones, reflect oxidized conditions in the hydrothermal
systems (Rollinson, 1993) (Fig. 6E). Slightly decreased values of Ce/Ce* (0.95) in the
potassic zone implies to prevailed reduced conditions. Potassic alteration zone is
characterized by alkaline pH and reduced conditions.

Regarding to Wandlandt and Harrison (1979) mobility of REE and fractionation of

LREE/HREE are related to sulfate complexes and increasing water/rock ratio. It seems
that sulfate ions can be the most important complexing agents in the fluids (Wood,
1990, Lewis et al. 1998). The enrichment of LREE and depletion of HREE is the main
characteristic of potassic zone (Pirajno, 2009). The solubility of REE in the
hydrothermal systems is related to the fluid/rock ratio and availability of carbonate
complex.
The depletion of HREE with respect to LREE is present in the Sungun samples. This is
indicated by F‐ ion, which preferentially removes HREE during alteration (Lewis et al.
1997, Fulignati et al., 1999). Also the enrichment of LREE in the phyllic and potassic
zones is due to SO42‐ agents which preferentially make complexes taking up LREE
(Lewis et al. 1997). LREE complexes are more unstable than HREE, so these are
entering faster into secondary minerals.

Conclusion
The Sungun porphyry copper deposit is a part of Alpian‐Himalayan metallogenic belt.
Porphyry mineralization is related to the Miocene monzonite‐quartz monzonite and
diorite‐granodiorite intrusions.
The deposit is characterized by four alteration zones: potassic, phyllic, argillic and
propylitic. REE patterns in the alteration zones are similar to fresh rocks, with all
samples showing enrichment of LREE and depletion of HREE. The present of OH‐, F‐
complexes caused depletion of HREE in the alteration zones. The F‐ ligand is related to
the mica minerals like biotite. ∑REE in the phyllic zone is higher than potassic one,
which is attributed to the H+ agent and high fluid /rock ratio.
The negative Eu anomaly in the potassic zone is due to reducing fluid and alteration of
plagioclase. Negative Ce/Ce* anomaly also confirms the reducing conditions.
The ratios of Eu/Eu* and Ce/Ce* in the phyllic zone is higher than potassic one, due to
high oxidized conditions. The highest LREE/HREE fractionation is shown in the argillic
zone which relates to the high H+ activity and the presence of secondary minerals for
the uptake of these elements.

References:

14. Calagari, A. A., “Stable isotope (S, O, H and C) studies of the phyllic and potassic–phyllic alteration
zones of the porphyry copper deposit at Sungun, East Azarbaidjan, Iran”. Journal of Asian Earth
Sciences, 21,767–780 (2003).
15. Calagari, A. A., “Fluid inclusion studies in quartz veinlets in the porphyry copper deposit at Sungun,
East‐Azarbaidjan”, Iran. Journal of Asian Earth Sciences, 23, 179–189 (2004) .
16. Fulignati, P., A. Gioncada, A. Sbrana., “Rare‐earth element (REE) behaviour in the alteration
facies of the active magmatic–hydrothermal system of Vulcano (Aeolian Islands, Italy)” Journal
of Volcanology Geothermal Resource, 88, 325–342 ( 1999).
17. Hezarkhani, A., “Petrology of the intrusive rocks within the Sungun Porphyry Copper Deposit,
Azerbaijan, Iran”. Journal of Asian Earth Sciences 27, 326–340 (2006).
18. Irvine, T. N. and Baragar, W. R. A., “A guide to the chemical classification of the common volcanic
rocks”. Canadian Journal of Earth Sciences 8, 523‐548 (1971).
19. Lewis, A. J., Palmer, M. R., Sturchio, N. C. and Kemp, A. J., “The rare earth element geochemistry
of acid–sulphate and acid– sulphate–chloride geothermal systems from Yellowstone National
Park, Wyoming, USA”. Geochimica et Cosmochimca Acta 61, 695–706 (1997).
20. Mehrpartou, M., “Contributions to the geology, geochemistry, ore genesis, and fluid inclusion
investigations on Sungun Cu‐Mo porphyry deposit, NW of Iran”. PhD dissertation, University of
Hamburg, Germany (1993).
21. Mehrpartou, M. and Tarkian, M., “Contribution to the fluid inclusion studies, Sungun Cu‐Mo
porphyry (West of Ahr‐Azrbayzan)”. Journal of Geoscience 10: 2‐27(in Persian) (1993).
22. Michard, A., “Rare earth element systematic in hydrothermal fluid”. Geochimca et Cosmochimca
Acta 53, 745‐750 (1989).
23. Middlemost, E. A. K., “Naming materials in the magma, igneous system”. Earth‐Science Reviews
37: 215–224 (1994).
24. Pearce, J. A., Harris, N. W. and Tindle, A. G., Trace element discrimination diagrams for the tectonic
interpretation of granitic rocks. Journal of Petrology 25, 956–983 (1984).
25. Pirajno, F., “Hydrothermal mineral deposits, principle and fundamental concept for the
exploration geologist”, Springer (2009).
26. Rolland, Y., Cox, S., Boullier, A. M., Pennacchioni, G. and Mancktelow, N., “Rare earth and tracen
element mobility in mid‐crustal shear zones: insights from the Mont Blanc Massif (Western Alps)”.
Earth Planetary Science Letter, 214: 203‐219 (2003).
27. Rollinson, H. R., “Using geochemical data: evolution, presentation, interpretation”. London, UK,
(1993).
28. Wendlandt, R. F. and Harrison, W. J., “Rare earth partitioning between immiscible carbonate and
silicate liquids and CO2 vapour: results and implications for the formation of light rare‐earth‐
enriched rocks”, Contribution to Mineralogy and Petrology, 69, 409‐419 (1979).
29. Wood, S. A., “The aqueous geochemistry of the rare‐earth elements and Yttrium. 2. Theoretical
predictions of speciation in hydrothermal solutions to 350 C at saturation water vapor pressure”.
Chemical Geology, 88: 99‐125 (1990).


Figure 1: Geological map of the Sungun porphyry‐Cu deposit, (Hezarkhani, 2006), and
location in the Sahand‐Bazman belt.

Figure 2: a) Microgranular texture in the potassic zone, b) Pesudomorph and relict
textures in the potassic zone, c) association of jarosite and muscovite in the potassic
zone, d) Late alteration of jarosite to goethite.

Figure 3: Enrichment‐depletion diagrams of major elements in the a) potassic, b)
phyllic and c) argillic zones.
Figure 4: Variation binary diagrams: a) As versus Mo, b) As versus Cu, c) Ag versus Cu
and d) Cu/As versus Cu/Ag in the potassic, phyllic and argillic zones.

Figure 5: REE pattern diagrams for potassic, phyllic, argillic altered zones in
comparison with corresponding paterns of unaltered host rocks.

Figure 6: Modified variation binary diagrams of a) Eu/Eu*, b) La/Sm, c) Tb/Yb, d) La/Yb
and e) Ce/Ce*, which are indicated of REE differentiation during hydrothermal
alteration process from intrusive rock to altered rocks.


RARE EARTH ELEMENT SIGNATURES IN MANGANESE

ORES ASSOCIATED WITH TETHYAN OPHIOLITES, IRAN;
INDICATORS OF PRIMARY ORE FORMING PROCESSES
Alireza Zarasvandi¹, Mohsen Rezaei1, Martiya Sadeghi2
1
Department of Geology, Shahid Chamran University (SCU), Ahavz, Iran
2
Department of Mineral Resources, Geological Survey of Sweden (SGU), Uppsala, Sweden
Zarasvandi_a@scu.ac.ir, mn‐rezaei@phdstu.scu.ac.ir, martiya.sadeghi@sgu.se
Abstract
The ophiolitic rocks along the Zagros orogenic and metallogenic belt, a segment in the
western part of the Tethyan collisional zone, are characterized by the widespread
occurrence of manganese and ferromanganese deposits. These deposits are closely
associated with radiolarian cherts and basaltic rocks, which cap the ophiolite
sequences. The present work provides a set of Rare Earth Element (REE) data for the
manganese‐bearing layers and nodules of three of these deposits: the Nasirabad and
Abadeh Tashk manganese deposits (associated with the Neyriz ophiolite) and the
Gugher manganese and ferromanganese deposit, associated with the Baft ophiolitic
mélange. The data are used to elucidate the primary processes that control the
deposition of manganese ores. Low REE contents, positive Eu anomalies and Lan/Ndn
ratios (˃3) indicate a proximal hydrothermal source for the Gugher and Abadeh Tashk
manganese and ferromanganese deposits. In comparison a relatively low REE
abundance, negative Eu anomalies and position on a Lan/Cen versus Al2O3/(Al2O3 +
Fe2O3) discrimination plot indicate a distal hydrothermal source for the Nasirabad ore‐
bearing layers and nodules. Moreover a weak positive Ce anomaly and elevated Ce/La
values indicate the contribution of terrigenous materials in the formation of Nasirabad
ore‐beraring layers. Finally, the comparison of the Gugher, Abadeh Tashk and
Nasirabad manganese and ferromanganese deposits with other manganese deposits
elsewhere in the world indicate that the REE patterns of Zagros manganese deposits
are analogous to those typical of hydrothermal deposits.

Introduction
More than 45 manganese deposits of economic importance have been reported in Iran
[1]
. These deposits have diverse origins and geological settings, and range in age from
Precambrian/Early Cambrian to Late Miocene‐Pliocene [1, 2, 3]. Many of these
manganese deposits lie within the Zagros orogenic and metallogenic belt, which is
located in the western part of the Tethyan collisional zone. This belt contains excellent

exposures of ophiolite sequences, including the Naien, Shahr Babak, Baft, Neyriz and
Kermanshah ophiolites (Fig. 1). The manganese deposits are commonly associated
with radiolarian cherts and basaltic rocks, capping the ophiolite sequences; examples
of such deposits include Abadeh Tashk, Nasirabad (associated with the Neyriz
ophiolite), Kamyaran, Sorkhvand (associated with the Kermanshah ophiolites), Gugher
and Gushk (associated with the Baft ophiolites). The present paper provides a set of
REE data for the manganese‐bearing layers and nodules of the Nasirabad, Abadeh
Tashk and Gugher manganese and ferromanganese deposits. The data are used to
provide a general overview of the primary processes that control the deposition of
manganese ores.

Figure 1: Map showing locations and distribution of different ophiolite complexes in
Zagros orogeny and the location of the Nasirabad, Abadeh Tashk and Gugher
manganese and ferromanganese deposits. Modified after Ghasemi and Talbot [4],
Abbreviations; Kh: Khoy, Kr: Kermanshah, Ny: Neyriz, Ba: Baft, Sh: Shahre‐Babak, Na:
Nain.

Geological setting
The Zagros orogenic and metallogenic belt can be divided into three main tectonic
trends, namely, the Zagros Fold and Thrust Belt (ZFTB), the Sanandaj‐Sirjan
Metamorphic Zone (can be subdivided into the north and south SSZ), and the Urumieh‐
Dokhtar magmatic arc (UDMA)[5]. The ophiolitic rocks along the Zagros orogenic belt
are classified into two groups of different ages: the Nain‐Shahr Babak‐Baft ophiolite
belt and the Neyriz‐Kermanshah ophiolite belt [4]. The former belt is a remnant of the
Naien‐Baft palaeo‐ocean that existed along the northern margin of the South SSZ

before the end of the Cretaceous, and the Neyriz‐Kermanshah ophiolite belt is a
remnant of the Neo‐Tethys ocean that was closed during the Miocene (Table 1). The
locations of the Nasirabad, Abadeh Tashk and Gugher manganese and
ferromanganese deposits are shown in Fig. 1.

Table 1: The age (formation) and main characteristics of Zagros ophiolitic complexes
Ophiolitic Rocks Formation Age Characteristics
(Ma)
Kermanshah 95–98 (Ar/Ar) Highly dismembered harzburgite, IAT, OIB
Neyriz 93–95 Ma (Ar/Ar) Harzburgite, E‐MORB (?), IAT
96–98 Ma (Ar/Ar)
Nain 99–101 (Ar/Ar) Harzburgite‐lherzolite, small ocean basin
around Lut Block
Shahr‐e‐Babak 93 Ma (Ar/Ar) Harzburgite‐lherzolite, small ocean basin
around Lut Block, IAT
from Ghazi [6].

The Nasirabad manganese occurrence is located to the south of the Neyriz ophiolite.
In this region the extremely folded and fractured ore‐bearing layers and manganese
nodules occur as interlayers with reddish radiolarites in the upper parts of the
Pichakun zone. This zone represents the Late Triassic to Middle Cretaceous abyssal
facies of the Neo‐Tethys ocean basin. It is thought that the Neyriz ophiolite has been
thrust over the Pichakun zone [7]. The lower parts of this zone are composed of Upper
Triassic megalodon limestone turbidites, dark marl, and serpentinite diapirs. These
grade upwards into thinly‐bedded (<5 cm) cherty radiolarites, alternating with
medium‐to thick‐bedded (up to 5 m) detrital limestone and green siliceous shale.
Above this lies a 500 m‐thick sequence of radiolarites. Radiolarites of the upper parts,
which host the manganese mineralization, are younger than Middle Jurassic in age [8].
Finally, the anhydritic limestone of the Tarbur Formation (Campanian‐Maastrichtian)
unconformably overlies the radiolarites.

The Abadeh Tashk manganese deposits are located to the NE of Tashk lake between
29o 47’‐55″ N and 53o 45́‐56″. In this region, the exposure of the Neyriz ophiolite
sequence is not complete. The ophiolite forms a discontinuous linear outcrop belt
about 16 km long and 11 km wide, between the high Zagros and the main Zagros thrust
fault (Fig. 1). Mn‐mineraliziation in Abadeh Tashk is associated with radiolarite cherts,
belonging to the Late Triassic to Early Cretaceous [9]. The chert sequences here include
both proximal bedded facies, and distal sedimentary facies. However, the bulk of Mn‐
mineralizations are spatially related to reddish brown radiolarian cherts (part of the
proximal bedded facies) that overlie the serpentinized ultramafic rocks of the ophiolite
[10]
. These radiolarites are overlain unconformably by massive limestones [9]. The

Abadeh Tashk manganese deposits represent three main textural features, namely,
syngenetic, diagenetic and epigenetic [10]. The syngenetic textures typically are consist
of thinly banded (2‐3 cm) and disseminated Mn‐minerals (mainly braunite and
psilomelane) within the radiolarite cherts interlayers, which cap the pillow lava basalts
and serpentinized ultramafic rocks. In diagenetic structures, the re‐mobilization of ore
minerals during deformation events have been resulted in accumulation of braunite,
pyrolusite and bixbyite in fold hinges. Moreover, Mn‐mineralization as lenses and
boudin are common features of diagenetic deposits [10]. The epigenetic features
represent the post depositional supergene enrichment processes, within which
pyrolusite, ranceite and aurorite were mainly deposited as open‐space filling textures
in the upper parts of chert sequences [10].

Comparison of the Neyriz ophiolite with the Baft ophiolite sequences indicates that
the latter is strongly imbricated and sheared by closely spaced sub‐vertical faults and
shear zones. It is referred to as a colored mélange complex [11]. The Baft ophiolitic
mélange consists of the following rock units from bottom to top: (1) ultramafic – mafic
complex (serpentinized harzburgite, gabbro, massive diabase, diabase dyke, spilitized
pillow lava and basic flow breccia with fragments of spilites), (2) sedimentary units
(Cretaceous limestone, sandstone and conglomerates) and (3) metamorphic units that
are slabs from the SSZ, tectonically overturned on top of the ophiolitic mélange. These
units are mainly composed of calcschist, chlorite‐sericite schist (amphibole schist),
biotite schist, marble and crystalline limestone of Paleozoic age [4, 12]. The Gugher
manganese and ferromanganese ores are located in the northwest of the Baft
ophiolite mélange. In these regions pyrolusite, hematite, goethite, pyrochroite,
tephroite and Fe‐Mn silicate occur as lenticular, layered and sometimes cross cutting
veins. These ores are commonly hosted by basalt and iron chert or massive limestone
[12]
. The indicative textures of sub‐marine hydrothermal processes such as cockade
(amorphous silica enclosed by Fe‐Mn oxyhydroxide shale) and colloform (pyrolusite
nucleous surrounded by microcrystalline quartz) have been recognized, and are
described in Heshmatbehzadi and Shahabpour [12].

Methodology
From the literature we compiled the data of three manganese deposits, consist of
Nasirabad, Abadeh Tashk and Gugher. For the Nasirabad, samples were collected from
the manganese‐bearing layers, nodules and radiolarian cherts, hosting manganese
mineralization. For petrochemical analysis, 5 relatively unaltered and fractured
samples from manganese‐bearing layers and 3 samples from nodules were chosen.
These were crushed using an iron pestle and were subsequently pulverized using a
tungsten carbide swing mill. Concentrations of rare earth elements (REEs) were
obtained by lithium metaborate fusion with nitric digestion followed by inductively
coupled plasma‐mass spectrometry (ICP‐MS) and ICP‐emission spectrometry (ICP‐ES)

at ACME Analytical Laboratories, Vancouver, Canada. The concentrations of REEs for
the Nasirabad samples are given in Table 2. Samples from the Gugher were processed
in Alschemex Co. [12], whereas samples from the Abadeh Tashk were processed by
AMDEL Analytical Laboratories [10] (Table 2).
Geochemistry

The assessment of rare earth element (REE) geochemistry in ores offers insights into
their genesis [13], with Mn ores being no exception, as has been described for
numerous occurrences of modern and ancient ferromanganese deposits in the deep
sea and terrestrial environments [14]. The REE data for the Nasirabad ore‐bearing layers
(n: 5) and nodules (n: 3) are listed in Table 1, and compared with published data for
the Gugher manganese and ferromanganese deposit (n: 6), and Abadeh Tashk deposit
(n: 3). All selected ore samples are characterized by relatively low REE abundance. The
chondrite‐normalized relative REE abundance patterns indicate a decrease of average
ƩREE from Gugher ores (ƩREE= 38.79 ppm), to Abadeh Tashk samples (ƩREE= 33.09
ppm), Nasirabad ore‐bearing layers (ƩREE= 19.7 ppm) and Nasirabad nodules (ƩREE=
8.94 ppm). In general, REE patterns exhibit distinct LREE enrichment relative to HREE
and the LREE/HREE value decreases on average from Nasirabad ore‐bearing layers
(9.27), to the Abadeh Tashk (9.12), Nasirabad nodules (7.04) and Gugher ore samples
(4.17), respectively. According to Ruhlin and Owen [15], the pronounced enrichment of
the LREE relative to HREE can be attributed to the greater stability of HREE complexes
in hydrothermal solutions.

In this study Cerium (Ce/Ce*) and Europium anomalies (Eu/Eu*) were calculated from:
Ce/Ce*=Cen/(Lan×Prn)1/2, Eu/Eu*=Eun/Smn×Gdn)1/2 [16], using chondrite‐normalized
abundances [17]. Taken as a whole, with the exception of the Nasirabad ore‐bearing
layers, ore samples from the selected areas have negative Ce anomalies (Fig. 2).
Moreover, some have weak negative Eu anomalies (Eu/Eu*= 0.67 and 0.8 for the
Nasirabad ore layers and nodules, respectively) whilst others have positive Eu
anomalies (Gugher; average, Eu/Eu*= 1.20 and Abadeh Tashk (average, Eu/Eu*=2.79).
It is important to note that the low REE concentration, positive Eu anomalies and weak
or absent Ce anomalies are considered as common geochemical features of
hydrothermal manganese and ferromanganese deposits [18, 19].

In hydrothermal manganese and ferromanganese crusts Lan/Ndn ratio is generally
between 3.0–7.4 [20]. These ratios in the selected areas are analogous to those typical
of hydrothermal crusts and decrease on average from the Nasirabad nodules (5.49) to
Abadeh Tashk (4.14), Gugher (3.9) and the Nasirabad ore‐bearing layers (3.62),
respectively. The correlation diagrams indicate that there are strong positive
correlations between the LREE and HREE, suggesting that the same mechanism was

Table 2: The concentrations of REEs (ppm) and some major oxides (Wt. %) in the selected manganese deposits.
Gugher* Abadeh Nasirabad Nasirabad
Tashk** nodules layers
Sample G5 GN1 BL3 KU4 BU6 K4 1 2 6 N1 N2 N3 LZ‐5 LZ‐2 La4 LA‐10 LA‐1
La 19.3 8.8 7 21.2 5 13.9 ˂10 ˂10 ˂10 2.4 2.5 4.4 6.9 5.8 3 3 5.1
Ce 20.5 9.6 2.7 1.5 0.5 0.7 7.6 8.9 15.4 2 2.4 4.1 11.6 8.5 8.2 5.9 13.2
Pr 4.2 4.2 4.2 4.2 4.2 4.2 0.6 1.22 0.98 0.3 0.33 0.55 1.1 1.17 0.52 0.47 0.98
Nd 11 5.6 6.4 10.3 1.3 7.7 3.2 5.96 6.17 0.7 1 1.7 3.6 3.7 1.2 1.5 3.7
Sm 2.07 1.14 1.24 2.21 0.2 1.31 1.12 2.42 1.25 0.15 0.14 0.57 0.81 0.8 0.53 0.28 0.78
Eu 0.78 0.48 0.4 0.72 0.12 0.64 1.42 2.77 0.22 0.04 0.06 0.11 0.19 0.17 0.07 0.07 0.19
Gd 2.61 1.55 1.42 1.9 0.33 1.7 1.69 1.35 1.19 0.21 0.23 0.39 0.8 0.84 0.43 0.25 0.71
Tb 0.88 0.88 0.88 0.88 0.88 0.88 0.11 0.2 0.17 0.03 0.04 0.07 0.12 0.14 0.08 0.05 0.12
Dy 2.39 1.47 1.28 1.65 0.33 1.46 0.45 0.89 0.65 0.15 0.22 0.45 0.82 0.86 0.27 0.2 0.78
Ho 1.1 1.1 1.1 1.1 1.1 1.1 0.13 0.28 0.17 0.05 0.05 0.08 0.16 0.17 0.12 0.04 0.12
Er 1.64 1.1 0.74 1.21 0.21 0.86 0.27 0.51 0.28 0.14 0.15 0.26 0.53 0.39 0.27 0.15 0.38
Tm 0.42 0.42 0.42 0.42 0.42 0.42 0.06 0.13 0.05 0.03 0.03 0.05 0.09 0.06 0.06 0.01 0.06
Yb 1.29 0.94 0.55 0.72 0.2 0.65 0.32 0.6 0.23 0.13 0.15 0.34 0.51 0.39 0.32 0.05 0.45
Lu 0.19 0.13 0.07 0.12 0.03 0.09 0.17 0.12 0.04 0.03 0.03 0.06 0.11 0.06 0.08 0.02 0.07
Ce/La 1.06 1.09 0.38 0.07 0.1 0.05 0.76 0.89 1.54 0.83 0.96 0.93 1.68 1.46 2.73 1.96 2.58
Lan/Ndn 3.39 3.04 2.11 3.98 7.45 3.49 6.05 3.25 3.13 6.64 4.84 5.01 3.71 3.03 4.84 3.87 2.67
Eu/Eu* 1.02 1.1 0.92 1.07 1.48 1.31 3.15 4.68 0.6 0.68 1.02 0.71 0.72 0.63 0.44 0.8 0.78
Ceanom ‐0.29 ‐0.28 ‐0.68 ‐1.39 ‐1.24 ‐1.54 ‐0.41 ‐0.35 ‐0.16 ‐0.38 ‐0.33 ‐0.34 ‐0.13 ‐0.18 0.01 ‐0.08 0.00
MnO2 45.4 35.8 33.9 22.2 19.9 17.8 ‐ 52.69 ‐ 6.07 4.57 3.7 17.8 10.22 3.22 3.16 4.5
Fe2O3 12.8 1.04 8.49 17.7 2.44 16.1 ‐ 1.06 ‐ 0.33 0.33 0.39 0.99 0.95 0.65 0.58 0.11
Al2O3 1.49 1.68 0.23 1.45 0.08 0.19 ‐ 0.83 ‐ 0.67 0.48 0.47 1.52 1.74 1.2 0.91 1.8
*[12]
, **[10].

responsible for REE uptake during the ore formation in in Nasirabad, Gugher and Abadeh
Tashk manganese and ferromanganese deposits[20] (Fig. 3). Positive correlations were also
observed between the ƩREE (ppm) and Fe2O3+MnO (wt. %) of selected samples from the
Nasirabad and Gugher deposits, indicating progressive REE entrance into the Mn and Fe
oxyhydroxides during the oxidation process (Fig. 3).

Figure 2: Chondrite‐normalized REE patterns of ore samples from the selected areas, with
chondrite values from [17].

According to Wright and Hosler [21], the Ce anom [Ceanom.= Log (3× Cen /(2×Lan+Cen))] can be
used as a good index for the determination of oxic or anoxic nature of the water body of
sedimentation. Ceanom. > ‐0.1 indicates anoxic condition during sedimentation, whereas
Ceanom. < ‐0.1 represents oxidative conditions during deposition. The Ce anom. values for the
Gugher, Abadeh Tashk, and Nasirabad ore bearing layers and nodules are listed in Table 2.
With the exception of the Nasirabad ore bearing layers (samples La4, La10 and La 1), all
studied manganese deposits are characterized by Ceanom. < ‐0.1, reflecting oxidative
conditions during the deposition of manganese ores.

Maynard [22] has indicated that Al is insoluble in seawater at surface temperatures and can be
used as a good indicator for the entrance of detrital materials into depositional basins. There
is significant positive correlation between the ƩREE and Al2O3 contents of the Nasirabad ore‐
bearing layers, representing the effect of terrigenous materials during the formation of the
Nasirabad manganese‐bearing layers. Another indication of this may be seen in the Ce/La
values. The Ce/La ratios increase from the Gugher deposit (average 0.46), to the Nasirabad
nodules (average 0.90), Abadeh Tashk (average 1.06) and the Nasirabad ore‐bearing layers
(average 2.08). According to Dubinin and Volcov [23], increasing contribution from
carbonaceous biogenic and terrigenous materials can elevate the Ce/La values. The
geodynamic setting of Neyriz ophiolite has been discussed in detail by Shahabpor [24] and
Ghasemi and Talbot [4]. The new geochemical data indicate the Neyriz was an intra‐oceanic

island arc during the Late Triassic to Late Cretaceous time. Rezaei [1] and Zarasvandi [25]
pointed out that there are several lines of evidence, including Pb

Figure 3: Examples of good linear correlation between LREE, HREE, MnO and (Fe2O3+MnO)
contents.
isotopes and major and trace inter‐element relations, to show the contribution of mafic
terrigenous matrials (derived from the eruptions and/or erosion of Neyriz island arc) in the
formation of Nasirabad ore sequences. Combining the REE results with Al2O3/(Al2O3 + Fe2O3)
ratios is useful to delineate the depositional environment of manganese ores [19]. In a Lan/Cen
versus Al2O3/(Al2O3 + Fe2O3) discrimination plot [26], the average values of Nasirabad nodules
and ore‐bearing layers mainly plot in the pelagic field, However, the average values of Gugher
in the Baft area plot in the field of spreading ridge proximal deposits (Fig. 5).

Figure 4: Strong linear correlation between the Al2O3 (wt. %) and ƩREE in the Nasirabad ore‐
bearing layers.

Figure 5: Plot of Lan/Cen versus Al2O3/(Al2O3 + Fe2O3) for the Gugher in the Baft area, and the
Nasirabad ore‐bearing layers and nodules, modified after Murray [26]. Normalizing values are
from Evensen [17].
Discussion and conclusion
The ophiolitic rocks along the Zagros orogenic belt, a part of the Tethyan realm can be
classified into two categories, the Naien‐Shahr Babak‐Baft and Neyriz‐Kermanshah ophiolite
belts [4]. These ophiolitic belts are characterized by widespread occurrence of manganese and
ferromanganese deposits. It is important to note that, ophiolitic rocks hosting Mn and Fe
bearing deposits are abundant along the Tethyan realm for example; Antalya complex,
Turkey; northern Apennine ophiolitic complex, Italy; Troodos Massif, Cyprus; Semail nappe,
Oman; Wazirestan comlpex, Pakistan. It is now generally understood that these deposits were
commonly formed by hydrothermal processes similar to those of the present‐day mid‐
oceanic spreading centers [1, 20, 27, 28].
According to Jach and Dudek [18], marine manganese deposits can be classified into three
genetic types: Hydrothermal, Hydrogenous and diagenetic. However, these deposits can form
by a combination of the aforementioned processes. Moreover, other parameters such as
microbial processes and selective enrichment of bioessential elements and contribution of
detrital materials in formation of deposits can give some special characteristics to manganese
and ferromanganese deposits [1, 29]. The Fe‐Mn deposits from Gugher, Nasirabad and Abadeh
Tashk (associated with the Baft and Neyriz ophiolitic rocks, respectively) are defined by low
contents of REE, negative Ce anomalies (the Nasirabad ore bearing layers being the
exception), LREE enrichment relative to HREE and Lan/Ndn values that resemble the

composition of hydrothermal manganese deposits formed by low temperature hydrothermal
fluids [20, 30]. It has long been recognized that sediments precipitated from hydrothermal vent
fluids yield pronounced positive Eu anomalies, with little or no negative Ce anomaly. The
Gugher and Abadeh Tashk samples have positive Eu anomalies (Eu/Eu*=1.2 and 2.79,
respectively) as is commonly found in hydrothermal settings. In contrast, the Nasirabad
samples show negative Eu anomalies (Eu/Eu*= 0.8 and 0.67 for the manganese nodules and
layers, respectively). Such negative anomalies may result from high dilution of hydrothermal
solutions with oceanic water due to the long distance between hydrothermal discharge and
depositional environment in the Nasirabad area [31]. Moreover the lack of volcanics in the
Nasirabad rock sequences provides further evidence for a remote source for the
hydrothermal exhalations (distal hydrothermal source) [25, 32]. Another indication of this
conclusion may be seen in the Lan/Cen versus Al2O3/(Al2O3 + Fe2O3) discrimination plat, where
average values of Gugher samples plot in the proximal ridge district and Nasirabad samples
plot in the pelagic district. The contribution of terrigenous materials, especially in the
formation of Nasiabad manganese layers can be traced in the Ce/La values. The Nasirabad
ore bearing layers are characterized by relatively high Ce/La ratio (2.08). Furthermore, such
high Ce/La values and strong linear correlations between Al2O3 and REE (Fig. 4) provide an
insight into the observed weak positive Ce anomaly in the Nasirabad ore‐bearing layers (Fig.
2).
Figure 6: The compassion of Abadeh Tash, Gugher and Nasirabad manganese and
ferromanganese deposits with Tokoro belt hydrothermal deposits [23], Peru basin
Hydrogenous nodules [34] and Eymir deposit [20].

A comparison of REE patterns of Gugher, Aabadeh Tashk and Nasirabad manganese and
ferromanganese deposits with Tokoro belt hydrothermal deposits [33], Peru basin
Hydrogenous nodules [34] and Eymir deposit [20] is shown in Fig. 6. It seems that the REE
pattern of manganese deposits associated with Zagros ophiolites are consistent with Tokoro
belt and Eymir hydrothermal deposits.

References

1‐ M. Rezaei, “The Investigation of Geology, Geochemistry and Genesis of Nasirabad Manganese deposit,
Neyriz, Fars province”, Unpublished M. Sc dissertation, Shahid Chamran University, Ahvaz, Iran (2012).
2‐ J. Shahabpour, Economic Geology, Shahid Bahonar University of Kerman Publications, 2002. (In
Persian).
3‐ M. Ghorbani, The economic geology of Iran: Mineral Deposits and Natural Resources. Springer geology,
2013.
4‐ A. Ghasemi and C.J. Talbot, “A new tectonic scenario for the Sanandaj–Sirjan Zone (Iran)”, Journal of
Asian Earth Sciences, 26 683–693 (2006).
5‐ M. Alavi, “Regional stratigraphy of the Zagros fold‐thrust belt of Iran and its proforeland evolution”,
American Journal of Science, 304 1‐20 (2004).
6‐ A.M. Ghazi, A.A. Hassanipak, J.J. Mahoney and R.A. Duncan, “Geochemical characteristics, 40Ar–39Ar
ages and original tectonic setting of the Band‐e‐Zeyarat/Dar Anar ophiolite, Makran accretionary prism,
S.E. Iran”, Tectonophysics, 393 175– 196 (2004).
7‐ H.A. Babaie, M.A. Ghazi, A. Babaie, T.E. La Tour and A.A. Hassanipak, “Geochemistry of arc volcanic
rocks of the Zagros Crush Zone, Neyriz Iran”, Journal of Asian Earth Sciences, 19 61–76 (2001).
8‐ M. Tangestani, L. Jaffari, R. Vincent and B.B. Maruthi Sridhar, “Spectral characterization and ASTER‐
based lithological mapping of an ophiolite complex: A case study from Neyriz ophiolite, SW Iran”,
Remote Sensing of Environment, 115 2243–2254 (2011).
9‐ M.A Rajabzadeh and N. Zamansani, “ Mineralization study of manganese deposits from Abadeh Tashk
area, Fars Province by using mineralogical and geochemical data”, in Proceeding of 3nd symposium of
Society of Economic geology of Iran, Ahvaz, Iran, 2012.
10‐ M.A Rajabzadeh and N. Zamansani,“ The investigation of manganese mineralization associated with
Neyriz ophiolitic mélange in Abadeh Tashk area, Fars Province: using mineralogical and geochemical
data”, Journal of Economic Geology, 2 201‐214 (2013) (In Persian with English abstract).
11‐ J. Stocklin, “Possible ancient continental margin in Iran”, in C.A. Burke and C.L. Drake (Eds.). The
Geology of Continental Margins, Springer, New York, 873‐887, 1974.
12‐ K. Heshmatbehzadi and J. Shahabpour, “Metallogeny of Manganese and Ferromanganese Ores in Baft
Ophiolitic Mélange, Kerman, Iran”, Australian Journal of Basic and Applied Sciences 4, 302‐313 (2010).
13‐ M. Bau and P. Möller, “REE systematics as source of information on minerogenesis”, in: Pagel, M.,
Leroy, J.L. (Eds.), Source, transport and deposition of metals. Balkema, Rotterdam, 17–20, 1991.
14‐ D. Chetty and J. Gutzmer, “REE redistribution during hydrothermal alteration of ores of the Kalahari
Manganese Deposit”, Ore Geology Reviews 47, 126–135 (2012).
15‐ D.E. Ruhlin and R.M. Owen, “The rare earth element geochemistry of hydrothermal sediments from
the East Pacific Rise: examination of a seawater scavenging mechanism”, Geochim. Casmochim. Acta,
50 393–400 (1986).

16‐ S.R. Taylor, S.M. Mclennan, The Continental Crustal: Its Composition and Evolution, Blackwell, Oxford
312, 1985.
17‐ M.N. Evensen, P. Hamilton and R.K. O’Nions, “Rare‐earth abundances in chondritic meteorites”,
Geochim. Cosmochim Acta, 42, 1199 (1978).
18‐ R. Jach and T. Dudek, “Origin of a Toarcian manganese carbonate/silicate deposit from the Krízna unit,
Tatra Mountains, Poland”, Chemical Geology, 224 136–152 (2005).
19‐ J. Xie, W. Sun, J. Du, W. Xu, L. Wu, S. Yang and S. Zhou, “Geochemical studies on Permian manganese
deposits in Guichi, eastern China: Implications for their origin and formative environments”, Journal of
Asian Earth Sciences, 74 155–166 (2013).
20‐ N. Oksuz, “Geochemical characteristics of the Eymir (Sorgun‐Yozgat) manganese deposit”, Journal of
Rare Earth, 29 287 (2011).
21‐ J. Wright and W.T. Hosler, “Paletedox variations in ancient oceans recorded by rare earth elements in
fossil apatite [J]”, Geochim. Cosmochim. Acta, 51 631 (1987).
22‐ J. Maynard, “The chemistry of manganese ores through time: a signal of increasing diversity of earth‐
surface environments”, Economic Geology, 105, 535–552 (2010).
23‐ A.V. Dubinin and L.L. Volkov, “Rare earth elements in metalliferous sediments of the East Pacific Rise [
J ] ”, Geokhimija, 5, 645 (1986).
24‐ J. Shahabpour, “Tectonic evolution of the orogenic belt in the region located between Kerman and
Neyriz ”, Journal of Asian Earth Sciences, 24, 405‐417 (2005).
25‐ A. Zarasvandi, M. Rezaei, H. Pourkaseb and A. Saki, “Investigation on primary and secondary processes
in Nasirabad manganese deposit, south of Neyriz: using mineralogy and Pb isotope geochemistry”,
Journal of Economic Geology 5, 37‐47 (2013). (In Persian with English abstract).
26‐ R.W. Murray, “Chemical criteria to identify the depositional environment of chert: general principles
and applications”, Sedimentary Geology, 90, 213‐232 (1994).
27‐ A. Karakus, B. Yavuz and S. Koc, “Mineralogy and major trace element geochemistry of the Haymana
manganese mineralizations, Ankara, Turkey ”,. Geochemistry International 48, 1014–1027 (2010).
28‐ M.T. Shah and A. Khan, “Geochemistry and origin of Mn‐deposits in the Waziristan ophiolite complex,
north Waziristan, Pakistan”, Mineralium Deposita, 34, 697–704 (1999).
29‐ M. Polgári, J.R. Hein, T. Vigh, M. Szabó‐Drubina, I. Fórizs, L. Bíró, A. Müller and A.L. Tóth, “ Microbial
processes and the origin of the Úrkút manganese deposit, Hungary”, Ore Geology Review, 47, 87–109
(2012).
30‐ M.T. Shah and C.J. Moon, “Manganese and ferromanganese ores from different tectonic settings in the
NW Himalayas, Pakistan”, Journal Asian Earth Science 29, 455–465 (2007).
31‐ N. Sabatino, R. Neri, A. Bellanca, H.C. Jenkyns, D. Masetti and G. Scopelliti, “ Petrography and high‐
resolution geochemical records of Lower Jurassic manganese‐rich deposits from Monte Mangart, Julian
Alps”, Palaeogeography, Palaeoclimatology, Palaeoecology, 299 97‐109 (2011).
32‐ A. Mücke, K. Adjimah and A. Annor, “Mineralogy, petrography, geochemistry and genesis of the
Paleoproterozoic Birimian manganese‐formation of Nsuta/Ghana”, Mineralium Deposita, 34 297–311
(1999).
33‐ J.H. Choi and Y. Hariya, “Geochemistry and depositional environment of Mn oxide deposits in the
Tokoro Belt, Northeastern Hokkaido, Japan”, Economic Geology, 87 1265–1274 (1992).
34‐ V. Stackelberg, Growth History of Manganese Nodules and Crusts of the Peru Basin, Geological Society,
Special Publications, London, 119–153, 1997.


Session VII: REE Recycling
Technologies


RECOVERY OF RARE EARTH METALS FROM END‐OF‐LIFE

PERMANENT MAGNET SCRAP: AN OVERVIEW
Yongxiang YANG
Department of Materials Science and Engineering,
Delft University of Technology, Mekelweg 2, 2628 CD Delft, The Netherlands

Email: Y.Yang@TUDelft.NL
Abstract
NdFeB permanent magnets have different life cycles and wide range of sizes, from less than a
gram to as big as 1000 ‐ 2000 kg. Recycling of rare earth elements (REEs) contained in this type
of magnets from the End‐of‐Life (EOL) products will play important and complementary role for
the primary supply of the rare earth metals in the future. However, collection and recovery of
the magnets from small consumer electronics imposes great social and technological
challenges. This paper gives a board overview about the sources of NdFeB permanent magnets
related to their applications, followed by summary of various available technologies to recovery
the rare earth metals in the magnets, including physical separation, and metallurgical recovery.
At present, no commercial operation has been identified for REE recycling from EoL NdFeB
permanent magnets. Most of the processing methods are still under research and development
stage.
Introduction
Neodymium‐iron‐boron (NdFeB) permanent magnets are the strongest and best available
magnets in the market since their finding in 1984, and are highly efficient and suitable for
lightweight mobile applications1. They are widely in wind turbines, hybrid and electric vehicles
(HEV and EV), household electrical appliances, computer hard disk drives (HDDs), and many
small consumer electronic devices.
NdFeB permanent magnets have different life cycles, depending on the applications: from as
short as 2‐3 years in consumer electronics to 20‐30 years in wind turbines. The size of the
magnets range from less than a gram in small consumer electronics to over 1 kg in EV and
HEVs, and can be as big as 1000 ‐ 2000 kg in the generators of modern wind turbines. NdFeB
magnet contains about 31‐32% REEs, mainly Nd+Pr plus a few minor (but expensive) heavy
REEs such as Dy, Tb and Ga, depending on the applications. Recycling of REEs in the NdFeB
magnets from the End‐of‐Life (EOL) products will be complementary for the primary supply of
the rare earth metals in the future. However, collection and recovery of the magnets from
small consumer electronics imposes great social and technological challenges. Moreover,
most EOL household electrical and electronic devices are currently shredded, the magnetic

content, mainly containing iron and neodymium, tend to end up in the generic ferrous metal
stream, which is extremely diluted for economic REE recovery.
At present, no industrial operation and mature technologies have been identified for recycling
the EOL NdFeB permanent magnets and the recovery of the associated REEs. It used to be
simply cheaper to buy newly manufactured magnets than to reprocess the complex scrap
material 2,3, especially for small‐sized products where manually separating the magnet is
challenging.
A few physical separation methods are recently reported for the treatment of the NdFeB
permanent magnets from EOL products in computer hard disk drives (HDDs) and air
conditioners, using mechanical shaking or hydrogen decrepitation4‐6. The reported processing
methods are still under intensive development for future commercial application. The
recovered magnet scrap still require metallurgical recovery and refining. From metallurgical
aspects, there are various possibilities to recover the REEs in the NdFeB magnet scrap,
including hydrometallurgical processing7,8, and pyrometallurgical processing such as slag
extraction9,10, liquid metal extraction11, molten salt extraction12‐14, chemical vapour
transport15,16. However, the majority of these techniques were developed to recover the REEs
from pre‐consumer (manufacturing) scrap or relatively clean and homogenous scrap with high
REE concentration. The recovery of REEs from the complex post‐consumer magnet scrap has
not been well studied before.
This paper will give an overview on the sources of EOL permanent magnets and characteristics
of their occurrence and applications, technical challenges and potential physical and
metallurgical separation methods for the recovery of the magnets and the associated REEs. It
is estimated that from 2030, the recycled REEs from the EOL permanent magnets will play a
significant role in the total REE supply in the magnet sector, if efficient technologies are
developed and used in practice17.
Applications of NdFeB magnets and the source for recycling
NdFeB magnet production and market demand
Global annual REE production remains at the 120,000 tons level in recent years (counted as
oxides ‐ REOs). Approximately 26,000 tons REEs (counted as REO) are used as NbFeB
permanent magnets which form the largest applications among all REEs both in tonnage and
in market value18,19, as is shown in Figure 1.
The great versatility of the REE magnet applications is also contributed by the fact that fine‐tuning
their chemical composition by adding some Dy, Tb, Nb, Co and Al can modify some of the physical
and magnetic properties to suit wide application needs20. Globally, about 76% of neodymium
(Nd), 69% praseodymium (Pr) and gadolinium (Ga), and 100% dysprosium (Dy) have been used
for permanent magnet production in recent years 21. In 2008, about 26,300 tons REEs including

Nd, Pr, Dy, Ga and Tb, counted as oxides, were used for magnet production.
Compared to other REEs, NdFeB permanent magnets and dysprosium are the most demanded
materials, largely due to the fast growing green technologies for clean energy and
transportation such as wind turbines and electric vehicles22.

Figure 1: Market share of REE magnets in volume and monetary value
(Created based on data of British Geological Survey18 ).
REE permanent magnet contains about 31‐32% REEs (mainly 21‐31% (Nd+Pr), 0‐10% Dy plus
small amount of Ga and Tb)23. Dy is critical for high temperature applications such as in
EVs/HEVs. It is expected that the NdFeB permanent magnet market will continue to grow, and
Figure 2 illustrates the global NdFeB permanent magnet production, and Table 1 lists the
magnet production in year 2012 and prediction for year 2015. As is clearly illustrated in Figure
2 and Table 1, Japan and China have been the two main REE magnet producing countries, and
in recent years China is by far the dominant producing country.
Table 1: Production of NdFeB permanent magnets (x 1000 metric tons)24

Production by country/region 2012 2015 prediction
China 50 65
Japan 10 8
Europe 1 1
USA 0 2

All others 2 2
Total REE magnet 63 78
Total REE in magnet (31%), estimates 19.5 24.2

Applications of NdFeB magnets
Figure 3 shows the relative REE magnet applications, predicted for the year end 2012 by Shaw
and Constantinides19. Computer hard disc drives (HDD) and CD/DVDs used to be the number 1

applications of the REE permanent magnets, and as is seen, motors and generators are
becoming the most important applications in recent years.

(a) Total gloabal NdFeB magnet production and prediction: 2005 – 202019

19
(b) Total global REE demands in the permanent magnets
Figure 2: Global production of NdFeB permanent magnets and the demand for the REEs

Figure 3: Prediction of REE permanent magnet application for 2008 ‐ 201219,25
According to Heli Magnetic Technology Co., Ltd. [http://www.helimagnet.com], the
proportion of NdFeB magnets used in different applications are estimated as follows:
 Over 40% for the electric bike motors, industrial motors, car motors and other
electrical products
 30% for microphones and sound amplification equipment

 10% for mobile phones, MP3, CD, DVD, headphones, and other consumer electronics
 8% for household electrical appliances
 5% photocopiers and other office equipment
 and 7% for the applications in other areas.

Table 2 (next page) lists the main applications of NdFeB magnets and the relative market size
and the total estimated annual demand for 2012/2013. The total estimated demand for the REE
magnets is about 63,000 tons, excluding the household appliances.
Recycling potential of NdFeB permanent magnets
According to the estimations of Du and Gradael26, global in‐use stocks for four rare earth
elements in NdFeB permanent magnets neodymium (Nd), praseodymium (Pr), terbium (Tb),
and dysprosium (Dy), amount to a total of 97.0 kt: 62.6 kt Nd, 15.7 kt Pr, 15.7 kt Dy, and 3.1 kt
Tb,. These stocks, if efficiently recycled, could provide a valuable supplement to geological
stocks as they are almost four times the 2007 annual extraction rate of the individual
elements. Different applications of the four REEs as permanent magnets estimated by Du and
Gradael26 are listed in Table 3.
Table 3: Estimated global in‐use stocks for the rare earth elements in NdFeB
permanent magnets in 200726
Stocks (kt) Nd Pr Dy Tb Total Share (%)
Computers 21.2 5.3 5.3 1.1 32.8 33.8
Audio systems 15.1 3.8 3.8 0.8 23.4 24.1
Wind turbines 10.1 2.5 2.5 0.5 15.7 16.2
Automobiles 9.8 2.5 2.5 0.5 15.2 15.7
Household 3.3 0.8 0.8 0.2 5.1 5.3
MRI 3.0 0.8 0.8 0.2 4.7 4.8
Total 62.6 15.7 15.7 3.1 97.0 ~100

Rademarker et al.17 estimated the recycling potential of NdFeB permanent magnets for 3
application sectors by considering the manufacturing and life span of the applications:
computer hard disc drives, wind turbines and automotive industry. The results show that for
some time to come, waste flows from permanent magnets will remain small relative to the
rapidly growing global REE demand. Policymakers therefore need to be aware that during the
next decade recycling is unlikely to substantially contribute to global REE supply. In the long
term, waste flows will increase sharply and will meet a substantial part of the total demand.
Figure 4 shows the recycling potential of Nd and Dy for the 3 studied application sectors.

Table 2: Applications of NdFeB permanent magnets and the estimated annual demand (2012/2013)
Sectors Product Concrete use Mass/ unit Total use (ton) Remarks
Computer HDDs, CDs, DVDs VCM, SP motors 10‐20 grams 14,200 (22.5%) Based on references 19,25.
Consumer Home electronics Small electric motors Speakers in sound system, mobile
Varies greatly No statistics
electronics & electrical appliances* and speakers phones and music players; washing
machines, refrigerators and air
conditioners etc.
200 kg/MW for hybrid drive, 600
Wind turbines Generation IV (only) Generators 400 kg/MW 16,000 (25.4%) kg/MW for direct drive25.
Based on Shaw and Constantinides 19.
Vehicles Conventional automobiles Small motors (40) 250 grams 20,900 (33.2%) Cars at different ages contain 220 ‐560
and sensors (20) g REE magnets/car 27.
83.5 million new vehicles in 2013
[247wallst.com].
HEVs, PEHVs and EVs Electric motors 1.25 kg** 2,200 (3.5%) 1,776,543 sales in 2013 [EVsRoll.com]
Electrical bikes (EB) Electric motors 300‐350 grams 9,700 (15.4%) 30 million/year in 201225
This figure does not include the

Total (incomplete) 63,000 (100%) magnet consumption from major
household appliances.
* Large household appliances include refrigerators (containing 40−60 g NdFeB magnets per unit), washing machines (80−180 g NdFeB/unit), and air
conditioners (60−400 g NdFeB/unit)27. Suppose an average REE magnet of 50‐100 g/unit and a global sales of 200 million unit per year: total magnet
consumption will be 10,000 – 20,000 ton/year for this sector.
**This figure does not include the magnets in the small motors and sensors like in the conventional cars25.
The main application sectors are conventional automotive (33.2%), wind turbines (25.4%), computer HDDs/CDs/DVDs (22.5%), E‐bikes (15.4%), and
EVs/HEVs (3.5%), according to this study. It can be seen that this differs from the estimations by Constantinides25 and Shaw and Constantinides19, however,
the classifications of the application areas are different. When recycling is concerned, the wide range of life span of different products need to be carefully
considered.

Figure 4: Recycling potential of Nd and Dy from EOL permanent magnets
(2011 – 2030 prediction) adapted from Rademarker et al.17
Rademarker et al.17 emphasized that for consumer products, one of the recycling challenges
is the physical dismantling and up‐concentration of small NdFeB magnets in the diversified
scrap. REE magnets in wind turbines and EV/HEV vehicles are more easily dismantled and
physically concentrated, and even reuse is possible after refurbishing. In general, efficient
metallurgical separation and refining processes remain the main challenges. Future REE
recycling efforts should, therefore, focus on the development of recycling technology and
build‐up of the infrastructure. It is estimated that about five to ten years is required to set up
a recycling practice.
Physical separation technologies
Destination of permanent magnets in different EOL products
Except for generators and motors in wind turbines and EV/HEVs, the REE permanent magnets
are not pre‐dismantled, due to their relatively small size, from their attached EOL products in
computer HDDs and other home electronics such as mobile phones, music players and
speakers in the sound systems. The EOL electronic products (e‐waste) are generally shredded
as a whole after removal of the hazardous components/materials. Because of the strong
permanent magnetism, the small magnets will stick to ferrous scrap (steel) after shredding
and physical separation using such as magnetic separators, or to a much smaller extent
trapped to the non‐ferrous scrap if the magnets are not liberated and still stuck to the non‐

ferrous components. At present, almost all small permanent magnets used in consumer
electronics (at relatively low collection rate), after collection and shredding operation, are lost
into ferrous scrap and reported into slag during metallurgical smelting or refining processes
of the major metals.
For the generators in wind turbines the REE permanent magnets, after reaching their end of
first life, can still be manually dismantled and reused or recovered for the REEs much more
easily due to their large size (250 – 600 kg/MW power). For the conventional automobiles,
there are around 40 magnets in small motors and actuators and 20 sensors in a typical
passenger car, and on average there are about 250 grams of NdFeB magnets and 10‐20 grams
SmCo magnets19. These small permanent magnets are not pre‐dismantled before car
shredder in the present industrial practice, and they are simply lost to the ferrous scrap similar
to the magnets in the consumer electronics. The REE permanent magnets in the electric
motors of EVs and HEVs are relatively larger than in normal cars (on average 1.25 kg, see Table
2) but still difficult to dismantle manually, and thus are more likely lost in the ferrous scrap.
Depending on the market value, pre‐dismantling of the REE magnet‐containing components
in major consumer products should be an option for further mechanical and metallurgical
recovery in the future. Therefore, there is an increasingly great demand for the design of
magnet for dismantling in the consumer products.
Analysis of REE magnets in the shredder material streams
Conventional shredding for End‐of‐Life Vehicles (ELVs) and household appliances
As is discussed earlier, most of the permanent magnets nowadays go to shredders together with
the bonded EOL products. A recent study by Bandara et al.27 gives a good overview about the
destinations and concentrations of the permanent magnets and the associated REEs. Their
study for shredding ELVs and household appliances indicates that the ferrous scrap of the
shredder products attracts most of the magnetic materials and contains about 130 – 290 grams
of Nd/ton (ppm). This concentration is even much lower than the REEs content in some mine
tailings of up to 1000 – 1500 ppm. Without pre‐concentration and novel technologies, the low
REE content in the scrap is very difficult to recover. Thus, technologies for pre‐shredder
separation are needed for any feasible metallurgical recovery.
Separated shredding of pre‐dismantled HDDs and air conditioners
Shredding and physical upgrading of pre‐dismantled computer HDDs
Pre‐dismantling of computer hard disk drives have been reported by some companies mainly
for data security reasons, although this is not a standard practice. The dismantled hard disk
drives are shredded and normally the ferrous and non‐ferrous scrap as well as the printed
circuit boards (PCBs) are recovered. In most cases, the REE magnets particles after shredding
are not recovered although the REE concentration is much higher than those in the shredder

ferrous products from the whole EOL product (e.g. computers).
In the Netherlands, Van Gansewinkel Group operates a small shredder for HDD shredding
(data destruction). TU Delft has recently developed a REE recover process from the magnet
concentrated HDD shredder residues10,28. The separately collected magnet residues go
through a thermal demagnetization process followed by grinding and screening. The highly
concentrated magnet particles (containing up to 25% Nd) separated from the ductile metal
pieces can go through either hydrometallurgical extraction or first processed with slag
extraction followed by REE leaching. The REE product of fluoride or oxide can be prepared for
final RE metal production. Figure 5 illustrates the developed processing flowsheet. The
magnet concentrate particles of less than 1 mm size range represent 63% of the total collected
shredder residue, and a total of 95% recovery rate has been achieved for all the REEs in the
collected residues through physical processing.

Figure 5:
Processing
flowsheet for
the REE recovery from EOL HDDs at TU Delft.
Hydrogen decrepitation of pre‐dismantled computer HDDs
University of Birmingham has developed a hydrogen processing method to selectively convert
sintered NdFeB magnet into hydride fine particles at room temperature and atmospheric
pressure, or high temperature and pressure (170°C and 7 bar) for Ni alloy coated magnets4,5.
The formed Nd hydride has a 5% volume expansion which makes the Nd hydrides into fine
demagnetized powders. The fine powders can be easily separated from the main structure of
the HDDs. The Nd hydrides can be converted to REE metal (alloy) at elevated temperatures by
releasing hydrogen (a desorption process). If the source of the magnets is very clean, the REE
magnet powder could be re‐sued directly for magnet manufacturing. If the source of magnets
is contaminated, the magnet powder could be re‐processed through chemical and
metallurgical technologies back to pure REE or their alloys. Further research is going on to
treat the low concentration EOL magnet scrap in a recently EU funded project “EREAN”
(European Rare Earth Magnet Recycling Network: http://erean.eu).

Mechanical separation of magnets from HDDs and air conditioners
Hitachi Group developed a mechanical dismantling and separation technique for NdFeB
magnets in HDDs and air conditioners, using a rotational dram. Through vibration and impact
by tumbling, the screws fixing the HDD casing become loose, and the magnets can be separated
out of the casing in about 30 minutes. The apparatus can process 100 HDDs per batch and 200
HDDs per hour, much faster than manual dismantling of 12 units per hour6,29. Hitachi and
Mitsubishi Materials have also developed similar technology to recovery the NdFeB magnets
from the compressors in air conditioners. After cutting off the casing, the rotors in the
compressor can be removed and treated by either non‐thermal demagnetization (Hitachi) or
thermal demagnetization at 400 – 500°C (Mitsubishi). Hitachi may use the similar rotating drum
for HDDs disintegration to loosen the magnets from the air conditioner, and Mitsubishi uses
other equipment (e.g. drills). Furthermore, Mitsubishi Materials is also trying to separate the
NdFeB magnets from direct drive (DD) motors in washing machines through thermal
demagnetization and removing the resin bounding6,30.
Metallurgical extraction and separation
There have been many studies for the recovery of Nd and other associated REEs (Pr, Dy. Tb)
using various types of metallurgical processes: hydrometallurgical, pyrometallurgical and
electrochemical, or their combinations. However, most of these studies have focused on the
highly concentrated and relatively clean waste magnets of manufacturing scrap or production
waste: the new solid scrap and cutting sludge “swarf”31 which accounts for 20‐30% of the
starting alloy for manufacturing the magnets6. The solid scrap is normally re‐melted (after
coating removal) within the magnet manufacturing plant for direct re‐use. However, the swarf
requires chemical or metallurgical separation and refining mostly through hydrometallurgical
leaching route to produce rare earth fluorides or oxides as raw materials for conventional metal
production (molten salt electrolysis or metallothermic reduction). Recently published two
comprehensive reviews on rare earths recycling by Binnemans et al.32 and Tanaka et al.6 are
good sources of reference. Binnemans et al.32 gave a good comparison for different metallurgical
methods as is partially listed in Table 4 below. It should be noted that except for direct melting
of the relative clean magnet scrap, all the listed methods convert the REE magnet in the scrap
or waste into REE halides or oxides, which need to be reduced to RE metal or alloys through
either molten salt electrolysis or metallothermic reduction, like the treatment of “swarf”.
Table 4: Comparison of different metallurgical methods for REE recovery from NdFeB
magnet scrap, extracted from Binnemans et al.32
Methods Advantages Disadvantages
 Generally applicable to all types  Many process steps required
of magnet compositions before obtaining new magnets
Hydrometallurgical  Applicable to non‐oxidized and  Consumption of large amount
oxidized alloys of chemicals

 Same processing steps as those  Generation of large amount of
for extraction of rare earths from waste water and effluents
primary ores
 Generally applicable to all types  High energy input required
of magnet compositions  Direct smelting and liquid metal
Pyrometallurgical:  No generation of waste water extraction cannot be applied to
Liquid‐phase and effluents oxidized magnets
processing  Fewer processing steps vs.  Glass slag method generate
hydrometallurgical methods large amounts of solid waste,
 Direct melting for obtaining REE‐bearing slag still requires
master alloys further hydrometallurgical
 Liquid metal extraction allows for processing.
obtaining REEs in metallic state
 Generally applicable to all types  Consumption of large amount
Pyrometallurgical: of magnet compositions of chlorine gas
gas‐phase  Applicable to non‐oxidized and  Aluminum chloride is very
extraction oxidized alloys corrosive
 No generation of waste water and 
effluents
For the small magnets used in the EOL consumer products, little recycling efforts could be
identified and it was commonly thought that less than 1% of the EOL REEs (including REE
magnets) has been recycled, mainly due to the their small size, tight bonds with the product
and the multiple type mixture of different magnets (ferrites, SmCo and NdFeB)6. It is one of
the main objectives of this paper to identify feasible metallurgical recovery techniques for
extraction of the small sized magnets in their bound EOL products. Although the past research
has focused mainly on the manufacturing scrap and residue (swarf) recycling, the developed
methods could be adapted to the EOL magnet scrap for REE recovery.
As is discussed in previously, the magnet or REE concentration would be too low for
metallurgical recovery if the EOL products (automobiles, home electrical appliances and
consumer electronics) are directly shredded. Pre‐dismantling and up‐concentration through
physical processing are critical for a viable chemical or metallurgical extraction. Recovery of
the minor amount of REEs from the magnet scrap is both technically and economically very
challenging. Allocating the REEs originated from the magnet scrap in the by‐product or waste
streams of smelting processes for main metals is a potential possibility, including slags, flue
dust, solid leach residues and waste solutions. Therefore, a total materials recovery from EOL
products (containing REE magnets) would be the direction for future research and
development, in parallel with magnet pre‐dismantling approach.
Hydrometallurgical methods
The key steps for hydrometallurgical route are leaching of the magnet scrap or waste, followed
by precipitation of the REE salts or compounds and further conversion to REE fluorides or oxides,
quite often with REE separation or up‐concentration in between. The leaching can be performed
in three different ways: (1) complete leaching (with or without roasting, (2) roasting – selective

REE leaching, and (3) selective conversion of REEs. Furthermore, conversion of REEs in the
magnet scrap into REE compound precipitates based on solubility of REE salt at different
temperatures or under hydrothermal conditions is also an option. Table 5 lists the major
hydrometallurgical methods reported in the literature.
The challenges for the hydrometallurgical leaching process to recover the RREs in the EOL
permanent magnets, in particular the REE dilute waste streams, are the selective dissolution,
up‐concentration and separation of REE species from the major species, and total recovery of
all major metals at the same time. Therefore, new separation technologies such as novel
solvent extractants, ionic liquid, ion exchange resin or the alike, are strongly needed to extract
REE species from low concentration leaching solutions. Alternatively, anodic dissolution of
NdFeB permanent magnet scrap is a potential option to dissolve REE species
electrochemically33, separating non‐REEs in the scrap into anode slimes or precipitated onto
the cathode, through accurate control of electrode and cell potentials and current density on
the electrodes, as well as proper selection of electrolyte.
Pyrometallurgical methods
The main purpose of high temperature processing of NdFeB magnet scrap is to selectively
convert REEs in the magnet into another phase which separate the main non‐REE
components. The separated REEs in another phase would be much more concentrated and
can be used for RE metal production through molten salt electrolysis or metallothermic
reduction. The identified pyrometallurgical extraction methods can be divided into the
following groups: (1) selective chlorination, (2) liquid metal extraction, (3) molten salt
extraction, (4) molten slag extraction, and (5) electrochemical processing.
Table 5: Summary of main hydrometallurgical REE recovery methods
from NdFeB magnet scrap
Methods Main features References

Leaching or conversion
Complete leaching With H2SO4 leaching at room temperature, Layman and
(with/without all components in NdFeB magnet will be Palmer8,
roasting) dissolved. Nd can be precipitated as double Abrahami et al.28
slat which can be converted to NdF3 or Nd2O3.
Large amount of Fe dissolved in the solution
needs to be properly disposed (e.g. jarosite
precipitation). Yoon et al.34
Roasting at 500°C for sintered scrap and at
700°C for bonded scrap followed by H2SO4 (2
mole/l) leaching at 50°C for 2 h and double
salt precipitation. Nd yield was over 99.9%.

Selective leaching Roasting in air at 700°C followed by H2SO4 (4 Lee et al.35
(with roasting) mole/l) at 70°C for 3 hours with pulp density
of 100 kg/m3, with 70% Nd recovery.
Roasting at 900°C for 6 h followed with HCl Koyama et al.36,
(0.02 mole/l) leaching at 180°C for 2 h. More Koyama and
than 99% REE and less than 5% Fe recovery. Tanaka37
Fractional A simultaneous dissolution and precipitation
crystallization process by making use the decreased Wei et al.38,
(Nd sulfate conversion) solubility of Nd2(SO4)3 at elevated Sato et al.39
temperatures. Adding ethanol to H2SO4
facilitates the precipitation. With 97.1% Nd
yield and 96.8% purity of Nd sulfate hydrate.
Hydrothermal Using mixture of HCl (3 mole/l) and oxalic acid Itakula et al.7
conversion (0.2 mole/l) to recover Nd as oxalate.
Operating condition: 110 °C and 6 h. Over 99%
of Nd in the magnet recovered as solid
precipitate of Nd2(C2O4)3 with of 99.8 purity %.

REE separation & concentration
Solvent extraction Conventional process with many different Xie et al.40
organic extractants for REE separation in
primary REEs production. This technique
would be applicable to the REE recycling in the
hydrometallurgical route after leaching for REE
separation from non‐REEs (may require
necessary modifications) and separation of
individual REEs.
Ionic liquid extraction Potential alternative to organic solvent Baba et al.41,42,
extractants due to low vapor pressure and Binnemans43
stability. Ionic‐liquid based liquid membrane
with permeability to REE ions.

Selective chlorination
For selective chlorination of REEs in the magnet, various chlorination agents have been
investigated: using metal halides such as MgCl2 to convert the REEs into molten NdCl3 and DyCl3
at 1000°C12,13, using FeCl2 at 800°C16, using NH4Cl at much lower temperature of 300°C15 , or
using Cl2 gas and AlCl3 (the so called chemical vapor transport process)44. In most of these cases,
the RECl3 (NdCl3 and DyCl3) were distilled at elevated temperatures to separate from Fe‐based
residue. For all chlorination approach, selectivity and the reaction rate would be key issues, in
particular when treating highly contaminated and partially oxidized magnet scrap.
Liquid metal extraction
The Ames Laboratory has developed a liquid metal extraction process to recycle REEs from
NdFeB and other alloys45, and later studied further by other researchers using Mg as extractant
11, 46, 47
. However, this process is very slow (24 – 72 hours) and cannot be applied to (partly)
oxidized NdFeB scrap.

Molten salt extraction
Okabe and Shirayama12,13 used molten MgCl2 to selectively dissolve Nd and Dy from the NdFeB
magnet scrap at 1000°C, and about 90% of REEs can be extracted in about 12 hours. The
obtained molten mixture of Mg – MgCl2 – NdCl3 – DyCl3 goes through a distillation process to
remove Mg and MgCl2 as vapor, and the separated REE chloride can be used for RE metal
production. Very recently, Hua et al.14 developed an integrated process using molten chloride
slat (MgCl – KCl) extraction to selectively dissolve Nd and other REEs in the magnet scrap at
temperature of 700 ~ 1000°C. An REE recovery of approximately 90% has been obtained, and
the formed NdCl3 in MgCl2 – KCl salt can be subsequently electrolyzed to produce directly Mg‐
Nd alloy. Tanaka et al.6,48 proposed a process by using molten fluoride to remove the oxidized
rare earth in the NdFeB magnet scrap. In this process, the molten fluoride reacts with the rare
earth oxides and the formed REO – fluoride mixture is used in molten salt electrolysis for RE
metal production. The un‐oxidized NdFeB scrap remain unreacted and is thus cleaned as master
alloy for magnet preparation again. It should be noted that molten chloride salts can be used for
treatment of un‐oxidized magnet scrap only, while molten fluoride salts are suitable for the
treatment of oxidized magnet scrap only.
Molten slag extraction
Two slag treatment methods are reported in the literature. In the glass slag method, the REE
alloy is brought into contact with a molten B2O3 that is able to selectively oxidize and dissolve
the rare earths from the alloys to form Nd2O3‐B2O3 slag9. The slag can be subsequently leached
with acid to extract Nd precipitate such as Nd(OH)3, but with relatively low Nd recovery49.
The second slag method is the use of CaO‐SiO2‐Al2O3 or CaO‐CaF2 fluxes to extract the REEs in
the complex magnet scrap at about 1500°C10. This method is very effective for the EOL magnet
scrap such as HDD shredder residues with more than 99% REE recovery. It can separate
completely the REEs in the magnet from the ferrous materials in the NdFeB magnet and the
mixed scrap. The REE‐rich slag can be used for hydrometallurgical extraction, and the
separated iron‐rich alloy can be a good source of raw materials for steelmaking. The
advantages of molten slag extraction is its capability of effective separation of REEs and iron,
and the recovery of both fractions at the same time.
According to Tanaka et al.6, Yamamoto’s group reported an in‐plant waste (sludge) separation
process where the magnet waste separates into liquid iron and solid rare‐earth oxide phases at
1550°C under argon atmosphere with low O2 partial pressure without any flux additions. It was
found that all of the REEs in the magnet waste were converted to the oxide phase, and the
formed REOs can be converted to RE metal through conventional technologies.
Electrochemical process
The electrochemical processes can be applied to the metallic magnet scrap based on electro‐
refining principles in the molten salt6,50,51. In a molten salt electro‐refining cell, the impure
REE magnet scrap used as starting anode can be selectively dissolved for rare earth elements

(Nd and/or Dy) in a molten chloride or fluoride electrolyte, and on the cathode only REEs are
preferably deposited forming Nd‐based alloys. In the cell a rare earth – iron group metal alloy
is used as bipolar electrode (diaphragm).
As can be seen from the above discussion, most of the pyrometallurgical methods are
applicable to only highly concentrated magnet scrap (production waste: off‐specs and swarf),
and adaptation is required to treat the dilute EOL magnet scrap. Slag extraction may be a very
promising method to separate and up‐concentrate the REEs in the low concentration magnet‐
bearing scrap such as shredder residues, however, subsequent hydrometallurgical extraction
such as REE leaching is needed.

Concluding remarks
This paper addresses the recycling potential of NdFeB permanent magnets and the identified
physical and metallurgical separation and recovery technologies. It is indicated that the annual
consumption and demand of the REE magnets are continuously increasing, and the recycling
of the REEs in the EOL permanent magnets will be complementary to the supply shortage in
the long run. REE recovery from the manufacturing scrap and waste (swarf) of the NdFeB
magnets is already a good industrial practice with available technologies.
However, recycling and recovery of REEs (Nd, Pr, Dy, Tb) from EOL magnets are challenging
due to their relative small size used in consumer electrical and electronics, as well as in the
conventional automotive industry. Shredding without pre‐dismantling of the EOL consumer
products will pose great difficulties for up‐concentration of the permanent magnets.
Development of advanced and low cost of pre‐dissembling technologies or equipment will be
crucial for the metallurgical recovery of REEs. This has been demonstrated by the
developments for computer hard disk drives and air conditioners by Hitachi and Mitsubishi,
Birmingham University, Delft University of Technology, and active European initiatives such as
EREAN (http://erean.eu) and REEcover consortia and other global government supports in
Japan and USA reported in recent years.
There are many different types of metallurgical recovery methods identified from the
literature, however, most of them have been studied for relatively clean and highly
concentrated new or manufacturing magnet scrap and waste. Adaptation of these
technologies to the dilute and complex EOL magnet scrap is highly needed. At the same time,
recovery of the major metals and materials in the dilute magnet waste stream (e.g. shredder
products) is an important factor for the total success of the permanent magnet and REE
recycling. It is believed that no single metallurgical process can be successful for the efficient
recovery of the REEs from their EOL secondary resource, and a combination of
hydrometallurgical, pyrometallurgical and/or electro‐chemical technologies will be the future

solution. It is expected that through the globally increasing efforts (Europe, Japan, USA, China)
for the research and development, establishment of technologically and economically feasible
REE magnet recycling industry could be realized in the near future, if the NdFeB permanent
magnet market continue to develop.

References
1. Jiles, D., Introduction to Magnetism and Magnetic Materials, second edition, Chapman & Hall, 1998, New
York.
2. Schüler D., Buchert M., Liu R., Dittrich S. and Merz C., Study on Rare Earths and Their Recycling – Final report
for the Greens/EFA Group in the European Parliament, Öko‐Institute e.V., January 2011.
3. Moss R.L., Tzimas E., Kara H., Willis P. and Kooroshy J., Critical Metals in Strategic Energy Technologies:
Assessing Rare Metals as Supply‐chain Bottlenecks in Low Carbon Energy Technologies, 2011. European
Commission JRC Institute for Energy and Transport.
4. Zakotnik M., Harris I.R. and Williams A.J., Possible methods of recycling NdFeB‐type sintered magnets using
the HD/degassing process. J. Alloys Compd., 2008, 450. 525‐531.
5. Walton A. and Williams A., Rare earth recovery. Materials World. 2011, 19. 24‐26.
6. Tanaka M., Oki T., Koyama K., Narita H. and Oishi T., Chapter 255 ‐ Recycling of rare earths from scrap. In:
Handbook on the Physics and Chemistry of Rare Earths, 43, 2013, Elsevier.
7. Itakura T., Sasai R., Itoh H., Resource recovery from Nd‐Fe‐B sintered magnet by hydrothermal treatment. J.
Alloys Compd., 2006, 408–412. 1382‐1385.
8. Lyman J. W. and Palmer G.R., Recycling of rare earths and iron from NdFeB magnet scrap. High Temperature
Materials and Processes, 1993,11 (1‐4). 175‐187.
9. Saito T., Sato H., Ozawa S., Yu J. and Motegi T., The extraction of Nd from waste NdFeB alloys by the glass
slag method, J. Alloys and Compounds, 2003, 353. 189‐193.
10. Yang Y., Abrahami S.T. and Xiao Y., Recovery of rare earth elements from EOL permanent magnets with
molten slag extraction. In: Proceedings of the 3rd International Slag Valorisation Symposium. 19‐20 March
2013, Leuven, Belgium. 249‐252.
11. Okabe T. H., Takeda O., Fukuda K. and Umetsu, Y., Direct extraction and recovery of neodymium metals from
magnet scrap, Mater. Trans., 2003, 44. 798‐801.
12. Shirayama S. and Okabe T.H., Recovery of Nd and Dy from rare earth magnet scrap. Presentation at The 4th
Workshop on Reactive Metal Processing (RMW4), March 14 15, 2008, Cambridge, MIT.
(http://www.okabe.iis.u‐tokyo.ac.jp/core‐to‐core/rmw/RMW4 /RMW4_23_Shirayama_Q.pdf)
13. Okabe T.H. and Shirayama S., Method and Apparatus for Recovery of Rare Earth Element, US Patent
Application Publication: US2011/0023660A1.
14. Hua Z., Wang L., Yang Y., Wang J., Li X. and Wang H., Recycling of rare earths in NdFeB magnet scraps with
molten chloride. Poster presentation in: ICRE’2013 ‐ The 7th International Conference on Rare Earth
Development and Application. August 10‐13, 2013, Ganzhou, China.
15. Itoh M., Miura K., and Machida K.‐I., Novel rare earth recovery process on Nd‐FeB magnet scrap by selective
chlorination using NH4Cl, J. Alloys Compd., 2009, 477. 484‐487.
16. Uda T., Recovery of rare earths from magnet sludge by FeCl2. Materials Transactions, 2002,43: 55‐62.
17. Rademaker J. H., Kleijn R. and Yang Y., Recycling as a strategy against rare earth element criticality: a systemic
evaluation of the potential yield of NdFeB magnet recycling. Environmental Science & Technology, 2013, 47.
10129‐10136.
18. British Geological Survey (BGS). Rare Earth Elements. 2010. http://www.MineralsUK.com
19. Shaw S. and Constantinides S. Permanent magnets: the demand for rare earths. Presentation in: 8th
International Rare Earths Conference. 13‐15 November 2012, Hong Kong.
20. Sagawa M., Fujimura S., Yamamoto H., Matsuura Y. and Hiraga K., Permanent magnet materials based on

the rare earth‐iron‐boron tetragonal compounds. IEEE Transactions on Magnetics, 1984,20 (5). 1584‐1589.
21. Goonan T. G.. Rare Earth Elements ‐ End Use and Recyclability. USGS Scientific Investigations Report 2011–
5094. 2011.
22. Alonso E., Sherman A. M., Wallington T. J., Everson M. P., FField. R., Roth R. and. Kirchain R. E, Evaluating
rare earth element availability: A case with revolutionary demand from clean technologies. Environ. Sci.
Technol., 2012, 46. 3406−3414.
23. Kozawa S., Trends and Problems in Research of Permanent Magnets for Motors ‐ Addressing Scarcity
Problem of Rare Earth Elements. Science & Technology Trends. Quarterly Review. 2011 (January ) No. 38.
40‐54.
24. Benecki W. T., The Permanent Magnet Market – 2015. Presentation in: Magnetics 2013 Conference.
February 7‐8, 2013 Orlando, Florida.
25. Constantinides S. The Magnetic Material Challenge. Presentation in: ARPA‐E Workshop: Rare Earth and
Critical Materials. December 6, 2010 in Arlington, VA.
26. Du X. and Graedel T. E., Global rare earth in‐use stocks in NdFeB permanent magnets. Journal of Industrial
Ecology, 2011, 15(6). 836‐843.
27. Bandara H. M. D., Darcy J. W., Apelian D., and Emmert M. H., Value analysis of neodymium content in
shredder feed: towards enabling the feasibility of rare earth magnet recycling. Environ. Sci. Technol., 2014,
48. 6553−6560.
28. Abrahami S.T., Xiao Y. and Yang Y., Rare‐earth elements recovery from post‐consumer hard‐disc drives. Paper
accepted for publication in: Mineral Processing and Extractive Metallurgy, 2014.
29. Baba K., Nemoto T., Maruyama H., Taketani N., Itayagoshi K. and Hirose Y., Hitachi’s Involvement in Material
Resource Recycling. Hitachi Review 2010, 59. 180–187.
30. Arai Y., Recycling of rare earth magnet from used home appliances. Presentation in: Symposium on Recycling
of Rare Metals, July 9, 2010. Japan Society of Material Cycles and Waste Management, Tokyo
(http://jsmcwm.or.jp/recycle/7_Arai_MMC.pdf, accessed 20 June 20140).
31. Bounds C.O., The recycle of sintered magnet swarf. In: Metals and Materials Waste Reduction, Recovery and
Mediation, K.C. Liddell, et al. (ed.), Pennsylvania, 1994. 173‐186.
32. Binnemans K., Jones P. T., Blanpain B., Van Gerven T., Yang Y., Walton A. and Buchert M., Recycling of rare
earths: a critical review. Journal of Cleaner Production, 2013, 51. 1‐22.
33. Prakash V., Sun Z.H.I., Sietsma J. and Yang Y., Electrochemical recovery of rare earth elements from magnet
scraps ‐ a theoretical analysis. Paper to be presented in: ERES2014 ‐ 1st European Rare Earth Resources
Conference, 04‐07 September 2014, Milos, Greece.
34. Yoon H.S., Kim C.J., Lee J.Y., Kim S.D. and Lee J.C., Separation of neodymium from NdFeB permanent magnet
scrap. J. of Korean Inst. of Resources Recycling, 2003, 12( 6). 57‐63.
35. Lee J.‐c., Kim W.‐b., Jeong J., Yoon I.J., Extraction of neodymium from Nd‐Fe‐B magnet scraps by sulfuric acid.
J. Korean Inst. Met. Mater., 1998, 36(6). 967‐972.
36. Koyama K., Kitajima A. and Tanaka M., Selective leaching of rare‐earth elements from an Nd‐Fe‐B magnet,
Kidorui (Rare Earths), 2009, 54. 36‐37.
37. Koyama, K., Tanaka, M., In: The Latest Technology Trend and Resource Strategy of Rare Earths, Machida, K.
(Ed.), CMC Press, 2011, Tokyo. 127‐131.
38. Wei Y., Sato N. and Nanjo M., Solubility of samarium sulfate and neodymium sulfate in sulfate solutions ‐
fundamental study on the recycling of rare earth magnet materials. Shigen to Sozai (J. MMIJ), 1989, 105.
965‐970.
39. Sato N., Wei Y., Nanjo M. and Tokuda M., Shigen to Sozai (J. MMIJ), 1997, 113. 1082‐1086.
40. Xie F., Zhang T.A., Dreisinger D. and Doyle F., A critical review on solvent extraction of rare earths from
aqueous solutions. Minerals Engineering, 2014, 56. 10‐28.
41. Baba Y., Kubota F., Kamiya N. and Goto, M., Recent advances in extraction and separation of rare‐earth
metals using ionic liquids. J. Chem. Eng. Jpn., 2011, 44. 679‐685.
42. Baba Y., Kubota F., Kamiya N. and Goto M., Selective recovery of dysprosium and neodymium ions by a
supported liquid membrane based on ionic liquids. Solvent. Extract. Res. Dev., 2011, 18. 193‐198.
43. Binnemans, K., Lanthanides and actinides in ionic liquids. Chem. Rev., 2007, 107. 2592‐2614

44. Murase K., Machida K. and Adachi G., Recovery of rare metals from scrap of rare earth intermetallic material
by chemical vapour transport. J. Alloys and Compounds, 1995, 217. 218‐225.
45. Ellis T.W. and Schmidt F.A., Recycling of rare earth metals from rare earth‐transition metal alloy scrap by
liquid metal extraction. U.S. Patent 5,437,709 (1995).
46. Xu Y., Chumbley L.S. and Laabs F.C.. Liquid metal extraction of Nd from NdFeB magnet scrap, J. Mater. Res.,
2000, 15: 2296‐2304.
47. Takeda O., Okabe T. H. and Umetsu Y., Recovery of neodymium from a mixture of magnet scrap and other
scrap. J. Alloys and Compounds, 2006, 408–412. 387‐390.
48. Tanaka M., Sato Y., Huang Y., Narita H. and Koyama K., In: Proceedings of the 10th International Symposium
on East Asian Resources Recycling Technology (EARTH2009). The Korean Institute of Resources Recycling. 2–
6 Nov. 2009, Seoul. 200‐203.
49. Saito T., Sato H. and Motegi T., Recovery of rare earths from sludges containing rare‐earth elements. J. Alloys
and Compounds, 2006, 425: 145‐147.
50. Oishi T., Konishi H., Nohira T., Tanaka M. and Usui T., Kagaku Kogaku Ronbunshu, 2010, 36. 299.
51. Kobayashi S., Kobayashi K., Nohira T., Hagiwara R., Oishi T. and Konishi H., Electrochemical Formation of Nd‐Ni
Alloys in Molten LiF‐CaF2‐NdF3. J. Electrochem. Soc., 2011, 158. E142‐E146.


LEACHING OF RARE EARTH ELEMENTS: PAST AND PRESENT
Sebastiaan PEELMAN1, Zhi H.I. SUN1, Jilt SIETSMA1, Yongxiang YANG1
1
Department of Materials Science and Engineering, Delft University of Technology, Mekelweg 2, 2628
CD Delft, The Netherlands
E‐mail: s.peelman@tudelft.nl
Abstract
The recycling of Rare Earth Elements (REE) from mine tailings and REE containing End‐of‐Life
(EoL) scrap, the two main secondary resources, offers great opportunities to secure REE supply
in Europe. The relatively low concentration of REE in mine tailings (1000‐1500 ppm) and the
variety in contaminants in the EoL scrap have made the extraction of REE from these resources
very challenging. This paper provides a review of the past and present technologies for REE
leaching used in primary REE production and in REE recycling, as an important part of
hydrometallurgical REE processing. Detailed studies of existing processes are essential to
properly understand and resolve the difficulties in REE recycling from these secondary
resources. The known processes range from acid leaching with H2SO4, HCl or HNO3 for primary
ores, to leaching with NaCl or (NH4)2SO4 of ion adsorbed clays and combined base and acid
leaching for EoL lamp phosphors. A comprehensive understanding of these processes is the
key to applying them to REE recycling from secondary resources.
I. Introduction
The rare earth elements (REE) are a group of 17 chemically similar elements consisting of the
lanthanides, Y and Sc. Their unique physical and chemical properties have made them
essential in state‐of‐the‐art equipment (e.g. magnets, catalysts, batteries). A potential supply
risk of REE, environmental concerns and economic benefits promote research on processing
of secondary materials, such as mine tailings or electronic waste, in the EU. Although
significant investigations on REE extraction from secondary resources have been carried out
and a variety of technologies have been developed or proposed, most of them are still in the
stage of research or only suitable for some specific secondary resources. The REE extraction
efficiency and selectivity, together with the cost and engineering during process design, still
require substantial optimisation before further commercialisation. At the same time, the
understanding of interactions between REE minerals in a low concentration and the leaching
media, as well as the behaviour of selective leaching of different REE phases in magnet scraps,
is still not sufficient. In order to have a better view on the state‐of‐the‐art and to improve
engineering possibilities of REE extraction from secondary resources, a review of the most
prevailed leaching technologies that are currently in use in REE production is needed.

Table 1: Summary of leaching technologies in primary REE production
REE Ref.
Mineral Process Remarks Status
yield
1) HCl leach to remove non REE carbonate
85‐90% The oldest way to process bastnaesite concentrates outdated [3]
2) Calcination of residue to form REO
Acid choice depends on further processing:
Digestion with HNO3 or H2SO4 98% solvent extraction  HNO3 outdated [3]
precipitation  H2SO4
1) Roast at 620°C to drive of CO2 Ce+III oxidises to Ce+IV during roasting  will not leach
Bastnaesite ‐‐ outdated [3]
2) 30% HCl leach REE fluorides will not leach, residue is marketable
1) Alkaline conversion REF3  RE(OH)3 Process can be preceded with HCl leach to extract REE
‐‐ In use [3]
2) HCl leach carbonates before alkaline conversion
1) Sulphuric acid roast
Precipitates are converted to chlorides for further
2) NaCl solution leach ‐‐ In use [3]
purification with solvent extraction
3) Precipitation as Na double sulphates
Process conditions determine what is leached: only LREE or
Digestion in hot H2SO4 ‐‐ LREE+HREE+Th outdated [3]
Yields no pure product
1) Digestion in hot 60‐70% NaOH Ce cannot be leached if Mn is present
Monazite 2) Washing residue with hot water 98% Th is leached together with REE In use [3‐4]
3) Leach with mineral acid of choice Na3PO4 is marketable by‐product
1) Heat under reducing and sulphidizing Requires no fine grinding
atmosphere with CaCl2 and CaCO3 89% Th does not leach, remains in residue as ThO2 In use [5]
2) Leach with 3% HCl No Mn problem
Salt leach with (NH4)2SO4 80‐90% Targets physisorbed REE through cation exchange In use [7]
Ion clay Leach with seawater 40% Inefficient but cheap process R&D [8]
Acid leach with strong acid (pH<1) ALL Dissolves entire clay, incurs significant additional costs Not used [7]

II. Leaching technologies in primary REE production
The main REE minerals used in primary REE production are bastnaesite and monazite. Next
to these, ion adsorbed clays are lower grade resources, which due to their easy processing
are becoming more popular as a primary resource. Table 1 offers a concise overview of the
leaching technologies, both past and present, used in primary REE production.
Bastnaesite
Bastnaesite is a rare earth fluorocarbonate mineral1, RE(CO3)F, which primarily contains light
rare earth elements. After physical upgrading, bastnaesite ore concentrates contain between
40 to 60% REE2.
One of the main concerns in past bastnaesite processing technologies (see table 1) was the
inability to extract the REE fluorides. It has been successfully resolved in the current day
technologies in two different ways: pre/post‐treatment with alkaline or sulphuric acid
roasting. The alkaline treatment, developed by Kruesi and Duker of Molycorp3, is a three step
process, using the following reactions:
9 → 2 3 3 3 (1)
3 → 3 (2)
3 → 3 (3)
Step I ‐ a mild HCl solution to extract the REE carbonate, step II ‐ reaction with NaOH at 96°C
to convert the REE fluorides to hydroxides, which are then dissolved by leaching with HCl in
step III. Alternative processes exist, which skip the first leaching step and instead the
bastnaesite is first treated with alkaline and then leached with HCl3.
Sulphuric acid roasting is the other main process currently used in market to process
bastnaesite. In the process, bastnaesite concentrate is heated in a 98% H2SO4 solution to 400
to 500°C for several hours. This decomposes the fluorocarbonate matrix of the bastnaesite,
leading to the release of the CO2 and HF gas. The REE are converted to their sulphates and
can be selectively precipitated as Na double sulphates by leaching the residue with NaCl
containing water. The sulphate process is currently in use at the Bayan Obo mine in China,
while the alkaline method was used by Molycorp at the Mountain Pass mine before it closed3.
Monazite
Monazite is a rare earth phosphate mineral, RE(PO4), containing mostly light REE and some
heavy REE (more than bastnaesite1). Monazite can contain up to 70% REE, primarily Ce and
La as well as significant amounts of Nd, Pr and Sm. The Th content is also quite high, ranging
from 4 to 12% which is a concern in monazite processing.
As shown in Table 1, the alkaline method is currently one of the main leaching technologies
for monazite. The reactions during alkaline leaching are:
3 → (4)
12 →3 4
(5)
After the monazite is digested in 60‐70% NaOH at 140‐150°C for 4 hours, the residue is
dissolved into hot acidic solution, i.e. HNO3 for solvent extraction using TBP and H2SO4 for
solvent extraction using amines of the solution. This process requires proper grinding of the

monazite ore prior to treatment (particle size below 45 µm) so that extraction rates of 98%
can be achieved even with relatively low grade ores (Australian monazite 48.6% REE3).
However, Th is leached at the same time, which brings radioactive concern of this process.
Another concern is the presence of Mn+IV during alkaline processing, which oxidises Ce and
form CeO2, which will not dissolve in HCl4.
An alternative method has been proposed by Merritt5, in which the monazite ore is heated
with CaCl2 and CaCO3 under a reducing and sulphidizing atmosphere. This leads to the
conversion of REE phosphates to REE oxysulphides (RE2O2S) and oxychlorides (REOCl), while
creating a stable Th oxide and chloropatite (Ca5Cl(PO4)3). From this mixture, the REE can be
selectively leached with 3% HCl. This process has three advantages over the alkaline process:
(1) the conversion step is shorter (45 min vs 3‐4 h in the alkaline digestion), (2) there is no
necessity for fine grinding, and (3) Th is stabilised as ThO2. However the REE recovery rate is
lower than the alkaline method (89% vs 98%) and the Na3PO4 by‐product is not present.
Ion adsorbed clays
Ion adsorbed clays are alumina‐silicate clays on which REE ions have been adsorbed. Although
these clays have an average REE concentration of only 0.05‐0.2 wt%, their ease to process
and relatively high heavy REE fraction make them a viable REE resource1. These clays require
no prior beneficiation process and contain very little radioactive elements, a constant concern
with monazite processing.
As given in Table 1, salt or low concentration acidic leaching of these clays to recover REEs is
most frequently applied7. (NH4)2SO4 and NaCl are commonly used as the leachants and the
leaching reaction is as follows7:
• RE 3 → • NH (6)
The kinetics of the leaching process is very fast, equilibrium is achieved in around 10 min and
the total REE extraction is between 80‐90%. There have also been leaching trials in
Madagascar using seawater as a leachant8. However, these trials yielded recovery rates of
around 40%, vastly inferior to the 80‐90% achieved with (NH4)2SO4. The industrial process
currently used in China uses an ion clay with an REO concentration between 0.08 and 0.8 wt%
and a leachant of 7% NaCl and 1‐2% (NH4)2SO4 at a pH of 4. A recovery rate of up to 95% REO
is achieved9.
III. Leaching technologies in new and upcoming REE resources
REE recovery in the phosphoric acid industry
The main resource for phosphorous in the phosphoric acid industry is the apatite mineral.
This mineral, Ca5(PO4)3(Cl, F, OH), is known to contain 0.1 to 1% REE10. The REE in apatite are
present as either REE3+ ions substituted on the Ca2+ ion sites of the apatite lattice (balanced
with Na+ ions) or as REE mineral inclusions, e.g. monazite inclusions. As such many of the
phosphoric acid producers are seeking to extract the REE as a side stream of their process.
The REE are most commonly found in the fluorine variant of the apatite mineral, as such
Ca5(PO4)3F will be used as the representative formula in this review.
The main reaction in the phosphoric acid production is as follows11:
5 → 3 5 ∙ ↓ ↑ (7)

where x=0 for calcium sulphate anhydrate, x=0.5 for hemihydrate and x=2 for dihydrate (most
common). As this completely dissolves the apatite, all REE are transferred into the solution.
However, 80% of the REE in solution is lost, due to the precipitation of the insoluble CaSO4
(gypsum). Considering the amount of gypsum formed under typical processing conditions
where 5 tonnes of gypsum are formed per tonne of P2O512, it has promoted several different
approaches by phosphoric acid producers to recover the REE.
The first approach was to ignore the REE losses to the gypsum and focus on the REE remaining
in solution. This method focusses on the purification of the crude P2O5 (27%) to the
commercial grade (54%)13. During this process a sludge of CaSO4•½H2O (hemihydrate) forms,
which contains the REE. This sludge is leached with HNO3 with a leaching efficiency of around
80%. However, this process is inherently flawed for the production of REE, as most of the REE
are lost to the gypsum by‐product in the first step. This leads to the second approach, the
hemihydrate process12, where the process parameters were adapted in order to minimise the
REE losses. Since the REE could be leached from the hemihydrate, the process was adapted
such that instead of gypsum the hemihydrate is formed during the digestion of the apatite. It
was also found that when precipitating the hemihydrate almost all REE are incorporated in
the crystals, instead of 80% in the case of gypsum11. The hemihydrate is filtered and then
leached with diluted H2SO4. This dissolves the hemihydrate and brings the REE into solution.
The REE in solution inhibit the re‐precipitation of gypsum, allowing for them to be removed
through solvent extraction12.
Some of the phosphoric acid producers have opted to completely redesign their process in
order to make REE recovery easier. Instead of dissolving the apatite using H2SO4, it is dissolved
with HNO314, with Ca(NO3)2 as a by‐product. The advantage of this approach is that the
solubility of Ca(NO3)2 can easily be controlled, allowing for the REE to be removed from the
solution before it is co‐precipitated. Also compared to gypsum, Ca(NO3)2 is a marketable
product for the fertilizer industry.
Parallel to these developments to extract the REE during the production of phosphoric acid,
there have been attempts to process the copious amount of REE containing gypsum already
produced by the industry worldwide. In some countries (like Poland15) the dumped gypsum
represents the largest national REE resource. The most basic process was by leaching the
gypsum with 0.5‐1M H2SO4 at room temperature10. This process leaches about 50% of the
REE from the gypsum without destroying the gypsum crystal structure. This process was
partially improved by mechanical activation through ball milling the gypsum before
leaching16. An alternative process uses the following reaction10:
→ (8)
All REE are incorporated in CaCO3, which is easily leached with HNO3, also producing useful
Ca(NO3)2. Alternatively the CaCO3 can be calcined to CaO and leached with (NH4)Cl leaving an
REE rich residue.
Recycling of lamp phosphor
Fluorescent lamp phosphors are a valuable REE resource, especially for Y, Eu and Tb. The main
REE compounds in these lamps are: Y2O3:Eu3+ (YOX), LaPO4:Ce3+,Tb3+ (LAP),

(Gd,Mg)B5O12:Ce3+,Tb3+ (CBT), (Ce,Tb)MgAl11O19 (CAT) and BaMgAl10O17: Eu3+ (BAM). Some
also contain chlorapatite ((Sr,Ca,Ba,Mg)5(PO4)3Cl:Eu3+) and halophosphate
(Sr,Ca)10(PO4)(Cl,F)2.
Among these REE compounds, the leaching of REE from YOX was found to be the easiest17, as
they dissolve in relatively diluted acids (0.5M H2SO4). In comparison, the REE in the other
phosphors can only be leached at a sulphuric acid concentration of 18M17. This is due to the
fact that the REE in YOX are present as oxides, while the REE in the other phosphors have
much stronger chemical bonds. It was found that for these compounds the same leaching
processes used in monazite processing were relatively effective18. Alternatively a leachant of
4M HCl with H2O2 also proved moderately effective18.
A typical leaching process of REEs from waste phosphor contains three stages. The ground
phosphors are leached with 1.5M H2SO4 to dissolve Y and Eu. As it also dissolves some
impurities (e.g. Ca, P, Mn, etc.), aqueous ammonia is added to the second stage. This keeps
the impurities in the solution while converting the remaining undissolved REE into
hydroxides, which precipitate out. The precipitates are leached with HCl in the third stage19.
A different approach is the process developed by OSRAM A.G. with a patent in 201120. In this
process, the multistep leaching targets specific compounds in the phosphors: 1) leaching with
diluted HCl below 30°C leaches only the halophosphates; 2) increasing temperature to 60‐
90°C, the diluted HCl leaches YOX (alternatively dilute H2SO4 can also be used); 3) LAP is then
dissolved with concentrated H2SO4 above 120°C (but below 230°C); 4) CAT and BAM are
dissolved in 30% NaOH at 150°C in autoclave or in molten alkali. For acids, H2SO4 is preferred
as it dissolves fewer impurities (Ca and Sr) compared to HCl or HNO317. Also ultra sound
increases the efficiency, regardless of the leachant21.
Recycling of magnet scrap
There are two major REE to be recovered from REE magnets: Nd from NdFeB magnets and
Sm from SmCo magnets. Most recycling efforts of magnets are currently focused on the
production waste (so called swarf). As such the input streams for magnet leaching are
relatively pure. The leaching of SmCo swarf is relatively easy. It can completely dissolve in 3M
HCl, HNO3 or H2SO419. Processes for SmCo leaching have not been further developed as the
magnets have fallen out of favour with the rise of the NdFeB magnets.
For the NdFeB magnets two different leaching routes are established: a total leaching route
and a selective leaching route. In the total leaching route the whole swarf is dissolved with
the aim to separate the Nd afterwards. Similar to the SmCo swarf, this is relatively
straightforward as NdFeB easily dissolves in the three mineral acids. The choice of acid is
mostly dependent on the subsequent separation process: H2SO4 for selective precipitation22
and HCl for solvent extraction22. HNO3 is avoided since it produces nitrated waste water22.
Note that the solubility of rare earth elements decreases with increasing temperature,
leading to lower leaching efficiencies with increasing temperature. As the leaching efficiency
is more important than leaching rate in magnet recycling, room temperature is preferred22.

In the selective leaching process,
Nd is extracted from the magnets
without dissolving Fe and B. This is
achieved through roasting and
leaching. The roasting is based on
the Pourbaix diagram (fig. 1).
A joint stability region of Nd3+ and
Fe2O3 within the pH range from
around 1 to 7 in the diagram can
be observed. Using roasting
condition of 6 h at 900°C in air and
followed by a 0.02M HCl leaching,
an extraction rate of 99% of Nd
and less than 0.5% of Fe was
achieved19, while the same
Figure 1: Pourbaix diagram Fe‐H2O and Nd‐H2O
process without the roasting
system19
leached over 50% of the Fe.

IV. Recent progress and new leaching technologies for REE leaching
Progress in bastnaesite leaching
The environmental pollution caused by fluorine emissions during H2SO4 roasting of
bastnaesite processing in China is becoming an increasing concern23. Due to this, processes
have been developed to prevent the emission of fluorine. A first method revolves around only
leaching the carbonate REE while leaving the REE fluorides in the residue. This is achieved by
thermally activating the ore (400°C for 3 h) and then leaching it with HCl. The thermal
activation enables the leaching of the carbonates at conditions in which the fluorides are
unaffected. The reported leaching efficiency of this process is 94.6% for the carbonates and
0.07% for the fluorides23.
A different method involves progress in the air roasting process mentioned in section II. This
process had already been proven ineffective in leaching the fluoride components, but the
oxidation of Ce+III to Ce+IV prevented Ce from being leached together with the other REE and
led to purification issues. The addition of thiourea24 offered a solution to this problem.
Thiourea prevents the oxidation of Ce, keeping it trivalent after roasting and thus allowing it
to be leached with HCl together with other REE. This allowed for the recovery of Ce and the
non‐fluoride bonded REE. These methods are inherently wasteful however, since not all REE
present in the bastnaesite are leached.
Another more efficient process involves the mechanochemical activation of bastnaesite by
milling it with NaOH powder25. In this process the bastnaesite concentrate is milled together
with NaOH powder, followed by washing with water to remove the Na compounds and then
leaching with HCl. These steps are performed at room temperature and can lead to a leaching
efficiency of around 90%.

Bioleaching
The field of bioleaching is being explored for REE extraction from low grade sources. The REE
concentration of the sources is often below the 1% level, mostly around 0.5%. These sources
include old mine tailings and ion adsorbed clays. As total leaching of these sources, using
strong mineral acids and/or bases, leads to large amounts of waste and is very inefficient,
alternative low cost routes are being explored. In Egypt the possibility of using
‘Acidithiobacillus ferrooxidans’ (a bacteria often used in the bioleaching of copper26) to
bioleach low grade Gibbsite ore to recover the REE (0.49%) and U (0.05%) has been
investigated27. Their initial results show a leaching efficiency of about 55% for REE and 49%
for U. Other bacteria that are tested are ‘Aspergillus ficuum’ and ‘Pseudomonas aeruginosa’28
leading to slightly higher leaching efficiencies for REE’s, around 75%.
Researchers in Japan have investigated the use of a blue‐green algae named ‘phormidium’ in
combination with (NH4)2SO4 for the extraction of REE from ion adsorbed clays29. In this
process the ammonium ions displace the adsorbed REE3+ ions in the clays bringing them into
solution, as well as several other adsorbed ions, mainly Al, Mn and Si. This solution is then
brought into contact with dried and ground algae, which selectively adsorb the REE ions.
Leaching efficiencies between 40% (Dy, Gd) and 70% (Nd, Sm), dependent on the REE species,
could be obtained. The advantage of this process is that temperature and pH control is easy
and the leaching time is relatively short compared to other bioleaches (3 h vs several days).
Also, in Japan ‘phormidium’ is easily obtained as it needs to be removed from the local reefs
to preserve them. The waste of this process is also minimal and easily detoxified.
Microwave assisted leaching
Microwave assisted leaching is frequently used for improving mineral leaching efficiency30.
For REE extraction however no concrete results have been found yet. The principle of
microwave assisted leaching is based on the fact that metal‐containing minerals are less
transparent for microwaves than gangue minerals such as CaO, CaCO3 and SiO230. This leads
to on‐site heating at and around the metal‐containing mineral, thus locally changing the
leaching kinetics. As the leaching kinetics generally increase with increasing temperature this
would lead to increased leaching rates at the metal‐containing minerals, allowing the leaching
to be finished sooner and with less unwanted dissolved species originating from the gangue.
Another effect of the localised and rapid heating is that it can fracture the metal‐containing
mineral surfaces due to thermal stresses, thereby effectively increasing the surface area31.
While microwave leaching has not yet been applied to REE leaching, it has been applied to Cu
leaching from chalcopyrite and to Au leaching31. The effects observed there, e.g. removing
reaction product from the surface through convective streams (Cu) or activating finely
distribute metal‐containing areas (Au), can be beneficial for REE leaching. Considering these
current applications there is potential in applying microwave heating to REE leaching,
especially for low grade sources where the REE are finely distributed in the material,
sometimes in hard to leach compounds (e.g. phosphates), such as in old mine tailings.


V. Conclusions
This brief review shows that there is a variety of leaching technologies that have been
developed in the past years, both for primary REE minerals and secondary resources. These
in‐market technologies were developed according to the mineralogy, REEs occurrence and
engineering feasibility. The main features are that these are all multi‐step processes and that
both acid and alkaline are often interactively used in a single process. In view of the increasing
importance of secondary REE resources, this overview of current state‐of‐the‐art
technologies can provide a basis for developing more efficient processes for REE recovery
from secondary resources, i.e. mine tailings and magnet scraps. For example, old mine tailings
often contain a sizeable fraction of apatite and/or monazite. Drawing inspiration from current
monazite processing and phosphoric acid production can lead to a workable process for REE
recycling from these tailings. As for WEEE, many individual elements (e.g. lamp phosphors
and magnet swarf) of WEEE have been looked at for REE recycling, but not a general WEEE
stream. However, knowledge of these individual components will be invaluable to develop a
process for a mixed WEEE stream. The main challenges will be overcoming the dilution effect
of REE, both in the tailings and in the mixed WEEE, and the variety of contaminants. Here
advancements in microwave leaching or bioleaching show promise.
Acknowledgements
The authors acknowledge the EU project FP7 REEcover (Project ID: 603564) for financial
support.
References
1. [1] Y. Kanazawa and M. Kamitani, “Rare earth minerals and resources in the world,” Journal of Alloys and
Compounds, vol. 408–412, pp. 1339–1343, Feb. 2006.
2. [2] Z. FENG, X. HUANG, H. LIU, M. WANG, Z. LONG, Y. YU, and C. WANG, “Study on preparation and
application of novel saponification agent for organic phase of rare earths extraction,” J. Rare Earths, vol.
30, no. 9, pp. 903–908, Sep. 2012.
3. [3] N. Krishnamurthy and C. K. Gupta, Extractive metallurgy of rare earths. CRC press, 2004.
4. [4] V. I. Kuzmin, G. L. Pashkov, V. G. Lomaev, E. N. Voskresenskaya, and V. N. Kuzmina, “Combined
approaches for comprehensive processing of rare earth metal ores,” Hydrometallurgy, vol. 129–130, pp.
1–6, Nov. 2012.
5. [5] R. R. Merritt, “High temperature methods for processing monazite: I. Reaction with calcium chloride
and calcium carbonate,” J. Common Met., vol. 166, no. 2, pp. 197–210, Nov. 1990.
6. [6] G. A. Moldoveanu and V. G. Papangelakis, “Recovery of rare earth elements adsorbed on clay minerals:
I. Desorption mechanism,” Hydrometallurgy , vol. 117–118, pp. 71–78, Apr. 2012.
7. [7] G. A. Moldoveanu and V. G. Papangelakis, “Recovery of rare earth elements adsorbed on clay minerals:
II. Leaching with ammonium sulfate,” Hydrometallurgy , vol. 131–132, pp. 158–166, 2013.
8. [8] G. A. Moldoveanu and V. G. Papangelakis, University of Toronto ‐ Tantalus Report_January 2013.pdf.
9. [9] Shi, Fu., “Rare earth extraction technologies.,” Beijing, Metallurgical industry publishing, 2009. p 55‐
57.
10. [10] F. Habashi, “The recovery of the Lanthanides from phosphate rock,” J Chem Tech Biotechnol,
vol. 35A, pp. 5–14, 1985.
11. [11] L. Wang, Z. Long, X. Huang, Y. Yu, D. Cui, and G. Zhang, “Recovery of rare earths from wet‐
process phosphoric acid,” Hydrometallurgy, vol. 101, no. 1–2, pp. 41–47, Feb. 2010.

12. [12] S. Zielinski, A. Szczepanik, M. Buca, and M. Kunecki, “Recovery of lanthanides from Kola apatite
in phosphoric acid manufacture,” J. Chem. Tech. Biotechnol., vol. 56, no. 4, pp. 355–360, 1993.
13. [13] J. S. Preston, P. M. Cole, W. M. Craig, and A. M. Feather, “The recovery of rare earth oxides
from a phosphoric acid by‐product. Part 1: Leaching of rare earth values and recovery of a mixed rare earth
oxide by solvent extraction,” Hydrometallurgy , vol. 41, no. 1, pp. 1–19, 1996.
14. [14] H. Li, F. Guo, Z. Zhang, D. Li, and Z. Wang, “A new hydrometallurgical process for extracting
rare earths from apatite using solvent extraction with P350,” Journal of Alloys and Compounds , vol. 408–
412, pp. 995–998, Feb. 2006.
15. [15] A. Jarosiński, J. Kowalczyk, and C. Mazanek, “Development of the Polish wasteless technology
of apatite phosphogypsum utilization with recovery of rare earths,” J. Alloys Compd., vol. 200, no. 1–2, pp.
147–150, Oct. 1993.
16. [16] D. Todorovsky, A. Terziev, and M. Milanova, “Influence of mechanoactivation on rare earths
leaching from phosphogypsum,” Hydrometallurgy, vol. 45, no. 1–2, pp. 13–19, May 1997.
17. [17] F. Yang, F. Kubota, Y. Baba, N. Kamiya, and M. Goto, “Selective extraction and recovery of rare
earth metals from phosphor powders in waste fluorescent lamps using an ionic liquid system,” Journal of
Hazardous Materials, vol. 254–255, pp. 79–88, Jun. 2013.
18. [18] K. Binnemans, P. T. Jones, B. Blanpain, T. Van Gerven, Y. Yang, A. Walton, and M. Buchert,
“Recycling of rare earths: a critical review,” Journal of Cleaner Production, vol. 51, pp. 1–22, Jul. 2013.
19. [19] M. Tanaka, T. Oki, K. Koyama, H. Narita, and T. Oishi, “Chapter 255 ‐ Recycling of Rare Earths
from Scrap,” in Handbook on the Physics and Chemistry of Rare Earths, vol. Volume 43, Jean‐Claude G.
Bünzli and Vitalij K. Pecharsky, Ed. Elsevier, 2013, pp. 159–211.
20. [20] R. Otto and A. Wojtalewicz‐Kasprzak, “Method for recovery of rare earths from fluorescent
lamps,” US20120027651 A1, 02‐Feb‐2012.
21. [21] C. Tunsu, C. Ekberg, and T. Retegan, “Characterization and leaching of real fluorescent lamp
waste for the recovery of rare earth metals and mercury,” Hydrometallurgy, vol. 144–145, pp. 91–98, Apr.
2014.
22. [22] C.‐H. Lee, Y.‐J. Chen, C.‐H. Liao, S. R. Popuri, S.‐L. Tsai, and C.‐E. Hung, “Selective Leaching
Process for Neodymium Recovery from Scrap Nd‐Fe‐B Magnet,” Metallurgical and Materials Transactions
A, vol. 44, no. 13, pp. 5825–5833, Sep. 2013.
23. [23] X. Bian, S. H. Yin, F. Y. Zhang, W. Y. Wu, and G. F. Tu, “Study on Leaching Process of Activation
Bastnaesite by HCl Solution,” Adv. Mater. Res., vol. 233–235, pp. 1406–1410, May 2011.
24. [24] A. Yörükoğlu, A. Obut, and İ. Girgin, “Effect of thiourea on sulphuric acid leaching of
bastnaesite,” Hydrometallurgy, vol. 68, no. 1–3, pp. 195–202, Feb. 2003.
25. [25] Q. Zhang and F. Saito, “Non‐thermal process for extracting rare earths from bastnaesite by
means of mechanochemical treatment,” Hydrometallurgy, vol. 47, no. 2–3, pp. 231–241, Jan. 1998.
26. [26] K. Nowaczyk, A. Juszczak, F. Domka, and J. Siepak, “The use of Thiobacillus ferrooxidans
bacteria in the process of chalcopyrite leaching,” Pol. J. Environ. Stud., vol. 7, pp. 307–312, 1998.
27. [27] H. A. Ibrahim and E. M. El‐Sheikh, “Bioleaching Treatment of Abu Zeneima Uraniferous Gibbsite
Ore Material for Recovering U, REEs, Al and Zn,” Res. J. Chem. Sci., vol. 20111, no. Vol. 1, pp. 55–66, Jul.
2011.
28. [28] W. A. HASSANIEN, O. A. DESOUKY, and S. S. HUSSIEN, “Bioleaching of some Rare Earth
Elements from Egyptian Monazite using Aspergillus ficuum and Pseudomonas aeruginosa.” [Online].
Available: http://wjst.wu.ac.th/index.php/wjst/article/view/481.
29. [29] J.‐A. Kim, G. Dodbiba, Y. Tanimura, K. Mitsuhashi, N. Fukuda, K. Okaya, S. Matsuo, and T. Fujita,
“Leaching of rare‐earth elements and their adsorption by using blue‐green algae,” Materials Transactions,
vol. 52, no. 9, pp. 1799–1806, 2011.
30. [30] K. E. Haque, “Microwave energy for mineral treatment processes—a brief review,” Int. J.
Miner. Process., vol. 57, no. 1, pp. 1–24, Jul. 1999.
31. [31] M. Al‐Harahsheh and S. W. Kingman, “Microwave‐assisted leaching—a review,”
Hydrometallurgy, vol. 73, no. 3, pp. 189–203, Jun. 2004.

Extraction of lanthanides from spent polishing agent
Alexander Poscher1, Stefan Luidold2, Holger Schnideritsch3 and Helmut
Antrekowitsch4
1,2,3,4
Chair of Nonferrous Metallurgy, University of Leoben, Franz‐Josef‐Str. 18, 8700 Leoben, Austria
alexander.poscher@unileoben.ac.at, stefan.luidold@unileoben.ac.at,
holger.schnideritsch@unileoben.ac.at , helmut.antrekowitsch@unileoben.ac.at
Abstract
The polishing of glass substrates or wafers requires the usage of fine grained particles based
on lanthanides. During processing the sludge becomes enriched with different elements
depending on the specific production program. However, concentrations of impurities can
reach critical values and may influence the surface quality of the substrates in a negative way.
Therefore, the polishing agents have to be removed from the system and recycled
meaningfully. The paper describes ways of processing such secondary RE resources using
different hydrometallurgical methods including mineral acids for the extraction. Finally, the
focus was set on the evaluation of precipitation methods for the production of a rare earth
concentrate.
Introduction
Polishing agents and additives used in the glass industry are estimated to represent a global
demand of rare earths (RE) of approximately 10 % in terms of economic value but mainly
consist of light rare earths cerium and lanthanum 1,2,3. These most abundant RE have suffered
from continuously falling prices (Figure 1) in the last years which makes economic recycling
difficult.

Figure 1: Relative price development of various REO from 2012 to 2014 4

Based on the price level in January 2012 which was set to one, the changes in price are
visualized as dimensionless values (€xx/201x/€01/2012). As can be seen from the chart the costs
for many of the REOs are less than 40 % of the price at the beginning of 2012. Therefore, it
seems to be of high relevance to design easy manageable, environmentally friendly and low‐
cost recycling processes. In contrast the strategic independence of the European Union and
many other countries outside Europe from China´s RE supply is a strong driving force for
numerous recycling projects. Kato 5 suggested an alkaline treatment of spent polishing
material using 4 mol/kg sodium hydroxide solution and temperatures between 50 to 60 °C.
Thus, impurities such as silicon dioxide and alumina dissolve and can be separated from the
solution by heating and precipitation as zeolites 5. Matsui 6 also treated abrasives with
solutions of sodium hydroxide to achieve dissolution of for example SiO2. Further processing
includes a precipitation respectively a promoted settling of fine disperse cerium oxide
particles with for example aluminium sulphate. Subsequent to the removal of the
supernatant the precipitates are washed several times. Additionally, the pH of the resulting
fraction is adjusted to low acidic or neutral in order to remove soluble impurities 6. An
advanced process is proposed by Moon 7 also dealing with the removal of contaminants from
a spent polishing agent by a caustic soda solution. Small amounts of sodium fluoride and
sodium carbonate support the purification process 7.
A dry process was introduced by Ozaki 8 using carbo‐chlorination techniques to achieve
extraction of the rare earths from a material basically consisting of 38 wt.‐% CeO2, 28 wt.‐%
La2O3 and 10 wt.‐% Nd2O3 together with alkaline earth metal oxides, silicon dioxide and
fluorine. Subsequent chemical vapour deposition experiments using Al2Cl6 as complexing
agent showed that specific temperature gradients may lead to a separation of rare earth
chlorides from formed aluminium and iron chlorides. Enrichment of single rare earth
chlorides can be obtained by a repeated CVT process 8.
Nevertheless, processes using solutions of sodium hydroxide face problems concerning the
quantitative dissolution of silicon dioxide and metallic components which originate from
semiconductor polishing. The application of chlorine in the process may result in the
separation of the rare earths, but careful material selection of the reactor lining has to be
done due to the highly corrosive media. Therefore, acidic hydrometallurgical methods seem
to be favourable. One available technology is used at Hydrometal in Belgium 9. The rare earth
compounds of spent polishing agents are dissolved by nitric acid in combination with
hydrogen peroxide acting as a reducing agent for Ce(IV) in this case. The addition of oxalic
acid leads to the precipitation of the RE oxalates 9. As it is shown later, our own experiments
confirm the possibility of good cerium extraction by these chemicals, but also illustrated
limited lanthanum dissolution.

Experimental
Input material
The material of interest was a typical polishing material based on lanthanides with an
approximate composition according to Table 1. The major impurities which make a recycling
necessary are the oxides of silicon, zinc, sodium and calcium because they may cause
decreasing surface qualities of the glass substrate at critical values.

Table 1: Chemical composition of the used abrasive material

Oxide wt.‐% Oxide wt.‐%
CeO2 69 SiO2 11
La2O3 8 ZnO 2
Nd2O3 1 P2O5 2
CaO 1 Na2O 1
Experimental set‐up
All leaching and extraction experiments were carried out at atmospheric pressure in three‐
neck round‐bottom flasks using a heating mantle and temperature control. Furthermore,
condensers and dropping funnels were employed. The experimental set‐up for the leaching
tests can be seen in Figure 2. Nitric and hydrochloric acid tests were carried out under
continuous agitation. Sulphuric acid experiments show onward solidification of the sludge
consisting of the polishing agent and the concentrated acid at raising temperatures.

Figure 2: Experimental set‐up for the extraction experiments (Figure drawn with
ACD/ChemSketch)


Precipitation experiments were performed from hydrochloric and sulphuric acid solutions.
These were carried out in beakers at room temperature under continuous stirring. Frits of
porosity 4 (10‐16 µm mesh) separated the leaching residues and precipitation products from
the liquid phase. Both solid fractions were dried at 120 °C and calcined if it was assumed that
the obtained compounds differ from oxides.
Nitric acid leaching
The experiments included a number of leaching experiments to determine the most
important parameters. As nitric acid ‐ an oxidizing acid ‐ was used, hydrogen peroxide had to
be added in order to reduce tetravalent cerium from CeO2 to the cerous state. The
experiments started with the addition of nitric acid to 50 g of spent polishing material. After
a heating period 100 g H2O2 (30 wt.‐%) were added. Different amounts and concentrations of
nitric acid were tested and the temperature and leaching time were varied (Table 2).
Subsequently, filtration and washing of the filter cakes were done. Additionally, the solid
residue was dried, the dissolved mass measured and the material was analysed by XRF.

Table 2: Basic parameters for HNO3 experiments

Concentrations Additional Temperatures tested Leaching
Acid
tested [wt.‐%] chemicals during leaching [°C] time [h]
HNO3 47, 65 H2O2 70, 90 1, 6, 12
Hydrochloric acid leaching
Hydrochloric acid leaching tests started in the same way as the nitric acid leaching
experiments. As HCl is a reducing acid, the strength of the chlorine ion was assumed to be
strong enough for complete reduction of Ce4+. Therefore, the amount of hydrogen peroxide
was varied as well. An overview of tested parameters in the leaching step is given in Table 3.

Table 3: Basic parameters for HCl leaching experiments

Concentrations Additional Temperatures Leaching
Acid
tested [wt.‐%] chemicals tested [°C] time [h]
HCl 16, 21, 32 H2O2 80, 95 4, 8

The pregnant leach solutions were adjusted to a volume of 500 ml. Subsequently,
precipitation experiments using the total amount of the leaching liquor and oxalic acid were
done. This step was supported by Design of Experiments which makes a statistical analysis
possible. While the leaching parameters were held on a constant level (Table 4) the amounts
and concentrations of oxalic acid solutions were modified. After the conversion of the
precipitation product to a mixed rare earth oxide by thermal treatment at 750 °C and the XRF
analysis of the product, mass balances were calculated. The yields for each element represent

the raw data for the statistical analysis. As a consequence of the constant leaching
parameters, the result of the statistical analysis gives accurate data about the precipitation
step.

Table 4: Leaching parameters for subsequent precipitation

Pol. Agent [g] c(acid) [wt.‐%] H2O2 [g] Temperature [°C] Leaching time [h]
50 32 100 80 4

A flowsheet of the entire HCl leaching and precipitation process can be found in Figure 3.
Figure 3: Process flowsheet of hydrochloric acid leaching and oxalic acid precipitation
Test execution for experiments using sulphuric acid
The experimental procedure and set‐up correspond to the above described in many ways.
One major difference between nitric or hydrochloric leaching experiments from the sulphuric
acid process is the extraction of lanthanides in two steps including a digestion and subsequent
elution of formed RE sulphates. The digestion was performed in round‐bottom flasks as it was
described earlier. During this process at temperatures between 100‐200 °C the slurry
consisting of polishing agent and concentrated sulphuric acid suffers from onward
solidification. Therefore, contrary to HNO3 and HCl experiments stirring during the entire
digestion period became impossible. The digested materials were subsequently leached at
room temperature with sufficient amount of water and transferred to frits. The filter cakes
were washed several times, dried, calcined and finally analysed.

The pregnant red‐orange leaching solutions were subsequently transferred to beakers and
treated with different precipitation agents to achieve mixed RE‐compounds. The universally
valid process flowsheet can be seen in Figure 4.

Figure 4: Sulphuric acid digestion and production of a mixed REO concentrate
Results
Nitric acid leaching experiments
Nitric acid leaching experiments showed the possibility of cerium dissolution up to 90 %.
Nevertheless, it could be found that lanthanum extraction is insufficient and limited to 60‐70
%. The investigations also revealed the results from Figure 5a and 5b. Decreasing solid/liquid
ratios induced increasing reduction of the initial mass of the input material which correlates
to the extraction of cerium and lanthanum and therefore higher extraction rates can be
achieved. Figure 5b shows the effect of extended leaching periods.


Figure 5: Influence of solid/liquid ratio and duration on the dissolution of the input material

Another comparison between the experiments visualises the effect of a change in leaching
temperature. An increase in temperature from 70 to 90 °C does not lead to intensified
dissolution. Moreover, an increase in acid concentration shows no effect on the extraction
(Figure 6).

Figure 6: Influence of temperature and acid concentration on the dissolution of the input
material
Hydrochloric acid leaching and precipitation experiments
Some obtained results from nitric acid leaching experiments coincide with data from HCl
leaching. On the one hand, good extraction of cerium is possible under optimum leaching
conditions. On the other hand the extraction of lanthanum is limited to 60‐75 % as well. XRD
analysis of the leaching residues revealed that mainly fluoride and phosphate compounds of
lanthanum remain undissolved. Further investigations of the leaching step showed
decreasing influence of H2O2 at larger excess and concentration of HCl. 50 % excess of acid
equals a mass of 43.75 grams of HCl whereas the amount of added solutions varied depending
on their concentration. As a result 50 % excess required 136.73 grams of 32 wt.‐% HCl.

Figure 7: Ce extraction as a function of acid concentration and amount of H2O2

Precipitation experiments required leaching operations at constant parameters. Therefore,
these tests were performed using solutions won by leaching of 50 g of polishing agent with
273.45 g HCl (32 wt.‐%) which represents an excess of 200 % and 100 g H2O2 (30 wt.‐%). After
filtration the solutions were adjusted to volumes of 500 ml and transferred to oxalic acid
processing. 30 to 50 g of oxalic acid were used in this process in concentrations of 50, 62.5
and 80 g/l H2C2O4 corresponding to approximate excesses of the chemical of 0 to 66 %. The
calculation took into account that on average 97 % of cerium, 70 % of lanthanum and 54 % of
zinc present in the raw material can be extracted under the applied leaching conditions and
form oxalates. Later the XRF analysis of the precipitation products showed that zinc remains
in the solution. Subsequent to the above described thermal treatment and XRF analysis of the
obtained mixed RE‐oxide the elemental yields were determined. The result of the statistical
analysis is shown in Figure 8.


Figure 8: Prediction contour plots for cerium and lanthanum yields [10]

As can be seen from the plots above the excess of oxalic acid added to the leach liquor
strongly affects the yields of Ce and La whereas there are some statistical uncertainties
concerning the influence of the concentration. Nevertheless, increasing acid concentrations
tend to support the formation of Ce oxalates, whereas the isolation of lanthanum decreases.
The low yields for lanthanum are caused by the above mentioned limitation of the extraction.
A representative chemical composition of a final product won by the described leaching and
precipitation process can be found in Table 5.

Table 5: Representative XRF analysis of an oxide product

Oxide wt.‐%
CeO2 90.73
La2O3 8.57
ZrO2 0.08
Fe2O3 0.074
CaO 0.07
SiO2 0.04
Sulphuric acid leaching and precipitation experiments
The experiments revealed that sufficient sulphuric acid and temperature must be applied to
the process for quantitative digestion of the spent abrasive. Furthermore, complete elution
of formed sulphates requires enough water and repeated washing of the filter cakes.
Optimum digestion and leaching conditions led to extraction rates of almost 99 % for Ce and
98 % for La. Figure 9 shows the filtration and washing of a digested polishing agent. However,
the solutions won by this process show the disadvantage of low stability and precipitation of
lanthanide salts within a couple of hours. Therefore, the further treatment with precipitation
agents had to be done quickly. The products won by the application of commercial
precipitation agents met the specifications for additional processing.

Figure 9: Digestion and elution of lanthanide compounds
Conclusions
Three ways of processing polishing materials were tested. The focus was mainly set on the
extraction of lanthanides from the raw material, whereas all mineral acids of technical
relevance were taken into account. Nitric and hydrochloric acid experiments showed strongly
limited dissolution of lanthanum compounds. The extraction of cerium as most of it was
present in its tetravalent stage necessarily requires a reducing agent in the case of HNO3
leaching. Hydrogen peroxide was chosen as reducing agent because the resulting leach liquor
is thereby not contaminated by foreign ions. Under certain conditions almost complete
cerium extraction could be achieved by hydrochloric acid leaching without the use of H2O2.
Especially high HCl concentrations affect the dissolution of cerium in a positive way. Oxalic
acid as a well‐known precipitation agent for rare earth elements was chosen because of its
advantage to form precipitation products at low pH values. Therefore, the contamination of
the obtained products from hydrochloric solutions could be reduced. Additional experiments
using concentrated sulphuric acid pointed out higher lanthanum extraction than was
achieved for the previously described tests. However, digestion of the polishing material and
subsequent leaching of formed water‐soluble cerium and rare earth sulphates require
significantly higher temperatures (up to 200 °C) and larger amounts of washing water
respectively.
Acknowledgements
The authors want to thank the company D. Swarovski KG, the Austrian Research Promotion
Agency (FFG) and the Federal Ministry of Science, Research and Economy (bmwfw) for the
financial support of this work.

References
1. D. Schüler, M. Buchert, R. Liu, S. Dittrich and C. Merz, “Study on Rare Earths and Their Recycling“, Final
Report for The Greens/EFA Group in the European Parliament (2011)
2. N. Curtis, “Rare Earths ‐ We can touch them every day”, J.P. Morgan Australia Corporate Access Days, New
York (2010)
3. K. Binnemans, P.T. Jones, B. Blanpain, T.V. Gerven, Y. Yang, A. Walton and M. Buchert, “Recycling of rare
earths: a critical review” in Journal of Cleaner Production, 51 1‐22 (2013)
4. Deutsche Rohstoffagentur, Internet: www.deutsche‐rohstoffagentur.de
5. K. Kato, T. Yoshioka and A. Okuwaki, “Study for Recycling of Ceria‐Based Glass Polishing Powder”, in Ind.
Eng. Chem. Res., 39 943‐947 (2000)
6. H. Matsui, D. Harada and M. Takeuchi, “Method for recovery of cerium oxide”, Patent US2013/0152483
7. W.‐J. Moon, S.‐O. Na and H.‐Y. Oh, “Method for recycling cerium oxide abrasive”, Patent US2011/0219704
8. T. Ozaki, K.‐I. Machida and G.‐Y. Adachi, “Extraction and Mutual Separation of Rare Earths from Used
Polishes by Chemical Vapor Transport”, in Met. Mat. Trans. B, 30B 45‐51 (1999)
9. P. Henry, S. Lamotte, J.‐M. Bier, “Recycling of rare earth materials at Hydrometal (Belgium)”, in Proceedings
of Materials Science & Technology 2013 (MS&T´13), Montréal, Canada, 2013
10. A. Poscher, S. Luidold and H. Antrekowitsch, ” Extraction of cerium and lanthanum from spent glass
polishing agent” in Proceedings of Materials Science & Technology 2013 (MS&T´13), Montréal, Canada,
2013


THE REE POTENTIAL IN “URBAN” ORE DEPOSITS: AN

EVALUATION ON CONTENTS AND PROSPECTING TOOLS
FROM ITALIAN MUNICIPAL SOLID WASTE INCINERATORS.
Valerio FUNARI1, Syed Nadeem Hussain BOKHARI2, Thomas MEISEL2, Luigi
VIGLIOTTI3, Roberto BRAGA1
1
BiGeA Dept. (Geology Division) – University of Bologna, Piazza di Porta San Donato 1, Italy
2
General and Analytical Chemistry – Montanuniversität Leoben, Franz‐Josef‐Str. 18, Austria
3
Institute of Marine Sciences – National Research Council, Via Piero Gobetti 101, Italy

valerio.funari@unibo.it
Abstract
Solid residues from Municipal Solid Waste Incinerators (MSWI) may represent an
unconventional source of REE and other critical raw materials as defined by the European
Commission. In this contribution we present the total REE content in bottom and fly ashes,
which are the common MSWI outputs, and their estimated annual flow are significant.
Samples were digested with sodium peroxide sintering which is highly effective analytical
technique and the analysis was performed with ICP‐MS. Bottom and fly ashes have an average
concentration of 103 and 54 mg/kg ƩREY, respectively. For fly ashes the expected annual flow
is 1.1 t/a LREE and 0.3 t/a HREE, whereas for bottom ashes is 22.5 t/a LREE and 5.2 t/a HREE.
Moreover, we show that simple magnetic susceptibility measurements can be a promising
prospecting method when exploring for REE in urban ores is concerned.
Introduction
The European Commission defined a list of “critical” raw materials on the basis of their
relative economic importance and supply risk [1]. Rare Earth Elements (REE) are among
critical raw materials because of their importance for the new and green technologies, e.g. in
magnets for wind turbines, hybrid car engines, etc [2].
Traditionally, the main sources of REE are natural ore deposits, usually divided into primary
(e.g., igneous rocks, hydrothermal related deposits), secondary deposits (e.g., placers,
lateritic soils) and by‐product productions from other mining operation.
The major REE ore deposits are located in a few countries namely China and Mongolia [1, 2]:
restriction policies on the REE export from these countries may increase the supply risk for
EU countries as occurred during the 2011 crisis. For these reasons there is an increasing effort
to evaluate other REE sources. In a recent work it has been demonstrated that solid residues
from incinerated municipal waste host interesting amounts of critical raw materials [3]. Here
we present the total elemental chemistry of bottom and fly ashes from two Italian
incineration plants and the calculated REE annual flow [4] and magnetic data results.

Materials and Methods
Sampling
Final solid residues from two waste incinerator plants from Northern Italy, named plant A and
B were collected. The selected incinerators have similar incineration systems consisting of
two lines that drive the collected waste, about 120000 t/a, in the furnace that operates at
temperatures between 850 and 1100 °C. More than 90% of the solid waste input is made of
unsorted municipal solid waste.
The main outputs of the incineration process are slags, bottom and fly ashes. Belt conveyors
transport the slags and bottom ashes (hereafter BA) to a temporary outdoor storage site
where the BA are piled up. Directly from the BA storage site, 7‐8 kg primary sample was taken;
the batch was split in four portions, the opposite portions were mixed together and again
split for three times, to ensure homogeneity. Fly ash and APC residues (hereafter, FA) have
been separately sampled at the different devices from the two incinerators. Where it was
possible, FA were collected at the furnace, after the mechanical waggle and after the bag
filter (with soda or lime additives). In both incinerator plants, about 5 kg FA primary sample
was collected with a random sampling method from big bags and blended from a large
number of increments in order to reach representativeness. The overall process ensures that
there is no mixing between FA and BA.
Analytical techniques
The complete sample digestion was obtained with the sodium peroxide sintering mainly of
protocol as in [5] and major and trace elemental analysis was made with ICP‐MS (7500ce
Agilent Technologies). Analyses were performed at the department of General and Analytical
Chemistry of Montanuniversität Leoben (Austria).
Magnetic susceptibility measurement were carried out by using a dual frequency MS2B
Bartington meter available at the Institute of Marine Sciences (CNR, Bologna). Dried and
milled (<40 μm) samples of MSWI residues were laid down and gently compacted in the
sample holder. The exact sample mass was recorded and magnetic susceptibility values on
MS2B display were noted. Mass specific magnetic susceptibility () was calculated dividing ,
i.e. the ratio between the induced magnetisation and external magnetic field, by the sample
mass. , previously corrected for the drift, is expressed in 10‐5 SI units, whereas  in m3kg‐1.
Substance flow analysis
An estimation of substance flow [4] for plant A and B residues was carried out to determine
the REE annual flow within output products. The spatial system boundary for the experiment
was the MSWI process itself, from the waste collection to outputs products, and included the
sampling. The temporal system boundary was defined as 1 year, assuming that the sampled
material is representative for one year process activity. For the calculation of total annual
flow, F in kg/a, for BA and FA we used the following equation:

F= Ci ∙ O (Eq. 1)
where: Ci is the measured concentration of element i in mg/kg and O is the total output mg/a
for the BA and FA.
Statistical data screening
The variation of REE in each sample, both for BA and FA, has been analysed using the box
plots (Fig. 1). The general variability may reflect the heterogeneity of the samples, resulting
from different waste input, treatment steps and combustion condition between the two
incineration facilities. In BA samples the majority of REE have concentration from around 0
mg/kg (25th percentile) to 15 mg/kg (75th percentile) while in FA from 0 (25th percentile) to 5
mg/kg (75th percentile) (Fig. 2a). Fig. 2b shows that light REE (LREE: Sc, La, Ce, Pr, Nd, Sm, Eu,
Gd) and heavy REE (HREE: Y, Tb, Dy, Ho, Er, Tm, Yb, Lu) are higher in BA residues than in FA,
probably due to the wide range of process treatments (e.g. different cooling conditions,
chemical additives).

Boxplot of REY
REY (mg/kg)

Figure 1: The REY variability in the studied samples. The boundaries of the box indicate the
25th percentile and the 75th percentile, the continuous line within the box marks the median
and the red dot marks the mean. Whiskers above and below the box indicate the 90th and
10th percentiles, whereas starred dots are the outliers (mainly Ce).
a) Boxplot of REE b) Boxplot of REE;HREE;LREE
REE (mg/kg)

Figure 2:a)Comparison of REE contents in BA and FA samples between incinerators A and B;
outliers, mainly Ce and La, are shown with starred dots. b) ∑REE, ∑HREE and ∑LREE in BA and
FA samples. For the significance of box‐plots see Fig.1 .

In order to understand the difference of REE concentration between plants A and B
independent sample t‐test was conducted (Tab. 1).
Considering the total of MSWI residues and FA residues the mean contents of REE for the two
incinerators are statistically equal. On the other hand the REE means in the BA residues is
significantly (at 0.05 significance level) different between plant A and B, in spite of the similar
incineration process. This is probably due to a more impacting nugget effect.

Table 1: Comparison of the REE mean content in MSWI residues from plants A and B. Tests
for FA and bulk MSWI (BA+FA) residues have been calculated but not reported.

Bartlett's test for homogeneity of variance t‐Test for equality of the mean
p‐value Mean Mean 95% confidence interval
REE Equal variance k‐squared p‐value t
(two‐tailed) A B Lower Upper
Bottom Ash
Ce Not assumed 1.75 0.19 2.80 0.01 32.93 30.53 0.60 4.19
Dy Assumed 8.34 0.00 7.11 0.00 2.06 1.63 0.30 0.56
Er Not assumed 0.84 0.36 6.12 0.00 1.33 1.07 0.17 0.35
Eu Assumed 3.98 0.05 5.31 0.00 0.97 0.84 0.08 0.17
Gd Assumed 8.22 0.00 4.43 0.00 2.18 1.93 0.13 0.37
Ho Assumed 5.24 0.02 8.76 0.00 0.42 0.32 0.08 0.13
La Not assumed 0.91 0.34 4.68 0.00 19.30 16.13 1.75 4.58
Lu Assumed 16.00 0.00 5.25 0.00 0.18 0.13 0.03 0.06
Nd Not assumed 2.86 0.09 ‐0.93 0.36 13.07 13.68 ‐1.97 0.76
Pr Assumed 6.54 0.01 2.39 0.03 3.57 3.20 0.05 0.68
Sc Assumed 7.55 0.01 3.50 0.00 12.92 9.93 1.20 4.78
Sm Assumed 4.38 0.04 4.91 0.00 2.35 2.05 0.17 0.42
Tb Not assumed 0.69 0.41 2.94 0.01 0.43 0.39 0.01 0.07
Tm Assumed 8.82 0.00 6.38 0.00 0.17 0.13 0.03 0.05
Y Not assumed 2.35 0.13 5.44 0.00 14.97 12.53 1.50 3.38
Yb Assumed 10.90 0.00 6.69 0.00 1.21 0.90 0.21 0.41
[Numbers smaller than 0.05, significance level, are in bold character]
Correlations with major and trace elements
Correlations between ∑REE and selected major and trace elements (Fig. 3), demonstrate that
volatile and soluble components (e.g. Cd and K2O, respectively) do not control the REE
contents, but allow to distinguish BA from FA. The most of volatile elements, e.g., Bi, As, Cd,
Sb, Sn has a similar behaviour. Conversely, major elements like P2O5 (of likely anthropogenic
introduction) and Al2O3 (both geogenic and anthropogenic component) show a positive
correlation with ∑REY.


Figure 3: Correlations of ∑REE and selected major and trace elements. Regression lines by
groups are provided for ∑REE vs. P2O5 and Al2O3.
Correlation with magnetic properties
Heavy metals are known to have affinity to form metallic bonding with ferrous material,
which leads to increase the magnetic susceptibility. Correlations between magnetic data and
heavy metals concentration are often observed, e.g. [6, 7]. In case of MSWI residues, mass
specific magnetic susceptibility is also related to iron oxide mass fraction. Furthermore we
found positive correlation with REE, both LREE and HREE (Fig. 4). This correlation could be
linked to the presence of REE‐bearing devices in waste input (e.g., NdFeB magnets). A better
comprehension of the relation between the REE and susceptibility requires further study; at
this stage the aim is to understand whether the magnetic properties can be a quick and
inexpensive indicator of the REE presence in MSWI residues.
Each scatter plot allows to distinguish BA from FA samples, which are higher and lower than
1000∙10‐8 m3/kg, respectively. FA from bag filters (FAS and FAL) have the lowest REE content,
which usually correspond to lowest  values. High magnetic values do not necessarily reflect
an high REE content; rather around the 2000  value there is the highest density of samples
with high amounts of REE.

Figure 4: ƩLREE and ƩHREE vs. mass specific magnetic susceptibility (). Data set is grouped
by kind (BA=bottom ash; FAE=fly ash from ESP; FAL=fly ash from bag filter with Ca‐additive;
FAS= fly ash from bag filter with Na‐additive; FAU=untreated fly ash) and provenance (A=plant
A; B= plant B).
Mass flow analysis
With the simple mass balance given by the equation 1, we estimated the yearly flows of REE,
in kg/a, for the incinerators A and B (Tab. 2). The measured concentrations are provided by
ICP‐MS analyses. The owners of the incinerator plant provided the total BA and FA output.
We assumed that the sampled material is representative for one‐year process activity (i.e.,
2013).

Figure 5: Pie charts of LREE and HREE; the partitioning of mean concentrations and mean
flows between the different outputs are reported.
The REE mean concentrations are quasi‐equally distributed between the different kind of
MSWI residues. However, the estimated yearly mean flows is extremely high in BA (Fig. 5). An

individual estimation of BA flows is indispensable for each incinerator because of the high
heterogeneity of sample materials and sampling bias and as a consequence of t‐test results.
In the following table we decided to report BA flows for plants A and B with their standard
deviation values (on the basis of the samples batch). Flows from each kind of FA and the mean
flow of FA residues from the two plants are reported.

Table 2: The estimated annual flow (kg/a) of selected residues.

FAS (A) FAU (A) FAL (A) FAU (B) FAU (B) FAE (B) BulkFA (A,B) BA (A) BA (B)
REE kg/a kg/a kg/a kg/a kg/a kg/a kg/a M ‐kg/a S.D. M ‐kg/a S.D.
Ce 50 455 359 669 827 600 464 7590 ±600 10149 ±497
Dy 2 19 28 35 43 26 23 453 ±48.5 546 ±8.50
Er 1 14 15 19 24 15 14 303 ±29.9 373 ±12.7
Eu 1 17 13 22 26 19 15 219 ±19.2 275 ±17.1
Gd 2 29 67 42 124 31 42 501 ±47.0 647 ±33.3
Ho 0 7 5 6 8 5 5 94 ±10.2 107 ±0.20
La 21 276 251 373 440 324 271 4400 ±540 5290 ±760
Lu 0 1 1 2 3 2 1 39 ±6.60 44 ±1.60
Nd 12 153 182 253 322 206 172 2900 ±570 4576 ±420
Pr 3 40 44 70 88 63 47 805 ±130 1100 ±12.7
Sc 0 113 45 141 181 138 102 2400 ±630 2960 ±15.3
Sm 1 28 30 45 55 33 30 540 ±50.3 684 ±6.30
Tb 0 6 6 9 12 8 6 100 ±10.8 127 ±17.6
Tm 0 1 2 2 3 2 1 40 ±4.80 44 ±1.40
Y 24 236 289.0 316 388 272 238 3400 ±390 4207 ±200
Yb 1 10 13.0 19 25 15 13 270 ±37.5 300 ±3.50
LREE 84 1114 994.0 1618 2066 1418 1146 19300 ±2050 25700 ±1700
HREE 30 298 363.0 412 508 348 306 4700 ±510 5800 ±200
REE 90 1176 1068.0 1714 2187 1494 1213 20600 ±2200 27200 ±1700
REY 114 1413 1357.0 2031 2575 1767 1452 24000 ±2400 31400 ±1900
[FAS=Fly ash from bag filter with Na‐additive; FAU=Fly ash untreated; FAL= Fly ash from bag filter with Ca‐additive; FAE=Fly
ash from electrostatic precipitator (ESP); bulkFA=mean flow of fly ashes; BA=Bottom ash; A and B are the selected WTE plants]

Considering to fig. 5 and table 2, the FA residues from bag filters (with chemical additives)
have very low potential of REE flow. Bottom and fly ashes have an average concentration of
103 and 54 mg/kg ƩREY, respectively. For fly ashes the expected annual flow is 1.1 t/a LREE
and 0.3 t/a HREE, whereas for bottom ashes is 22.5 t/a LREE and 5.2 t/a HREE. The data of
annual flows are in good agreement with recent works [3].
As a semi‐quantitative index to evaluate the economic feasibility, we compare the typical
cutoff values used in the mining industry. The minimum cut‐off grade for naturally occurring
REE oxides is 0.2 wt.% [9], while this study shows a REE grade of 0.01 wt.%. Numbers seem
unfavourable, but the urban ore deposit has the advantage of being already in granular form,
in contrast with the natural ore deposit that may require lengthy and expensive processing in
order to reach the overburden and concentrate the ore from the gangue. However, to
dissolve vitrified MSWI residues is a very difficult task requiring highly aggressive chemical
conditions using strong acids. We think that low‐temperature pre‐concentration steps even
under high pH conditions (since the residues have strongly basic pH) can be explored as an
alternative extraction route in order to upgrade the REE contents from the starting urban ore.

Conclusions and outlook
Bottom and fly ashes from MSWI have an average concentration of 103 and 54 mg/kg ƩREY,
respectively. Mass flow analyses indicate that the expected annual flow is 1.1 t/a LREE and
0.3 t/a HREE, whereas for bottom ashes is 22.5 t/a LREE and 5.2 t/a HREE. Therefore MSWI
residues might represent a low concentration stream of REE. The independent sample t‐test
reveals that REE flows in FA and bulk MSWI residues (BA+FA) are statistically equal for the
two incinerators, while the estimation of REE flows in BA differ from the two plants. Major
elements like P2O5 and Al2O3 show a positive correlation with REE and also magnetic
susceptibility values can be used as proxies for REE occurrence, although magnetic data still
remains to be understood and other magnetic properties of MSWI residues should be
investigated.

References
1. European Commission, “Critical Raw Materials for EU”. (2010).
2. A. R. Chakhmouradian and F. Wall, "Rare Earth Elements: Minerals, Mines, Magnets (and More)."
Elements, 8(5): 333‐340 (2012).
3. L. S. Morf, R. Gloor, O. Haag, M. Haupt, S. Skutan, F. Di Lorenzo and D. Boni, "Precious metals and rare
earth elements in municipal solid waste‐‐sources and fate in a Swiss incineration plant." Waste
Management, 33(3): 634‐644 (2012).
4. P. H. Brunner and H. Rechberger, Practical Handbook of Material Flow Analysis. CRC Press LLC, Boca
Raton London New York Washington, D.C. 2004.
5. T. Meisel, N. Schoner, V. Paliulionyte and E. Kahr, "Determination of Rare Earth Elements, Y, Th, Zr, Hf,
Nb and Ta in Geological Reference Materials G‐2, G‐3, Sco‐1 and WGB‐1 by Sodium Peroxide Sintering
and Inductively Coupled Plasma‐Mass Spectrometry." Geostandards Newsletter 26: 53‐61 (2002).
6. G. Maier and R. Scholger, “Demonstration of connection between pollutant dispersal and atmospheric
boundary layers by use of magnetic susceptibility mapping, St. Jacob (Austria).” Physics and Chemistry
of the Earth, 29(13‐14): 997‐1009 (2004).
7. B. A. Reyes, F. Bautista, A. Goguitchaichvili and O. Morton, "Magnetic monitoring of top soils of Merida
(Southern Mexico)." Studia Geophysica et Geodaetica, 55(2): 377‐388 (2011).
8. Northwest Territories Geoscience Office ‐ Online Reports. Viewed on the 26th June 2014
(http://www.nwtgeoscience.ca/normin/cutoffs.html).



European REE Conference

Uploaded by

Copyright:

Available Formats

European REE Conference

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

European REE Conference

Uploaded by

Copyright:

Available Formats

RARE³: Research Platform for the Advanced

Shaking table Shaking table Shaking table Shaking table

Nutralization Dried 105°C

Filtration Burned 850°C

Session II: European Rare Earth

Greenland Carbonatite 5.9 Mt@ 1.8% TREO

Praseodymium applications are concentrated in the consumer products industries,

 “Terms like ‘risk’ and ‘critical’ are emotive,” and the end result could be panic and

Participation and stakeholder involvement in policy design and

ERES2014: 1st European Rare Earth Resources Conference|Milos|04‐07/09/2014

ERES2014: 1st European Rare Earth Resources Conference|Milos|04‐07/09/2014

ERES2014: 1st European Rare Earth Resources Conference|Milos|04‐07/09/2014

ERES2014: 1st European Rare Earth Resources Conference|Milos|04‐07/09/2014

ERES2014: 1st European Rare Earth Resources Conference|Milos|04‐07/09/2014

ERES2014: 1st European Rare Earth Resources Conference|Milos|04‐07/09/2014

ERES2014: 1st European Rare Earth Resources Conference|Milos|04‐07/09/2014

ERES2014: 1st European Rare Earth Resources Conference|Milos|04‐07/09/2014

ERES2014: 1st European Rare Earth Resources Conference|Milos|04‐07/09/2014

ERES2014: 1st European Rare Earth Resources Conference|Milos|04‐07/09/2014

ERES2014: 1st European Rare Earth Resources Conference|Milos|04‐07/09/2014

ERES2014: 1st European Rare Earth Resources Conference|Milos|04‐07/09/2014

A Framework for the Sustainable Development of Rare Earth

ERES2014: 1st European Rare Earth Resources Conference|Milos|04‐07/09/2014

ERES2014: 1st European Rare Earth Resources Conference|Milos|04‐07/09/2014

ERES2014: 1st European Rare Earth Resources Conference|Milos|04‐07/09/2014

Application of the generic framework in REEs Sustainable

ERES2014: 1st European Rare Earth Resources Conference|Milos|04‐07/09/2014

ERES2014: 1st European Rare Earth Resources Conference|Milos|04‐07/09/2014

ERES2014: 1st European Rare Earth Resources Conference|Milos|04‐07/09/2014

ERES2014: 1st European Rare Earth Resources Conference|Milos|04‐07/09/2014

ERES2014: 1st European Rare Earth Resources Conference|Milos|04‐07/09/2014

ERES2014: 1st European Rare Earth Resources Conference|Milos|04‐07/09/2014

ERES2014: 1st European Rare Earth Resources Conference|Milos|04‐07/09/2014

1.2 -0.8 Nd2O3 in 77LiF-23NdF3

ERES2014: 1st European Rare Earth Resources Conference|Milos|04‐07/09/2014

System R2 a [103 K] b

Y2O3 in 80LiF‐20YF3 0.986 ‐4.1  0.2 2.7  0.2

ERES2014: 1st European Rare Earth Resources Conference|Milos|04‐07/09/2014

Y2O3 in LiF-YF3 at 1023K

Y2O3 in LiF-YF3 at 1173K

ERES2014: 1st European Rare Earth Resources Conference|Milos|04‐07/09/2014

ERES2014: 1st European Rare Earth Resources Conference|Milos|04‐07/09/2014

1.6 LiF-MgF2-15NdF3 at 1073K

ERES2014: 1st European Rare Earth Resources Conference|Milos|04‐07/09/2014

ERES2014: 1st European Rare Earth Resources Conference|Milos|04‐07/09/2014

ERES2014: 1st European Rare Earth Resources Conference|Milos|04‐07/09/2014

ERES2014: 1st European Rare Earth Resources Conference|Milos|04‐07/09/2014

ERES2014: 1st European Rare Earth Resources Conference|Milos|04‐07/09/2014

Preliminary studies on lanthanum electro‐recovery from

ERES2014: 1st European Rare Earth Resources Conference|Milos|04‐07/09/2014

ERES2014: 1st European Rare Earth Resources Conference|Milos|04‐07/09/2014

ERES2014: 1st European Rare Earth Resources Conference|Milos|04‐07/09/2014

ERES2014: 1st European Rare Earth Resources Conference|Milos|04‐07/09/2014

ERES2014: 1st European Rare Earth Resources Conference|Milos|04‐07/09/2014

ERES2014: 1st European Rare Earth Resources Conference|Milos|04‐07/09/2014

Wt% 23‐25 3.5‐5 0.05‐ 62‐69 1 0‐10 0‐0.14 0‐0.1 1‐2

ERES2014: 1st European Rare Earth Resources Conference|Milos|04‐07/09/2014

Nd 3e ⇔ Nd ‐671.6 ‐2.32 E = ‐2.32 + 0.020 log [Nd3+]

Nd 3H2 O⇔Nd OH 3 3H 3e ‐563.3 ‐1.94 E = ‐1.94 +0.059 pH

Nd3 3H2 O⇔Nd OH 3 3H 108.3 3 pH = 19 – log [Nd3+]