Economic Processing of High Silica Bauxites - Exis

See discussions, stats, and author profiles for this publication at: https://www.researchgate.
net/publication/237828211
Economic Processing of High Silica Bauxites - Existing and Potential

Processes
Article
CITATIONS READS
6 594
1 author:
Peter Smith
The Commonwealth Scientific and Industrial Research Organisation
37 PUBLICATIONS 419 CITATIONS
SEE PROFILE
All content following this page was uploaded by Peter Smith on 13 November 2016.
The user has requested enhancement of the downloaded file.

Submission to the APP as part of the project (ATF-06-4)
High Silica Bauxite Processing
Economic Processing of High Silica Bauxites –

Existing and Potential Processes
Peter Smith
Parker Centre, CSIRO Light Metals Flagship
December 2008
Table of Contents
1. Introduction................................................................................................................ 3
1.1 Scope ............................................................................................................. 3
1.2 Bauxite types.................................................................................................. 3
1.3 Current world bauxite reserves ...................................................................... 4
1.4 The Bayer process ......................................................................................... 6
1.5 Structure of DSP ............................................................................................ 7
1.6 Strategies for caustic loss minimisation.......................................................... 9
2. Potential Technologies – Strategy One – Reduce Reactive Silica Input ................. 10

2.1 Screening / washing ..................................................................................... 10
2.2 Gravity separation ........................................................................................ 11
2.3 Flotation ....................................................................................................... 12
2.4 Bio-leaching ................................................................................................. 13
2.5 Magnetic separation ..................................................................................... 14
2.6 Mechano-chemical treatment of bauxite with lime........................................ 14
2.7 Roast-leach process..................................................................................... 15
3. Potential Technologies – Strategy Two - Modify the Process.................................. 17

3.1 Sinter processes........................................................................................... 17
3.2 Combination process (Bayer-sinter) ............................................................. 18
3.3 Bauxite activation ......................................................................................... 19
3.4 Double digestion .......................................................................................... 20
3.5 Low soda DSPs............................................................................................ 22
3.6 Hydrothermal lime treatment ........................................................................ 24
3.7 Differential extraction processes .................................................................. 25
4. Potential Technologies – Strategy Three - Soda Recovery Processes ................... 28

4.1 Lime sinter of DSP ....................................................................................... 28
4.2 Weak acid leaching of DSP.......................................................................... 29
4.3 Complex causticisation (mud causticisation) ................................................ 30
4.4 Mechano-chemical treatment of residue ...................................................... 31
5. Summary and Conclusions...................................................................................... 32

5.1 Strategy One options.................................................................................... 32
5.2 Strategy Two options.................................................................................... 32
5.3 Strategy Three options ................................................................................. 33
6. Acknowledgements ................................................................................................. 34
Economic Processing of High Silica Bauxites – Existing and Potential Processes
Peter Smith (Parker CRC, CSIRO Light Metals Flagship)
1. Introduction
1.1 Scope
This document will critically review existing and potential processes for the economic
processing of high silica bauxites. High silica bauxites are usually considered to be
those with reactive silica contents 1 in excess of ~8% by weight of the bauxite.
“Economic processing” is meant to imply production of alumina at a cost comparable to
that from low silica bauxite by a conventional Bayer process.
This review is limited to bauxites as feedstocks for the processes. Bauxites are those
ores in which the valuable (extractable) aluminium is in the form of hydroxides or oxy-
hydroxides of aluminium (i.e. gibbsite, boehmite and diaspore). In addition the review
will be restricted to alkali leaching of bauxites – which includes all forms of Bayer
process variations. The review will not address alumina production from non-bauxitic
materials such as clays, shales, nephelines or fly ash. It will also not consider alumina
production using non-alkaline processes such as those using mineral acids, reduction,
or bacterial processes. These have been adequately reviewed (until the mid 1980’s) by
O’Connor (O'Connor 1988).
There is an assumption that the reader is familiar with the basics of the Bayer process,
and North American liquor terminology. In brief, the Free Caustic content of a liquor is
defined as g/L of sodium hydroxide, Caustic (or Total Caustic) is defined as Free Caustic
plus sodium aluminate, and Soda is defined as Caustic plus sodium carbonate - all
expressed as g/L of equivalent sodium carbonate. Aluminium in solution is expressed
as g/L Al2O3.
1.2 Bauxite types
Bauxites are usually considered to be of two major types; (1) “lateritic” (sometimes
called equatorial) and (2) “karst” bauxites. Both are weathered products from underlying
parent rock; lateritic bauxites derive from primary aluminosilicate rocks whereas karst
bauxites from interbedded carbonate and aluminosilicate rocks.
Lateritic bauxites, formed in equatorial climates comprise ~90% of the world’s

exploitable bauxite reserves (Freyssinet, Butt et al. 2005). The weathering process has
resulted in a typical profile in which the valuable aluminous material lies atop of an
aluminosilicate base (often clay) and has formed from it through the leaching of silica.
Bardossey and Aleva (Bardossy and Aleva 1990) have described 6 types of lateritic
bauxites, differentiated mainly by the age of the deposit and the different weathering
conditions it was exposed to. The main silicate mineral is kaolinite which is often
associated with goethite as the iron mineral. Aluminous minerals are predominately
gibbsite and to a lesser extent boehmite.
1
Reactive silica is defined later in this document.
HSB review 2008_240209 This version 24 February 2009 Page 3 of 42

Karst bauxites have a different typical mineralogical composition which stems from the
presence of carbonates in the parent rock and different weathering conditions to those
for lateritic bauxites. Silicate minerals are again dominated by kaolinite, but may include
more difficult to process minerals such as chamosite. Aluminous minerals are more
likely to be boehmite and diaspore, with the changes to the mineralogy compared to
lateritic bauxites sometimes attributed to different oxidative conditions experienced
during weathering (Bardossy 1982). Karst bauxites are largely found in Eastern Europe
and Northern Asia.
The different mineral composition of these two main bauxite types has influenced the
way they are processed. In general lateritic bauxites are easier to digest than karst
bauxites and are treated by Bayer process variations using less severe conditions of
caustic concentration, temperature and/or holding times. The difference in the
“processibility” of these two bauxite types will also influence the strategies for developing
viable processes for treating high silica versions of these ores.
The following table describes the approximate compositional differences between

lateritic and karst bauxites. It is not intended to be definitive.
Table 1 Approximate mineralogical compositions of lateritic and karst bauxites
Element Lateritic Karst
Al2O3 Gibbsite, Boehmite Boehmite, Diaspore
SiO2 Kaolinite, Quartz Kaolinite, Quartz, Chamosite, Illite
Fe2O3 Goethite, Hematite Hematite, Goethite Maghemite, Magnetite
TiO2 Anatase, Rutile Anatase, Rutile, Ilmenite
CaO Calcite, Apatite Calcite, Apatite, Crandallite
A classification of bauxites, particularly from the point of view of Bayer processing has
been given by Hill and Robson (Hill and Robson 1981).
1.3 Current world bauxite reserves
Table 2, sourced from 2008 US Geological Survey “Mineral Commodities Summaries”,

gives world bauxite production and reserves by country. The table gives 2006 and 2007
production figures (sorted from highest to lowest 2007 production) and two indicators of
bauxite reserves.
According to the USGS report (Bray 2008) the “Reserve Base” is the part of the
resource that meets minimum physical and chemical criteria related to grade, quality
thickness and depth, and includes those resources that are currently economic,
marginally economic and some that are current sub-economic. By contrast the
“Reserve” is a sub-class of this definition, restricted to the part that is economically
recoverable with current technologies.

Table 2 World bauxite production and reserves (Bray 2008).
Production
Country Reserve Reserve base
2006 (106 T) 2007 (106 T) (106 T) (106 T)
Australia 62.30 64.00 5,800 7,900
China 21.00 32.00 700 2,300
Brazil 21.00 24.00 1,900* 2,500*
Guinea 14.50 14.00 7,400 8,700
Jamaica 14.90 14.00 2,000 2,500
India 12.70 13.00 770 1,400
Russia 6.60 6.00 200 250
Venezuela 5.50 5.50 320 350
Suriname 4.92 5.00 580 600
Kazakhstan 4.80 4.90 360 450
Guyana 1.40 2.00 700 900
United States - - 20 40
Other 5.46 6.80 3,400 4,000
countries
World Total 178.00 190.00 25,000 32,000
N.B. sorted by 2007 production. * suspected to not include the “new” Paragominas deposit which could easily double
this reserve, T=metric ton.
Alternative figures from the most recent Roskill report (Roskill 2008) are slightly different,
but cannot be reproduced due to copyright issues. In particular they indicate lower
Chinese and higher Guinea production for 2007, with a world total production for 2007 of
~209 MT. Gu (Gu 2008) puts the Chinese reserve at 540 MT and the Chinese bauxite
production at ~40 MT for 2007. Vietnam is not listed by the Bray report, probably
because the deposits there are only now being evaluated, but this will undoubtedly be a
significant source of bauxite in the future.
The Bray report estimates world bauxite resources totalling 55–75 billion tons.
Resource is defined as the concentration of bauxite such that economic extraction is
potentially feasible. Resources presumably include non-identified reserves, as well as
holdings that are considered non-economic with present technology. The report gives
the breakdown of world reserves by region and these are reproduced in Table 3.
Table 3 World bauxite resources by region (total world estimate 55-75 billion tons).
World region % of world bauxite resource

Africa 33
Oceania 24
South America and Caribbean 22
Asia 15
Others 6
Resources that are currently not part of the demonstrated reserves or reserve bases are
almost certainly dominated by bauxites that contain high levels of reactive silica. The
55–66% of world resources that are not currently described as being within the

economic reserve indicates the potential for new processes to exploit these holdings and
bring them into the realm of economic reserves.
1.4 The Bayer process
The Bayer process is the principal method for production of alumina from bauxite
worldwide. The modern version of the process (developed in the 1880’s) still maintains
the key steps of dissolution of alumina-rich minerals into hot caustic solution, separation
of the insoluble phases, followed by gibbsite precipitation and calcination of the gibbsite
to alumina (Al2O3). A good summary of the process and some variants is given by Misra
(Misra 1986). A schematic is shown in Figure 1. Operations with a red background
represent those involving either the bauxite or mud, and thus represent the so-called
“red-side” of the Bayer process. Operations with a white background relate to processes
in the absence of these solid materials, and represent the so-called white-side
operations (eg precipitation, evaporation).
Bauxite
Pre-desilication
Digestion
Evaporation
Oxalate
Clarification
Classification
/Washing
Mud
Washing
Precipitation
Seed Residue
Disposal
Gibbsite Product
Figure 1 Schematic of the basic Bayer process (Calcination of gibbsite to alumina not shown).
Bauxite is first reduced through crushing and grinding (usually in a portion of “spent”
liquor) at ~60oC. Silicate minerals, principally kaolinite, start to dissolve in the spent
liquor.
3 Al2 Si2O5 (OH ) 4 + 18 NaOH → 6 Na2 SiO3 + 6 NaAl (OH ) 4 + 3H 2O eq.1
In many modern operations there is a slurry storage or “pre-desilication” stage in which

slurry is kept at high volume-fraction at near atmospheric boiling for several hours to
encourage the re-precipitation of the solubilised silica to DSP (desilication products).

6 Na2 SiO3 + 6 NaAl (OH ) 4 + Na2 X → Na6 [ Al6 Si6O24 ].Na2 X + 12 NaOH + 6 H 2O eq.2
The above equation represents the formation of Bayer sodalite (in which X represents a
variety of inorganic anions, most commonly sulphate, carbonate, chloride, aluminate and
hydroxide.
Slurry storage usually transforms ~80–90% of the reactive silica into DSP, with the
remainder being converted in digestion. Digestion conditions are tailored to the
aluminous phase distribution of the bauxite, such that if gibbsite is the only source of
recoverable alumina, the digestion is performed at temperatures 140–155oC. Digestion
times are often determined, not by the kinetics of gibbsite dissolution, but by the kinetics
of residual DSP formation, and more importantly, the reduction of the soluble silica in
solution. Bauxites containing non-trivial boehmite concentrations are mostly digested at
temperatures 220–270oC. Again, time in digestion is often not dictated by alumina
dissolution kinetics, but kept to a minimum to avoid excessive dissolution of quartz which
provides additional reactive silica.
Since DSP is usually discarded with the red mud residue, the two equations above
demonstrate the loss of sodium hydroxide from the liquor (at least 1 mole NaOH per
mole of reactive silica) 1 . This loss of soda from the liquor is the source of the economic
penalty and is linear with the reactive silica content of the bauxite. Bauxites with
reactive silica contents greater than ~8% by weight (A/S=6.25) are usually considered to
be uneconomic.
Bauxite is charged to digestion to achieve a specific alumina supersaturation. The level

is chosen to ensure that the supersaturated liquor is stable through the flash train and
subsequent clarification stages.
After clarification the liquor is sent to precipitation (gibbsite crystallisation) and the
insoluble components (usually referred to as red mud) are sent to a counter-current
washing system. The counter-current washing system returns liquor to the circuit but
does not provide conditions suitable to recover the so-called “fixed soda” from the DSP.
It is worth noting here that refineries which currently process bauxites with relatively high
levels of reactive silica also carry the legacy of high fixed soda in their residue areas.
This may have residue disposal and/or environmental consequences.
1.5 Structure of DSP
The Bayer sodalite described in the previous section is only one of several mineral
phases that precipitate silica from solution and can therefore be termed “desilication
products” 2 . The others that can form include cancrinite (formed at digestion
temperatures of ~220oC and above) and also silica substituted tri-calcium aluminates
(hydrogarnets) which are the consequence of the dissolution of lime compounds, either
from the bauxite itself or from lime added to control liquor impurities (e.g. phosphate).
1
In addition, if quartz supplies part of the reactive silica, then alumina is also lost from the liquor.
2
There are many other forms of “desilication product” that occur during pyrometallurgical processing of bauxite or
residue. These are discussed in the relevant potential processes.

The structure of Bayer sodalite is given in Figure 2.
Figure 2 Bayer sodalite schematic showing cubic structure. Full occupancy is achieved with a monovalent
ion in each cage, or a divalent ion in every other cage.
The aluminosilicate cage structure has a net three negative charge which is balanced by
three sodium ions. At full occupancy Bayer sodalite has a monovalent anion (and
associated sodium) in each cage, or a divalent anion in every other cage (see
stoichiometry in equation 2). Soda can be recovered from this structure only if the
sodalite cage structure is broken down, or if the included ions can be removed through
the openings of the sodalite cage structure (Smith, Wingate et al. 2008).
At high digestion temperatures cancrinite, another cage structured aluminosilicate can

form from sodalite. Cancrinite has the same cage stoichiometry as sodalite, but has a
different structure Figure 3.
cage
channel
Figure 3 Structure of cancrinite (hexagonal). Cancrinite has two possible sites for included ions, either in
cages which are smaller than those in sodalite, or in channels which are wider than the sodalite cage
dimensions.
Cancrinite has two possible locations for included ions, either in the cage structures
(which are smaller than those in sodalite) or the larger linear channels that form part of
the hexagonal structure. The size and location of the channels mean that the preferred
inclusion for cancrinite is calcium carbonate (Na6[AlSiO4]6.2CaCO3). Thus cancrinite
has 25% less sodium than sodalite and the added bonus of favouring a non-useful
anion.
Other compounds that can be considered desilication products (which are particularly
relevant to options for processing high silica bauxites) are the hydrogarnet series of
minerals. These are silica substituted grossulars of the general formula

Ca3Al2(SiO4)n(OH)12-4n. Natural minerals exist for integer n values, 0 (hydrogrossular or
tri-calcium aluminate, TCA), 1 (katoite), 2 (hibbschite) and 3 (grossular). Isomorphous
substitution of Si4+ for 4H+ gives rise to a range of solid solution compounds under Bayer
conditions, ranging from 0.6>n>0 (Whittington 1996b).
1.6 Strategies for caustic loss minimisation
In a Bayer type process, reactive silica will dissolve in the caustic aluminate solution and
re-precipitate as a variety of (mostly) sodium aluminium silicates which are then
discarded as part of the red mud residue. The opportunities to reduce the soda lost to
the system therefore fall into three broad strategies:
1. Reduce the input of reactive silica to the process. This could take the form of
beneficiation which may be preceded by a pre-treatment to allow separation of
silica from the bauxite, or some other form of pre-treatment that effectively
renders the silica unreactive during the subsequent Bayer digestion.
2. Modify the process to produce a lower soda residue. This could be by reducing
the fraction of reactive silica transforming to DSP, or by the formation of
alternative DSPs with lower soda content.
3. Recover soda from DSP. This would constitute a process developed to return to
the Bayer liquor stream valuable caustic derived from the re-processing of either
residue or a concentrate high in DSP produced from the residue.
Depending on what technology is being considered/developed in these three strategies,

then (1) and (3) could be considered to be “add-ons” to the Bayer process (either at the
front or back of the process) whereas (2) would constitute a change to the Bayer
process itself. However there are undoubtedly some options which would blur this
distinction.
The remainder of this document will concentrate on potential technologies (either

previously considered or previously used commercially). They will be grouped under the
broad strategies described above. For each technology, a brief description of the
principle will be given, together with a status or state of development, and an evaluation
of the techno-economic and (where appropriate) ecological hurdles to implementation.

2. Potential Technologies – Strategy One – Reduce Reactive Silica Input
2.1 Screening / washing
Principle:
Screening and washing is effective in reducing silica in the bauxite feed material if the
silica is preferentially concentrated in the smaller particles size fractions or fines.
Bauxite is “wet screened” on inclined or vibrating screens. The wet screening rejects
pisolites lower than a certain size, and, at the same time, washes the fine material
(adhering to the larger pisolites) which is also high in silica concentration.
Status:
Weipa bauxite mined by Rio Tinto in Queensland Australia is pisolitic, with the silica
level (both kaolinitic and quartz) decreasing with increasing pisolite size. The
beneficiation plant operated by Rio Tinto “wet” screens at a specific pisolite size and the
undersize (tailings) are rejected. A schematic of the operations is given in Figure 4
(Roberts and Dunne 1980). Although the schematic is dated the essentials are thought
to be unchanged. In particular it shows a cut-off size of ~1.7 mm (Weipa ore) and
~0.4 mm (Andoom).
Figure 4 Schematic of Weipa and Andoom beneficiation operations at Weipa (From Roberts and Dunne, 1980)

Sometimes washing and screening will target quartz only. Rousseaux et al.
(Rousseaux, Verschuur et al. 2006) examined the use of drum washing (and
subsequent cyclones) to remove quartz known to be in Pijiguaos bauxite in the size
range 0.1–3.0 µm. Amer (Amer and Abdel-Aal 1996) separated ~82% of quartz from an
Egyptian bauxite by simply pulverising and screening at 63 µm.
Hurdles to implementation:
The application of this technology will depend heavily on the micro-mineralogy of the
bauxite. In the case of the pisolitic bauxite, kaolinitic silica is sometimes more prevalent
in the outer layers of the pisolite, and thus the silica concentration increases with
increasing surface area and thus decreasing size. Owada (Owada, Okajima et al. 2005)
examined the possibility of “surface grinding” to remove a kaolin rich outer layer in Gove
pisolitic bauxite. When quartz is the target, then the effectiveness of beneficiation
depends on the size distribution of that mineral and also its liberation – i.e. the degree to
which the quartz can be separated from the valuable aluminous minerals.
At Weipa, the rejection of the undersize pisolites results in the return of this material to
the minesite, essentially unchanged. If pre-processing is needed to liberate the target
minerals (i.e. grinding before screening), then this is likely to be done as a slurry, and
the rejected material might constitute a tailings disposal issue. In all cases, it is
assumed that water is available to aid the screening process. Dry screening may be
used in arid mining areas, but will be less effective than wet screening.
2.2 Gravity separation
Principle:
Gravity separation of phases in bauxite relies on the density between the valuable
alumina minerals (e.g. gibbsite) and the silicate minerals (e.g. kaolinite). It is rarely used
because this difference is small. If this technology has been considered, then the effects
of density differences have often been magnified by the use of hydrocyclones. A recent
paper by Gao (Gao, Li et al. 2008) describes conditions of the operation of the cyclone
equipment to optimise alumina recovery and silica content (A/S ratio).
100 8
80 6
Recovery (%)
Recovery of Al2O3 in underflow

Ratio of A/S
Ratio of A/S in underflow

60 Recovery of SiO2 in overflow 4
Ratio of A/S in overflow
40 2
20 0
5 7 9 11 13
Apex diameter (mm)
Figure 5 Effect of hydrocyclone geometry on grade and recovery of bauxite (re-drawn from (Gao, Li et al.
2008)).

Gravity separation for beneficiation has been used to target valuable by-products from
the reprocessing of Bayer residue. Rao (Rao, Das et al. 1996) examined the possibility
of separating titanium minerals (rutile and ilmenite) from Bayer residue sand using
standard Wilfley tables. The work was successful partly because of the high density of
the titanium minerals which would be heavier than sodalite, goethite and quartz, but
lighter than hematite. Picaro (Picaro, Pei et al. 2002) examined the upgrading of DSP
and TiO2 minerals in QAL residue, concluding that a ~40% DSP concentrate could be
produced at 89% recovery, but that the richest TiO2 fraction was ~20-36% at only 7%
recovery.
Status:
None of the authors whose work is summarised above have indicated any commercial
use of this technology.
It is unlikely that gravity separation (or hydrocyclones) could ever be used economically
to reduce silica in lateritic bauxites. For these bauxites the alumina and silica minerals
are intimately mixed so that adequate liberation is not possible. Sometimes in non-
lateritic bauxites (like the diasporic bauxites found in China) the silicate minerals are less
intimately mixed. This may explain why Gao has had some success with hydrocyclones.
2.3 Flotation
Principle:
Bauxite can be floated in a suitable medium (usually near neutral aqueous solutions)
using proprietary chemicals to absorb on the target minerals to either activate them
towards collection/attachment by bubbles or to depress them. In the first variation,
valuable minerals preferentially report to the froth phase and are thus separated from
silicate minerals which become part of the gangue. In the alternative strategy, the
valuable alumina minerals are depressed and therefore report to the tailings (reverse
flotation).
Flotation is particularly suited to diasporic ores in which there is sufficient liberation of

alumina from silica minerals, and the silicate minerals are of a particle size suitable to
flotation ~(10–200 µm).
Status:
After the initial work by the Russians (Andrew, Anishchenko et al. 1973), (Ishchenko,
Korus et al. 1974) it has been the Chinese researchers who have developed this area,
from the initial work of Li (Li and Chen 1979) to a great body of work in the years 2000
onwards by authors such as Hu (Hu, Ouyang et al. 2008) and Wang (Wang, Hu et al.
2004). A good review is given by Xu (Xu, Plitt et al. 2004) which covers the literature in
reverse flotation of diasporic bauxites up until 2004, and more recently Zhong (Zhong,
Liu et al. 2008). The western literature perhaps does not reflect the depth of this work.
For example it is suspected that a great deal of work has been done by both Chalco and
BGRIMM which is less publicised in English and therefore less well known to western
researchers. It is also possible that more of their work appears in the patent literature
e.g. (Chen and Li 2006).

The industrial implementation of bauxite flotation has been through the Chalco owned
Zhongzhou Light Metals Research Institute and associated refinery which has
demonstrated this technology with both Shanxi and Henan bauxite (Li, Song et al. 2002).
This technology is already being used commercially, and its use is expected to increase
for those bauxites which are applicable (mainly diasporic bauxites). A hurdle is
expected to be problems with the organic input associated with flotation chemicals
especially their effect on downstream processes e.g. digestion (Loginova, Koryukov et
al. 1979).
As mentioned before, this technology is not expected to be useful for lateritic bauxites
where the silicate and alumina minerals cannot be sufficiently liberated to achieve good
upgrading and recoveries. A good explanation of liberation with respect to beneficiation
options is given by Solymar (Solymar, Madai et al. 2005).
2.4 Bio-leaching
Principle:
Bacterial strains are mixed with bauxite and a growth medium (usually a glucose or
related carbon source) from which certain elements of the minerals are extracted into
solution. A key mechanistic question for all mineral bio-beneficiation is whether the
solubilisation of elements from minerals involves a form of direct attachment of the
bacteria onto the mineral surface, or by the simple chemical action of bacterial
metabolites (usually organic acids) on the minerals (Ehrlich 1992).
Status:
Anand (Anand, Modak et al. 1996) examined the leaching of calcium and iron from
bauxites, i.e. phases most likely to be solubilised by the action of organic acids.
Comparison of the action of the bacteria strain to the leaching produced by the
metabolites only (i.e. no bacteria) indicated that the solubilisation was due to both
metabolite and direct attachment action. These results were later confirmed by Vasan
(Vasan, Modak et al. 2001).
Groudeva and Groudev (Groudeva and Groudev 1983) examined the specific use of
“silicate” bacteria to remove silicates from bauxite. The origin of the bacteria was not
clear, in that they were said to have been isolated from “soil and rock samples”, but the
most successful strain was one related to Bacillus circulans. A common problem
encountered when trying to “solubilise” silica is the co-leaching of alumina (either from
kaolin or from gibbsite) which will suppress the silica solubility in the medium. Bacillus
circulans is known to produce “mucilaginous capsules” which evidently can concentrate
the alumina and silica dissolved by the bacterial action. Best results were obtained with
growing organisms with a glucose food source and fresh medium (i.e. not containing
spent nutrient components which can inhibit bacterial action). A continuous flow-through
system was tested in which 55% of silica could be removed from the bauxite at
~100 g/L over a period of 5 days.
The hurdles to implementation are, most obviously, the time required for pre-treatment.
Other considerations are the non specific nature of the dissolution (both aluminium and
silicon are dissolved), the requirements for a growth medium, and the separation of the

dissolved silica and the bacteria from the desilicated bauxite. The time issue has been
addressed in other bacterial applications by the use of heap leaching technology.
2.5 Magnetic separation
Principle:
Magnetic separation (either wet or dry) is used to produce two product streams, one
magnetic (or weakly magnetic) and one non-magnetic. Magnetic separation may or may
not be preceded by a pre-treatment designed to enhance the magnetic susceptibility of
some minerals (e.g. a reducing roast to encourage the reduction of iron minerals to
produce magnetic phases).
Status:
There has been some work on the use of magnetic separation for the reduction of iron
contents in bauxites for the production of refractory bauxites (see e.g. (Bhagat, Banerjee
et al. 2001) (Rao 1996) (Sadler and Venkataraman 1991)), however there does not
appear to be any industrial application of this technology, even in India where the
majority of this work has been done. Bartosik (Bartosik, Strnad et al. 1974) indicates
that magnetic separation has been used (in conjunction with flotation) to remove both
calcite and corundum from bauxite to produce a feed suitable for the Bayer process.
There has been more interest in the use of magnetic separation as part of residue re-
use programs, in order to produce an iron based by-product (e.g. (Mishra, Staley et al.
2002)) but again there is no evidence that this has been implemented commercially.
There does not appear to be any work in the literature specifically detailing the use of
magnetic separation to remove silica from bauxite.
There is some possibility that magnetic separation could be useful for silica reduction on
the basis of mineral associations. For instance, if the reactive silica minerals
preferentially reported to the magnetic iron fraction by virtue of their associations then an
upgrading may be possible. Similarly if boehmite could be separated from quartz, then
the possibility may exist of producing a feed lower in reactive silica at higher Bayer
digestion temperatures.
2.6 Mechano-chemical treatment of bauxite with lime
Principle:
Mechano-chemical treatment (activation) is a form of ultra-fine grinding in which certain
chemical reactions occur that would not in the absence of the grinding. It is thought that
the energy required to initiate these chemical reactions is provided by the collisions of
the grinding media (perhaps in the form of high localised temperatures) which are then
transferred to the reactants.
Status:
The application of mechano-chemical treatment to bauxite was first suggested by
Pawlek (Pawlek, Kheiri et al. 1992). They found that ultra-fine ground bauxite promoted
high alumina loadings at low temperatures during subsequent digestion, and that the

addition of lime reduced the concentration of dissolved silica exiting digestion.
McCormick (McCormick, Picaro et al. 2002) refined this observation and noted that the
addition of lime promoted the formation of silica substituted TCA (hydrogarnets) from
both kaolinite and quartz, which reduced caustic losses (albeit at high lime charges,
>~9% CaO on bauxite). A patent on this technology also exits (Picaro and McCormick
1997).
A recent paper by Fortin and Forté (Fortin and Forté 2007) re-examines the claims of
McComick but the authors were unable to reproduce TCA formation using ultra-fine
grinding, and thus also did not see any reduction in caustic loss. It is not clear whether
this is due to a difference in the methods or materials, or whether there was a
fundamental problem with either group’s results.
The advantage of this technology compared to the “Lime Bayer” technology promoted by
Chinese researchers (q.v.) appears to be the formation of hydrogarnets without side
reactions. As noted in previous sections, hydrogarnets can be an effective, if inefficient
method of reducing caustic losses (typically the molar ratio of Ca:Si in hydrogarnets
produced in the Bayer process is 5:1).
Mechano-chemical treatment appears to convert most, if not all silica into hydrogarnets,
including some or all of the quartz that would otherwise pass through digestion
unreacted. This adds to the inefficiency of the process. These shortfalls combined with
the substantial costs associated with ultra-fine grinding of bauxite suggest that this could
not be an economical way of reducing caustic losses.
2.7 Roast-leach process
Principle:
When kaolin in bauxite is heated to ~550oC it dehydrates to metakaolinite. Further
heating to 980oC causes a sudden “splitting” of the kaolin into 2 or more products.
Amorphous silica (SiO2) is certainly produced. There is some conjecture about the other
products: a transition alumina (possibly γ-alumina) and/or and alumino-silicate spinel are
produced (Brown, MacKenzie et al. 1985) (Chakraborty and Ghosh 1978) (Santos,
Campos et al. 2005). The uncertainty is exacerbated by the fact that both the spinel
identified 2Al2O3.3SiO2 and γ-alumina have similar structures and therefore X-ray
diffraction patterns.
After the roast, the bauxite is subjected to a weak alkaline leach that dissolves the
amorphous silica (to form a sodium silicate solution). The weakness of the caustic
solution ensures that the alumina from the kaolin, and the alumina from the other
alumina values (which have also been transformed into transition aluminas by the
roasting process) are not dissolved. The “desilicated” bauxite is then fed to a Bayer
process variation capable of digesting the transition aluminas produced.
Status:
The process appears to have its origins in Eastern Germany (Schoenfelder and
Ginsberg 1960) – which has references to earlier (~1955) East German patents. Some
further development occurred in Czechoslovakia (Klan 1959) where it was apparently
used commercially for a while.

After a long period, the process has received further attention (Lin, Malts et al. 1998)
(Rayzman, Pevzner et al. 2003) (Qiu, Jiang et al. 2004). Rayzman describes the
operation of this process on a pilot scale at the facilities of VAMI in St Petersburg. It is
not clear whether this process has ever been used commercially within Russia or other
former Eastern block countries.
Bauxite Crushing Roasting
Alumina
concentrate Leaching Roasted bauxite
Sodium silicate Weak caustic

solution solution
Bayer
Lime
Process
Causticisation Calcium silicate
Alumina Red mud
Figure 6 A schematic of the Roast-Leach process (after Rayzman (Rayzman, Pevzner et al. 2003)).
Roasting bauxite to ~1000oC is energy intensive and would be difficult to justify
economically. However in theory, the process has the potential to reduce soda losses
significantly. In addition a portion of the alumina once associated with the kaolin is now
potentially available for recovery (boosting the effective “available alumina”). The
savings in soda and extra alumina produced need to cover not only the energy of the
roast, but the re-activation of the sodium silicate solution – and this has caused much of
the problems experienced at VAMI. The silicate solution is limed to produce calcium
silicate (a saleable by-product) but this process is compromised by the leaching of small
quantities of alumina which reduce its efficiency.
In addition the desilicated bauxite contains transition aluminas instead of the original
hydrated (oxy) hydroxides. This necessitates more aggressive Bayer digestion
conditions (temperatures over 260oC and retention times exceeding 30 minutes). The
suitability of the particular bauxite for this process has to be considered carefully, since if
the bauxite contains substantial quartz, the extra DSP formed by additional quartz attack
can reduce or even eliminate the caustic saving advantages sought.
In spite of these hurdles there is potential in this approach, especially if a way can be
found to offset the additional energy costs.

3. Potential Technologies – Strategy Two - Modify the Process
3.1 Sinter processes
Principle:
In all of the different variations of the sinter process, the common principle is that the raw
material is heated, together with another chemical (usually an alkali) to induce a solid
state reaction. An alkaline leach is then performed to separate the aluminous material
from the insolubles. Of the sinter process variants, two are within the scope of this study
– the lime sinter process and the lime-soda sinter process.
In the lime sinter process the additive is a calcium salt (usually calcite – CaCO3). The
mixture is heated to above the decomposition temperature of calcite (~825oC) and solid
state reactions occur between the lime and the alumina values (gibbsite and boehmite)
and also with alumino-silicates (kaolinite). This requires temperatures of ~1150oC or
above. The mixture ratios and the sintering conditions are designed to optimise the
production of various calcium (aluminium) silicates, which, after cooling, are amenable to
leaching with an alkaline solution, either caustic, or more commonly sodium carbonate.
Gibbsite is subsequently recovered either by seeding or by carbonation (to regenerate
the sodium carbonate for leaching). The equations below describe the primary reactions
with silica and with alumina, but many more are also possible, leading to a wide range of
products -
2CaCO3 + SiO2 → 2CaO.SiO2 + 2CO2

12CaCO3 + 7 Al2O3 → 12CaO.7 Al2O3 + 12CO2
In the lime-soda sinter process the bauxite is heated as before with calcite, but also with
a source of soda (usually solid sodium carbonate – soda ash). The relative mixtures of
all three components are designed to promote the reaction of the aluminous material
with soda to produce sodium aluminate and the reaction of lime with silica to form
calcium silicates (Ca:Si 2:1 and 1:1). Soda also reacts with iron minerals to form sodium
ferrites 1 .
2 Al (OH ) 3 → Al2O 3 +3H 2O

Al2O3 + Na2CO3 → 2 NaAlO2 + CO2
Fe2O3 + Na2CO3 → Na2O.Fe2O3 + CO2
4CaCO3 + Al2O3 .2 SiO2 → 2(CaO ) 2 .SiO2 + Al2O3 + 4CO2
Again these reactions are not comprehensive, and are given in oxide form for the sake
of clarity. Both the (dehydrated) sodium aluminate and the sodium ferrite are soluble in
water or weak caustic solution. The sodium ferrite almost immediately reverts to
hematite, producing sodium hydroxide which helps stabilise the sodium aluminate.
Status:
The literature on sinter processes is vast. A comprehensive review of many of the
process variants is given by O’Connor, including versions designed to process non-
1
Sodium may also react with titanium minerals. Also a number of tertiary compounds involving Na2O, CaO, Al2O3 ,
Fe2O3 and TiO2 will form depending on the mixture ratios and the reaction conditions (Eremin 1974).

bauxitic feedstocks (O'Connor 1988). The specific use of sinter processes for the
production of alumina from low quality (high silica) bauxites is given in a series of articles
collected in a 1974 issue of Travaux (ICSOBA). In particular the papers by Bartosik
(Bartosik, Strnad et al. 1974), Eremin (Eremin and Shmorgunenko 1974) and Tomka
(Tomka 1974) are useful. Although these papers represent snapshots of technology
that was in use over 30 years ago, the sinter process in the Eastern block countries was
then in widespread use, and these papers represent well established process variants.
More recently the efficiency of phase formation (including the avoidance of the
undesirable α-alumina is discussed by Lu and Feng (Lu, Fang et al. 2003) and general
optimisation of the sinter process by Li et al. (Li, Liu et al. 2004).
The commercial operations still using this (and related) technology are those processing
low grade, often diasporic ores in both Eastern Europe (Russia, Ukraine, Turkey) and
China. Refineries still using the sinter process include Zhongzhou, Nanting and
Shandong in China, Bogoslovsk, Tikhvin and Kammensk in Russia and Dnjepr in the
Ukraine.
Hurdles to implementation
The main disadvantage of these processes is energy costs. Whilst typical Bayer
refineries, processing high grade lateritic bauxites can expect to produce alumina at
~11–13 GJ/t Al2O3, the sinter process energy costs are ~38 GJ/t Al2O3 (Arlyuk 1995).
New refineries are very unlikely to adopt such technology, and the trend has been to
adapt older refineries to more efficient Bayer technology, where this is seen to be
appropriate for the bauxite used. In some cases there are limited options for domestic
bauxite supplies, and this explains the continued use of sinter and combination
processes (see next section) especially in China.
An inherent technical problem with these technologies is the high level of dissolved silica
that is generated by the presence of many and diverse calcium (aluminium) silicates,
some of which have greater solubilities than the corresponding phases produced in a
standard Bayer process (Liu, Li et al. 2003). Product quality concerns (over silica and
associated soda) have prompted refinery modifications to include green liquor deep
desilication (seeding with TCA) (Bartosik, Strnad et al. 1974) (Li, Liu et al. 2003) which
again adds to the cost of production.
3.2 Combination process (Bayer-sinter)
Principle:
The combination process first uses a regular Bayer process to extract alumina values
from the high silica bauxite. Soda and alumina losses are allowed to occur in this first
stage and report to the residue. The residue is then subjected to a lime-soda sinter, in a
similar fashion to that already described. Often the exit leach solution from the second
stage provides the starting liquor for the Bayer stage with fresh bauxite.
The advantage of this combination (compared to a single lime-soda sinter) is the energy
saving associated with heating a reduced mass of residue compared to the bauxite from
which it was produced. Malts (Malts, Shmorgunenko et al. 1984) also claims that the
melt conditions for mixtures containing red mud is more advantageous than those for
bauxite. In addition, and dependent on the bauxite and process variant used, the
combination process can be used to achieve greater alumina loadings in the process

liquor (greater liquor productivity) than could be achieved by a single lime-soda sinter
process.
Status:
A good overview of the East European versions of the combination process is given in
the papers by Prokopov and Malts (Prokopov and Malts 1974) and Reisner (Reisner
1974). The main concern in recent researches is one of energy, and Liu (Liu, Aye et al.
2006) has indicated the steps taken between 1995–2000 at the Zhongzhou refinery to
reduce the energy consumption of the combination process by ~7 GJ/t Al2O3. The table
below (taken from the Liu paper) shows the relative energy intensity requirements of the
sinter process, the combined process and straight Bayer.
Table 4 Comparison of energy intensities of alumina refinery variants (from (Liu, Aye et al. 2006)).
Refinery name Production method Ore type Energy Intensity
GJ/t Al2O3.
Zhengzhou Combined Diaspore 34.15
Shanxi Combined Diaspore 37.28
Guizhou Combined Diaspore 43.31
Zhongzhou Combined Diaspore 52.17
Pinguo Bayer Gibbsite and diaspore 15.10
Shandong Sinter Diaspore 40.50
(un-named) France Bayer Diaspore 13.52
Pinjarra (Australia) Bayer Gibbsite 11.21
Shennigola (Greece) Bayer Diaspore 14.86
Straford* (Germany) Bayer Gibbsite 9.60
* This is likely to be the Stade refinery since there is no Straford refinery in Germany.
Other developments in the combination method have examined the optimisation of

materials flow (Liu, Yu et al. 2002).
Sinter and Bayer-sinter methods are generally considered to be reliable but old
technologies. They are still used because, for some bauxites, there are no alternatives
(or the Bayer alternative is even more expensive due to the loss of caustic soda).
Improvements that have been made during the past ~ten years have been aimed at
process optimisation, especially to reduce energy consumption. To be economically
competitive with Bayer processing of low silica bauxites, energy consumption would
have to be cut by at least 50% and perhaps 65%. This appears unlikely without step
change technology development.
3.3 Bauxite activation
Principle:
The Rio Tinto (Comalco) bauxite activation process involves a thermal pre-treatment
designed to make the valuable alumina phases (gibbsite and boehmite) amorphous
(Bhargava, Allen et al. 2004). The benefit of this process is that subsequent Bayer
processing can be done under less severe conditions (especially temperature). This

allows more effective extraction of the boehmitic phases, reduced operating
temperatures and costs, and the elimination of most, if not all, organic phases and the
problems they bring with them. The heat treatment is relatively low temperature (~500–
550oC) for a short time (i.e. flash calcining) and thus the kaolin is unaffected (a portion of
the kaolin might start dehydration to meta-kaolinite, but this does not significantly affect
its reactivity and desilication characteristics). This there is thus no benefit for caustic
loss minimisation from this reaction.
The benefits for high silica bauxite processing come from the reduction of digestion
temperature and the consequent reduction (or elimination) of quartz attack that would
normally occur under the temperatures commonly used to digest boehmitic bauxite.
This process can be advantageously combined with double digestion (see next section)
to further enhance the extractability of the alumina bearing minerals, since, according to
Hollitt (Hollitt, Kisler et al. 2002) there may be difference in the solubilisation of
amorphous alumina from boehmite, compared to amorphous alumina from gibbsite.
Comalco (Rio Tinto) has a number of patents over this process (Hollitt, Grocott et al.
2000a), and (Hollitt, Grocott et al. 2000b).
Status:
The bauxite activation process was developed by Comalco at the time when the plans
for the future Yarwun refinery (Gladstone, Queensland) were well advanced. It is the
author’s view that the process was considered for the refinery, but on balance the
owners decided on more proven technology 1 .
Rio-Tinto continues to develop this technology for possible use in the future.
There are many benefits to this technology, which are listed in great detail in the patents
and the AQW paper (Hollitt, Kisler et al. 2002). The only benefit directly related to
caustic consumption is the reduction in processing temperature and hence quartz attack.
A similar reduction in the high temperature digestion may be achievable using Double
Digestion technology only, but a detailed comparison would be required. An additional
hurdle surrounds the conditions for the roasting itself. Moisture control (to avoid
problems with amorphous alumina rehydration) and alumina mineral transformation
during the very small reaction times (order of 10’s of seconds) can provide issues of
stability.
3.4 Double digestion
Principle:
Double digestion is the application of the principle of counter-current digestion to the
Bayer process (Doucet 1993) (Doucet, Hendricks et al. 2002). In this technology the
bauxite and liquor streams move in opposite directions. Fresh bauxite is contacted with
liquor at ~150oC that has already been through a high temperature “boehmite digestion”.
The gibbsite from the fresh bauxite dissolves producing a highly loaded liquor. The
slurry is separated under high pressure and temperature using a pressure decanter.
1
Even though the current technology at the Yarwun refinery does include a number of new aspects, it was
considered less radical than bauxite activation.

The liquor is flashed and goes to precipitation, whereas the solids are next contacted
with spent liquor at a higher temperature to dissolve the boehmite. After this second
digestion the slurry is again separated, with the solids flashed and sent to residue, and
the liquor directed to the first digestion.
Post-desilication
Liquor to
precipitation
Gibbsite
digestion
Boehmite
Pressure
Bauxite digestion
decanter
Pressure
Spent liquor decanter
Residue
Figure 7 Double Digestion circuit schematic (adapted from Doucet (Doucet 1993)).
As with the bauxite activation technology, the key advantage of this process with respect
to caustic loss is the reduced high temperature processing. The saving in soda
consumption will clearly depend on the bauxite used, and would be suitable particularly
to boehmitic bauxites with high quartz contents.
Status:
Double digestion is proven technology that has been implemented in several refineries.
The enabling decanter technology was first developed by Alcan (Iida, Stewart et al.
1994).
It is interesting to note that although double digest technology has proven benefits, there
has been a growing trend for new (Greenfield) refineries to opt for the leaner tube
digester technology which is more energy efficient and simpler (de Boer, Edwards et al.
2005).
As a proven technology, the hurdles to implementation are those related to economics.
Retrofitting double digest technology is expensive but has been achieved recently for the
Gove refinery in the Northern Territory Australia (Valenti and Ho 2008). Although the
primary driver in this case was to recover boehmite that was increasing in their bauxite
feeds, the operation of double digestion compared to a single high temperature digest
will undoubtedly save some soda.
Two particular drawbacks that have been identified with double digestion are boehmite
reversion in the first digestion stage where the alumina loadings in liquor are high, and
solid boehmite is present, and the need for green liquor desilication (Lamerant and
Ferret 2002). Both of these disadvantages have been targeted in a patent by Rodda
(Rodda and Shaw 1996) in which the gibbsite digestion is reduced to 2-5 minutes (to
avoid boehmite reversion), followed by solid-liquid separation and green liquor
desilication. Interestingly, the Rodda patent describes similar gibbsite digestion
temperatures to Sumitomo (see section 3.7) but does not claim any reduction in reactive
silica dissolution (this may be as a result of higher caustic concentrations). The two

stage post-digestion desilication that follows does result in soda savings through the
precipitation of hydrogarnets, but the primary objective of this technology is to reduce
silica in liquor and its detrimental downstream effects.
3.5 Low soda DSPs
Principle:
The benchmark for soda content in DSP is that of Bayer sodalite. The formula
Na6[AlSiO4]6.Na2X.yH2O (where X can be a variety of inorganic anions) indicates a molar
Na2O/SiO2 of 2/3. Low soda DSPs are therefore defined to be DSPs in which the molar
soda to silica ratio is less than 2/3.
In either hydrometallurgical or pyrometallurgical methods of forming DSPs, emphasis

has been almost exclusively in producing forms of calcium (aluminium) silicates which
may or may not contain some soda. The most desirable end-products in this series are
calcium silicates in which the Ca:Si molar ratio is either 1:1 or 2:1. These DSPs contain
no soda and represent an efficient consumption of calcium. Although these products are
the goal of some thermal processes (sinter, combined process), calcium silicates have
rarely been reported to occur in hydrometallurgical processes with Bayer liquor (a
notable exception is the hydrothermal lime process, see section below).
The main non-sodalite DSPs produced within the Bayer process are the hydrogarnet
series Ca3Al2(SiO4)n(OH)12-4n, and cancrinite Na6[AlSiO4]6.2CaCO3.yH2O. Whittington
(Whittington 1996b) concludes that under hydrothermal formation conditions in Bayer
liquors n is restricted to <0.1–0.8. Iron can also substitute in the hydrogarnets, but only
at high digestion temperatures (>250oC) (Zoldi, Solymar et al. 1987). Hydrogarnets
therefore represent an effective means of reducing soda losses, but are inefficient
(compared to calcium silicates) in terms of their lime consumption (Ca:Si ratio). They
also constitute a significant loss of alumina from the liquor.
Cancrinites offer the potential of a 25% reduction in soda in DSP (compared to sodalite).
However it should be remembered that the calcium carbonate inclusion in cancrinite
effectively replaces the Na2X in sodalite. If X=sulphate or chloride ion (for example) then
although cancrinite formation reduces “soda loss”, this soda is not, in fact, useful caustic
soda (i.e NaOH or NaAl(OH)4). Cancrinite formation is only observed under high
temperature digestion conditions (Whittington 1996a).
Status:
The addition of lime to form hydrogarnets specifically to reduce soda losses has been
pioneered by Chinese researchers (Zhao, Hu et al. 2002) (Gu, Yin et al. 2002).
Additions of up to ~16% CaO on bauxite have been used (Zhao, Hu et al. 2002) which
can reduce the soda content of the residue to ~2.5 wt% (this small content was
considered to be from liquor adhered to residue and not a soda containing mineral
phase). “Lime-Bayer” is currently used in several Chinese Bayer refineries including
Shanxi and Henan (Gu 2008). Hydrogarnets are very stable under typical Bayer
digestion conditions (Ni, Goldman et al. 1979) and thus the process is quite robust.
The description of the iron-hydrogarnets goes back to the pioneering work of Ni (Ni,
Bunchuk et al. 1968), but it is the extensive work of Solymar and co-workers from
Aluterv that have promoted the use of these “catalytic” additives in high temperature
digestions for the reduction of soda losses (Zoldi, Solymar et al. 1987). The technology
was first introduced to control iron problems, especially those related to goethite to

hematite transformations (Boros, Csillag et al. 1978) but was soon after suggested to
offset the alumina losses associated with normal hydrogarnet formation (Solymar, Orban
et al. 1983). An important paper that summarises the Hungarian experience with iron
hydrogarnets for soda loss minimisation is given by Solymar et al. (Solymar, Steiner et
al. 1997). The technology is based on the production of iron hydrogarnets with the
general formula (NB compare to the hydrogarnet formula in section 1.5).
Ca3 [ Al , Fe]2 ( SiO4 ) 3− x (4OH ) x
Where ideally x=1 or 2. The theory of hydrogarnet formation and stability in these
conditions is given in the book by Ni (Ni, Goldman et al. 1979) 1 . Recently the
technology has been revisited in a patent (Medvedev, Kiselev et al. 2002) and a paper
(Medvedev, Akhmedov et al. 2003). This technology is very attractive, since if the iron
hydrogarnet can be produced with x=1 (above) then the Ca:Si molar ratio becomes 1.5:1
compared to the hydrothermal aluminium hydrogarnets (5:1). In addition the iron
hydrogarnets represent a smaller loss of aluminium. According to the paper by
Medvedev (Medvedev, Akhmedov et al. 2003) five refineries (one in Bosnia, 2 in Russia
and 2 in the Ukraine) are using the iron hydrogarnet technology.
The economic feasibility of the “Lime Bayer” process depends heavily on the availability
and cost of lime compared to caustic soda. Generally lime charges of ~16% CaO on
bauxite would not be considered to be economic outside of China. But in that country,
where lime is plentiful and cheap, the process does find acceptance, especially when its
economics are compared to, say, the combination process.
Cancrinite formation is favoured by high digestion temperatures, long digestion times

(suggesting that cancrinite can form from sodalite and/or hydrogarnets) and the
presence of lime. Unfortunately some of these conditions also favour quartz attack
which generally increases soda loss by increasing DSP production.
There is no known commercial process specifically based on soda savings from

cancrinite production. If a way could be found to encourage the formation of cancrinite
under the conditions normally used for high temperature digestion (especially to
encourage sodalite to cancrinite transformation) then there would be some merit in this
approach.
The iron hydrogarnet technology, although very attractive, also requires significant
operating costs. In particular, in the Medvedev patent (Medvedev, Kiselev et al. 2002)
there appears to be a complex internal cycle to regenerate very low A/C liquor for the
mud digestion (a similar requirement appears in the hydrothermal lime process, next
section). Also, the mud digestion requires the addition of an active iron source (sodium
ferrite) possibly to allow the reaction to occur at lower temperatures than would
otherwise be the case. It is not clear from Medvedev’s publications how or where the
sodium ferrite is produced.
1
This book is in Russian (no translation available) and the author includes it here only from references by other
authors.

3.6 Hydrothermal lime treatment
Principle:
This patented process, developed by Comalco in the early eighties 1 , digests Bayer
residue and lime at ~280oC in very low A/C liquor (low alumina) for ~20 minutes. These
conditions lead to the formation of calcium silicates (mostly di-calcium silicate) which are
stable in Bayer liquor and efficient at saving caustic. The process is therefore a
hydrothermal equivalent of the combination process.
The problems that are immediately apparent with this technology is that the re-digestion
requires the Bayer digested slurry to be flashed to atmospheric pressure, deliquored, re-
slurried in low A/C liquor and heated up to the second digest temperature. The cost of
this step, along with the requirement for a high caustic low A/C stream (which Comalco
suggested could made by salting out NaAl(OH)4 from spent liquor) made the process
unattractive at the time (see Figure 8).
Status:
The process was patented by Comalco in 1982 (Cresswell and Milne 1982a) and
described by Cresswell in the Bayer literature (Cresswell and Milne 1982b) (Cresswell
and Milne 1984). There does not appear to be any development or use of the process
(at least none published outside Comalco/RioTinto) since this time.
Several years after the Comalco patent, Sagin (Sagin 1988) described a similar
hydrothermal process designed to replace residue sintering in the combined process.
The advance that Sagin proposed was to design the lime digestion to encourage the
formation of the desilication product Na2.Ca2Si2O5.(OH)2 which, it is claimed, is superior
to di-calcium silicate in two important ways – firstly it requires half the lime charge, and
secondly it can be easily re-processed to recover soda without returning silica to the
process. A similar desilication product has also been mentioned by Rayzman and
Fillipovitch in the pyrometallurgical equivalent – “High Alkaline Sintering Method”
(Rayzman and Fillipovich 1999).
At the time that the Comalco and Sagin processes were developed double digestion was
presumably not available (or at least not in commercial use). Double digestion
technology could be used to save considerably on energy costs of re-digestion. The
problem of how to create a high caustic low A/C stream still remains however, as well as
how to integrate this process into a regular Bayer cycle (particularly washing).
1
The original idea appears in a paper by Klan (Klan 1971).

Bauxite
Bayer
Clarification
digestion
Flash Flash
cooling cooling
Settling Washing Lime digestion
Sod. aluminate
Crystallisation Lime
dissolution
Bayer
Evaporation
precipitation
Calcination Alumina
Figure 8 Process flowsheet for Hydrothermal lime treatment of residue and its incorporation into the Bayer
cycle (after Cresswell 1984).
3.7 Differential extraction processes
Principle:
The main principle of these processes (for the processing of high silica gibbsitic ores) is
exploiting the difference in extraction rates of gibbsite and kaolin. In the most recent
variation (the Sumitomo process), at low temperature digestion (130oC), gibbsite can be
almost totally extracted into Bayer liquor at good loadings (A/C~0.69) in 2–5 minutes. In
the same amount of time, only a fraction of reactive silica (kaolin) dissolves and almost
none of this has precipitated into DSP (Figure 9).

100
Al2O3
extraction
Reaction Rate (%)

Re.SiO2
extraction
Caustic loss
50 to DSP
Sumitomo
Digestion Conventional
Digestion
0
2-5 30 - 60
Time (minutes)
Figure 9 Schematic of the principle of the Sumitomo process (adapted from Harato et al. (Harato, Ishida et al.
1996)).
After this short digestion, the slurry is rapidly solid-liquid separated using a pressure
decanter, and the kaolin-containing residue is subjected to high rate washing to prevent
further reactive silica transformation to DSP. The pregnant liquor has a high
concentration of dissolved silica which necessitates a seeded, post-digestion desilication
prior to hydrate precipitation.
When operated under “careful control” the process is claimed to reduce caustic
consumption by ~50% compared to a conventional Bayer process (Harato, Ishida et al.
1996).
Status:
The original work in this area was done by Takahashi (Takahashi 1962) who considered
the use of two bauxites, one high silica and one low silica. The high silica bauxite is
digested under conditions of low temperature (80–105oC i.e. atmospheric digestion) and
solid-liquid separated under conditions to maximise the amount of undigested reactive
silica in the residue. The liquor from the atmospheric digestion is then used to digest
further low silica bauxite in a conventional digestion. The second digestion therefore
increases the alumina loading and desilicates the liquor.
Other variations are by Grubbs (Grubbs 1986) and Iwase (Iwase and Murase 1988).
They found that when bauxite is ground and separated into fine and coarse fractions, the
fine fraction is richer in silica and poorer in alumina than the coarse fraction. The
principle of Takahashi can then be applied to a single bauxite, separated into two
fractions.
Improvements to this concept (principally in the area of solid-liquid separation) were

provided by Oku (Oku and Suzuki 1973), Iwase (Iwase 1987), where the slurry was
flash-cooled before separation and by Harato (Harato, Ishida et al. 1993) where a
pressure decanter was used. The Sumitomo variation (Harato, Ishida et al. 1993) was
operated continuously from 1994–2000 at the Ehime refinery, but was discontinued
when the company strategy changed, and the refinery was used to produce specialty
alumina.
Fulford (Fulford, Chinloy et al. 1991) also patented a version of this process using
double digestion technology in which it is claimed the second digestion stage in the
double digester can be used as the post-digestion desilicator. Fulford does not claim
soda savings, but it is clear that the application of this technology to a Sumitomo type
process would be advantageous.
Banvolgyi (Banvolgyi and Siklosi 1998) proposed another variation in that gibbsite is
charged to very high Al2O3 loadings (close to the solubility of the gibbsite) at which the
free hydroxide concentration in the liquor is very low. This free hydroxide concentration
is the driver for not only gibbsite dissolution but also kaolin dissolution, and hence low
free hydroxide concentrations further retard kaolin dissolution. The high A/C (alumina
loadings) produced by this variation would normally cause problems with alumina
reversion during solid-liquid separation and even though the digestion time is not so
critical as with the Sumitomo version, Banvolgyi recommends pressure decantation be
conducted at or near digestion temperature. The process is patented (Banvolgyi, Zoldi
et al. 1992) and the theory behind the process has been described (Banvolgyi 1992).
The Sumitomo process is restricted to gibbsitic bauxite ores because of the low
digestion temperatures. There are a number of potential problems with all versions of
this process including the operation of the pressure decanters, especially the
effectiveness of flocculants, the operation of the high rate washers, and the
effectiveness of the post-digestion desilication, especially aspects of seeding (Harato,
Ishida et al. 1996). However, developments in these operations over the period since
this technology was introduced are likely to provide at least partial solutions. Of
particular importance would be the combination of modern double digest technology with
the Sumitomo variation.
The Sumitomo process remains the only commercially used hydrothermally-based

Bayer process variation specifically designed to reduce caustic consumption. As such it
would be one of the process options with the highest potential.

4. Potential Technologies – Strategy Three - Soda Recovery Processes
4.1 Lime sinter of DSP
Principle:
There are several variations of the Bayer process in which a DSP concentrate is formed
from the precipitation of DSP from a silica saturated liquor, either before or after
digestion. The DSP concentrate can then be re-processed to recover caustic.
Reprocessing of the DSP can be by the use of any sinter process of known technology
(e.g. the lime-soda sinter). In this respect the process is similar to the combination
process, but it is differentiated from it by the use of a DSP concentrate in the sinter stage
rather than sintering the whole residue solids.
Status:
In the various processes already described in the sections on differential extraction and
double digestion, post-digestion desilications will provide a “pure” DSP concentrate.
Some of this will be returned to seed the desilication, but it also provides a source of
DSP concentrate for re-processing.
In a process patented by Comalco (Hollitt, Crisp et al. 1998) the whole of the spent
liquor stream is used to treat the raw bauxite with the aim of completely dissolving the
reactive silica. The solids charge is designed to ensure that the resulting level of silica in
liquor is not high enough to trigger the precipitation of DSP at the particular temperature
of the dissolution (~100oC). After dissolution, the slurry is solid-liquid separated. The
high silica liquor is seeded with DSP and desilicated, producing a DSP concentrate
which is separated. The liquor is recombined with the silica leached bauxite and sent to
regular Bayer processing. There is no evidence of any commercial application of the
Comalco technology.
The Comalco patent appears to have a number of antecedents, in which the silica is
extracted into spent liquor before digestion – all are Japanese. A selection are:
(Kanehara, Morishige et al. 1983) (Kokoi, Sakamoto et al. 1993) and (Tanjo, Yudate et
al. 1994). The objective in all of these works appears to be the production of Bayer
residue low in soda (and silica) and higher in iron for re-use in the cement industry. In
most cases there is no indication of the use of the DSP concentrate to recover soda.
The benefit of this process compared to the combination process would be the energy
savings associated with re-processing a 100% active concentrate rather than a total
solids residue. However the extra capital and operating costs associated with the
production of the concentrate would not be expected to be economically favourable
when compared to the conventional Bayer process with soda losses included.
These economics are significantly influenced by the economics of the sinter, the
products of the sinter, and how easily a concentrated caustic stream can be produced
and returned to the Bayer process. A paper by Rayzman (Rayzman and Fillipovich
1999) summarises a particular VAMI technology designed to reduce sinter costs by
using coal in the sintering process of high iron bauxites. However this relatively dirty fuel
source brings its own environmental problems.

4.2 Weak acid leaching of DSP
Principle:
Weak acids can be used to leach or dissolve soda from sodalite DSP. CO2 reaction with
sodalite can leach the caged soda from DSP (caged soda represents 25% of the total
sodalite soda). The equation for this process for hydroxide inclusion (hydroxy-sodalite)
has been given by Cooling et al. (Cooling, Hay et al. 2002) -
Na6 [ AlSiO4 ]6 .2 NaOH + 2CO2 ↔ Na6 [ AlSiO4 ]6 + 2 NaHCO3
The mobility of the soda is significantly dependent on the included anion, and since
Bayer sodalite from high silica refineries is more likely to contain aluminate and
carbonate, this leaching may not be as efficient as with hydroxy-sodalite (Smith, Wingate
et al. 2008). It is not known whether H+ in carbonic acid could ion-exchange for sodium
which is part of the cage structure itself.
Treatment with a stronger (sulphurous) acid can dissolve the DSP itself liberating silica
into the acid solution. Amorphous silica can be precipitated from the leach solution,
which is then causticised with lime to produce a caustic stream (recycled to the refinery)
and solid calcium sulphite (which is calcined to regenerate CaO and SO2). The process
has been patented by Comalco (Cresswell, Grayson et al. 1986) (Cresswell 1992).
Status:
Carbonation of residue has been developed over a number of years by Alcoa in
Australia, primarily to ameliorate the residue (both chemically and physically) to improve
red mud disposal practices. The release of soda from sodalite by this process is
considered to be part of the neutralisation process and its recovery is not a priority.
There is little in the literature on the dissolution of DSP with sulphurous acid, but the use
of residue to scrub SO2 from stack gases has received attention (Yamada, Fukunaga et
al. 1979) (Yamada and Harato 1982) (Leoni and Penco 2002). The reactions produce
both sodium sulphite and sodium bisulphite which are rapidly oxidised to sodium
sulphate. It is unclear what happens to the silica phase, although a spinel phase is most
likely. In a soda recovery process the challenge would be to separate the sodium
sulphate and causticise it to recover caustic value.
It is unlikely that a soda recovery process based on carbonic acid leaching could be
viable. The maximum recovery of soda is ~25% and the leaching is unlikely to produce
a caustic stream that is concentrated enough to return to the refinery.
Recovery of soda by sulphurous acid leaching however targets all of the DSP soda
although 100% recovery is never achieved (Yamada and Harato 1982). Causticisation
by lime and the calcination of CaSO3 would be significant process costs. Despite these
shortcomings, this process has potential.

4.3 Complex causticisation (mud causticisation)
Principle:
Complex causticisation was a term coined in the 1980’s by Hungarian researchers to
describe a two stage hydrothermal soda recovery process using lime (Baksa, Vallo et al.
1986) (Baksa, Boros et al. 1984). In the first stage lime is added to a residue slurry at
~95oC. The lime reacts firstly with alumina in solution to form hydrogarnets (or TCA
depending on the amount of silica in solution). Thereafter the lime attacks solid Bayer
sodalite, liberating soda, alumina and silica. Lime combines with the released alumina
to form more hydrogarnet, and, when the alumina is exhausted, lime reacts with silica in
liquor to form (amorphous) calcium silicates. Soda, once associated with the sodalite, is
liberated as useful caustic. To minimise the lime reacting with liquor alumina, this
process is usually performed on last washer underflow where liquor strength is low.
However this also means that the caustic produced from the breakdown of the sodalite
is also of low strength, and needs significant evaporation to be useful.
The solids from the first stage are re-slurried in a concentrated solution of sodium
carbonate. The high carbonate concentration causes the breakdown of hydrogarnets to
form calcite (CaCO3) liberating sodium aluminate which can also be returned to the
process. Apparently the amorphous calcium silicate produced in the first stage is stable
through the second stage and does not revert to sodalite. An optimisation of this second
stage has been described by Liu (Liu, Zhang et al. 2004).
Washer (n) Washer (n+1)
Soda
Lime
solution
Slurry
heater
Figure 10 Schematic of the complex causticisation of red mud residue - adapted from Baksa (Baksa, Vallo et
al. 1986).
Status:
According to “Alumina refineries and producers of the world”, this process has been
used commercially in the Hungarian refinery at Ajka from the mid 80’s until at least 1998
(Aluminium-Verlag 2007). There is no information on whether other installations use this
process.
Hurdles to implementation
There are several hurdles to implementation with this process. The first is high lime
consumption, in that significant lime is used to scavenge alumina in solution before the
DSP can be broken up, and this detracts from lime efficiency. Additionally the recovery
of caustic from the first stage depends heavily on the efficiency of washing and thus the
integration of the process into the back-end of the Bayer process is critical.

Another potential problem is that the second stage needs a reasonably concentrated
stream of sodium carbonate. Local trona (sodium carbonate ore) has been suggested
by the inventors for this purpose, but of course this is not available to all refineries. The
production of such a stream from within the Bayer process itself is not seen as
economical.
4.4 Mechano-chemical treatment of residue
Principle:
The recovery process is based on the principle of ultra-fine grinding of residue. The
mechanism of soda recovery is not clear, but as with similar grinding of the bauxite
(section 2.6) it is thought to be the energy of the collisions (ore with ore or ore with
grinding media) that produces phase changes that would not normally occur without the
grinding.
Picaro (Picaro 2000) claims that the ultra-fine milling of red mud residue (with or without
additives such as lime) can result in caustic recovery from DSP. Alumina is also
recovered but it is not clear whether this comes from DSP or from unextracted or re-
precipitated alumina values. Recoveries of ~40% of the soda from DSP are claimed
without additives, and up to 70% recovery for residue milling with lime.
The mechanism for soda recovery is not clear and has not been disclosed. Since XRD
patterns of the sodalite become less well defined after milling, the speculation in the
patent was that the soda is leached through a partial breakdown / amorphitisation of the
sodalite structure.
Status:
There does appear to be any disclosed further work since the patent.
The process does not appear to be tested under rigorous conditions. This needs to be
done and the soda recoveries confirmed. There is a concern that if silica is released into
solution that might trigger the re-formation of DSP which would be counter-productive,
and so the fate of the silica is important to determine.
The cost of ultra-fine grinding must be weighed up with the benefit of soda recovery. In
this respect, the beneficiation of residue to produce a DSP concentrate (or the
production of pure DSP, see section 4.1) or the use of new grinding technology may be
advantageous.

5. Summary and Conclusions
5.1 Strategy One options
Options in this strategy are concerned with beneficiation of reactive silica or bauxite pre-
treatments. They are, by their nature, largely dependent on the mineralogy, liberations
and associations of the minerals in the bauxite ore. As such, they will find specific
applications for particular bauxites; for example the washing and screening of Weipa
pisolites and the flotation of kaolin from diasporic ores in China.
There is no beneficiation process that will be suitable for bauxites in general. However,
the particle size distribution for quartz in bauxite is almost always coarser than for the
alumina and alumino-silicate minerals such as gibbsite and kaolin. Therefore, removal
of quartz from bauxite (by flotation or another beneficiation technique) could be a
general strategy to reduce reactive silica input for boehmitic bauxites.
Two pre-treatment options that are relatively independent of mineralogical associations

are the mechano-chemical treatment of bauxites, and the roast-leach pre-treatment.
Both these options need to be examined in more detail, especially with respect to cost
and the economics of the soda saved.
5.2 Strategy Two options
Options within this strategy are less bauxite dependent than those in strategy one.
Both sinter and combination processes are robust, reliable technologies that have been
used extensively to process high silica ores for many years in the former Eastern Block
and in China. Although they have been the subject of much research with respect to
reducing their energy consumption, it is clear that they can never be reduced to the level
to compete with standard Bayer processing of high quality ores.
Both bauxite activation and double digestion (although different process variations) have
the same objective with respect to high silica bauxite processing, i.e. they both allow the
digestion of bauxite at temperatures sufficiently lower than normal for boehmitic ores, to
reduce the degree of quartz attack (and therefore soda loss). These are variations
which are known to work, and are currently available. Although bauxite activation has
other benefits (e.g. organics destruction) it would not provide any extra benefit with
respect to soda loss, and therefore would not be as attractive as double digestion. It
should be pointed out that although these process variations can clearly save soda, their
cost/benefit needs to be established.
Low soda DSPs potentially offer the most exciting prospect with respect to soda saving.
However sodalite and cancrinite (for the sodium alumino-silicates) and hydrogarnets (for
the calcium aluminium silicates) appear to be overwhelmingly the most stable
desilication products. It is likely that in the unpublished literature of the alumina
companies and other research organisations has examined different alumino-silicate
mineral structures, and other DSP cations to encourage DSP phases other than those
mentioned above. However since there is no disclosed advances in this field, it is also
likely that these phases are not stable enough under Bayer processing conditions to be
useful. If this is the case, then the opportunities for low soda DSPs appear to be
restricted to cancrinite and (iron) hydrogarnets.

Hydrothermal lime treatment (the hydrothermal version of the combination process)
shows great potential, in that the conditions needed to produce calcium silicates (some
of the most efficient desilication products) are not far from those used in the Bayer
process. The temperatures needed (~280oC) may have been unusual when the process
was being developed, but are reasonable now (the new Rio Tinto refinery at Yarwun
uses tube digestion at ~270oC). In addition, double digestion technology (and pressure
decantors) could make the second (lime) digestion reasonably energy efficient. A
remaining problem would the supply of a high caustic, low alumina stream for this
digestion.
The Sumitomo (and related) processes provide an excellent option for the processing of
high silica gibbsitic bauxites. The hurdles to the use of such a process, including
engineering problems associated with the operation of high-rate washers, pressure
decantors are likely to have diminished with improvements in technology and control
strategies over the last 10 years.
5.3 Strategy Three options
Options in this category are also less dependent on bauxite, and have the added benefit
of being somewhat separate from the workings of the Bayer process itself. As such,
some could be treated as “swing processes”, i.e. used only at times that justified their
use (e.g. when caustic prices are high), or for other operational reasons detailed below.
There appears to be little scope to justify the concept of a pre-digestion desilication

stage on the economics of soda recovery by lime sinter alone. However this process
results in a Bayer residue almost free of non-quartz silica and soda which would have
some environmental benefits (legacy and potential future licence to operate issues).
Recovery of soda from the use of weak acids such as the aqueous forms of CO2 and
SO2 with residue are possible but, in common with the lime sinter of DSP processes, are
likely to have bigger benefits from the reduction of residue alkalinity. Of the two, the SO2
process is the most likely to be useful in terms of returning soda to the process.
The chemistry of the complex causticisation process (mud causticisation) process has
never been adequately explained (especially the fate of the silica in the second stage).
However it is unlikely that the lime efficiency in this process could be elevated to a point
of being economical without a step change in the technology (especially the recovery of
alumina in the second stage with normal Bayer process streams). The technology is
appealing though, especially since it is independent of both the bauxite and the process
itself (it would essentially be a “bolt-on” rear end process applicable to most operations).
It is possible that the economics could change if this process was combined with another
to produce a DSP concentrate.
Ultra-fine grinding of the mud would need more evaluation to be dismissed, but at
present there is no clear indication of the fate of the other DSP components (especially
silica) if and when soda is recovered. The stability of the changed desilication product
(after soda removal) is critical in determining the robustness of any soda recovery
process.

6. Acknowledgements
The author would like to thank Takuo Harato (CSIRO Minerals) for valuable comments.
This work was undertaken as part of a project receiving funding from the Australian
Government as part of the Asia-Pacific Partnership on Clean Development and Climate.
Permission to publish is gratefully acknowledged.
The views expressed herein are not necessarily the views of the Commonwealth, and
the Commonwealth does not accept responsibility for any information or advice
contained herein.

Aluminium-Verlag (2007). Alumina refineries and producers of the world, Aluminium-
Verlag, Dussseldorf.
Amer, A. M. and E. A. Abdel-Aal (1996). "Mineralogical studies and hydrometallurgical

processing of Egyptian bauxite." Transactions of the Indian Institute of Metals 49(1-2):
25-28.
Anand, P., J. M. Modak, et al. (1996). "Bio-beneficaition of bauxite using bacillus

polymyxa - calcium and iron removal." International Journal Mineral Processing 48(1-2):
51-60.
Andrew, P. I., N. M. Anishchenko, et al. (1973). "Mechanism of the anionic flotation of

chamosite and gibbsite." Tsvet. Metall. 116 (6): 16-20.
Arlyuk, B. (1995). "Comparative Evaluation of Efficiency of Using Bauxite and Nepheline

Ores for Production of Alumina". Light Metals, The Minerals, Metals & Materials Society,
420 Commonwealth Dr , Warrendale, Pennsylvania 15086, USA: 121-131.
Baksa, G., J. Boros, et al. (1984). "Process for the Reduction of the Sodium Hydroxide
losses of the Bayer-type alumina production." Patent US4486393
Baksa, G., F. Vallo, et al. (1986). "Complex Causticization: An Effective Means for the
Reduction of NaOH Losses in an Alumina". Light Metals, New Orleans, LA, USA,
Metallurgical Soc of AIME, Warrendale, PA, USA.
Banvolgyi, G. (1992). "Reactions of gibbsite and kaolinite in the Bayer liquor: a

comprehensive kinetic model and an improvement of the low temperature digestion." 7th
International Congress for Study of Bauxites, Alumina and Aluminium.: 1-16.
Banvolgyi, G. and P. Siklosi (1998). "The improved low temperature digestion (ILTD)
process: an economic and environmentally sustainable way of processing gibbsite
bauxites". Light Metals. San Antonio, Texas, The Minerals, Metals & Materials Society,
420 Commonwealth Dr , Warrendale, Pennsylvania 15086, USA: 45-53.
Banvolgyi, G., J. Zoldi, et al. (1990). "Prodn. of alumina from gibbsite bauxite - by
digestion in caustic- aluminate liquor to produce liq. phase with controlled reactive
hydroxyl ion content" Patent AU8943731-A. ; CA2000832-A (27 Apr 1990); BR8905475-
A (29 May 1990); ES2019173-A (01 Jun 1991); US5122349-A (16 Jun 1992);
CA2000832-C (11 May 1993)
Bardossy, G. (1982). Karst bauxites. Amsterdam, Elsevier.
Bardossy, G. and G. J. J. Aleva (1990). Lateritic Bauxites. Amsterdam, Elsevier.
Bartosik, L., K. Strnad, et al. (1974). "Technical and economic problems in the
processing of low grade bauxites for the production of aluminium oxide." Travaux
ICSOBA(12): 175-194.
Bhagat, R. P., B. Banerjee, et al. (2001). "Beneficiation tests on an Indian bauxite

incorporating magnetic separation." Transactions of the Institution of Mining and
Metallurgy Section C-Mineral Processing and Extractive Metallurgy 110: C165-C168.

Bhargava, S., M. Allen, et al. (2004). "Thermal activation of bauxite." Chemistry in
Australia 71: 6-8.
Boros, J., Z. Csillag, et al. (1978). "Catalyst for bauxite digestion - for the conversion of
goethite to haematite." Patent DE2754289-A; JP53071700-A; FR2373489-A; HU15955-
T; BR7800124-A; SU691073-A; US4226838-A; JP81013651-B; GB1597006-A;
CA1114173-A; RO76285-A; DE2754289-C; IT1087523-B
Bray, E. L. (2008). Mineral Commodity Summaries, US Geological Survey.
Brown, I. W. M., K. J. D. MacKenzie, et al. (1985). "Outstanding problems in the

kaolinite-mullite reaction sequence by 29Si and 27Al solid-state nuclear magnetic
resonance; II, high-temperature transformations of metakaolinite." Journal of the
American Ceramic Society 68(6): 298-301.
Chakraborty, A. K. and D. K. Ghosh (1978). "Re-examination of kaolinite-to-mullite

reaction series." Journal of the American Ceramic Society 61(3-4): 170-173.
Chen, X. and W. Li (2006). "Parallel floatation process of desiliconization for bauxite"

Patent CN1869258-A
Cooling, D., P. S. Hay, et al. (2002). "Carbonation of bauxite residue". Sixth International
Alumina Quality Workshop. Brisbane, Qld: 185-190.
Cresswell, P. J. (1992). "Chemical processing of bauxite tailings". International Bauxite

Tailings Workshop, Perth.
Cresswell, P. J., I. L. Grayson, et al. "Values recovery from bayer process red mud - to
obtain directly recyclable sodium aluminate soln" Patent DE3545388-A; FR2575149-A;
AU8551016-A; US4668485-A; DE3545388-C; IT1188213-B
Cresswell, P. J. and D. J. Milne. "Alumina and soda recovery esp. from red mud - by
digestion in presence of lime" Patent AU8288102-A; US4483830-A AU8288102-A 17
Mar 1983 198317.
Cresswell, P. J. and D. J. Milne (1982b). "A Hydrothermal Process for Recovery of Soda
and Alumina From Red Mud." Light Metals, Minerals, Metals and Materials Society
(TMS), 184 Thorn Hill Road, Warrendale, PA, USA: 227-238.
Cresswell, P. J. and D. J. Milne (1984). "Hydrothermal recovery of soda and alumina

from red mud". Light Metals, Minerals, Metals and Materials Society (TMS), 184 Thorn
Hill Road, Warrendale, PA, USA: 211-221.
de Boer, D. J., M. Edwards, et al. (2005). "Design, start-up and operational aspects of
the new digestion process at Comalco alumina refinery, Gladstone, Queensland".
Seventh International Alumina Quality Workshop. Perth, W.A.: 32-36.
Doucet, J. (1993). "Double digestion: Technology that leads towards quality and
efficiency". Third International Alumina Quality Workshop. Hunter Valley, NSW: 93-101.
Doucet, J., C. Hendricks, et al. (2002). "Pressure decantation technology: the Kaiser
Gramercy experience". Sixth International Alumina Quality Workshop. Brisbane, Qld: 94-
99.

Ehrlich, H. L. (1992). "Metal extraction and ore discovery". Encyclopedia of Microbiology,
Academic Press. 3: 75-80.
Eremin, N. I. and N. S. Shmorgunenko (1974). "Some theoretical and practical questions

of low-grade bauxite processing by the sinter process and the in-series combined Bayer-
sinter process." Travaux ICSOBA(12): 213-222.
Fortin, S. and G. Forté (2007). "Mechano-activated bauxite behaviour". Light Metals.

Orlando, Florida, Minerals, Metals and Materials Soc (TMS), Warrendale, PA, USA: 87-
92.
Freyssinet, P. H., C. R. M. Butt, et al. (2005). Ore forming processes related to lateritic
weathering. Colorado, SEG.
Fulford, G. D., D. R. Chinloy, et al. (1990). "Bayer alumina prodn. with has enhanced red
mud sepn. stage - carried out at temp. above slurry liq. phase atmospheric boiling temp."
Patent EP382383-A2; EP382383-A (16 Aug 1990); AU9048782-A (09 Aug 1990);
BR9000421-A (15 Jan 1991); US4994244-A (19 Feb 1991); CN1045254- A (12 Sep
1990); EP382383-A3 (04 Mar 1992); CA1330865-C (26 Jul 1994); EP382383-B1 (08
May 1996); ES2086368-T3 (01 Jul 1996); IE81147-B (03 May 2000)
Gao, S., X. Li, et al. (2008). "Beneficiation of low-grade diasporic bauxite with
hydrocyclone." Transactions of Nonferrous Metals Society of China 18(2): 444-448.
Groudeva, V. I. and S. N. Groudev (1983). "Bauxite dressing by means of Bacillus

circulans." Travaux (ICSOBA) 13(18): 257-263.
Grubbs, D. K. (1986). " Parallel purification of alumina with physical pretreatment -

Alumina Extraction From Low Grade Ores." Patent US4614641
Gu, S. (2008). "Chinese Alumina Industry - Development, Prospect and Challenge in the
Future". Eighth Alumina Quality Workshop Darwin, N.T.
Gu, S., Z. Yin, et al. (2002). "Intensifying method of Bayer digestion process of diasporic
bauxite in China". Light Metals Seattle, WA, USA, Metals and Materials Society/AIME
Warrendale, PA: 83-88.
Harato, T., T. Ishida, et al. (1996). "The development of a new Bayer process that
reduces the desilication loss of soda by 50 % compared to the conventional process".
Fourth International Alumina Quality Workshop, Darwin, N. T.
Harato, T., T. Ishida, et al. (1993). "Process for producing aluminium hydroxide from
alumina-containing ore" Patent EP564659-A; WO9309062-A; WO9309062-A1;
JP5193931-A; AU9227992-A; EP564659-A1; BR9205404-A; AU658235-B; EP564659-
A4; US5545384-A; EP564659-B1; DE69228865-E; CA2099279-C; JP3438240-B2
Hill, V. G. and R. J. Robson (1981). "The Classification of Bauxites from the Bayer Plant
Standpoint." Light Metals: 15-27.
Hollitt, M., S. Grocott, et al. (2000). "Treatment of an alumina process feedstock includes
controlling the contact time of the solid alumina feedstock during heating" Patent
WO200010919-A. ; WO200010919-A1 (02 Mar 2000); AU9954977-A (14 Mar 2000);

AU727363-B (14 Dec 2000); EP1109740-A1 (27 Jun 2001); CN1324328- A (28 Nov
2001); JP2002523326-W (30 Jul 2002); US6616902-B1 (09 Sep 2003)
Hollitt, M., J. P. Kisler, et al. (2002). "The Comalco bauxite activation process". Sixth
Alumina Quality Workshop. Brisbane: 115-122.
Hollitt, M. J., A. J. Crisp, et al. (1998). "Removal of silica from Bayer process bauxite
feedstock - by contacting with caustic liquor then separating the silica bearing liquor from
the solids" Patent WO9822390-A. ; EP950020-A; WO9822390-A1 (28 May 1998);
AU9749357-A (10 Jun 1998); EP950020-A1 (20 Oct 1999); AU716537-B (24 Feb 2000);
CN1242756-A (26 Jan 2000); JP2001503727-W (21 Mar 2001); US6309615-B1 (30 Oct
2001)
Hollitt, M. J., S. Grocott, et al. (2001). "Treating alumina process feedstock by heating
via direct contact with combustion gas, cooling, and controlling the contact time to
ensure tri- and monohydrate alumina decomposition" Patent WO200018685-A. ;
WO200018685-A1 (06 Apr 2000); AU9963206-A (17 Apr 2000); AU725611-B (12 Oct
2000); EP1135333-A1 (26 Sep 2001); CN1320102- A (31 Oct 2001); JP2002525261-W
(13 Aug 2002); US6582670-B1 (24 Jun 2003)
Hu, Y. H., K. Ouyang, et al. (2008). "Flotation of kaolinite and diaspore with hexadecyl
dimethyl benzyl ammonium chloride." Journal of Central South University of Technology
15(3): 378-381.
Iida, F., F. Stewart, et al. "Pressure decanter for solid separation from a liq. slurry - with
an incoming pressurised liquid more efficiently separated without depressurisation"
Patent EP696225-A; WO9426383-A; WO9426383-A1; AU9466427-A; US5407561-A;
FI9505329-A; BR9406391-A; EP696225-A1; EP696225-B1; DE69401206-E;
ES2096468-T3; AU681657-B; CN1122581-A; HU76198-T; HU215983-B; CA2160346-C;
KR329711-B; CN1051244-C
Ishchenko, V. V., V. M. Korus, et al. (1974). "Flotation of silica from bauxite." Tsvet
Metall. 17(3): 7-11.
Iwase, H. and A. Murase (1988). "Alumina extn. in Bayer process without sol. silica extn.
- includes classifying bauxite in fine and coarse granules, treating in sodium aluminate,
sepg. residue etc" Patent JP63190709-A; JP93001212-B
Iwase, M. (1987). "Prodn. of aluminium hydroxide from bauxite - by extracting alumina

with soda aluminate, using tubular reactor, flashing resultant mixt., quickly cooling etc"
Patent JP62230613-A
Kanehara, M., H. Morishige, et al. (6 May 1983). "Modification method of red mud"
Patent JP S62-230613-A
Klan, P. (1959). "Removal of silicon oxide from roasted bauxites by caustic soda" Patent
PV 5373-59
Kokoi, H., A. Sakamoto, et al. (9 Jul 1993). "Production method of red mud containing
low soda" Patent JP H05-170434-A

Lamerant, J.-M. and Y. Ferret (2002). "Boehmitic reversion in a double digestion process
on a bauxite containing trihydrate and monohydrate". Light Metals Seattle, WA, USA,
Metals and Materials Society/AIME: 181-184.
Leoni, F. and C. Penco (2002). "Bauxite residue desulphurisation system (BRDS) at

Eurallumina". Sixth International Alumina Quality Workshop: 162-167.
Li, W., P. Song, et al. (2002). "A new process technology of bauxite concentration-tube
digestion for alumina production". Light Metals Seattle, WA, USA, Metals and Materials
Society/AIME Warrendale, PA: 95-100.
Li, X. B., X. Liu, et al. (2003). "Study and application of intensified sintering process for
alumina production." Chinese Journal of Non-ferrous metals 14(6): 1031-1036.
Li, X. B., X. M. Liu, et al. (2004). "Study and application of intensified sintering process
for alumina production." Chinese Journal of Nonferrous Metals 14(6): 1031-1036.
Li, Y. and D. Chen (1979). "Beneficiation of diaspore-kaolinite bearing bauxite by

flotation." Jinshu Xuebao/Acta Metallurgica Sinica 15(3): 319-322.
Lin, I. J., N. Malts, et al. (1998). "The complex chemical treatment of alumina-silica-
containing materials." Journal of Materials Synthesis and Processing (USA) 6(1): 27-35.
Liu, G., X. Li, et al. (2003). "Stability of calcium silicate in basic solution." Transactions of
Nonferrous Metals Society of China 13(5): 1235-1238.
Liu, G., Y. Zhang, et al. (2004). "Alumina recovery from sodium hydrate alumino-
silicate." Chinese Journal of Nonferrous Metals 14(3): 499-503.
Liu, L.-R., Q.-B. Yu, et al. (2002). "Influence of material flow in alumina manufacturing
process with Bayer-sinter combination method on its energy intensity." Dongbei Daxue
Xuebao/Journal of Northeastern University 23(10): 944-947.
Liu, L., L. Aye, et al. (2006). "Analysis of the overall energy intensity of alumina refinery
process using unit process energy intensity and product ratio method." Energy 31(8-9):
1167-1176.
Loginova, I. V., V. N. Koryukov, et al. (1979). "Effect of some flotation reagents on the
digestion of bauxites." Soviet non-ferrous metals research 7: 634-637.
Lu, S., R. L. Fang, et al. (2003). "Study on the extraction of aluminium oxide from low
grade bauxite ore by an environmentally friendly method." Chemical Industry
Engineering Progress 22(3): 391-394.
Malts, N. S., N. S. Shmorgunenko, et al. (1984). "Production of alumina by combination

Bayer-sintering method". Bauxite, Los Angeles, CA, USA, Soc of Mining Engineers of
AIME, New York, NY, USA.
McCormick, P. G., T. Picaro, et al. (2002). "Mechanochemical treatment of high silica

bauxite with lime." Minerals Engineering 15(4): 211-214.
Medvedev, V. V., S. N. Akhmedov, et al. (2003). "Hydrogarnet technology of processing

the bauxite raw material as up-to-date alternative to Bayer-sintering technique."
Tsvetnye Metally(11): 58-62.

Medvedev, V. V., A. I. Kiselev, et al. (2002). "Method for hydrochemical processing of
aluminosilicate feedstock" Patent RU2193525-C1
Mishra, B., A. Staley, et al. (2002). "Recovery of value-added products from red mud."
Minerals & Metallurgical Processing 19(2): 87-94.
Misra, C. (1986). Industrial Alumina Chemicals. Washington, American Chemical

Society.
Ni, L. P., L. V. Bunchuk, et al. (1968). "A study of hydrogarnets formed in the Al2O3-
CaO-Fe2O3-Na2O-SiO2-H2O system." Russian Journal of Inorganic chemistry 13(11):
1585-1587.
Ni, L. P., M. M. Goldman, et al. (1979). Processing of iron-rich bauxites, Moscow

Metallurgica.
O'Connor, D. J. (1988). Alumina Extraction from Non-Bauxitic Materials. Dusseldorf,

Aluminium-Verlag.
Oku, T. and A. Suzuki (1973). "Sepn of aluminates - from slurries,in prodn of alumina"
Patent JP73037678-B1; GB1307316-A; US3716617-A; JP73037678-B
Owada, S., D. Okajima, et al. (2005). "Two approaches for reducing wasted 'red mud':
possibility of upgrading bauxite and 'red mud'". Seventh International Alumina Quality
Workshop. Perth: 205-209.
Pawlek, F., M. J. Kheiri, et al. (1992). "The leaching behaviour of bauxite during
mechano-chemical treatment". Light Metals, Minerals, Metals and Materials Society
(TMS), 184 Thorn Hill Road, Warrendale, PA, USA: 91-95.
Picaro, T. "Red mud processing" Patent EP880467-A; WO9729992-A; WO9729992-A1;

AU9715853-A; EP880467-A1; BR9707549-A; AU719126-B
Picaro, T. and P. G. McCormick. "Bauxite treatment" Patent EP880468-A; WO9729993-

A; WO9729993-A1; AU9715854-A; EP880468-A1; BR9707543-A; AU719117-B
Picaro, T., B. Pei, et al. (2002). "Separation and mineralogical analysis of Bayer red
mud." Developments in Chemical Engineering and Mineral Processing 10(5-6): 475-489.
Prokopov, I. V. and N. S. Malts (1974). "Industrial realisation and perspectives of further

development for low grade bauxite processing." Travaux ICSOBA(12): 237-244.
Qiu, G. Z., T. Jiang, et al. (2004). "Activation and removal of silicon in kaolinite by
thermochemical process." Scandinavian Journal of Metallurgy 33(2): 121-128.
Rao, M. B. S., B. Das, et al. (1996). "Beneficiation of titanium minerals from sand rejects
of an alumina refinery plant by Gravity separation." Powder Handling and Processing
8(4): 351-353.
Rao, R. B. (1996). "Novel approach for the beneficiation of ferruginous bauxite by

microwave heating." Minerals and Metallurgical Processing 13(3): 103-106.

Rayzman, V. L. and I. K. Fillipovich (1999). "Integrating coal combustion and red mud
sintering at an alumina refinery." JOM-Journal of the Minerals Metals & Materials Society
51(8): 16-18.
Rayzman, V. L., I. Z. Pevzner, et al. (2003). "Extracting silica and alumina from low
grade bauxite." JOM-Journal of the Minerals Metals & Materials Society 55(8): 47-50.
Reisner, K. H. (1974). "Studies on the reactive behaviour of Bayer red mud in sintering
with lime and soda." Travaux ICSOBA(12): 299-312.
Roberts, R. H. and J. F. Dunne (1980). "Bauxite Mining and Beneficiation by Comalco

Ltd. at Weipa, Qld." Mining and Metallurgical Practices in Australia J. T. Woodcock.
Parkville Victoria, Australian Institute and Mining and Metallurgy: 368-371.
Rodda, D. P. and R. W. Shaw. "Extraction of alumina from bauxite by double digestion -

including fast low temp. digestion of gibbsite fraction, solid/liq. sepn., digestion of
boehmite fraction and two-stage post-desilication" Patent EP777628-A; WO9606043-A;
WO9606043-A1; AU9532481-A; EP777628-A1; BR9508761-A; KR97705519-A
Roskill (2008). The Economics of Bauxite and Alumina, Roskill Information Services.
Rousseaux, J.-M., H. Verschuur, et al. (2006). "Beneficiation of High Quartz Content

Bauxite from Los Pijiguaos". Light Metals Minerals, Metals and Materials Society (TMS),
184 Thorn Hill Road, Warrendale, PA, USA: 47-52.
Sadler, L. Y. and C. Venkataraman (1991). "A process for the enhanced removal of iron
from bauxite ores." International Journal of Mineral Processing 31(3-4): 233-246.
Sagin, V. C. (1988). New hydro-chemical processing of combined treatment of alumino-

silicates and high silica bauxites, Moscow Metallurgica.
Santos, H. D., T. W. Campos, et al. (2005). "Thermal phase sequences in

gibbsite/kaolinite clay: electron microscopy studies." Ceramics International 31(8): 1077-
1084.
Schoenfelder, H. and H. Ginsberg (June 7 1960). "Method for reducing the silica content
of alumina-containing materials of the clay type" Patent US2,939,764
Smith, P., C. Wingate, et al. (2008). "Mobility of included soda in sodalite". Eighth
International Alumina Quality Workshop. Darwin N.T.: 27-30.
Solymar, K., F. Madai, et al. (2005). "Effect of Bauxite Microstructure on Beneficiation

and Processing". Light Metals. San Francisco, CA, Minerals, Metals and Materials
Society (TMS), 184 Thorn Hill Road, Warrendale, PA, USA: 47-52.
Solymar, K., M. Orban, et al. (1983). "Methods for Reducing NaOH Losses in the
Hungarian Alumina Plants." Travaux ICSOBA 13(18): 377-390.
Solymar, K., J. Steiner, et al. (1997). "Technical peculiarities and viability of

hydrothermal treatment of red mud". Light Metals, Minerals, Metals and Materials
Society (TMS), 184 Thorn Hill Road, Warrendale, PA, USA: 49-54.
Takahashi, T. (14 Jul 1962). "Method to use high silica bauxite by extracting alumina
from high silica bauxite" Patent JP37-008257-B

Tanjo, M., O. Yudate, et al. (21 Jun 1994). "Process for producing iron-making material
composition comprising bauxite dissolution residue" Patent JP H6-172876-A WO
93/20251.
Tomka, L. (1974). "Technology of aluminium oxide production from low modulus

bauxites by the sintering method." Travaux (ICSOBA)(12): 195-211.
Valenti, R. and P. Ho (2008). "Rio Tinto Alcan Gove G3 experience on pre-assembled

modules". Eighth International Alumina Quality Workshop Darwin, N.T.: 1-5.
Vasan, S. S., J. M. Modak, et al. (2001). "Some recent advances in the bioprocessing of
bauxite." International Journal of Mineral Processing 62(1-4): 173-186.
Wang, Y., Y. Hu, et al. (2004). "Reverse flotation for removal of silicates from diasporic-
bauxite." Minerals Engineering 17(1): 63-68.
Whittington, B. I. (1996a). "The chemistry of CaO and Ca(OH)2 relating to the Bayer
process." Hydrometallurgy 43: 13-35.
Whittington, B. I. (1996b). "Quantification and characterisation of hydrogarnet and

cancrinite present in desilication product (DSP) by powder X-ray diffraction". Fourth
International Alumina Quality Workshop. Darwin, N.T.: 413-422.
Xu, Z., V. Plitt, et al. (2004). "Recent advances in reverse flotation of diasporic ores - A
Chinese experience." Minerals Engineering 17(9-10): 1007-1015.
Yamada, K., T. Fukunaga, et al. (1979). "Process of sulfur dioxide removal from waste
gas by red mud." Light Metals, Minerals, Metals and Materials Society (TMS), 184 Thorn
Hill Road, Warrendale, PA, USA: 69-79.
Yamada, K. and T. Harato (1982). "SO2 removal from waste gas by red mud slurry -
pilot plant test and results of plant operation " Kagaku Kogaku Robunshu 8(1): 32--38.
Zhao, H., H. Hu, et al. (2002). "Digestion of diasporic bauxite with mass ratio of
Al2O3/SiO2 no greater than 7 by Bayer process with an excessive addition of lime".
Light Metals. Seattle, WA, Minerals, Metals and Materials Society (Warrendale,
Pennsylvania): 101-104.
Zhong, H., G. Liu, et al. (2008). "Flotation separation of diaspore from kaolinite,
pyrophyllite and illite using three cationic collectors." Minerals Engineering 21(12-14):
1055-1061.
Zoldi, J., K. Solymar, et al. (1987). "Iron hydrogarnets in the Bayer process". Light
Metals, Minerals, Metals and Materials Society (TMS), 184 Thorn Hill Road, Warrendale,
PA, USA: 105-111.
View publication stats

Economic Processing of High Silica Bauxites - Exis

Uploaded by

Copyright:

Available Formats

Economic Processing of High Silica Bauxites - Exis

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

Economic Processing of High Silica Bauxites - Exis

Uploaded by

Copyright:

Available Formats

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

Economic Processing of High Silica Bauxites - Existing and Potential

The user has requested enhancement of the downloaded file.

High Silica Bauxite Processing

Economic Processing of High Silica Bauxites –

2. Potential Technologies – Strategy One – Reduce Reactive Silica Input ................. 10

3. Potential Technologies – Strategy Two - Modify the Process.................................. 17

4. Potential Technologies – Strategy Three - Soda Recovery Processes ................... 28

5. Summary and Conclusions...................................................................................... 32

1.2 Bauxite types

Lateritic bauxites, formed in equatorial climates comprise ~90% of the world’s

HSB review 2008_240209 This version 24 February 2009 Page 3 of 42

The following table describes the approximate compositional differences between

Element Lateritic Karst

Al2O3 Gibbsite, Boehmite Boehmite, Diaspore

SiO2 Kaolinite, Quartz Kaolinite, Quartz, Chamosite, Illite

Fe2O3 Goethite, Hematite Hematite, Goethite Maghemite, Magnetite

TiO2 Anatase, Rutile Anatase, Rutile, Ilmenite

CaO Calcite, Apatite Calcite, Apatite, Crandallite

1.3 Current world bauxite reserves

Table 2, sourced from 2008 US Geological Survey “Mineral Commodities Summaries”,

HSB review 2008_240209 This version 24 February 2009 Page 4 of 42

World region % of world bauxite resource

HSB review 2008_240209 This version 24 February 2009 Page 5 of 42

1.4 The Bayer process

3 Al2 Si2O5 (OH ) 4 + 18 NaOH → 6 Na2 SiO3 + 6 NaAl (OH ) 4 + 3H 2O eq.1

In many modern operations there is a slurry storage or “pre-desilication” stage in which

HSB review 2008_240209 This version 24 February 2009 Page 6 of 42

Bauxite is charged to digestion to achieve a specific alumina supersaturation. The level

1.5 Structure of DSP

HSB review 2008_240209 This version 24 February 2009 Page 7 of 42

At high digestion temperatures cancrinite, another cage structured aluminosilicate can

HSB review 2008_240209 This version 24 February 2009 Page 8 of 42

1.6 Strategies for caustic loss minimisation

Depending on what technology is being considered/developed in these three strategies,

The remainder of this document will concentrate on potential technologies (either

HSB review 2008_240209 This version 24 February 2009 Page 9 of 42

2.1 Screening / washing

HSB review 2008_240209 This version 24 February 2009 Page 10 of 42

2.2 Gravity separation

Recovery of Al2O3 in underflow

Ratio of A/S in underflow

HSB review 2008_240209 This version 24 February 2009 Page 11 of 42

Flotation is particularly suited to diasporic ores in which there is sufficient liberation of

HSB review 2008_240209 This version 24 February 2009 Page 12 of 42

HSB review 2008_240209 This version 24 February 2009 Page 13 of 42

2.5 Magnetic separation

2.6 Mechano-chemical treatment of bauxite with lime

HSB review 2008_240209 This version 24 February 2009 Page 14 of 42

2.7 Roast-leach process

HSB review 2008_240209 This version 24 February 2009 Page 15 of 42

Bauxite Crushing Roasting

Sodium silicate Weak caustic

Alumina Red mud

HSB review 2008_240209 This version 24 February 2009 Page 16 of 42

3.1 Sinter processes

2CaCO3 + SiO2 → 2CaO.SiO2 + 2CO2

2 Al (OH ) 3 → Al2O 3 +3H 2O

HSB review 2008_240209 This version 24 February 2009 Page 17 of 42

3.2 Combination process (Bayer-sinter)