Hydrometallurgy 134–135 (2013) 102–109

Contents lists available at SciVerse ScienceDirect

Hydrometallurgy
journal homepage: www.elsevier.com/locate/hydromet

Mineralogical characterisation of Indonesian laterites prior to and post

atmospheric leaching
Rong Fan, Andrea R. Gerson ⁎
Minerals and Materials Science & Technology, Mawson Institute, University of South Australia, Mawson Lakes, SA 5095, Australia

a r t i c l e i n f o a b s t r a c t

Article history: Nickel laterite is of increasing importance as an economically viable source of Ni. However, practical strate-
Received 11 August 2012 gies for low cost atmospheric leaching of laterites have been hampered in part by poor recovery and the for-
Received in revised form 12 February 2013 mation of potentially environmentally harmful residues, for instance jarosite. It is important therefore to
Accepted 12 February 2013
define the mineralogical forms of nickel present, using readily accessible laboratory-based approaches, in
Available online 16 February 2013
the ores and residues, and the form of secondary precipitates in the residues, as a means for future optimisa-
Keywords:
tion of leach parameters. To this end a detailed mineralogical study of an Indonesian laterite and residues
Mineralogy of nickel laterite resulting from atmospheric acid leaching has been undertaken. The leach process involved the sequential re-
Atmospheric leaching actions of first, limonite (3 h) and then saprolite (11 h) in sulphuric acid leach liquor. The acid, dry limonite
Indonesian laterites and saprolite ores were added in the weight ratio of 1.4:1:1. At 100 °C, over 80% of the total Ni was extracted
Quantitative XRD into solution during the whole leach process.
The limonite ore was found to contain Ni (1.2 wt.%) predominantly within Fe oxy/hydroxides, serpentine
and phyllomanganate whilst in the residue Ni (0.4 wt.%) is present in highly crystalline leach resistant min-
erals, such as spinels. The majority of the Ni (1.6 wt.%) in the saprolite ore was present in serpentine with a
small fraction also present in phyllomanganate. This association is further confirmed by the similar saprolite
leach extraction curves of Ni and Mg. Goethite and jarosite are both found to be concentrated in the saprolite
residue, as is consistent with the decrease in Fe extraction during the saprolite leach. For the final 11 h res-
idue of the saprolite leach, the Ni (0.3 wt.%) is mainly present in undissolved lizardite and leach resistant py-
roxene and spinel minerals.
Crown Copyright © 2013 Published by Elsevier B.V. All rights reserved.

1. Introduction is strongly controlled by climate, topography, drainage and other

hydro-geological factors together with the mineralogy and structure
It is estimated that approximately 70% of global nickel resources are of the ultramafic bedrock. Hence, the profile of nickel laterites can
contained within lateritic ores, yet laterites accounted for only about be highly variable with variations also occurring even within laterite
50% of global annual nickel production in 2009 (Mudd, 2010). However, profiles. Despite the diversity of laterite profiles, some common fea-
nickel extraction from lateritic deposits did provide the capacity for tures can be found enabling laterite types to be classified as oxide,
most of the expansion in Ni production over the five years prior to clay or silicate laterites (Brand et al., 1998).
2009 (Kuck, 2009). Nickel laterite ores are often treated by high pres- Weathering, resulting from contact between ultramafic rock and
sure acid leaching (HPAL). Drawbacks of HPAL are that the process is acid surface water, gives rise to geochemical break-down and the re-
extremely complex and requires high initial capital cost (Whittington lease of mobile elements. Fe and Al are less soluble than Mg and Si
and Muir, 2000). To overcome this considerable effort has gone into and hence are enriched in the zone forming the uppermost oxide li-
the development of atmospheric pressure leach processes for the ex- monite layer. Oxide laterites mainly comprise goethite, hematite
traction of Ni and Co (McDonald and Whittington, 2008). and Mn oxides. Ni is associated with goethite typically by substitution
As lateritic profiles and the factors affecting their genesis have for Fe (de Carvalho-e-Silva et al., 2003; Manceau et al., 2000) thus re-
previously been extensively reviewed (Gleeson et al., 2003), we quiring destruction of the mineral matrix for Ni recovery by leaching
present here only a summary of their main features. Laterites are (Canterford, 1979; Chander, 1982; Senanayake and Das, 2004). Ni and
formed by intense tropical weathering of ultramafic rocks, such as pe- Co are also strongly incorporated and precipitated into Mn oxides, e.g.
ridotites and serpentinites, that generally contain up to 0.3 wt.% Ni asbolane (Elias, 2001). Similarly, extraction of Ni and Co from
within the mineral olivine (Golightly, 1981). This weathering process Mn-bearing nodules also requires complete dissolution (Canterford,
1986).
⁎ Corresponding author. Tel.: +61 422 112 516; fax: +61 8 83025545. Some Ni percolates downward from the overlying limonite layer
E-mail address: Andrea.Gerson@unisa.edu.au (A.R. Gerson). and is incorporated into the underlying serpentine (Mg–Fe silicate)

0304-386X/$ – see front matter. Crown Copyright © 2013 Published by Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.hydromet.2013.02.004
 R. Fan, A.R. Gerson / Hydrometallurgy 134–135 (2013) 102–109 103

layer. Silicate laterites are formed after a long period of aggressive (Mastersizer 2000) indicated that approximately 80 vol.% of the limo-
weathering and are generally found in this underlying saprolite nite ore particles was less than 30 μm in size (Fig. 1). In contrast, the
zone (Elias, 2001). Clay laterites contain predominantly the smectite particle size of the saprolite ore is not as fine, with 50 vol.% of the sap-
clay nontronite and are formed under less severe weathering condi- rolite particles being less than 50 μm in size, and around 35 vol.% of
tions (Gaudin et al., 2005). It has been reported that compared to ser- the saprolite particles being larger than 100 μm. Large rocks of diameter
pentine minerals which are leached relatively easily (Liu et al., 2010; greater than 5 cm are also found in the saprolite ore.
Luo et al., 2009, 2010), other minerals present in saprolite ores such The size fractions of the limonite and saprolite ores used for
as magnetite/maghemite and chromite, are more difficult to dissolve leaching were b75 μm and 38–75 μm, respectively. Both ores were
by acidic attack, and may remain even after a long period of leaching wet sieved to achieve the appropriate size fraction. The > 75 μm frac-
(Canterford, 1986; Rice and Strong, 1974). For saprolite ore, it has tions were crushed and re-sieved and combined with the previous
been found that the leach residues contained mainly amorphous silica fractions. As the majority of the limonite particles are very fine it
(Hirasawa and Horita, 1987). This interpretation can also be made by was not possible to obtain sufficient sample to carry out leaching
examining the incongruent dissolution behavior of serpentine min- using a 38 to 75 μm sized fraction.
erals, where the metal cations are released without appreciable re-
lease of silica (Jurinski and Rimstidt, 2001; Kosuge et al., 1995). 2.2. Leaching tests
Lateritic leach waste management is also of concern with jarosite
being a common secondary precipitate. The unstable nature of jarosite, Atmospheric pressure leaching of the sized Indonesian limonite
compared to goethite, requires long term management for conversion and saprolite ores was carried out based on the patent Liu et al.
to a more stable iron oxide form thus avoiding acid drainage upon (2002) (Fig. 2). Leaching was performed in a 2 L glass reaction vessel
weathering (Li et al., 2007). It is hence of considerable importance to with an impeller agitator operating at 550 rpm. 800 g of 25 wt.%
optimise not only leach conditions and recovery but also to ensure slurries of the two ores was prepared. The leach process involved the
that the leach waste products are not environmentally deleterious. sequential reactions of first, limonite with 155 mL 98 wt.% sulphuric
Indonesian nickel laterite ores have been the subject of a number of acid at 100 °C and atmospheric pressure for 3 h (weight ratio of
previous mineralogical studies. The chemical composition and infrared acid to dry limonite 1.4:1), followed by saprolite leaching at approxi-
spectrum of Ni- and Fe-substituted serpentine sourced from a predom- mately pH 3.0, 100 °C and atmospheric pressure for 11 h. The limo-
inantly saprolitic laterite profile located at Sorowako, Indonesia have nite residue was removed prior to the addition of the saprolite slurry
been examined (Golightly and Arancibia, 1979). Another nickeliferous to the leach liquor. The redox potential of the slurry was adjusted by
ore sampled from the same location was found to consist of mainly adding lithium sulphite (Li2SO3) solution after acid addition at the be-
Mg silicates, Mg–Fe silicates and goethite. Scanning electron microscopy ginning of the limonite leach, to enhance the dissolution of Ni but also
(SEM) and electron microprobe examination indicated that Ni was pres- particularly Co, from the Mn oxide minerals. The lithium sulphite solu-
ent in serpentine, olivine, chlorite and amphibole, and in smaller tion was prepared by bubbling SO2 gas into lithium carbonate slurry.
amounts in goethite and Mn oxide. Cobalt was also detected in the Mn Limestone was used for neutralisation after the saprolite leach. Sam-
oxide (Chen et al., 2004). Ni has also been found to be associated with pling was carried out every hour by taking approximately 10 mL of
goethite in the Sebuku Island limonitic laterites, wherein the Ni concen- slurry. The pH and Eh (SHE) of the slurry were recorded. The liquid
tration is 0.30 wt.% and the Fe concentration is 50.88 wt.% (Purwanto et and solid were separated, by centrifuging, for chemical and XRD
al., 2003). Electrochemical leaching of these low-grade Sebuku Island li- analyses.
monitic laterites also suggests that Ni could be associated with goethite Solution speciation calculations are employed to understand the
due to similar dissolution behaviour of Fe and Ni. The chemistry and ki- speciation and saturation indices for an aqueous solution at given
netics of HPAL of a limonitic laterite from Indonesia have been system- thermodynamic conditions. Chemical equilibrium is determined by
ically examined by Georgiou and Papangelakis (1998). Transmission solving a set of linear equations consisting of the equilibrium con-
electron microscopy (TEM) analysis of goethite within the limonitic lat- stants and mass balance constraints. The PHREEQC computer pro-
erite indicates that Ni exists as a substituent of the goethite lattice rather gram was used for the solution speciation calculations reported
than being bound on its surface by adsorption mechanisms (Georgiou herein (Parkhurst, 1995). Eh, pH, temperature, elemental concentra-
and Papangelakis, 1998). tions, atmospheric CO2 pressure and the redox couple Fe 2+/Fe 3+
Even though there has been considerable effort previously devot- were input to calculate the equilibrium solution speciation for each
ed to the atmospheric and pressure leaching of laterites, little empha- solution sample. A calculated saturation index of less than 0 indicates
sis has been placed on understanding in detail the Ni mineralogy as a that the given mineral is under-saturated with respect to the solution
function of leaching (Liu et al., 2009). A general lack of knowledge re-
garding the mineralogy and chemistry of leach products also hampers
the understanding of leaching behavior of the other mineral compo-
nents present in the ore. We report the mineralogical and chemical
characterisation, using readily accessible laboratory-based tech-
niques, of an Indonesian limonitic and saprolitic laterite and their re-
spective atmospheric leach residues. The mineralogical information,
of both ore and leach residues, is examined in detail to understand
physical and chemical characteristics of the atmospheric leaching
process. The characterisations also provide quantitative analysis of
jarosite precipitates, which are of major environmental concern.

2. Experimental

2.1. Materials

Samples were obtained from Raja Ampat, West Papua, Indonesia.

The limonite and saprolite ores were selected from the upper Fe-rich Fig. 1. Particle size analyses for the limonite and saprolite ores with dashed lines show-
ore and the deeper Mg-rich ore, respectively. Particle size analysis ing cumulative volume distributions.
104 R. Fan, A.R. Gerson / Hydrometallurgy 134–135 (2013) 102–109

3. Results and discussion

3.1. Mineralogical analyses of the limonite ore

SEM/EDX analyses confirmed the limonite ore to be heterogeneous

and fine grained, as was indicated by particle size analysis and as has
been observed for other limonite ores (Fan and Gerson, 2011). Ni is
found to be heterogeneously distributed within different sized particles
and was associated with grains containing relatively high concentra-
tions of Fe suggesting that a Ni-containing mineral could be goethite
or other Fe oxide/oxyhydroxide (Fig. 3 and Table 2; grain G in Fig. 4b
and Table 3).
EDX analysis suggests that the Ni in the particles shown in Fig. 3a
Fig. 2. Diagram of atmospheric leach of nickel laterite. is probably associated with phyllomanganate (atomic percentages: Ni
1.6%, Mn 15.8%, Co 6.6%, Fe 2.8%, Al 10.7% and O 62.1%). SEM/EDX
so that the mineral might dissolve if present, whereas a saturation analysis also suggests that Ni is found associated with large Mg and
index greater than 0 indicates that the mineral may precipitate. Si and small Fe concentrations in some of the selected Indonesian li-
monite grains (grains D and E in Fig. 4b and Table 3). These grains
probably contain Ni serpentine minerals.
2.3. Analytical methods Rietveld analysis of the XRD data indicates that the limonite ore is
composed of 78±5 wt.% goethite, 2 ± 1 wt.% lizardite (serpentine
The sized ore and solid leach samples were dry ground to a fine group mineral), 1 ± 1 wt.% maghemite (spinel), 2 ± 1 wt.% gibbsite,
powder (b 10 μm) in an agate mortar and pestle for elemental composi- 3 ± 1 wt.% quartz, 6 ± 2 wt.% clinopyroxene, 6 ±2 wt.% talc, 1 ±
tion and thermal analysis (described below). Trace metal concentra- 1 wt.% hercynite (spinel), 0.5 ± 0.8 wt.% enstatite and a small compo-
tions in the solids were measured after HCl/HNO3/HF analytical nent of (approx. 1 wt.%) unidentified material (final converged weight-
microwave assisted digestion (EPA Method 3052). The resulting solu- ed R-factor was 2.3%, Fig. 5). Phyllomanganate was not observed in the
tions were then analysed using inductively coupled plasma atomic bulk XRD analysis, due to its relatively small concentration (as
emission spectroscopy (ICP-AES; Trace Scan, Thermo Jarrel Ash). Bulk suggested by 1.6 wt.% Mn, Table 1) but is indicated as being present
assays for the limonite and saprolite ores were obtained (Table 1). by SEM/EDX analysis (Fig. 4a). It is estimated that the limonite ore con-
Morphological and mineralogical studies of the ore samples were tains approximately 5 wt.% phyllomanganate based on the assumption
performed using a Philips XL30 SEM equipped with an energy disper- that all the 1.6 wt.% Mn is associated within phyllomanganate.
sive X-ray (EDX) system. Quantitative mineralogical analyses of the Thermal analysis of the limonite ore gave rise to three endothermic
ores and leach samples were conducted using Rietveld analysis of lab- events at 40 °C, 257 °C and 560 °C (the latter two are labelled A and B in
oratory XRD data. XRD data collection was performed on a Bruker D4 Fig. 6), corresponding to the release of moisture, the decomposition of
Endeavor diffractometer with Co Kα radiation (1.7902 Å) at 30 kV goethite, and the decomposition of serpentine to amorphous forsterite,
and 20 mA. Diffraction patterns were collected with θ/2θ geometry respectively (Dutrizac and Jambor, 2000). Weight loss of 9 wt.% and
from 5 to 80° 2θ at 0.02° 2θ increments with a rotating sample stage 2 wt.% that occurred during the latter two events, over the temperature
to ameliorate preferred orientation effects. Corundum (α-Al2O3, ranges of 150 °C to 350 °C and 400 °C to 700 °C respectively, indicates
20 wt.%) was added as an internal standard. All the samples were that the original material contains a mass composition of 84 wt.% goe-
ground to − 38 μm for XRD data collection. Phase quantification thite or other Fe oxide/oxyhydoxide and 7 wt.% lizardite. An exothermic
was carried out using the DiffracPlus EVA software (Bruker) with ap- event (labelled C in Fig. 6) is due to recrystallisation of forsterite
plication of the PDF2 database as supplied by JCPDS (Joint Committee (Brindley and De Souza, 1975; Tartaj et al., 2000). The 7 wt.% lizardite
on Powder Diffraction Standards) — International Centre for Diffrac- indicated from thermal analysis, is likely to be overestimated, in com-
tion Data. Bruker-AXS's TOPAS V4.2 software was then used to quan- parison to the 2 wt.% derived from XRD Rietveld analysis. This is may
tify each phase (Cheary and Coelho, 1992). Crystal structure be due to the presence of another phyllosilicate mineral, talc, which
information for all the minerals was obtained from the Bruker Struc- could also contribute to the two thermal events (B and C in Fig. 6). It ap-
ture Database. By comparing the quantitative results with the known pears that the Indonesian limonite sample contains a significant quanti-
content of corundum, it is possible to quantify the amorphous content ty of goethite which is consistent between the analyses, i.e. 84 wt.% total
of each sample. A reasonable estimate of the uncertainty of the goethite from DTA–TG compared to 78 wt.% goethite from XRD Rietveld
resulting quantitative XRD analyses at the 95% level is ±X 0.35, analysis. The remaining phases consist of silicate minerals (lizardite,
where X is the weight % of a given phase (Geelhoed et al., 2002). clinopyroxene, talc and enstatite), spinels (maghemite and hercynite),
Thermal analysis (differential thermal analysis and thermal gra- metal oxide/hydroxides (phyllomanganate and gibbsite) and quartz.
vimetry, DTA–TG) was conducted using a temperature increase of
5 °C per minute in an argon atmosphere. A thermal analysis curve 3.2. Mineralogical analysis of the saprolite ore
for corundum was used for instrument calibration and background
subtraction. The DTA–TG and elemental composition results were Figs. 7 and 8 provide SEM images of the Indonesian saprolite ore.
reconciled with the quantitative XRD results to provide improved EDX analysis shows that most of the Ni is associated with Mg and Si
mineralogical assessment. (grains A, B, E, F and H in Fig. 7 and Table 4, and grains B and C in
Fig. 8 and Table 5), possibly as lizardite. Some Ni also seems to be as-
sociated with Mn (grains C and C1 in Fig. 7 and Table 4) possibly as
Table 1 phyllomanganate. Phyllomanganate is not detected by bulk XRD
Elemental composition of the limonite and saprolite ore samples (wt.%). probably due to its small concentration as confirmed by the bulk
assay of less than 0.1 wt.% Mn (Table 1). Ni is also found to be associ-
Sample Al Co Cr Fe Mg Mn Ni S Si
ated with Fe (G in Fig. 7 and Table 4, and E in Fig. 8 and Table 5)
Limonite (b75 μm) 5.4 0.3 1.0 42.3 0.2 1.6 1.2 0.3 3.4 where Mg and Si are also present but at relatively small atomic per-
Saprolite (38–75 μm) 1.3 0.0 0.2 6.2 16.6 0.0 1.6 0.0 22.3
centages. In addition Ni is detected in Si and Fe rich grains where
 R. Fan, A.R. Gerson / Hydrometallurgy 134–135 (2013) 102–109 105

a b c

50 µm 50 µm 50 µm

d e f

100 µm 100 µm 50 µm

Fig. 3. SEM images of selected limonite Ni-containing grains. EDX analysis (Table 2) suggests that the Ni is probably present associated with Fe oxide/oxyhydroxide.

Mg is not found (D in Fig. 7 and Table 4, and A in Fig. 8 and Table 5). to 1.57 due to dilution of the acidity. The pH reached a maximum
These grains could be composed of pyroxene or other silicate upon 2 h of leaching and decreased thereafter. This implies that fur-
minerals. ther acid generation occurred during the saprolite leach process
The DTA–TG trace from the saprolite ore is similar to that from the most probably due to secondary precipitation of goethite (the reverse
limonite ore. Three endothermic events at 40 °C, 257 °C (A, Fig. 6) of Eq. (1)) and jarosite (Eq. (2)). The Eh (SHE) was stable during the
and 560 °C (B, Fig. 6), and an exothermic event at 812 °C (C, Fig. 6) limonite leach being between 750 mV and 770 mV and being be-
are observed. The first endothermic peak is associated with the loss tween 710 mV and 720 mV during saprolite leach.
of moisture and the second one with goethite decomposition. The
þ 3þ
weight loss during the second event, over the temperature range of FeOOHðsÞ þ 3H ⇆Fe þ 2H2 O ð1Þ
150 °C to 350 °C, is 1 wt.%, indicating the presence of 8 wt.% goethite
in the saprolite ore. The third endothermic peak is due to serpentine 3þ 2− þ
3Fe þ 2SO4 þ 7H2 O→ðH3 OÞFe3 ðSO4 Þ2 ðOHÞ6ðsÞ þ 5H ð2Þ
dehydroxylation and the exothermic peak at 812 °C is associated with
forsterite recrystallisation. Weight loss of 6.8 wt.% during serpentine Fig. 9 shows a comparison of the percentage extractions of Ni, Mg,
dehydroxylation, over the temperature range of 400 °C to 700 °C, Fe, Mn, Co and Cr during both leaching stages. The limonite leaching
suggests that approximately 53 wt.% lizardite is present in the sapro- kinetics is were fairly rapid with 80% Ni extraction on 1 h of leaching
lite ore. as compared to 85% at the end of 3 h. High extractions (above 95%)
Rietveld XRD analysis indicates that the saprolite ore is composed are also achieved for Co, Mn and Fe. These rapid and high recoveries
of 27 ± 3 wt.% lizardite, 21 ± 3 wt.% goethite, 10 ± 2 wt.% enstatite, are probably due to the rapid dissolution of Fe oxide/hydroxide and
3 ± 1 wt.% diopside, 5 ± 2 wt.% olivine, and 5 ± 2 wt.% montmorillon- Mn oxides phases both of which are readily attacked by high acidity.
ite (final weighted R-factor of the Rietveld analysis was 5.4 %, Fig. 5). Approximately 70% of the Cr is also dissolved from the limonite ore.
The ore also contains minor quantities of quartz (1 ± 1 wt.%) and spi- Fig. 10 shows a comparison of XRD patterns of the limonite ore and
nels (2 ± 1 wt.%). The phase of 27 wt.% of the material is unidentified. residues and their Rietveld analyses. There was 35 wt.% goethite found
The calculation of Si composition derived from Rietveld XRD results is in the residue extracted after 3 h of leaching, suggesting that much of
15 wt.%, compared to the observed value of 22 wt.%. Furthermore, the the goethite in the limonite was dissolved. Silicate minerals, e.g.
difference between the 53 wt.% of lizardite derived from DTA–TG and lizardite, enstatite, clinopyroxene and talc were nearly completely
the 27 wt.% lizardite derived from Rietveld analysis suggests that a dissolved after 3 h of leaching. The spinel hercynite (FeAl2O4, ICDD
substantial portion of the unidentified material is likely to be poorly 34-0192) was identified in the residues with d-spacings of 4.71, 2.88,
crystalline serpentine. In comparison to DTA–TG results, the Rietveld 2.46, 2.04, 1.66, 1.57 and 1.44 Å (21.9, 36.1, 42.6, 52.1, 65.0, 69.5 and
XRD analysis suggests an overestimation of goethite, which is con- 76.7° 2θ, respectively). Only small concentrations of hercynite and
firmed by a relatively low concentration of Fe, 6.2 wt.% in the sapro- quartz (1± 1 wt.% and 3 ± 1 wt.%, respectively) were observed in the
lite ore. XRD pattern from the limonite ore. However, after the majority of the
predominant Fe oxide/oxyhydroxide, Mn oxide phases and silicates
3.3. Atmospheric leaching minerals are dissolved, these phases become more concentrated with
18±3 wt.% for quartz and 16 ± 3 wt.% for hercynite present after 1 h
The pH increased from − 0.20 to −0.14 during the limonite leach of leaching. Cr is found to be mainly associated with chromite (a spinel
stage due to acid consumption mainly due to goethite dissolution mineral). The main diffraction peaks of chromite minerals,
(Eq. (1)). Upon addition of the saprolite slurry the pH value jumped

Table 2
a b G
Atomic percentages (determined by EDX) of the particles shown in Fig. 3. A F
O Mg Al Si Fe Ni
B
a 62.9 0.5 7.1 2.1 26.7 0.9
b 44.6 0.6 3.1 1.2 18.2 0.6 C E
c 65.7 1.8 3.4 2.5 25.7 0.9
d 60.7 13.0 1.0 22.3 0.5 20 µm 10 µm
D
e 61.5 7.4 3.9 25.6 0.8
f 48.3 2.9 1.6 45.2 1.2
Fig. 4. SEM images of selected limonite Ni-containing grains.
106 R. Fan, A.R. Gerson / Hydrometallurgy 134–135 (2013) 102–109

Table 3
Atomic percentages, determined by EDX, of the particles shown in Fig. 4b.

O Mg Al Si Cr Fe Ni

A 55.4 2.2 3.5 2.5 36.4

B 59.6 1.2 3.1 2.0 34.1
C 67.9 0.6 7.5 1.9 1.2 20.6
D 58.9 15.3 3.2 16.2 6.1 0.3
E 60.5 17.5 1.6 13.2 6.2 1.0
F 69.5 8.1 1.8 0.9 19.2
G 55.1 7.9 3.2 0.6 32.2 1.1

e.g. cochromite, have d-spacings of 2.95 and 2.51 Å (35.34 and 41.72 2θ,
respectively), which are indistinguishable from those of maghemite
which is also a spinel. Maghemite (and/or chromite) behaves in a sim-
ilar manner to hercynite and quartz, becoming concentrated from 1 ±
1 wt.% wt.% in the ore, to 7 ± 2 wt.% in the 3 h residue. 0.4 wt.% of Ni
is detected in the final limonite residue, where the Ni is likely to be Fig. 6. DTA–TG curves of the limonite and saprolite ores with solid lines showing DTA
present in the leach resistant spinel minerals. (temperature difference) curves and dashed lines showing TG (sample weight) curves.
The kinetics of leaching for the saprolite ore is slower than that for
the limonite ore with respect to metal dissolution. Only 50% of the
total Cr (that is from both the limonite and saprolite ores) is extracted
during the saprolite leach stage. The similar extraction curves of Mn
and Co during the saprolite leach imply that they are likely to be H
co-located. This is reasonable as Mn and Co have been reported to be
generally associated within Mn oxides (Chen et al., 2004; Llorca, G
1993). Similar extraction profiles are found for Ni and Mg during the
F
saprolite leach, indicating that most of the Ni in the saprolite is associat- E
ed with Mg most likely within serpentine. Some of the Ni is also possibly
associated with Mn and Co in phyllomanganate minerals, as was ob- B
served in our recent geochemical study of a Philippine laterite (Fan
C
and Gerson, 2011). The final overall extractions are around 80% for Ni
D
and 45% for Mg, suggesting significant Mg remained undissolved.
The saprolite Fe extraction profile contrasts those of the other
A
metals. The decreasing extraction of Fe during the saprolite leach
50 µm
(from 66% at 1 h to 32 % at 11 h) is most likely caused by the lower
acidity in the saprolite leach, as compared to the limonite leach, Fig. 7. SEM image of selected saprolite Ni-containing grains.
resulting in the precipitation of secondary Fe-containing minerals,
such as goethite and jarosite. Fig. 11 shows the saturation indices cal- precipitate, i.e. not redissolve, than Na-jarosite as the reaction con-
culated for the saprolite leachates. It is clear that both jarosite and tinues, as its saturation index remains above 0.
goethite are supersaturated and may precipitate soon after the addi- The formation of fine grained particles, e.g. goethite, is confirmed
tion of the saprolite slurry. H-jarosite is more likely to remain as a by particle size distribution analyses of two saprolite residues
(Fig. 12). It is clear, by comparing the 2 h residue to the 11 h residue,
that the coarse particles decrease in number but the concentration of
particles smaller than 5 μm increases significantly.
XRD analysis of the saprolite leach residues compliments the Fe
extraction measurements and solution speciation calculations. The
diffraction peak heights of lizardite decrease in relative intensity

F G
D
H
C
A
B

Fig. 5. Experimental and Rietveld calculated traces for the limonite and saprolite
ores. G — goethite (ICDD 29-0713), L — lizardite (09-0444), C — corundum E
(10-0173), M — maghemite (25-1402, or other spinel minerals), T — talc
(73-0147), Q — quartz (46-1045), Gb — gibbsite (76-1782), E — enstatite 100 µm
(88-2377), D — diopside (80-1869), Mt — montmorillonite (13-0135), O — olivine
(79-1191), Cp — clinopyroxene (78-1391). Corundum was added as an internal
standard. Fig. 8. SEM image of selected saprolite grains.
 R. Fan, A.R. Gerson / Hydrometallurgy 134–135 (2013) 102–109 107

Table 4
Atomic percentages of the particles shown in Fig. 7. The particle centres were chosen
for EDX analysis.

O Mg Al Si Mn Cr Fe Ni

A 49.4 19.8 2.8 19.7 0.8 5.2 2.3

B 58.5 19.3 0.6 17.0 3.5 1.2
C 57.3 7.1 2.7 10.7 4.7 13.9 3.6
C1 20.1 7.1 5.6 26.9 28.7 11.6
D 60.4 2.8 17.7 1.3 15.5 2.2
E 58.2 18.6 1.1 17.2 3.9 1.0
F 53.6 15.1 2.5 20.6 0.7 5.9 1.7
G 36.8 2.9 7.2 49.3 3.8
H 53.7 15.0 3.2 19.6 0.9 5.5 2.2

Table 5
Atomic percentages of the particles shown in Fig. 8. The particle centres were chosen
for EDX analysis.

O Mg Al Si Cr Fe Ni

A 66.2 3.3 21.6 0.5 5.3 2.8

B 60.1 17.6 1.4 17.4 0.2 2.2 1.1
C 51.9 22.7 17.9 5.7 1.9
D 55.6 3.9 14.9 25.6
E 50.5 5.2 8.1 33.7 2.6
F 28.4 10.6 15.9 37.3 7.8
G 46.3 8.4 19.4 23.6 2.4
H 56.4 3.5 12.1 26.1 2.0

due to dissolution (Fig. 13a) whereas the diffraction peak heights of

goethite and jarosite seemingly increase in relative intensity.
The Rietveld method has been utilised to obtain quantitative phase
information for the saprolite residues (Fig. 13b). It is found that the con-
centrations of goethite and jarosite in the residues increase during the
leach and, in contrast, the concentration of lizardite decreases. The ‘un- Fig. 10. (a) XRD of the limonite ore (bottom) and residues upon 1 h of leaching (middle)
accounted’ for amorphous material indicated in Fig. 13b is present at a and 3 h of leaching (top), and (b) the change in concentration of the major mineral phases
relatively low concentration (12 ±2 wt.% for the 2 h residue and 16± during limonite leaching as suggested by Rietveld analysis of the laboratory XRD data
from the limonite leach residues. G — goethite, L — lizardite, C — corundum, Q — quartz,
3 wt.% for the 11 h residue) during the saprolite leach. This material is H — hercynite, M — maghemite. Hercynite (FeAl2O4) and maghemite (Fe2O3), found in
unlikely to be amorphous Fe oxy/hydroxide as the calculated total Fe the residues, are spinel minerals. Corundum was added into the ore as an internal stan-
concentration in goethite and jarosite derived from Rietveld analysis dard and not included in the Rietveld quantification.
of 25 wt.% in the 11 h residue, reasonably matches the measured con-
centration of 20 wt.%. This implies most of the ‘unaccounted’ for mate- 4. Conclusions
rial is probably amorphous silica. It is likely that some of the Ni present
in the final 11 h (0.3 wt.%) is present within undissolved lizardite and Improved definition of the forms of Ni present in both laterite ore
leach resistant spinels and pyroxene minerals. and leach residues is essential for better understanding of their leach
mechanisms and thus the design of cost-effective nickel concentration
and leaching strategies. Mineralogical characterisation of a selected

Fig. 9. The extractions of Co, Mn, Ni, Cr, Mg and Fe. The % extractions calculated for the
saprolite leach are based on the total metal content, for each specific metal, within both Fig. 11. Saturation indices as determined from the solution speciation calculations for
the limonite and the saprolite slurries. the saprolite leachates.
108 R. Fan, A.R. Gerson / Hydrometallurgy 134–135 (2013) 102–109

serpentine and phyllomanganate. The rapid dissolution of these Ni-

bearing minerals is confirmed by the predominance of highly crystalline
spinel minerals in the limonite residue, the likely location of the
remaining Ni.
For the saprolite ore and the final 11 h residue, the Ni percentages
are 1.6 and 0.3 wt.%, respectively. The majority of the Ni is located with-
in serpentine while some of the Ni is associated with phyllomanganate
as indicated by the Ni extraction profile. Formation of goethite and
jarosite during the saprolite leach is observed by XRD. These observa-
tions are in agreement with the solution speciation calculations. It is
also found that a small amount of amorphous silicate is probably
formed during the leach process. The remaining Ni, 0.3 wt.%, in the
final residue, is probably associated with well crystalline, leach resistant
spinels and pyroxene minerals.

Acknowledgements

Fig. 12. Particle size analyses of the saprolite ore and two residues with dashed lines
This research has been supported by the Premier's Science and Re-
showing cumulative volume distributions.
search Fund of South Australia and was co-sponsored by BHP-Billiton
and Rio Tinto. Rong Fan acknowledges support from the University of
Indonesian laterite (limonite and saprolite) and the resultant atmo- South Australia for the award of postgraduate UniSA President's
spheric leach products is described. By combining SEM investigations, Scholarship.
thermal analysis, particle size distribution and quantitative XRD studies,
it is possible to elucidate the evolution of limonite and saprolite during References
leaching using standard readily assessable laboratory techniques.
Brand, N.W., Butt, C.R.M., Elias, M., 1998. Nickel laterites: classification and features.
The concentrations of Ni in the Indonesian limonite ore and the final
AGSO J. Aust. Geol. Geophys. 17 (4), 81–88.
3 h residue are 1.2 and 0.4 wt.%, respectively. Three locations are Brindley, G.W., De Souza, J.V., 1975. Nickel-containing montmorillonites and chlorites
suggested for Ni within the limonite ore examined: Fe oxyhydroxide, from Brazil, with remarks on schuchardtite. Mineral. Mag. 40 (310), 141–152.
Canterford, J.H., 1979. The sulphation of oxidized nickel ores. In: Evans, D.J.I.,
Shoemaker, R.S., Veltman, H. (Eds.), International Laterite Symposium. Society of
Mining Engineers, American Institute of Mining, Metallurgical, and Petroleum
Engineers Inc. (SME-AIME), New York, pp. 636–677.
Canterford, J.H., 1986. Acid leaching of chromite-bearing nickeliferous laterite from
Rockhampton, Queensland. AusIMM Bull. Proc. 291 (6), 51–56.
Carvalho-e-Silva, M.L., Ramos, A.Y., Tolentino, H.C.N., Enzweiler, J., Netto, S.M., Alves,
M.C.M., 2003. Incorporation of Ni into natural goethite: an investigation by X-ray
absorption spectroscopy. Am. Mineral. 88 (5–6), 876–882.
Chander, S., 1982. Atmospheric pressure leaching of nickeliferous laterites in acidic
media. T. Indian I. Metals 35 (4), 366–371.
Cheary, R.W., Coelho, A., 1992. A fundamental parameters approach to X-ray line-
profile fitting. J. Appl. Crystallogr. 25 (2), 109–121.
Chen, T.T., Dutrizac, J.E., Krause, E., Osborne, R., 2004. Mineralogical characterization of
nickel laterites from New Caledonia and Indonesia. International Laterite Nickel
Symposium, Proceedings of Symposium held during the TMS Annual Meeting,
Charlotte, NC, United States, Mar. 14–18, 2004, pp. 79–99.
Dutrizac, J.E., Jambor, J.L., 2000. Jarosite and their application in hydrometallurgy. Rev.
Mineral. Geochem. 40, 405–451.
Elias, M., 2001. Nickel laterite deposits — geological overview, resources and exploita-
tion. In: Cooke, D.R., Pongratz, J. (Eds.), Giant Ore Deposits: Characteristics, Genesis
and Exploration. : CODES Special Publication. Centre for Ore Deposit Research, Uni-
versity of Tasmania.
Fan, R., Gerson, A.R., 2011. Nickel geochemistry of a Philippine laterite examined by
bulk and microprobe synchrotron analyses. Geochim. Cosmochim. Acta 75 (21),
6400–6415.
Gaudin, A., Decarreau, A., Noack, Y., Grauby, O., 2005. Clay mineralogy of the nickel lat-
erite ore developed from serpentinized peridotites at Murrin Murrin, Western
Australia. Aust. J. Earth Sci. 52 (2), 231–241.
Geelhoed, J.S., Meeussen, J.C.L., Hillier, S., Lumsdon, D.G., Thomas, R.P., Farmer, J.G.,
Paterson, E., 2002. Identification and geochemical modeling of processes control-
ling leaching of Cr(VI) and other major elements from chromite ore processing res-
idue. Geochim. Cosmochim. Acta 66 (22), 3927–3942.
Georgiou, D., Papangelakis, V.G., 1998. Sulfuric acid pressure leaching of a limonitic laterite:
chemistry and kinetics. Hydrometallurgy 49 (1–2), 23–46.
Gleeson, S.A., Butt, C.R.M., Elias, M., 2003. Nickel laterites: a review. SEG Newsletter 54,
11–18.
Golightly, J.P., 1981. Nickeliferous laterite deposits. Econ. Geol. 75, 710–735.
Golightly, J.P., Arancibia, O.N., 1979. The chemical composition and infrared spectrum
of nickel- and iron-substituted serpentine from a nickeliferous laterite profile,
Soroako, Indonesia. Can. Mineral. 17 (4), 719–728.
Hirasawa, R., Horita, H., 1987. Dissolution of nickel and magnesium from garnierite ore
in an acid solution. Int. J. Miner. Process. 19 (1–4), 273–284.
Jurinski, J.B., Rimstidt, J.D., 2001. Biodurability of talc. Am. Mineral. 86 (4), 392–399.
Fig. 13. (a) XRD patterns of saprolite ore (bottom) and residues upon 2 h of leaching Kosuge, K., Shimada, K., Tsunashima, A., 1995. Micropore formation by acid treatment
(middle) and 11 h of leaching (top), and (b) the change in concentration of the major of antigorite. Chem. Mater. 7 (12), 2241–2246.
mineral phases during saprolite leaching as suggested by Rietveld analysis of the labora- Kuck, P.H., 2009. Mineral Commodity Summaries: Nickel. U.S. Geological Survey.
tory XRD data from the saprolite leach residues. G — goethite, J — jarosite, L — lizardite, Li, J., Smart, R.S.C., Schumann, R.C., Gerson, A.R., Levay, G., 2007. A simplified method
C — corundum, E — enstatite, D — diopside, Mt — montmorillonite, O — olivine, Q — quartz. for estimation of jarosite and acid-forming sulfates in acid mine wastes. Sci. Total
Corundum was added as an internal standard. Environ. 373 (1), 391–403.
 R. Fan, A.R. Gerson / Hydrometallurgy 134–135 (2013) 102–109 109

Liu, H., Gillaspie, J.D., Lewis, C.A., Neudorf, D., Barnett, S., 2002. Atmospheric pressure Mudd, G.M., 2010. Global trends and environmental issues in nickel mining: sulfides
leaching process for lateritic nickel ore. International Patent Number. versus laterites. Ore Geol. Rev. 38 (1–2), 9–26.
Liu, K., Chen, Q., Hu, H., 2009. Comparative leaching of minerals by sulphuric acid in a Parkhurst, D.L., 1995. User's guide to PHREEQC-a computer program for speciation,
Chinese ferruginous nickel laterite ore. Hydrometallurgy 98 (3–4), 281–286. reaction-path, advective-transport, and inverse geochemical calculations. US
Liu, K., Chen, Q., Hu, H., Yin, Z., 2010. Characterization and leaching behaviour of Geol. Surv. Water Resour. Inv. Rep.
lizardite in Yuanjiang laterite ore. Appl. Clay Sci. 47 (3–4), 311–316. Purwanto, H., Shimada, T., Takahashi, R., Yagi, J.I., 2003. Recovery of nickel from selec-
Llorca, S., 1993. Metallogeny of supergene cobalt mineralization, New Caledonia. Aust. tively reduced laterite ore by sulphuric acid leaching. ISIJ Int. 43 (2), 181–186.
J. Earth Sci. 40, 377–385. Rice, N.M., Strong, L.W., 1974. Leaching of lateritic nickel ores in hydrochloric acid. Can.
Luo, W., Feng, Q., Ou, L., Zhang, G., Lu, Y., 2009. Fast dissolution of nickel from a Metall. Q. 13 (3), 485–493.
lizardite-rich saprolitic laterite by sulphuric acid at atmospheric pressure. Hydro- Senanayake, G., Das, G.K., 2004. A comparative study of leaching kinetics of limonitic
metallurgy 96 (1–2), 171–175. laterite and synthetic iron oxides in sulfuric acid containing sulfur dioxide. Hydro-
Luo, W., Feng, Q., Ou, L., Zhang, G., Chen, Y., 2010. Kinetics of saprolitic laterite leaching metallurgy 72 (1–2), 59–72.
by sulphuric acid at atmospheric pressure. Miner. Eng. 23 (6), 458–462. Tartaj, P., Cerpa, A., Garcia-Gonzalez, M.T., Serna, C.J., 2000. Surface instability of ser-
Manceau, A., Schlegel, M.L., Musso, M., Sole, V.A., Gauthier, C., Petit, P.E., Trolard, F., pentine in aqueous suspensions. J. Colloid Interface Sci. 231 (1), 176–181.
2000. Crystal chemistry of trace elements in natural and synthetic goethite. Whittington, B.I., Muir, D.M., 2000. Pressure acid leaching of nickel laterites: a review.
Geochim. Cosmochim. Acta 64 (21), 3643–3661. Miner. Process. Extr. Metall. Rev. 21, 527–599.
McDonald, R.G., Whittington, B.I., 2008. Atmospheric acid leaching of nickel laterites
review. Hydrometallurgy 91 (1–4), 35–55.