Journal of Asian Earth Sciences 93 (2014) 74–88

Contents lists available at ScienceDirect

Journal of Asian Earth Sciences

journal homepage: www.elsevier.com/locate/jseaes

Mineralogical and geochemical characteristics of a serpentinite-derived

laterite profile from East Sulawesi, Indonesia: Implications for the
lateritization process and Ni supergene enrichment in the tropical
rainforest
Wei Fu a,b,⇑, Jianwen Yang a,c, Mengli Yang a, Baocheng Pang a,b, Xijun Liu a,b, Hujie Niu a, Xiaorong Huang a
a
College of Earth Sciences, Guilin University of Technology, Guilin, Guangxi 541004, China
b
Guangxi Key Laboratory of Hidden Metallic Ore Deposits Exploration, Guilin, Guangxi 541004, China
c
Department of Earth and Environmental Sciences, University of Windsor, Windsor, Ontario N9B3P4, Canada

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

Article history: To evaluate the lateritization process and supergene Ni enrichment in the tropical rainforest, a well
Received 14 September 2013 developed laterite profile over the serpentinite in the Kolonodale area of East Sulawesi, Indonesia, has
Received in revised form 31 May 2014 been investigated using field geology methods, mineralogical and geochemical techniques. Three litho-
Accepted 30 June 2014
stratigraphic horizons over the bedrock are distinguished from bottom to top: the saprolite horizon,
Available online 16 July 2014
the limonite horizon, and the ferruginous cap. In general, the profile is characterized by (1) a depth-
related pH ranging from 5.56 to 8.56, with a higher value in the saprolite horizon and a lower value in
Keywords:
the ferruginous cap, (2) a highly variable organic matter concentration from 1.11% to 4.82%, showing a
Lateritization process
Ni supergene enrichment
increasing trend from bottom to top, (3) a progressive mineral assemblage transition from the silicate
Serpentinite mineral dominant (mainly serpentine) to the Fe-oxyhydroxide dominant (mainly goethite), and (4) a typ-
Garnierite ical laterite geochemical pattern with an increase of Fe, Al, Mn, Cr and Ti but a decrease of Mg, Ca, Na and
Indonesia K upward from the bedrock. The highest concentration of Ni (up to 11.53%NiO) occurs in the saprolite
horizon, showing nearly 40 times richer compared to the bedrock. The mineral evolution during the lat-
eritization process shows various paths from the primary minerals to altered minerals, which is integrally
affected by the nature of the primary minerals and environmental conditions. Garnierite, as a significant
ore mineral formed by the secondary precipitation processes in the study profile, is identified as a mix-
ture of talc- and serpentine-like phases. The mass-balance calculation reveals that there are diversified
elemental behaviors during the serpentinite lateritization under the rainforest conditions. In particular,
Ni, as the ore-forming element in the laterite profile, is associated closely with the pH environment,
organic matter concentration and mineral evolution during the lateritization process.
The findings of the present study support a four-stage evolutional model for the lateritization process:
the ferruginous saprolite development (stage I), the limonite development (stage II), the silicate saprolite
and ferruginous cap development (stage III), and the precipitation of secondary minerals (stage IV). Due
to this multistage process, there is a progressive Ni-enrichment in the laterite profile.
Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction
Laterite soils derived from ultramafic rocks have been well doc-
umented, not only due to the scientific interest in understanding
the supergene processes on the surface of the Earth (Wilson,
2004; Garnier et al., 2009), but also to the economical significance
⇑ Corresponding author at: College of Earth Sciences, Guilin University of
Technology, Guilin, Guangxi 541004, China. Tel./fax: +86 0773 5897019.
E-mail address: fuway59@163.com (W. Fu).
of the laterite-type nickel ore (Elias, 2002; Gleeson et al., 2004;
Dalvi et al., 2004; Mudd, 2010; Butt and Cluzel, 2013). Nickel
deposit of this type accounts for over 60% of the world’s nickel
resource and nearly 40% of the annual global nickel production
(Berger et al., 2011). Globally, there are extensive outcrops of
ultramafic rocks in tropical or subtropical environments at present
and along the geological history, and these places tend to develop
Ni-laterite soils, such as in New Caledonia, Western Australia, West
Africa, Central America, Brazil and Balkans. Such Ni-laterite soils
have received a great attention in economic geology, and a large

http://dx.doi.org/10.1016/j.jseaes.2014.06.030
1367-9120/Ó 2014 Elsevier Ltd. All rights reserved.
 W. Fu et al. / Journal of Asian Earth Sciences 93 (2014) 74–88
number of related articles have been published in the last decade
(e.g., Gleeson et al., 2004; Freyssinet et al., 2005; Fouateu et al.,
2006; Lewis et al., 2006; Wells et al., 2009; Thorne et al., 2009,
2012; Golightly, 2010; Roqué-Rosell et al., 2010; Butt and Cluzel,
2013).
In Asia, Ni-laterite soils overlying the ultramafic rocks are well
developed in the countries of Southeast Asia, especially Philippines
and Indonesia (Brand et al., 1998; Elias, 2002), where numerous
obducted ophiolite sheets are exposed to the intensive chemical
weathering conditions of high rainfall and warm temperatures.
Many large scale Ni-laterite deposits have been discovered, such
as Sorowako, Pomalaa and Surigao in this area, which contributes
to make Philippines and Indonesia the third and fourth biggest
individual Ni-laterite ore reserves in the world, respectively
(Dalvi et al., 2004). However, the geological characteristics of the
Ni-laterite soils over the ultramafic rocks in Southeast Asia,
particularly in Indonesia, have not been fully studied. Until now,
Indonesia has focused the attention mainly on the Sorowako area,
Sulawesi Island (Golightly, 1979, 1981; Sufriadin et al., 2011). In
addition, the knowledge of lateritization process related to the
formation of the Ni-laterite is still tenuous since their geological
characteristics vary from one to the other (Golightly, 2010).
The present study deals with a newly discovered profile derived
from serpentinite, which is located in Kolonodale area, Sulawesi
Island, Indonesia. We present a preliminary evaluation of our
new data about field geology, mineralogy and geochemistry, pro-
viding a detailed information about the Ni-laterite soil in Indone-
sia. Also, our work helps to get a deep insight into the
lateritization process and Ni enrichment mechanism during the
serpentinite weathering in the tropical rainforest, which is useful
for the future exploration of Ni resources in this region.
2. Geological and geographical setting
The K-shaped Sulawesi Island is at the convergence zone of
three tectonic plates: Eurasian, Pacific, and Indian–Australian.
Due to strong tectonic activities, the geological structure of
Sulawesi Island is very complicated. Four lithotectonic belts are
identified in this island (Fig. 1): the West Sulawesi Volcano-Plu-
tonic Arc Belt, the Central Sulawesi Metamorphic Belt, the East
Sulawesi Ophiolite Belt, and the Continental Fragments of Bang-
gai-Sula, Tukang Besi, and Buton (Mubroto et al., 1994). Our study
site, the Kolonodale area in Central Sulawesi province, is located in
the East Sulawesi Ophiolite Belt.
The East Sulawesi Ophiolite Belt represents a part of the Circum
Pacific Phanerozoic multiple ophiolite belts, which were emplaced
through convergent plate processes in Cretaceous to Miocene (Hall
and Wilson, 2000). It extends 700 km from north to south and out-
crops more than 15,000 km2 (Kadarusmana et al., 2004). Ultra-
mafic bodies are widely exposed within this belt, consisting of
lherzolite, harzburgite, and peridotite, with some dunite, pyroxe-
nite and gabbroic dikes.
In the study area, peridotite is the dominant rock type in the
ultramafic complex, and it has been widely metamorphosed at a
moderate-high grade to form serpentinite. There are two other
lithological units occurring in the surroundings of the ultramafic
complex: alluvial and sedimentary lacustrine rocks of Quaternary
and Cretaceous sedimentary rocks (Kadarusmana et al., 2004).
Faults, fractures and joints, striking mainly in NNW direction, are
largely observed in the ultramafic complex due to the regional tec-
tonic activities.
Geographically, the climatic condition in the study area is
typical of humid tropical region, with a rainfall of about
2000 mm per year and an average temperature of about 26 °C.
The high temperature and abundant rainfall support a lush vegeta-
75
tion over the entire region. The terrain of the study area is charac-
terized by a low-moderate relief with gently sloping surfaces.
Under this environment, the outcrop of ultramafic rocks is subject
to chemically weathering and forms thick lateritic soils.
3. Field regolith geology
In the study area, serpentinite, as the main bedrock, outcrops
along the edges of the highlands and it can also be observed in
the bottom of some shallow exploration wells. The outcrops of ser-
pentinite are characteristics of dark green to black, massive texture
with smooth surfaces, and usually broken up into irregular frag-
ments along the fractures and joints.
From the bedrock upward, the laterite soil is widely developed
mainly along the gentle slopes, ridges, and platforms of the high-
land. It occurs with a variable thickness and may be continuous
or not. The average thickness is about 10 m, and the thickest part
may exceed 30 m. A typical well developed laterite profile can be
divided into three lithostratigraphic horizons based on the varia-
tions in color, texture and mineralogy. They are termed as saprolite
horizon, limonite horizon and ferruginous cap from bottom to top
(Fig. 2). The details of each horizon are as follows.
3.1. The saprolite horizon
It overlies the bedrock with an irregular and gradual contact.
This horizon is commonly well developed along the gentle slopes,
but to a less extent in the highland. It is fairly heterogeneous with a
variable thickness ranging from 1 to 5 m, and with mixed colors
such as yellowish-brown, grayish-yellow and grayish-green. The
saprolite horizon is loose, porous and friable, and contains different
proportions of fine-grained groundmass (earthy saprolite) and
coarse-grained fragments or blocks (rocky saprolite) (Fig. 3C).
The primary structure of the bedrock can be partly observed in
the lower part of the saprolite horizon where the rocky blocks
are abundant, whereas it nearly disappears in the higher part of
this horizon in which the earthy groundmass are dominant due
to a stronger weathering.
The silicate minerals (verified as the serpentine and smectite in
the following analyses) are the dominant component of the
saprolite horizon, which is similar to many silicate-type Ni-laterite
profiles (Brand et al., 1998). Thus, this saprolite horizon is simply
termed as silicate saprolite in this study, as mentioned by
Freyssinet et al. (2005). It should be noted that there is another
type of saprolite horizon, which is found locally in the shallow
laterite profiles and its earthy saprolite is dominated by the ferrig-
unous minerals (verified as the goethite and hematite in the fol-
lowing analyses). For comparison, the latter is termed as
ferruginous saprolite corresponding to the description in the
Cawse profile, Australia (Brand et al., 1998).
Garnierite, as a striking weathering product, is found in the sap-
rolite horizon of our study profile. It shows a jade-green color and a
vein-like texture (Fig. 3D), and mainly occurs in the joints and frac-
tures of the lower part of the saprolite horizon. Locally, the garni-
erite vein may extend into the underlying bedrocks along the fault
zone.
3.2. The limonite horizon
This horizon is developed above the saprolite horizon with a
visible contact in color. Its thickness ranges from 1 to 10 m, and
it is preferentially developed in the flat relief. This horizon is
brownish-red to brownish-yellow in color, soft, porous, and fine-
grained. Main features of the original mineral components and
the structures of the bedrock have been destroyed. The mineralogy
76 W. Fu et al. / Journal of Asian Earth Sciences 93 (2014) 74–88
Fig. 1. Geological setting of Sulawesi Island (Indonesia), showing the location of the Kolonodale Ni-laterite profile (modified after Kadarusmana et al., 2004).
of this horizon is dominated by goethite and hematite. Addition-
ally, silica veins in stockwork can be observed locally in the bottom
of this horizon (Fig. 3B).
3.3. The ferruginous cap
It is the uppermost unit, directly overlying the limonite horizon
with a clear contact. This horizon is well preserved in the highland,
while it is commonly eroded away in the deep slope. The weather-
ing products are soft and incoherent, being reddish brown in color
and with massive or colloidal textures. Some ferruginous concre-
tions occur broadly in this horizon, showing variable diameters
ranging from 2 to 20 cm (Fig. 3A), and the diameter becomes smal-
ler and smaller downward. These concretions are dominantly
made up of Fe-oxyhydroxides.
4. Samples and analytical methods
Nine samples were collected from the study profile along the
different horizons from bottom to top as follows: the bedrock
sample (L1-7), the saprolite samples (L1-4, L1-5, L1-6), the garnier-
ite sample (K1-2) in the saprolite horizon, the limonite samples
(L1-2, L1-3), the silica stockwork sample in the limonite horizon
(D1-3) and the ferruginous cap sample (L1-1), as shown in Fig. 2.
All the samples were examined under optical microscopy and also
analyzed by pH and organic matter (OM) measurements, X-ray
diffraction (XRD) and bulk geochemical techniques. Some solid
samples including saprolite (L1-5), garnierite (K1-2) and bedrock
(L1-7) are selected for mineral chemistry study by electron micro-
probe analysis (EMPA).
Petrographical observations were performed in thin sections
using a Leica DM EP optical microscope under the plane- and
crossed-polarized light. In order to preserve the integrity of the
samples containing free particles, they were included in epoxy
resin.
Prior to the pH and organic matter measurements, the samples
were air dried, crushed and sieved through a 2–0.25 mm sieve. The
pH values were measured at a soil to water mass ratio of 1:2.5
using the Mettler Toledo acidity meter with a precision better than
0.01. The organic matter was determined by the dichromate
 W. Fu et al. / Journal of Asian Earth Sciences 93 (2014) 74–88 77

Fig. 2. The laterite profile from the study site at the Kolonodale area, including the location of the samples. L-1, Purple-red, friable, porous, low density and generally with
earthy texture, easy forms aggregates after long exposure; L-2, Brownish-red, friable, porous, low density and generally earthy or nodular texture, easy forms aggregates; L-3,
Brownish, friable, low density and mainly earthy texture; D1-3, Whitish, moderately dense, showing yellow clayey materials cemented by intensive white silica stockworks;
L-4, Yellow, friable, moderately dense, earthy and massive texture with few rocky fragments; L-5, Sage green, moderately hard and dense, earthy and massive texture, mixed
with some rocky fragments; K1-2, Green, moderately dense and generally massive texture, some clayey materials are trapped; L-6, dark brown, moderately hard and dense ,
mixed with many clayey and rocky fragments; L-7, Black, high density and massive texture. (For interpretation of the references to color in this figure legend, the reader is
referred to the web version of this article.)

oxidation method (Nelson and Sommers, 1982). Briefly, one gram
of soil was oxidized at 150 °C for 1 h with 25 ml of a mix of potas-
sium dichromate and sulfuric acid. Then, a reverse titration of Cr2-
O2
7 ions was made using an acidified ferrous ammonium sulfate
solution.
X-ray diffraction data were obtained by using a Philips X’ Pert
MPD diffractometer, working on Cu target at 40 kV and 40 mA.
The range of 2h scanning was from 5° to 80°. The scan step and step
duration were 0.05° and 3 s, respectively. A standard grinding and
mounting of the sample was performed prior to XRD. In order to
confirm the presence of the expandable clay silicates (e.g., smec-
tite), one sample (L-6) was further analyzed under different treat-
ments, including a ethylene glycol saturation treatment (EG) and a
heating treatment to 550 °C(HT).
All nine samples for whole rock geochemical analyses were first
dried at 70 °C, and then both the dried weathering products and
the fresh bedrock were ground into powder. The powder was
baked at 105 °C to remove adsorbed water before analysis. Major
elements were measured using a Rigaku ZSX100e X-ray fluores-
cence (XRF). Analytical precision for major elements is less than
1%, and the determination limits for the major elements are gener-
ally better than 30 ppm. Trace elements were measured using a
Perkin–Elmer Elan 6000 ICP-MS with detection limits of about sev-
eral pg/g in solution, corresponding to a determination limit of
about 10 ppb in solid samples for trace elements. Analytical preci-
sion for trace elements is less than 5%.
Selected mineral phases were analyzed in situ on polished sec-
tions for their chemical compositions using JXA-8230 electron
microprobe with a wavelength dispersive system. The measure-
ment conditions were 20 kV accelerating voltage, 15 nA probe cur-
rent and 2 lm beam diameter. Calibrations were performed using
natural and synthetic reference materials: forsterite (Mg, Fe and
Si), anorthite (Ca), albite (Na, Al), chromite (Cr), phlogopite (K),
rutile (Ti), NiO (Ni), CoO (Co) and MnO (Mn).
The petrography, pH, organic matter, XRD and EMPA were all
conducted in the Key Laboratory of the Guangxi Hidden Metallic
Ore Deposits Exploration in China. In addition, the whole rock
geochemical analyses were carried out in the Key Laboratory of
Geochronology and Geochemistry, Chinese Academy of Sciences.
5. Results
The analytical results of pH, organic matter (OM) and whole
rock geochemical data are listed in Table 1. The mineral chemistry
data are listed in Table 2, and the XRD results are illustrated in
Fig. 4.
5.1. pH
The overall pH values of the samples fall in the range of 5.56–
8.56 with the highest value in the ferruginous cap (L1-1) and the
lowest value in the bedrock (L1-7). Fig. 5 shows the depth-related
pH distribution pattern along the profile. There is a change from a
lower pH value (5.58–5.8) in the ferruginous cap and limonite hori-
zon to a higher pH value (6.98–8.56) in the saprolite horizon and
bedrock, indicating the soil environment is transformed from the
acidity to the slightly alkalinity with the profile downward.
5.2. Organic matter
The organic matter concentration varies broadly from 1.11%
to 4.82% among the weathered samples with the highest value
in the ferruginous cap (L1-1) and the lowest value in the limo-
78 W. Fu et al. / Journal of Asian Earth Sciences 93 (2014) 74–88

Fig. 3. Macrophotographs and microphotographs of the study samples from the Kolonodale laterite profile Macrophotographs: (A), large-size (>5 cm in diameter) ferruginous
concretions developed on the surface of the laterite crust; (B), silica stockwork are frequent at the lower part of the limonite horizon; (C), rocky saprolite fragments
surrounded by earthy saprolite; (D), garnierite occurs as vein fillings along joints; (E), serpentinite with intensively joints; Microphotographs: (F–G), mixtures of Gth and Hem
in the ferruginous cap; (H), a small amount of residual Srp in the limonite horizon are altered to Gth and Hem; (I), Gth and Hem are cemented by Qz ; (J–K), the fragments of
bedrock containing Px and Srp; Srp, Gth and Hem are coexisting in the saprolite sample; (L), Srp with the booklet structure, and Ker with the nodule and zoned structure are
cemented by Qz in the garnierite sample; (M), two stages of Ker aggregates occur as circles with different colors; (N), the primary minerals in the bedrock including Srp, Ol
and Px; (O), Ol in the bedrock are partly altered to Idn. (Px-pyroxene, Ol-olivine, Srp-serpentine, Idn-iddingsite, Ker-kerolite, Qz-quartz, Gth-goethite, Hem-hematite).

nite sample (L1-3). The vertical distribution of the organic work sample (D-3) contains a higher organic matter concentra-
matter displays a decreasing trend, showing a great drop from tion than that of its neighbouring samples (L1-3 and L1-4), and
the ferruginous cap (4.82%) to the limonite horizon (avg. that the garnierite sample (K1-2) exhibits a similar feature as
1.93%) and the saprolite horizon (avg. 2.2%). The silica stock- well.
 W. Fu et al. / Journal of Asian Earth Sciences 93 (2014) 74–88 79

Table 1
Whole rock analyses of the samples in the Kolonodale laterite profile, including organic matter (OM, wt.%), major element (wt.%) and trace element (in ppm).

Horizon Ferruginous cap Limonite Saprolite Bedrock

Sample L1-1 L1-2 L1-3 D1-3 L1-4 L1-5 K1-2 L1-6 L1-7
pH 5.56 5.61 5.8 7.66 8.03 7.03 6.98 7.19 8.56
OM 4.82 2.74 1.11 2.22 2.11 2.7 2.67 1.75 3.36
Al2O3 11.6 11.96 8.75 0.19 2.61 4.56 0.16 4.22 1.87
CaO 0.03 0.03 0.03 0.06 0.41 0.69 0.04 0.1 1.49
Fe2O3 63.45 63.22 66 8.36 22.65 19.13 3.58 32.5 8.91
K2O 0.02 0.03 0.01 0.02 0.01 0.03 0.03 0.01 0.02
MgO 0.55 0.44 0.65 0.45 21.81 16.45 15.9 7.2 35.97
MnO 0.28 0.36 0.32 0.13 0.28 0.21 0.08 0.51 0.14
Na2O 0.24 0.45 0.08 0.1 0.12 0.54 0.41 0.1 0.2
P2O5 0.02 0.02 0.01 0.01 b.d. 0.01 b.d. b.d. 0.01
SiO2 3.71 3.21 2.79 87.48 35.89 45.29 60.34 39.65 42.09
TiO2 0.38 0.41 0.19 0.04 0.05 0.11 0.06 0.09 0.04
Cr2O3 5.36 4.99 6.58 0.44 1.11 1.1 0.24 2.5 0.62
NiO 0.72 0.68 1.1 0.13 3.25 2.46 11.53 1.77 0.26
L.O.I 13.27 14.02 13.52 2.41 11.75 9.85 7.86 11.35 9.19
Total 99.63 99.79 100.04 100.07 99.93 100.41 100.23 99.9 100.78
Cs 0.04 0.03 0.02 0.07 0.08 0.06 0.29 0.04 0.09
Rb 0.33 0.23 0.11 1.34 0.09 0.1 0.81 0.11 1.25
Ba 1.17 2.88 10.65 11.47 12.81 15.24 1.82 17.39 4.22
Th 2.47 3.16 0.31 0.03 0.01 0.01 0.01 0.01 0.05
U 0.98 0.85 0.14 0.29 0.24 0.35 0.01 0.06 0.02
Ta 0.18 0.21 0.03 0.02 0.01 0.01 0.02 0.15 0.07
Nb 2.24 2.8 0.36 0.13 0.05 0.05 0.14 0.07 0.16
Sr 2.49 1.65 0.04 2.95 0.52 0.75 0.1 0.69 0.47
Hf 1.24 1.49 0.2 b.d. 0.05 0.08 b.d. 0.05 0.02
Zr 43.92 55.41 5.55 0.02 0.22 0.56 0.04 0.33 0.18
Y 2.83 2.95 2.46 0.73 64.92 9.62 1.18 2.75 1.33
Pb 7.23 9.62 4.67 0.95 0.94 0.45 2.52 0.21 3.01
Co 339.9 327.1 294.8 16.57 352.1 187.8 75.85 517.5 102
Cu 72.28 69.06 84.62 8.59 26.57 46.99 4.42 48.12 15.41
Ga 12.8 12.91 8.28 0.29 2.54 4.45 0.28 4.29 1.73
Ge 3.49 3.42 3.81 1.38 3.6 2.47 17.2 2.57 1.09
Sc 69.97 63.78 76.28 3.41 23.87 28.48 2.59 32.42 12.51
V 264.8 257 258.6 28.69 81.16 85.39 20.92 151.1 55.56
Zn 189.1 175.5 251.4 21.73 445.7 89.98 248 165.3 37.43

Note: b.d. means below detection.

5.3. Mineralogy
Petrographical observations show that the serpentine, most
likely lizardite, is the dominant primary mineral (>75%) in the
bedrock. It is characterized by a strip-like mineral shape. Minor
primary minerals are olivine (<10%), orthopyroxene (<15%), and
in some cases, clinopyroxene occurring as exsolution in the ortho-
pyroxene. These minerals crosscut each other and are surrounded
by the serpentine veins, forming a mesh (Fig. 3N). Some olivine
crystals are altered to iddingsite (Fig. 3O), and the large pyroxene
crystals are usually cut into many irregular networks of fractures
in which serpentine occurs. In addition, trace chromite grains can
be observed among the serpentine.
Both the petrographical observations and XRD analyses reveal
that there is a mineral variation within the saprolite horizon. In
the lower part of this horizon, serpentine is the most abundant
mineral type, which is inherited from the bedrock and relatively
well preserved. However, olivine and pyroxene is only slightly
preserved in this position, showing as the relict mixed with the
serpentine (Fig. 3K). It is worthy to note that a significant amount
of smectite was detected in two saprolite samples (L-5 and L-6) by
the XRD analysis. It is identified based on its characteristic peak at
16.5 Å for the normal sample and a swelling peak at 18.7 Å after
treatment with ethylene glycol. By heating up to 550 °C, this char-
acteristic peak collapsed and gave a broad peak at 10 Å (Fig. 8). In
contrast, there is a different mineral assemblage in the upper part
of the saprolite horizon, where goethite, hematite and serpentine
are coexisting together (Fig. 3J).
Result of the XRD analysis show that the garnierite sample (K-2)
is a mixture of talc- and serpentine-like phase, evidenced by the
0
presence of a strong peak respectively at 10 Å A and 7 Å. The charac-
teristic peak of the talc-like phase shows a better match with ker-
olite-pimelite series, a hydrated talc with general formula
(Mg,Ni)3Si4O10(OH)2nH2O, and that of the serpentine-like phase
corresponds to serpentine-nepouite series with general formula
(Mg,Ni)30 Si2O5(OH)4 (Brand et al., 1998; Freyssinet et al., 2005).
The 10 Å A peak with a broad range also indicates that the talc-like
mineral has a relatively low crystallinity (Fig. 4). Under the micro-
scope, the serpentine-like mineral is characterized by a lath-shape
and bundle texture (Fig. 3L), whereas the talc-like mineral exhibits
a colloidal texture with quartz crystals filling voids (Fig. 3M).
In contrast to the saprolite horizon, the limonite horizon has a
relatively simple mineral assemblage, mainly consisting of fine-
grained goethite and hematite. A small amount of serpentine has
been preserved in this horizon (Fig. 3H). Locally, secondary quartz
may be abundant in the silica stockwork (Fig. 3I). The ferruginous
cap is dominated by hematite and goethite (Fig. 3F and G), accom-
panied by minor chromite and quartz.
In general, the mineral composition in the Kolonodale profile,
from bottom to top, shows a progressive transition from the
silicate dominant to the Fe-oxyhydroxide dominant, with some
sudden changes (like garnierite and quartz) in some specific hori-
zons (Fig. 5).
5.4. Mineral chemistry
The EMPA data reveal that the mineral chemistry of the primary
minerals (serpentine, pyroxene and olivine) in the bedrock mainly
exhibits a variation in Mg, Fe and Si concentration. The olivine is
richer in Mg (avg. 48.33% MgO) and Fe (avg. 9.15% FeO), whereas
80 W. Fu et al. / Journal of Asian Earth Sciences 93 (2014) 74–88

Table 2
Electron microprobe composition of the primary and weathered minerals in the Kolonodale profile (wt.%).

Mineral types Chemical composition

MgO Al2O3 Na2O CaO NiO K2O CoO FeO MnO Cr2O3 TiO2 SiO2 Total
Primary minerals in the bed rock Serpentine 1 33.92 4.87 0 0.05 0.08 0 0 6.72 0.11 0.59 0.06 45.31 91.71
2 25.4 0 0 0.58 0.49 0 0 14.78 0.05 0.02 0.03 47.45 88.8
3 24.05 0 0.02 1.1 0.28 0 0 14.3 0.01 0.32 0.01 52.1 92.18
Avg. 27.79 1.62 0.01 0.58 0.28 0 0 11.93 0.06 0.31 0.03 48.29 90.9
Pyroxene 1 15.93 6.16 0.29 22.15 0.03 0 0 2.41 0.05 1.13 0.19 52.2 100.54
2 32.57 5.33 0.02 0.87 0.1 0 0 5.77 0.14 0.71 0.1 55.99 101.61
3 32.55 5.02 0.02 0.87 0.07 0 0 5.98 0.15 0.6 0.08 55.53 100.88
Avg. 27.02 5.5 0.11 7.97 0.07 0 0 4.72 0.12 0.81 0.12 54.57 101.01
Olivine 1 48.41 0 0 0.01 0.33 0 0 8.93 0.07 0.02 0.01 42.06 99.81
2 47.86 0 0 0.01 0.38 0 0 9.17 0.13 0 0.04 41.19 98.78
3 48.72 0.01 0 0.02 0.47 0 0.01 9.36 0.12 0.02 0.03 40.98 99.74
Avg. 48.33 0 0 0.01 0.39 0 0 9.15 0.11 0.01 0.03 41.41 99.44
Weathered minerals in the Serpentine 1 33.01 0.51 0.03 0.09 1.83 0.01 0.03 7.76 0.14 0 0 43.22 86.62
saprolite horizon
2 33.53 0.3 0.03 0.11 2.11 0.02 0.01 7.5 0.03 0.07 0 43.07 86.78
3 33.09 0.99 0.01 0.09 1.9 0 0 7.29 0.01 0.05 0.01 42.81 86.24
Avg. 33.21 0.6 0.02 0.1 1.95 0.01 0.01 7.52 0.06 0.04 0 43.03 86.55
Minerals in the garnierite Talc-like mineral 1 16.21 0.01 0.03 0.18 17.97 0.04 0 0.61 0.03 0.03 0.09 53.44 88.64
2 18.79 0.03 0 0.19 15.09 0 0 0.81 0.01 0 0 52.05 86.97
3 18.41 0.01 0.06 0.09 14.91 0.03 0.01 0.59 0 0 0.16 55.68 89.94
Avg. 17.8 0.02 0.03 0.15 15.99 0.02 0 0.67 0.01 0.01 0.08 53.72 88.52
Serpentine-like 1 31.84 0.03 0.01 0.11 6.21 0 0.02 5.25 0.04 0.07 0 41.66 85.22
mineral
2 30.65 0.02 0.05 0.09 5.88 0.02 0 5.23 0 0 0.03 41.88 83.85
3 28.69 0.02 0.08 0.11 5.81 0.02 0 5.74 0 0.03 0.09 42.3 82.88
Avg. 30.39 0.02 0.04 0.1 5.97 0.01 0.01 5.4 0.01 0.03 0.04 41.95 83.98

pyroxene and serpentine contain more Si (avg. 54.57% and 48.29%
SiO2, respectively). Apart from Mg, Si and Fe, a small amount of Al
concentrations (<6%) is also present in pyroxene and serpentine. Ni
is the significant trace element in these three primary minerals
with 0.08–0.49% NiO (avg. 0.29%) in serpentine, 0.03–0.1% NiO
(avg. 0.07%) in pyroxene and 0.33–0.47% NiO (avg. 0.39%) in oliv-
ine, respectively. Co is poor in all these primary minerals (generally
near its detection limit).
There is no significant mineral chemistry difference in major
composition (Mg, Si and Fe) between the residual serpentine from
the saprolite horizon and the primary serpentine from the bedrock.
However, their Ni concentrations vary greatly. The residual serpen-
tine contains a larger Ni concentration (1.83–2.11% NiO, avg.
1.95%), about 6.7 times higher than that of the primary serpentine
(0.08–0.49% NiO, avg. 0.28%).
A high Si, Mg and Ni concentration (especially Ni) is the marked
chemical characteristic of the serpentine- and talc-like minerals in
the garnierite sample. The serpentine-like minerals contain
5.81–6.21% NiO (avg. 5.97%), and the talc-like minerals contain
14.91–17.97% NiO (avg. 15.99%). Their Ni concentration is nearly
10–600 times higher than that of the primary minerals (0.03–
0.49% NiO) in the bedrock. Apart from Si, Mg and Ni, a significant
amount of Fe concentration (avg. 5.4% FeO) is also present in the
serpentine-like minerals, while it is lower in the talc-like minerals
(avg. 0.67% FeO). Other elements detected in the serpentine-like
and talc-like minerals are present at near trace amounts (i.e.,
<0.1%).
In comparison with previous works (Brindley and Hang, 1973;
Manceau and Calas, 1985; Wells et al., 2009), the chemical compo-
sition of the serpentine-like minerals in our garnierite sample plot
within the field between the kerolite-pimellite end members, and
that of the talc-like minerals plot within the field between the ser-
pentine-nepouite end members. According to the 50% rule for
defining the solid solution minerals (Nickel, 1992) and the conven-
tion for classification of the Ni-bearing minerals in the laterite soils
(Brand et al., 1998) , the Ni/Mg mole ratio in the serpentine-like
minerals is in the range of 0.1–0.11, which is less than 1 and
indicates these serpentine-like minerals can be defined as
Ni-serpentine. Similarly, the talc-like minerals examined in the
garinierite sample can be defined as Ni-kerolite for their Ni/Mg
mole ratios ranging in 0.43–0.6.
5.5. Whole-rock geochemistry
The vertical distributions of major elements and some interest-
ing trace elements (e.g., Ni and Cr) through the study profile are
shown in Fig. 6.
The bedrock (L1-7) is characterized by a high concentration of
Mg (35.97% MgO) and Si (42.09% SiO2), coupled with a low concen-
tration of Fe (8.91% Fe2O3) and Al (1.87% Al2O3). The Ni concentra-
tion of the bedrock is about 0.26% (NiO), similar to that of other
ultramafic rocks in East Sulawesi Ophiolite Belt (Kadarusmana
et al., 2004). The saprolite samples (L1-4, L1-5 and L1-6) show a
marked decrease in Mg (avg. 15.15% MgO) and an increase both
in Fe (avg. 24.76% Fe2O3) and Al (avg. 3.80% Al2O3). Its average Ni
concentration (2.49% NiO) is the highest among the different litho-
stratigraphic horizons. In particular, a single garnierite sample (K-
2) in the saprolite horizon exhibits a striking Ni concentration
(11.53% NiO), which is over 44 times than that of the bedrock.
The limonite samples (L1-2 and L1-3) exhibit a further decrease
in Mg (avg. 0.55% MgO) and Si (avg. 3.0% SiO2) and an increase in
Fe (avg. 64.61% Fe2O3) and Al (avg. 10.36% Al2O3) compared to
the saprolite samples. The average Ni concentration of the limonite
samples is 0.89% (NiO), which is slightly low among the weathered
samples. In the uppermost horizon, the ferruginous cap sample
(L1-1) shows a comparable chemical composition with that of
the limonite horizon, with a high Fe (63.45% Fe2O3) and Al (11.6%
Al2O3) but a low Mg (0.55% MgO) and Si (3.71% SiO2) concentration.
The Ni concentration of this horizon is only 0.72% (NiO).
 W. Fu et al. / Journal of Asian Earth Sciences 93 (2014) 74–88 81

Fig. 4. XRD patterns of the samples from different horizons of the laterite, showing their mineral components. (Gth-goethite, Hem-hematite, Sme-smectite, Srp-serpentine,
Qz–quartz, Tlc-talc).

Fig. 5. Variations of the pH value, organic matter concentration (OM) and mineralogy composition along the laterite profile. (Gth-goethite, Hem-hematite, Sme-smectite, Srp-
serpentinite, Gar-garnierite, Qz–quartz, Ol-olivine, Px-pyroxene).

Fig. 7 demonstrates the enrichment or depletion status of the 6. Discussions

major and trace elements in the weathered products in comparison
with the bedrock. It is clear that most of the major and trace ele-
ments show an enrichment with various extents in all the litho-
stratigraphic horizons, represented by Fe, Ti, Cr, Zr and Hf,
whereas Mg, Ca, Si, Rb and Cs are depleted. There is also a fluctuat-
ing pattern regarding P, Ba, Pb and Th, that is, an enrichment in one
horizon but a depletion in the other.
6.1. Evolution of minerals
Primary minerals in the ultramafic rock may have various
weathering rates and sequences during the lateritization process
due to their different mineralogical nature (Wilson, 2004). Olivine
and pyroxene, in spite of occurring as the minor primary minerals
82 W. Fu et al. / Journal of Asian Earth Sciences 93 (2014) 74–88
Fig. 6. Variations of the major elements and some interesting trace elements through the laterite profile. (A-Ferruginous cap; B-Limonite horizon; C-Saprolite horizon; D-
Bedrock).
of the bedrock in the Kolonodale profile, are the earliest minerals
experiencing alteration when they exposed to the surface environ-
ment. The iddingsite observed by the petrographical analysis
confirms that the transformation from olivine to ferric mineral
has taken place. This transformation is most frequent when drain-
age in the laterite profile is good (Troly et al., 1979). As for pyrox-
enes, there is no direct petrographical evidence in our study for
illustrating what it has been altered to. However, previous studies
have reported that the general weathered product of pyroxenes is
smectite (Nahon et al., 1982; Nelson and Sommers, 1982; Colinf,
1990), and it is also identified by the XRD analysis in our saprolite
samples. Thus, the alteration from pyroxene to smectite may have
happened in the Kolonodale profile.
The alteration of serpentine generally commences after olivine
and pyroxene have disappeared (Golightly, 1981), and its alter-
ation path may be diversified depending on the supergene environ-
mental conditions. In a mildly alkaline environment, part of the
octahedral Mg in serpentine may be replaced by Ni due to the ionic
exchange reaction between the serpentine and weathering solu-
tion (Golightly, 1981; Elias et al., 1981). From the EMPA data men-
tioned above, we find that the Ni concentration of the residual
serpentine of the saprolite horizon is 6.7 times higher than that
of the primary serpentine of the bedrock. This evidence indicates
that the alteration of serpentine by the ionic exchange reaction
widely exists in the saprolite horizon of the Kolonodale profile.
The presence of a large amount of smectite in the lower part of
the saprolite horizon may have a close association with serpentine.
Alteration from serpentine to smectite in the laterite profile has
been documented in many laterite profiles (e.g., Elias et al., 1981;
Caillaud et al., 2009; Yongue-Fouateu et al., 2009). Such a weather-
ing alteration preferentially occurred where the water table is high
or drainage is poor (Golightly, 1981; Gleeson et al., 2003). In our
study area, the serpentinization of bedrock, together with the
low topographic relief, may contribute to the poor drainage
 W. Fu et al. / Journal of Asian Earth Sciences 93 (2014) 74–88 83

Fig. 7. Enrichment-depletion diagrams of the major and trace elements in the weathered products compared to the bedrock.

Fig. 8. The mass transfer coefficient (s) of some representative elements relative to the immobile element Ti along the laterite profile. (A-Ferruginous cap; B-Limonite
horizon; C-Saprolite horizon).
84 W. Fu et al. / Journal of Asian Earth Sciences 93 (2014) 74–88
environment, facilitating the formation of smectite. During the
process of the serpentine weathering, the smectite may be an
intermediate phase because it is unstable in a low pH environment
(Golightly, 1981), such as in the upper part of the laterite profile.
Thus, the smectite is liable to be replaced ultimately by Fe-oxyhy-
droxide (Oliveira et al., 1992).
The serpentine may have been directly altered to Fe-
oxyhydroxide (e.g., goethite) under an oxidized and acidic environ-
ment without undergoing an intermediate stage (Thorne et al.,
2009). Our mineralogical evidence suggests that the alteration
from serpentine to Fe-oxyhydroxide probably happened at the
Kolonodale profile, since there is a small amount of serpentine
coexisting with hematite and goethite in the lower part of the
limonite horizon, while no serpentine but only hematite and
goethite are preserved in the upper part of this horizon.
The Fe-oxyhydroxides are the common minerals during the
weathering of the ultramafic rocks. Previous studies proposed that
hematite is more stable than goethite in a high Eh and low pH envi-
ronment, and goethite may be turned into hematite by dehydra-
tion (Golightly, 1981). Thus, we think that the hematite may
represent a final weathered mineral in the lateritization process
of the Kolonodale profile, which is favored to keep stable in the fer-
ruginous cap due to its low pH and high OM environment. The lack
of either kaolinite or gibbsite in the Kolonodale profile might sug-
gest that the pH value at study site was not low enough for these
minerals to crystallise.
Garnierite is generally regarded as a product of the secondary
precipitation process from the weathering solution at a low tem-
perature condition (Brindley et al., 1977; Golightly, 1981;Cluzel
and Vigier, 2008; Wells et al., 2009; Gali et al., 2012). In the study
area, there are a number of geological and environmental factors
which have played the positive roles for the formation of the gar-
niertite. The serpentinite with a relative high Ni concentration
(0.29%NiO) provided the abundant Ni source. The rainforest cli-
mate condition is favorable for Ni transportation and accumulation
in the weathering solutions. The slightly alkaline (pH > 7) condi-
tions in the saprolite horizon is also in favor for the precipitation
of garnierite from the Ni-saturated weathering solution
(Golightly, 1981; Gali et al., 2012). Moreover, the joints and frac-
tures in the lower part of the profile may act as a good pathway
for the percolation of the Ni-saturated solution, or locally as the
precipitation space for the garnierite, although currently there
are no sufficient evidences to define them as the active syn-
weathering structures similar to those in New Caledonia (Cluzel
and Vigier, 2008).
6.2. Mobilization of elements and enrichment of Ni
The geochemical characteristics of the Kolonodale laterite pro-
file generally conforms the trends reported by previous studies
(e.g., Golightly, 1981; Trescases, 1986; Bárdossy and Aleva, 1990;
Das et al., 1999; Gleeson et al., 2004; Fouateu et al., 2006;
Thorne et al., 2009). During the lateritization process, the rock-
forming elements from the bedrock show complex behaviors,
which are affected by oxidation, hydrolysis, and carbonation pro-
cesses (Golightly, 1981; Gleeson et al., 2004; Fouateu et al.,
2006). Herein, we apply the mass balance calculation (e.g.,
Schellmann, 1989; Braun et al., 1993) to investigate the mobiliza-
tion of elements during the weathering process.
The study profile fits well into the application conditions of the
mass balance calculation proposed by Nahon and Merino (1996).
The saprolite horizon, the thickest horizon in the Kolonodale pro-
file, preserves the parent-rock texture and can be considered as
the product of isovolumetrically weathering. According to Braun
et al. (1993) and Anderson et al. (2002), the element mass transfer
coefficient s is given by the following relationship:
C j ; wC i ; p
s¼ 1
C j ; pC i ; w
where Cj is the concentration in weight percent of element j, Ci is
the concentration in weight percent of immobile element i, and
the subscripts w and p refer to weathered material and parent rock,
respectively. If s = 0, then the element j in a sample has no mobility
related to the parent rock. However, s > 0 denotes the external addi-
tion of element j. If s < 0, then the element j in a sample is lost dur-
ing weathering, and s = 1 means that the element j is completely
removed.
With respect to the selection of the immobile elements, it
depends on the type of the profile. Ti, Fe, Al, Zr, Nb, Ta, Sc and Th
are potential targets because of their conservative geochemical
behaviors (Grant, 1986; Braun et al., 1993; Baumgartner and
Olsen, 1995; Amundson, 2003; Brown et al., 2003). For the mafic
and ultramafic rocks, Cr, Zr, and Th have been used as the immobile
elements in previous works (Schellmann, 1989; Ma et al., 2007;
Gong et al., 2011). However, the choose of Cr as an immobile ele-
ment is critical for laterite soils, because of the high mobility of
Cr (e.g., Bergera and Frei, 2014).
By applying the Isocon-Diagram (Grant, 1986), we find that Ti is
a representative element in the ‘‘isocon’’ line and can be viewed as
the immobile element for the mass balance calculation in this
study profile. The calculated data are shown in Table 3, where sign
‘‘’’ represents the loss of elements, and ‘‘+’’ represents the gain. It
is worthy to note that our data from the mass balance calculation
are only interpreted semi-quantitatively in order to avoid any
over-interpretation, as proposed by Schellmann (1989).
As illustrated in Fig. 8, Mg shows an absolute loss pattern in the
laterite profile, which is lost over 77% (s = 0.77) in the saprolite
horizon, and even totally removed (s = 1) in the limonite horizon
and ferruginous cap. Besides Mg, many other major (e.g., Ca, K and
P) and trace (e.g., Pb) elements also display an absolute loss pat-
tern, and their depletion extent roughly follows an increasing trend
from the lower horizon to the upper horizon. The high loss of such
elements might be related to the alteration of the primary mineral
and the leaching process. The dissolution of olivine, pyroxene and
serpentine can lead to release considerable Mg, Ca and K, that are
highly mobile and can be easily removed out of the laterite profile
by aqueous solution, except for that part of the dissolved elements
may participate in the formation of secondary minerals such as
serpentine and smectite (Gleeson et al., 2003).
Some major (e.g., Si and Mn) and trace (e.g., Co and Ni) elements
show a notable gain in some specific horizons (Fig. 8). These fea-
tures indicate that there is a strong redistribution of elements
within the laterite profile. As for Si, it is lost nearly 99%
(s = 0.99) in the ferruginous cap, while it has an external addition
over 1 time (s = 1.03) in the lower part of the limonite horizon. This
result indicate that Si is released and transported by aqueous solu-
tions in the upper part of the profile, and part of the transported Si
may precipitate in joints and fractures of the lower part of the pro-
file, especially forming the silica stockwork.
Ni, as the most striking element in the Kolondale laterite soil,
shows a loss with a range of 17–76% (s = 0.17–0.76) in the ferrugi-
nous cap and limonite horizon, while it exhibits a marked gain
with a range of 1.93–28.09 times (s = 1.93–28.09) in the saprolite
horizon. The variation of Ni mass transfer coefficient along the pro-
file indicates a complex geochemical behavior of Ni during the
lateritization.
Ni is probably derived from olivine in the early stage of weath-
ering for the Kolonodale laterite profile, not only because olivine
has the highest Ni concentration among the primary minerals
based on our EMPA data, but also it is the earliest mineral
experiencing alteration. Besides olivine, the primary serpentine
may also have provided an extra Ni source for the Ni enrichment
 W. Fu et al. / Journal of Asian Earth Sciences 93 (2014) 74–88 85

Table 3
The element mass balance coefficient (s) along the profile.

Horizon Ferruginous cap Limonite Saprolite

Sample L1-1 L1-2 L1-3 D1-3 L1-4 L1-5 K1-2 L1-6
Al2O3 0.39 0.42 0.09 0.90 0.02 0.16 0.94 0.03
CaO 1.00 1.00 1.00 0.96 0.81 0.84 0.98 0.97
Fe2O3 0.30 0.36 0.44 0.08 0.78 0.26 0.74 0.56
K2O 0.89 0.88 0.87 0.02 0.61 0.43 0.00 0.72
MgO 1.00 1.00 1.00 0.99 0.57 0.84 0.71 0.91
MnO 0.81 0.77 0.55 0.07 0.38 0.48 0.64 0.54
Na2O 0.88 0.79 0.92 0.50 0.58 0.05 0.38 0.78
P2O5 0.70 0.78 0.75 0.20 0.83 0.69 0.66 0.74
SiO2 0.99 0.99 0.99 1.03 0.40 0.63 0.07 0.60
Cr2O3 0.15 0.27 1.08 0.30 0.27 0.39 0.74 0.74
NiO 0.73 0.76 0.17 0.50 7.83 2.28 28.09 1.93
Cs 0.95 0.97 0.97 0.20 0.39 0.76 1.12 0.83
Rb 0.97 0.98 0.98 0.05 0.95 0.97 0.58 0.96
Ba 0.97 0.94 0.51 1.66 1.13 0.24 0.72 0.77
Th 3.84 4.69 0.20 0.35 0.89 0.92 0.84 0.93
U 3.79 2.84 0.36 12.99 7.49 4.96 0.80 0.20
Ta 0.75 0.72 0.93 0.69 0.86 0.97 0.86 0.07
Nb 0.37 0.58 0.56 0.18 0.79 0.90 0.45 0.82
Sr 0.49 0.69 0.98 5.09 0.23 0.46 0.86 0.38
Hf 4.54 5.11 0.74 0.91 0.47 0.19 0.94 0.12
Zr 22.62 26.47 4.94 0.88 0.15 0.07 0.85 0.21
Y 0.79 0.80 0.64 0.46 33.30 1.50 0.42 0.11
Pb 0.76 0.71 0.70 0.69 0.78 0.95 0.45 0.97
Co 0.67 0.71 0.44 0.84 1.42 0.37 0.52 1.17
Cu 0.54 0.60 0.07 0.45 0.21 0.05 0.81 0.34
Ga 0.28 0.33 0.07 0.84 0.03 0.11 0.90 0.06
Ge 0.69 0.72 0.32 0.24 1.32 0.22 9.29 0.01
Sc 0.45 0.54 0.19 0.05 0.34 0.22 0.18 0.11
V 0.53 0.58 0.09 0.49 0.02 0.47 0.75 0.17
Zn 0.51 0.58 0.31 0.43 7.36 0.17 3.31 0.89

in the laterite soils due to its wide presence in the bedrock in spite
of its lower Ni concentration.
With the alteration of olivine or serpentine, Ni is released from
the crystal lattice of the primary minerals to aqueous solution. A
previous study has confirmed that there is a marked Ni concentra-
tion in the leaching solution of the laterite profile (Gleeson et al.,
2004), and Ni was likely to be transported as a bicarbonate or com-
plexed form with an organic acid (Gleeson et al., 2003; Retallack,
2010). In the Kolonodale profile, a high Ni mobility may have taken
place in the limonite horizon where is verified as a low pH and a
rich organic matter condition, because the organic matter may play
a significant role in the dissolution and transportation of Ni under
the acidic environment (Barker et al., 1997).
The sorption of Ni in the weathered products from the leaching
solution may be performed by various minerals at different depths
(Golightly, 1981). In the upper part of the Kolonodale laterite pro-
file, the leached Ni may be fixed in goethite due to its large surface
area and open channel structure (Carvalho-e-silva et al., 2003), or
can be adsorbed by organic matter (Shi et al., 2012). In the lower
part of the study profile, part of the leached Ni may be fixed in sil-
icate minerals, as indicated by the high Ni concentrations of the
residual serpentine from our EMPA study. The Ni2+ in the weather-
ing solution may exchange for Mg2+ in the structure of serpentine
due to their similar ionic radius, and this ion exchange reaction is
greatly favorable under the alkaline condition (Golightly, 1981;
Burns, 1993). In the saprolite horizon, Ni may also precipitate with
Si and Mg when the weathering solution is over-saturated, forming
the garnierite. This precipitation process occurs locally but leads to
the richest Ni-bearing weathered product in the study profile.
Besides Ni, Co is another economical element in the ultramafic
rock-derived laterite profile. The elemental behavior of Co is gener-
ally associated with Mn during the lateritization process (Llorca
and Monchoux, 1991; Lambiv et al., 2013). Often Co preferentially
enters the structure of low-crystallinity Mn-oxyhydroxydes by
substitution in the lower part of the limonite horizon (Manceau
et al., 1987; Fouateu et al., 2006; Roqué-Rosell et al., 2010). How-
ever, the highest concentration of Co in the Kolondale profile is
only 0.05%, much lower compared with many Ni laterite profiles
such as Nkamouna profile, Cameroon, with up to 0.84% Co
(Lambiv et al., 2013) and Bulong profile, Australia, with up to 2%
Co (Elias et al., 1981). Thus, Co is not a potential ore-forming ele-
ment for further exploration in the study area.
6.3. Models of the Kolonodale Ni-laterite formation
The study area is a relatively juvenile terrain, such as Sorowako
(Golightly, 1979) in Sulaweisi, Indonesia, which has probably been
formed directly under tropical rainforest conditions that have
changed little since exposure in the late Miocene. According to
Elias (2002), the general understanding of the ultramafic rock lat-
eritization is based on the sequence of the lithostratigraphic hori-
zons along the profile: that is, the lowest horizon of the laterite
profile reflects early stages of weathering of the bedrock, and each
horizon further up represents a transformation of that lying imme-
diately below it, displaying progressively advancing stages of the
processes.
However, the Elias’s model cannot properly explain the forma-
tion of the Kolondale Ni-laterite profile. For example, the lowest
horizon of the profile is the saprolite horizon dominated by the sil-
icate minerals, which cannot reflect early stages of weathering of
the bedrock because the most favorable weathering products are
the ferruginous minerals at the beginning of the lateritization
under an oxidized condition (Golightly, 1981; Freyssinet et al.,
2005). Thus, it is insufficient to reconstruct a rational lateritization
model only depending on the lithostratigraphic sequence of the
profile.
Noticing that the formation of a laterite profile can be regarded
as a advancement history of a weathering front moving from the
86 W. Fu et al. / Journal of Asian Earth Sciences 93 (2014) 74–88

Fig. 9. Evolution history of the Kolonodale Ni-laterite profile. Stage A: Formation of the ferruginous saprolite horizon firstly after the serpentinite exposure, which retain the
structure of bedrock partly but many primary minerals have altered to ferric minerals under the intensive oxidizing environment; Stage B: Formation of limonite horizon at
the top of the ferruginous saprolite horizon because of the advancing of the oxidation process. Meanwhile, the ferruginous saprolite horizon extended downward; Stage C:
Ferruginous cap was developed in the top of weathering crust and most of the ferruginous saprolite was oxidized to limonite by the advancing of the oxidation process. At the
same time, with the weathering front moving downward to a higher pH and a lower Eh environment in the lower part of the profile, the silicate saprolite began to develop
with the movement of water table and ion-exchange process; Stage D: Secondary precipitation process have taken place as a result of the long-term leaching process in the
laterite profile.

surface to the water table, therefore by using the advancement of
the weathering front as a time path, it is feasible to interpretate
the formation of various lithostratigraphic horizons and weathered
products in the study profile, and reconstruct the lateritization
process ultimately. Based on our field data, and mineralogical
and geochemical analyses mentioned above, a four-stage evolu-
tional model for the Kolonodale profile is hence proposed as fol-
lows (see Fig. 9):
6.3.1. Stage I: Ferruginous saprolite development
This stage may exist temporally at the beginning of the lateriti-
zation in the Kolonodale profile based on three aspects of evi-
dences: (1) the ferruginous minerals are the earliest weathered
products in our study profile; (2) the formation of a saprolite is
very common at the early stage of the lateritization in many later-
ite profiles (e.g., Freyssinet et al., 2005) due to the differential
weathering; (3) the ferruginous saprolite is observed locally in
some shallow profiles despite of its rare presence in most of the
deep profiles. Thus, we suspect that the ferruginous saprolite can
reflect the initial stage products in the latertization.
This weathering stage may begin along joints and faults firstly,
where the primary minerals of the bedrock were quickly altered to
ferruginous minerals due to the intensive hydrolysis and oxidation
processes at the surface, but the bedrock distal to joints and fac-
tures were not decomposed and maintained the primary feathures.
Most of the ferruginous saprolites may have been destroyed com-
pletely by further weathering processes. This is why no specific
lithostratigraphic horizon is preserved in the present deep profiles.
The Ni enrichment effect is uncertain in this stage because the
weathering degree is relatively weak and inhomogeneous.
6.3.2. Stage II: Limonite development
This stage is based on the fact that there is a stable limonite
horizon in the upper part of the laterite profile. It can be explained
by the movement of the weathering front and continuous weather-
ing processes after the stage I.
With the weathering front continued to extend downward,
most of the bedrock blocks or fragments retained in the upper part
of the previously formed ferruginous saprolite were greatly
destroyed by the oxidation and hydrolysis. The new formed weath-
ering products dominated by the ferruginous minerals were accu-
mulated with a typical earthy structure. These changes indicated
that a stable limonite horizon has separated from the previously
ferruginous saprolite horizon. Significant geochemical changes
have also taken place in this stage, showing an increase in Fe and
Al, and a strong decrease in Mg, Si and Ca. The Ni enrichment effect
was notable in this stage. The goethite may adsorb Ni2+ released
from the primary minerals and promote the Ni concentration of
the limonite samples increase to around 0.5–1.0%.
 W. Fu et al. / Journal of Asian Earth Sciences 93 (2014) 74–88
6.3.3. Stage III: Silicate saprolite and ferruginous cap development
This stage is inferred from the fact there is a thick silicate
horizon underlying the limonite horizon. The silicate saprolite
with a lower occurrence position in the study profile may indi-
cate that its development need a specific environment where
the weathering front have passed the formation field of the
limonite and further moved to a deeper level. Thus, the silicate
saprolite reflects a weathering product after the limonite
formation.
There is a regional uplift setting caused by plate movement
from Eocene to recent in Sulawesi island (Golightly, 1979;
Kadarusmana et al., 2004). This geological background is favorable
for enhancing the relief, lowering the water table and promoting
the weathering front to move deeper (Golightly, 2010; Butt and
Cluzel, 2013). As the depth increases, oxidation may have a weak
influence on the lateritization process, but the groundwater activ-
ities and rock-water interaction may play a more important role.
The oscillation of the groundwater level can increase the structural
decomposition of the bedrock and promote mineral alteration.
Olivine was still the first mineral suffering alteration and the prod-
uct was iron oxide. However, iron oxide cannot be preserved for
long and was easy to be dissolved by groundwater under a high
pH and a low Eh environment. Most of serpentine was preserved,
but the ion exchange reactions were favorable between the serpen-
tine and the leaching solution under an alkaline environment
(Golightly, 1981). Moreover, when the weathering front reached
to near the water table, serpentine may be altered to smectite
where the drainage condition was poor. In this stage, the leached
Ni in solution may substitute Mg of serpentines or smectite by
the ion exchange reactions, holding a remarkable amount of Ni-
rich silicate minerals. This process can promote the Ni concentra-
tion of the saprolite horizon to rise to 1.4–2.5%, and form so-called
silicate ore (Brand et al., 1998).
At the same time with the development of the silicate saprolite,
the ferruginous cap was formed at the top of the weathering pro-
file. Its formation may be ascribed to the dehydration process from
goethite to hematite under a strong acidic surface condition.
6.3.4. Stage IV: Precipitation of secondary minerals
This stage is represented by the occurrence of some secondary
precipitation products in the profile, such as the garnierite and sil-
ica stockwork. The garnierite may be formed after the silicate sap-
rolite due to their crosscut relationship.
After stage-II and stage-III, the surface of the laterite profile has
accumulated abundant organic matter and produced a low pH
environment, leading to the dissolution and recrystallization of
Ni-bearing goethite. Ni was released again and transported to the
deeper horizon. In the saprolite horizon, the pH environment was
alkaline and favorable for Ni precipitation from the saturated
leaching solutions, forming the Ni-rich garnierite in joints and
cracks. By this process, the concentration of Ni in the garnierite
may exceed 10%.
The leached Si may also precipitate in the lateritization because
of a locally poor-drained environment or a dry season condition
(Golightly, 1981; Thorne et al., 2009). Generally, the silica precipi-
tation diluted the Ni concentration of weathering products and
acted as an negative factor for the Ni enrichment. However, the sil-
ica stockwork may play an important role in protecting the laterite
profile from erosion (Golightly, 2010).
7. Conclusions
(1) The intensive weathering of serpentinite in the Kolonodale
area yielded a Ni-laterite profile with three specific horizons
(the saprolite horizon, the limonite horizon, and the
87
ferruginous cap) overlying the bedrock. The pH (5.56–8.56)
and organic matter concentration (1.11–4.82%) of the
weathered products show variations associated with the
depth.
(2) There is a progressive mineral assemblage transition from
the silicate mineral dominant (mainly serpentine) to the
Fe-oxyhydroxide dominant (mainly goethite) along the lat-
erite profile. Garnierite is identified as a mixture of talc-
and serpentine-like phases. The mineral evolution during
the lateritization process shows various paths from the pri-
mary minerals to altered minerals, which is integrally
affected by the nature of the primary minerals and environ-
mental conditions.
(3) The whole rock geochemistry shows a fairly typical laterite
pattern, in which Mg and Si are depleted toward the top of
the sequence, whereas Fe, Al and Cr increase. The highest
Ni concentration (11.53%NiO) is shown in the garnierite
sample. The mass-balance calculations reveal that Mg, Ca
and K show an absolutely loss within the whole laterite pro-
file, while Si, Ni, Mn and Co show a gain in specific horizon.
There are diversified elemental behaviors during the ser-
pentinite lateritization under the rainforest conditions.
Supergene behavior of Ni is closely associated with the pH
environment, organic matter, and mineral evolution.
(4) A four-stage lateritization process may have taken place in
the Kolonodale laterite profile as follows: the ferruginous
saprolite development (stage I), the limonite development
(stage II), the silicate saprolite and ferruginous cap develop-
ment (stage III), and precipitation of secondary minerals
(stage IV). Due to this multistage process, there is a progres-
sive Ni-enrichment in the laterite profile.
Acknowledgments
This research was financially supported by the National Natural
Science Youth Foundation (41102051), Guangxi Natural Science
Foundation (2012GXNSFAA053187, 2012GXNSFCA053007) and
the Program to Sponsor Teams for Innovation in the Construction
of Talent Highlands in Guangxi Institutions of Higher Learning.
We acknowledge the support of the PT. Pan China Indon Co., Ltd.
for our field investigation. We also thank Zeng Nanshi of Guilin
University of Technology (China), S. A. Gleeson of University of
Alberta (Canada), and J. Paul Golightly of Golightly Geoscience
Ltd (Canada) for their valuable help.
References
Amundson, R., 2003. Soil formation. Treatise Geochem. 5, 1–35.
Anderson, S.P., Dietrich, W.E., Brimhall, G.H., 2002. Weathering profiles, mass-
balance analysis, and rates of solute loss: linkages between weathering and
erosion in a small, steep catchment. Geol. Soc. Am. Bull. 114, 1143–1158.
Bárdossy, G., Aleva, G.J.J., 1990. Lateritic Bauxites: Developments in Economic
Geology. Elsevier Scientific Publication, pp. 1–624.
Barker, W.W., Welch, S.A., Banfield, J.F., 1997. Biogeochemical weathering of silicate
minerals. In: Baneld, J.F., Nealson, K.H. (Eds.), Geomicrobiology: Interactions
Between Microbes and Minerals, (Reviews in Mineralogy, vol. 35).
Mineralogical Society of America, Washington, DC, pp. 391–428.
Baumgartner, L.P., Olsen, S.N., 1995. A least square approach to mass transport
calculations using the isocon method. Econ. Geol. 90, 1261–1270.
Berger, V.I., Singer, D.A., Bliss, J.D., Moring, B.C., 2011. Ni–Co laterite deposits of the
world—database and grade and tonnage models. U.S. Department of the Interior
& U.S. Geological Survey, Open-File Report 2011-1058, pp. 1–26.
Bergera, A., Frei, R., 2014. The fate of chromium during tropical weathering: a
laterite profile from Central Madagascar. Geoderma 213, 521–532.
Brand, N.W., Butt, C.R.M., Elias, M., 1998. Nickel laterites: classifications and
features. AGSO J. Australian Geol. Geophys. 17, 81–88.
Braun, J.J., Pagel, M., Herbillon, A., Rosin, C., 1993. Mobilisation and distribution of
REEs and thorium in syenitic lateritic profile: a mass balance study. Geochim
Cosmochim Acta 57, 4419–4434.
88 W. Fu et al. / Journal of Asian Earth Sciences 93 (2014) 74–88
Brindley, G.W., Hang, P.T., 1973. The nature of garnierites-i structures, chemical
compositions and color characteristics. Clays Clay Miner. 21, 27–40.
Brindley, G.W., Bish, D.L., Wan, H.M., 1977. The nature of kerolite, its relation to talc
and stevensite. Mineral. Mag. 41, 443–452.
Brown, D.J., Helmke, P.A., Clayton, M.K., 2003. Robust geochemical indices for redox
and weathering on a granitic laterite landscape in Central Uganda. Geochim
Cosmochim Acta 67, 2711–2723.
Burns, R.G., 1993. Mineralogical Applications of Crystal Field Theory. Cambridge
University Press, pp. 551.
Butt, C.W., Cluzel, D., 2013. Laterite nickel ore deposits: weathered serpentinites.
Elements 9, 123–128.
Caillaud, J., Proust, D., Philippe, S., Fontaine, C., Fialin, M., 2009. Trace metals
distribution from a serpentinite weathering at the scales of the weathering
profile and its related weathering microsystems and clay minerals. Geoderma
149, 199–208.
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, 876–882.
Cluzel, D., Vigier, B., 2008. Syntectonic mobility of supergene nickel ores of New
Caledonia (Southwest Pacific). evidence from Garnierite Veins and Faulted
Regolith. Resource Geol. 58, 161–170.
Colinf, 1990. Lateritic weathering of pyroxenites at Niquelandia, Goias, Brazil: the
supergene behavior of nickel. Econ. Geol. 85, 1010–1023.
Dalvi, A.D., Bacon, W.G., Osborne, R.C., 2004. Past and future of nickel laterite
projects. In: Imrie, W.P., Lane, D.M. (Eds.), International Nickel Laterite
Symposium TMS. 133 Annual Meeting and Exhibition, Charlotte, N. Carolina,
pp. 23–50.
Das, S.K., Sahoo, R.K., Muralidhar, J., Nayak, B.K., 1999. Mineralogy and geochemistry
of profiles through lateritic nickel deposits at Kansa, Sukinda, Orissa. J. Geol. Soc.
India 53, 649–668.
Elias, M., 2002. Nickle laterite deposits-geological overview, resources and
exploitation, in giant ore deposit: characteristics, genesis and exploration.
Centre Ore Deposit Res., Univ. Tasmania, Spec. Pub. 4, 205–220.
Elias, M., Donaldson, M.J., Giorgetta, N., 1981. Geology, mineralogy, and chemistry
of lateritic nickel–cobalt deposits near Kalgoorlie, Western Australia. Econ.
Geol. 76, 1775–1783.
Fouateu, R.Y., Ghogomua, R.T., Penaye, J., Ekodecka, G.E., Stendalc, H., Colin, F., 2006.
Nickel and cobalt distribution in the laterites of the Lomié region, south-east
Cameroon. J. Afr. Earth Sc. 45, 33–47.
Fouateu, R.Y., Yemsfack, M., Wouatong, A.S.L., Ndjigui, P.D., Bilong, P., 2009.
Contrasted mineralogical composition of the laterite cover on serpentinites of
Nkamouna-Kongo, southeast Cameroon. Clay Miner. 44, 221–237.
Freyssinet, P., Butt, C.R.M., Morris, R.C. et al., 2005. Ore-forming processes related to
laterite weathering. In: Economic Geology 100th Anniversary Volume, pp. 681–
722.
Gali, S., Soler, J.M., Proenza, J.A., Lewis, J.F., Cama, J., Tauler, E., 2012. Ni enrichment
and stability of Al-free garnierite solid-solutions: a thermodynamic approach.
Clays Clay Miner. 60, 121–135.
Garnier, J., Quantin, C., Guimarães, E., Garg, V.K., Martins, E.S., Becquer, T., 2009.
Understanding the genesis of ultramafic soils and catena dynamics in
Niquelândia, Brazil. Geoderma 151, 204–214.
Gleeson, S.A., Butt, C.R.M., Elias, M., 2003. Nickle laterites: a review. Newslett. Soc.
Econ. Geol. 4, 12–18.
Gleeson, S.A., Herrington, R.J., Durango, J., Velázquez, C.A., 2004. The mineralogy and
geochemistry of the Cerro Matoso S.A. Ni laterite deposit, Montelíbano,
Colombia. Econ. Geol. 99, 1197–1213.
Golightly, J.P., 1979. Geology of Soroako nickeliferous laterite deposits. In: Evans,
D.J.I., Shoemaker, R.S., Veltman, H. (Eds.), International Laterite Symposium.
AIME, New York, pp. 38–56.
Golightly, J.P., 1981. Nickeliferous laterite deposits. In: Economic Geology, 75th
Anniversary Volume, pp. 710–735.
Golightly, J.P., 2010. Progress in understanding the evolution of nickel laterites. In:
Goldfarb, R.J., Marsh, E.E., Monecke, T. (Eds.), The Challenge of Finding new
Mineral Resources—Global Metallogeny, Innovative Exploration, and New
Discoveries, vol. 15. Society of Economic Geologists Special Publication, pp.
451–485.
Gong, Q.J., Deng, J., Yang, L.Q., Zhang, J., Wang, Q.F., Zhang, G.X., 2011. Behavior of
major and trace elements during weathering of sericite–quartz schist. J. Asian
Earth Sci. 42, 1–13.
Grant, J.A., 1986. The isocon-diagram—a simple solution to Gresens’ equation for
metasomatic alteration. Econ. Geol. 81, 1976–1982.
Hall, R., Wilson, E.J., 2000. Neogene suture in Eastern Indonesia. J. Asian Earth Sci.
18, 781–808.
Kadarusmana, A., Miyashitab, S., Maruyamaa, S., Parkinsonc, C.D., Ishikawad, A.,
2004. Petrology, geochemistry and paleogeographic reconstruction of the East
Sulawesi Ophiolite, Indonesia. Tectonophysics 392, 55–83.
Lambiv, D.G., Gleeson, S.A., Schofield, P.F., 2013. Mineralogical characterization of
the Nkamouna Co–Mn laterite ore, southeast Cameroon. Miner. Deposita 48,
155–171.
Lewis, J.F., Draper, G., Proenza, J.A., Espaillat, J., Jimenez, J., 2006. Ophiolite-related
ultramafic rocks (Serpentinites) in the Caribbean Region: a review of their
occurrence, composition, origin, emplacement and Ni-laterite soils formation.
Geol. Acta 4, 237–263.
Llorca, S., Monchoux, P., 1991. Supergene cobalt minerals from New Caledonia. Can.
Mineral. 29, 149–161.
Ma, J.L., Wei, G.J., Xu, Y.G., Long, W.G., Sun, W.D., 2007. Mobilization and re-
distribution of major and trace elements during extreme weathering of basalt in
Hainan Island, South China. Geochim. Cosmochim. Acta 71, 3223–3237.
Manceau, A., Calas, G., 1985. Heterogeneous distribution of nickel in hydrous
silicates from New Caledonia ore deposits. Am. Mineral. 70, 549–558.
Manceau, A., Llorca, S., Calas, G., 1987. Crystal chemistry of cobalt and nickel in
lithiophorite and asbolane from New Caledonia. Geochim Cosmochim Acta 51,
105–113.
Mubroto, B., Briden, J.C., McClelland, E., Hall, R., 1994. Paleomagnetism of the
Balantak ophiolite, Sulawesi. Earth Planet. Sci. Lett. 125, 193–209.
Mudd, G.M., 2010. Global trends and environmental issues in nickel mining:
sulfides versus laterites. Ore Geol. Rev. 38, 9–26.
Nahon, D., Merino, E., 1996. Pseudomorphic replacement versus dilation in
laterites: petrographic evidence, mechanisms, and consequences for
modeling. J. Geochem. Explor. 57, 217–225.
Nahon, D., Paquet, H., Delvigne, J., 1982. Lateritic weathering of ultramafic rocks and
the concentration of nickel in western Ivory Coast. Econ. Geol. 77, 1159–1175.
Nelson, D.W., Sommers, L.E., 1982b. Total carbon, organic carbon and organic
matter methods of soil analysis. ASA and SSSA, Madison, Wisconsin USA, pp.
534–580.
Nickel, E.H., 1992. Solid solutions in mineral nomenclature. Mineral. Mag. 56, 127–
130.
Oliveira, S.M.B., Trescases, J.J., Melfi, A.J., 1992. Lateritic nickel deposits of Brazil.
Miner. Deposita 27, 137–146.
Retallack, G.J., 2010. Lateritization and bauxitization events. Econ. Geol. 105, 655–
667.
Roqué-Rosell, J., Mosselmans, J.F.W., Proenza, J.A., Labrador, M., Galí, S., Atkinson,
K.D., Quinn, P.D., 2010. Sorption of Ni by ‘‘lithiophorite–asbolane’’
intermediates in Moa Bay lateritic deposits, eastern Cuba. Chem. Geol. 275,
9–18.
Schellmann, W., 1989. Composition and origin of lateritic nickel ore at Tagaung.
Miner. Deposita 24, 161–168.
Shi, Z.Q., Peltier, E., Sparks, D.L., 2012. Kinetics of Ni sorption in soils: roles of soil
organic matter and Ni precipitation. Environ. Sci. Technol. 46, 2212–2219.
Sufriadin, Idrus, A., Pramumijoyo, S., Warmada, I.W., Imai, A., 2011. Study on
mineralogy and chemistry of the saprolitic nickel ores from Soroako, Sulawesi,
Indonesia: implication for the lateritic ore processing. J. Southe. Asian Appl.
Geol. 3, 23–33.
Thorne, R.L., Herrington, R., Roberts, S., 2009. Composition and origin of the Çaldağ
oxide nickel laterite deposit, W. Turkey. Miner. Deposita 44, 581–595.
Thorne, R.L., Herrington, R., Roberts, S., 2012. Climate change and the formation of
nickel laterite deposits. Geology 40, 331–334.
Trescases, J.J., 1986. Nickeliferous laterites: a review of the contributions of the last
10 years. Geol. Surv. India, Memoirs 120, 51–62.
Troly, G., Esterle, M., Pelletier, B., Reibell, W., 1979. Nickel deposits in New
Caledonia, some factors influencing their formation. In: Evans, D.J.I., Shoemaker,
R.S., Veltman, H. (Eds.), International Laterite Symposium. Society of Mining
Engineers, New York, pp. 85–119.
Wells, M.A., Ramanaidou, E.R., Verrall, M., Tessarolo, C., 2009. Mineralogy and
crystal chemistry of ‘‘garnierites’’ in the Goro lateritic nickel deposit, New
Caledonia. Eur. J. Mineral. 21, 467–483.
Wilson, M.J., 2004. Weathering of the primary rock-forming minerals: processes,
products and rates. Clay Miner. 39, 233–266.