The occurrence and behaviour of rare earth and associated elements in lateritic regolith profiles in Western Australia XIN DU BSc (EnvSc) This thesis is presented for the degree of Doctor of Philosophy in Soil Science of the University of Western Australia School of Earth and Environment Faculty of Natural and Agricultural Sciences 2012
267
Embed
The occurrence and behaviour of rare earth and associated … · The occurrence and behaviour of rare earth and associated elements in lateritic regolith profiles in Western Australia
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
The occurrence and behaviour of rare earth and associated
elements in lateritic regolith profiles in Western Australia
XIN DU
BSc (EnvSc)
This thesis is presented for the degree of Doctor of Philosophy
in Soil Science of the University of Western Australia
School of Earth and Environment
Faculty of Natural and Agricultural Sciences
2012
i
Statement of candidate contribution
This thesis contains published work and/or work prepared for publication, some of
which has been coauthored. The bibliographical details of the work and where it
appears in the thesis are outlined below (with percentage contributions from coauthors
in parentheses).
i. Du, X. (70%), Rate, A.W. (15%), & Gee, M. (15%), 2012. Redistribution and
mobilization of titanium, zirconium and thorium in an intensely weathered lateritic
profile in Western Australia. Chemical Geology 330-331, 101-115.
ii. Du, X. (70%), Rate, A.W. (15%), & Gee, M. (15%), 2012 (In press). Particle size
fractionation and chemical speciation of REE in a lateritic weathering profile in
Western Australia. Explore.
iii. Du, X. (70%), Rate, A.W. (15%), & Gee, M. (15%), 2011. Translocation and
fractionation of rare earth elements within intensely weathered lateritic profiles in
Western Australia. Mineralogical Magazine 75, 784. Oral presentation in the 2011
Goldschmidt Conference, Prague, Czech Republic. available at:
REE mobilization (Johannesson et al., 1996); conversely, phosphate tends to removal
REE from solution (Johannesson et al., 1995). The complex Ln(CO3)2- (Ln denotes
REE) is strongly enriched in HREE over LREE whereas LnCO3+ (stable in seawater) is
the opposite (Cantrell and Byrne, 1987). At pH 6.5-9.5, LnCO3
+ predominates whereas
at pH ≥9.5 the Ln(CO3)2- complex is favoured (Wood, 1990). When the pH is 2-6.5,
REE occur mainly as simple ions and sulphate complexes (Wood, 1990). However,
lack of systematic experimental data for: (i) all REE across a wide range of pH, (ii) the
presence of multi-complex phases, (iii) the limited investigations of the complexation
of Ce4+
and Eu2+
and stability constants for REE phosphate complexes, and (iv)
conflicting data about the thermodynamics of REE hydrolysis (Wood, 1990), all
restrain further understanding of the mobilization of REE during water/rock interaction
under natural supergene weathering conditions.
Occurrence and behaviour and rare earth and associated elements in lateritic regolith profiles in Western Australia
18
2.5.3 Fractionation of REE during weathering
Although fractionation of REE may occur during weathering processes (Koppi et al.,
1996), exactly how weathering intensity exerts effects on REE fractionation, and the
precise sequence of events and fate of REE during weathering is not fully understood.
This information is necessary in order to use REE as tracers for supergene weathering,
pedogenesis and sedimentation processes.
Differences in weathering rates and the formation of more element-specific secondary
minerals (Tyler, 2004) during weathering may also result in REE fractionation in
weathering profiles. However, diverging views exist regarding at which stage the
movement and differentiation of REE start during rock weathering and soil formation
(Zhang et al., 2007). Banfield and Eggleton (1989), Price et al. (1991) and Sharma and
Rajamani (2000) illustrated that REE contents change dramatically during the initial
stages of weathering, while Middelburg et al. (1988) and Duzgoren-Aydin and Aydin
(2009) proposed that the migration and differentiation of REE occurs at advanced
stages of weathering. In early and intermediate weathering, mineral abundances may
control REE abundances (Banfield and Eggleton, 1989; Nesbitt, 1979) but for
advanced weathering in laterite, the relatively greater mobility of the HREE appears to
be more significant (Braun et al., 1993; Braun et al., 1990; Brown et al., 2003).
The lower mobility of LREE compared to HREE commonly results in a significant
relative enrichment of LREE and depletion of HREE in weathering products after
extensive weathering (Braun et al., 1993; Braun et al., 1990; Compton et al., 2003;
Koppi et al., 1996; Nesbitt, 1979), however higher mobility of LREE over HREE have
also been reported in intensely weathered environments (e.g. Beyala et al., 2009; Braun
et al., 1990; Ndjigui et al., 2009; Nesbitt and Markovics, 1997). Light REE are known
to be hydrolysed more easily than HREE, whereas HREE preferentially form more
stable inorganic complexes than LREE, particularly with carbonate, fluoride,
hydroxide or sulphate anions in alkaline solutions (Åström and Corin, 2003) and are
more likely to desorb from clay minerals than LREE. This may explain why HREE are
more prone to mobilization than LREE and translocation to the lower parts of regolith
Chapter Two: Literature review
19
profiles (Aubert et al., 2004; Aubert et al., 2001; Cantrell and Byrne, 1987; Ma et al.,
2007). However, this process is believed to be controlled by pH and the type of REE
complexation (Cantrell and Byrne, 1987; Johannesson et al., 1995; Johannesson et al.,
1996; Wood, 1990).
The relationship between weathering intensity and the mobilization and fractionation
of REE is also unclear. Some previous studies have shown that there is no significant
correlation between the degree of REE fractionation and any of the following;
chemical weathering intensity (by CIA) (Caspari et al., 2006), physical weathering
intensity (by particle size fraction index) (Caspari et al., 2006) and rock weathering
and soil formation (Minarik et al., 1998; Zhang et al., 2007). However, the degree of
inter-horizon transport of REE has been proposed to have great potential to become an
index of weathering intensity (Aide and Christine-Aide, 2012). Then this poses the
question that what is the association between the mobility and fractionation of REE
and the weathering intensity? What factors control REE fractionation during
weathering? And what mechanisms are involved during initial and advanced
weathering? An important research issue, the mobilization and fractionation
mechanisms of REE during weathering, is therefore discussed in detail in this thesis.
2.5.4 Mineral transformation of REE during weathering
The transformation of REE minerals during weathering processes depends on their
susceptibility and the weathering conditions. Rare earth elements-bearing minerals can
be subdivided into three groups according to weathering susceptibility: (i) strongly
resistant to weathering e.g. xenotime and zircon; (ii) moderately resistant to weathering,
e.g. monazite; and (iii) weakly resistant to weathering, e.g. allanite and the
fluorocarbonates (bastnäsite and parisite) (Bao and Zhao, 2008). During initial and
moderate stages of weathering, the weathering-susceptible minerals: feldspar, biotite,
allanite, epidote, apatite etc., preferentially dissolve in low pH solutions. The REE
released during this process may: (i) be lost from the weathering profile via transport in
solution (Condie et al., 1995), (ii) form secondary minerals (Braun and Pagel, 1994),
(iii) be incorporated into or adsorbed onto clay minerals (Vos et al., 2006), and (iv) be
Occurrence and behaviour and rare earth and associated elements in lateritic regolith profiles in Western Australia
20
adsorbed onto Fe/Mn oxyhydroxides (Aide and Smith-Aide, 2003; Aide et al., 1999).
Simultaneously, weathering-resistant heavy minerals such as zircon and rutile are
retained as residuum in weathered regolith, and the REE (especially HREE) included
in these heavy minerals are not expected to be highly mobilized during pedogenesis,
except under intense or extreme weathering. Under extreme weathering conditions,
both secondary and weathering-resistant heavy minerals may partially or completely
alter (Braun et al., 1993; Taunton et al., 2000a; Taunton et al., 2000b). In this process
quartz, being relatively resistant to weathering, acts as a diluent in the regolith (Hardy
and Cornu, 2006). Incorporation or reformation of secondary florencite, rhabdophane
and/or churchite (Braun and Pagel, 1994) is believed to be the main pathway for LREE
retention in weathered regolith (Nedachi et al., 2005).
2.6 Geochemical pathways of REE during lateritization
2.6.1 Definition of lateritic profiles
The study of regolith spans many disciplines of the Earth Sciences (Anand and Paine,
2002), and thus many definitions are confusing and may lead to misunderstanding.
Therefore, it is necessary to define terms from the perspective of regolith geochemistry.
In this thesis the term ‘laterite’ refers to Fe-rich weathering profiles which have
undergone intense supergene weathering (Anand and Butt, 2000; Anand and Butt,
2003; Anand and Butt, 2010; Anand and Paine, 2002); and the other key terminology
for deeply weathered profiles used in this thesis is summarised below, following the
definitions published by Anand and Paine (2002).
A typical laterite profile commonly includes saprolite, mottled clay zone, ferruginous
zone and surface soil. Saprolite refers to nearly isovolumetrically weathered bedrock
retaining the fabric and structure of the parent rock, pseudomorphically replacing the
primary minerals. The mottled zone has macroscopic segregations of subdominant
colour that differ from the surrounding matrix and mottled clay zone is dominantly
composed by secondary clay minerals. The ferruginous zone is composed
predominantly of secondary oxides and oxyhydroxides of Fe (goethite, hematite,
Chapter Two: Literature review
21
maghemite), hydroxides of aluminum (e.g. gibbsite, boehmite) and kaolinite, with or
without quartz. The upper lateritic profile consists of ferruginous mottled zone,
ferruginous duricrust and loose gravel. Ferruginous mottled zone has a goethite rich
halo with sharp or diffuse boundaries, whereas ferruginous duricrust is a cemented
hard layer composed of various Al-Fe secondary segregations that originated from
underlying parent rock. Distinct from the ferruginous duricrust, ferricrete is a product
of cementation and conglomeration of surficial sands and gravel by Fe oxides, where
no genetic relationship between the Fe and the underlying mottled and saprolite zones
is inferred (Anand and Butt, 2010). In this thesis, ‘duricrust’ is used in short for
‘ferruginous duricrust’ since the ferruginous duricrust is the only type of duricrust
present in the lateritic profiles studied. Regolith is used as a collective term for the
weathered and transported materials covering fresh rock, which have been formed by
various geochemical processes e.g. weathering, erosion, transport and/or deposition of
older material.
Laterite represents one of the most common superficial formations in the tropics,
covering approximately 30% of the continents (Dequincey et al., 2002), the formation
of which can extend of 10’s of Ma (Dequincey et al., 2006). Lateritic profiles are
commonly thought to have formed in tropical climates with relatively high
temperatures and seasonal rainfall. However, laterite can also form in wet, cool to cold
climates given sufficient time (Gozzard, 2007). In contrast to common pedogenesis,
during lateritic weathering, lateritic regolith is intensely depleted in base cations and
enriched in iron either in some layers, or throughout the profile, commonly forming
ferruginous zones. In Western Australia, seasonally high rainfall and alternating
arid/humid weathering conditions may have further enhanced the lateritization process
and resulted in a widespread distribution of lateritic profiles. Therefore, studying of
abundance and redistribution of REE in lateritic regolith is significant for
understanding the behaviour and mode of occurrence of REE under intense weathering
conditions and advanced lateritization.
Occurrence and behaviour and rare earth and associated elements in lateritic regolith profiles in Western Australia
22
2.6.2 A typical lateritic profile
Zoning in lateritic profiles occurs at variable depths (Bourman, 1993), but a typical
lateritic profile would have: fresh bedrock, saprock, saprolite, a mottled clay zone and
a ferruginous zone (Anand and Paine, 2002). In real situations, however, one or more
zones may be missing from the profile. The saprock and saprolite form the lower part
of the regolith and retain the primary mineral constituents of the bedrock. At this depth,
weathering has been less intense and is nearly isovolumetric, whereas the mottled clay
and ferruginous zones comprising the upper part of the profile have been subjected to
stronger, non-isovolumetric weathering, leaching, cementation and soil-forming
processes, and probably precipitation and erosion (Anand and Butt, 2010; Anand and
Paine, 2002). Thin, depleted (commonly quartz-rich) topsoil may often be present
above the ferruginous zone. In a ferruginous zone, nodules and pisoliths may be
present and they are distinguished by their morphology: nodules are irregular, with
re-entrant surfaces, whereas pisoliths are ellipsoidal or spherical. As the sphericity of
nodules increases they merge with pisoliths (Anand and Butt, 2010) and nodules can
also be formed from cementation of one or more pisoliths. In this study, both nodules
and pisoliths are presented in the Jarrahdale regolith profile (JG) studied, and ‘iron
nodules’ is used to denote both types to simplify the terminology.
2.6.3 Lateritization
The geochemical processes for formation of lateritic profiles/landscape are collectively
called lateritization. Intense weathering is an important process during lateritization,
leading to disaggregation, breakdown of original silicate minerals, dissolution of
primary minerals, and leaching of base cations. At the onset of weathering, any
carbonates are dissolved, sulfides are oxidised, easily weathered Fe-Mg silicates are
hydrolysed, and then most of the readily weathered minerals such as feldspar alter to
kaolinite. Resistant minerals, such as quartz and zircon, remain relatively unaltered.
The intense leaching of base cations and formation of secondary clay minerals leads to
the formation of mottled zone (sometimes known as the ‘pallid zone’ if depleted in Fe).
Secondary clay mineral formation by chemical weathering of the primary minerals is
Chapter Two: Literature review
23
termed ‘kaolinization’ and is one of the fundamental processes during lateritization.
With more intense weathering, silicates are increasingly leached out of the profile, and
more secondary oxides, Fe oxyhydroxides, Al hydroxides, and kaolinite are formed.
Continuous dissolution, precipitation, cementation, and erosion results in formation of
the ferruginous zone (Anand and Paine, 2002); note that enrichment of Fe oxides and
oxyhydroxides during lateritic weathering is referred as ferruginization. This process
leads to accumulation of crystalline Fe oxyhydroxides (e.g. goethite), Fe oxides (e.g.
hematite and maghemite), Al hydroxides (e.g. gibbsite), and Al oxyhydroxides
(e.g. boehmite), in the intensely weathered regolith. Though the geochemistry and
genesis of lateritization have been widely investigated (Bourman, 1993; Brimhall et al.,
1991; Schellmann, 1994), the translocation and fractionation of REE during
lateritization in various secondary mineral phases is still not fully understood (Feng,
2011; Ma et al., 2007).
2.6.4 Geochemical behaviour of REE during lateritization
Quantitative understanding of the nature of the migration-fixation and fractionation
mechanisms of REE caused by sorption of clay or Fe oxyhydroxide is currently
inadequate. Such information is important for interpreting the behaviour of REE as
potential clues tracing the processes of lateritization and weathering.
Mottled clay formed during kaolinization is believed to act as a potential reservoir of
REE in weathered lateritic profiles because of adsorption of REE onto the clay surface
(Laveuf and Cornu, 2009). This is an important secondary REE enrichment process
(Bao and Zhao, 2008; Ohlander et al., 1996) during weathering, however, the main
enrichment of REE takes place during precipitation of secondary REE bearing
minerals (Braun and Pagel, 1994; Braun et al., 1993). Adsorption of REE by clay is
controlled by the nature of the clay minerals, pH, ionic strength, the presence of
additional ligands such as carbonate or organic complexes, surface coverage, and
effects specific to the individual REE (Coppin et al., 2002; Fendorf and Fendorf, 1996;
Koeppenkastrop and Decarlo, 1992; Koeppenkastrop and Decarlo, 1993; Laveuf and
Occurrence and behaviour and rare earth and associated elements in lateritic regolith profiles in Western Australia
24
Cornu, 2009; Piasecki and Sverjensky, 2008; Takahashi et al., 1999). As well as these
controls, differences in clay mineralogy can affect fractionation of REE (Coppin et al.,
2002; Laveuf and Cornu, 2009), potentially explaining the apparently contradictory
signatures of REE adsorbed by clay minerals. Usually, REE adsorption increases with
increasing pH (Coppin et al., 2002), and in high ionic strength solutions HREE are
more sorbed than LREE (Coppin et al., 2002). Clay minerals have a strong affinity for
all REE except Ce (Duzgoren-Aydin and Aydin, 2009). As the most important clay
mineral in lateritic regolith, kaolinite has considerably variable REE concentrations
(Laveuf and Cornu, 2009) and the fractionation of REE by kaolinite sorption is still not
fully understood. The concentrations of REE in different horizons of lateritic regolith
profiles are listed in the Table 2.3.
The upper ferruginous zones in weathered lateritic profiles are commonly depleted in
all REE except Ce, although it has been reported that Fe oxides have high surface areas
rendering them very efficient sinks for heavy metals (Nedel et al., 2010; Singh and
Gilkes, 1992). Iron oxides are known to contain REE, but the concentration does not
correlate with Fe content (Laveuf and Cornu, 2009). It is rare for REE to substitute for
Fe in the lattice of Fe oxides at ambient temperatures and pressures; however,
transformation from ferrihydrite to goethite with Lu3+
and Eu3+
substitution and
incorporation was reported at 70 ºC and pH 13 (Dardenne et al., 2003).
Scavenging of REE by Fe oxides and oxyhydroxides is believed to be mainly affected
by surface complexing, which is strongly pH-dependent (Bau, 1999; Marmier et al.,
1999; Marmier and Fromage, 1999; Piasecki and Sverjensky, 2008), hence sorption of
REE usually increases with increasing pH within a range of 5-7 (Marmier et al., 1999;
Marmier et al., 1997; Marmier and Fromage, 1999; Piasecki and Sverjensky, 2008).
Surface complexing and pH do not affect Y as much as REE (Bau, 1999).
Fractionation of REE by Fe oxyhydroxides shows MREE enrichment at pH>5 in low
salinity solutions when other strong complex ligands are absent (Bau, 1999), indicating
La, Gd, Y and possibly Lu preferentially remain in the solution rather than being
surface-complexed onto Fe oxyhydroxides (Bau, 1999).
Chapter Two: Literature review
25
However, the fractionation between LREE, MREE and HREE in Fe oxides is subject
to debate (Laveuf and Cornu, 2009) and varied fractionation with enrichment of LREE
(Koeppenkastrop and Decarlo, 1993), MREE (Bau, 1999; Land et al., 1999) or HREE
(Elderfield and Greaves, 1981; Marker and Deoliveira, 1994) have been observed. The
differences in REE fractionation induced by Fe oxides probably arise from the
presence of various proportions of different types of Fe oxides and the presence of
other complex ligands (Laveuf and Cornu, 2009). For example, in solutions at
pH 4.0-7.1 with carbonate present, REE sorption by amorphous Fe oxyhydroxides
initially increases with increasing carbonate concentration and then decreases; this
effect was more pronounced for HREE than LREE (Quinn et al., 2006). In addition,
REE are fractionated during adsorption by Fe oxyhydroxides when humate complexes
are present, resulting in MREE enrichment (pH 5.2) rather than the non-preferential
adsorption by Fe oxyhydroxides or humate complexes (Davranche et al., 2004).
Generally, REE contents in amorphous Fe oxyhydroxides are higher than in crystalline
Fe oxides (Land et al., 1999; Laveuf and Cornu, 2009), and operationally defined
amorphous and crystalline Fe phases displaying enrichment of MREE were reported
by Land et al., (1999).
2.6.5 Anomalies of Ce in lateritic regolith
Cerium is one of the ‘unusual’ REE because it can occur in nature as Ce3+
like the
majority of lanthanides or as Ce4+
in oxidizing conditions. Cerium also has very low
elemental mobility, due mainly to the stability and low solubility of CePO4 and CeO2.
In lateritic profiles, topsoil and clay zones usually have either slight positive or no
apparent Ce anomalies, ferruginous zones commonly have a positive Ce anomaly, and
saprolite may show a negative or positive Ce anomaly (Angelica and Dacosta, 1993;
Braun et al., 1993; Braun et al., 1990; Braun et al., 1998; Ndjigui et al., 2009). In the
ferruginous zone, Fe acts as a redox sensitive element, and strong redox mediated
associations between the oxyhydroxide phases are expected within weathered profiles.
During the primary redox change in lateritic weathering, soluble Ce3+
is released by
Occurrence and behaviour and rare earth and associated elements in lateritic regolith profiles in Western Australia
26
chemical weathering from REE-bearing minerals and is oxidized to Ce4+
where pH
ranges from 2.5 to 7.5, and Eh from -0.4 to 1.1V; precipitation as CeO2 in the
ferruginous zones (Angelica and Dacosta, 1993; Braun et al., 1990; Braun et al., 1998;
Takahashi et al., 2000) is most likely in lateritic profiles, however, adsorption by
Fe/Mn oxyhydroxides also has the potential to produce a Ce enrichment (Ndjigui et al.,
2008; Nedel et al., 2010; Ohta et al., 2009; Quinn et al., 2006). The variability of
observed Ce anomalies in saprolite may relate to a change in redox condition induced
by fluctuation of groundwater or movement of the weathering front (Braun et al., 1990;
Ndjigui et al., 2009).
2.6.6 Anomalies of Eu in lateritic regolith
Although Eu anomalies are variable in lateritic weathered profiles (Table 2.3), they are
less studied than Ce anomalies. Negative Eu anomalies in the saprolite zone, or even
throughout the profile (e.g. Braun et al., 1993; Braun et al., 1998), may be caused by
breakdown of plagioclase (Panahi et al., 2000), sphene and allanite (Condie et al.,
1995), or the tetrad effect1 (Feng, 2011). Although the tetrad effect in REE patterns
has been reported widely in different geological samples (e.g. Liu and Zhang, 2005;
Monecke et al., 2002; Takahashi et al., 2002), it is still under debate (McLennan, 1994).
Positive Eu anomalies also exist in lateritic regolith (e.g. Braun et al., 1998; Ndjigui et
al., 2009), and they may relate to the type of parent rock and the redox conditions
during weathering (Ndjigui et al., 2009).
2.7 Summary
Although the geochemical behaviour of REE in supergene settings has been widely
investigated since the 1980s, compared with the studies of REE under high
temperature and high pressure settings it is far less studied, especially during intense
lateritic weathering. Many issues remain unresolved or not fully understood, such as
1tetrad effect: a split of chondrite-normalized REE patterns into four rounded segments which either are convex or
concave and formed M-shaped and W-shaped lanthanide pattern Masuda, A., Kawakami, O., Dohmoto, Y., Takenaka, T., 1987. Lanthanide tetrad effects in nature: two mutually opposite types, W and M. Geochemical Journal 21(3), 119-124.
Chapter Two: Literature review
27
migration-fixation mechanisms and mode of occurrence of REE in lateritic regolith,
the impact of Fe oxyhydroxides on translocation and fractionation of REE during
lateritization, the distribution of REE into different particle size solid phases and the
influence of weathering intensity on mobilization and fractionation of REE etc.. In
addition, many results are controversial and have not been interpreted unambiguously
yet; for example, the preferential mobilization of LREE or HREE during weathering,
and at which stage of weathering mobilization and fractionation of REE starts.
Therefore, the study of the geochemical behaviour and fractionation of REE under
intensely lateritic weathering is important and worthy of further research. This thesis
will improve the understanding of the mode of occurrence, fractionation mechanism
and geochemical behaviour of REE during weathering and lateritization in supergene
settings.
28
Table 2.1 Summary of REE in common minerals in granitoid rocks
2Ferruginous zone is defined by the dominant composition of secondary Fe oxides and oxyhydroxides, Al hydroxides and kaolinite; and the
3duricrust refers to a hard cemented
layer consisting secondary segregations. In this thesis all ferruginous zones consist of duricrust; however, in the references listed in the table, duricrust has been separated out to
emphasize the concentration variations of REE in different zones.
4ΣREE is the total concentrations of REE;
5Ce
*=(Ce/CePR)/[(La/LaPR)
0.5×(Pr/PrPR)
0.5];Eu
*=(Eu/EuPR)/[(Sm/SmPR)
0.5×(Gd/GdPR)
0.5];
6Data of Pr is missing, so Nd is used when calculating Ce
*;
7w refers to whole rock analysis.
33
3 Description of the study areas
3.1 General geology and climate
The study areas were located in the south-western part of Western Australia and lie
within the Darling Range, slightly east of the Darling Fault and Perth Basin (Figure
3.1). The geological history of the Darling Range can be traced back at least 2600
million years and possibly even further (Gozzard, 2007). Since Paleogene, deep and
intense weathering of exposed rocks of the Darling Plateaus, resulted in a widespread
cover of lateritic materials, and this weathering has continued until geologically recent
times (Gozzard, 2007). This area is part of the vast Yilgarn Craton - an ancient region
of varied rock types that occupies much of the south-western part of Western Australia
(Anand et al., 2006; Anand and Paine, 2002; Gozzard, 2007). Large volumes of
granitoids intruded the metamorphic rocks and other rocks of the Yilgarn Craton
between 2700 and 2600 million years ago and dolerite dykes intrude the granitoids
during development of the Darling Fault in the Mesoproterozoic and Neoproterozoic
(Gozzard, 2007).
The area currently has a Mediterranean climate, with a cool wet season from May to
September and a warm, dry season from November to March, with transition periods
in April and October. Average annual rainfall was ca. 1239.5 mm in the Dwellingup
(Darling Range) from 1934 to 2011 and rainfall mostly occurs in winter (Bureau of
Meterology, 2012).
The vegetation of the areas studied shows marked regional changes based largely on
climate, with local variations of geology, soils, topography and drainage. The Darling
Range in the high-rainfall area currently has open eucalypt forests of jarrah
(Eucalyptus marginata) and marri (Eucalyptus calophylla) (Anand and Paine, 2002).
3.2 Sampling and profile description
The profiles studied are Fe-rich lateritic weathered profiles near outcrops of granitoids
Occurrence and behaviour and rare earth and associated elements in lateritic regolith profiles in Western Australia
34
and dolerites. The dolerites are sub-vertical intrusions that cross cut the metamorphic
fabric of the granitoids and thus called dolerite dykes here. Samples of outcropping
fresh granitoid and dolerite were sampled ca. 5 kg separately at each study site.
Different zones of regolith in each profile were identified based on the physical
properties (e.g. texture, colour, coarse fragment content) and chemical properties (e.g.
mineralogy, Al/Fe concentrations, and organic matter contents). One to two ten-cm
blocks of bulk undisturbed regolith samples per horizon were collected at different
depth and sealed in plastic bags or boxes for transport to the laboratory.
Great Eastern Profile: The first intensely weathered lateritic profile investigated was
adjacent to Great Eastern Highway, Western Australia (31°22'30.95"S, 118°41'27"E),
with very good bedrock (granitoid/dolerite dyke) exposure (Figure 3.1&Figure 3.2).
Sample identities from this location are prefixed with GE. Regolith samples from
different horizons, the parent granitoid and the intrusive dolerite dyke were collected
on 3rd
April, 2009. One outcropped coarse grained granitoid sample (GEPR1B) was
crossed by a late-stage sub-horizontal pegmatite vein (GEPR1A), thus these two
subunits were analysed separately. The profile was ca.12 m deep, including saprolite,
mottled clay, ferruginous duricrust, and A horizon regolith. The saprolite formed from
weathered bedrock with horizons above showing progressive loss of rock fabric
upwards as porosity and the proportion of clay increases. The mottled clay zone was a
pale white kaolinite-rich zone, ca. 3 m thick, with distinct upper boundary with the
ferruginous zone. The ferruginous zone was composed of loose lateritic ferruginous
materials with ferruginous gravel at ca. 7 m depth (GE5), and the cemented
ferruginous duricrust which was a dark red, dense, and hard layer without gravel at ca.
3.5 m depth (GE6).
Mountain Quarry Profiles: The second and third profiles studied were located in
Mountain Quarry (31°54'54" S, 116°3'44" E), on the southern slope of Greenmount
Hill, Western Australia (Figure 3.1). Sample identities are prefixed with MQ. The
second profile (MQ I profile, Figure 3.3) was 3.6 m deep; samples of regolith from
different horizons based on colour and texture were collected on 29th
, May, 2009. The
Chapter Three: Description of the study areas
35
third profile (MQ II profile) was ca. 2 m deep, 10 m away from the second profile,
developed overlying granitoid with stonelines preserved from quartz veins which
imply in-situ weathering below the sub-horizontal component of the stone line at 0.6 m
depth (samples MQ10 to MQ13). Samples of outcropping granitoid and dolerite and
the regolith were also collected at the MQ sampling site.
Jarrahdale profile: The fourth profile was located at the Jarrahdale Railway cutting
(32°17'46"S, 116°5'40"E) at an average elevation of 270 m above sea level in the
Darling Range, 80 km south-east of Perth, Western Australia (Figure 3.1). Sample
identities were prefixed with JG. The lateritic JG profile was ca. 12 m deep overlies
metamorphic basement consisting of granitoid intruded by a dark-coloured dolerite
dyke. The location of intrusive dykes in the granitoids can be mapped from the
overlying duricrust and the profiles developed on granitoid and dolerite are distinctly
different (Gozzard, 2007). The JG regolith is developed overlying granitoids and
zone, ferruginous duricrust, upper ferruginous zone and the A horizon. The mottled
clay is pale white kaolinite-rich, consisting of a lower zone (JG2) at 8.6 m depth and an
upper zone (JG3) at 6.5 m depth. The ferruginous duricrust (3 m depth) is gibbsite and
goethite rich. The upper ferruginous zone (JG6, 1.5 m depth) is rich of red iron nodules.
In contrast, the horizon A regolith (JG7-10, <1 m depth) is gravely sandy soil rich of
dark brown to black loose nodules. The sampling of the profile was conducted on the
6th
August 2009. Each zone was identified based on different properties (e.g. texture,
colour, coarse fragment content), sampled at the depth given above in a
ca. 10×10×10 cm cube, put into a sealed plastic box, transported to the laboratory and
air dried. Photographs of each horizon were not taken and thus are not presented in this
thesis.
A number of studies have investigated the geological, morphological and geochemical
characteristics of lateritic bauxite regolith in the Darling Range, and at Jarrahdale the
regolith is widely accepted to have undergone in-situ intense weathering and
lateritization (Anand and Butt, 2010; Anand et al., 1991; Anand and Paine, 2002;
Occurrence and behaviour and rare earth and associated elements in lateritic regolith profiles in Western Australia
36
Brimhall et al., 1992; Brimhall et al., 1994; Gozzard, 2007; Kew and Gilkes, 2007;
Sadleir and Gilkes, 1976). Note that both fresh and weathered euhedral zircons from
lateritic bauxite profiles at Jarrahdale have been dated at ca. 2650 Ma (Brimhall et al.,
1994) indicating upward lithological continuity of the parent meta-granitoid through
the bauxite profile. Rounded zircon, ilmenite, and rutile have been found in the
surficial meter-depth regolith and are proposed to have been transported by wind and
predominantly from a different, much younger source (ca. 700-1150 Ma) (Brimhall et
al., 1992; Brimhall et al., 1994; Brimhall et al., 1988).
37
Figure 3.1 Sampling sites (a, labelled as box) and sketches of the profiles sampled (b). On the map (a) the dashed line labelled Darling Fault represents
the western margin of the Darling Range. In the sketch of regolith profile (b) ‘m’ denotes matrix and ‘g’ denotes gravel.
38
GE1, 12.5 m depth, saprolite GE3, 10 m depth, mottled clay GE5, 7 m depth, lower ferruginous
zone
Figure 3.2 Photographs of regolith from selected horizons of the GE profile.
MQ1, 3.6 m depth, C horizon MQ2, 3.3 m depth, lower B horizon MQ4, 2.2 m depth, upper B horizon
Figure 3.3 Photographs of regolith from selected horizons of the MQ I profile
39
4 Redistribution of major elements in lateritic profiles during
intensive weathering in Western Australia
4.1 Abstract
In order to understand the geochemical behaviour of major elements in different solid
phases in laterite and to investigate geochemical pathways of lateritic weathering, the
redistribution of major elements in matrix (<2 mm) and gravel (>2 mm) in four
intensely weathered lateritic profiles (GE, MQ I, MQ II and JG) in Western Australia
was investigated.
The GE and JG regolith samples were highly weathered with chemical indices of
alteration ca. 99%, nearly complete loss of Na, Ca and Si, and enrichment of Fe in the
ferruginous zone. In the GE and JG profiles, Fe mainly occurred as goethite, hematite
and maghemite, while Al mainly occurred as secondary clay minerals (kaolinite) and
gibbsite in ferruginous zone; gravel was more enriched in Al and Fe but more depleted
in Si than the matrix, which was consistent with gravel having higher weathering
intensity and degree of lateritization than the matrix.
The regolith samples from both MQ profiles were less weathered than the GE and JG
profiles, lack of gibbsite and hematite, and showed weak lateritization. The chemical
indices of alteration ranged from 55% to 92% in both MQ profiles and gravel had
higher concentrations of Si than matrix. The presence of a pedogenic discontinuity in
both MQ profiles was identified from the molar ratio Na/K, concentration ratios
Al2O3/Fe2O3 and Ti/Zr, implying that mass movement had occurred in the upper part of
both profiles during weathering.
Significant depletion of base cations and Si, coupled with enrichment of Fe and Al,
reveal that intense leaching of cations, kaolinization, desilication and ferruginization
took place in lateritic regolith during weathering and lateritization.
4.2 Key words
Major elements; laterite; weathering; mass balance; Western Australia;
Occurrence and behaviour and rare earth and associated elements in lateritic regolith profiles in Western Australia
40
4.3 Introduction
Chemical weathering is one of the critical processes in the geochemical cycling of
elements and translocation of elements from crust to sediments. During early
pedogenesis, the chemical composition of a soil will be strongly controlled by the
composition of geological parent materials, though this influence diminishes in
importance with time (Schaetzl and Anderson, 2005; Thanachit et al., 2006). The
development of a soil reflects the weathering processes associated with the dynamic
environment in which it has formed. The mobilization and redistribution of elements
during weathering follows various pathways as elements behave differently during
various pedogenic processes, including: dissolution of primary minerals, formation of
secondary minerals, redox processes, transport of material and ion exchange
(Middelburg et al., 1988).
Lateritic regolith represents one of the most common superficial formations in the
tropics, and is commonly diachronous, extending over tens of millions of years
(Dequincey et al., 2006). In contrast to common pedogenesis, during lateritic
weathering, regolith is intensely weathered and enriched in Fe, either in some layers or
throughout the profile, commonly forming a hard cap of ferruginous duricrust or
ferruginous gravel. Though many studies have been conducted on lateritic regolith
profiles (e.g. Beauvais, 1999; Brimhall et al., 1991; Brown et al., 2003; Costa, 1997;
Dequincey et al., 2002; Dequincey et al., 2006; Fernández-Caliani and Cantano, 2010),
the mobilization and redistribution of major elements into different grain size fractions
of lateritic regolith during intense weathering are not yet fully understood. A holistic
understanding of elemental behaviour during weathering and lateritization processes
cannot be achieved solely by determination of total elemental concentrations in bulk
regolith. It is essential to determine the relative elemental concentrations in different
solid phases as well, since partitioning of elements into matrix (<2 mm) or gravel
(>2 mm) may reflect the weathering history and weathering processes involved. The
analysis of the geochemical and mineralogical features of lateritic regolith, including
matrix and gravel, has the potential to improve our understanding of weathering and
lateritization (Beauvais, 1999).
A number of studies have been investigated the geological, geographical, morphological
and geochemical characteristics of lateritic bauxite regolith in the Darling Range
(Anand and Butt, 2010; Anand et al., 1991; Anand and Paine, 2002; Brimhall et al.,
Chapter Four: Redistribution of major elements in lateritic profiles during intensive weathering in Western Australia
41
1992; Brimhall et al., 1988; Kew and Gilkes, 2007; Sadleir and Gilkes, 1976).
Accordingly, this study investigates the mobilization and redistribution of major
elements into different solid phases from lateritic regolith building on previous work.
The objective of this study is to investigate the key geochemical and mineralogical
pathways of lateritic weathering and the relative partitioning of major elements into
matrix and gravel in four intensely weathered lateritic profiles in Western Australia.
This information will be helpful to understand the genesis of ferruginous materials
during lateritization.
4.4 Materials and methods
4.4.1 Analytical methods
In this study, chemical compositions of matrix (<2 mm, represented by suffix ‘m’) and
gravel (>2 mm, represented by suffix ‘g’) were analysed separately. Exceptions were
the duricrust in the GE profile (GE6), a very hard cemented material without
corresponding loose matrix or iron nodules, and the saprolite (JG1) and mottled clay
(JG2&3) in the JG profile, both soft pale materials without gravel. These four samples
were crushed and/or ground to ≤ 200 µm and oven dried at 105 °C overnight prior to
chemical analysis.
Pre-treatment of regolith samples included hand-picking roots/rhizomes and sieving
through 2 mm plastic mesh for separation of gravel (>2 mm) from the matrix (<2 mm)
and weighing each subsample separately. The matrix fraction (<2 mm) was used to
determine the pH, total carbon and particle size distribution (Table 4.1). Soil pH was
determined potentiometrically at 23 °C in the supernatant in a 1:5 suspension of soil:
deionised water and 1:5 suspension of soil: 0.01 M CaCl2 (Rayment and Higginson,
1992). Total carbon was determined by Elementar (Vario Macro, Hanau, Germany).
Subsamples of matrix and gravel were ground to ≤ 200 µm and oven dried at 105 °C
overnight. The bulk raw regolith matrix (< 2 mm) from MQ II profile was separated
further into three size fractions without crushing or grinding: clay (<2µm), silt (2-20 µm)
and sand (> 20 µm) by the sedimentation and wet sieving method (Day, 1965) in order
to understand the behaviour of major elements in different particle size fractions. The
particle size fraction limit recommended by the International Society of Soil Science
(ISSS) has been adopted in Australia (Marshall, 1947; Marshall, 2003; Prescott et al.,
1934). Different particle size fractions were rinsed with MilliQ water three times, oven
Occurrence and behaviour and rare earth and associated elements in lateritic regolith profiles in Western Australia
42
dried at 105 °C overnight and ground to ≤ 200 µm.
Fusion beads for elemental analyses were made by mixing 0.1 g (to an accuracy of
0.1 mg) of finely ground sample or reference material with 0.7 g 12:22 Norrish flux
(lithium metaborate:lithium tetraborate=12:22) and heating in a muffle furnace at
1050 °C for 40 minutes. Duplicate fusion beads were made on 10% of samples to check
preparation errors. After cooling, the fusion beads were dissolved in 100 mL of 10%
analytical grade HCl. The major elements were determined by inductively coupled
plasma-optical emission spectroscopy (ICP-OES, Perkin-Elmer Optima 7300DV) at the
University of Western Australia. Certified international standard materials, including
stream sediment reference standards STSD-2, STSD-4 (Canada Centre for Mineral and
Energy Technology, CANMET), an in-house standard reference and 12 blanks were
prepared in the same way as the samples and analysed together with samples to check
the accuracy and precision. The variation between the tested and expected values of the
standards was within 5% (Appendix 11.2). The concentrations of major elements in
matrix and gravel are listed in Table 4.2.
Primary minerals in the weathering products were identified by means of random
powder X-ray diffraction (XRD) from 4 to 70 2 using CuKα radiation and a Philips
PW 1830 diffractometer with a diffracted beam graphite crystal monochromator, after
grinding to <63 µm and homogenisation. The proportion of mineral phases (not include
amorphous or poorly-crystalline phases) were identified semi-quantitatively using the
software Traces (GBC Scientific Equipment). All primary mineral phases were
identified manually and cross-checked with the dataset of the International Centre for
Diffraction Data (ICDD). The main-peak area of each primary mineral was measured,
and the mineral proportion was calculated using the main-peak areas of each primary
mineral divided by the sum of main-peak areas of all primary minerals identified by the
Traces software. The clay fraction (<2 µm) was separated by dispersion and
sedimentation, basally oriented on ceramic slides, air-dried, and then scanned at
1° 2/min from 1 to 30° 2/min. For clay mineral identification, the oriented aggregates
were treated with ethylene glycol. Mineral names were abbreviated according to
Whitney and Evans (2010).
Chapter Four: Redistribution of major elements in lateritic profiles during intensive weathering in Western Australia
43
Table 4.1 Selected physical and chemical properties of matrix fractions (<2mm) of the
refers to the weight proportion of gravel (g) or matrix (m) in percentage; ‘W
2’ refers to this sample without corresponding gravel/matrix and thus determined by
whole-rock analysis; 2d.l. refers to detection limit.
3GEPR3 and
4 MQPR3 are dolerite, whereas other rock samples are granitoid.
4Batch refers to each time determination of element concentrations based on the profile;
5RSD is the range of relative standard deviations (precision) of the duplicates/triplicates analysed by ICP-OES.
51
Figure 4.1 Ternary A-CN-K and A-FM-CNK plots of regolith samples from four lateritic profiles (GE, MQ I, MQ II, JG) based on chemical
compositions of matrix and gravel samples. Dashed line with arrow indicates weathering trend for MQ profiles. (a) and (e) GE profile; (b) and (f) JG
profile; (c) and (g) MQ I profile; (d) and (h) MQ II profile. Triangles represent granitoids and circles represent dolerite, whilst squares indicate regolith
gravel and diamonds indicate matrix samples.
Occurrence and behaviour and rare earth and associated elements in lateritic regolith profiles in Western Australia
52
4.5.2 Mineralogical properties
In the GE profile, the parent granitoid (GEPR2) was characteristically identified by
Figure 4.3 Mass balance of major elements in regolith samples from four lateritic profiles, based on weighted average concentrations of major elements
in matrix and gravel at each depth, using Zr as the reference element (Brimhall et al., 1991): (a) the GE profile; (b) the MQ I profile; (c) the MQ II
profile; (d) the JG profile (the vertical dashed line refers to τ(Zr) = 1.0 without depletion or accumulation relative to the parent granitoid; trend of Ca
was similar to Na, and Mg to K, so not plotted above).
57
Figure 4.4 Depth functions of the molar ratio Na/K and concentration ratio Al2O3/Fe2O3 for MQ two profiles and concentration ratio (Ti/Zr)/10 for four
profiles illustrating the pedogenic discontinuity (shown by the boxed area) at 1.1 m depth in MQ I profile and 0.6 m depth in MQ II profile.
58
Figure 4.5 Major element concentrations in grain size fractions of the regolith samples from the MQ II profile.
Chapter Four: Redistribution of major elements in lateritic profiles during intensive weathering in Western Australia
59
4.5.4 Depth functions of pedogenic discontinuities
Pedogenic discontinuities represent zones of change in physical and chemical
properties primarily originating from processes during soil development, rather than
the parent lithology, which is more commonly referred as lithological discontinuities
(Schaetzl and Anderson, 2005; Tsai and Sang, 2000). In both MQ profiles, the
distribution of major elements in A horizon regolith is not correlated to the weathering
trend (Figure 4.1). In addition, a stone line observed within the MQ II profile may
indicate either the presence of an erosional episode at/near a discontinuity (Parsons and
Herriman, 1966; Ruhe, 1958) or a mass movement with biogenic involvement
(Johnson, 1990; Johnson and Balek, 1991; Lichte, 2000). Either of these reasons
indicates the presence of pedogenic discontinuity in both MQ profiles. Therefore,
depth functions for various pedogenic parameters, including molar ratio Na/K,
concentration ratios Al2O3/Fe2O3 and (Ti/Zr)/10 were prepared for the MQ I and MQ II
profiles (Figure 4.4).
In the MQ profiles, abrupt changes of Na/K, Al2O3/Fe2O3 and (Ti/Zr)/10 occurred at
1.1 m depth of the MQ I profile and at 0.6 m depth of the MQ II profile, including
gravel and matrix. The (Ti/Zr)/10 first increased upwards and then decreased until
surface soil, showing great variations (variation up to > 70% in MQ I profile and > 50%
in MQ II profile relative to the parent granitoid). However, below 1.1 m depth of the
MQ I profile and at 0.6 m depth of the MQ II profile, (Ti/Zr)/10 was relatively
consistent. The depth of abrupt change coincided with the presence of stone line in the
MQ II profile, implying that a mass movement (i.e. erosion and re-deposition with or
without biological recycling) had occurred during the weathering history of these two
profiles.
In the GE and JG profiles, (Ti/Zr)/10 was relatively consistent in the saprolite and
mottled clay zone, close to the parent granitoid (variation < 30%), including both
gravel and matrix; then Ti and Zr in gravel and matrix started to fractionate in the
ferruginous zone, and significantly differentiated from each other in the A horizon. The
variation of (Ti/Zr)/10 in the ferruginous zone showed that Ti and Zr partitioned
Occurrence and behaviour and rare earth and associated elements in lateritic regolith profiles in Western Australia
60
between gravel and matrix at the advanced stage of weathering and lateritization and
further details will be discussed in the next chapter using the JG profile as an example.
4.5.5 Grain size distribution of major elements in MQ II profile
The distribution of major elements into different grain size fractions from the MQ II
profile (Figure 4.5) showed that Al, Fe, Mg, Ti and Na were enriched in the clay and
silt fractions. Potassium and Si, however, were enriched in the sand and gravel
fractions. The silt fraction had the highest Zr concentrations. In the A horizon, gravel
had the highest concentrations of Ca; however, in the B and C horizons, sand was the
main host for Ca.
4.6 Discussion
4.6.1 Significant processes during lateritization
Based on the geochemical and mineralogical data, the regolith from these four lateritic
profiles has experienced moderate to extreme weathering. The mass balance
calculations (Figure 4.3) reflect the substantial loss of alkaline and earth-alkaline
elements in the saprolite of the GE and JG profiles, in agreement with the depletion of
plagioclase (Figure 4.2). Compared with Na and Ca, K and Mg were less depleted and
corresponded to the residual occurrence of muscovite and potassium feldspar. Both
alkalis and Si continued to be lost as weathering intensified. Kaolinite was further
altered into gibbsite and Fe was enriched in the ferruginous zones as goethite and
hematite (Figure 4.2). Similar trends have been reported by Anand and Paine (2002).
At pH above 5, Si (even from quartz) has a higher solubility than Al- and Fe- oxides
(Breemen and Buurman, 1998), and thus would be preferentially removed from the
system. The near-complete loss of Na and Ca and significant enrichment of Al and Fe
in the ferruginous zone indicate that the regolith has undergone an advanced stage of
lateritization. Only residual or partially weathered quartz and some
weathering-resistant minerals remained in the ferruginous regolith.
Both MQ profiles were less weathered than GE and JG profiles (Figure 4.1), and the
Chapter Four: Redistribution of major elements in lateritic profiles during intensive weathering in Western Australia
61
chemical composition of regolith showed intermediate depletion of Si, Na and K and
enrichment of Al and Fe (Figure 4.3). The plots of SiO2 vs. Al2O3 and Fe2O3 vs. Al2O3
(Figure 4.6) clearly show that (i) the regolith was weathered in-situ from granitoids; (ii)
silicon consistently decreased whereas Al and Fe increased with increasing weathering
intensity. Aluminium was residually enriched by dissolution of feldspar and mica and
the formation of kaolinite at the expense of primary minerals (kaolinization) in the B
horizon of both MQ profiles and in the mottled clay zone of the GE and JG profiles
(Figure 4.2). Idealized weathering reactions for dissolution of silicates, e.g. feldspar
and muscovite, and for formation of kaolinite at the early stages of weathering are
presented below:
CaAlSi3O8+H2O+2H+⇌Al2Si2O5(OH)4+Ca
2+ (loss of Ca)
2KAl3Si3O10(OH)2+3H2O+2H+⇌3Al2Si2O5(OH)4+2K
+ (loss of K)
4.6.2 Genesis and sources of Fe redistribution
The nature of formation of iron oxides is generally more dependent on the
environmental conditions at the time of formation than on the particular structures of
the primary mineral from which the Fe was released (Anand and Paine, 2002). Under
anaerobic conditions, Fe can mobilize and redistribute over a range of spatial scales.
This mobilization can be vertical or lateral, at a horizon, or even a landscape, scale.
In the GE and JG profiles, residual accumulation of Fe is likely to be the result of
continuous cyclic dissolution-precipitation processes with redox changes occurring as
a result of changes in regolith water regime (Anand and Butt, 2010; Anand and Paine,
2002; Tripathi and Rajamani, 2007). Redistribution of Fe in the duricrust may be
related to a capillary effect triggered by seasonal fluctuation of the water table (Braun
et al., 1993; Ndjigui et al., 2009). In addition, in the JG profile, the location of intrusive
dykes in the granitoids can be mapped from the overlying duricrust and the profiles
developed on dolerite and granitoid are distinctly different; both of which suggest that
duricrust has formed in-situ (Anand and Paine, 2002; Sadleir and Gilkes, 1976). The
regolith from the GE profile was also formed residually from weathering and
pedogenic processes; the complete profile components (including saprolite, mottled
Occurrence and behaviour and rare earth and associated elements in lateritic regolith profiles in Western Australia
62
clay and ferruginous zone) occur without any discontinuity, erosion or missing zones
and no apparent enrichment of any trace elements derived from external sources in
subsurface regolith (regolith <1 m depth given aeolian input) was observed. Therefore,
Fe enrichment in the GE profile is more likely a result of vertical residual
accumulation rather than lateral movement.
In the MQ profiles, transition zones (at 1.1 m depth in MQ I profile and 0.6 m depth in
MQ II profile) showed the lowest Al2O3/Fe2O3 ratio. It is possible that the Fe released
from primary minerals was partially oxidized by oxygen in penetrating
rainwater/atmosphere and, as a result, was precipitated in the upper part of the regolith
during weathering. With continued weathering and the influence of soil creep or
colluviation, and alternative wetting and drying of the soil due to seasonality or
longer-term changes between humic and arid climates, translocation and redistribution
of Fe was facilitated.
Chapter Four: Redistribution of major elements in lateritic profiles during intensive weathering in Western Australia
63
Figure 4.6 The distribution of Al2O3 vs. SiO2 and Al2O3 vs. Fe2O3 in matrix and gravel
from four lateritic profiles (A_g: gravel of A horizon; A_m: matrix of A horizon;
regolith_g: gravel of subsurface regolith; regolith_m: matrix of subsurface regolith; in
(g) and (h), dolerite had Fe2O3 concentration 13.7 wt%, so not showed).
Occurrence and behaviour and rare earth and associated elements in lateritic regolith profiles in Western Australia
64
4.6.3 Degrees of lateritization
To classify the degree of lateritization and demonstrate the effects of weathering
intensities on lateritic regolith, ternary SiO2-Al2O3-Fe2O3 diagrams (Schellmann, 1981)
were plotted (Figure 4.7). The progression from fresh parent granitoids to loss of Si
and relative enrichment of Fe and Al involves various degrees of weathering intensity.
Al2O3 and Fe2O3 only separate under the extreme weathering (strong lateritization).
The GE and JG profile have apparently undergone strong lateritization, whereas both
MQ profiles are still within the ‘weak lateritization’ status, but have both undergone
intense kaolinization.
Figure 4.7 Schellmann SiO2-Al2O3-Fe2O3 diagrams showing different degrees of
lateritization of weathered regolith from four lateritic profiles: (a) the GE profile; (b)
the MQ profiles; (c) the JG profile; regolith include matrix and gravel.
4.6.4 Principal components analysis
In order to systematically analyse the data, chemical compositions of all regolith from
four profiles are standardised and subjected to principal component analyses. Two
principal components extracting 76.0% of variance in major element data are
recognized and plotted in Figure 4.8 and the factor score of each regolith sample is
plotted in Figure 4.9.
Principal Component 1 was controlled by base cations (positive loadings) and
conservative elements such as Ti, Zr, Al and Fe (negative loadings). Parent granitoids
were enriched in base cations (ellipse a in Figure 4.9), with moderate weathered MQ
regolith having some base cation depletion (ellipse b). In contrast, GE and JG regolith
Chapter Four: Redistribution of major elements in lateritic profiles during intensive weathering in Western Australia
65
was characterized by more residually accumulated Zr, Ti, Al and Fe (ellipse c, d, e and
f). This reflects the relative mass flux depletion of base cations in weathered regolith
and relative conservation of Al and Fe in ferruginized regolith during lateritic
weathering. Principal Component 2 was mainly affected by Si, P and Mn. The regolith
of mottled clay and ferruginous mottled zones from the GE and JG profiles contained
high concentrations of Si and separated from the other regolith, indicating initial
residual enrichment of Si due to depletion of base cations without desilication. In
ferruginous zone of the GE and JG profiles, this Si enrichment was weakened by
desilication and formation of gibbsite. The ferruginous surface gravel from the GE and
JG profiles had lower concentrations of Si but higher concentrations of P and Mn than
corresponding matrix, and thus, separated from the matrix and the subsurface regolith.
The C horizon regolith from the MQ I profile and the saprolite from the JG profile (see
arrows) were close to parent granitoids, reflecting: (i) incipient weathering conditions;
(ii) weak depletion of base cations; and (iii) the genetic relationship with the granitoid.
In addition, different types of granitoids were slightly differentiated from each other;
e.g., the coarse grained granite was separated from the late-stage sub-horizontal
pegmatite vein.
Occurrence and behaviour and rare earth and associated elements in lateritic regolith profiles in Western Australia
66
Figure 4.8 Principal component analyses of major elements in regolith samples and
parent granitoids from four lateritic profiles. Compositional data were transformed
using centered log-ratios.
Figure 4.9 Principal component factors of regolith samples and parent granitoids from
four lateritic profiles calculated using major element composition.
Chapter Four: Redistribution of major elements in lateritic profiles during intensive weathering in Western Australia
67
4.6.5 Mineralogy and element grain size distribution
The distribution pattern of major elements into different subgroups can be interpreted
by regolith mineral characteristics. In the GE and JG profiles, gravel was more
enriched in Al and Fe but more depleted in Si, whereas matrix was more enriched in Si
but more depleted in Al and Fe (Figure 4.10). This corresponds to the weathering
intensity and mineral composition of gravel and matrix. Both profiles are more
intensely weathered; cyclic reducing and oxidizing conditions likely lead to mottle
formation, with Fe oxides and oxyhydroxides migrating into clay matrix or voids.
Repeated dissolution and cementation results in the formation of gravel, which are
dominated by gibbsite, goethite and hematite. The occurrence of maghemite in
near-surface regolith in the JG profiles may be induced by a combination of heat (bush
fire) to dehydroxylate goethite and organic matter (Anand and Gilkes, 1987; Anand
and Paine, 2002; Perrier et al., 2006). However, in both MQ profiles, Si in gravel was
higher than in matrix, but Al and Fe varied with the depth. The MQ gravel containing
higher Si likely results from less weathering intensity and relatively fast drainage. The
absence of gibbsite in both MQ profiles reveals that kaolinite did not further alter into
gibbsite and no apparent partitioning of Al occurred between matrix and gravel in MQ
regolith profiles. Variation of Fe with depth correlates to the weathering intensity of
matrix and gravel samples. Therefore, the element distribution between matrix and
gravel is a reflection of the mineralogical composition, which is more dependent on the
weathering environment and weathering intensity than the parent lithology.
In addition, the enrichment of Al, Fe, Mg, Ti and Na in silt and clay fractions indicates
that Al released from feldspar, muscovite, etc., in the parent granitoids formed clay
minerals and sesquioxides in the regolith. Iron released from altered magnetite and
other iron-bearing phases in the parent granitoids produced goethite in the regolith.
Due to different charge/ionic radius and hosting minerals, the distribution patterns of
Na and K, Mg and Ca were not the same. These elements inherited from the parent
granitoids, separate from each other during weathering and pedogenic processes
depending on the mineral weathering rate and weathering conditions. The differences
of charge/ionic radius also affect the ability of ion exchange of these elements onto
Occurrence and behaviour and rare earth and associated elements in lateritic regolith profiles in Western Australia
68
clay minerals, Al and Fe oxides/oxyhydroxides and organic matter.
In addition, Si was residually enriched in the sand fraction as quartz, and Zr was
residually enriched in the silt fraction as zircon. Enrichment of Ti in the fine fractions
indicated the physical, and possibly chemical, weathering of ilmenite and rutile or
newly formed anatase (Anand et al., 1991; Anand and Paine, 2002). The geochemical
distribution trends of major elements in different grain size fractions are therefore
mainly controlled by the physical characteristics and chemical stabilities of hosting
minerals inherited from the parent rock and secondary formed minerals in the regolith,
which in themselves are constrained by the weathering conditions.
4.6.6 Mobility of Ti and Zr
Although Ti and Zr are commonly considered to be and most frequently used as the
least immobile elements in weathered profiles (Beyala et al., 2009; Braun et al., 1993;
Brimhall et al., 1991; Nesbitt and Markovics, 1997; Taboada et al., 2006b),
mobilization of Zr induced by eroded zircon and redistribution of Ti during extreme
weathering have been discussed (Anand and Paine, 2002; Anand et al., 2010; Braun et
al., 1993). Titanium occurs in rocks mainly as rutile, ilmenite and sphene, or in the
structure of silicates such as micas, amphiboles and pyroxenes. The susceptibility of
silicate minerals to weathering results in release of some Ti in the early stage of
weathering of igneous rocks and continuing release as weathering proceeds (Anand
and Paine, 2002). Zirconium occurs in rocks largely as zircon, which is very resistant
to weathering and hence Zr is not considered to be transported under low temperature
and low pressure conditions (Henderson, 1984; Linnen et al., 2005; Vos et al., 2006).
In intensely weathered profiles (e.g., laterite), it is difficult to identify the most
appropriate immobile element(s) not subject to dissolution and physical translocation
given different scales.
In the GE and JG profiles, a variation of (Ti/Zr)/10 from the ferruginous zone upwards
was observed, illustrated that either or both Zr or Ti were relatively mobile under
extreme weathering conditions. In the Jarrahdale bauxitic lateritic profile, both rutile
Chapter Four: Redistribution of major elements in lateritic profiles during intensive weathering in Western Australia
69
and zircon have been introduced by aeolian input, inducing higher concentrations of Ti
and Zr in the surface soils (<1 m depth) (Brimhall et al., 1988; Brimhall et al., 1991;
Foo, 1999; McLennan, 1995). The euhedral grains of zircon, observed by scanning
electron microscopy in the regolith of four profiles showed zircon’s stability. However,
eroded ilmenite and rutile, and poorly crystalline Ce, Zr and Th (hydr)oxides, were
observed in the ferruginous zone of the JG profile, suggesting mobility of both Ti and
Zr at the sampling scale during the advanced stages of weathering and lateritization.
The mobilization and redistribution of Ti, Zr and Th will be investigated further in the
next chapter taking the JG profile an example.
4.7 Summary of the chapter
In this chapter, the bulk geochemistry of major elements in matrix and gravel from
four intensely weathered lateritic profiles (GE, MQ I, MQ II and JG) was investigated.
The regolith from the GE and JG profiles had undergone intense weathering and strong
lateritization with the CIA above 99% in the ferruginous zone. Both MQ profiles were
less weathered than the GE and JG profiles with only weak lateritization.
In the GE and JG profiles, gravel was more enriched in Al and Fe but more depleted in
Si than matrix. In both MQ profiles, however, gravel had higher Si than the associated
matrix with varied concentrations of Al and Fe. In the ferruginous gravel of the GE and
JG profiles, Fe was mainly enriched as goethite, hematite and maghemite, while Al
was mainly enriched as secondary clay minerals and gibbsite. In both MQ profiles,
however, lower weathering intensities and relatively fast drainage resulted in the
absence of gibbsite and hematitie and low concentrations of goethite.
Using Zr as the reference element, Ca and Na were near-completely depleted and large
loss of K, Mg and Si also occurred in the GE and JG profiles, indicating breakdown of
plagioclase and incongruent dissolution of potassium feldspar, muscovite and quartz at
advanced stages of weathering. The increasing proportion of gibbsite in the duricrust
and ferruginous gravel of the GE and JG profiles revealed that kaolinite was further
altered into gibbsite and cemented with iron oxides with weathering intensifying.
Occurrence and behaviour and rare earth and associated elements in lateritic regolith profiles in Western Australia
70
Taking the MQ II profile as an example, Al, Fe, Mg, Ti and Na were enriched in clay
and silt fractions, K and Si was enriched in sand and gravel fractions and Zr was
enriched in the silt fractions as zircon. The redistribution of elements into different
grain size fractions is mainly controlled by the physical characteristics and chemical
stabilities of hosting minerals inherited from the parent rock and newly formed
secondary minerals in the regolith. Intense leaching of cations, kaolinization,
desilication and ferruginization were identified as significant processes during
lateritization using principal component analysis, and these mechanisms were further
substantiated by the geochemical mass balance calculations and mineralogical
analyses.
71
Figure 4.10 Calculated τ values of Al, Fe and Si referenced to Zr in matrix and gravel from four lateritic profiles (the dashed line indicates τ(Zr) = 0,
without enrichment or depletion).
73
5 Redistribution and mobilization of Ti, Zr and Th in an intensely
weathered lateritic profile in Western Australia
5.1 Abstract
The mobility of titanium, zirconium and thorium, elements commonly considered
insoluble during supergene weathering, is still not well understood, especially in
intensely weathered regolith. Thus, an intensely weathered lateritic profile (JG)
developed on meta-granitoids in Jarrahdale, Western Australia, was investigated. The
mobility of Ti, Zr and Th has been assessed at both mineral assemblage and profile
scale and the mode of occurrence has been investigated through the combined use of
geochemical data from bulk regolith, particle size fractions and sequential extractions,
with in-situ data determined by electron probe micro-analyzer and synchrotron X-ray
powder diffraction.
Neoformed poorly crystalline phases containing trace to minor amounts of Zr, Ce and
Th unassociated with silicates or phosphates were identified on the walls of Al/Fe-rich
pores in the ferruginous duricrust. This implies that some mobilization and
redistribution of Zr and Th occurs within a sample scale. Breakdown of primary thorite
and rare earth element rich fluorocarbonates is thought to be the source for Zr and Th in
the neoformed phases rather than zircon. Thus, the mineral hosts of Zr, Ti and Th in the
parent rock and their relative susceptibility to weathering are the fundamental controls
on subsequent mobility during initial weathering. Trace amounts of Th in secondary
phases, such as rhabdophane and florencite, shows translocation of Th at the mineral
scale; whilst strong partitioning of Th into gravel rather than matrix reflects
redistribution of Th at the profile scale. The absence of primary sphene from the
regolith and the presence partially dissolved ilmenite and rutile grains in the ferruginous
mottled zone suggest mobilization and translocation of Ti at a mineral assemblage scale.
Furthermore, the fluctuation of Ti/Zr in the ferruginous zone is in contrast to the
consistency of Zr/Hf throughout the profile in general (within the range of parent
meta-granitoid). This suggests that Ti and Zr fractionate from each other and partition
Occurrence and behaviour and rare earth and associated elements in lateritic regolith profiles in Western Australia
74
between gravel and matrix during extreme weathering and advanced lateritization. This
study demonstrates that Ti, Zr and Th are mobile at a variety of scales, an important
consideration that is often overlooked when calculating element mass flux in intensely
Occurrence and behaviour and rare earth and associated elements in lateritic regolith profiles in Western Australia
82
5.5 Results
5.5.1 Bulk Ti, Zr and Th concentrations in regolith
5.5.1.1 Abundance of Ti, Zr and Th in the parent rock and regolith
All regolith except saprolite had a higher Ti concentration than the parent
meta-granitoid (0.13 wt%; Figure 5.1a). Furthermore, in the A horizon (0-0.4 m depth)
Ti was more abundant in gravel (0.67-0.89 wt%) than in matrix (0.65-0.74 wt%),
whereas in the ferruginous zone (1.5-5 m depth) Ti was more concentrated in matrix
than gravel, excepting the upper ferruginous zone (JG6, 1.5 m depth). The concentration
of Zr varied significantly in the profile (Figure 5.1b), 105 ppm in the saprolite (10 m
depth), 164-341 ppm in the mottled clay (6.5-8.6 m depth), 291-482 ppm in the
ferruginous zone and 346-506 ppm in the A horizon. With the exception of the saprolite
(105 ppm), the regolith samples were enriched in Zr compared with the parent
meta-granitoid (159 ppm). The concentrations of Zr in the A horizon were higher than
in the regolith below, and the concentrations in matrix were higher than in gravel. Both
Ti and Zr concentrations in matrix generally increased upwards, but Th was extremely
enriched in the lower mottled clay matrix (167 ppm) at 8.6 m depth (Figure 5.1c). This
extreme enrichment of Th is not thought to be an analytical error, as analyses of the
particle size fractions from the lower mottled clay showed similar enrichments in all
fractions (sand, silt and clay). Relative to the average concentration in the parent
meta-granitoid (17 ppm), Th was significantly enriched in ferruginous gravel (up to
196 ppm in duricrust) in contrast to weak enrichment in ferruginous matrix (average
26 ppm), implying strong partitioning of Th into gravel during weathering and
lateritization.
5.5.1.2 Variation of ratios of Ti, Zr and Th with depth
During intense weathering and lateritization processes it is difficult to define an
‘immobile’ element, and thus it is instructive to examine element ratios for potential
immobile elements. The (Ti/Zr)/10 value (Figure 5.2a) varies little from saprolite (1.0)
to ferruginous mottled zone (1.1 in matrix and 1.3 in gravel) and remains close to that of
Chapter Five: Redistribution and mobilization of Ti, Zr and Th in an intensely weathered lateritic profile in WA
83
the meta-parent granitoid (ca. 0.8). In the duricrust and upper ferruginous zone,
however, values in both gravel (1.0-2.8) and matrix (0.9-1.5) are more variable,
suggesting Ti and Zr fractionate from each other and are partitioned between gravel and
matrix at advanced stages of lateritization. This fractionation is most apparent in the
upper ferruginous zone. Compared with (Ti/Zr)/10, (Zr/Hf)/10 remains largely constant
and is within the range of parent meta-granitoids (2.9-3.9) throughout the profile
(Figure 5.2b). In contrast, (Ti/Th)/100 is more variable from the saprolite to horizon A
regolith, especially in gravel from the upper part of the profile (Figure 5.2c). The
constant ratios of Ti/Zr in the lower part of the profile and Zr/Hf throughout the profile
suggest that either these elements have undergone a relative mass flux change at a
similar rate, or they may remain residual during weathering at the investigated scale. It
is difficult to envisage similar rates of element mass flux under persistent intense
supergene weathering, and thus it is more likely that Ti, Zr and Hf are effectively
conservative during initial and moderate weathering. Therefore, as Ti/Zr and Zr/Hf
appear less affected by external processes than Ti/Th, Ti/Zr and Zr/Hf may be more
suitable discrimination ratios for moderate weathering. The fractionation between Ti
and Zr in the ferruginous zone suggests that Ti and Zr partition between gravel and
matrix that subject to extreme weathering and strong lateritization. Given the constant
ratio of Zr/Hf consistent with the parent meta-granitoid and the relatively higher
concentrations than Hf, resulting in robust estimates of mass balance, Zr is used as the
reference element. Similar mass balance calculations have been reported in previous
studies in the lateritic bauxitic profiles in Jarrahdale (Brimhall et al., 1992; Brimhall et
al., 1994).
Occurrence and behaviour and rare earth and associated elements in lateritic regolith profiles in Western Australia
84
Figure 5.1 Variation of Ti, Zr and Th with depth in the JG profile (gravel, ‘g’ is
represented by a triangle and matrix, ‘m’ by a circle in this figure and Figure 5.2). Note
that Th is strongly partitioned into gravel and the increase in Th concentration in the
lower mottled clay matrix does not correlate with any similar spike (positive or negative)
in the Zr or Ti.
Figure 5.2 Variation of Ti/Zr, Zr/Hf and Ti/Th with depth in the JG profile. In order to
aid comparison the ratios have been divided by either 10 or 100 as indicated in the
graphs. Note that Zr/Hf remains within the range defined by the parental meta-granitoid
(bounded by dashed lines; the Ti/Zr range is so small it is represented by one line at this
scale).
Chapter Five: Redistribution and mobilization of Ti, Zr and Th in an intensely weathered lateritic profile in WA
85
5.5.2 Mass balance of Ti and Th
Using Zr as the reference element, Ti and Th were enriched throughout the regolith
profile (Figure 5.3). Values of τ(Zr,Ti) slightly increased from 0.24 in the saprolite to 0.56
in the ferruginous duricrust, and then sharply increased to 2.1 in the upper ferruginous
zone, and averaged 1.3 in the A horizon. Compared with Ti, τ(Zr,Th) showed that Th was
significantly enriched in the duricrust and extremely enriched in the lower mottled clay.
In the ferruginous duricrust (3 m depth), ε(Ti) = −0.5, ε(Zr) = −0.2 and ε(Th) = −0.7, all
suggesting regolith collapse. Collapse can be inferred from the increase in
concentrations of the immobile elements (Ti, Zr and Th) because the loss of mobile
elements is not exactly compensated by an inversely proportional decrease in bulk
density during intense weathering and lateritization (Brimhall et al., 1992).
Figure 5.3 Mass balance calculations of Ti and Th against depth in the JG profile, based
on weighted average concentrations in matrix and gravel, using Zr as the reference
element. As τ(Zr,Ti) and τ(Zr,Th) are both above 0 throughout the profile, this implies mass
flux increase of Ti and Th relative to Zr.
Occurrence and behaviour and rare earth and associated elements in lateritic regolith profiles in Western Australia
86
5.5.3 Mineralogical characteristics of Ti, Zr and Th in the JG profile
5.5.3.1 Occurrence of Ti, Zr and Th in parent meta-granitoids
Representative accessory mineral data from the parent meta-granitoid are presented in
Table 5.3. Titanium was predominantly partitioned into ilmenite (FeTiO3, 28-32 wt%,
Figure 5.4a & b) and sphene (CaTiSiO5, also known as titanite, 20-22 wt%, Figure 5.4b).
Zircon grains (ZrSiO4, Figure 5.4c) had high concentrations of Zr (ca. 45-47 wt%), a
minor amount of Th (ca. 0.1 wt%) and varied concentrations of REE (up to 0.3 wt%
total REE). Thorite (ThSiO4, Figure 5.4d) is the main host for Th (18.9-32.9 wt%), and
also contained 7.7-13.7 wt% Zr, 0.6-0.9 wt% Ti and 3.5-5.2 wt% total REE.
In addition to their main host minerals, significant concentrations of Ti, Zr and Th also
occurred in many widely disseminated accessory minerals, for example, REE-rich
fluorocarbonates (Figure 5.4e & f) contained varied concentrations of Ti
(0.02-0.09 wt%), Zr (up to 0.27 wt%) and Th (0.6-6.4 wt%); magnetite 0.03-0.05 wt%
Ti and up to 0.02 wt% Th; and allanite ca. 0.02 wt% Ti and ca. 0.08 wt% Th.
Another Zr-hosting mineral, probably zirkelite, was observed as a string some hundreds
of microns long and one micron wide associated with quartz in the parent
meta-granitoid (Figure 5.4g & h). As this mineral size range is below the spatial
resolution of the electron microprobe, the chemical data could be separated from the Si
interference originating from the quartz.
87
(a) (b) (c) (d)
(e) (f) (g) (h)
Figure 5.4 Backscatter electron images of Ti-, Zr- and Th- hosting phases in parent meta-granitoids of the JG profile: (a) ilmenite surrounded by apatite;
(b) ilmenite intergrown with sphene; (c) zoned euhedral zircon crystal; (d) thorite crystal rich in Zr (a fracture resulted from electron beam impact); (e)
and (f) REE-bearing fluorocarbonates containing Zr and Th; (g) and (h) probable zirkelite ‘string’ associated with quartz. Compositional analyses of
minerals in (b), (c), (d), (e) and (f) are listed in Table 5.3 (Ap: apatite; Fsp: feldspar; Ilm: ilmenite; Py: pyrite; Qz: quartz; Spn: sphene; Zrk: zirkelite).
Occurrence and behaviour and rare earth and associated elements in lateritic regolith profiles in Western Australia
88
Table 5.3 Element concentrations of minerals in Figure 5.4 and Figure 5.5 based on
EPMA in parent meta-granitoids and lateritic regolith in the JG profile
*Y, La, Pr, Nd, Sm, Eu, Tb, Dy, Ho, Er, Tm, Yb, Lu, Mg, Sr, Na, K, P and As were below the detection
limit of the microprobe; the analysis spot was located at the brightest areas of the Ce-mapping
corresponding with the highest concentrations of Ce.
15
20
25
30
35
3 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8
d spacing
Co
un
ts (
tho
us
an
d)
10
15
20
25
Ant
Ilm
RtJG5
JG4
Chapter Five: Redistribution and mobilization of Ti, Zr and Th in an intensely weathered lateritic profile in WA
93
(a)
(b)
Figure 5.8 Neoformed poorly crystalline Zr-hosting phases associated with Ce on pore
walls around Al/Fe matrix in the duricrust of the JG profile. This co-occurrence of Zr
and Ce is unassociated with silicates or phosphates and thus is most likely (hydr)oxides
(Table 5.4): (a) edge of a ca. 2 mm nodule cemented with clay matrix in the duricrust;
(b) quartz surrounded by and cemented with Al/Fe matrix in the pore system of
duricrust; (CP is the backscatter image).
Occurrence and behaviour and rare earth and associated elements in lateritic regolith profiles in Western Australia
94
5.5.4 Grain size distribution of Ti, Zr and Th in the lateritic regolith
In order to exclude the externally sourced Ti and Zr in the A horizon, regolith samples
from 1.5 m to 10 m depth were used to investigate the distribution of Ti, Zr and Th in
different grain size fractions. The results are shown graphically in Figure 5.10 and the
data are presented in Table 5.2.
The silt fraction had the highest concentration of Ti in all regolith samples (Figure
5.10a). In the ferruginous zone, Ti concentration in the clay increased with depth from
1.5 m to 5 m, and then decreased from the mottled clay (6.5 m depth) to the saprolite
(10 m depth).
In all regolith samples except the lower mottled clay zone (8.6 m depth), the silt fraction
had the highest concentration of Zr, whereas the clay fraction contained the lowest
concentration of Zr (Figure 5.10b). In the lower mottled clay zone, silt still had the
highest concentration of Zr and sand had the lowest.
In the ferruginous zone (1.5-5 m depth) gravel had the highest concentrations of Th
whereas the silt fraction contained the highest concentration of Th in the mottled clay
(6.5-8.6 m depth). In the saprolite (10 m depth), the concentration of Th in clay was
slightly higher than both sand and silt (Figure 5.10c).
95
(a) (b)
(c) (d)
Figure 5.9 Forms of Th persisting in regolith samples of the JG profile with/as: (a) secondary REE-bearing mineral rhabdophane; (b) secondary
REE-bearing mineral florencite; (c) substituted with minor concentration in zircon; (d) micron-size grain of thorium orthosilicate mineral (ThSiO4); the
EDS spectra plots in the column on the right correspond to the minerals circled in the SEM backscatter images on the left.
96
Figure 5.10 Grain size distribution of Zr, Ti and Th in the JG profile.
Chapter Five: Redistribution and mobilization of Ti, Zr and Th in an intensely weathered lateritic profile in WA
97
5.5.5 Partition of Ti, Zr and Th into different extraction species
Sequential extraction showed that Ti predominantly occurred in the Res, with only trace
amounts occurring in WAE and FeAm species in the upper part of the mottled clay
(Table 5.5). In contrast, Zr was more enriched in Org and FeCry than WAE and FeAm in
all regolith samples, though the Res still contained the highest concentration of Zr.
Unlike both Ti and Zr, the WAE and FeCry species contained more Th than Org and
FeAm and the highest concentration of Th was in the Res. In saprolite, a high amount of
Th was also determined in the WAE.
Table 5.5 Concentrations of Zr, Ti and Th in different sequential extraction species
Element concentrations (ppm)
Ti Zr Th
Detection limit 2.00 0.002 0.001
Method ICP-OES ICP-MS ICP-MS
Saprolite
JG1m_WAE b.d. 0.40 8.78
JG1m_Org b.d. 2.03 3.93
JG1m_FeAm b.d. 0.01 0.38
JG1m_FeCry b.d. 3.26 1.42
JG1m_Res 1592 140 9.30
Upper mottled clay
JG3m_WAE 2.00 0.26 2.67
JG3m_Org b.d. 1.32 0.56
JG3m_FeAm 3.00 0.02 0.19
JG3m_FeCry b.d. 8.90 6.69
JG3m_Res 3317 242 29.70
Duricrust
JG5m_WAE b.d. 0.29 1.14
JG5m_Org b.d. 1.72 0.65
JG5m_FeAm b.d. 0.04 0.06
JG5m_FeCry b.d. 7.30 1.62
JG5m_Res 4175 225 36.30
5.6 Discussion
5.6.1 Mode of occurrence of Zr and Th in the lateritic regolith
In the ferruginous duricrust, some Zr occurs in poorly crystalline phases associated with
Ce and Th, forming a rim or coating around Al/Fe matrix in the pore systems (Figure
5.8). This occurrence of Zr and Ce is not associated with Si as zircon (ZrSiO4) or P as
rhabdophane (LnPO4, where Ln denotes REE; Table 5.4), demonstrating that Zr was
Occurrence and behaviour and rare earth and associated elements in lateritic regolith profiles in Western Australia
98
precipitated as oxides or hydroxides (ZrO2∙nH2O or Zr(OH)4) in addition to being
present in zircon in intensely weathered regolith. The geochemical behaviour of Zr
including complexation, mobilization and precipitation, depends on pH and the
presence or absence of organic matter. At a pH of ca. 5, the Zr hydroxy-bicarbonate
(Zr(OH)4-HCO3-H2O) complex, which may be the most significant Zr complex in
natural water, is unstable and possibly decomposes to form Zr(OH)4 (Salminen, 2005;
Vos et al., 2006). When a high organic component is present Zr can also be adsorbed as
colloidal oxides or hydroxides and translocated in the profile (Duvallet et al., 1999).
The sequential extraction (Table 5.5) showed that as well as being hosted by zircon in
residue, Zr was also present in the species of FeCry (7.3 ppm) and Org (1.7 ppm) in the
matrix of ferruginous duricrust (JG5m). Thus, in this case, it is likely that released Zr
was included in neoformed crystalline (hydr)oxides and attached onto the walls of
Al/Fe-rich pores, a process that was enhanced by low pH and the presence of organic
matter.
It is accepted that the geochemical behaviour of Th is dominated by the Th4+
ion
(Langmuir and Herman, 1980), and thus it shows affinity with other tetravalent
elements such as Ce and Zr. This behaviour is seen in this study by the distribution of
Th as a trace component in secondary REE-bearing phosphates (e.g. rhabdophane and
florencite in Figure 5.9) in weathered lateritic regolith. The high concentrations of Th
(5 wt% in Table 5.4) determined in neoformed poorly crystalline phases in ferruginous
duricrust, suggest the formation of insoluble Th (hydr)oxides associating with Ce and
Zr. Trace amounts of Th, associated with the WAE (1.14 ppm) and FeCry species
(1.62 ppm), in the matrix of ferruginous duricrust (Table 5.5) suggest that Th was
affected by sorption or co-precipitation with Al- and Fe- oxides.
5.6.2 Sources of Zr in poorly crystalline phases in duricrust
In this study, the formation of poorly crystalline (hydr)oxide phases containing Zr
indicates some mobility of Zr in the supergene weathering environment. However, the
distance that Zr is mobilized from its original location is not clear, but it is reasonable to
propose that this was less than the sampling scale (centimetre scale) due to: (i) the very
low solubility of Zr(OH)4; (ii) an almost constant Zr/Hf ratio consistent with the parent
meta-granitoid; and (iii) the absence of eroded zircon grains. This then poses the
question: where did the mobilized Zr come from? Limited remobilization of Zr in
supergene environments may occur in strongly acidic and organic-rich media in podzols,
Chapter Five: Redistribution and mobilization of Ti, Zr and Th in an intensely weathered lateritic profile in WA
99
or F/Cl-rich coastal profiles (Colin et al., 1993). The susceptibility of zircon to
weathering can also be enhanced by mechanical fracturing during deformation and
damage to the crystal lattice (metamictization) from radioactive decay of incorporated U
and Th (Rubin et al., 1993). Therefore, old zircons (with high U and Th concentrations)
in paleosols are more likely to be susceptible to dissolution than undamaged zircon.
The appearance of metamict areas in zircon resembles the effects of chemical
weathering (Rubin et al., 1993) and is visible using a petrological microscope. The
majority of zircon grains in this regolith profile, however, still retained typical crystal
morphology and only two containing metamict areas were observed (Figure 5.7). Thus,
it is unlikely that the Zr contained in the neoformed (hydr)oxides was released from
zircon breakdown. In contrast, other igneous phases incorporating trace amounts of Zr,
such as thorite (up to 13 wt% Zr) and REE-bearing fluorocarbonates (ca. 0.07 wt% Zr
and 7.3 wt% F), would be more likely to break down, thus releasing Zr from the parent
meta-granitoids at the initial stages of weathering. The mobility of Zr would be further
enhanced by the F-rich solution released by the REE-rich fluorocarbonates during
breakdown. Therefore, the residence of Zr not only in zircon but also in other primary
igneous minerals, and the amount and distribution of these mineral phases in the parent
rock, are likely to be significant controls on the mobility of Zr in the lateritic regolith.
5.6.3 Partitioning of Th between gravel and matrix
Unlike Zr, Th has a more complicated mode of occurrence in the lateritic regolith.
Though most Th was contained in the resistant mineral phases (revealed by significant
concentrations of Th hosted by the Res in Table 5.5), trace to minor amounts of Th were
also detected in the WAE, Org, FeAm and FeCry species. In addition, strong partitioning
of Th into gravel in the ferruginous zone reflects the local translocation and
redistribution of Th in the profile. Similar enrichment of Th in iron nodules (gravel) was
also observed in the Nsimi lateritic profile, Cameroon (Braun et al., 2005). However,
the enrichment of Th in gravel is not consistent with the concentrations of Ti and Zr
determined in this study, as concentrations of Ti and Zr were not consistently higher in
gravel than matrix. This indicates that zircon, ilmenite, rutile and anatase were not the
only hosts for Th in gravel. In the duricrust Th also precipitated as Th (hydr)oxides
associating with Zr and Ce (Table 5.4); in the upper ferruginous zone and the A horizon
Th-hosting REE-rich phosphates distributing into iron nodules had been found (Figure
5.9). In addition, from the ferruginous mottled zone to the A horizon, pH varies from
Occurrence and behaviour and rare earth and associated elements in lateritic regolith profiles in Western Australia
100
4.4 to 5.3, concentrations of Al and Fe oxides in gravel are higher than in matrix; both
features may also contribute to the strong partitioning of Th into surface gravel. This is
in agreement with the strong sorption of Th by hematite and gibbsite within a similar
pH range (Cromieres et al., 1998; Zhang et al., 2006). In the mottled clay and saprolite
the silt size fraction had the highest concentrations of Th, which is likely to be the result
of the Th-hosting minerals being dominantly within the range of the silt size fraction
originally (e.g. secondary REE-phosphates). However, the reason for the abnormal
enrichment of Th in the lower mottled clay zone is not clear. To understand more fully
the abnormal accumulation of Th in the lower mottled clay and to evaluate whether this
is observed elsewhere, further research is needed.
5.6.4 Mobility of Ti in the JG profile
The weathering-resistant minerals ilmenite and anatase are the main hosts of Ti in the
profile (Figure 5.5); the absence of the igneous mineral, sphene (present in the parent
meta-granitoid), and the presence of fractured and eroded grains of ilmenite and rutile
are all evidence that Ti is mobile at the mineral assemblage scale. Trace amounts of Ti
were found in the WAE (2 ppm) and FeAm (3 ppm) species in the upper mottled clay
(Table 5.5) as well as with neoformed Zr-(hydr)oxide phases in ferruginous duricrust
(0.07-0.20 wt% in Table 5.4). These amounts, however, are negligible in comparison
with the concentration of Ti in the Res in the upper mottled clay (3317 ppm) and in the
duricrust (4175 ppm). A similar result, where Ti-phases were trapped within neoformed
clay minerals, was noted by Malengreau et al. (1995). In addition, enrichment of Ti
increased in the fine soil fractions (silt and clay) with increasing weathering intensity
(decreasing depth). Enrichment of Ti in the fine fraction was also found in ten
weathering and pedogenetic soil profiles developed on granitic rocks by Taboada et al.
(2006a). The relatively mostly constant Ti/Zr from saprolite to ferruginous mottled zone
suggests that Ti remains largely conservative at the sampling scale in the lower part of
the profile, although mobility of Ti at the mineral scale has been revealed. This implies
that the sphene and ilmenite in parent meta-granitoid break down and are replaced by
secondary ilmenite and rutile in-situ (or nearly so) in the lower part of lateritic regolith
during initial to moderate weathering. The secondary ilmenite and rutile are then altered
into anatase and thus constrain any further mobility of Ti. This is supported by the
intergrowth of ilmenite and Ti oxides in the ferruginous mottled zone (Figure 5.5c).
Similar mineral transformations in intensely weathered lateritic regolith, from
Chapter Five: Redistribution and mobilization of Ti, Zr and Th in an intensely weathered lateritic profile in WA
101
ilmenite/rutile to anatase, have been reported before by Anand et al. (1991; 2002).
Fluctuation of Ti/Zr (in comparison with almost constant Zr/Hf) in the ferruginous
duricrust and upper ferruginous zone indicates that Ti fractionates from Zr during
extreme weathering and advanced lateritization. High concentrations of Ti partitioned
into iron nodules in the upper ferruginous zone (Table 5.2) are consistent with
cementation of Fe oxides and formation of iron nodules in the upper ferruginous zone.
This appears to have resulted in further alteration of ilmenite and rutile into anatase,
which was cemented with Al/Fe oxides and incorporated into iron nodules. Although
zircon and rutile from aeolian input are relatively stable in the upper part of regolith
profile (Brimhall et al., 1992; Brimhall et al., 1988), redox change during lateritization,
and associated changes in pH and Eh have a profound effect on the mobilization and
translocation of Ti that is further enhanced by the involvement of organic matter. This is
supported by the experiment of Thompson et al., (2006) who reported that 10% of total
Ti in a basaltic soil was mobilized as colloids at peak dispersion (related to a change in
pH accompanying redox oscillation); furthermore, Ti and Zr were also observed to be
mobile in the uppermost meter of lateritic regolith in Cameroon, attributed to the
presence of organic colloids (Braun et al., 2005).
5.6.5 Geological parent mineralogy vs. weathering conditions
It is widely known that the factors influencing the mobility of trace elements include: (i)
initial concentration and mineralogical host in the parent rocks, (ii) the susceptibility of
these hosts to subsequent alteration, and (iii) the ability of the solution to transport the
elements released (Rubin et al., 1993). In this study, the susceptibility of primary
igneous hosts of Zr, Ti and Th to weathering (e.g. thorite for Th and Zr, sphene for Ti
and Th, REE-rich fluorocarbonates for REE, Zr and Th) fundamentally controls the
subsequent mobility of these elements, and changes their abundance at the early stages
of weathering. Breakdown of thorite, sphene and REE-rich fluorocarbonates releases Ti,
Zr and Th into solution. Once in solution, Ti, Zr and Th re-enter the solid phase by
formation of new secondary minerals (e.g. ilmenite, rutile, rhabdophane, (hydr)oxide).
Formation of these secondary minerals with very low solubility further limits
mobilization of Ti, Zr and Th. As weathering proceeds, the initial control by igneous
host minerals in the parent rocks diminishes in importance, rather the weathering
intensity and characteristics of the solutions present play an increasingly important role
in translocation of Ti, Zr and Th. Low pH (3.2-5.3 throughout the profile), extreme
Occurrence and behaviour and rare earth and associated elements in lateritic regolith profiles in Western Australia
102
weathering intensity (CIA 65%-99%) and the presence of organic matter enhance the
mobility of Ti, Zr and Th. However, the greater stability of zircon relative to ilmenite
and rutile results in alteration of ilmenite and rutile into anatase, leaving zircon as a
residual mineral. The presence of organic matter, clay minerals and Fe
oxides/oxyhydroxides enhances the mobility of Th by formation and sorption of
complexes at the profile scale.
5.7 Summary of the chapter
A geochemical and mineralogical study of the mobility and mode of occurrence of Ti,
Zr and Th in the intensely weathered lateritic profile at Jarrahdale, Western Australia,
was conducted. The mobilization and redistribution of Zr and Th at the sampling scale
was revealed by neoformed poorly crystalline Zr, Ce and Th (hydr)oxide phases
attaching onto the walls of Al/Fe-rich pores in the ferruginous duricrust. The source for
Zr and Th in these neoformed phases is proposed to be the breakdown of thorite and
REE-rich fluorocarbonates during the early stages of weathering. Distribution into
secondary REE-bearing phosphates (e.g. rhabdophane and florencite) as a trace
component in the regolith showed translocation of Th at the mineral assemblage scale,
whereas strong partitioning of Th into gravel rather than matrix reflects redistribution of
Th at the profile scale. Absence of primary sphene in the regolith and dissolution of
ilmenite and rutile in the ferruginous mottled zone suggest mineral transformation from
sphene, ilmenite and rutile to anatase at the mineral assemblage scale during intense
weathering. The limited range of Ti/Zr from saprolite to ferruginous mottled zone
indicates that Ti is mostly conservative during moderate weathering despite varying in
concentration. The fluctuation of Ti/Zr in the duricrust and upper ferruginous zone
suggests that Ti and Zr fractionate from each other and partition between gravel and
matrix during extreme weathering and advanced lateritization. Therefore, these
commonly considered immobile elements are mobile at a variety of scales, and special
attention should be paid when using these elements to calculate the flux mass,
especially under intensely weathered conditions or where there are particle size sorting
transport processes.
103
6 Distribution and fractionation of REE in intensely weathered
lateritic profiles in Western Australia
6.1 Abstract
Three intensely weathered lateritic profiles (GE, MQ I and MQ II) developed on
granitoids with dolerite dykes in Western Australia were studied to investigate
geochemical behaviour and fractionation mechanisms of rare earth elements (REE)
during intense weathering and lateritization. In three profiles, regolith developed from
the granitoid rather than the dolerite was confirmed by chondrite normalized REE
distribution patterns. Substantial depletion of REE in the regolith was observed,
especially in the GE profile. Chondrite normalized REE distribution patterns of regolith
from three profiles showed light REE (LREE)-enrichment, coupled with higher
depletions of LREE than HREE relative to the parent granitoids.
Monazite, allanite, apatite, zircon, ilmenite and sphene were important REE-hosting
mineral phases in the parent granitoids, with REE-rich fluorocarbonates restricted to the
MQ parent granitoids. Residual monazite and secondary rhabdophane were important
phosphates for retention of LREE, whereas zircon and ilmenite were significant HREE
selective hosts, in weathered MQ regolith. REE released by breakdown of
easy-weathering LREE-rich allanite and fluorocarbonate in parent granitoids at the early
stages of weathering may be partially leached away, or alternatively, be retained in the
regolith by formation of secondary phosphates, e.g. rhabdophane, and hence limited
further mobility.
In addition to being hosted by mineral phases, REE were also retained in weathered
regolith by association with clay minerals, Fe oxides/oxyhydroxides and organic matter.
Among five species of sequential extraction, the water soluble (including adsorbed and
exchangeable) species hosted up to 7.9% of total REE in the C horizon regolith of MQ I
profile; the amorphous Fe oxyhydroxide species contained 3.7% of total REE in the
duricrust of GE profile. A positive Ce anomaly (Ce*=6.1) in the duricrust of GE profile
was likely related to the redox change during formation of the duricrust. All of these
observations suggest that REE can mobilize during weathering and lateritization, to the
extent of becoming highly depleted in intensely weathered lateritic regolith. The
abundance, stability and composition of secondary LREE-rich phosphate minerals and
residual HREE-selective weathering-resistant minerals may control the fractionation of
Occurrence and behaviour and rare earth and associated elements in lateritic regolith profiles in Western Australia
3Subscript PR refers to the parent granitoids: GEPR2, average MQPR1 and MQPR2 were used as the parent granitoids in corresponding profile;
4b.d. refers to below detection limit; ‘g’ represents gravel and ‘m’ represents matrix.
5RSD refers to the range of relative standard deviations of the duplicates/triplicates analysed by ICP-MS.
12
1
Figure 6.5 SiO2-Al2O3-Fe2O3 ternary plots and associated variation of REE concentrations and ratios against the S/SAF weathering index for the GE
profile.
12
2
Figure 6.6 Mass balance calculations of REE against depth for three lateritic profiles, based on weighted average concentrations of REE in matrix and
gravel, using Zr as the reference element: (a) GE Profile; (b) MQ I Profile; (c) MQ II Profile (vertical dashed line refers to mass balance τ(Zr,REE) = 0;
Only selected REE are plotted here, as the remaining REE have similar patterns).
Chapter Six: Distribution and fractionation of REE in intensely weathered lateritic profiles in WA
123
6.5.2 Mineralogy of REE in the parent rock
The parent granitoids of the GE profile contained accessory minerals such as monazite
and allanite, which controlled the abundance and distribution of REE (Figure 6.7 and
Table 6.2). Monazite (<0.05 wt%) was usually incorporated into quartz or feldspar or
intergrown with apatite in GE parent granitoids. However, two different types of
monazite were identified by EPMA (Figure 6.7 and Table 6.2). Type 1 monazite (Figure
6.7a & b) contained an average 52 wt% ΣREE, 5.6 wt% Th and 0.3 wt% U with high
(La/Yb)PR (average 5.9, up to 19.5). Type 2 monazite (Figure 6.7c & d) was
characterized by a much higher concentration of Th (average 23 wt%) than the Type 1,
but lower concentrations of ΣREE (average 23 wt%) and (La/Yb)PR (average 1.0). Both
types of monazite had no apparent Ce anomalies (Ce* ranged from 0.9-1.1) and
moderate Eu anomalies (Eu* ranged from 0.3-0.7).
In addition to monazite, allanite (<0.03 wt%) was another important REE-rich
accessory mineral in GE parent granitoids (Figure 6.7e). The average concentration of
ΣREE in allanite was ca. 18 wt%, lower than Type 1 monazite, and with 0-0.02 wt% Th
and/or U. The ΣREE concentration was dominated by LREE with an average (La/Yb)PR
of 4.9 and without Ce anomaly, but had a moderate negative Eu anomaly (average 0.6).
Other accessory minerals such as zircon and ilmenite hosted trace to minor
concentrations of ΣREE. Zircon (<0.02 wt%) contained 0.15-3 wt% ΣREE, with HREE
Tb and Ho were below detection limit (b.d.); Na was also below the detection limit except No. 18 (0.57 wt%); Fc: REE-rich fluorocarbonate; Mnz: monazite; Aln: allanite; Ilm:
ilmenite; Spn: sphene; Ap: apatite; Zrn: zircon.
13
0
Table 6.3 Concentrations of REE and associated elements from EPMA analyses of representative minerals in parent granitoids from the MQ profile
had a preference for Yb (up to 0.06 wt%). Feldspars contained up to 0.05 wt% ΣREE.
Occurrence and behaviour and rare earth and associated elements in lateritic regolith profiles in Western Australia
162
7.5.3 Mode of occurrence of REE in lateritic regolith
7.5.3.1 Mineral phases containing REE in the lateritic regolith
Selected REE-bearing minerals in the duricrust and iron nodules were analysed and
mapped by electron microprobe, and the resulting images are presented in Figure 7.8.
The REE mainly existed in two types of mineral phases in the intensely weathered
regolith: secondary REE-bearing phosphates and primary weathering-resistant minerals;
both are discussed in more detail below.
Secondary REE-bearing phosphates (Figure 7.8 and Table 7.3) are important hosts for
REE in intensely weathered lateritic regolith, especially LREE, as these elements are
essential structural components in these minerals. These REE-bearing phosphates in
regolith are believed to be the secondary weathered products of fluorocarbonates,
allanite, and especially apatite, because no REE-bearing phosphates e.g. monazite, were
observed in the JG parent meta-granitoids. These secondary phosphates were identified
as rhabdophane and florencite based on EPMA analyses, and were the main form of
REE-bearing mineral phases observed in the regolith (Table 7.3), playing an important
role in trapping REE during weathering. Secondary rhabdophane and florencite are
predominantly LREE hosts, as ((La/Yb)PR ranged from 1.2-7.5. Conversely, xenotime is
significant for retaining HREE (Bea, 1996) and was observed as micron-size crystals in
the duricrust; however, this size range is below the spatial resolution of the electron
microprobe, and thus the compositional results cannot be separated from the
interference of Al, Si and Fe in the nearby clay, quartz and Fe oxides. Most of the
REE-bearing secondary phosphates are in the size range 1-10 µm and are distributed in:
(i) in the clay layers rather than the iron cores of iron nodules in the ferruginous zones
(Figure 7.8, Figure 7.9 & Figure 7.10); and (ii) in the clay matrix of iron nodules from
the A horizon (Figure 7.8). Note that iron nodules from the A horizon are different from
the iron nodules in the ferruginous zone, as they are non-concentric, having a kaolinitic
matrix cemented with Fe oxides without layering (Appendix 11.4).
These secondary rhabdophane minerals usually contained a 103-fold enrichment of
ΣREE over the average JG parent meta-granitoids. Rhabdophane lacked an apparent Ce
Chapter Seven: Mode of occurrence of REE in an intensely weathered lateritic profile in WA
163
anomaly, but had variable Eu anomalies (Eu* ranged from 0.4-1.3), reflecting the strong
mineralogical control on the redistribution of REE in the intensely weathered lateritic
regolith, similar to Braun’s study (1993). High Th concentrations in the secondary
rhabdophane and florencite in JG regolith were also determined, consistent with high Th
concentrations in LREE-rich fluorocarbonates in meta-granitoids.
Weathering-resistant minerals (Table 7.4) such as zircon, ilmenite, rutile and anatase,
present in trace to minor concentrations in lateritic regolith, are also important hosts for
REE. The REE, especially HREE, included in these weathering-resistant minerals are
not commonly expected to be extensively mobilized during pedogenesis, although
erosion and dissolution may occur under very intense weathering (Taunton et al.,
2000a). In weathered regolith, and especially in extremely weathered lateritic regolith,
zircon is the most important of these minerals to host significant concentrations of REE.
Zircon in regolith contained up to 3.4 wt% ΣREE, with a preference for HREE or Ce
(0.05 wt%). Zircon also contained varied concentrations of Th (0-0.56 wt%) and U
(0.04-0.18 wt%). In addition, ilmenite was an important HREE-selective mineral in the
ferruginous regolith, especially for Dy and Yb, containing up to 2 wt% Dy and/or
0.08 wt% Yb whereas other REE were below detection limits. Ilmenite and rutile
concentrated Yb, containing up to 0.03 wt%. Thorite (ThSiO4), though very rare in the
lateritic regolith (only one grain was observed), contained 0.64 wt% ΣREE with a
preference for HREE.
16
4
(a) (b) (c) (d)
(e) (f) (g) (h)
Figure 7.6 Backscatter electron images of REE-bearing accessory minerals in parent meta-granitoids of the JG profile: (a) micron-size fluorocarbonate
intergrown with 100 µm-size fluorocarbonate; (b) and (c) REE-rich fluorocarbonates; (d) thorite rich in REE and Zr; (e) sphene intergrown with
ilmenite surrounded by feldspars; (f) magnetite surrounded by quartz; (g) apatite intergrown with a tiny crystal REE-bearing fluorocarbonate; (h)
apatite intergrown with ilmenite; (Qz-quartz; Ap-apatite; Ilm-ilmenite; Spn-sphene; Fsp-feldspar; Mag-magnetite).
Chapter Seven: Mode of occurrence of REE in an intensely weathered lateritic profile in WA
165
Figure 7.7 REE distribution patterns of fluorocarbonate and thorite normalized by the
parent meta-granitoids in the JG profile (Fc: fluorocarbonate; Thr: thorite).
0.0
0.2
0.4
0.6
0.8
La Ce Pr Nd Sm Eu Gd Yb
RE
E/p
are
nt
me
ta-g
ran
ito
id
Fc
Fc
Fc
Fc
Thr
Occurrence and behaviour and rare earth and associated elements in lateritic regolith profiles in Western Australia
166
(a) (b)
(c) (d)
(e) (f)
Figure 7.8 Backscatter electron images of REE-bearing secondary phosphate minerals
in regolith samples of the JG profile: (a) and (b) secondary rhabdophane in the clay
layer of iron nodules in the ferruginous zone at 1.5m depth; (c) and (d) secondary
rhabdophane in clay matrix of iron nodules in A horizon at 0.4m depth; (e) and (f) are
secondary florencite locating in the clay layer of iron nodules in ferruginous zone at
1.5m depth; (Zrn-zircon; dark circles are secondary REE-bearing phosphates).
Chapter Seven: Mode of occurrence of REE in an intensely weathered lateritic profile in WA
167
(a)
(i
(ii
Occurrence and behaviour and rare earth and associated elements in lateritic regolith profiles in Western Australia
168
(b)
(i
(ii
Figure 7.9 Images of REE-bearing secondary phosphates located in the clay layer of
iron nodules at 1.5m depth in the JG profile; element compositions of the phosphate are
listed in the Table 7.3: (a) labelled as the No. 63 rhabdophane; (b) labelled as the No. 64
rhabdophane; (i is backscatter image: the rectangular box indicates the area mapped by
EPMA, and the bright spot in the rectangular box is the fine-grained (<10 µm)
secondary rhabdophane; and (ii is the corresponding microprobe mapping: CP refers to
backscatter scan image, and SL refers to secondary scan image.
16
9
(a) (b) (c)
Figure 7.10 Mapping of secondary rhabdophane in iron nodule at 1.5 m depth of the JG profile. (a) backscatter image; (b) elemental mapping of the
rectangular black box in (a) by EPMA; (c) RGB post-imaging of the rectangular white box in (a), collected by SXFM with Maia 384/96 detector, using
Geopixe software with Ni foils as the in-house reference standards; mapping area 3.6×0.8 mm, scan duration 30 min, Maia run number 18745; (the
blue circles in (a) are the bright spots of Ti minerals in (b); the green circles in (a) are the green spots of Ti minerals in (c); red circle is the micron-size
secondary rhabdophane; element compositions are listed in the Table 7.3 and labelled as the No. 65 rhabdophane).
Occurrence and behaviour and rare earth and associated elements in lateritic regolith profiles in Western Australia
170
7.5.3.2 Mode of occurrence of Ce in the duricrust and iron nodules
Cerium was fractionated from the other REE and significantly enriched in iron nodules
(up to 200 ppm). Element mapping of the duricrust and iron nodules shows that, in
addition to being hosted by secondary phosphate and weathering-resistant minerals, Ce
also precipitated as poorly crystalline phases: (i) as a rim along the Fe-rich pores in the
duricrust (Figure 7.11); (ii) as a rim along the boundary between clay layers (Al-rich)
and iron layers (Fe-rich) in iron nodules (Figure 7.12); and (iii) as joint matrix between
two iron cores within one large nodule (Figure 7.13). Quantitative microprobe analysis
revealed that the concentration of Ce in the rims was up to 1.5 wt% and varied with
location, whereas most of other REE except Gd, Sm and Nd were below detection
limit (Table 7.5). However, the size range of the rims (width<1 µm) is below the spatial
resolution of the microprobe, and therefore the compositional results cannot be
separated from the interference of Al and Fe in the matrix. Despite this problem,
quantitative analysis by microprobe of selected several areas revealed that Ce did not
always exist with P or Si, and the sum of oxides was <100, likely reflecting a hydrous
Ce (hydr)oxide. In addition, a correlation between Th, Zr and Ce was also identified in
both element mapping and chemical analyses (Chapter Five).
In addition to amorphous Fe, crystalline Fe was also observed to be important for
retention of REE, including Ce. A crystalline Fe oxide in the duricrust contained
(sub)micron-size Ce-rich phosphates (Figure 7.14). This may suggest that crystalline Fe
can partially control Ce redistribution by secondary mineral intergrowth, whereas
amorphous Fe controls Ce occurrence by sorption and coprecipitation of Ce
(hydr)oxide.
Chapter Seven: Mode of occurrence of REE in an intensely weathered lateritic profile in WA
171
(a)
(b)
Figure 7.11 Cerium fractionated from other REE and occurring as a rim along the
Al/Fe-rich pores in the duricrust; (a) is backscatter image with rectangular box
indicating the area mapped by EPMA; and (b) presents the corresponding EPMA
mapping; element concentrations of spot analysis of Ce-rich rim are listed in Table 7.5.
Occurrence and behaviour and rare earth and associated elements in lateritic regolith profiles in Western Australia
172
(a)
(b)
Figure 7.12 Cerium fractionated from other REE and occurring as a rim along the
boundary between clay and iron layers in iron nodules; (a) is backscatter image with
rectangular box indicating the area mapped by EPMA; and (b) presents the
corresponding EPMA mapping; element concentrations of spot analysis of Ce-rich rim
are listed in Table 7.5; the slightly paler rectangular zone in (a) resulted from previous
beam scans by the electron microprobe.
Chapter Seven: Mode of occurrence of REE in an intensely weathered lateritic profile in WA
173
(a)
(b)
Figure 7.13 Cerium fractionated from other REE and occurring as joint matrix between
two iron cores within one large nodule; (a) is backscatter image with rectangular box
indicating the area mapped by EPMA; and (b) presents the corresponding EPMA
mapping; element concentrations of spot analysis of Ce-rich rim are listed in Table 7.5;
the slightly paler rectangular zone in (a) resulted from previous beam scans by the
electron microprobe.
17
4
(a) (b) (c)
Figure 7.14 Backscatter electron images of crystalline Fe oxides intergrown with micron-size Ce-rich secondary phosphates in the duricrust: (a)
crystalline Fe oxides; (b) and (c) Ce-rich secondary phosphates.
Table 7.5 Element concentrations in Figure 7.11, Figure 7.12 & Figure 7.13 of the duricrust and iron nodules in the JG profile
*Apart from Ce, Gd and Yb, other REE and Y were below the microprobe detection limit, so are not listed above; the spots for analysis were located at the brightest areas of the
Ce-mapping images.
Chapter Seven: Mode of occurrence of REE in an intensely weathered lateritic profile in WA
175
7.5.3.3 Determination of Yb onto iron oxide core
In order to understand more fully the effect of Fe oxides on the distribution and
fractionation of REE, 16 random spots in the iron core and clay layer of iron nodules
from the A horizon (0.4 m depth) and the upper ferruginous zone (1.5 m depth) were
analysed by EPMA. Voids in the iron nodules were avoided. Trace concentrations of Yb
(0.02-0.12 wt%) were determined in seven spots, including six spots in the iron core
and one spot in the clay layers (Table 7.6). Only one spot, in the clay layer, contained
trace concentrations of Pr.
Table 7.6 REE concentrations of random spots in iron core and clay layer in iron
nodules from the A horizon and upper ferruginous zone of the JG profile
El.
Concentration (wt%)
iron
core
iron
core
iron
core
iron
core
iron
core
iron
core
clay
layer
Si 0.27 0.26 0.60 0.24 0.26 0.54 0.83
Zr 0.03 b.d. 0.03 0.03 b.d. b.d. b.d.
Ti 0.27 0.08 0.27 0.43 0.83 0.04 0.44
Pb 0.03 0.06 0.04 b.d. 0.03 0.04 b.d.
Th 0.01 b.d. 0.02 0.02 0.08 0.03 b.d.
U 0.03 0.02 0.03 0.03 0.03 0.04 b.d.
Al 3.22 2.99 6.87 15.3 8.48 4.32 35.7
Pr b.d. b.d. b.d. b.d. b.d. b.d. 0.02
Nd b.d. b.d. b.d. b.d. b.d. b.d. b.d.
Yb 0.11 0.04 0.07 0.04 0.02 0.12 0.05
Lu b.d. b.d. b.d. b.d. b.d. b.d. b.d.
Fe 63.0 63.9 58.1 49.8 54.4 57.8 2.42
Ca 0.01 b.d. 0.02 0.01 b.d. 0.02 0.03
Sr b.d. b.d. b.d. b.d. b.d. b.d. 0.01
Na b.d. b.d. b.d. b.d. 0.02 b.d. b.d.
K b.d. b.d. 0.02 0.01 b.d. 0.01 b.d.
P 0.01 0.02 0.03 0.01 0.02 0.01 b.d.
S 0.05 0.02 0.07 0.05 0.14 0.10 0.03
As 0.07 0.06 0.06 0.02 0.04 0.06 b.d.
F b.d. b.d. 0.13 b.d. b.d. 0.15 0.11
O 21.6 21.4 23.8 28.6 24.3 21.2 33.7
total 88.7 88.8 90.2 94.6 88.6 84.57 73.36
Occurrence and behaviour and rare earth and associated elements in lateritic regolith profiles in Western Australia
176
7.6 Discussion
7.6.1 Geochemical pathways and fractionation of REE
Based on bulk geochemistry and the mode of occurrence of REE, abundant easily
weathered accessory minerals in the parent meta-granitoids e.g. fluorocarbonates,
thorite, apatite, are thought to have broken down during the early stages of weathering,
greatly changing the abundance and distribution pattern of REE. In the saprolite (10 m
depth) ca. 94% ΣREE released by dissolution of the accessory minerals has been
leached away or transported via solutions, and only ca. 6% ΣREE is retained; of that 6%,
ca. 5.1% REE is retained in the saprolite by mineral phases (e.g. secondary phosphates
or weathering-resistant minerals), and the remaining ca. 0.9% ΣREE has been retained
by association with other phases e.g. clay minerals, organic matter and Fe
oxides/oxyhydroxides (revealed by sequential extraction in Chapter Eight). Strong
depletion of ΣREE was also shown in the mottled clay zones (6.5-8.6 m depth,
68%-83%) and the A horizon (32%-53%), except for the duricrust (3 m) and the
ferruginous mottled zone (6.5 m depth) because of their anomalous enrichment of Ce
(Table 7.1). Therefore, significant amounts of REE, excluding Ce, have been leached
out of the profile rather than being translocated at the profile scale, suggesting high
mobility of REE under advanced weathering and strong lateritization. This is in contrast
to commonly reported accumulation of REE at the base of lateritic profiles (Beyala et
al., 2009; Braun et al., 1993; Dequincey et al., 2006; Nesbitt, 1979). The acidic
condition found in the weathered matrix of this study (a range of pH 3.2-5.3 from the
saprolite to the ferruginous zone) is lower than or close to the pH of natural rainfall (ca.
4.5-5.6, Charlson and Rodhe, 1982), which may have enhanced strong leaching of REE
during weathering.
Formation of secondary phosphate minerals e.g. rhabdophane and florencite, constrains
further mobility of REE (Braun et al., 1993), especially LREE (because of their LREE
selectivity, with average (La/Yb)PR = 3.2), and play an important role in redistribution
and fractionation of REE. Though Tripathi and Rajamani (2007) proposed that
secondary minerals are not particularly known to produce strong REE fractionation, a
strong preference for LREE in secondary rhabdophane and florencite was observed in
the lateritic JG regolith (Table 7.3). Unlike the LREE-hosting minerals, most
HREE-selective minerals are weathering-resistant and have undergone residual
accumulation in the weathered regolith. Thus, the low values of (La/Yb)PR in the saprolite
Chapter Seven: Mode of occurrence of REE in an intensely weathered lateritic profile in WA
177
(0.2) and the lower mottled clay (0.4) suggests stronger depletion of LREE than HREE
relative to the parent meta-granitoids. As weathering intensifies, however, some
HREE-rich minerals, e.g. ilmenite, may be partially dissolved. The dissolution and even
removal of ilmenite (Chapter Five) under intense-extreme weathering and strong
lateritization may change the fractionation of REE and increase (La/Yb)PR. This might
be the reason for (La/Yb)PR in the upper mottled clay being higher than the lower
mottled clay (1.2 and 0.4 respectively), as the weathering of upper mottled clay
(CIA=94%) is more intense than lower mottled clay (CIA=86%). Further support for
the partial breakdown of ilmenite changing the fractionation of REE was provided by
the sequential extraction experiments (Chapter Eight): in the saprolite, where ca. 92%
ΣHREE was hosted by mineral phases (Res species), and this value has decreased to
ca. 88% in the upper mottled clay and ca. 82% in the duricrust. This reflects the mineral
control on the translocation of REE and the important effects of weathering on
redistribution and fractionation of REE. Therefore, the abundance and fractionation of
REE in regolith are essentially weighted mean of the abundances and compositions of
LREE-rich secondary phosphates and HREE-rich weathering-resistant minerals, which
are predominantly controlled by the weathering conditions (including weathering
intensity, weathering time, accessibility to solution and pH).
7.6.2 Enrichment mechanism of Ce in ferruginous zone
In the JG profile, significant Ce anomalies were observed in the ferruginous zone (Ce*
ranged from 1.5-25.3), especially in the duricrust (Ce*=25.3 in gravel) relative to the
parent meta-granitoids (Table 7.1); except Ce, the other REE are commonly depleted
(Figure 7.5). This enrichment of Ce is consistent with the total Fe enrichment and the
occurrence of neoformed Ce-(hydr)oxide phases rimming along Fe-rich pores in the
duricrust. The co-accumulation of both Fe (III) and Ce (IV) probably reflects a redox
boundary and existence of oxidising conditions. Similar situations have been reported
previously (Angelica and Dacosta, 1993; Braun et al., 1990; Braun et al., 1998).
During lateritic weathering, goethite, hematite, and maghemite form in the ferruginous
zone. Repetitive dissolution-precipitation of Fe oxyhydroxides produces the duricrust
and concentric iron nodules by seasonal growth during alternative wet-dry periods
(Chapter Four). Under oxidizing conditions, Ce is likely to be fractionated from the
other REE and precipitated as Ce (IV) and attaching onto Fe oxyhydroxides. The net
result of these processes will be positive Ce anomalies in the ferruginous zone
Occurrence and behaviour and rare earth and associated elements in lateritic regolith profiles in Western Australia
178
(including the duricrust).
No apparent Ce anomalies were observed in secondary phosphates in this study (Ce*
ranged from 0.95-1.02), suggesting that the replacement of accessory REE-bearing
minerals by secondary phosphates is not related to a redox gradient. Dissolution of
accessory minerals released REE at the early stages of weathering; some REE
(including Ce) precipitate as secondary phosphates, or complexed by different phases
e.g. clay minerals and retained in regolith, whereas some REE may dissolve in solutions.
Once in solution, pH, Eh and ligand concentrations are important controls on the
solubility of Ce. Fluctuation of water tables can induce redox change; consequently, Ce
may be oxidised to form (hydr)oxides or precipitated with ferric minerals in the
duricrust. Similar situations in lateritic regolith have been reported by Braun et al.,
(1990). Alternatively, repetitive dissolution and precipitation of Fe oxyhydroxides may
shift pH and Eh conditions; in consequence, Ce may fractionate from other REE by
surface precipitation with Fe oxyhydroxides during seasonal growth of iron nodules.
This is supported by the occurrence of Ce as a rim coating along the boundary between
Al-rich and Fe-rich layers in iron nodules.
7.6.3 Effects of Fe oxides/oxyhydroxides on mode of occurrence of REE
Although Fe oxides are known to be efficient sinks for heavy metals due to their large
surface areas (Nedel et al., 2010; Singh and Gilkes, 1992), 58%-82% ΣREE excluding
Ce are depleted in the ferruginous zone of the JG profile, similar to the mottled clay
(67%-83% ΣREE depletion excluding Ce) and the saprolite (93% ΣREE depletion
excluding Ce). This may be the result of persistent intense acidic leaching of
REE-bearing minerals under alternative wet-dry periods.
The mass proportion of amorphous Fe oxyhydroxides was 2.0%, lower than the mass
proportion of crystalline Fe oxides (4.7%), however, amorphous Fe oxyhydroxides
contained a higher mass proportion of ΣREE (4.5%, excluding Ce) than the crystalline
Fe oxides (mass proportion of ΣREE 1.0%, excluding Ce) in the duricrust matrix
(Chapter Eight). This suggests that amorphous Fe oxyhydroxide is more efficient at
scavenging REE than crystalline Fe oxide, a finding that is supported by previous
studies (Compton et al., 2003; Land et al., 1999; Laveuf and Cornu, 2009). In the JG
duricrust (Figure 8.4), amorphous Fe oxyhydroxide has higher (La/Sm)PR (0.16) and
(La/Yb)PR (0.25) than the crystalline Fe oxide (La/Sm)PR (0.37) and (La/Yb)PR (0.49),
Chapter Seven: Mode of occurrence of REE in an intensely weathered lateritic profile in WA
179
reflecting the tendency for amorphous Fe oxyhydroxides to be more selective for Sm
(MREE) and Yb (HREE). The (La/Yb)PR < 1.0 in the amorphous and crystalline Fe
extractions is in consistent with trace concentrations of Yb (0.02-0.12 wt%) determined
in the iron cores and the clay layers (Table 7.6). This finding is supported by Marmier et
al.’s experiments (1997; 1999) showing that surface complexation of Yb occurred on
hematite and magnetite between pH5 and 7.
Although REE showed different degrees of association with Fe oxides and
oxyhydroxides in this study, the mechanism of these associations is proposed to be
different:
(i) Poorly crystalline Ce (hydr)oxide phases as a rim along Fe-rich pores in the duricrust
reflect the oxidation and surface precipitation of Ce with Fe oxyhydroxides when redox
changes.
(ii) The association of REE with extracted Fe species and the determination of trace
concentrations of Yb (0.02-0.12 wt%) in the iron core and the clay layers are likely to
be the result of surface complexation, substitution and/or co-precipitation.
(iii) Minor amounts of REE in the extracted crystalline Fe oxides are likely the result of
surface precipitation of secondary REE-bearing phosphates with Fe oxides during
duricrust formation. This is supported by the observation of micron-size REE-bearing
phosphate crystals occurring within the crystalline Fe oxides in the duricrust (Figure
7.14).
(iv) Alternatively, the occurrence of REE-bearing phosphates in the clay layer of the
iron nodules might result from iron nodules sequestering secondary REE-bearing
phosphates during their seasonal growth and cementation at advanced stages of
lateritization.
Therefore, Fe oxides and oxyhydroxides can play important roles in redistribution and
fractionation of REE during intense weathering and lateritization.
7.7 Summary of the chapter
A lateritic profile (JG) locating in Jarrahdale, Western Australia was investigated for the
mode of occurrence and geochemical behaviour of REE under intense weathering and
advanced lateritization.
In the saprolite, ca. 94% ΣREE was released by dissolution of accessory minerals in the
Occurrence and behaviour and rare earth and associated elements in lateritic regolith profiles in Western Australia
180
parent meta-granitoids e.g. fluorocarbonates, thorite and apatite; ca. 5.1% ΣREE was
retained by formation of secondary phosphates and residual accumulation of
weathering-resistant minerals, and ca. 0.9% ΣREE was retained by association with clay
minerals, organic matter and Fe oxides/oxyhydroxides. The formation of secondary
phosphate minerals e.g. rhabdophane and florencite, which are absent in the parent
meta-granitoids, constrains further mobility of REE, especially LREE. The residual
accumulation of weathering-resistant minerals e.g. zircon and rutile/anatase are
important hosts for retention of HREE, especially in extremely weathered ferruginous
zones. Thus, the abundance and stability of LREE-rich secondary phosphates and
HREE-rich weathering-resistant minerals control the fractionation of REE in intensely
weathered lateritic regolith.
In the ferruginous zone, Ce fractionated from the other REE and was abnormally
enriched (Ce*=25.3 in the duricrust gravel). This Ce enrichment is in agreement with
the occurrence of poorly crystalline Ce (hydr)oxide phases as rims along Al/Fe-rich
pores in the duricrust, or along the boundary of Al/Fe-rich layers in iron nodules of the
ferruginous zone. The fractionation of Ce is the result of surface precipitation of Ce (IV)
phases with Fe oxyhydroxides under oxidization and/or changes of pH and Eh
conditions during advanced stages of lateritization.
Trace concentrations of Yb (0.02-0.12 wt%) determined in the iron cores and clay
layers are consistent with the association between REE and amorphous and crystalline
Fe extracted species. They suggest that Fe phases are effective for retention of REE in
lateritic regolith. Fine-grained (<10 µm) REE-bearing phosphates occurred in the clay
layer of iron nodules, most likely to be the result of sequestering by the iron nodules
during their formation.
Therefore, the significant mineralogical control and high mobility of REE during
intense lateritic weathering are important considerations when using REE as tracers of
geochemical processes in intensely weathered environments. The sensitivity of REE to
weathering conditions, especially Ce to redox change, suggests a potential for REE to
be used as complementary geochemical clues along with Fe to investigate the
lateritization processes and to understand the role of Fe oxides and oxyhydroxides on
scavenging trace metals.
181
8 Particle size fractionation and chemical speciation of REE in a
lateritic profile in Western Australia
8.1 Abstract
The rare earth elements (REE) are commonly used as indicators of geochemical and
pedological processes. To better understand the distribution and partitioning of REE in
different particle size fractions and chemical species, an intensely weathered lateritic
profile developed on meta-granitoids in Jarrahdale, Western Australia was investigated.
High concentrations of REE were found in silt and clay fractions. Given the variation in
mass percentages of different particle size fractions, however, gravel and sand contained
56%-98% of the mass of REE in the ferruginous zone. In the saprolite and mottled clay,
clay had the highest mass loading of light REE (LREE) in contrast to the highest mass
loadings of heavy REE (HREE) found in sand. In the ferruginous zone, gravel was the
predominant host for Ce, whereas most of other REE were contained in the gravel and
sand fractions, suggesting that Ce fractionated from other REE and precipitated with, or
was adsorbed by, Fe oxides/oxyhydroxides during formation of duricrust and iron
nodules. The residual species contained the highest percentages of total REE revealed
by sequential extraction, indicating that the abundance and distribution of REE are
controlled by weathering-resistant minerals in intensely weathered regolith. Water
soluble (including adsorbed and exchangeable) species was the fraction hosting the
second highest percentages of total REE, suggesting the important effect of adsorption
by clay and potential bio-availability. The low pH of the profile is believed to account
for the high proportion of REE in this species. The amorphous Fe oxyhydroxide and
crystalline oxide extractions preferentially hosted LREE and MREE over HREE,
whereas the organic matter species was important in complexing HREE. The
distribution and fractionation of REE in different particle size fractions and chemically
extractable species can be used to better understand geochemical behaviour of REE in
(iv) crystalline Fe-Mn oxide bound (FeCry); and (v) residual species (Res). Since
carbonates were unlikely to be present in the regolith being studied here due to low pH,
species WAE is considered to include mainly water soluble, adsorbed or exchangeable
elements. Sulfides are also scarce in the lateritic regolith, therefore it is assumed that
species Org is mainly hosted by organic matter complexes. A brief summary of the
method is shown in Table 5.1 in Chapter Five. The residual samples and reference
materials were rinsed with MilliQ water three times, oven dried at 105 °C overnight and
ground to ≤200 µm prior to fusion in order to determine trace element concentrations.
The fusion beads were made by mixing 0.1 g (to an accuracy of 0.1 mg) of finely
ground sample or reference material with 0.7 g 12:22 Norrish flux (lithium
metaborate:lithium tetraborate) and heating in a muffle furnace at 1050 °C for 40
Occurrence and behaviour and rare earth and associated elements in lateritic regolith profiles in Western Australia
184
minutes. Duplicate fusion beads were also made on 10% of samples to check
preparation errors. After cooling, the fusion beads were dissolved in 100 mL of 10%
analytical grade HCl. The trace elements including REE were determined by
inductively coupled plasma-mass spectrometry (ICP-MS) in Genalysis Laboratory
Services of Intertek Commodities in Maddington, Western Australia. Certified
international standard materials, including stream sediment reference material STSD-2,
STSD-4 (Canada Centre for Mineral and Energy Technology, CANMET) and an
in-house standard material were prepared in the same way as the samples and analysed
together with samples to check the accuracy and precision. The variation between tested
values and expected values was within 10% of the certified values. The concentrations
of REE in different particle size fractions and chemical species are given in Table 8.1
and Table 8.2 respectively.
8.4.2 Calculation methods
8.4.2.1 Fractionation of REE
In order to study the fractionation of REE, three groups are identified (Henderson,
1984): the light REE (LREE; from La to Nd), the middle REE (MREE: from Sm to Ho)
and the heavy REE (HREE: from Gd to Lu). The normalized ratios (La/Sm)PR and
(La/Yb)PR were used for identifying fractionations between LREE-MREE and
LREE-HREE using the average composition of parent meta-granitoids as a reference.
8.4.2.2 Calculation of REE mass loading in particle size fraction
To index an element’s partitioning into different particle size fractions, a mean element
mass loading was calculated based on the element’s concentration in a selected grain
size of known mass percentage (Sutherland, 2003).
GSFloading 100 (X i GSi
X i GSii1
n
)
Where:
Xi is the concentration of REE (ppm) in an individual grain size fraction (e.g. <2 µm);
GSi is the mass percentage of an individual fraction, which has limits of 0-100%.
GSFloading is the element mass loading in a selected grain size and the summation of
GSFloading indices for each soil sample equals 100%.
In the ferruginous zone, four classes of particle sizes (clay, silt, sand and gravel) were
Chapter Eight: Particle size fractionation and chemical speciation of REE in the JG profile in WA
185
defined and three in the mottled clay zone and the saprolite (clay, silt and sand). Thus, if
the REE concentration for a given fraction is very high but it forms only a small portion
of the overall sample mass, the contribution of this fraction to the total sample REE
loading will be minimal.
8.5 Results
8.5.1 Concentrations of REE in different particle size fractions
In the lateritic JG profile, silt and clay fractions generally contained the highest
concentrations of REE, except in the saprolite (Figure 8.1). In the ferruginous zone
(from JG6 to JG4, 1.5-5 m depth), clay contained the highest concentrations of LREE
(from La to Nd), followed by the silt fraction. In the duricrust (3 m depth) and
ferruginous mottled zone (5 m depth), however, gravel was abnormally enriched in Ce.
Concentrations of LREE in matrix were slightly higher than in sand in the ferruginous
zone. In the mottled clay (6.5-8.6 m depth) and the saprolite (10 m depth), the relative
concentrations of LREE from high to low were: silt > clay > sand.
MREE (from Sm to Ho) had different distribution patterns between the particle size
fractions. From Sm to Gd, closer to LREE, the highest concentrations were in the clay
fraction in the ferruginous zone but in the silt fraction in the mottled clay zone. From Tb
to Ho, closer to HREE, silt fraction had the highest concentrations except duricrust and
upper ferruginous zone.
HREE (from Er to Lu) and Y showed mostly consistent distribution patterns. The silt
fraction contained the highest concentration of HREE throughout the profile followed
by the clay fraction in the ferruginous zone. In the saprolite and mottled clay, both clay
and sand fractions had similar HREE concentrations.
8.5.2 Mass loading of REE in different particle size fractions
Given the mass percentage of each particle size, the mass loading of selected REE in
each particle size fraction was plotted in Figure 8.2. Although silt and clay fractions had
the highest concentrations of REE, their relatively low mass percentage compared with
other fractions minimized the enrichment.
In the ferruginous zone, gravel dominated the distribution and abundance of Ce, with up
Occurrence and behaviour and rare earth and associated elements in lateritic regolith profiles in Western Australia
186
to 84% Ce in the duricrust. In the upper ferruginous zone (1.5 m depth), gravel and sand
accounted for more than 95% mass of REE, decreasing to ca. 80% in duricrust (3 m
depth) and ca. 60% in ferruginous mottled zone (5 m depth). In the duricrust, the mass
loading of each REE was higher in the clay fraction than the silt fraction. In the
ferruginous mottled zone, however, REE were fractionated: the mass loadings of LREE
and MREE were higher in the clay fraction whereas the mass loadings of HREE were
higher in the silt fraction.
From the upper mottled clay to the saprolite (JG3-JG1, 6.5-10 m depth), the regolith
does not contain gravel. The clay fraction was the most important host for LREE in
these zones (6.5-10 m depth), especially in the upper mottled clay zone (6.5 m depth)
with ca. 48%-50% LREE was in the clay fraction. Higher mass loadings of HREE
(46%-61%), however, were found to be in the sand fraction in the saprolite and mottled
clay. The mass loading of REE in the silt fraction increased with depth from upper
mottled clay to saprolite.
8.5.3 Speciation of REE from sequential extraction
The sequential extraction experiment revealed the percentages of ΣREE (the total REE
concentration) in each chemical species of representative lateritic regolith in the JG
profile (Figure 8.3). Generally, the ΣREE distribution percentage followed the order:
Res > WAE > FeAm > FeCry and Org. The Res and WAE species dominated the
distribution and abundance of REE, accounting for 89%-98% ΣLREE, 87%-97%
ΣMREE and 91%-98% ΣHREE. The saprolite Res had higher percentages of ΣMREE
(85%) and ΣHREE (92%) than upper mottled clay (75% ΣMREE and 88% ΣHREE) and
duricrust (66% ΣMREE and 82% ΣHREE) and the percentages decreased from saprolite
to duricrust. In addition, the saprolite WAE had higher percentage of ΣLREE (13%)
than upper mottled clay (9%) and duricrust (9%). The percentages of ΣMREE (12%)
and ΣHREE (7%) in the saprolite WAE were lower than in the WAE of upper mottled
clay (20% ΣMREE and 9% ΣHREE) and duricrust (21% ΣMREE and 9% ΣHREE).
The duricrust Org had higher percentage of ΣREE (1.5%) than the Org in the saprolite
(0.6%) and upper mottled clay (0.2%), especially HREE. The percentages of total REE
hosted in the FeAm phase of the duricrust (6.6%) were also higher than in the FeAm
phases of the saprolite (1.4%) and upper mottled clay (1.7%). Similarly, the FeCry phase
in the duricrust also had higher percentages of total REE (3.3%) than the total REE
Chapter Eight: Particle size fractionation and chemical speciation of REE in the JG profile in WA
187
percentage in the FeCry of saprolite (0.8%) and upper mottled clay (0.6%). In addition,
in the duricrust the percentage of REE in the FeAm phase (6.6%) was higher than the
percentages of REE in the FeCry (3.3%) and Org (1.5%) phases.
18
8
Figure 8.1 Concentrations of REE in grain size fractions in the JG profile
Figure 8.2 Mass loading of REE in grain size fractions in the JG profile (JG6-upper ferruginous zone, 1.5 m depth; JG5- duricrust, 3 m depth;
JG4-ferruginous mottled zone, 5 m depth; JG3-upper mottled clay, 6.5 m depth; JG2-lower mottled clay zone, 8.6 m depth; JG1-saprolite, 10 m depth.
Only selected REE are plotted here; other REE showed similar patterns).
18
9
Figure 8.3 Distribution of REE percentages in sequential extractions of the representative regolith of the JG profile. (Res: residual; FeCry: crystalline Fe
oxides; FeAm: amorphous Fe oxyhydroxides; Org: organic matter; WAE: water soluble, adsorbed and exchangeable. JG5- duricrust, 3 m depth;
JG3-upper mottled clay, 6.5 m depth; JG1-saprolite, 10 m depth. (Some REE concentrations were below the detection limit of ICP-MS and are not
presented here).
19
0
Table 8.1 Concentrations of REE in grain size fractions of the JG profile
Aln: allanite; Ap: apatite; Fc: REE-rich fluorocarbonate; Fsp: feldspar;Ilm: ilmenite; Mnz: monazite, I and II refers to Type 1 or Type 2; Rbp: rhabdophane; Spn: sphene; Thr:
thorite; Zrn: zircon.
24
5
Appendix 11.9 Concentrations of REE in grain size fractions of the MQ II profile