Contents lists available at ScienceDirect

Geoderma

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

Surficial weathering of kaolin regolith in a subtropical climate: Implicationsfor supergene pedogenesis and bedrock argillization

Qian Fanga,b, Hanlie Honga,c,d,⁎, Harald Furnese, Jon Choroverb, Qing Luof, Lulu Zhaoa,b,Thomas J. Algeoc,d,g,⁎⁎

a School of Earth Sciences, China University of Geosciences, Wuhan 430074, ChinabDepartment of Soil, Water and Environmental Science, University of Arizona, Tucson, AZ 85721, USAc State Key Laboratory of Geological Processes and Mineral Resources, China University of Geosciences, Wuhan 430074, Chinad State Key Laboratory of Geobiology and Environmental Geology, China University of Geosciences, Wuhan 430074, Chinae Department of Earth Science, Centre for Geobiology, University of Bergen, Allegt. 41, 5007 Bergen, Norwayf Center for Exploration of Earth Resources of Jiangxi Province, Nanchang 330000, Chinag Department of Geology, University of Cincinnati, Cincinnati, OH 45221-0013, USA

A R T I C L E I N F O

Handling Editor: M. Vepraskas

Keywords:RegolithChemical weatheringChemical gradientClay mineralKaolinite

A B S T R A C T

Regolith, or in situ weathered material overlying bedrock, develops through pedogenic processes such as clay-mineral formation (“argillization”). Research on products of argillization, such as kaolin, is commonly focusedon its economic value rather than on an integrated understanding of the pedological, mineralogic, and geo-chemical processes taking place in the regolith. Here, we analyzed three kaolin-regolith drillholes from a sub-tropical climate zone in southern China using X-ray diffraction (XRD), major and trace element analyses, H andO isotopes, differential thermal-thermogravimetric analysis (DTA-TG), and scanning electron microscopy (SEM).Many kaolinite aggregates within regolith have fan-shaped stacks, a morphology that is closely associated withtransformation of muscovite plates. Hypogene indicators such as a shallow level of granite emplacement, re-presentative structural controls on kaolinization, mineral zoning, and hypothermal minerals cannot be observed.Mineral assemblages and morphologies, elemental binary plots (e.g., Zr vs. TiO2 and P2O5 vs. SO3), as well aschemical profiles all show characteristics typical of surfical weathering. The δ18O and δ2H values of the clayfractions range from +16.8 to +18.7‰ and from −70 to −51‰, respectively, suggesting that supergeneweathering has played a key role in forming the kaolin regolith. Dominance of kaolinite over halloysite implies anear-surface, freely draining environment. Adopting the underlying weakly altered granite (saprolith) as theparent material, the kaolin regolith exhibits four elemental profile patterns as revealed by mass transfer coef-ficients: (1) depletion (e.g., Na), (2) depletion-enrichment (e.g., Al), (3) enrichment (e.g., Mn), and (4) biogenic(e.g., Ca). These profiles reflect a combination of chemical, geologic, and biologic processes that are typical ofrelatively thin, in situ regolith profiles, and that is not necessarily similar to those typically associated with deep(thick) granite weathering profiles. We propose that supergene kaolin regolith is intrinsically more similar toshallow, biologically active residual soil deposits, rather than deeply weathered granite-hosted regoliths.

1. Introduction

Regolith is defined as in situ weathered material overlying bedrock,and it is present to varying degrees across the Earth (Wilford et al.,2016). It represents the entire unconsolidated and secondarily re-cemented cover that overlies relatively intact bedrock, and it is formedby erosion, transport/deposition, and weathering of pre-existing ma-terial (Anand and Paine, 2002; Keeling et al., 2003). The regolith is host

to many of the resources that maintain our society and living standards.It is also in the upper part of the regolith that our soils are formed; theregolith is thus the host for life and agriculture (Taylor and Eggleton,2001). The distribution and nature of regolith are controlled by variousfactors, reflecting long-term interactions among the atmosphere,pedosphere, lithosphere, biosphere and hydrosphere (Brantley andWhite, 2009; Buss et al., 2017). Regolith develops through a variety ofprocesses such as argillization, i.e., the formation of clay minerals in the

https://doi.org/10.1016/j.geoderma.2018.09.020Received 15 July 2018; Received in revised form 5 September 2018; Accepted 9 September 2018

⁎ Correspondence to: H. Hong, School of Earth Sciences, China University of Geosciences, Wuhan 430074, China.⁎⁎ Correspondence to: T.J. Algeo, Department of Geology, University of Cincinnati, Cincinnati, OH 45221-0013, USAE-mail addresses: honghl8311@aliyun.com (H. Hong), thomas.algeo@uc.edu (T.J. Algeo).

Geoderma 337 (2019) 225–237

Available online 21 September 20180016-7061/ © 2018 Elsevier B.V. All rights reserved.

T

soil environment (Gilkes et al., 1973), and clay minerals commonlyprovide insights into the nature of these processes (Dill, 2017; Lopezet al., 2018).

One type of argillization process is kaolinization, which generatesthe mineral kaolin (Ekosse, 2001; Anand and Paine, 2002; Dill, 2017).Kaolin deposits can be formed in situ by hydrothermal processes duringlate-stage pluton cooling (hypogene), by (near-)surface weathering re-actions (supergene), or by a combination of these processes (Dudoignonet al., 1988; Murray, 1988; Dill et al., 1997; Galán et al., 2016). Hy-pogene processes play a prominent role in alteration of granitic rocksintruded at relatively shallow depths, and it is commonly associatedwith high acidity and high temperatures (Marfil et al., 2005). In su-pergene deposits, chemical weathering is the dominant process of for-mation of kaolin minerals (e.g., kaolinite, halloysite, and dickite) from avariety of parent rocks (e.g., granite, gneiss, and slate; Chorover andSposito, 1995; Scott and Bristow, 2002; Keeling, 2015; Hong et al.,2016). With time, regolith composition diverges progressively from thatof the parent material owing to the cumulative influences of vegetation,biota, topography, and, in particular, climate (Thanachit et al., 2006;Zhao et al., 2017). In the present study, “kaolin regolith” is defined as aresidual kaolin deposit that has undergone substantial chemicalleaching and physical weathering processes such as regolith creep/flowand erosion by water (Taylor and Eggleton, 2001).

Kaolin genesis has a direct relation to its industrial applications,with supergene kaolins generally being more concentrated and havinghigher economic value than hypogene ones (Ekosse, 2001; Njoya et al.,2006; Fernandez-Caliani et al., 2010). As a raw material for chinaware,kaolin has substantial economic value, but its economic significance isbroader in that kaolin is found in many products for daily use, includingpaper coating, ceramics, plastics, rubber, cosmetics, and pharmaceu-ticals (Zhou and Keeling, 2013; Lu et al., 2016; Pruett, 2016). Relativelysystematic work on kaolin argillization has been undertaken in Aus-tralia (Gilkes et al., 1973; Anand and Paine, 2002), USA (Schroederet al., 2004), and Europe (Galán et al., 2016). China represents one ofthe most important producers of kaolin (especially for ceramics, sani-tary ware and porcelain; Wilson, 2004). Indeed, the first use of kaolin inpottery manufacture emerged in Jingdezhen, Jiangxi Province, Chinaover 2000 years ago (Schroeder and Erickson, 2014), and the firstmining site for gaoling tu (kaolin clay) was the Gaoling area, located50 km northeast of Jingdezhen. Despite its cultural importance, onlylimited research concerning the formation of kaolin regolith particu-larly related surficial weathering process has been conducted in China.

A detailed study of kaolin deposits in China has the potential toimprove our understanding of mineral alteration and elemental mi-gration during supergene processes and kaolin accumulation world-wide. The Chongyi kaolin regolith in Jiangxi Province, southern China,is located in a subtropical climate zone with warm temperatures andhigh rainfall, contributing to its intensely weathered condition. Here,we investigate the Chongyi kaolin regolith using an integrated ap-proach based on XRD, SEM, XRF, ICP-MS, DTA-TG, and H and O iso-topic analyses. The goals of this study were to: (1) systematicallycharacterize the kaolin regolith using a variety of mineralogic andgeochemical techniques; (2) gain insights into the geochemical beha-vior of major and trace elements during supergene kaolin argillization;(3) determine the genetic relationships between kaolin regolith and thetypical residual soil deposit; and (4) propose a model for kaolin argil-lization in Earth's surface environments.

2. Materials and methods

2.1. Study location and climatic setting

The Chongyi kaolin regolith is located in Chongyi County, south-western Jiangxi Province, southern China (Fig. 1). The landscape con-sists of low mountains and hilly terrane. Within Chongyi County, ele-vations fall from northwest (~1000meters above sea level - masl) to

southeast (~500masl), with rivers thus flowing in a southeasterly di-rection. The study area is within a subtropical monsoonal climate zonethat is characterized by large seasonal variations in air temperature,wind direction, and rainfall, with>80% of precipitation falling duringthe period from May to September. Mean annual temperature (MAT) is17.6 °C, and mean annual precipitation (MAP) is 1600mm. TheChongyi kaolin regolith is located within the Nanling metallogenic belt,which was generated during the Jurassic-Cretaceous (~200–134Ma)Yanshanian magmatic event. The Yanshanian event resulted in NE–SW-trending faults and associated pluton emplacement in the study area(Zhou et al., 2006; Wang et al., 2018). The bedrock of the kaolin re-golith is a muscovite granite of early Yanshanian age (~200–170Ma),and the original argillization process may have been driven by mag-matic heat.

According to layering of typical regolith profiles by Taylor andEggleton (2001) and Anand and Paine (2002), the Chongyi regolith canbe divided into a lower saprolith layer and an upper pedolith layer(Fig. 2): (1) the saprolith contains saprorock below and saprolite above;and (2) the pedolith contains a plasmic zone, a mottled zone, and anoverlying surface layer. The regolith can be further vertically sub-divided into four units based on differences in color and lithology(Fig. 2; listed from top to bottom): the surface layer, the mottled kaolinlayer, the homogenous kaolin layer, and the saprolith.

2.2. Field sampling

Geological field work included mapping in outcrop and descriptionand sampling of drillcore and drill-cuttings. In each of three drillcores,five samples were collected: one in each regolith layer (surface layer,mottled kaolin, homogenous kaolin, and saprolith) with a secondsample taken in the homogenous kaolin layer (the two samples of thehomogeneous layer are designated “upper kaolin” and “lower kaolin”).This yielded a total of 15 samples for geochemical analyses (Fig. 2).Clay fractions (< 2 μm) were extracted from 12 samples (all but thethree saprolith samples) by conventional sedimentation and cen-trifugation methods. The< 45 μm fractions were extracted from threekaolin samples (i.e., CY 1–3, CY 2–3, and CY 3–3). The grain-sizefraction of 63–600 μm was used for analysis of heavy minerals.

2.3. X-ray diffraction (XRD)

The samples of bulk kaolin (including mottled kaolin and homo-geneous kaolin), fine fractions (< 45 μm and < 2 μm), and heavy mi-nerals were mineralogically analyzed by XRD. These samples werepretreated to remove organic matter with 30% H2O2 at 60 °C. Air-driedoriented clay mounts for the< 2 μm fractions were prepared by dis-persing the clay slurry onto glass slides. XRD patterns were recordedusing a D/max-3B X-ray diffractometer with Cu Kα radiation at 40 kV,35mA (resolution ratio= 0.02°2θ; scan rate= 4° 2θ/min; scanrange=3–65°2θ). The formamide test (Churchman et al., 1984) wascarried out for the<2 μm fractions to distinguish halloysite fromkaolinite. Bulk-rock mineral contents were determined on randomlyoriented powders using reference intensity ratio(RIR) together with theso-called 100% approach (Hillier, 2000). RIR is also known as I/Icor,which represents the ratio of intensities (areas) of specific XRD peaks(mineral: corundum=1:1). The term 100% approach means that thesum of all mineral phases identified is 100%. Five minerals (muscovite,feldspar, quartz, kaolinite, and gibbsite) that were involved in contentcalculation have RIR values (refer to corundum 113 peak) of 0.23(~10.0 Å), 2.07 (~3.20 Å), 0.89 (~4.25 Å), 0.53 (~3.57 Å), 1.91(~4.85 Å), respectively. Considering this is a semi-quantitative methodwith a calculation error of 5–10%, we used ‘—’ (not detected), ‘*’(< 15%), ‘**’ (15–40%), and ‘***’ (> 40%) to describe different mi-neral contents in order to present more reasonable data (Table S1).

Q. Fang et al. Geoderma 337 (2019) 225–237

226

2.4. X-ray fluorescence (XRF)

All collected samples were analyzed for major-, trace-, and rare-earth element (REE) concentrations. Major-element concentrationswere measured by XRF spectrometry using a Shimadzu XRF sequentialspectrometer. The analytical precision for major-element concentra-tions was< 1% and the detection limit was 0.01 wt%.

2.5. Inductively coupled plasma-mass spectrometry (ICP-MS)

Trace-element and REE concentrations were determined using an

Agilent ICP-MS, with an analytical precision of better than 4% for REEsand 4–10% for other trace elements. Prior to ICP-MS analysis, eachsample was leached for ~20min in 1-N HCl, after which the residueand leachate were separated by continuous centrifugation, and the re-sidue was rinsed with milli-Q water.

2.6. H and O isotopic analyses

The< 2 μm fractions of nine samples (i.e., the six homogenouskaolin plus three mottled kaolin samples) were analyzed for H and Oisotope compositions. For analysis of δ18O values, oxygen in the clay

Fig. 1. Location and regional geology of Chongyi kaolin. The study area is located in Jiangxi Province, southern China. The granites derived from Yanshanianmagmatic event are widely distributed in this area. The study section is hosted in albitization muscovite granite, located in Nanling metallogenic belt.

Fig. 2. Simplified kaolin profiles and the photo of the studied kaolin. The study samples are collected from three drill cores in three representative kaolin-regolithdrillholes (CY1, CY2, and CY3), and the detailed sampling positions are marked beside the kaolin profiles.

Q. Fang et al. Geoderma 337 (2019) 225–237

227

fractions was extracted through reaction with BrF5 at 600 °C in a nickelchamber for ~24 h and then converted to CO2 over a graphite furnace(Clauer et al., 2015). For H isotope analysis, the clay fractions wereheated at 1500 °C, and the released water was converted to H2 at900 °C. The converted CO2 and H2 were measured for oxygen and hy-drogen isotopes, respectively, using a Finnigan-MAT253 mass spectro-meter. Isotopic ratios were reported by δ notation as deviations in permille versus the international Vienna standard mean ocean water(VSMOW) standard, with reproducibility of± 0.5‰ for δ18O and ±1‰ for δ2H values.

2.7. Differential thermal and thermogravimetric analysis (DTA-TG)

Differential thermal-thermogravimetric analysis (DTA-TG) wasperformed using a Setaram Labsys thermal analyzer, operating under anitrogen flow of 60ml/min. Both the bulk kaolin and< 45 μm kaolinsamples were selected for DTA-TG analysis. The DTA-TG curves wereobtained using 10mg of powdered sample, heated from room tem-perature to 1400 °C at rates of 10, 20, 30, or 40 °Cmin−1. Aluminumwas used as a reference material.

2.8. Scanning electronic microscopy (SEM)

Both clay fractions and heavy minerals were selected for SEManalysis. These heavy minerals were hand-picked under binocular mi-croscope after heavy-liquid (solution of lithium heteropolytungstates inwater) and magnetic separation of 63–600-μm fractions. An EDS-equipped FEG-SEM (HITACHI-SU8010) was used to observe the mi-cromorphology of the clay minerals and heavy minerals of re-presentative samples and analyze their chemical compositions. Thesesamples were sputter-coated with thin films of platinum prior to ana-lysis. The SEM instrument was operated at a voltage of 10–15 kV, acurrent of 3–5 nA, and a 15mm working distance, with observationsperformed in secondary electron mode.

3. Results

3.1. Mineralogic characteristics

The mineralogic compositions of representative samples are shownin Fig. 3, and data for all samples (including bulk kaolin,< 45 μmfraction, and< 2 μm fraction) are given in Table S1. The mineralogiccomposition varies with particle size: the clay (< 2 μm) fractions aredominated by kaolinite with subordinate amounts of muscovite, gibb-site, and quartz, whereas bulk kaolin samples contain larger amounts ofquartz than kaolinite. Kaolinite is identified by diagnostic narrow peaksat 7.16 Å and 3.58 Å. The non-basal sharp and less intense reflections ofdoublets at 4.48/4.38 Å and 2.57/2.50 Å and triplets at 2.38/2.34/2.30 Å are linked to well-crystallized kaolinite (Fig. 3B, C; Erkoyun andKadir, 2011). Kaolinite is prevalent in the fine fractions (< 45 μmand<2 μm), in which quartz content is significantly lower.

Comparing the XRD patterns of the air-dried oriented clay mountswith the formamide-treated mounts, there is no change in intensity ofthe 7 Å/10 Å peaks after formamide saturation, suggesting that negli-gible halloysite is present in the Chongyi kaolin, even though this mi-neral is detectable by SEM (Churchman et al., 1984). Muscovite in-herited from the parent granite has concentrations of< 15% in the bulkkaolin samples (Table S1). A ~4.86 Å peak is observed in the XRDpatterns of bulk and fine-fraction samples, suggesting the presence ofgibbsite (Fig. 3).

3.2. Major- and trace-element compositions

The major-element concentrations of the analyzed samples aregiven in Table S2. The three study sections are similar in their overallchemical compositions. The saprolith is characterized by high SiO2

(mean 66.6%) and low CaO (mean 0.09%) and MnO (mean 0.10%).SiO2 (45.7–75.0%) and Al2O3 (15.0–38.4%) are the major constituentsof both the bulk kaolin samples and finer fractions (< 45 μm). Fe2O3

varies from 0.63 to 2.89%, K2O varies from 1.83 to 4.81%, and theremaining major-element oxides show concentrations of< 1%. Thefiner fractions (< 45 μm) in all three sections show increases in Al2O3(up to 38.4%) and decreases in SiO2 (mean 46.2%) compared to thebulk kaolin samples (Table S2). The major element concentrations ofthe samples correspond well to the mineralogic compositions de-termined by XRD.

Trace elements such as Zn, Zr, and Ba exhibit high concentrations(~100–3600 ppm), V, Ga, Sr, Nb, Ta, and Pb have intermediate con-centrations (~10–100 ppm), and Sc, Y, Sn, Th, and U show low con-centrations (< 10 ppm). The UCC-normalized trace-element and REEdistributions of the kaolin samples show similar features for all samples(Fig. 4), including depletions of many elements (e.g., Y, Zr, Th, andmedium-REE Dy) that are usually considered to be resistant to re-mobilization during weathering, enrichments of some relatively mobileelements (e.g., Ba and Ga), and negative Ce anomalies. These featuresshould be mainly inherited from parent granite rocks, rather than as aresult of chemical weathering, because the distribution patterns aresimilar for weakly weathered saprolith and strongly weathered kaolinsamples. These features are typical of trace-element and REE distribu-tions of Mesozoic granitoids in South China (Zhou et al., 2006). Actu-ally it has been widely suggested that chemical weathering can oftenexert a limited influence on REE distributions of weathered profiles(e.g., Galán et al., 2016; Hong et al., 2016).

3.3. H and O isotopic compositions

Before evaluating the H and O isotopic results, it is necessary toestimate the contributions of impurity minerals to the measured iso-topic data. As pointed out by Savin and Lee (1988), the isotope

Fig. 3. Representative X-ray diffraction patterns of (A) bulk kaolin, (B) finerfractions (< 45 μm), (C) clay fractions (< 2 μm), and (D) heavy minerals.

Q. Fang et al. Geoderma 337 (2019) 225–237

228

fractionation factors of weathering-formed (secondary) muscovite andgibbsite are similar to and slightly smaller than those of kaolinite, re-spectively. However, at least a small part of the muscovite in oursamples may be residual primary muscovite from the parent granite,which formed at high crystallization temperatures and has an unknownisotopic composition. As noted before, the halloysite content is negli-gible and should have contributed little to the measured isotopic data.An empirical approach by Clauer et al. (2015) suggests that the po-tential influence of contaminant quartz can be disregarded. Assessmentof the contributions of impurity minerals to the H isotopic results yieldssimilar conclusions as for O isotopes (cf. Clauer et al., 2015). In con-clusion, the minor amounts of impurity minerals in our samples arelikely to have only a minor influence, at most, on measured isotopicvalues.

The δ18O and δ2H values of the clay (< 2 μm) fractions range from+16.8 to +18.7‰ and from −70 to −51‰, respectively (Fig. 5). In aplot of δ18O vs. δ2H, the Chongyi kaolin samples mostly lie close to thekaolinite weathering line (representing the isotopic composition ofkaolinite in equilibrium with meteoric water at 20 °C; Fig. 5).

Supergene kaolin samples of Ramon-Fazouro (Clauer et al., 2015) andChubut/San Cruz (Cravero et al., 2001) and hypogene samples of Seilizand Kemmliz (Gilg et al., 2003) are plotted for comparison (Fig. 5).

3.4. SEM observations

3.4.1. Kaolin mineralsThe kaolin minerals occur either as discrete plates or as composite

stacks, and they exhibit one of two morphologies: relatively porousaggregates and well-developed crystals (Fig. 6). The porous aggregatesconsist of pseudo-hexagonal plates with poorly developed lateral sur-faces (010) and (110), yielding irregularly ragged outlines (Fig. 6A).The clay particles range in size from 1 to 5 μm. The second morphologyconsists of well-developed euhedral pseudo-hexagonal crystals, usuallyarranged face-to-face in elongated book-like stacks (Fig. 6B). The lateralsurfaces (010) and (110) are flat and clearly defined, and the particlesizes are larger, ranging from 3 to 10 μm. Many of the kaolin aggregatesare vermiform or fan-shaped (Fig. 6C–E), and crystal elongation parallelto the c-axis (to 1.5mm) was retained within the composite stacks. This

Fig. 4. Upper continental crust (UCC) - normalized trace element diagrams for mottled kaolin and surface layer (A), bulk homogeneous kaolin and< 45 μmhomogeneous kaolin fractions (B), and saprolith (C). UCC-normalized REE distributions for mottled kaolin and surface layer (D), bulk kaolin and<45 μm kaolinfractions (E), and saprolith (F).

Q. Fang et al. Geoderma 337 (2019) 225–237

229

suggests that growth of the composite stacks was initiated from one sideof a muscovite layer and then propagated inward to the other side of thelayer (Chen et al., 1997). Fine tubular crystals of halloysite (an elongatekaolin mineral; 0.3–3 μm long), which are common in weathered kaolindeposits (Churchman et al., 2010; Keeling, 2015), cover the surfaces ofblocky and book-like kaolinite aggregates (Fig. 6). However, halloysiteis not abundant in our samples, as shown by both SEM and XRD results.

3.4.2. Heavy mineralsTi-oxides are abundant in the heavy-mineral fraction and diverse in

their morphology and composition (Fig. 7A, B). The TiO2 content of thebulk samples varies from 0.10 to 0.13 wt%. Ti-oxide minerals alwayscontain a certain amount of niobium (Nb) (Dill, 2017), which can beseen in the EDS spectrum (Fig. 7A) and is an indication (along with anXRD peak at 3.51 Å) that the grain consists of anatase. Some Ti-oxidegrains show indentations and notches along their crystal boundaries,which may reflect intergrowth of rutile with other minerals such asquartz (Fig. 7B). Iron-oxides are rare in the homogenous kaolin layerbut relatively abundant in the mottled kaolin layer. Altered magnetite ispresent as dense massive aggregates, and it contains a certain amount ofchromium (Cr) (Fig. 7C). Altered magnetite may also be present asnodular/earthy aggregates (Fig. 7D). Both types of magnetite have re-latively high Cl concentrations (Fig. 7C, D), consistent with its inter-action with the volatile component of a magmatic system, e.g., duringmagma cooling and crystallization (Wang et al., 2014). Tourmaline ispresent as prisms, irregular fragments, and sub-round to round grains(e.g., Fig. 7E). The presence of iron and magnesium and the dark colorare consistent with Fe, Mg-tourmaline (Henry and Guidotti, 1985).Wolframite, which generally occurs as tabular grains, is a solid solutionbetween the endmembers huebnerite (MnWO4) and ferberite (FeWO4)and shows widely varying chemical compositions. In our samples,Mn≫ Fe suggests a composition close to the huebnerite endmember(Fig. 7F).

3.5. DTA-TG analysis

The DTA-TG curves for bulk kaolin samples (Fig. 8A) and< 45 μmfractions (Fig. 8B) show similar thermal reactions and are consistentwith the results of XRD, SEM, and geochemical analyses. The samplesshow an endothermal peak at a temperature of ~90–100 °C attributedto the loss of physically adsorbed water in pores, on mineral surfaces(Kakali et al., 2001). Weight loss over the temperature range of90–350 °C (−0.93% for bulk samples and −2.07% for the<45 μmfractions) is associated with a pre-dehydration process, which occurs asa result of reorganization of the octahedral sheet of kaolinite (Balek andMurat, 1996; Manju et al., 2001). Weight loss over the temperaturerange of 400–650 °C and an endothermal peak at ~525 °C are evidenceof endothermic reactions linked to dehydroxylation of kaolinite andformation of metakaolinite (Cravero et al., 2016). Weight loss over thetemperature range of ~600–800 °C may represent complete dehydrox-ylation of kaolinite and conversion to metakaolinite, or possibly de-composition of alunite. However, SO3 concentrations are low (Table S2)and, assuming that all SO3 is present in alunite, the maximum alunitecontent would be< 0.1% and, thus, unlikely to be a factor in sampleweight loss. An exothermic reaction at ~970–990 °C is due to trans-formation of metakaolinite to spinel and amorphous silica, and exo-thermic peaks at 1100–1400 °C are attributed to transformation ofmetakaolinite to mullite and cristobalite.

4. Discussion

4.1. Supergene kaolinization and associated characteristics of the regolith

None of hypogene indictors (hydrothermal mineral assemblages,shallow igneous activity, mineral zoning, and typical structures such asfunnel-shaped, pipelike or elongated zones of kaolinization) are ob-served in the present study units, and, thus, hypogene kaolinization canbe ruled out. Instead, the Chongyi kaolin regolith shows a variety ofmineralogic, elemental, and isotopic characteristics typical of surficialweathering, as discussed below.

Kaolinization in the Chongyi regolith is usually associated withminerals produced through low-temperature chemical weathering, e.g.,gibbsite, anatase, and iron-oxides. For example, gibbsite, which is ty-pically found in highly weathered soils/sediments, is generated throughalteration of aluminosilicate or aluminous minerals under intenseweathering conditions (Ramasamy et al., 2014).In addition, tungstenmineralization (here as wolframite) can also be observed. The presenceof tourmaline is indicative of high-temperature fractionation within thegranite, inconsistent with conditions for kaolinite formation. Theseminerals can be a weathering residuum from a primary high-tempera-ture igneous intrusive body.

SEM observations show that Chongyi kaolinite aggregates arecharacterized by a variety of morphological features typical of kaoli-nites produced through supergene weathering processes (Chen et al.,1997; Churchman and Lowe, 2012). Book-like pseudohexagonal kaoli-nite crystals are commonly due to retention of the original morphologyof feldspar crystals (Ekosse, 2001).

Normalized element-distribution patterns do not reveal much aboutchemical weathering but rather about the parent granite rocks (Fig. 4).However, elemental compositions of the Chongyi kaolin show featurestypical of surfical weathering. Although Ti and Zr can be released fromprimary minerals during alteration, they are relatively immobile ele-ments under Earth-surface conditions, and are concentrated more effi-ciently through chemical weathering (Marfil et al., 2005). In a Zr vs.TiO2 plot, for example, the homogenous kaolin and mottled kaolinsamples of the Chongyi deposit fall very close to the supergene field,suggesting a significant surficial weathering contribution (Fig. 9A).Alunite and some aluminum-phosphate-sulfate minerals are common inhypogene deposits, thus these deposits are usually enriched in P and S(Dill et al., 1997). As a result, the rather low P2O5 and SO3

Fig. 5. Plot of δ18O vs. δ2H values showing results for Chongyi kaolin andmottled kaolin samples relative to the supergene/hypogene line and the linemarking isotopic composition of kaolinite in equilibrium with meteoric water at20 °C (Savin and Epstein, 1970). Reference data for Seiliz/Kemmlitz, Ramon-Fazouro, and Chubut/San Cruz kaolin deposits are from Gilg et al. (2003),Clauer et al. (2015), and Cravero et al. (2001), respectively.

Q. Fang et al. Geoderma 337 (2019) 225–237

230

concentrations of the Chongyi deposit further demonstrate few influ-ences of hydrothermal alteration (Fig. 9B), consistent with the fieldobservations and XRD analysis. Besides, vertical elemental profiles canalso well present supergene chemical characteristics of the Chongyikaolin regolith (see Section 4.3).

The H and O isotopic compositions of kaolinite provide constraintson the geologic conditions of its formation, assuming no post-deposi-tional alteration has occurred (Savin and Lee, 1988; Gilg et al., 2003;Baioumy, 2013). The oxygen isotopic composition of the kaolin sug-gests the moderate depth of weathering. In δ18O vs. δ2H crossplots, theChongyi samples plot entirely within the supergene field, suggestingsignificant contribution from chemical weathering (Fig. 5). The global

kaolinite weathering line, whose equation is δ2H=4.85× δ18O− 152,represents the isotopic composition of kaolinite in equilibrium withmeteoric water at 20 °C (Savin and Epstein, 1970). Most of the Chongyisamples plot very close to the global kaolinite weathering line (Fig. 5),indicating the influence of meteoric waters (Ma et al., 2017). The re-lationship of δ18O to δ2H in kaolinites from weathered profiles maypermit calculation of a linear equation and isotope fractionation factor(Savin and Lee, 1988; Gilg et al., 2003). The best-fit line for the Chongyisamples is δ2H=4.85× δ18O− 150. Although the δ18O and δ2H va-lues data of the full sample set are poorly correlated (r= 0.124, n=9,p > 0.1), the samples with δ2H < −60‰ (which were used to definethe best-fit line) exhibit a strong correlation (r= 0.940, n=5,

Fig. 6. SEM images of: (A) poorly crystallized kaolinite (Kaol) with irregular outlines typical of chemical weathering; (B) face-to-face in elongated book-like stacks;(C–E) kaolinite aggregates with vermiform or fan-shaped morphologies; (F) tubular crystals of halloysite (Ha); (G) a sketched diagram shows the progressivetransformation of muscovite to kaolinite stacks, involving fan-shaped composite stacks (after Chen et al., 1997). Note that halloysite attached on kaolinite is muchless abundant and more sporadic than kaolinite under SEM.

Q. Fang et al. Geoderma 337 (2019) 225–237

231

p < 0.01).These observations may suggest that minor residual mineral im-

purities (e.g., quartz and muscovite) with higher δ2H values(e.g., >−60‰) have exerted a measurable influence on the isotopiccompositions of some Chongyi samples.

Deviations from the global kaolinite weathering line are potentiallyattributable to post-depositional modifications in weathered profiles(Baioumy, 2013). Such modifications can result from the interactionsbetween earlier-formed kaolinite and downward-percolating meteoricwaters (Baioumy, 2013). Another possible cause of deviations is con-tributions from high-temperature volatiles. Although evidence for suchinfluences is limited in the Chongyi deposit, the presence of Cl-enrichedmagnetite may be significant in this regard (Fig. 7C, D).

4.2. Kaolin regolith subdivision and mineral weathering

The Chongyi kaolin was developed on a monolithologic and

homogeneous protolith (i.e., muscovite granite) without significantsediment inputs. The study area has a tropical monsoonal climate withwarm temperatures and high precipitation. Therefore, the Chongyikaolin deposit can be considered as a “kaolin regolith” (i.e., a residualsoil deposit) formed under intense weathering conditions.

Before further discussion of mineral transformations and chemicalgradients within the Chongyi regolith, it is necessary to verticallysubdivide it and compare its layering with that of typical deeplyweathered profiles, as given in Taylor and Eggleton (2001) and Anandand Paine (2002). Regolith consists of a lower saprolith layer and anupper pedolith layer (Fig. 2). The saprolith consists of two layers, sa-prorock below and saprolite above. As suggested by Taylor andEggleton (2001), the pedolith is more strongly modified through soil-forming processes and is commonly composed (from bottom to top) of aplasmic zone (also referred to as a pallid zone because of its whitecolor), a mottled zone, and a laterite zone (the terms ferruginous dur-icrust and lateritic residuum may also be used). The pedolith of the

Fig. 7. SEM images with EDS spectra of various heavy minerals: (A) anatase grain with major Ti and minor Nb; (B) rutile with some “indentations” and notchesobserved along grain boundaries; (C) altered magnetite is present as dense massive aggregates; (D) altered magnetite is present as nodular/earthy aggregates, andthese magnetite grains have relatively high Cl concentrations; (E) a prismatic crystal of Fe, Mg-tourmaline; (F) tabular grain of huebnerite with Mn≫ Fe.

Q. Fang et al. Geoderma 337 (2019) 225–237

232

Chongyi profile contains a plasmic zone, a mottled zone, and an over-lying surface layer (Fig. 2), but the laterite zone that is commonly abovethe mottled zone is absent. This vertical layering pattern is similar tothat of the Yalanbee profile reported by Gilkes et al. (1973). Below thisparagraph, we try describing the mottled zone and plasmic zone in theChongyi regolith in more details with emphasis on some related im-portant mineralogic and geochemical reactions.

The mottled zone, which is found in the middle-upper part of thepedolith (Fig. 2), has cemented segregations of subdominant colordifferent from the surrounding matrix within the Chongyi regolith. Themottles in the Chongyi kaolin have sharp and distinct boundaries,ranging in size from ~5 to ~100mm. Mottles frequently form due tomigration of Fe2+ in soil waters and precipitation as Fe3+ upon contactwith an oxidizing region within the regolith (depending on the depth),which are a common influence on the development of local redoxvariation within soils and, hence, soil mottling (Taylor and Eggleton,2001). Based on Taylor and Eggleton (2001), long-term seasonal riseand fall of the groundwater table can lead to reductive mobilization ofFe from the saprolite zone and subsequent precipitation higher in thesoil profile upon oxidation at the groundwater table. As a result, thesaprolite zone is a major source of Fe for mottle formation.

Below the mottled zone is the homogeneous kaolin zone (Fig. 2),which corresponds to the plasmic zone and is transitional to the un-derlying saprolite (Gilkes et al., 1973; Anand and Paine, 2002). Thiszone, which consists mainly of kaolinite, is mesoscopically homogenousowing to destruction of the original sediment fabric (Taylor andEggleton, 2001). Kaolinization of feldspar by chemical weathering isthe main source of kaolin (e.g., Gilg et al., 1999; Churchman et al.,2010). Many kaolinite aggregates within regolith (and especially in the

plasmic zone) have fan-shaped stacks (Fig. 6C–E), a morphology that isclosely associated with transformation of muscovite plates. Feldsparmay not directly transform into kaolinite but rather first into muscoviteas an intermediate stage under certain geologic conditions (Chen et al.,1997; Lu et al., 2016), which may have been the case for the Chongyikaolin.

As illustrated in Fig. 6G, the transformation process involves theparallel growth of kaolinite originating from the edges of muscovitegrains and subsequent topotactic growth, with the development of fan-shape composite stacks or coalesced stacks. The edges of the kaolinitestacks may split apart into discrete plates (Fig. 6G; Chen et al.,1997).Muscovite of the parent granite can be another important sourceof kaolinite (Gilkes et al., 1973).

Thermodynamically, the Gibbs free energy of kaolinite formation isslightly lower than that of halloysite formation, so halloysite is lessstable than kaolinite (Jeong, 2000; Keeling, 2015). For this reason, ifmeta-stable halloysite forms, it should do so in advance of formation ofthe more stable kaolinite phase, due to its lower activation energy offormation (Stumm, 1992). In the Chongyi deposit, however, halloysiteis observed to have crystallized after the formation of kaolinite. Thisapparently contradictory relationship can be ascribed to the presence ofweathered mica grains, the surfaces of which provide favorable nu-cleation sites for kaolinite by significantly reducing its activation en-ergy (Jeong, 2000; Keeling, 2015).

Water-saturated soils in tropical/subtropical climates can lead toabundant halloysite formation (Joussein et al., 2005). Dominance ofkaolinite over halloysite in the Chongyi deposit implies a near-surface,freely draining environment (Churchman and Gilkes, 1989; Churchmanet al., 2010). However, it is possible that the tubular halloysite crystals

Fig. 8. Representative differential thermal (green) and thermogravimetric (blue) curves of (A) bulk kaolin and (B)< 45 μm kaolin samples. (For interpretation of thereferences to color in this figure legend, the reader is referred to the web version of this article.)

Q. Fang et al. Geoderma 337 (2019) 225–237

233

within the weathered kaolin directly crystallized from soil watersduring periodic waterlogging of microcavities and microfissures, whenthe activity of aqueous Al and Si did not lead to nucleation of kaoliniteon existing mineral surfaces (Chen et al., 1997; Joussein et al., 2005).

4.3. Chemical gradients in the kaolin regolith

The weathering of rock-forming minerals (e.g., feldspar and mus-covite) of igneous parent rocks to kaolin is mainly controlled by thechemical composition of the parent rock and the climatic and hydro-logic conditions of the supergene environment (Fang et al., 2017).Elemental distributions are related to retention and leaching of variouselements and to mineral transformations in the principal regolith layers(Tardy et al., 1973; Anand and Paine, 2002; Scott and Pain, 2008).Here, we examine chemical gradients in the Chongyi regolith to in-vestigate how supergene processes influenced and transformed theoriginal kaolin deposit, and to compare them to typical regolith models.

Elemental concentrations in weathered profiles are typically nor-malized to the concentration of an immobile element such as Ti, Nb, orZr (White, 1995). In the present case, Ti and Nb are present in rutileand/or anatase, so their concentrations may be influenced by variationsin the relative content of these minerals in the host rock. Zr is presentalmost exclusively in zircons, making it the most suitable element fornormalization. The relative gains and losses of elements were calculatedbased on the assumption of immobility of Zr.

The mass transfer coefficient (τi,j) quantifies elemental changes inregolith profiles, accounting for both volume changes and relative gainsor losses of elements (Anderson et al., 2002):

= −

CC

CC

τ 1j w

j p

i p

i wi,j

,

,

,

, (1)

where C represents concentration of mobile (j) or immobile (i) elementsin parent (p) or weathered (w) material. Negative τi,j values suggestdepletion of element j, positive values mean enrichment, and whenτi,j = 0, the concentration of an element is identical to that of the parentmaterial. We used the underlying weakly altered granite (saprolith;Anand and Paine, 2002) as the parent material in calculating the ele-mental profiles of Fig. 10.

Elemental profiles are one of the most important lines of evidencefor assessing pedogenic processes in regoliths. The patterns of elementalchanges with depth can be assigned to specific endmember types(Brantley et al., 2008; Brantley and White, 2009), of which four arerecognized: (1) Depletion profiles develop for elements that are stronglydepleted in the pedolith (mostly in the upper levels of the regolith),usually due to mineral dissolution, and without reprecipitation at depth(Fig. 10A). For example, Na commonly exhibits a depletion profileowing to weathering of plagioclase feldspar (White, 1995) and a lack ofreprecipitation in illuviated oxides or clay minerals. (2) Depletion-en-richment profiles develop for elements that are leached in an upper re-golith layer but reprecipitated at depth (Fig. 10B). For example, Al iscommonly mobilized as colloids and organic complexes in the surfacelayer and then redeposited deeper in a regolith profile (e.g., mottledkaolin zone), where concentrations of organic ligands are lower(Brantley and White, 2009). (3) Addition profiles develop for elementsthat are added to the regolith surface layer through atmospheric de-position, and which are redistributed downward through leaching ormixing processes (Fig. 10C). Addition profiles are characterized by peakconcentrations at the top of the regolith profile and declining con-centrations at depth. For example, Mn profiles can be dominated by wetor dry deposition from the atmosphere (Herndon et al., 2011). (4)Biogenic profiles exhibit trends that are opposite to depletion-enrichmentprofiles (Fig. 10D). They show peak enrichments at or close to the top ofthe regolith profile owing to the effects of organic acid secretion byplants or fungi and the consequent dissolution and uptake of nutrientelements (Brantley and White, 2009). If the released nutrients (e.g., Ca)are taken up by roots, or recycled in the upper layer of soil, a biogenicprofile develops (Jobbagy and Jackson, 2001).

These four types of elemental profiles tend to be distinctly devel-oped in thin in situ regoliths, whereas deeply weathered (thick) re-goliths formed on granite host rocks may show different chemicalprofiles (e.g., Gilkes et al., 1973; Davy and El-Ansary, 1986; Anand andPaine, 2002; Scott and Pain, 2008). The weathering behavior of ele-ments in thick granite-derived regoliths has been summarized by Davyand El-Ansary (1986) and Anand and Paine (2002). Here, we comparefour elements (i.e., Na, Al, Mn, and Ca) representing different types ofelemental profiles in the Chongyi regolith with those from publishedsources (Fig. 10).

Elemental profiles of mobile elements such as Na exhibit similarpatterns in both the Chongyi kaolin and deeply weathered profiles(reference profiles; Fig. 10A). These elements are characterized byleaching and depletion throughout the regolith, without reprecipitationat depth. Elemental profiles of elements like Al show different patternsin the two types of regolith: they are leached from upper levels (e.g.,mottled kaolin zone) and reprecipitated at depth in the Chongyi kaolin,whereas they are generally residually enriched towards the surfacelayer in the reference profiles (Fig. 10B). Elements such as Mn areleached and depleted at deeper levels (e.g., in the saprolith) in bothtypes of regolith, and they are enriched in the upper part of the Chongyiregolith (e.g., mottled kaolin zone); but they are residually enrichedthroughout the reference profiles (Fig. 10C). Elemental profiles of ele-ments like Ca exhibit similar patterns in both types of regolith, andthese elements are reprecipitated/enriched at or close to the top of theprofile (Fig. 10D).

Fig. 9. (A) A Zr vs. TiO2 plot suggests a supergene origin of the Chongyi kaolin,given samples fall more close to those supergene samples with relatively high Zrand Ti concentrations; (B) rather low P2O5 and SO3 concentrations in theChongyi kaolin also point to supergene kaolinization processes. Due to thecommon presence of alunite and some aluminum-phosphate-sulfate minerals inhypogene deposits, they are usually enriched in P and S (Dill et al., 1997).Reference plots of supergene, hypogene, and mixed-type kaolin samples arefrom Kitagawa and Koster (1991), Dill et al. (1997), and Dill (2015).

Q. Fang et al. Geoderma 337 (2019) 225–237

234

4.4. Implications for formation of kaolin regolith and supergene pedogenesis

Previous studies on various argillization, in particular kaolin, focusmore commonly on its economic value, compared to systematic studiesregarding pedological, mineralogic, and geochemical processes takingplace in kaolin regoliths (Ekosse, 2001; Anand and Paine, 2002; Scottand Pain, 2008). This would limit our understanding of kaolin argilli-zation linked to supergene processes in the Earth's surface environ-ments. In this study, we demonstrated that the kaolin regolith fromsubtropical China was derived from supergene weathering processesusing integrated mineralogic and geochemical analyses.

While the kaolin regolith can be subdivided into several layers withreference to typical subdivision of the deeply weathered profile, theirelemental profiles are not necessarily uniform with those of typicaldeep granite weathering profiles. Instead, they are generally consistentwith those of relative thin in situ soil systems as widely reported pre-viously. The four types of elemental profiles reflect chemical gradientsin the kaolin regolith as well as some biological effects (depending onthe depth).These elemental patterns are affected by a combination ofchemical, geologic, and biologic processes that control soil formationand evolution (Brantley and White, 2009). That is to say, the supergenekaolin regolith is intrinsically more similar to shallow, biologicallyactive residual soil deposits, rather than the deep granite weatheringprofiles.

Based on the discussion above, we propose a model to illustrategeological evolution, mineralogic transformation, and elemental mi-gration within the kaolin regolith in particular linked to supergeneprocesses (Fig. 11). Kaolin may form in semi-humid to humid mid-la-titude climate zones with moderate to intense weathering intensity, but

chemical weathering alone in such settings are usually unable to initiatesupergene kaolinization. The initial processes of weathering could bephysical, for example, the development of physical openings alongjoints, which may produce small fragments moving locally (Fig. 11;Taylor and Eggleton, 2001). Strong weathering conditions typical ofhumid subtropical/tropical climate zone with high MAT and MAP (e.g.,the Chongyi region) favor supergene argillization. As chemical weath-ering occurs, solutes are moved by solutions passing through thegranite. The exposed kaolin experienced enhanced weathering undersuch climatic conditions, resulting in new regolith formation (Fig. 11).

5. Conclusions

Three kaolin-regolith drillholes from a subtropical climate zone insouthern China were investigated using an integrated mineralogic,elemental, and isotopic geochemical approach. Neither shallow level ofgranite emplacement, representative structural controls on kaoliniza-tion, mineral zoning, nor hypothermal minerals can be observed. Themineral assemblages display typical supergene characteristics, andkaolinite mineral aggregates are characterized by a variety of mor-phological features suggestive of supergene weathering. Geochemicalbinary plots (e.g., Zr vs. TiO2, P2O5 vs. SO3, and δ18O vs. δ2H) also pointto a supergene origin for the kaolin regolith. Numerous kaolinite ag-gregates within regolith have fan-shaped stacks, a morphology that istightly associated with transformation of muscovite plates. Dominanceof kaolinite over halloysite implies a near-surface, freely draining en-vironment.

Four types of regolith elemental profiles can be identified in thekaolin regolith: depletion (e.g., Na), depletion-enrichment (e.g., Al),

Fig. 10. Elemental profiles (orange solid lines) of the Chongyi regolith can be assigned to four specific endmember types, i.e., (A) depletion, (B) depletion-en-richment, (C) addition, and (D) biogenic profiles. These profiles tend to be well-developed in thin in situ regoliths. Elemental profiles (black dotted lines) commonlyobserved in deeply weathered (thick) profiles developed on granite are also shown for comparison (Davy and El-Ansary, 1986; Anand and Paine, 2002).

Q. Fang et al. Geoderma 337 (2019) 225–237

235

addition (e.g., Mn), and biogenic (e.g., Ca) profiles. These profiles re-flect chemical gradients in the kaolin regolith as well as some biologicaleffects that are typical of relative thin in situ soil systems. However, theelemental profiles are not necessarily consistent those of typical deep(thick) granite weathering profiles. Strong weathering conditions, ty-pical of humid subtropical/tropical climate zone with high MAT andMAP favor supergene argillization, resulting in new kaolin regolithformation.

Acknowledgments

This study was supported by the Special Funding for SoilMineralogy (CUG170106), National Science Foundation of China(41772032 and 41472041), and the Fundamental Research Funds forNational Universities (CUG-Wuhan). TJA acknowledges support fromthe NASA Exobiology program (NNX13AJ1IG) and the China Universityof Geosciences (SKL-GPMR 201301 and SKL-BGEG BGL21407). QF andLZ acknowledge the China Scholarship Council (CSC) for financialsupport (201706410017 for QF and 201706410006 for LZ). LZ thanksDr. David Dettman for providing her an opportunity to visit theUniversity of Arizona and for his kind guidance and experimentalsupport. The authors wish to thank Xinrong Lei, Yao Liu, Hetang Hei,and Shuling Chen for help with laboratory work, and Ke Yin, ChaowenWang, Hongkun Dai, and Jiacheng Liu for helpful discussion. Threeanonymous reviewers and Prof. M. Vepraskas provided meticulouscomments that greatly improved this work.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.geoderma.2018.09.020.

References

Anand, R.R., Paine, M., 2002. Regolith geology of the Yilgarn Craton, Western Australia:implications for exploration. Aust. J. Earth Sci. 49, 3–162.

Anderson, S.P., Dietrich, W.E., Brimhall, G.H., 2002. Weathering profiles, mass balanceanalysis, and rates of solute loss: linkages between weathering and erosion in a small,steep catchment. Geol. Soc. Am. Bull. 114, 1143–1158.

Baioumy, H.M., 2013. Hydrogen and oxygen isotopic compositions of sedimentary kaolindeposits, Egypt: Paleoclimatic implications. Appl. Geochem. 29, 182–188.

Balek, V., Murat, M., 1996. The emanation thermal analysis of kaolinite clay minerals.Thermochim. Acta 282, 385–397.

Brantley, S.L., White, A.F., 2009. Approaches to modeling weathered regolith. Rev.Mineral. Geochem. 70, 435–484.

Brantley, S.L., Bandstra, J., Moore, J., White, A.F., 2008. Modelling chemical depletionprofiles in regolith. Geoderma 145, 494–504.

Buss, H.L., Lara, M.C., Moore, O.W., Kurtz, A.C., Schulz, M.S., White, A.F., 2017.Lithological influences on contemporary and long-term regolith weathering at theLuquillo Critical Zone Observatory. Geochim. Cosmochim. Acta 196, 224–251.

Chen, P.Y., Lin, M.L., Zheng, Z., 1997. On the origin of the name kaolin and the kaolindeposits of the Kauling and Dazhou areas, Kiangsi, China. Appl. Clay Sci. 12, 1–25.

Chorover, J., Sposito, G., 1995. Surface charge characteristics of kaolinitic tropical soils.Geochim. Cosmochim. Acta 59, 875–884.

Churchman, G.J., Gilkes, R.J., 1989. Recognition of intermediates in the possible trans-formation of halloysites to kaolinites in weathering profiles. Clay Miner. 24,579–590.

Churchman, G.J., Lowe, D.J., 2012. Alteration, formation, and occurrence of minerals insoils. In: Huang, P.M., Li, Y., Sumner, M.E., 2nd ed. (Eds.), Handbook of Soil Sciences.Properties and Processes Vol. 1. CRC Press (Taylor & Francis), Boca Raton, Florida,pp. 20.1–20.72.

Churchman, G.J., Whitton, J.S., Claridge, G.G.C., Theng, B.K.G., 1984. Intercalationmethod using formamide for differentiating halloysite from kaolinite. Clay ClayMiner. 32, 241–248.

Churchman, G.J., Pontifex, I.R., McClure, S.G., 2010. Factors influencing the formationand characteristics of halloysites or kaolinites in granitic and tuffaceous saprolites inHong Kong. Clay Clay Miner. 58, 220–237.

Clauer, N., Fallick, A.E., Galán, E., Aparicio, P., Miras, A., Fernández-Caliani, J.C., Aubert,A., 2015. Stable isotope constraints on the origin of kaolin deposits from Variscangranitoids of Galicia (NW Spain). Chem. Geol. 417, 90–101.

Cravero, F., Dominguez, E., Iglesias, C., 2001. Genesis and applications of the Cerro Rubiokaolin deposit, Patagonia (Argentina). Appl. Clay Sci. 18, 157–172.

Cravero, F., Fernández, L., Marfil, S., Sánchez, M., Maiza, P., Martínez, A., 2016.Spheroidal halloysites from Patagonia, Argentina: some aspects of their formation

Fig. 11. A sketched model for geological evolution of the kaolin regolith (after Anand and Paine, 2002). At first, the granite was uplifted by regional tectonicmovement, and the erosion base level was reaching the surface of the granite. Then primary minerals of granite were weathered to secondary minerals with theformation of saprolith and pedolith. Further weathering occurred with migration and accumulation of Fe as mottles in mottled kaolin zone. With the supergeneprocess going on, the primary granite was gradually changed to a kaolin regolith. Also shown is typical elemental leaching and precipitation (e.g., Na, Al, Mn, and Ca)along the Chongyi regolith.

Q. Fang et al. Geoderma 337 (2019) 225–237

236

and applications. Appl. Clay Sci. 131, 48–58.Davy, R., El-Ansary, M., 1986. Geochemical patterns in the laterite profile at the

Boddington gold deposit, Western Australia. J. Geochem. Explor. 26, 119–144.Dill, H.G., 2015. The Hagendorf-Pleystein Province: the center of pegmatites in an ensialic

orogen. In: Dilek, Y., Pirajno, F., Windley, B. (Eds.), Modern Approaches in SolidEarth Sciences. Springer, New York (475 pp).

Dill, H.G., 2017. Residual clay deposits on basement rocks: the impact of climate and thegeological setting on supergene argillitization in the bohemian massif (CentralEurope) and across the globe. Earth Sci. Rev. 165, 1–58.

Dill, H.G., Bosse, R., Henning, H., Fricke, A., 1997. Mineralogical and chemical variationsin hypogene and supergene kaolin deposits in a mobile fold belt the Central Andes ofnorthwestern Peru. Mineral. Deposita 32 (2), 149–163.

Dudoignon, P., Beaufort, D., Meunier, A., 1988. Hydrothermal and supergene alterationsin the granitic cupola of Montebras, Creuse, France. Clay Clay Miner. 36, 505–520.

Ekosse, G., 2001. Provenance of the Kgwakgwe kaolin deposit in southeastern Botswanaand its possible utilization. Appl. Clay Sci. 20 (3), 137–152.

Erkoyun, H., Kadir, S., 2011. Mineralogy, micromorphology, geochemistry and genesis ofa hydrothermal kaolinite deposit and altered Miocene host volcanites in the Hallaclararea, Usak, western Turkey. Clay Miner. 46, 421–448.

Fang, Q., Hong, H., Zhao, L., Furnes, H., Lu, H., Han, W., Liu, Y., Jia, Z., Wang, C., Yin, K.,Algeo, T.J., 2017. Tectonic uplift-influenced monsoonal changes promoted homininoccupation of the Luonan Basin: insights from a loess-paleosol sequence, easternQinling Mountains, central China. Quat. Sci. Rev. 169, 312–329.

Fernandez-Caliani, J.C., Galán, E., Aparicio, P., Miras, A., Marquez, M.G., 2010. Originand geochemical evolution of the Nuevo Montecastelo kaolin deposit (Galicia, NWSpain). Appl. Clay Sci. 49, 91–97.

Galán, E., Aparicio, P., Fernández-Caliani, J.C., Miras, A., Márquez, M.G., Fallick, A.E.,Clauer, N., 2016. New insights on mineralogy and genesis of kaolin deposits: theBurela kaolin deposit (Northwestern Spain). Appl. Clay Sci. 131, 14–26.

Gilg, H.A., Hülmeyer, S., Miller, H., Sheppard, S.M.F., 1999. Supergene origin of theLastarria kaolin deposit, South-Central Chile, and paleoclimatic implications. ClayClay Miner. 47, 201–211.

Gilg, H.A., Weber, B., Kasbohm, J., Frei, R., 2003. Isotope geochemistry and origin ofillite-smectite and kaolinite from the Seilitz and Kemmlitz kaolin deposits, Saxony,Germany. Clay Miner. 38, 95–112.

Gilkes, R.J., Scholz, G., Dimmock, G.M., 1973. Lateritic deep weathering of granite. Eur.J. Soil Sci. 24, 523–536.

Henry, D.J., Guidotti, C.V., 1985. Tourmaline as a petrogenetic indicator mineral - anexample from the staurolite-grade metapelites of NW Maine. Am. Mineral. 70, 1–15.

Herndon, E.M., Jin, L., Brantley, S.L., 2011. Soils reveal widespread manganese enrich-ment from industrial inputs. Environ. Sci. Technol. 45, 241–247.

Hillier, S., 2000. Accurate quantitative analysis of clay and other minerals in sandstonesby XRD: comparison of a Rietveld and a reference intensity ratio (RIR) method andthe importance of sample preparation. Clay Miner. 35, 291–302.

Hong, H., Fang, Q., Cheng, L., Churchman, G.J., Wang, C., 2016. Microorganism-inducedweathering of clay minerals in a hydromorphic soil. Geochim. Cosmochim. Acta 184,272–288.

Jeong, G.Y., 2000. The dependence of localized crystallization of halloysite and kaoliniteon primary minerals in the weathering profile of granite. Clay Clay Miner. 48,196–203.

Jobbagy, E.G., Jackson, R.B., 2001. The distribution of soil nutrients with depth: globalpatterns and the imprint of plants. Biogeochemistry 53, 51–77.

Joussein, E., Petit, S., Churchman, J., Theng, B., Righi, D., Delvaux, B., 2005. HalloysiteClay Minerals—A Review.

Kakali, G., Perraki, T., Tsivilis, S., Badogiannis, E., 2001. Thermal treatment of kaolin: theeffect of mineralogy on the pozzolanic activity. Appl. Clay Sci. 20, 73–80.

Keeling, J.L., 2015. The mineralogy, geology and occurrence of halloysite. In: Pasbakhsh,P., Churchman, G.J. (Eds.), Natural Mineral Nanotubes, Properties and Applications.Apple Academic Press, Inc., Oakville, Canada, pp. 95–116.

Keeling, J., Mauger, A.J., Scott, K.M., Hartley, K., 2003. Alteration mineralogy and acid-sulphate weathering at Moonta copper mines, South Australia. Adv. Regolith230–233.

Kitagawa, R., Koster, H.M., 1991. Genesis of the Tirschenreuth kaolin deposit in Germanycompared with the Kohdachi kaolin deposit in Japan. Clay Miner. 26, 61–79.

Lopez, T., Antoine, R., Darrozes, J., Rabinowicz, M., Baratoux, D., 2018. Developmentand evolution of the size of polygonal fracture systems during fluid-solid separation

in clay-rich deposits. J. Earth Sci. 1–16.Lu, Y., Wang, R., Lu, X., Li, J., Wang, T., 2016. Reprint of Genesis of halloysite from the

weathering of muscovite: insights from microscopic observations of a weatheredgranite in the Gaoling Area, Jingdezhen, China. Appl. Clay Sci. 119, 59–66.

Ma, Z., Li, X., Zheng, H., Li, J., Pei, B., Guo, S., Zhang, X., 2017. Origin and classificationof geothermal water from Guanzhong Basin, NW China: geochemical and isotopicapproach. J. Earth Sci. 28, 719–728.

Manju, C.S., Nair, V.N., Lalithambika, M., 2001. Mineralogy, geochemistry and utilizationstudy of the Madayi kaolin deposit, north Kerala, India. Clay Clay Miner. 49 (4),355–369.

Marfil, S.A., Maiza, P.J., Cardellach, E., Corbella, M., 2005. Origin of kaolin deposits inthe ‘Los Menucos’ area, Río Negro Province, Argentina. Clay Miner. 40, 283–294.

Murray, H.H., 1988. Kaolin minerals: their genesis and occurrences. In: Bailey, S.W. (Ed.),Hydrous Phyllosilicates (Reviews in Mineralogy 19). Mineralogical Society ofAmerica, Washington, DC, pp. 67–89.

Njoya, A., Nkoumbou, C., Grosbois, C., Njopwouo, D., Njoya, D., Courtin-Nomade, A.,Martin, F., 2006. Genesis of Mayouom kaolin deposit (western Cameroon). Appl. ClaySci. 32, 125–140.

Pruett, R.J., 2016. Kaolin deposits and their uses: Northern Brazil and Georgia, USA. Appl.Clay Sci. 131, 3–13.

Ramasamy, V., Sundarrajan, M., Suresh, G., Paramasivam, K., Meenakshisundaram, V.,2014. Role of light and heavy minerals on natural radioactivity level of high back-ground radiation area, Kerala, India. Appl. Radiat. Isot. 85, 1–10.

Savin, S.M., Epstein, S., 1970. The oxygen and hydrogen isotope geochemistry of clayminerals. Geochim. Cosmochim. Acta 34, 25–42.

Savin, S.M., Lee, M.L., 1988. Isotopic studies of phyllosilicates. Rev. Mineral. Geochem.19, 189–223.

Schroeder, P.A., Erickson, G., 2014. Kaolin: from ancient porcelains to nanocomposites.Elements 10, 177–182.

Schroeder, P.A., Pruett, R.J., Melear, N.D., 2004. Crystal-chemical changes in an oxida-tive weathering front in a Georgia kaolin deposit. Clay Clay Miner. 52, 211–220.

Scott, P.W., Bristow, C.M., 2002. Field excursion to study aspects of the St Austell Graniteand its related kaolinization and other mineralization. Geosci. South West England10, 370–372.

Scott, K.M., Pain, C.F., 2008. Regolith Science. CSIRO Publishing, Victoria, pp. 461.Stumm, W., 1992. Chemistry of the Solid-Water Interface. John Wiley & Sons, New York

(428 pp).Tardy, Y., Bocquier, G., Parquet, H., Millot, G., 1973. Formation of clay from granite and

its distribution in relation to climate and topography. Geoderma 10, 271–284.Taylor, G., Eggleton, R.A., 2001. Regolith Geology and Geomorphology. Wiley,

Chichester, UK (392 pp.).Thanachit, S., Suddhiprakarn, A., Kheoruenromne, I., Gilkes, R.J., 2006. The geochem-

istry of soils on a catena on basalt at Khon Buri, northeast Thailand. Geoderma 135,81–96.

Wang, L., Marks, M.A.W., Keller, J., Markl, G., 2014. Halogen variations in alkaline rocksfrom the Upper Rhine Graben (SW Germany): insights into F, Cl and Br behaviorduring magmatic processes. Chem. Geol. 380, 133–144.

Wang, X., Yang, Z., Chen, N., Liu, R., 2018. Petrogenesis and ore genesis of the lateYanshanian granites and associated porphyry-skarn W-Mo deposits from the Yunkaiarea of South China: evidence from the zircon U-Pb ages, Hf isotopes and sulfide S-Feisotopes. J. Earth Sci. 29, 939–959.

White, A.F., 1995. Chemical weathering rates of silicate minerals in soils. In: White, A.F.,Brantley, S.L. (Eds.), Chemical Weathering Rates of Silicate Minerals, Reviews inMineralogy. 31. Mineralogical Society of America, pp. 407–461.

Wilford, J.R., Searle, R., Thomas, M., Pagendam, D., Grundy, M.J., 2016. A regolith depthmap of the Australian continent. Geoderma 266, 1–13.

Wilson, I.R., 2004. Kaolin and halloysite deposits of China. Clay Miner. 39, 1–15.Zhao, L., Hong, H., Fang, Q., Yin, K., Wang, C., Li, Z., Torrent, J., Cheng, F., Algeo, T.J.,

2017. Monsoonal climate evolution in southern China since 1.2 Ma: new constraintsfrom Fe-oxide records in red earth sediments from the Shengli section, ChengduBasin. Palaeogeogr. Palaeoclimatol. Palaeoecol. 473, 1–15.

Zhou, C.H., Keeling, J., 2013. Fundamental and applied research on clay minerals: fromclimate and environment to nanotechnology. Appl. Clay Sci. 74, 3–9.

Zhou, X.M., Sun, T., Shen, W.Z., Shu, L.S., Niu, Y.L., 2006. Petrogenesis of Mesozoicgranitoids and volcanic rocks in South China: a response to tectonic evolution.Episodes 29, 26–33.

Q. Fang et al. Geoderma 337 (2019) 225–237

237