Geoderma 235–236 (2014) 19–29

Contents lists available at ScienceDirect

Geoderma

j ourna l homepage: www.e lsev ie r .com/ locate /geoderma

Silicon isotopes record dissolution and re-precipitation of pedogenic clayminerals in a podzolic soil chronosequence

Jean-Thomas Cornelis a,b,⁎, Dominique Weis a, Les Lavkulich c, Marie-Liesse Vermeire b,Bruno Delvaux b, Jane Barling a,1

a Pacific Centre for Isotopic and Geochemical Research, Department of Earth, Ocean and Atmospheric Sciences, University of British Columbia (UBC), 6339 Stores Road, Vancouver,BC V6T 1Z4, Canadab Soil Science and Environment Geochemistry, Earth and Life Institute, Université catholique de Louvain, Croix du Sud 2/L7.05.10, B-1348 Louvain-la-Neuve, Belgiumc Soil Science, University of British Columbia (UBC), 127-2357 Main Mall, Vancouver, BC V6T 1Z4, Canada

⁎ Corresponding author at: Earth and Life Institute (ELouvain (UCL), Croix du Sud 2, L7.05.10, 1348 Louvain-la-

E-mail address: jean-thomas.cornelis@uclouvain.be (J1 Now at Department of Earth Sciences, University of O

OX1 3AN, United Kingdom.

http://dx.doi.org/10.1016/j.geoderma.2014.06.0230016-7061/© 2014 Elsevier B.V. All rights reserved.

a b s t r a c t

a r t i c l e i n f o

Article history:Received 8 February 2014Received in revised form 20 June 2014Accepted 22 June 2014Available online xxxx

Keywords:PodzolSilicon isotopesSoil formationClay mineralsBiogeochemical cycles

By providing the largest part of the reactive surface area of soils, secondary minerals play a major role in terres-trial biogeochemical processes. The understanding of the mechanisms governing neo(trans-)formation of pedo-genic clay minerals in soils is therefore of the utmost importance to learn how soils evolve and impact thechemistry of elements in terrestrial environments. Soil-forming processes governing the evolution of secondaryaluminosilicates in Podzols are however still not fully understood. The evolution of silicon (Si) isotope signaturein the clay fraction of a podzolic soil chronosequence can provide new insight into these processes, enabling totrace the source of Si in secondary aluminosilicates during podzol-forming processes characterized by themobi-lization, transport and precipitation of carbon, metals and Si. The Si isotope compositions in the clay fraction(comprised of primary and secondary minerals) document an increasing light 28Si enrichment and depletionwith soil age, respectively in illuvial B horizons and eluvial E horizon. The mass balance approach demonstratesthat secondary minerals in the topsoil eluvial E horizons are isotopically heavier with δ30Si values increasingfrom −0.39 to +0.64‰ in c.a. 200 years, while secondary minerals in the illuvial Bhs horizon areisotopically lighter (δ30Si = −2.31‰), compared to the original “unweathered” secondary minerals in BC hori-zon (δ30Si = −1.40‰). The evolution of Si isotope signatures is explained by the dissolution of pedogenic clayminerals in the topsoil, which is a source of light 28Si for the re-precipitation of new clay minerals in the subsoil.This provides consistent evidence that in strong weathering environment such as encountered in Podzols, Si re-leased from secondary minerals is partially used to form “tertiary clay minerals” over very short time scales(ca. 300 years). Our dataset demonstrates the usefulness to measure Si isotope signatures in the clay fractionto discern clay mineral changes (e.g., neoformation versus solid state transformation) during soil evolution.This offers new opportunity to better understand clay mineral genesis under environmental changes, and theshort-term impact of the dissolution and re-precipitation of pedogenic clay minerals on soil fertility, soil carbonbudget and elemental cycles in soil–plant systems.

© 2014 Elsevier B.V. All rights reserved.

1. Introduction

Soil is a precious but threatened resource (Banwart, 2011). In orderto protect it for the future we need a better understanding of the soil-forming processes controlling the evolution of newly-formed minerals(secondary minerals). Soil formation progressively modifies parentrockmaterial and controls the pathways of primarymineralweatheringand secondary mineral synthesis in the clay fraction (Chadwick and

LI-e), Université catholique deNeuve, Belgium..-T. Cornelis).xford, South Parks Road, Oxford

Chorover, 2001). The secondary minerals consist of layer-typealuminosilicates (called pedogenic clay minerals) and Fe-, and Al-oxyhydroxides, both of which play a major role not only in soil fertility,but also in the transfer of elements and pollutants from land to oceangiven their high surface reactivity (Sposito, 2008). Moreover, thecapacity of charged mineral surfaces to form adsorption complexescan stabilize organic carbon (OC) in soils through the formation oforgano-mineral associations, partly controlling global C budget (Parfittet al., 1997; Torn et al., 1997).

The formation of secondary minerals and their evolution during pe-dogenesis have been studied for over a half century (Wilson, 1999). Theproportion and the chemistry of minerals in the clay fraction changewith soil evolution (Egli et al., 2002; Righi et al., 1999; Turpault et al.,2008). Some environmental changes (vegetation type, agricultural

20 J.-T. Cornelis et al. / Geoderma 235–236 (2014) 19–29

practices, land-use, climate and drainage) can amplify the modificationof clay mineralogy on very short time-scales (10–1000 years) (Caneret al., 2010; Collignon et al., 2012; Cornu et al., 2012; Mareschal et al.,2013). These rapid clay modifications occur in chemically reactive soilmicro-environments, i.e. the part of the soil influenced by roots andearthworms (Calvaruso et al., 2009; Jouquet et al., 2007), and can playa key role in geochemical balance of several minor and major elementsin soils and sediments (Michalopoulos and Aller, 1995; Velde andMeunier, 2008). However, the origin of elements involved in clayneo(trans-)formation is still not well understood.

Podzol, the focus of this study, is a type of soil that covers more than3% of the Earth's land surface. The low stock of weatherable minerals,the acidic conditions and complexing capacity of organic acids in the en-vironment where Podzols developed are responsible for mobilization,transport and precipitation of carbon (C), metals (Fe, Al) and silicon(Si) in the soil profile (Lundström et al., 2000). A fully developed Podzolconsists of a leached gray subsurface eluvial E horizon contrasting withthe accumulation of elements in the dark illuvial B horizons. The topsoilis characterized by the production of organic acids that form solubleorgano-metallic complexes enhancing weathering in the eluvial E hori-zon. This E horizon overlies the dark C-enriched Bh horizon and reddishFe-, Si-, and Al-enriched Bhs/Bs horizons (Lundström et al., 2000). Giventhe very acidic conditions in Podzols, besides theweathering of primaryminerals, secondary clay minerals can be dissolved in the podzolicweathering front (Ugolini and Dahlgren, 1987; Zabowski and Ugolini,1992), which describes the soil depth where minerals dissolve fasterthan they form. A podzolic soil chronosequence, i.e. in which all soil-forming factors remain constant except time; represents an ideal natu-ral system for the study of the effect of time on pedogenic clay mineralsbehavior in soils.

Stable Si isotopes fractionate during silicate weathering and the bio-geochemical Si cycling (Opfergelt et al., 2010; Ziegler et al., 2005), andas such provide a means of tracing the bio-physico-chemical processesin terrestrial environments (Cornelis et al., 2011). In addition to its in-corporation in the mineral structure during the formation of crystallinelayer-type aluminosilicates, poorly-crystalline aluminosilicates andpedogenic opal, monosilicic acid (H4SiO4) released into soil solutioncan also be transferred into the biosphere to produce biogenic opal(phytoliths) or be adsorbed onto secondary Fe oxy-hydroxides. The in-corporation of Si inmineral structures through neoformation of second-ary pedogenic and biogenic precipitates and its adsorption onto thesurfaces of Fe oxides are two processes favoring the retention of light28Si in soils and contributing to the enrichment of rivers in heavy 30Si(Delstanche et al., 2009; Georg et al., 2007; Opfergelt et al., 2006;Ziegler et al., 2005). Clayminerals can also be unstable in organic and in-organic acidic environments where they dissolve (Sokolova, 2013;Zabowski and Ugolini, 1992), and enrich soil solutions (Cornelis et al.,2010) and rivers (Cardinal et al., 2010) in light 28Si. The naturally occur-ring mass-dependent Si isotopic fractionation is induced by dissolution,precipitation and adsorption but not by complexation as chemical bind-ing of Si to organic matter is negligible (Pokrovski and Schott, 1998). Ithas also been demonstrated that the Si isotopic compositions of second-ary clayminerals relates to climatic gradient and its control on claymin-eralogy (Opfergelt et al., 2012). However Si isotopes have never beenused to better understand clay mineral modifications induced by soil-forming processes under identical geo-climatic conditions. The rapidmodification of clay mineralogy in Podzol is well documented (Caneret al., 2010; Egli et al., 2002; Righi et al., 1999), but the fate of Si releasedin soil solution after clay modification has not yet been studied, eventhough it is of crucial importance for identifying the sources controllingthe formation of pedogenic clay minerals in soils.

In this study, we aim to use Si isotope signatures of the clay fractionin a podzolic soil chronosequence for gaining better insights into theorigin of Si in pedogenic clay minerals.

To achieve this goal, we analyzed Si isotopes, elemental (Ge/Si, Al/Si,Fe/Si) ratios and determined clay fraction mineralogy for an age

sequence of four soil profiles undergoing podzolization (Cox Bay onVancouver Island, Canada) (Fig. 1) and for a single Podzol pedon(Gaume, Belgium). The Cox Bay chronosequence offers an opportunityto study the variation of Si isotopic composition and elemental ratiosof the clay fraction in the vertical pedogenic scale: E, Bh, Bhs, Bs, Bwand BC horizons, and in the horizontal time-dependent scale: durationof pedogenesis from 0 to 335 years. We used the Belgian Podzol as a“natural duplicate” in temperate climate to corroborate the processesdocumented in the soil samples from the Cox Bay podzolic soilchronosequence.

2. Materials and methods

2.1. Sample collection and location

We sampled a soil chronosequence undergoing podzolization in CoxBay (CB), on the west coast of Vancouver Island (British Columbia,Canada). At the Cox Bay study site, three main vegetative associationsare identified in the chronosequence. These correspond to Sitka spruce(Picea sitchensis) in the younger site (CB-120 years), and Sitka spruce(P. sitchensis) and salal (Gaultheria shallon) in the sites of 175 and270 years (CB-175 and -270 years). The oldest site (CB-335 years) ischaracterized by Sitka spruce (P. sitchensis), Douglas fir (Pseudotsugamenziesii), salal (G. shallon) and western sword fern (Polystichummunitum). Heavy mean annual precipitation (3200 mm) coupled withfrequent fogs and sea sprays ensure an abundance of moisture and nu-trients year round in this maritime temperate climate (Cfb: without dryseason and with warm summer; Peel et al., 2007). The Tofino AreaGreywacke Unit is the source of the beach sand parent material, fromwhich soils have developed in the age sequence (Singleton andLavkulich, 1987). Sampling sites were located along a transect (0–94m) perpendicular to the present shoreline (Fig. 1). Dendrochronologyand geomorphology established surface duration of pedogenesis rang-ing from 0 to 335 years for the four selected pedons. Tree ages were de-termined counting the tree rings in the increment bores. Assuming thatthe beach built towards the ocean in a configuration parallel to theexisting shoreline and that a linear deposition rate occurred with timebetween successive oldest trees, the rate of advance of the beach frontwas estimated to be 0.26 m per year. At this rate, the 13-m strip ofsand containing tree seedlings would have accumulated in approxi-mately 50 years (Singleton and Lavkulich, 1987). With soil develop-ment, there was progressive deepening and differentiation of genetichorizons during podzolization, resulting in soil classification (WorldReference Base for Soil Resources — WRB) that ranged from DystricCambisol at the youngest sites (CB-120 years; CB-175 years) to a PlacicPodzol at the oldest site (CB-335 years) (Fig. 1). The 335-year-old Pod-zol is characterized by the following soil horizon development: eluvialalbic E horizon (strongly weathered horizon)→ illuvial spodic Bh hori-zon (enriched in organic matter) → Bhs horizon (enriched in Feoxyhydroxides and organic matter) → Bs horizon (enriched in poorly-crystalline aluminosilicates and Fe oxyhydroxides) → Bw horizon(development of color and structure without illuvial accumulation ofmaterials)→ BC horizon (weakly colored and structured; little affectedby pedogenic processes).

The sampling area of the Podzol in Gaume (Belgium), ranging in al-titude from 300 to 350 m above sea level, has an annual rainfall of1100 mm and a mean annual temperature of 7.7 °C (Herbauts, 1982),and is also characterized by a maritime temperate climate (Cfb; Peelet al., 2007). The Podzol is located on the Lower Lias outcrop in South-east Belgium (Gaume). The bedrock (calcareous sandstone of LowerLias age) is covered by a two-layered sheet: an autochthonous sandylayer, formed by the dissolution of the calcareous bedrock, is overlaidby a mixture of this sandy material with loessic silt-sized particles.The Belgian Podzol developed under heather (Calluna vulgaris) is char-acterized by a similar morphological profile as the Podzol in Cox Bay

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150

40

20

30

50

CB-120 yrs CB-175 yrs CB-270 yrs CB-335 yrs

0

-0.75 m

10

0

0,26 m/year

Vertical scale (m)

Horizontal scale (m)

Dystric Cambisol Haplic Podzol Placic Podzol

BC

E

Bw

BC

E

Bh

Bw

BC

EBh

BhsBs

Bw

BC

C

CB-0 yr

Parent material

Beach

Fig. 1. Cross section of the Cox Bay study area showing site locations and soil horizons, depending on their respective age of soil formation: CB-0 year, CB-120 years, CB-175 years,CB-270 years and CB-335 years.

21J.-T. Cornelis et al. / Geoderma 235–236 (2014) 19–29

sequence (CB-335 years) with the following horizons: E–Bh–Bhs–Bs–Bw–BC.

2.2. Physico-chemical characterizations

The soil samples were air-dried, then sieved and homogenized. Thecontent of free iron oxides was assessed after selective dissolution ofFe oxides usingNa-dithionite–citrate–bicarbonate and ammoniumoxa-late–oxalic acid (Fedcb= crystalline Fe oxides, Feox=poorly-crystallineFe oxides). The content of Si bound to poorly crystalline aluminosilicatesand weakly-ordered Fe oxyhydroxides was estimated on fine earth byextraction with ammonium oxalate–oxalic acid (Siox). Al complexedwith organic ligands was assessed using the complexing agent Na-pyrophosphate at pH 10 (Alp). The total organic carbon (OCtot) contentwas measured on ground samples using CNS analyzer.

The clay fraction (b2 μm) was separated using a ‘clean procedure’without any oxidative treatment. Air-dried soil was dispersed in deion-ized water and sonicated. The suspension was then separated on a50 μm sieve, re-suspended in deionized water and sonicated and sieveduntil the supernatant was clear after sonication. The fraction retained inthe sieve was collected as the N50 μm sand fraction. Clay (0–2 μm) andsilt (2–50 μm) fractions were then collected by gravimetric sedimenta-tion after dispersion using an ultrasonic probe and Na+ as a dispersionagent.

2.3. X-ray diffraction patterns

XRD analyses were carried out on the clay-sized fraction (b2 μm) ofsoil horizons sampled in the Cox Bay chronosequence (120, 175, 270and 335 years), using CuKα radiation in a Bruker Advance diffractome-ter. After removal of the organic matter by treating the sample with 6%H2O2 at 50 °C, and removal of Fe-oxyhydroxides using dithionite–

citrate–bicarbonate, eight standard treatments were applied to deter-mine mineralogy of the clay fraction: K-saturation (KCl 1 N) followedby drying and heating at 20, 105, 300 and 550 °C, and Mg-saturation(MgCl2 1 N) followed by drying at 20 °C and saturation with ethylene-glycol (eg). XRD analysis was also performed on powder samples ofthe clay-sized fraction after removal of organic matter and Feoxyhydroxides but without any further treatment for quantifying min-eralogy of the clay fraction using the Siroquant software V4.0 (SietronicsPty Ltd), and for the following horizons: BC horizon (CB-120 years), Ehorizons (CB-175, 270 and 335 years), Bh horizon (CB-335 years) andBhs horizon (CB-335 years).

2.4. Isotopic and geochemical analyses

Silicon isotope compositions and elemental (Ge, Al, Fe and Si) con-centrations were measured on clay-sized fraction (b2 μm) extractedfrom all the horizons of the four soil profiles in Cox Bay (clay-CB120 years; clay-CB 175 years; clay-CB 270 years; clay-CB 335 years)and the undated podzolic soil profile in Gaume (clay-G), and also onparent material of soils in Cox Bay (sand fraction of the beach sand;Beach-CB 0 year). An alkaline digestion with 99.99% pure NaOH isused to transform solid samples into an aqueous HF-free solution(Georg et al., 2006). All dissolutions and chemical separations were car-ried out in Class 100 laminar flow hoods in Class 1000 clean labs, massspectrometric analyses were performed in Class 10,000 laboratories atthe Pacific Centre for Isotopic and Geochemical Research (PCIGR) atthe University of British Columbia (UBC). Al, Fe and Si contents of thedissolved NaOH fusions were analyzed by ICP-OES (Varian 725-ES)with Europium as the internal standard. For Ge measurements, the dis-solved NaOH fusions were dried and re-dissolved in 1% v/v HNO3 with10 ppb indium (In) for analysis by HR-ICP-MS (Element 2) in mediumresolution.

22 J.-T. Cornelis et al. / Geoderma 235–236 (2014) 19–29

The remaining dissolved NaOH fusion solution was purified for iso-topic analyses through cation exchange chromatography (Georg et al.,2006). The Si isotope compositions were measured on a Nu Plasma(Nu 021; Nu Instruments Ltd, UK) MC-ICP-MS in dry plasma modeusing type B cones and a Cetac Aridus II desolvating nebulizer system.Instrumental mass bias was corrected by simple sample-standardbracketing of measured Si isotope ratios, i.e. one sample measurementnormalized to the average of two bracketing NBS-28 standardmeasure-ments. Silicon isotopic compositions are expressed as deviations in 30Si/28Si relative to the NBS-28 reference standard using the delta (δ) permil(‰) notation: δ30Si = [(30Si / 28Sisample) / (30Si / 28SiNBS28)− 1] × 1000.Each samplewasmeasured at least twice during different analytical ses-sions. Silicon isotopic (δ30Si) values are reported as the mean ofreplicate isotopic analyses (n N 2) ± 2 standard deviations (SD). TheNBS-28 (quartz standard) which processed through the full analyticalprocedure, and analyzed over a period of 7 months during 5 data acqui-sition sessions gave a value of δ30Si = 0.01± 0.18‰ (2SD, n= 66). Ac-curacy and reproducibility were also checked on reference materials(diatomite and BHVO-2) at the beginning and at the end of each sampleseries. These gave values identical within error to previously publishedvalues: 1.24 ± 0.13‰ (2SD, n= 15) for diatomite and−0.29 ± 0.19‰(2SD, n = 6) for BHVO-2 (Reynolds et al., 2007; Savage et al., 2012).

3. Results

3.1. Soil mineralogy

The parentmaterial of the soil chronosequence (0–335 years) is CoxBay beach sand (Singleton and Lavkulich, 1987), which is comprised ofvery well-sorted glacial sands with little wearing off and smoothingsharp edges and corners. The primary minerals present in the beachsand C material identified by X-ray diffraction and microscopy arequartz, amphibole, pyroxene, olivine and feldspars, as well kaoliniteprecipitating in the dissolution pits of feldspars. The parent materialdoes not contain inherited clay minerals, except kaolinite present in

Table 1Summary of the major soil physical and chemical characteristics (for the fine earth b2 mm an

Horizon Depth pH Soil fractions Sioxa Sidcb Feo

Sand Silt Clay

cm % g.kg−1

Cox Bay 120 years (Dystric Cambisol)BC 0–75 5.9 99.2 0.6 0.3 0.1 0.4 1.

Cox Bay 175 years (Dystric Cambisol)E 0–3 5.4 90.2 7.1 2.7 0.1 1.1 1.Bw 3–44 5.8 99.0 0.6 0.4 0.1 0.3 1.BC 44–75 5.9 99.6 0.2 0.1 0.1 0.2 1.

Cox Bay 270 years (Haplic Podzol)E 0–7 4.6 90.8 6.1 3.1 0.1 0.8 1.Bh 7–23 5.1 97.2 1.7 1.0 0.2 0.4 2.Bw 23–57 5.3 97.4 1.8 0.8 0.2 0.4 2.BC 57–75 5.4 98.2 1.1 0.7 0.2 0.3 1.

Cox Bay 335 years (Placic Podzol)E 0–16 4.8 82.3 14.4 3.0 0.1 0.4 0.Bh 16–23 5.6 88.0 8.7 2.8 0.5 1.0 3.Bhs 23–24 Nd 90.0 6.8 2.9 1.1 1.6 21.Bs 24–28 5.1 94.9 4.2 0.9 2.9 1.0 3.Bw 28–60 5.1 96.1 2.4 1.4 2.5 0.9 2.

Gaume (Haplic Podzol)E 19–35 5.0 94.0 3.1 2.9 0.0 0.0 0.Bh 35–40 4.7 89.0 7.0 4.0 0.1 0.3 4.Bhs 40–47 4.8 90.0 6.0 4.0 0.3 0.3 5.Bs 47–58 5.1 91.6 3.4 5.0 0.5 0.4 0.BC 70–100 4.6 92.9 2.5 4.6 0.2 0.1 0.

Nd = not determined.a Dithionite- (dcb), oxalate- (ox) and pyrophosphate- (p) extractable contents of Fe, Al and Sib Total organic carbon.

the weathered feldspar. We observe an increase of oxalate-extractableSiox in Bhs, Bs and Bw horizons of the Podzol (CB-335 years)(Table 1). We also document a strong mobilization of Fe in Podzolafter 335 years, characterized by an accumulation of crystalline andamorphous Fe oxides in the Bhs horizon, which is related to an increaseof OC content. This co-accumulation of Fe oxides andOC is also observedin the Belgian Podzol in Bh and Bhs horizon. The content of clay-sizedminerals is quite constant in the Belgian Podzol while we observe an in-crease of clay content towards the topsoil in the Canadian podzolic soilchronosequence (Table 1). The content of clay-sizedminerals in the en-tire soil profiles increases over time in the chronosequence.

The mineralogy of the clay fraction in the Cox Bay podzolicchronosequence is dominated by quartz, amphiboles, chlorites, vermic-ulite, mixed-layers minerals (MLM), smectite, illite, and kaolinite andevolves depending on soil age and the development of soil horizons(Fig. 2). In the youngest soil profile (CB-120 years), the claymineralogyis characterized by the presence of quartz, Na-feldspar and amphibolesas primary minerals and kaolinite, chlorite and illite as pedogenic clayminerals (data not shown). XRD patterns display similar mineral com-positions in the E horizons of CB-175 years and CB-270 year profiles.In those soil horizons, peaks at 1.40, 1.00, 0.83 and 0.70 nm, correspondrespectively to chlorite, illite, amphibole and kaolinite (disappearanceof the 0.7 nm peak after K 550 °C treatment). A band at 1.40 nm(Mg20 °C treatment) that shifts to 1.60–1.70nmafterMg–eg treatmentdue to swelling indicates the presence of discrete smectite. In addition,the combination of the peaks at 1.40 nm afterMg-20 °C andMg–eg, andthe collapse of the peak from 1.10 to 1.00 nm due to the dehydrationafter a K-saturation followed by heating correspond to vermiculite. Fi-nally, the presence of a wide peak at 1.20 nm after Mg-20 °C treatmentthat shifts afterMg–eg treatment indicates irregularlymixed-layermin-erals (MLM).

In the CB-335 years profile, mineralogical differences were ob-served. In the E horizon, relative to the E horizons of CB-175 years andCB-270 year profiles, XRD patterns show a strong decrease of the abun-dance of kaolinite (the 0.70 nm peak has almost disappeared), absence

d the clay fraction b2 µm) of the investigated soils.

x Fedcb Alox Alp OCtotb Clay fraction

Si Al Fe Ge

% μg·g−1

7 1.8 0.7 0.5 9.5 15.8 8.6 12.0 2.9

4 2.8 0.4 0.3 35.2 22.0 8.3 8.1 5.77 2.0 0.9 0.7 4.3 13.5 7.4 10.5 2.00 1.3 0.6 0.4 2.7 16.1 8.9 9.4 3.1

2 2.5 0.6 0.4 13.3 23.3 9.1 5.6 8.93 2.6 1.3 0.9 16.1 15.5 8.7 12.1 3.01 2.3 1.1 0.8 10.4 16.1 8.8 12.0 2.87 2.2 1.4 1.0 7.6 13.1 10.0 11.3 2.4

2 0.5 0.6 0.5 10.7 26.5 10.5 1.9 12.93 4.4 8.8 5.0 36.8 16.9 14.3 6.2 5.85 44.0 5.2 4.4 17.8 6.2 9.4 30.2 4.67 4.0 7.8 1.3 5.2 14.2 19.6 9.3 3.01 2.9 6.4 1.1 3.8 13.2 20.3 7.5 3.8

1 2.1 0.04 Nd 1.3 12.6 7.6 10.9 2.86 16.7 1.3 Nd 14.4 11.0 7.8 14.5 2.95 16.8 2.0 Nd 6.1 7.7 9.6 19.7 1.57 6.8 2.2 Nd 3.6 9.9 12.7 12.9 1.91 2.8 0.6 Nd 0.7 9.8 13.9 13.8 1.4

.

A B

C D

CB-175 yrs : E

2 - theta [°]4 10

0.830.991.20

1.39

1.62

Mg eg

Mg 20°C

K 550°C

K 300°C

K 150°C

K 20°C

CB-335 yrs : E

2 - theta [°]4 10

0.700.830.99

1.39

1.19

1.64

Mg eg

Mg 20°C

K 550°C

K 300°C

K 150°C

K 20°C

CB-335 yrs : Bh

2 - theta [°]4 10

0.70

0.830.99

1.39

1.20

Mg eg

Mg 20°C

K 550°C

K 300°C

K 150°C

K 20°C

CB-270 yrs : E

2 - theta [°]4 10

0.830.99

1.39

1.19

Mg eg

Mg 20°C

K 550°C

K 300°C

K 150°C

K 20°C

Fig. 2.XRD patterns of the clay-sized fraction (b2 μm)of soils of the Cox Bay soil chronosequence after six treatments: K-saturation followed by drying at 20, 105, 300 and 550 °C, andMg-saturation followed by drying at 20 °C and saturation with ethylene-glycol. (A) CB-175 years E horizon, (B) CB-270 years E horizon, (C) CB-335 years E horizon, (D) CB-335 years Bhhorizon. Spacings of major reflections are in nanometers.

23J.-T. Cornelis et al. / Geoderma 235–236 (2014) 19–29

of chlorite (no peak at 1.40 nm after K treatments), and increase of therelative abundance of smectite compared to vermiculite (increase of thepeak at 1.60–1.70 nm and almost no peak at 1.40 nm after the Mg–egtreatment). In the Bh horizon relative to the E horizon of CB-335 yearprofile, XRD patterns show the presence of kaolinite and chlorite, ab-sence of smectite (no swelling after Mg–eg treatment), increase in theabundance of vermiculite, and a decrease of the abundance of MLM(smaller peak at 1.20 nm after Mg 20 °C treatment).

The mineralogy of the Belgian Podzol (Gaume) is compared to themineralogy of the Canadian Podzol. The primary minerals of the loesscontain quartz, feldspars, micas and small amounts of trioctahedralchlorites and amphiboles (Van Ranst et al., 1982). The mineralogy of

the clay fraction in the Belgian Podzol is comprised of vermiculite, smec-tite, hydroxyl-interlayered vermiculite, chlorite, MLM and kaolinite(Herbauts, 1982).

As we are not able to precisely quantify each type of 2:1minerals onthe powder of the clay fraction (chlorite, smectite, vermiculite, illite,MLM) with Siroquant software, we carried out the clay mineralogyquantification in the soil chronosequence by separating the mineralsin the clay fraction in 4 groups: quartz, amphiboles, kaoliniteand 2:1 minerals (Fig. 3). Compared to the mineralogy of BC horizon(CB-120 years) at the initial stage of soil formation (quartz = 15%,amphiboles = 63%, kaolinite = 5% and 2:1 minerals = 17%), thequantification of clay mineralogy indicates an increase of the relative

Fig. 3. Quantitative evolution of the mineralogy in the clay-sized fraction of the Cox Baysoil chronosequence. The clay-size mineralogy is comprised of primary minerals (quartzand amphiboles) and pedogenic clay minerals (kaolinite and 2:1 minerals). Chlorite,vermiculite, smectite, illite and mixed-layer minerals (MLM) are the 2:1 aluminosilicatesencountered in the podzolic chronosequence.

24 J.-T. Cornelis et al. / Geoderma 235–236 (2014) 19–29

abundance of kaolinite (+14%) and 2:1 minerals (+6%) in E horizon ofthe 175-year-old soil. Then we observe a strong decrease of the relativeabundance of kaolinite in older and more weathered E horizons:−12%in the 270-year-old soil and −18% in the 335-year-old soil, while therelative abundance of 2:1 minerals is constant between the two oldestsoils (=18%). In the Bh and Bhs horizons of the 335-year-old soil, wenote an increase of the relative abundance of kaolinite (+7% and+12%, respectively) compared to the stronger weathered E horizon(=1%). The evolution of primary clay-sized minerals is characterizedby a decrease of the relative abundance of amphiboles in the earlystage of soil formation (−21%), then by a relative increase of the abun-dance (+14 and+19%) in themoreweathered E horizons, which is re-lated to the decrease of kaolinite, while quartz content remains quiteconstant (=16 ± 2%) during pedogenesis.

3.2. Si isotopic modifications in the clay fraction over time

In our study, pedogenic clay minerals in the clay fraction of BC hori-zon are considered as “unweathered” secondary minerals compared topedogenic clay minerals in more weathered horizon (E, Bh, Bhs andBs) since BC horizon is not yet reached by the podzolic weatheringfront (Lundström et al., 2000). In the Cox Bay soil chronosequence, wetherefore compare the Si isotopic signatures of the clay fraction ineach soil horizon with those in the “unweathered” clay fraction in theBC pedogenic horizon.

In the Cox Bay chronosequence, Si in the “unweathered” clayfraction (BC horizon; δ30Si = −0.52 ± 0.16‰, 2SD, n = 3) isisotopically lighter compared to the primary lithogenic minerals in theparent beach sand material (C material; δ30Si = −0.27 ± 0.10‰, 2SD,n = 3) (Fig. 4A). In the early phase of soil formation, the difference ofSi isotope signature between the lithogenic primary minerals in thesand fraction of the C material and the clay fraction in BC material, 30ɛis −0.25‰ (min − max = −0.12 − 0.37‰). This is not the fraction-ation factor due to precipitation of pedogenic clay minerals as the clayfraction also comprises lithogenic primary minerals.

Relative to the “unweathered” BC clay fraction (δ30Si = −0.52 ±0.16‰, 2SD, n = 3), the clay fraction of the topsoil eluvial E horizonsshows depletion in light 28Si (i.e., less negative δ30Si values: from−0.33 ± 0.02‰ to−0.10 ± 0.22‰ Fig. 4B, C). The clay fraction in thesubsoil illuvial Bh–Bs horizons is isotopically lighter (i.e., enriched inlight 28Si) than “unweathered” BC clay fraction (δ30Si from −0.60 ±0.06‰ to −0.84 ± 0.08‰ ‰; Fig. 4B, C). The magnitude of lightSi depletion/enrichment in the clay fraction increases with soil age,with 30ΔSiE–BC varying from +0.20‰ (at t = 175 years) to +0.42‰

(at t = 335 years); and 30ΔSiB–BC varying from −0.17‰ (at t =175 years) to −0.32‰ (at t = 335 years).

A comparable depletion/enrichment in light 28Si in the clay fractionduring pedogenesis is found in the Belgian Podzol (30ΔSiE–BC =+0.29‰; 30ΔSiB–BC = −0.27‰) from a similar temperate climate butwith a different parent material and rainfall conditions (Fig. 4D).

3.3. Geochemical modifications in the clay fraction over time

As the clay fraction becomes relatively more depleted in Si, the clayfraction becomes more enriched in light 28Si (Fig. 5A). Our results showthat Si isotopic signature of the clay fraction becomes increasingly lightwith enrichment in Al (higher Al/Si ratio in the clay fraction) (Fig. 5B).

The enrichment in light 28Si (and the increase of Al/Si ratio)in the clay fraction also relates to an increase in the proportion ofpoorly-crystalline Si components in the clay fraction (estimated by theSiox/Siclay ratio). As the Si-bearing phases of the clay fraction accumu-lates poorly-crystalline aluminosilicates, the Si isotopic composition be-comes more enriched in light Si isotope (Fig. 5C). We observe also thatthe enrichment in light 28Si in the clay fraction is not systematically re-lated to a relative depletion in Ge, i.e. lower Ge/Si ratio (Fig. 5D).

4. Discussion

4.1. Evolution of clay-sized mineralogy

Different processes, such as transformation and neoformation, mod-ify the chemical composition of the clay mineral within soil profile andcontrol the clay content and mineralogy during pedogenesis. As wateracts tomediate chemical reactions and to transport reactants and prod-ucts from topsoil (Chadwick and Chorover, 2001), we observe thehighest content of pedogenic subproducts (clay-sized minerals) in thetop- and subsoils (0–24 cm). The depth where clay-sizedminerals con-centrate (~3%) increases over time, which highlights the deepening ofthe weathering front: 0–3 cm after 175 years, 0–7 cm after 270 years,and 0–24 cm after 335 years. We show that the chemical modificationsof clay mineral structure in the podzolic weathering front mobilize Al(and Fe) and Si from secondary minerals over time. The evolution ofAl/Si in the clay fraction substantiates the preferential mobilization ofAl, relative to Si, during the dissolution of secondary clay minerals, inparticular in the presence of organic acids with high complexing capac-ities, such as those encountered in Podzols (Sokolova, 2013; Stumm,1992). The clay mineralogy evolution (Fig. 2) in Podzols studied hereunder maritime temperate climate is very similar to the ones observedfrom postglacial moraines (Righi et al., 1999) and tills (Egli et al.,2002) in Switzerland. The aluminization of primary clay minerals,such as chlorites, leads to formation of irregularly-interstratified min-erals in the moderately acid B horizons. In the stronger weathering Esystem, Al-removal from interlayers by organic complexing agentsleads to the formation of vermiculite. Further alteration induces the for-mation of smectite-likeminerals in the E eluvial horizon. Finally, the Sioxcontent (Table 1) confirms that the formation of poorly-crystalline alu-minosilicates (ITM) occurs when the concentration of organic acids issufficiently low to allow the precipitation of Al with Si, as suggestedby Ugolini and Dahlgren (1987) in the fulvate bicarbonate theory ofpodzolization. The clay mineralogy evolves with increasing weatheringin the age sequence and formation of typical podzolic soil horizons(E, Bh, Bhs, Bs, Bw), which is in good agreement with the formation oftwo geochemical compartments during podzolization (Ugolini andSletten, 1991). The upper E-Bh compartment is controlled by organicacids as major proton donors and complexing metals, which leads todissolution of primary and secondary minerals. In the lower Bhs-Bscompartment, the absence of organic acids leads to a less aggressiveweathering system mainly controlled by inorganic acids (carbonic andnitric acids).

Fig. 4. Silicon isotopic signature (δ30Si‰; mean values± standard deviation represented by error bars) in the clay-sized fraction depending on soil ages and in primary lithogenicmineralsin the beach sand parental material. (A): 0- and 120-year-old soil fraction (δ30Si of primaryminerals in beach sand in black and δ30Si of the clay fraction of the 120-year-old BC horizon inblue), (B): 175- and 270-year-old clay fractions (175 years= redΔ; 270 years= green ◊), (C): 335-year-old clay fraction (purple○), and (D): clay fraction in an undated Belgian Podzol(brown□). After only 175 years (B), we observe the depletion in light 28Si in the clay fraction of the eluvial E horizon and enrichment in light 28Si in the clay fraction of deeper illuvial soilhorizon; respectively, relative depletion in light 28Si (+0.20‰) and relative enrichment in light 28Si (−0.17‰) compared to the original Si isotopic signature of the unweathered clay frac-tion in the BChorizon. The isotopic fractionation increasesover timewith anenrichment inheavy 30Si of+0.42‰ in the clay fraction of the E horizon anda concomitant enrichment in light28Si of−0.32‰ in the clay fraction of the Bhs horizon (after 335 years).We observe exactly the same tendency in the Belgian Podzolwith enrichment in light 28Si in the clay fraction of theBhs horizon of−0.27‰ compared to the unweathered clay fraction in BC horizon. (For interpretation of the references to color in this figure legend, the reader is referred to theweb ver-sion of this article.)

25J.-T. Cornelis et al. / Geoderma 235–236 (2014) 19–29

Four importantmineralogical evolutions are observed in the Cox Baysoil chronosequence, as a result of podzolization: (i) the neoformationof kaolinite, illite and chlorite from dissolution of primary minerals atthe very beginning of soil formation, (ii) the disappearance of kaolinitein the strongest weathered E horizon, then (iii) the increase of relativeabundance of kaolinite in Bh and Bhs horizons compared to E horizon(Fig. 3), and finally (iv) the accumulation of imogolite-type materialsin Bhs and Bs horizons (Table 1).

4.2. Dissolution and re-precipitation of pedogenic clay minerals duringpodzolization

Since the clay fraction of soils comprises aluminosilicates and Fe-,and Al-oxyhydroxides, Si in the clay fraction includes Si incorporatedin primary minerals (quartz and amphiboles), secondary minerals(kaolinite and 2:1 minerals) and Si adsorbed onto Fe oxyhydroxides.In the Bhs horizon of the 335-year-old soil, the high content of free Fe(Fedcb = 44 g·kg−1) is in the same order of magnitude than in aweathering sequence in Cameroon (20–85 g·kg−1) (Opfergelt et al.,2009), for which the variations of δ30Si values in the clay fraction dueto adsorption onto Fe oxides are known (Opfergelt et al., 2010). Wehave to take into account the pool of Si adsorbed onto Fe oxides in theclay fraction as this Si pool significantly influences the enrichment inlight 28Si in the clay fraction: the difference of the Si isotope signaturein the clay fraction of B horizons before and after dithionite-treatment(i.e., after the release of Si from the surface of Fe oxides) in the

Cameroon weathering sequence varies between 0.08 and 0.45‰(Opfergelt et al., 2010). However, all of the Fe in the Cameroonweathering sequence is in the clay fraction, while in the temperatesoils of the Cox Bay chronosequence, only 20% of the bulk Fe contentis in the clay fraction for Bhs horizon (=8.8 g·kg−1), where we observethe largest enrichment in light Si isotope. In eluvial E horizons, weobserve the largest depletion in light 28Si while the Fe content inthe clay fraction represents between 70 and 100% of the total Feconcentration in bulk soil (until 2.8 g·kg−1). The ratio of Fe oxides inthe clay fraction to Si content in the clay fraction is similar betweenBhs (14%) and E (13%) horizons, while the Si isotope composition inthe clay fraction follows opposite trends in these two horizons. As aconsequence, we assume that the δ30Si values of the clay fraction ofBelgian and Canadian temperate soils can be considered representativeof the Si isotopic composition of the primary and secondary silicates,and not significantly influenced by the fractionation of Si isotopesthrough adsorption onto Fe oxides. The role played by the Si adsorptiononto Fe oxides on Si isotope compositions of the clay fraction musthowever be further investigated.

It is well established that the preferential incorporation of light 28Siduring neoformation of secondary pedogenic minerals accounts fortheir isotopically lighter signature relative to primary lithogenic min-erals (Georg et al., 2009; Opfergelt et al., 2010; Ziegler et al., 2005).The Si isotope composition of the soil clay fraction depends on the de-gree of soil weathering and the evolution of the clay mineralogy(Opfergelt et al., 2010, 2012; Ziegler et al., 2005).

Fig. 5. Evolution of Si isotope composition with elemental composition (Si, Al, Ge) and the proportion of poorly crystalline Si (Siox/Siclay) in the clay fraction for the Cox Bay soilchronosequence (175-year-old soil = red Δ; 270-year-old soil = green ◊; 335-year-old soil = purple○) and for the Gaume Podzol (brown□). (For interpretation of the references tocolor in this figure legend, the reader is referred to the web version of this article.)

26 J.-T. Cornelis et al. / Geoderma 235–236 (2014) 19–29

Using the quantification of primary minerals (quartz and amphi-boles) and secondary minerals (kaolinite and 2:1 minerals) and the Siisotope signature of lithogenic primary minerals (−0.27‰), we cancompute δ30Si value of “unweathered” secondary clay minerals in theclay fraction of BC horizon (−1.40‰; Table 2). The isotopic fraction-ation factor between primary lithogenic minerals and secondary

Table 2Quantification of primary and secondaryminerals in the clay-sized fraction of the Cox Bay soil capproach.

Measured data

Primary minerals(% in the clay fraction)a

Secondary minerals(% in the clay fraction)

BC horizon(120 years)

78 22

E horizon(175 years)

59 42

E horizon(270 years)

75 25

E horizon(335 years)

81 19

Bh horizon(335 years)

72 28

Bhs horizon(335 years)

70 30

a Mineralogy of the clay fraction quantified using the Siroquant software V4.0; primar(vermiculite, smectite, illite, chlorite, mixed-layers minerals).

b The δ30Si of secondary minerals present in the clay fraction is computed as follows: δ30Simin

min II = secondary minerals and δ30Simin I = −0.27‰.c Si isotope discrimination between “unweathered” clayminerals in BC horizon and pedogen

δ30SiBC–δ30SiBh/Bhs.

pedogenic minerals (30ε = δ30Simin I − δ30Simin II) is therefore−1.13‰. Themass balance approach (Table 2) shows also a progressivedepletion in light 28Si in secondary minerals of the E horizon (from−0.51‰ to 0.64‰) and an enrichment in light 28Si in secondary min-erals of the illuvial horizons (until −2.31‰). In identical bio-geo-climatic conditions, the Si isotopic fractionation associated with the

hronosequence. The clay-sized quantification is then used for the Si isotopic mass balance

Computed

δ30Si(‰)in the clay fraction

δ30Si (‰)of pedogenic clay mineralsb

Δ30SiBC–x (‰)c

−0.52 −1.40 –

−0.32 −0.39 +1.01

−0.33 −0.51 +0.89

−0.10 +0.64 +2.04

−0.45 −0.92 +0.48

−0.84 −2.31 −0.91

y minerals = quartz + amphiboles; secondary minerals = kaolinite + 2:1 minerals

II = ((δ30Siclay fraction − %min I ∗ δ30Simin I) / % min II), wheremin I = primaryminerals,

ic clay minerals in the “x” horizon of interest (x = E, Bh or Bhs horizons): δ30SiE–δ30SiBC or

27J.-T. Cornelis et al. / Geoderma 235–236 (2014) 19–29

dissolution of primary lithogenic minerals and neoformation of second-ary pedogenic minerals should generate comparable Si isotopic signa-tures in the clay fraction in the entire soil profile with no evolutionover time given identical fractionation factor between the primary andsecondary Si pools. Here, we show that that the signature of secondaryminerals varies in the soil profile and the relative depletion/enrichmentin E and B horizons increases with time in the Cox Bay chronosequence.The dissolution of primaryminerals and precipitation of secondarymin-erals therefore cannot explain the increasing depletion/enrichment inlight 28Si in the clay fraction over time and with depth. This highlightsthat the evolution of δ30Si values in the clay fraction of the soil profilesobserved here rules out the weathering of primary minerals (lithogenicSi pool) as the sole source for the neoformation of secondaryminerals inthe clay fraction.

Germanium (Ge), a chemical analog of Si, generally follows similarinorganic geochemical pathways than Si (Froelich and Andreae, 1981).However, secondary pedogenic (clay) and biogenic (phytoliths) min-erals display contrasting Ge/Si ratios: neoformed clay minerals areenriched in Ge (higher Ge/Si) while biogenic silica polymerized inplants as phytoliths is depleted in Ge (lower Ge/Si) (Derry et al., 2005;Kurtz et al., 2002). Although there is a negative relationship betweenGe/Si ratios and δ30Si in the youngest soils (Cambisols) of the Canadiansoil chronosequence, the absence of a relationship between Ge/Si andδ30Si ratios in the oldest soil (Podzol) of the Canadian chronosequenceand in the Belgian Podzol (Fig. 5D) allows us to dismiss the dissolutionof phytoliths (biogenic Si pool) as a major source of Si for clayneoformation. This process would be characterized by enrichment inlight 28Si and depletion in Ge in secondary clay minerals relative tobeach sand parent material, as phytoliths are Ge-depleted (low Ge/Siratio) relative to primary minerals (Derry et al., 2005).

The mass balance approach (Table 2) shows that the enrichment inlight 28Si of secondary minerals of Bhs horizon (−2.31‰) compared tothe “unweathered” secondary minerals in the BC horizon (−1.40‰)partly explains the depletion in light 28Si of secondary minerals in theclay fraction of E horizon (+0.64‰) for the oldest soil (Podzol CB-335 years). Our data highlight that the isotopic fractionation due topreferential release of light 28Si during dissolution of secondaryminerals in the E horizon (Δ30SiE–BC = +2.04‰) partly accountsfor the enrichment in light 28Si during re-precipitation of new clayminerals in Bhs horizon (Δ30SiBhs–BC = −0.91‰). This combined withthe fact that kaolinite is progressively dissolved in the E horizon andis almost completely dissolved in the strongly weathered E horizon(CB-335 years) (Fig. 3), highlights that 28Si is redistributed in the soilprofile through re-precipitation of new pedogenic clay minerals deeperin the soil profile and leaching. As a part of Si precipitating during theneoformation comes from the dissolution of secondary clay minerals,we name those new clay minerals as “tertiary minerals”. This impliesthat the preferential lessivage of clay particles enriched in light 28Siand the resulting relative accumulation of primary clay-sized mineralsin topsoil cannot be responsible for the on-going enrichment in light28Si in the clay fraction. Indeed, the increasing enrichment in light 28Siin new tertiary minerals (tertiary kaolinite) in B horizons can only berelated to a Si source progressively enriched in light 28Si over time.Kaolinite seems to play a key role in the successive formation of clayminerals as the content of 2:1 clayminerals is quite constant during pe-dogenesis in the soil chronosequence (Fig. 3).

The preferential release and incorporation of light 28Si during disso-lution and re-precipitation of clay minerals in the pedogenic Si pool ac-count for the Si isotopic depletion/enrichment in the clay fraction overtime in the podzolic chronosequence. The preferential incorporation oflight 28Si during precipitation of Si released from the dissolution of ped-ogenic clay minerals (in E and Bh horizons) explains the increasing en-richment in light 28Si in newly-formed clay minerals (tertiary clayminerals in Bhs horizon) during podzolization. This is confirmed bythe fact that pedogenic clay minerals in E horizons are increasinglyheavier over time (Table 2), showing that the dissolution of pedogenic

clay minerals discriminate against the release of heavy 30Si as alreadydemonstrated for diatoms (Demarest et al., 2009) and crystalline basalt(Ziegler et al., 2005). Besides the lithogenic and biogenic Si pools, weprovide evidence that pedogenic Si pool is therefore involved in theneoformation of pedogenic clay minerals and as such in the evolutionof their Si isotope signatures (Fig. 6).

4.3. Implications for podzolization theory

For thefirst time,we document enrichment in light 28Si in secondaryclay minerals over time in a podzolic soil chronosequence. The highestenrichment in light 28Si and oxalate-extractable Siox in Bhs/Bs horizonsrelative to E/Bh horizons (Fig. 5C; Table 2) highlights that the dissolu-tion of secondary aluminosilicates in E/Bh horizons acts as a Si sourcefor formationof poorly-crystalline aluminosilicates (imogolite-typema-terials ITM) in Bhs/Bs horizons. The release of Si from the dissolution ofprimary and secondary clay minerals and precipitation of dissolved Siwith Al released by microbial decomposition from the organic ligands(Lundström et al., 1995) can explain the formation of ITM in Bhs/Bs ho-rizons (Ugolini and Dahlgren, 1987). During podzol development, ITMundergo additional dissolution for the re-precipitating Si as crystallinetertiary clay minerals in Bhs horizon. The evolution of Si isotopic signa-ture in pedogenic clay minerals of the podzolic soil chronosequencetherefore corroborates the process of dissolution and re-precipitationof aluminosilicate phases duringpodzolization (fulvate bicarbonate the-ory; Ugolini and Dahlgren, 1987). We can infer that low contents ofpoorly-crystalline ITM in the Bhs/Bs horizons play a key role in the evo-lution of Podzols and the progressive enrichment in light 28Si in pedo-genic clay minerals. The absence of ITM in the Bh horizon and thelighter δ30Si in Bhs/Bs indicates their high reactivity during podzoliza-tion, dissolving as organic-rich Bh horizon forms and precipitating asFe-, Si-, and Al-enriched Bhs/Bs horizons form. This is confirmed bythe high reactivity of ITM also reflected in Ge/Si and δ30Si patterns insoil solutions of the Santa Cruz soil chronosequence, which indicatesseasonal precipitation and dissolution of hydroxyaluminosilicates suchas allophane (White et al., 2012). The positive correlation betweenSiox/Siclay and δ30Si values in the clay fraction (Fig. 5C) highlights thatduring podzolization, pedogenic clay minerals become enriched inlight 28Si together with Al in the poorly-crystalline part of the clay frac-tion. Based on these findings, poorly-crystalline aluminosilicates can beregarded as a temporary reactive reservoir of light 28Si in Bs horizon.This reservoir acts as a source of light 28Si in tertiary crystalline claymin-erals, such as tertiary kaolinite, in Bhs horizon that will develop in thecurrent Bs horizon during podzolization. The dissolution and re-precipitation of pedogenic clay minerals are therefore an importantpodzol-forming process (Fig. 6).

4.4. Implications for tracing the effects of environmental changes on soils

In the Cox Bay soil chronosequence, we show that the production ofacidity (protons and complexing organic acids) in temperate forests andthe subsequent Podzol formation imply heavy 30Si enrichment in pedo-genic clay minerals of E horizons relative to the “unweathered” clayminerals in BC horizon; Δ30SiE–BC increasing from +1.01 to +2.04‰in ca. 200 years (Table 2). The preferential loss of light 28Si inweatheredclay minerals in E horizons compared to the “unweathered” clay min-erals in BC horizon is recorded in the Si isotope signature of pedogenicclay minerals on very short time-scale. Moreover, the Si isotope frac-tionation between the “unweathered” clayminerals in BC andpedogen-ic clay minerals precipitating in Bhs (Δ30SiBC–Bhs) of−0.91‰ highlightsthat a part of light 28Si released in topsoil is used for re-precipitation inthe subsoil (Table 2). As a consequence, Si isotope signatures in the clayfraction of soils should be tested in other systems to trace themodifica-tions of pedogenic clay minerals insoil–plant systems, such as devel-oped in highly weathered tropical and subtropical environments(Ferralsols, Lixisols, Nitisols, …), in frozen soils (Cryosols), in soils

Fig. 6. Conceptual representation of the contribution of Si released from the dissolution of primary and secondary Si pools (lithogenic, biogenic and pedogenic) to the re-precipitation ofnew “tertiary” clayminerals during podzolization. Phase I→ phase II (C→ BC)= transition from the parent Cmaterial to the pedogenic BC horizonwith neoformation of secondary clayminerals. Phase II= formation of typical podzolic soil horizons: E, Bh, Bhs, Bs and Bw. Phase II→ phase III = transition from young to older Podzol characterized by (i) the deepening ofthe E horizonwhere secondary clayminerals areweathered and enriched in heavy 30Si (EII→ EIII), and (ii) B horizons tertiary clayminerals re-precipitate and are enriched in light 28Si (BCII→ BhsIII).

28 J.-T. Cornelis et al. / Geoderma 235–236 (2014) 19–29

characterized by illuviation of clay minerals (Luvisols), in young soils(Cambisols) and in soils with high biological activity (Chernozems). Siisotope composition of pedogenic clay minerals can be useful to traceand quantify the impact of environmental changes (temperature, rain-fall, acid deposition, land use …) on pedogenic clay evolution. This iscentral to a better understanding of soil development and associatedterrestrial biogeochemical processes.

5. Conclusions

The process of dissolution of pedogenic clayminerals during podzol-ization is confirmed by the Si isotopic signature of the clay fraction in apodzolic soil chronosequence (Cox Bay, Vancouver Island). Our datasetshows Si isotopic, geochemical andmineralogical trendswith depth andas a function of pedogenic time, providing an orthogonal dataset whichsheds light on the origin and evolution of pedogenic clayminerals in theclay fraction. The depletion in light 28Si in pedogenic clay minerals intopsoil increases over time (from +1.01 to +2.04‰) and a part oflight 28Si released accounts for the relative enrichment in light 28Si inpedogenic clayminerals in subsoil (−0.91‰). This highlights that Si re-leased from the partial dissolution of secondary clay minerals in topsoilcontributes to the neoformation of tertiary clayminerals in subsoil. Claymineral dissolution has often been regarded as an irreversible process,while the increase of 28Si enrichment over time in the clay fraction doc-umented in this study indicates successive formation of clay minerals,which depends on the downward movement of the weathering frontin the soil. The continuous weathering of pedogenic clay minerals isan important process in the formation of Podzols as we show that theSi released in soil solution contributes to the reformation of clay min-erals deeper in soils over very short time scales (ca. 300 years). The re-cording of Si isotopic ratios in the clay fraction as a function of the age ofsoil formation is therefore an untapped resource for tracing pedogenicprocesses controlling the Si incorporation in pedogenic clay mineralsduring podzolization, and offering new perspectives for unraveling thegenesis of pedogenic subproducts in various soil types. This has impor-tant implications as the process of dissolution and re-precipitation ofpedogenic clay minerals would play a major role in several soil biogeo-chemical processes such as the retention of plant nutrients, the preser-vation of organic carbon frommicrobial decomposition, and the transferof elements and pollutants from land to ocean. Further investigationsare needed for quantifying the contribution of pedogenic Si pool to

newly-formed clay minerals (tertiary, quaternary …) compared to thecontribution of lithogenic and biogenic Si pools. Our dataset showsthat the Si isotope compositions of soils are influenced not only by bio-genic (phytolith formation/dissolution) and litho-, pedo-genic process-es (primary mineral dissolution and secondary mineral precipitation)but also by a more advanced weathering process, i.e. successive forma-tion of pedogenic clayminerals. This should be taken into accountwhenδ30Si values of the bulk soil and soil solutions are used for studying soilweathering degree and tracing dissolved and particulate Si transferredfrom soil–plant systems to the hydrosphere.

Acknowledgments

We thankA. Iserentant, C. Givron, P. Populaire, A. Lannoye, I. Caignet,P. Sonnet, M. Detienne (UCL), H. Schreier, S. Smukler, B. Kieffer (UBC),as well F. Talbot and A. Cornelis for field and laboratory assistance,V. Lai and M. Soon (UBC) for assistance in element analysis andK. Gordon (UBC) for assistance in Si isotopic analysis. We thankM. Brzezinski (University of California Santa Barbara) for providing usdiatomite. J-T.C. is supported by “Fonds National de la RechercheScientifique” of Belgium (FNRS; Postdoctoral Researcher Grant). This re-search was also supported by the “Fonds Spécial de Recherche” of theUCL and by D.W. NSERC Discovery Grant.

References

Banwart, S., 2011. Save our soils. Nature 474, 151–152.Calvaruso, C., Mareschal, L., Turpault, M.-P., 2009. Rapid clay weathering in the

rhizosphere of Norway Spruce and oak in an acid forest ecosystem. Soil Sci. Soc.Am. J. 73, 331–338.

Caner, L., Joussein, E., Salvador-Blanes, S., Hubert, F., Schlicht, J.-F., Duigou, N., 2010. Short-time clay-mineral evolution in a soil chronosequence in Oléron Island (France).J. Plant Nutr. Soil Sci. 173, 591–600.

Cardinal, D., Gaillardet, J., Hughes, H.J., Opfergelt, S., André, L., 2010. Contrasting siliconisotope signatures in rivers from the Congo Basin and the specific behavior oforganic-rich waters. Geophys. Res. Lett. 37, L12403. http://dx.doi.org/10.1029/2010GL043413.

Chadwick, O.A., Chorover, J., 2001. The chemistry of pedogenic thresholds. Geoderma 100,321–353.

Collignon, C., Ranger, J., Turpault, M.-P., 2012. Seasonal dynamics of Al- and Fe-bearingsecondary minerals in an acid forest soil: influence of Norway spruce roots(Picea abies (L.) Karst.). Eur. J. Soil Sci. 63, 592–602.

Cornelis, J.-T., Delvaux, B., Cardinal, D., André, L., Ranger, J., Opfergelt, S., 2010. Tracingmechanisms controlling the release of dissolved silicon in forest soil solutions usingSi isotopes and Ge/Si ratios. Geochim. Cosmochim. Acta 74, 3913–3924.

29J.-T. Cornelis et al. / Geoderma 235–236 (2014) 19–29

Cornelis, J.-T., Delvaux, B., Georg, R.B., Lucas, Y., Ranger, J., Opfergelt, S., 2011. Tracing theorigin of dissolved silicon transferred from various soil–plant systems towards rivers:a review. Biogeosciences 8, 89–112.

Cornu, S., Montagne, D., Hubert, F., Barré, P., Caner, L., 2012. Evidence of short-term clayevolution in soils under human impact. Compt. Rendus Geosci. 344, 747–757.

Delstanche, S., Opfergelt, S., Cardinal, D., Elsass, F., André, L., Delvaux, B., 2009. Silicon iso-topic fractionation during adsorption of aqueous monosilicic acid into iron oxide.Geochim. Cosmochim. Acta 73, 923–934.

Demarest, M.S., Brzezinski, M.A., Beucher, C.P., 2009. Fractionation of silicon isotopes dur-ing biogenic silica dissolution. Geochim. Cosmochim. Acta 73, 5572–5583.

Derry, L.A., Kurtz, A.C., Ziegler, K., Chadwick, O.A., 2005. Biological control of terrestrialsilica cycling and export fluxes to watersheds. Nature 433, 728–731.

Egli, M., Zanelli, R., Kahr, G., Mirabella, A., Fitze, P., 2002. Soil evolution and developmentof the clay mineral assemblages of a Podzol and a Cambisol in ‘Meggerwald’,Switzerland. Clay Miner. 37, 351–366.

Froelich, P.N., Andreae, M.O., 1981. The marine geochemistry of germanium: ekasilicon.Science 213, 205–207.

Georg, R.B., Reynolds, B.C., Frank, M., Halliday, A.N., 2006. New sample preparationtechniques for the determination of Si isotopic compositions using MC-ICP-MS.Chem. Geol. 235, 95–104.

Georg, R.B., Reynolds, B.C., West, A.J., Burton, K.W., Halliday, A.N., 2007. Silicon isotopevariations accompanying basalt weathering in Iceland. Earth Planet. Sci. Lett. 261,476–490.

Georg, R.B., Zhu, C., Reynolds, B.C., Halliday, A.N., 2009. Stable silicon isotopes of ground-water, feldspars, and clay coatings in the Navajo Sandstone aquifer, Black Mesa,Arizona, USA. Geochim. Cosmochim. Acta 73, 2229–2241.

Herbauts, J., 1982. Chemical and mineralogical properties of sandy and loamy–sandyochreous brown earths in relation to incipient podzolization in a brown earth–podzolevolutive sequence. J. Soil Sci. 33, 743–762.

Jouquet, P., Bottinelli, N., Lata, J.C., Mora, P., Caquineau, S., 2007. Role of the fungus-growingtermite Pseudacanthotermes spiniger (Isoptera Macrotermitinae) in the dynamic of clayand soil organic matter content, an experimental analysis. Geoderma 139, 127–133.

Kurtz, A.C., Derry, L.A., Chadwick, O.A., 2002. Germanium–silicon fractionation in theweathering environment. Geochim. Cosmochim. Acta 66, 1525–1537.

Lundström, U.S., van Breemen, N., Jongmans, A.G., 1995. Evidence for microbial decompo-sition of organic acids during podzolization. Eur. J. Soil Sci. 46, 489–496.

Lundström, U.S., van Breemen, N., Bain, D., 2000. The podzolization process: a review.Geoderma 94, 91–107.

Mareschal, L., Turpault, M.-P., Bonnaud, P., Ranger, J., 2013. Relationship between theweathering of clay minerals and the nitrification rate: a rapid tree species effect. Bio-geochemistry 112, 293–309.

Michalopoulos, P., Aller, R.C., 1995. Rapid clay mineral formation in Amazon Delta sedi-ments: reverse weathering and oceanic elemental cycles. Science 270, 614–617.

Opfergelt, S., Cardinal, D., Henriet, C., Draye, X., André, L., Delvaux, B., 2006. Silicon isotopefractionation by banana (Musa spp.) grown in a continuous nutrient flow device.Plant Soil 285, 333–345.

Opfergelt, S., de Bournonville, G., Cardinal, D., Andre, L., Delstanche, S., Delvaux, B., 2009. Im-pact of soil weathering degree on silicon isotopic fractionation during adsorption ontoiron oxides in basaltic ash soils, Cameroon. Geochim. Cosmochim. Acta 73, 7226–7240.

Opfergelt, S., Cardinal, D., André, L., Delvigne, C., Bremond, L., Delvaux, B., 2010. Variationsof δ30Si and Ge/Si with weathering and biogenic input in tropical basaltic ash soilsunder monoculture. Geochim. Cosmochim. Acta 74, 225–240.

Opfergelt, S., Georg, R.B., Delvaux, B., Cabidoche, Y.-M., Burton, K.W., Halliday, A., 2012.Silicon isotopes and the tracing of desilisication in volcanic soil weatheringsequences, Gaudeloupe. Chem. Geol. 326–327, 113–122.

Parfitt, R.L., Theng, B.K.G., Whitton, J.S., Shepherd, T.G., 1997. Effects of clay minerals andland use on organic matter pools. Geoderma 75, 1–12.

Peel, M.C., Finlayson, B.L., McMahon, T.A., 2007. Updated world map of the Köppen–Geiger climate classification. Hydrol. Earth Syst. Sci. 11, 1633–1644.

Pokrovski, G.S., Schott, J., 1998. Experimental study of the complexation of silicon and ger-maniumwith aqueous organic species: implications for germanium and silicon trans-port and Ge/Si ratio in natural waters. Geochim. Cosmochim. Acta 62, 3413–3428.

Reynolds, B.C., et al., 2007. An inter-laboratory comparison of Si isotope reference mate-rials. J. Anal. At. Spectrom. 22, 561–568.

Righi, D., Huber, K., Keller, C., 1999. Clay formation and podzol development from postgla-cial moraines in Switzerland. Clay Miner. 34, 319–332.

Savage, P.S., Georg, R.B.,Williams, H.M., Turner, S., Halliday, A.N., Chappell, B.W., 2012. Thesilicon isotope composition of granites. Geochim. Cosmochim. Acta 92, 184–202.

Singleton, G.A., Lavkulich, L.M., 1987. A soil chronosequence on beach sands, VancouverIsland, British Columbia. Can. J. Soil Sci. 67, 795–810.

Sokolova, T.A., 2013. Decomposition of clay minerals in model experiments and in soils:possible mechanisms, rates, and diagnostics (analysis of literature). Eurasian SoilSci. 46, 182–197.

Sposito, G., 2008. The Chemistry of Soils, Second edition. Oxford Univ. Press, New York.Stumm,W., 1992. Chemistry of the Solid–Water Interface. JohnWiley & Sons, Chichester,

Brisbane, Toronto, Singapore, New York.Torn, M.S., Trumbore, S.E., Chadwick, O.A., Vistousek, P.M., Hendricks, D.M., 1997. Mineral

control of soil organic carbon and turnover. Nature 389, 170–173.Turpault, M.-P., Righi, D., Utérano, C., 2008. Clay minerals: precise markers of the spatial

and temporal variability of the biogeochemical soil environment. Geoderma 147,108–115.

Ugolini, F.C., Dahlgren, R., 1987. Podzols and podzolization. The Mechanism of Podzoliza-tion as Revealed by Soil Solution StudiesINRA, Assoc. Franc. Etude Sol, Paris.

Ugolini, F.C., Sletten, R.S., 1991. The role of proton donors in pedogenesis as revealed bysoil solution studies. Soil Sci. 151, 59–75.

Van Ranst, E., De Coninck, F., Tavernier, R., Langohr, R., 1982. Mineralogy in silty to loamysoils of central and high Belgium in respect to autochtonous and allochtonous mate-rials. Bull. Soc. Belg. Géol. 91, 27–44.

Velde, B., Meunier, A., 2008. The Origin of Clay Minerals in Soils and Weathered Rocks.Springer-Verlag, Berlin Heidelberg.

White, A.F., Vivit, D.V., Schulz, M.S., Bullen, T.D., Evett, R.R., Aagarwal, J., 2012. Biogenicand pedogenic controls on Si distributions and cycling in grasslands of the SantaCruz soil chronosequence, California. Geochim. Cosmochim. Acta 94, 72–94.

Wilson, M.J., 1999. The origin and formation of clay minerals in soils: past, present andfuture perspectives. Clay Miner. 34, 7–25.

Zabowski, D., Ugolini, F.C., 1992. Seasonality in the mineral stability of a subalpineSpodosol. Soil Sci. 154, 497–504.

Ziegler, K., Chadwick, O.A., Brzezinski, M.A., Kelly, E., 2005. Natural variations of δ30Si ra-tios during progressive basalt weathering, Hawaiian Islands. Geochim. Cosmochim.Acta 69, 4597–4610.