Comparison of Pretreatment Methods for Organic-matter Removal and their Effects on the Hydrogen Isotope (δ²H) Composition of Kaolinite

Abstract

The hydrogen isotopic composition (δ²H) of authigenic clay minerals has been used extensively in paleoclimate studies. The separation of clay minerals from sediments/soils, using various chemicals, is a prerequisite for isotope ratio measurements, where carbonate, Fe-(oxyhydr)oxides, and organic matter are removed successively from the sediments for a greater clay yield. The commonly adopted organic matter-removal method using hydrogen peroxide (H₂O₂) is thought to either alter directly the pristine δ²H values of the smectite clay minerals or to introduce organic hydrogen-bearing impurities through the ineffective removal of organic matter. The objective of the present study was to test whether H₂O₂ treatment can alter the δ²H values of kaolinite (Kln) by comparing two organic matter-removal methods, namely, H₂O₂ and disodium peroxodisulfate (Na₂S₂O₈) combined with a neutral buffer. In doing so, kaolinite-rich, old (~56 Ma) sediment samples and pure kaolinite internal laboratory reference materials were used to understand the effectiveness and suitability of the above-mentioned methods in clay-sample preparation for δ²H measurements. The δ²H values of the H₂O₂-treated aliquots show smaller δ²H values than those for the Na₂S₂O₈-treated aliquots. Estimated ambient water δ¹⁸O values (−4‰) from the Na₂S₂O₈-treated aliquots agreed well with the bio-phosphate (fish vertebrae) based environmental water δ¹⁸O estimation (−3.3‰). The present study indicated, therefore, that δ²H values obtained after Na₂S₂O₈ treatment are likely to be more realistic for paleoclimate reconstruction.

Introduction

The geochemistry of authigenic (both neoformed and transformed; Pozo & Calvo, 2018) clay minerals, particularly oxygen (δ¹⁸O) and hydrogen (δ²H some time referred to as δD) isotopic compositions, is a potential tool for quantitative reconstruction of past climates (Andrzejewski & Tabor, 2020; Bukalo et al., 2019; Gilg, 2000; Gilg et al., 2004; Girard et al., 2000; Oyebanjo et al., 2018; Tabor & Montañez, 2005). Complete removal of organic matter from the soil/sediment samples without affecting the mineral phase is a prerequisite for studying the structure, physicochemical properties, and, most importantly, isotopic composition of clay minerals for paleoclimate study. The incomplete removal of organic matter from phyllosilicates hinders the determination of their mineralogy (Tributh & Lagaly, 1991), and isotope values (mainly δ²H, Hyeong & Capuano, 2000) may also affect inferences based on infrared (IR) and differential thermal analysis (Mitchell & Farmer, 1962). The well known organic matter-removal methods include treatment with hydrogen peroxide (H₂O₂, Jackson, 2005), disodium peroxodisulfate (Na₂S₂O₈) combined with various neutral buffers (Menegatti et al., 1999), sodium hypochlorite (NaOCl, Anderson, 1963), ozone, Br₂, low-temperature plasma ashing, and stepped-heating (for a summary see Gilg et al., 2004; Sheppard & Gilg, 1996). Among these methods, H₂O₂ treatment is the most popular and is used commonly during sample preparation for δ²H analysis (Andrzejewski & Tabor, 2020; Bechtel & Hoernes, 1993; Bechtel et al., 1999; Jackson, 2005; Mizota & Longstaffe, 1996; Oyebanjo et al., 2018). However, the efficacy of organic matter-removal by H₂O₂ has been questioned, especially for samples containing >1 wt.% total organic carbon (TOC; Hyeong & Capuano, 2000). Further, Hyeong and Capuano (2000) also suggested that H₂O₂ can alter significantly the δ²H values of clays, e.g. from ~−8‰ in pure smectites (e.g. smectite reference material H26) to ~−13‰ in smectite-rich shale samples (Hyeong & Capuano, 2000). Studies have shown that treating clays with Na₂S₂O₈ combined with various kinds of neutral buffers is more effective at organic-matter removal (>98 wt.% of the initial TOC, Meier & Menegatti, 1997) than H₂O₂ (Mikutta et al., 2005). An easier method of Na₂S₂O₈ treatment (using sodium bicarbonate as a buffer) showed that the δ²H values remained unchanged after leaching (with Na₂S₂O₈ only) for less susceptible illite and kaolinite but smectite was altered by up to ~−28‰ (Meier & Menegatti, 1997; Menegatti et al., 1999). No data exist on the effects of various organic matter-removal treatments for kaolinite, the most widely used clay mineral for paleoclimate studies. A possibility exists that H₂O₂ treatment may alter the δ²H values of kaolinite, as has been found for smectite (Hyeong & Capuano, 2000).The purpose of the present study was to investigate the effect of commonly used H₂O₂ and Na₂S₂O₈ treatments on δ²H values of kaolinite separated from the late Paleocene/early Eocene (~ 56–52 Ma) Cambay Shale samples.

Geology of the Study Area

The Cambay Shale sediments were deposited in India’s western margin rift basin. The rift basin formed when India was being separated from Madagascar between the early Jurassic and late Cretaceous. The late Paleocene–early Eocene (~56–52 Ma) Cambay Shale sediments, deposited in a marginal marine setup with an abundant terrestrial organic matter supply, directly overlie the late Cretaceous Deccan Trap flood basalt (see Samanta et al., 2013, 2016 for a detailed geological setup). The sediments were derived mostly from weathering of the Deccan Trap basalts, occurring below the thick canopy of the late Paleocene–early Eocene tropical rainforest (Bansal et al., 2022; Prasad et al., 2009). Despite their old depositional age, the Cambay Shale sediments were never buried deeply and remained close to the subsurface (<1–1.5 km depth), which is evident from the: (1) exceptionally well preserved original aragonitic shells of bivalves with intact outermost organic layer (periostracum); (2) low thermal maturity index (T _max values ranging between 400 and 423°C); and (3) preserved angiosperm biomarkers (see Dutta et al., 2012 for details).

Materials and Methods

Sediment samples were obtained from a well drilled near the village of Valia (latitude: 21°35′50″ N; longitude: 73°12′2″ E; see Samanta et al., 2013 and references therein for depositional age assignment). Shale samples from four discrete depths of the Cambay Shale were selected for kaolinite clay separation and hydrogen isotopic analysis. The shale samples were selected primarily based on their color (grayscale color) so that they cover a range of TOC values.

Separation of bulk clay from the sediment samples requires a few pretreatment steps to remove carbonate, iron (oxyhydr)oxide phases, and organic matter, which may act as binding agents (Mehra & Jackson, 1960). Sediment samples were treated successively with HCl (37%; Merck, India, EMPARTA® ACS grade), citrate-bicarbonate-dithionite (trisodium citrate [> 99%, Merck, India], NaHCO₃ [Merck, India, EMPLURA® grade], and sodium dithionite [purified, > 85%, Lobachemie, India], following Mehra & Jackson, 1960), and H₂O₂ (30%; Merck, India, EMPLURA®ISO grade) with repeated rinsing, centrifuging for removal of carbonate, iron (oxyhydr)oxide phases, and organic matter, respectively. Apart from the usual cleaning using neutral pH detergent (Extran®) and Elix® (Merck, 10—12 Mohm cm at 25°C) water, all laboratory glassware was cleaned with Milli-Q® (18.2 Mohm cm at 25°C) water before use, and the same water was used in sample and reagent preparation at various stages. The freshly prepared Milli-Q® water was used during the wet chemistry and never stored for >2–3 days.

Approximately 5 g of gently crushed samples (samples were not lithified) were treated with >100 mL of 0.5 N HCl at ~60°C in 250 mL borosilicate glass beakers until effervescence stopped. This de-carbonation process generally took a few hours, depending on the presence of carbonate types (Ca, Ca-Mg, Fe, or Mn). Once effervescence stopped, the samples were washed repeatedly with Milli-Q® water by centrifuging 4 or 5 times at 4000 rpm (15 cm rotor diameter, 2683 × g) for 10 min. The neutralized samples were then dried in a microprocessor-controlled hot air-circulation oven at 60°C, where the actual temperature never exceeded the set temperature by >2°C. A portion of each of the dried samples was then transferred to a clean 500 mL borosilicate glass beaker and treated with citrate-bicarbonate-dithionite (CBD) solution to remove iron (oxyhydr)oxide phases. The CBD solution was freshly prepared by adding 45 mL of 0.3 M sodium citrate solution, 5 mL of 1 M sodium bicarbonate solution, and 2 g of sodium dithionite to each 0.5 g sample. The whole CBD solution containing the samples was kept in a water bath at 80°C for a maximum of ~2 h with intermittent stirring. The color change of the sample from black/dark-gray to light gray was a good indicator of the completeness of the reaction; the time taken increased from ~2 to 4 h depending on the amount of iron (oxyhydr)oxide present in the samples (Meier & Menegatti, 1997; Menegattiet al., 1999; Mikutta et al., 2005). The samples were then allowed to cool to room temperature and centrifuged four or five times with Milli-Q® water at 4000 rpm (2683 × g) for 10 min. A portion of the dried sample was then treated with 30% H₂O₂ at room temperature (~25°C) to remove the organic matter. In doing so, a 250 mL beaker was wrapped in multiple layers of aluminum foil to prevent direct exposure to light and kept for a minimum period of 2 h, an intermediate period of 24 h, and a maximum period of 48 h, with intermittent shaking. The varied-time (2 to 48 h) H₂O₂ treatment was only conducted for the kaolinite internal laboratory reference materials. The core samples were treated for a total of 48 h.

As discussed above, several studies have suspected that H₂O₂ treatment may alter the δ²H values of the clay samples. Therefore, a portion of each of the CBD-treated samples was treated with sodium peroxodisulfate (Na₂S₂O₈, >95%, Merck, India) in combination with sodium bicarbonate (NaHCO₃ [Merck, India, EMPLURA® grade]) used as a buffer, in parallel with H₂O₂ treatment to remove organic matter. In doing so, each sample was dispersed in Milli-Q® water (70 mL of water for each 200 mg sample). A total of ~1:50 Na₂S₂O₈:OM mass ratio was slowly added to it, keeping the pH of the solution >7 by intermittently adding small aliquots of NaHCO₃ (NaHCO₃:Na₂S₂O₈ mass ratio ≈1.1) when the pH dropped below 5. This procedure was conducted at ~80°C for a maximum of ~1 h (see Meier & Menegatti, 1997; Menegatti et al., 1999 for details). However, except for the internal laboratory reference materials and one dark-colored sample (244 m), which were treated for ~1 h, the rest of the samples stopped changing color after ~20–30 min and were removed from the bath after ~30–40 min. After centrifuging four or five times at 4000 rpm (2683 × g) to remove sulfates, a few mL of 0.5 N HCl was added to the samples to remove any carbonates that formed during the reaction. The samples were again centrifuged at 4000 rpm (2683 × g) to remove any excess acid and salts.

After completion of HCl, CBD, and H₂O₂ or Na₂S₂O₈ treatments, the samples were transferred to a 250 mL beaker, and 5.5 g L^–1 sodium hexametaphosphate (Merck, India, 96%) was added to the samples for dispersion (Sperazza et al., 2004). A small amount of dispersed sample was poured into a 1 L borosilicate graduated cylinder with 1000 mL of Milli-Q® water for gravity settling. The depth of the water column, containing a particular size fraction, was calculated by ‘Stokes Law’ using Sedicalc (Krumm, 2006) software and later concentrated by a centrifuge method for making oriented glass slides. To check whether the grain size (in the clay size range) has any effect on the clay mineralogy, various size fractions of one Cambay Shale sample, i.e.: (1) bulk sample (having all size fractions between 2 and 0.2 μm); (2) 2 to 1 μm; (3) 1 to 0.5 μm; (4) 0.5 to 0.2 μm; and (5) < 0.2 μm were analyzed in a Rigaku desktop X-ray diffractometer operating with CuKα radiation between 2 and 60°2θ with a step size of 0.02°2θ and count time per step of 1 s (Fig. 1). The bulk clay samples containing the 2–0.2 μm size fractions were selected for the δ²H analysis. In addition, two kaolinite internal laboratory reference materials (BDLK2 and BDLK3) collected from a kaolinite deposit formed over a Precambrian basement rock were also processed. As mentioned, the reference materials were subjected to H₂O₂ treatment for a minimum of 2 h and a maximum of 48 h to assess its effect on δ²H values, if any. About 1–2 mg of vacuum-dried (70°C for 24 h) clay samples were packed into tin capsules and analyzed immediately by placing them onto the automated carousel atop a Temperature Conversion Elemental Analyzer (TC-EA; Thermo Scientific, Bremen, Germany; Fig. 2a; following Sharp et al., 2001). The sample capsule was plunged inside the graphite crucible housed in a glassy carbon reactor surrounded by a ceramic insulator tube under a continuously He-flushed (purity 99.999%, flow rate of ~100 mL min^–1) condition. The lower part of the reactor was filled with glassy carbon granules, pre-conditioned at ~1450°C. The base of the reactor was plugged with silver wool, which also adsorbed sulfur and other halogen gases produced during the pyrolysis. Gases produced by the high-temperature reduction were then allowed to pass through a gas chromatographic column (molecular sieve 5 Å) maintained at 80°C to separate other trace gases from the target H₂. The purified H₂ was introduced into a Delta^plusXP mass spectrometer (Thermo Scientific, Bremen, Germany) via a ConFlo interface housed at the National Facility for Stable Isotope Geochemistry, Department of Geology and Geophysics, Indian Institute of Technology (IIT) Kharagpur. Only ~10% of the H₂ from the reaction tube went into the open split (maintained at ~2 mL min^–1 He), which dilutes and sends the gas to the mass spectrometer, and the remaining ~90% of the gas was expelled through the vent. The ConFlo interface also generated reference H₂ pulses, so individual sample H₂ peaks could be measured against those of a reference gas (Fig. 2b). This procedure is comparatively fast, requires a small amount of sample, and provides 100% yield for hydrogen. Routine H₃ ⁺-correction was employed before analyzing each set of samples. Raw delta (δ) values were calculated with respect to a laboratory tank of H₂. The tank H₂ was used as a primary reference gas. Raw δ values were calculated using the following relationship:

(1) ${\delta }_X=\left(\frac{R_X}{R_R},-,1\right)\times 1000\left(i,n,,‰,,w,i,t,h,,r,e,s,p,e,c,t,,t,o,,V,S,M,O,W\right)$

where $R_X$ = D/H ratio in the sample and $R_R$ = D/H ratio in the reference (the δ²H of the laboratory tank was assumed to be 0‰).

Fig. 1 XRD data (CuKα) showing the presence of kaolinite only in all size fractions of a single Cambay Shale sample and δ²H values of bulk and the finest (<0.2 µm) size fractions treated with H₂O₂; note no appreciable difference in δ²H values between the size fractions studied

Fig. 2 a Schematic diagram of the high-temperature TC-EA reactor showing the flow of He carrier gas through the GC column into the mass spectrometer. b Typical chromatograms of the H₂ reference and sample gases

Before routine analysis of the clay samples, the reproducibility of the TC-EA-Delta^plusXP mass spectrometer system was monitored by analyzing the NBS-22 (oil) certified reference material. In doing so, many aliquots of NBS-22 were measured over a large range of H⁺ ion beam amplitude (mV), and the raw δ²H (vs. tank H₂) was estimated (Table 1; Fig. 3a). The raw δ²H values of NBS-22 exhibited good reproducibility (± 2‰) over a very large range of beam amplitude (mV) varying between ~2000 and ~31,000 mV, which is similar to those obtained elsewhere by both dual inlet and online continuous flow techniques (Sanyal et al., 2010; Sharp et al., 2001). Such good reproducibility of the NBS-22 δ²H values indicated the system’s stability over a wide range of ion currents. The reproducibility becomes better (~0.8‰) with a restricted/narrower range of H⁺ ion beam amplitude (mV), as was shown by Coplen and Qi (2010).

Table 1 Amplitude of the H⁺ ion beam and raw δ²H values vs. values from tank H₂ (raw)

Fig. 3 a The stability of the Elemental Analyzer-DeltaplusXP mass spectrometer system over a large range of ion-current (represented as collector voltage). b Results of δ²H stretching experiment using three reference materials having widely different δ²H values. Measured and calculated values for NBS-30 Biotite reference material treated as an unknown are also shown using dotted lines (see text for details; w.r.t. = with respect to)

Natural clays can have a wide range of δ²H values depending on the meteoric/environmental water hydrogen isotopic composition with which the clay had equilibrated. While glacial meltwaters at high altitudes or latitudes can be extremely depleted in deuterium, tropical waters, affected by evaporation, may be significantly enriched in deuterium (Dansgaard, 1964). Because the resistance of the amplifier of the mass spectrometer does not behave linearly with minor ion current, the system is required to be referenced with respect to reference materials of widely varying isotope compositions before measurement of any unknown samples (Coplen & Qi, 2010; Qi et al., 2014). During the present work, this was obtained by analyzing three reference materials having widely different δ²H values (Fig. 3b). These are NBS-22_VSMOW oil (− 116.9 ± 0.8‰; Coplen & Qi, 2010), AP-223_VSMOW (− 80‰), and AP-227_VSMOW (−49‰; internal cellulose nitrate reference materials obtained from Physical Research Laboratory, Ahmedabad). Using the slope and intercept obtained from recommended and measured δ²H values of the reference materials, NBS-22 − AP-223 − AP-227 (Fig. 3b), the final δ²H values (VSMOW) of multiple aliquots of ground biotite certified reference material NBS-30 were calculated as unknown samples. The δ²H values of the kaolinite samples were calculated in the same way as mentioned above. The accuracy of the analysis was checked by comparing the measured and recommended δ²H values of the primary NBS-30 certified reference material. Because δ²H values of NBS-30 show a grain-size dependency (see Qi et al., 2014 and references therein for details), the NBS-30 was powdered with the help of an agate mortar and pestle before being introduced into the TC-EA for high-temperature reduction. As observed by previous workers, the δ²H values of NBS-30 crystals showed a clear grain size-dependent trend, and the measured δ²H values became closer to recommended δ²H values (− 65.7 ± 2‰; Qi et al., 2014) only when the grain size was <74 µm (− 64‰, Qi et al., 2014). The measured values of NBS-30 (n = 3), as an unknown sample, yielded a δ²H value of −64.2 ± 0.3‰ (VSMOW) which agreed well with the recommended value of −65.7 ± 2‰ (see Qi et al., 2014). Hereafter, all δ²H values are reported with respect to VSMOW unless otherwise mentioned specifically.

Total organic carbon (in wt.%) of all decarbonated (HCl-treated), H₂O₂-treated, and Na₂S₂O₈-treated bulk clay samples (size fraction 2–0.2 μm) were measured in a Flash 2000 Elemental Analyzer (Thermo Scientific, Bremen, Germany) attached to a Thermo Delta V Advantage continuous flow isotope ratio mass spectrometer (Thermo Scientific, Bremen, Germany) at the Department of Geology and Geophysics, IIT Kharagpur, India (see Samanta et al., 2013, for details).

Results

The X-ray diffraction (XRD) data of all size fractions, i.e. (1) bulk sample (having all size fractions between 2 and 0.2 μm), (2) 2 and 1 μm, (3) 1 and 0.5 μm, (4) 0.5 and 0.2 μm, and (5) <0.2 μm suggested that irrespective of size, all size fractions represent kaolinite only (Fig. 1). The δ²H values of the bulk clay and the finest fraction (<0.2 μm) did not show an appreciable difference even when treated with H₂O₂.

Two samples of organic matter-free (below detection limit), pure kaolinite internal laboratory reference materials BDLK2 and BDLK3 showed similar δ²H values of −77 ± 2‰ and −80 ± 2‰, respectively, for H₂O₂-treated, Na₂S₂O₈-treated, and untreated samples (Table 2). No significant difference was observed even when H₂O₂ treatment duration was varied from a minimum of 2 h to a maximum of 2 days. The δ²H values of kaolinite extracted from the Cambay Shale using H₂O₂ treatment exhibited a range between −73 and −78‰ (average ~−75‰), whereas kaolinite extracted from the same samples using Na₂S₂O₈ treatment showed larger values, varying from −48 to −56‰ (average ~−53‰; Table 3; Fig. 4).

Table 2 δ²H values of kaolinite internal laboratory reference materials (BDLK2 and BDLK3) using various organic matter-removal techniques

Table 3 δ²H values (all values with respect to VSMOW) of kaolinite separated from Cambay shale using various organic matter-removal techniques and calculated meteoric water δ²H values under assumed equilibration temperatures of 28°C

The TOC, in % after HCl-assisted decarbonation, and Na₂S₂O₈ and H₂O₂ treatments are also provided

Fig. 4 Comparison of kaolinite δ²H_VSMOW (‰) values after treatments with H₂O₂ and Na₂S₂O₈. The total organic carbon (%) after HCl-assisted decarbonation ( ${\mathrm{TOC}}_{\mathrm{DC}}$ ), H₂O₂ ( ${\mathrm{TOC}}_{H_2{O}_2}$ ), and Na₂S₂O₈ ( ${\mathrm{TOC}}_{{\mathrm{Na}}_2{S_{s}O}_8}$ ) treatments are also shown for comparison

The TOC of the decarbonated samples varied from 0.12 to 5.15 wt.%. However, after organic-matter removal/oxidation, the TOC varied from 0.05 to 3.58 wt.% for the H₂O₂-treated samples and 0.03 to 0.07 wt.% for Na₂S₂O₈-treated samples (Table 3; Fig. 4).

Discussion

The δ¹⁸O and δ²H values of both detrital (having a nearby source; Oyebanjo et al., 2018) and authigenic (formed during the early diagenesis; Girard et al., 2000) clay minerals found in sediments can be excellent recorders of the ambient climate. However, depending on the source and mineralogy, the δ²H values of the clay minerals can show large variability. The general assumption is that the finer fraction (< 0.2 µm) of the clay represents the authigenic minerals formed during early water–rock interaction, whereas the coarser fraction generally represents the detrital fraction (Andrzejewski & Tabor, 2020; Bukalo et al., 2019; Oyebanjo et al., 2018; Tabor & Montañez, 2005). Therefore, a detailed size-fraction analysis of one representative sample was performed before the δ²H measurements to check whether the clays studied represent a mixture of both detrital and authigenic populations. All size fractions were characterized by kaolinite only (Fig. 1) and showed no appreciable difference in δ²H values between the bulk (2–0.2 μm) and finest (>0.2 μm) fraction, even when treated with H₂O₂. Hence, the 2–0.2 μm size fraction (i.e. bulk clay fraction) was selected for δ²H analysis for the rest of the samples.

Hyeong and Capuano (2000) and Menegatti et al. (1999) showed that both H₂O₂ and Na₂S₂O₈ treatment alter the δ²H values of their smectite reference materials. The authors suggested that this may be due to the exchange of H⁺ ions, adsorbed primarily on interlamellar surfaces of smectite, with H⁺ coming from H₂O₂ (intermediate cation in Fenton-like reactions, Mikutta et al., 2005 and references therein). The δ²H values of two pure kaolinite internal laboratory reference materials (BDLK2 and BDLK3) were measured without any pretreatment as well as after the H₂O₂ (after 2 h, 1 day, and 2 days of reaction time) and Na₂S₂O₈ treatments to evaluate the effect, if any, of the treatment, on the δ²H values (Table 2). No appreciable difference in δ²H values was observed between the treated and untreated kaolinite reference materials. The δ²H values obtained for these pure kaolinite reference materials subjected to these two treatments were −77 ± 2 and −80 ± 2‰, respectively, and are very similar within experimental error (Table 2). Even H₂O₂ treatment for 2 days did not alter the kaolinite δ²H values. Also important to note is that, before H₂O₂ and Na₂S₂O₈ treatment, the kaolinite internal laboratory reference materials were subjected to routine HCl and CBD treatment. One may assume, therefore, that, apart from H₂O₂ and Na₂S₂O₈, the HCl and CBD treatments did not alter the hydrogen isotopic composition of these kaolinite internal laboratory reference materials. The possibility exists that not enough free H⁺ was present during the HCl reactions (0.5 N HCl) to cause hydrogen isotopic exchange, i.e. all H⁺ may have been consumed during carbonate dissolution or the reaction rate was too slow to show any significant effect on the δ²H values.

When subjected to different pretreatment methods, however, a substantial difference in δ²H values was observed for the kaolinites separated from Cambay Shale. When comparing δ²H values of H₂O₂- and Na₂S₂O₈-treated kaolinite from these shale samples, H₂O₂-treated samples were ~20–30‰ less than the Na₂S₂O₈-treated aliquots (Fig. 4). TOC data highlight that the Na₂S₂O₈ treatment was more efficient for organic-matter removal (Table 3; Fig. 4). The δ²H values of the investigated shale samples, along with their TOC data, suggested that for sediment samples with high TOC (carbonaceous shale 244 m), the H₂O₂ treatment is not sufficient, resulting in partial retention of the organic phases, which may eventually contaminate the samples during their δ²H measurement. The present observation concerning the inefficient removal of organic matter during the H₂O₂ treatment has also been reported by many previous workers, e.g. Hyeong and Capuano (2000), Meier and Menegatti (1997), Menegatti et al. (1999), and Mikutta et al. (2005). Three possible reasons for the observed difference in the average δ²H values of ~22‰ between H₂O₂ and Na₂S₂O₈-treated samples could be: (1) the presence of residual organic matter after H₂O₂ treatment; (2) hydrogen isotopic exchange between H₂O₂ and kaolinite’s ‘outer hydroxyl’ group (Fig. 5; Ledoux & White, 1964); and (3) dissolution (and possible reprecipitation) of kaolinite during the H₂O₂ treatment (Fig. 5; Cama & Ganor, 2006; Chin & Mills, 1991; Mikutta et al., 2005; Yang & Steefel, 2008). So, a simple mass-balance estimation was performed to understand whether the presence of residual organic matter can cause the observed ~22‰ difference between H₂O₂- and Na₂S₂O₈-treated samples. In doing so, it was estimated from the chemical formula that kaolinite has a hydrogen mass fraction of 1.56 wt.% and organic matter (CH₂) has a hydrogen mass fraction of 14.29 wt.%. The organic-matter chemical formula, i.e. CH₂, was chosen on purpose to estimate the maximum possible depletion to be expected in δ²H values when samples are contaminated with organic matter. Any other natural soil/sedimentary organic matter, e.g. cellulose, lignin, etc., contains a much smaller amount of hydrogen, ~6%. If the clay δ²H value is ~−53‰ (average of the Na₂S₂O₈-treated samples) and the organic matter δ²H value is ~−120‰ ( $\Delta D_{C3\mathrm{Plant}\mathrm{OM}}^{\mathrm{Water}}$ ~−120‰, Leaney et al., 1985; ( $\Delta D_{\mathrm{sedimentary}\mathrm{organic}\mathrm{matter}}^{\mathrm{Water}}$ ~−80 to −110‰, Schimmelmann et al., 2006), the presence of 1 wt.% organic matter in the treated clay can shift the δ²H values by ~5–6‰. The observed difference of ~20‰ between H₂O₂- and Na₂S₂O₈-treated samples requires ~5 wt.% organic matter in the H₂O₂-treated samples. The TOC (wt.%) data suggested that the TOC of the H₂O₂-treated samples was < 1 wt.% (except for one sample), however. The presence of residual organic matter after H₂O₂ treatment cannot explain the observed difference in δ²H values. Interestingly, a study by Barnette et al. (2011) showed that a linear relationship exists between δ²H and δ¹⁸O values of H₂O₂ solutions purchased across the United States and in Mexico, which nearly parallels the Local Meteoric Water Line (LMWL) and the δ²H values can be as low as −235‰. So, a possibility exists that ²H-depleted H₂O₂ can cause a significant depletion in the clay δ²H values through hydrogen isotopic exchange of the kaolinite’s ‘outer hydroxyl’ group (Fig. 5a). However, the small ratio (~1:130; see Supplementary Material for detailed calculations) of the ~0.2 µm size kaolinite’s ‘outer hydroxyl’ and ‘inner hydroxyl’ groups can only cause a minimal shift (~−1.4‰) during the isotopic exchange between highly ²H-depleted H₂O₂ (−235‰), and most ²H-enriched Na₂S₂O₈-treated kaolinite (~−53‰). Therefore, following Mikutta et al. (2005), H₂O₂ treatment was suspected possibly to promote the organic-assisted dissolution of poorly crystalline minerals at low pH. During this dissolution (± reprecipitation) process, isotope exchange between mineral and H₂O₂ may cause the lowering of the δ²H values of kaolinite clays. Previous studies have shown that oxidation of organic matter (coal/lignite) using H₂O₂ produces a large quantity of low-molecular-weight organic acids, e.g. formic acid, acetic acid, glycolic acid, succinic acid, and malonic acid (Fig. 5; Doskočil et al., 2014; Miura et al., 1996). Because the low-molecular-weight organic acids are more efficient at dissolution and sometimes reprecipitation of silicates or kaolinite compared to the inorganic acids, even at low temperatures (Cama & Ganor, 2006; Chin & Mills, 1991; Yang & Steefel, 2008), the organic-assisted dissolution (± reprecipitation in selected cases) during the H₂O₂ treatment may be the reason for this discrepancy (Fig. 5). Na₂S₂O₈ treatment, on the other hand, is likely to be less harmful to the minerals (Mikutta et al., 2005) as it is performed at a neutral to basic pH and thus lacks free H⁺ to affect the δ²H of kaolinite..

Fig. 5 Schematic diagram showing a the projection of the kaolinite crystal structure on the (100) plane (modified after Ledoux & White, 1964) and b the proposed mechanism for the H₂O₂ oxidation of Brown coal and formation of low-molecular-weight organic acids (modified after Miura et al., 1996). Two important processes, i.e. hydrogen isotopic exchange between H₂O₂ and kaolinite’s “outer hydroxyl” group and dissolution (and possible reprecipitation) of kaolinite in the presence of low molecular-weight organic acids produced during the H₂O₂ treatment, are also shown

In a further attempt to understand which of the average kaolinite δ²H values (i.e. H₂O₂-treated [−75‰] or Na₂S₂O₈-treated [−53‰]) is more realistic, soil/environmental water δ²H values were estimated using a branched glycerol dialkyl glycerol tetraether-based paleotemperature estimation of ~28°C (Naafs et al., 2018) from nearby early Eocene lignite deposits at Matanomadh and Panandhro, and a hydrogen isotope fractionation factor of the kaolinite–water system provided by Gilg and Sheppard (1996; Table 3). The calculated soil/environmental water δ²H values for the H₂O₂ and Na₂S₂O₈ treated kaolinite were −45 and −22‰, respectively (Table 3). Using the relationship between the δ¹⁸O and δ²H values of global precipitation water, i.e. the global meteoric water line (Dansgaard, 1964), the estimated early Eocene equatorial precipitation δ¹⁸O for the H₂O₂- and Na₂S₂O₈-treated kaolinite are −7 and −4‰, respectively. Bera et al. (2010) reported a phosphate δ¹⁸O value of 18.1‰ from fish vertebrae recovered from the equivalent stratigraphic horizon (i.e. above the characteristic carbon isotopic excursion of the Paleocene–Eocene Thermal Maximum) of nearby Vastan Mine (see Samanta et al., 2016 for detailed stratigraphic correlation). Following the revised phosphate–water fractionation equation for biogenic apatite (Pucéat et al., 2010), the environmental (lagoonal) water δ¹⁸O value of −3.3‰ was estimated from the phosphate δ¹⁸O value (at 28°C). The estimated environmental water δ¹⁸O values from these two independent proxies, i.e. Na₂S₂O₈-treated kaolinite δ²H values (−4‰) and fish vertebrae phosphate (−3.3‰), agree well with each other. Alternatively, the environmental water δ¹⁸O value of − 7‰, obtained from the H₂O₂-treated kaolinite, requires an abnormally small d-excess (= δ²H – 8*δ¹⁸O; Dansgaard, 1964) value of ~−19‰. As such a low d-excess value is unrealistic for the equatorial marginal marine precipitation, the Na₂S₂O₈-treated kaolinite δ²H values are proposed to be more realistic.

An important observation is that such a difference in δ²H values between H₂O₂- and Na₂S₂O₈-treated samples was not seen for the kaolinite internal laboratory reference materials. The possibility exists that: (1) the very small organic matter content of these kaolinite reference materials was insufficient to cause organic-assisted dissolution (± reprecipitation); and/or (2) a better crystallinity (greater structural order) of the kaolinite internal laboratory reference materials, obtained from a kaolinite deposit formed over a Precambrian basement rock (Aparicio & Galan, 1999), prevented organic-assisted dissolution at low pH and possible disintegration/transformation (following Mikutta et al., 2005).

Based on these observations and also on the data of pretreatment experiments on kaolinite reference materials, the conclusion is that Na₂S₂O₈ treatment is a better method for both complete removal of organic matter and preservation of pristine δ²H values for paleoclimate analysis compared to H₂O₂ treatment.

Conclusions

The present study showed that the commonly used H₂O₂ pretreatment for extracting clay minerals from organic matter-rich sediments (e.g. carbonaceous shales) is inefficient at removing organic matter. The Na₂S₂O₈ treatment was more efficient. Further, the smaller δ²H values of the H₂O₂-treated aliquots compared to the Na₂S₂O₈-treated aliquots raised suspicion about the possible alteration of the pristine δ²H values during H₂O₂ treatment. Because H₂O₂ treatment of organic material-bearing sediment produces low-molecular-weight organic acids, which have the potential to cause organic-assisted dissolution (± reprecipitation) of the clay minerals, the possibility exists that H₂O₂ treatment can lead to alteration of the pristine isotopic composition. The excellent agreement between the estimated environmental water δ¹⁸O values from two independent proxies, i.e. biogenic phosphate and Na₂S₂O₈-treated kaolinite, further suggested that Na₂S₂O₈ treatment is less harmful to the clay minerals as the treatment is performed under neutral to mildly basic pH condition that helps to retain the pristine isotopic composition. The present study, therefore, recommends Na₂S₂O₈ treatment for clay mineral separation for isotopic study.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s42860-023-00221-z.