Volcanic ash (<2 mm particle size) from explosive eruptions is often altered into smectite-rich clay rocks (bentonites) during early diagenesis (Christidis & Huff, Reference Christidis and Huff2009; Huff, Reference Huff2016; Hong et al., Reference Hong, Algeo, Fang, Zhao, Ji and Yin2019), after which chemical modification and progressive illitization take place during late diagenesis (Fortey et al., Reference Fortey, Merriman and Huff1996). Altered volcanic ash deposits of marine facies containing mainly mixed-layer illite-smectite (I-S) clays with >3.5% K2O are usually termed ‘K-bentonites’ (Merriman & Roberts, Reference Merriman and Roberts1990; dos Muchangos, Reference dos Muchangos2006), whereas those of non-marine facies containing >50% kaolinite are termed ‘tonsteins’ (Spears, Reference Spears2012). In order to avoid inaccurate characterization, we refer to both as ‘altered ash beds (or layers)’ regardless of their depositional setting (Kiipli et al., Reference Kiipli, Kallaste and Nestor2010). Altered ash layers frequently represent beds corresponding to events that can be correlated in regional chronostratigraphic studies, as they are deposited quickly over wide areas (Huff et al., Reference Huff, Merriman, Morgan and Roberts1993; Siir et al., Reference Siir, Kallaste, Kiipli and Hints2015; Schindlbeck et al., Reference Schindlbeck, Kutterolf, Freundt, Alvarado, Wang and Straub2016).
Previous investigations showed that the geochemistry of altered ash beds carries information regarding the type of source magma and the tectonomagmatic setting of the source area, which may be inferred using magmatic and tectonic discriminant diagrams (Huff & Türkmenoğlu, Reference Huff and Türkmenoğlu1981; Ver Straeten, Reference Ver Straeten2004). Alteration of volcanic ash may result in notable changes in the abundances of mobile elements such as K, Na, Ca, Mg and Si (De La Fuente et al., Reference De La Fuente, Cuadros, Fiore and Linares2000), whereas Al, Ti, high-field-strength elements such as Zr, Nb, Hf and Ta and rare earth elements (REEs) are nearly immobile and remain largely unaffected during diagenesis. For this reason, the concentrations and ratios of these elements can be used to draw inferences regarding the source-magma chemistry and tectonic setting of the altered ash beds (Zielinski, Reference Zielinski1985; Göncüoğlu et al., Reference Göncüoğlu, Günal-Türkmenoğlu, Bozkaya, Ünlüce-Yücel, Okuyucu and Yilmaz2016). However, the distributions of these so-called immobile elements in ash beds can be affected by depositional and diagenetic processes (e.g. through mixing with non-volcanic materials and alteration via certain diagenetic pathways; Ver Straeten, Reference Ver Straeten2008). Thus, when using trace elements as source indicators or correlation tools, it is essential to investigate whether the immobile element fraction of volcanic ash has been modified.
Titanium is less mobile during sedimentary and alteration processes than other major elements (Condie et al., Reference Condie, Boryta, Liu and Quian1992), making it useful as an indicator of the provenance of ash beds (Christidis, Reference Christidis1998; Saylor et al., Reference Saylor, Poling and Huff2005; Batchelor, Reference Batchelor2014). However, the mobility of Ti depends on the pH, salinity and redox state of the depositional environment. In particular, extreme pH conditions (e.g. high pH in alkaline lakes or low pH in swampy environments due to the decay of organic matter) may facilitate dissolution of Ti-bearing minerals such as anatase and thus loss of Ti from ash beds (McHenry, Reference McHenry2009). Modification of Ti distributions through such processes may make it difficult to accurately identify the source-magma type and tectonic setting of an ash bed (Clayton et al., Reference Clayton, Francis, Hillier, Hodson, Saunders and Stone1996).
In South China, ash deposits occur widely in Permian–Triassic boundary (PTB) successions deposited in various depositional settings. The PTB transition interval commonly contains from one to three ash beds in terrestrial successions and from five to ten ash beds in marine successions, with individual ash beds having thicknesses of 2–40 cm (Yu et al., Reference Yu, Broutin, Chen, Shi, Li, Chu and Huang2015). The wide occurrence of ash deposits in South China provides an opportunity to investigate the influence of depositional environment on Ti distributions in devitrified ash and the use of ash beds for source-magma discrimination. A recent study showed that marine PTB ash beds experienced no changes in Ti concentrations after deposition, yielding TiO2/Al2O3 ratios similar to those of their intermediate–felsic source magmas, whereas the Ti concentrations of terrestrial facies were commonly modified through reworking and diagenesis (Hong et al., Reference Hong, Algeo, Fang, Zhao, Ji and Yin2019). In the present study, we investigated the distribution of Ti in the extracted clay fractions and remnants of PTB ash samples in two terrestrial sections (Tucheng and Chahe areas) and one shallow-marine section (Yanlou area). The objectives of the present study were to determine the influence of depositional environment on Ti distributions and evaluate the robustness of immobile elements of ash beds as source indicators and correlation tools in various depositional settings.
Geological background
The stratigraphy, mineralogy and geochemistry of ash deposits in many of the PTB successions in South China have been investigated extensively and described in earlier studies (Peng et al., Reference Peng, Zhang, Yu, Yang, Gao and Shi2005; Yu et al., Reference Yu, Broutin, Chen, Shi, Li, Chu and Huang2015). Submarine (especially deep-water) weathering of ash materials did not yield any separation of Ti from other immobile elements, and the original distribution of Ti in the altered ash materials remained largely unchanged (Huff et al., Reference Huff, Merriman, Morgan and Roberts1993; Kiipli et al., Reference Kiipli, Hints, Kallaste, Verš and Voolma2017). Therefore, the present study was limited to PTB ash beds representing terrestrial (including both paludal and lacustrine) and shallow-marine facies. We collected representative ~500 g ash samples from the ash beds of the three PTB stratigraphic successions.
The Yanlou section (26°32.26′ N, 126°60.15′ E) is located near Guiyang City, Guizhou Province (Fig. 1). It represents a shallow-marine setting and contains six ash beds with thicknesses ranging from 1 to 10 cm. The ash beds occur within siliceous shale of the Upper Permian Dalong Formation and calcareous mudstone and marl of the Lower Triassic Shabaowan Formation (Hong et al., Reference Hong, Algeo, Fang, Zhao, Ji and Yin2019).
Study sections and sampling. (a) Palaeogeographical maps of the South China Craton in the latest Permian with section locations; shallow-marine: Yanlou; lacustrine facies: Chahe; paludal facies: Tucheng. (b) Sampling of altered ash beds around the PTB, the stratigraphic position of which is approximate in terrestrial successions, known as the ‘Permian–Triassic boundary stratigraphic set’ (PTBSS). Modified from Hong et al. (Reference Hong, Algeo, Fang, Zhao, Ji and Yin2019).
The Tucheng section (25°36.12′ N, 104°24.18′ E) is located in Panxian County, Guizhou Province, at a distance of ~250 km south-west of the Yanlou section (Fig. 1a). It represents paludal (marine–terrestrial transitional) facies and contains three ash beds with thicknesses ranging from 15 to 40 cm. The ash layers are interbedded with muddy siltstone layers of the Upper Permian Xuanwei Formation and Lower Triassic Kayitou Formation (Peng et al., Reference Peng, Zhang, Yu, Yang, Gao and Shi2005).
The Chahe section (26°42.08′ N, 103°47.36′ E) is located between the 31st and 32nd kilometre milestones on the township road between Heishitou and Haila, Weining County, Guizhou Province, at a distance of ~100 km north-west of the Tucheng section (Fig. 1a). It represents fluvial and lacustrine facies that contain two ash beds with thicknesses of 8–10 cm. The ash layers occur in the Upper Permian Xuanwei Formation, in which lacustrine ash beds are interbedded with coal measures, and in the Lower Triassic Kayitou Formation (Yu et al., Reference Yu, Broutin, Chen, Shi, Li, Chu and Huang2015).
Methods
X-ray diffraction analysis
Whole-rock samples were dried at 60°C overnight and then ground using a mortar and pestle to a ~200 mesh powder that was re-dried at 105°C for 3 h. For clay mineral analysis, the <2 μm clay fraction was extracted using the sedimentation method described by Jackson (Reference Jackson1978), leaving a coarser counterpart referred to as the ‘remnant’. The extraction procedure is described briefly as follows: a total mass of ~5 g was placed in a 1000 mL beaker and was then mixed with 1000 mL of distilled water and subsequently stirred for 2 h. The suspension was allowed to settle for 6 h and the <2 μm clay fraction was collected by centrifuging the upper clear solution, whereas the remnant was collected by centrifuging the lower part of the suspension. Both the clay fraction and the remnant were collected for X-ray diffraction (XRD) analysis. Oriented mounts were prepared by carefully pipetting the suspension onto glass slides that were then air-dried at room temperature.
The XRD measurements were performed with a Panalytical X'Pert PRO diffractometer. The instrument was operated at 40 kV and 30 mA, with Ni-filtered Cu-Kα radiation with a 1° divergence slit, 1° anti-scatter slit and 0.3 mm receiving slit. The XRD traces were recorded from 3 to 65°2θ at a scan rate of 4°2θ min–1 with a resolution of 0.02°2θ.
X-ray fluorescence analysis
Determination of major element compositions was performed for both the clay fractions and remnants using X-ray fluorescence (XRF) (Hong et al., Reference Hong, Fang, Wang, Churchman, Zhao, Gong and Yin2017). Briefly, preparation of fused sample pellets entailed: (1) addition of 5 g of dilithium tetraborate to 1 g of sample powder followed by homogenization; (2) addition of four drops of 1.5% LiBr followed by further homogenization; (3) addition of 0.5 mL of polyvinyl alcohol followed by a final homogenization for 10 min; and (4) fusion of this mixture using a hydraulic press and heating with a Philips Perl'X 3 automatic bead machine. Measurements were carried out using a Shimadzu XRF-1800 sequential XRF spectrometer. The relative standard deviations of major elements were generally <1%, and the detection limits of the major elements were ~0.01%. Loss on ignition (LOI) was measured as the difference in sample weight before and after heating at 1000°C. The chemical index of alteration (CIA) of altered ash materials was obtained using the following equation:
$${\rm CIA} = {\rm A}{\rm l}_2{\rm O}_3/\lpar {{\rm A}{\rm l}_2{\rm O}_3 + {\rm CaO}^\ast + {\rm K}_2{\rm O} + {\rm N}{\rm a}_2{\rm O}} \rpar \times 100\percnt $$where CaO* is the CaO in silicates (Nesbitt & Young, Reference Nesbitt and Young1982).
Inductively coupled plasma mass spectrometry analysis
The trace element and REE concentrations of the clay fractions and remnants were analysed using inductively coupled plasma mass spectrometry (ICP-MS), as described by Hong et al. (Reference Hong, Fang, Wang, Churchman, Zhao, Gong and Yin2017). Briefly, the sample digestion procedure entailed: (1) adding a few drops of ultra-pure water to 50 mg of powdered sample in a Teflon bomb, followed by mixing with 1.5 mL HNO3 + 1.5 mL HF and heating at 190°C in an electric oven for 48 h; (2) heating the Teflon bomb at 115°C and allowing it to evaporate completely, then adding 1 mL HNO3 to the dried residue and allowing it to be dissolved and further evaporated completely; (3) adding HNO3 to dissolve the residue in a Teflon bomb and heating to 190°C for 16 h; and (4) diluting the resulting solution to 100 mL by addition of a 2% HNO3 solution. Analyses were performed using an Agilent 7500a ICP-MS spectrometer, and the relative standard deviations were usually <10% for trace elements and <4% for REEs.
Microscopic observations
The occurrence of anatase in ash samples was observed using scanning electron microscopy (SEM) and high-resolution transmission electron microscopy (HRTEM). A small amount of powdered clay fraction was fixed on a sample holder and then coated with gold. The SEM analysis was performed using a Quanta200 scanning electron microscope equipped with an X-ray energy-dispersive detector. The SEM instrument was operated at an accelerating voltage of 15–20 kV and a beam current of 10–20 nA, with a resolution of 3.5 nm for the secondary electron image. Prior to HRTEM analysis, the powdered clay sample was immersed in methanol solution and dispersed ultrasonically for 15 min. The clay sample was collected using a copper net and then dried under infrared light. The HRTEM analysis was carried out using a JEM 2010FEF transmission electron microscope equipped with an energy dispersive spectrometer (EDS) system at a resolution of 1 Å and an accelerating voltage of 200 kV.
Results
Mineralogical composition of the clay fraction and remnant
The clay fractions and remnants in all samples contain the same assemblage of minerals but in various proportions. In addition, ash samples from different facies yield differing mineral assemblages (Fig. 2). The Tucheng samples consist mainly of kaolinite and mixed-layer kaolinite-smectite (K-S) minerals, with minor quartz, lepidocrocite and anatase. The mixed-layer K-S was identified in previous works (Hong et al., Reference Hong, Fang, Wang, Churchman, Zhao, Gong and Yin2017, Reference Hong, Algeo, Fang, Zhao, Ji and Yin2019). The characteristic peaks of kaolinite, lepidocrocite and anatase are markedly stronger, and that of quartz markedly weaker, in the clay fraction compared to the remnants. Furthermore, the peaks of the (1
$\bar{1}$0) diffraction bands of kaolinite are better separated and stronger in the XRD trace of the remnant compared with that of the clay fraction, indicating that the crystal order of kaolinite is higher in the remnant.
XRD traces of the clay fraction and remnant. I-S = mixed-layer illite-smectite; K = kaolinite; Q = quartz; F = feldspar; A = anatase; I = illite; T-1C, T-2C, T-3C, Ch-2C, Y-1C, Y-2C, Y-3C = clay fractions of Tucheng, Chahe, and Yanlou sections, and T-1R, T-2R, T-3R, Ch-2R, Y-1R, Y-2R and Y-3R are the remnant counterparts.
The Yanlou and Chahe samples have similar mineral assemblages consisting mainly of mixed-layer I-S minerals and minor quartz and anatase. In addition, sample Y-2 contains minor smectite and kaolinite. All remnants of the ash samples contain relatively large amounts of quartz compared to their clay fractions, as reflected in the greater XRD peak intensity. The consistent presence of anatase in the clay fractions suggests a close association of this mineral with clay minerals.
Microscopic characteristics of clay minerals and anatase in altered ash beds
There are notable differences in the micromorphology of the clay minerals between ash samples from different sedimentary environments. Clay particles of the paludal ash samples displayed mainly poorly developed pseudo-hexagonal kaolinite flakes, with irregular ragged outlines and a chemical composition consisting mainly of Si and Al based on EDS analysis. Small amounts of clay grains showed poorly defined and irregularly curled edges, indicating the possible presence of mixed-layer K-S minerals (Fig. 3a). Anatase grains were not readily identified by SEM due to their small particle size and anhedral morphology, and their recognition depended on EDS surface scanning. These observations yielded only a few small anatase grains (10–50 nm in size) in close association with clay minerals (Fig. 3a).
Microscopic images showing the presence of anatase in altered ash beds. (a) Authigenic anatase in a nano-sized grain adsorbed by clay minerals in Yanlou ash (SEM). (b) Detrital anatase particle in a remnant of Tucheng ash (HRTEM). K-S = mixed-layer kaolinite-smectite; Ka = kaolinite; An = anatase.
Owing to their greater thickness, anatase grains usually display a darker background than the surrounding silicate minerals and are thus easily identified using HRTEM, which was further confirmed by EDS analysis. In the clay fraction, anatase was observed as anhedral grains with diameters of 10–50 nm located on the rims of clay particles or closely mixed with clay minerals, whereas the remnant yielded detrital anatase grains with broken outlines and diameters >100 μm (Fig. 3b).
Major element compositions of altered ash beds
The major element compositions of the clay fractions and their remnants for ash samples from the different sedimentary facies are given in Table 1. Both the clay fractions and their remnants exhibit rather uniform geochemical compositions with only minor differences within a given sedimentary facies. However, large compositional differences are apparent between the ash beds from paludal facies at Tucheng relative to those of lacustrine facies at Chahe and shallow-marine facies at Yanlou. Paludal ash samples yield LOI values of 13.08–13.80% for the clay fractions and 11.52–11.95% for the remnants, which are notably higher than those for the clay fractions (8.99–11.29%) and remnants (6.44–11.06%) of shallow-marine and lacustrine samples. Paludal ash samples have generally smaller SiO2 contents for the clay fractions (43.07–45.63%) and remnants (48.67–50.85%) compared to the shallow-marine and lacustrine ash samples (clay: 49.02–54.74%; remnant: 51.01–65.98%). Paludal ash samples exhibit relatively large Al2O3 contents (clay: 29.22–30.41%; remnant: 26.47–27.94%) compared to the shallow-marine and lacustrine samples (clay: 21.06–25.29%; remnant: 16.27–25.81%). Paludal ash samples contain markedly more TiO2 (clay: 3.76–4.03%; remnant: 4.27–4.56%) compared to shallow-marine and lacustrine samples (clay: 0.21–2.03%; remnant: 0.26–1.78%) (Table 1). Paludal ash samples have significantly smaller K2O contents (clay: 0.90–1.01%, remnant: 0.81–0.91%) than shallow-marine and lacustrine samples (clay: 4.10–4.61%; remnant: 3.39–4.76%). Finally, paludal ash samples contain markedly less MgO (clay: 0.67–0.80%; remnant: 0.51–0.57%) than shallow-marine and lacustrine samples (clay: 1.00–3.24%; remnant: 0.70–3.20%).
Major chemical compositions of the whole-rock, clay-fraction and remnant samples from the altered ash beds (wt.%).
Whole-rock data from Hong et al. (Reference Hong, Algeo, Fang, Zhao, Ji and Yin2019).
Trace element and REE concentrations
The concentrations of trace elements and REEs in whole-rock ash samples and their clay fractions are listed in Tables 2 and 3. In general, the upper continental crust-normalized distribution patterns of trace elements of the clay fractions are similar to those of the corresponding whole-rock samples, characterized by notable losses of Cr, Ni, Sr, Ba and Tl (Fig. 4). However, the concentrations of the immobile elements Y and Zr increased slightly, whereas that of Nb decreased in the clay fractions relative to the corresponding whole-rock samples. The lacustrine sample yielded a slightly different pattern: slightly less Y but more Zr and Nb in the clay fraction relative to the whole-rock sample. The concentrations of Y, Zr and Nb are nearly the same in the clay fractions and whole-rock samples of the shallow-marine samples, except for a slight decrease of Nb in the clay fractions.
Trace element distributions of clay fractions and bulk samples of altered ash beds normalized to upper continental crust (UCC; values from Taylor & McLennan, Reference Taylor and McLennan1985).
Concentrations of trace elements and REEs of the whole-rock altered ash beds (ppm).
Carbonaceous Ivuna (CI) chondrite values from Taylor and McLennan (Reference Taylor and McLennan1985); δEu = EuN / ((SmN + GdN) / 2), where EuN, SmN and GdN refer to their chondrite-normalized values.
Concentrations of trace elements and REEs of the clay fractions extracted from the altered ash beds (ppm).
The ΣREE values of the whole-rock samples from the paludal facies are 573 ppm (T-1), 520 ppm (T-2) and 493 ppm (T-3), which are slightly lower than the values of 657 ppm (T-1), 611 ppm (T-2) and 480 ppm (T-3) of their clay fractions. The ΣREE value of 1238 ppm for the lacustrine sample is markedly higher than 805 ppm for its clay fraction. The ΣREE values of whole-rock samples of the shallow-marine facies are 353 ppm (Y-1), 233 ppm (Y-2) and 284 ppm (Y-3), which are comparable to those of 398, 220 and 274 ppm, respectively, for their clay fractions.
In general, the bulk altered ash beds and their clay fractions display similar chondrite-normalized REE distributions, with only slight differences in abundances (Fig. 5). The REE distribution patterns of altered ash beds and clay fractions exhibit a general trend towards enrichment of light REEs (LREEs) and relatively flat heavy REE (HREE) patterns, with a moderately negative Eu anomaly. Samples from the shallow-marine facies show a notably larger negative Eu anomaly, with δEu values of 0.39–0.70 for bulk ashes and 0.41–0.69 for their clay fractions, whereas samples from the paludal and lacustrine facies show only a small negative Eu anomaly, with δEu values of 0.83–0.87 for bulk ashes and 0.79–0.84 for their clay fractions (Table 3). Europium is soluble under strongly reducing diagenetic conditions due to the transformation of Eu3+ to soluble Eu2+ (Bau, Reference Bau1991). The larger negative Eu anomaly in shallow-marine ash beds suggests a relatively large loss of Eu in reducing diagenetic environments, whereas the small negative Eu anomaly in paludal and lacustrine ash beds reflects preferential accumulation of Eu under oxidizing diagenetic conditions.
REE distributions of clay fractions and bulk samples of altered ash beds normalized to Carbonaceous Ivuna chondrite compositions (values from Taylor & McLennan, Reference Taylor and McLennan1985).
Discussion
Immobile elements during the alteration of ashes
The geochemical characteristics of altered volcanic materials depend largely on the environmental conditions in which the devitrification of volcanic ash takes place (Hong et al., Reference Hong, Algeo, Fang, Zhao, Ji and Yin2019). In closed conditions such as in marine phreatic environments, alteration of volcanic materials does not lead to fractionation of the immobile elements, and consequently their ratios remain unchanged (Bertine, Reference Bertine1974). For this reason, ratios of immobile elements can be used for testing correlations and interpreting source magmas and tectonic environments (Pearce & Peate, Reference Pearce and Peate1995; Göncüoğlu et al., Reference Göncüoğlu, Günal-Türkmenoğlu, Bozkaya, Ünlüce-Yücel, Okuyucu and Yilmaz2016). The Nb/Y vs Zr/TiO2 diagram of Winchester and Floyd (Reference Winchester and Floyd1977) is widely used to interpret source magmas of volcanic ash beds (Huff & Türkmenoğlu, Reference Huff and Türkmenoğlu1981; Batchelor & Clarkson, Reference Batchelor and Clarkson1993; Kiipli et al., Reference Kiipli, Hints, Kallaste, Verš and Voolma2017). However, the relative mobility of these elements in altered ash beds during alteration can vary as a function of the depositional environment and the diagenetic pathway (Zielinski, Reference Zielinski1985; Clayton et al., Reference Clayton, Francis, Hillier, Hodson, Saunders and Stone1996; Spears, Reference Spears2012). In addition, although most altered ash beds are considered as originating from in situ weathering of fine-grained, air-fall, pyroclastic volcanic ash, volcanic ashes in lacustrine, paludal and shallow-marine environments are often subject to subaerial erosion, transport and re-deposition of ash, which may result in mixing with terrestrial materials (Naish et al., Reference Naish, Nelson and Hodder1993). These processes may accelerate the dissolution and alteration of volcanic ash and result in the removal of the most mobile components (Christidis, Reference Christidis1998; Arslan et al., Reference Arslan, Abdìoğlu and Kadır2010). The hydrological setting of an ash deposit influences leaching intensity and transport (e.g. with strong leaching and downward flushing in ombrogenous paludal environments; Kiipli et al., Reference Kiipli, Kallaste and Nestor2010; Özdamar et al., Reference Özdamar, Ece, Uz, Boylu, Ercan and Yanik2014). In marine environments, ashes are subject to reworking and redistribution by currents, which can change the ratios of immobile elements as the separation of minerals of various densities is closely related to ambient environmental energy (Ver Straeten, Reference Ver Straeten2008).
For ash deposits in open and exoreic lacustrine and paludal environments, immobile element ratios may be modified by weathering, reworking and diagenesis, and careful analysis is required to recognize the post-depositional history of the ash beds (Clayton et al., Reference Clayton, Francis, Hillier, Hodson, Saunders and Stone1996; Laviano & Mongelli, Reference Laviano and Mongelli1996). In general, the nature of weathering processes and, therefore, the mineralogical and geochemical characteristics of the altered volcanic materials are strongly dependent on depositional facies (Christidis, Reference Christidis1998; McHenry, Reference McHenry2009; Hong et al., Reference Hong, Algeo, Fang, Zhao, Ji and Yin2019). Paludal environments are usually rich in organic acids that promote ash alteration (Zielinski, Reference Zielinski1985), leading to intense leaching and the production of mainly kaolinite (Fig. 2). In the present study, paludal ash beds (Tucheng) exhibit intense weathering, as shown by CIA values of 96.78–97.02 and a dominantly kaolinitic mineralogy for whole-rock samples. In contrast, lacustrine (Chahe) and shallow-marine (Yanlou) samples show only intermediate weathering, with CIA values of 81.11–83.70 and a dominance of mixed-layer I-S clays (Fig. 2, Table 1).
All ash samples of a single PTB section tend to display similar REE distributions, whereas various patterns are observed for various sections, even within a single depositional facies (Hong et al., Reference Hong, Algeo, Fang, Zhao, Ji and Yin2019). The PTB volcanic ash beds spread through various depositional settings from deep-sea to terrestrial facies within a limited geographical area in western Guizhou, South China. It has been confirmed that the ash beds are the source of intermediate–felsic magmas, deriving from subduction-zone volcanic arcs along the margins of the South China–Indochina and/or South China–Panthalassa plates, which had been active during the Late Permian–Triassic period (Yin et al., Reference Yin, Huang, Zhang, Yang, Ding, Bi and Zhang1989; Wu et al., Reference Wu, Ren and Bi1990; Stampfli & Borel, Reference Stampfli and Borel2002; Isozaki et al., Reference Isozaki, Shimizu, Yao, Ji and Matsuda2007; Zhang et al., Reference Zhang, Yuan, Zhao, Tong, Yang, Yu and Shi2009; Gao et al., Reference Gao, Zhang, Xia, Feng, Chen and Zheng2013; He et al., Reference He, Zhong, Xu and Li2014). The well-correlated PTB volcanic ash beds of various sedimentary facies, with a distance of tens to hundreds of kilometres between different sections in western Guizhou, South China, are reasonably expected to originate from intermediate–felsic rocks. The REE distributions of altered ash beds with intermediate–felsic origins usually exhibit a strong negative Eu anomaly (He et al., Reference He, Zhong, Xu and Li2014); therefore, the different REE distribution patterns between the ash beds of the different sedimentary facies can probably be attributed to distinct weathering, reworking and diagenesis processes in the geological environments.
In the present study, the paludal samples contain mainly kaolinite and have high REE concentrations, and their REE distributions, which are characterized by a notably small negative Eu anomaly and right-leaning shape, differ from those of lacustrine and shallow-marine facies (Fig. 5). The lacustrine sample displays similar features to those of the paludal samples, but the former has notably higher ΣREE abundances (1105 ppm) compared to the latter (493–573 ppm) (Fig. 5). In general, except for in saline–alkaline conditions, REEs tend to be relatively immobile during alteration in most environments (Wood, Reference Wood1990; Kiipli et al., Reference Kiipli, Hints, Kallaste, Verš and Voolma2017), and the REE distribution patterns of ash beds are largely inherited from the source rocks due to the rapid devitrification of volcanic glass into clay minerals and the sorption processes (Millero, Reference Millero1992; Arslan et al., Reference Arslan, Abdìoğlu and Kadır2010; Obst et al., Reference Obst, Ansorge, Matting and Huneke2015). Thus, ash beds of marine environments exhibit generally the same REE distributions as their source materials (Wray, Reference Wray1995). Compared to the shallow-marine ash beds (Tables 2 & 3), the notably high REE abundances and distinctive distribution patterns with a markedly small Eu anomaly of the paludal and lacustrine ashes are probably attributable to concentrations of rare accessory minerals through reworking rather than to elemental mobility during alteration and diagenesis (Zielinski, Reference Zielinski1985; Ver Straeten, Reference Ver Straeten2008). This is because in paludal and lacustrine environments, the common anoxic porewaters may cause the reduction and transport of Eu during diagenesis due to their significant sediment organic matter content, leading to a negative Eu anomaly (Bau, Reference Bau1991). In addition, REEs exhibit similar geochemical behaviours in supergene environments and experience negligible fractionation during processes such as weathering, transport and deposition in fine-grained siliciclastic facies (Taylor & McLennan, Reference Taylor and McLennan1985). In the present study, the paludal ash beds exhibit lamination textures and mixed-layer K-S is present (Hong et al., Reference Hong, Fang, Wang, Churchman, Zhao, Gong and Yin2017), observations which confirm that these ash beds experienced subaerial weathering and reworking prior to burial and diagenesis.
Titanium concentrations and source-magma discrimination
In the present study, the whole-rock and clay fractions of a given ash sample generally display similar REE distribution patterns and only small differences in ΣREE, reflecting a relatively homogeneous distribution of REEs (Fig. 5). However, the relative concentrations of the trace elements Nb and Y, as well as those of Zr and TiO2, show notable differences in the clay fractions of the paludal and lacustrine ash samples, suggesting that these elements are heterogeneously distributed in various grain-size fractions of the sediment. In general, Y tends to be concentrated in the clay fraction relative to Nb (Table 3). Weathering of ash leaches soluble ions from glass phases undergoing dissolution and precipitates clay minerals that can have varying adsorption capacities for certain trace elements, thereby resulting in differential retention patterns among trace elements released from the parent ash (Brookins, Reference Brookins, Lipin and McKay1989; Millero, Reference Millero1992).
Alteration of volcanic ash involves the devitrification of ash to clay minerals and leaching of alkalis and silica but uptake of Mg, Fe and Ca during the weathering process (Christidis, Reference Christidis1998; Arslan et al., Reference Arslan, Abdìoğlu and Kadır2010), and the clay species yielded by the ash beds are largely facies-dependent (Hong et al., Reference Hong, Algeo, Fang, Zhao, Ji and Yin2019). Al2O3 and TiO2 are highly immobile chemical components during chemical weathering, and their concentrations will increase with greater degree of weathering due to the net loss of mobile components (Göncüoğlu et al., Reference Göncüoğlu, Günal-Türkmenoğlu, Bozkaya, Ünlüce-Yücel, Okuyucu and Yilmaz2016). Ash beds of the paludal Tucheng section consist mainly of kaolinite and K-S minerals, while those of the shallow-water Yanlou section and the lacustrine Chahe section consist mainly of mixed-layer I-S minerals (Fig. 2). The different clay-mineral assemblages between the PTB sections indicate their various intensities of chemical weathering, which result in the notable differences in Al2O3 and TiO2 concentrations due to the varying degree of leaching of the ash beds (Table 1). Although theoretically the TiO2/Al2O3 ratios of the altered ash beds are inherited from their source magma, the TiO2/Al2O3 values will tend to decrease somewhat as chemical weathering proceeds, as Ti exhibits relatively high mobility in comparison with Al during devitrification of ash to clay minerals (Hodson, Reference Hodson2002; Arslan et al., Reference Arslan, Abdìoğlu and Kadır2010).
However, in paludal and lacustrine environments, reworking and re-deposition of volcanic ash can potentially lead to an increase in Ti concentration, as an open environment favours the possible migration of very small and light authigenic silicate minerals and the relative accumulation of heavy Ti-bearing phases due to the higher-energy conditions (Ver Straeten, Reference Ver Straeten2004). The paludal and lacustrine ashes are characterized by distinctly higher TiO2 concentrations (Table 1). Reworking and re-deposition of the volcanic material can lead to mixing with detrital material of terrestrial origin (Clayton et al., Reference Clayton, Francis, Hillier, Hodson, Saunders and Stone1996; Ver Straeten, Reference Ver Straeten2008). Titanium-bearing accessory minerals such as anatase, rutile, brookite and ilmenite have relatively high densities (4.0–5.0 g cm–3) compared to clay minerals (2.0–2.8 g cm–3) (Berry et al., Reference Berry, Mason and Dietrich1983). As a consequence, reworking and redistribution of volcanic material are liable to mechanical sorting, and these processes can cause significant loss of clay minerals and concurrent enrichment of TiO2 in the ash bed. In magmas, Ti is generally incorporated in early-crystallized silicate minerals such as biotite, pyroxene and amphibole (Abdel-Rahman, Reference Abdel-Rahman1994). During devitrification of volcanic ash into clay minerals, Ti is released and forms fine-grained anatase, which usually precipitates in clay-mineral aggregates (Weaver, Reference Weaver1976; Laviano & Mongelli, Reference Laviano and Mongelli1996; Ece & Nakagawa, Reference Ece and Nakagawa2003). Therefore, in marine diagenetic environments, ash-sourced Ti is largely retained in altered ash beds, and no separation of Ti from Al is observed, preserving the initial TiO2/Al2O3 value of the ash bed (Bertine, Reference Bertine1974; Hong et al., Reference Hong, Algeo, Fang, Zhao, Ji and Yin2019).
Organic-rich paludal and lacustrine environments may enhance the mobility of Ti released from primary ash minerals through the formation of Ti(OH)4 colloids (Brookins, Reference Brookins1988; Knauss et al., Reference Knauss, Dibley, Bourcier and Shaw2001). However, Ti(OH)4 colloids tend to dehydrate when pH >5 and to precipitate as crystalline TiO2 (Cornu et al., Reference Cornu, Lucas, Lebon, Ambrosi, Luizão and Rouiller1999). As a consequence of this process, Ti-bearing gels often precipitate as an occlusional phase during authigenesis of kaolinite (Malengreau et al., Reference Malengreau, Muller and Calas1995), limiting the range of Ti mobility to short distances (i.e. from the nanometre scale to the domain size of detrital Ti-bearing minerals; Tilley & Eggleton, Reference Tilley and Eggleton2005). The paludal and lacustrine ash beds have high TiO2 contents (2.62–4.85%), and their TiO2/Al2O3 ratios (0.15–0.21; Table 1) are notably higher than the averages for intermediate and felsic source magmas (0.055 and 0.022, respectively) (Le Maitre, Reference Le Maitre1976). The marked increase in TiO2 relative to Al2O3 probably involves mixing with underlying sediments and mechanical sorting as a consequence of reworking and re-deposition, which would have caused substantial gains of TiO2 in comparison with Al2O3 in the lacustrine and paludal ash beds (Table 1) (Hints et al., Reference Hints, Kirsimäe, Somelar, Kallaste and Kiipli2008; Hong et al., Reference Hong, Algeo, Fang, Zhao, Ji and Yin2019).
The authigenic anatase within an altered ash bed is usually present as nano- to micro-sized grains in close association with clay minerals (Fig. 3a), whereas Ti-bearing accessory heavy minerals of detrital origin are usually coarse and significantly resistant to alteration. The TiO2 contents of the clay fractions of the paludal ash samples are mostly lower than those of their remnants (Table 2). The clay fractions consist of kaolinite and mixed-layer K-S, and the remnants mainly of kaolinite (Fig. 2). However, as reflected by the well-separated (1
$\bar{1}$0) band peaks, the kaolinite of the remnant fraction is more ordered than that of the clay fraction, consistent with a relatively large particle size. This suggests that Ti-bearing heavy minerals are concentrated in the remnants (Fig. 3b), whereas the clay fraction contains mainly of authigenic anatase grains closely associated with clay minerals (Fig. 3a).
In addition to alteration in diagenesis, the concentration of Ti-bearing heavy minerals in altered ash is also related to the transport of sediments (Schroeder & Shiflet, Reference Schroeder and Shiflet2000). The significant TiO2/Al2O3 ratios of lacustrine and paludal ash beds reflect mainly the modification of reworking and diagenesis. As mentioned above, the source magmas of ash beds in the PTB successions of South China contain intermediate–felsic source materials based on the consistent presence of hexagonal dipyramidal quartz, zircon and apatite, as well as the associated sintering glass spherules in the ash beds within PTB successions (Yin et al., Reference Yin, Huang, Zhang, Yang, Ding, Bi and Zhang1989; Wu et al., Reference Wu, Ren and Bi1990; Zhang et al., Reference Zhang, Yuan, Zhao, Tong, Yang, Yu and Shi2009; He et al., Reference He, Zhong, Xu and Li2014; Gong et al., Reference Gong, Huff, Hong, Fang, Wang, Yin and Chen2018). In the present study, the whole-rock samples of the paludal facies plot in the mafic magma field of a TiO2 vs Al2O3 discriminant diagram, whereas their clay fractions are projected in the intermediate magma region (Fig. 6). The notable decrease in the TiO2/Al2O3 values of the clay fraction samples is attributable to scavenging of most of the Ti of terrestrial origin through clay-mineral extraction, as neoformed anatase remains in the clay fraction and is associated with clay minerals. Therefore, the Ti content of the clay fraction may serve as a more robust indicator of the source magma of the volcanic ash bed than that of the whole-rock sample (Laviano & Mongelli, Reference Laviano and Mongelli1996). However, during early diagenesis of paludal and lacustrine ashes, the organic-rich environments may enhance the dissolution of Ti from fine-grained detrital Ti-bearing phases due to reworking, and thus increase the uptake of colloidal Ti by clay minerals (Yusoff et al., Reference Yusoff, Ngwenya and Parsons2013). Therefore, although clay fractions may scavenge larger particles of detrital Ti-bearing minerals, the Ti content of the clay fraction can only reduce the influence of detrital Ti-bearing phases in the discrimination of their source rocks. As is shown in Table 1, the clay fractions generally contain less TiO2 but more Al2O3 than the whole-rock samples due to the exclusion of possibly incomplete altering Ti phases of the source material and more enrichment in clay phases; in this case, the clay fractions may provide partial information regarding the geochemical affinities of the parent rocks. Thus, for marine altered ash beds, and especially for deep-water ashes, the immobile chemical components of the whole-rock materials could be used for discrimination of the source magmas (Hong et al., Reference Hong, Algeo, Fang, Zhao, Ji and Yin2019).
TiO2 vs Al2O3 discrimination plots of the clay fractions and bulk samples of altered ash beds.
The Nb concentrations of the clay fractions are generally lower than those of the whole-rock samples, especially for the paludal ashes (Table 2), mirroring the Ti concentration trends, which is consistent with Nb substituting for Ti in minerals such as ilmenite, rutile, anatase and brookite (Berry et al., Reference Berry, Mason and Dietrich1983). In a Nb/Y vs Zr/TiO2 diagram, the paludal ashes have whole-rock compositions that are projected in the alkaline basalt zone, whereas those of their clay fractions plot in the andesite region (Fig. 7). The clay-fraction results are more robust in that they are consistent with evidence from the associated phenocrysts (Zhang et al., Reference Zhang, Yuan, Zhao, Tong, Yang, Yu and Shi2009). Mechanical sorting can increase the TiO2/Al2O3 ratio and the Nb/Y value of an ash bed, thereby influencing source-magma inferences based on TiO2 vs Al2O3 and Nb/Y vs Zr/TiO2 discriminant diagrams. Thus, for ash beds that have undergone reworking prior to final deposition, which is common for paludal ashes and especially tonsteins (Pearce & Peate, Reference Pearce and Peate1995; Spears, Reference Spears2012), the relative concentrations of immobile elements may differ from those of the original volcanic ash. This can lead to incorrect inferences regarding source magmas based on discriminant diagrams (Clayton et al., Reference Clayton, Francis, Hillier, Hodson, Saunders and Stone1996). For this reason, we recommend the use of clay-fraction rather than whole-rock immobile element concentrations for source-magma analysis.
Zr/TiO2 vs Nb/Y discrimination plots of the clay fractions and bulk samples of altered ash beds. Bsn/Nph = basanite/nephelinite; Alk-Bas = alkaline basalt; Trach/And = trachyte/andesite; Com/Pant = comendite/pantellerite (discrimination diagram after Winchester & Floyd, Reference Winchester and Floyd1977).
Conclusions
Differences in the mineralogical and geochemical characteristics between the clay fractions and remnants of altered ash beds are due to environment-specific differences in weathering and reworking processes. Paludal and lacustrine ash beds are characterized by higher TiO2 contents, and the presence of detrital Ti-bearing heavy minerals in the remnant is indicative of terrestrial origin. Reworking and re-sedimentation of the original volcanic materials can lead to differential concentrations of the immobile elements, with heterogeneous Nb/Y and Zr/TiO2 ratios for various grain-size fractions of paludal and lacustrine altered ash beds. Nevertheless, immobile elements are usually retained by clay minerals of the ash beds due to adsorption, and authigenic anatase is present as nano- to micro-sized grains precipitated in clay aggregates. Extraction of the clay fraction from reworked ash beds can exclude detrital heavy minerals, allowing the concentrations of immobile elements in the clay fraction to serve as a more robust proxy for source magmas than those of the whole-rock sample. However, the organic-rich paludal and lacustrine environments may enhance the dissolution of detrital Ti-bearing phases and thus increase the uptake of Ti by clay minerals; therefore, the Ti content of the clay fraction may reduce the influence of detrital Ti-bearing phases in the discrimination of their source rocks owing to scavenging the larger detrital Ti-bearing minerals. By contrast, for marine ash beds, due to the exclusion of the possibly incomplete altering Ti phases of the source magma, Ti of clay fractions may provide partial information regarding the geochemical affinities of the parent rocks; thus, the immobile chemical components of the whole-rock ash beds could be used directly for source discrimination.
Acknowledgements
The authors thank Drs J.X. Yu and S.Q. Cao for assistance with sample collection, and Prof G.E. Christidis, Principal Editor, and two anonymous reviewers for their insightful reviews, comments and suggestions.
Financial support
This work was supported by the National Natural Science Foundation of China (Projects 41972040 and 41472041) and the Fundamental Research Funds for the Central Universities, China University of Geosciences (Wuhan) (No. CUG170106).
