Mg and Sr Isotopes in Cap Dolostone: Implications for Oceanic Mixing after a Neoproterozoic Snowball Earth Event

Abstract

The snowball Earth (SBE) describes a state of the Earth’s climate with global or near-global ice cover. The cap dolostone at the base of the Ediacaran successions serves as useful archives for studying environmental change during the Marinoan Snowball Earth deglaciation in Neoproterozoic. The characteristic compositions in dolomite provide critical information on continental weathering and coastal water mixing after glacial retreat. However, valid methods for pristine dolomite separation remain challenging. In this study, four selected cap dolostone samples from the base of the Ediacaran Lantian Formation were used for establishing a new 3-step leaching method, to remove the secondary calcite and other impurities before determination of δ²⁶Mg and ⁸⁷Sr/⁸⁶Sr in dolomite. Non-destructive Raman, X-ray diffractometer (XRD) and scanning electron microscopy (SEM) were used to examine the distribution of dolomite and minor calcite/silicate in each sample. Micro-drill powders before each extraction procedure were examined in weight loss and mineralogical compositions, as well as the chemicals in the leaching solutions. Potential diagenetic artifacts were evaluated using Sr/Ca, Mn/Sr, ⁸⁷Sr/⁸⁶Sr and δ²⁶Mg in solutions. By applying a simple two-end member mixing between the seawater and the silicate sources (R² = 0.48, n = 23), the down-core variations of δ²⁶Mg and ⁸⁷Sr/⁸⁶Sr in cap dolostone can be used to gain a better understand of the temporal weathering intensity changes, as well as the coastal oceanic mixing processes, after the Marinoan deglaciation.

1. Introduction

The Snowball Earth (SBE) in Neoproterozoic era, 1000~539 Ma, is one of the most unrest periods in the early Earth. Particularly, the global glaciations from ~720 to ~635 Ma, covered nearly the entire Earth surface [1,2], evident found in the equatorial regions [3]. The SBE events changed not only the ocean, continent and atmosphere environments [4], but also impacted the early life evolution history. The Marinoan diamictite has been one of the most attractive subjects; however, its δ¹³C remains contentious in timing and sedimentation enviroments. Melezhik et al. [5] and Alene et al. [6] studied the cap dolostone deposition and proposed processes of biotic productivity, gas discharge and weathering may influence the δ¹³C compositions in dolostone. The characteristic compositions in dolomite will provide critical information on continental weathering and coastal water mixing after glacial retreat [7]. One of the main features of the ocean circulation in the SBE aftermath is the stratification that develops as a consequence of the freshwater inflow during the deglaciation [8]. The SBE hypothesis and the gas-hydrate destabilization (GHD) hypothesis are two major models for the Neoproterozoic global glaciation events [4], and they are opposite models. The deglaciation scenes described in SBE model involve complicated geological and biological processes, including volatile released in mid-ocean ridges, gas exchanged with crevasses, acidity triggered carbonate dissolution, glaciers melting, and low albedo lead to green-house climate [1,4]. After deglaciation, the weathered silicates increased alkalinity to facilitate precipitation at low latitude ocean, as well as volcanic activity to cause possible depleted δ¹³C carbonates [7,9]. A large amount of hydrate preserved in the ocean or shallow sea-bed [10,11] and subsequent magmatic activity may cause hydrate decomposition [12,13,14] to warm climate and to make further decomposition, in addition to the aerobic oxidation caused carbonate dissolution. On the other hand, the increased temperature and O₂ consumption led to oxygen deficit, as well as anaerobic oxidation of sulfate-reducing bacteria in hypoxic environments, eventually caused massive carbonate precipitation:

CH₄ + SO₄²⁻ → HCO₃⁻ + HS⁻ + H₂O

2HCO₃⁻ + Ca²⁺ → CaCO₃ + CO₂ + H₂O

The CO₂ and HCO₃⁻ in the equations were derived mainly from methane and depleted in δ¹³C (−5~−60‰). The carbonate ions, however, may derive from various sources to cause its wide δ¹³C distributions. Cold-seeps in the Dousantuo Formation was explained and agreed with the textures of deposition [15,16]. A plumeworld scenario describes that continental deglaciation initiated before the warming and the associated chemical weathering processes [17]. The energy absorbed at low latitudes was several times higher than the polar region, where glacial melts stayed above saline water to cause stratification and micro-organism derived carbonates formation at low latitudes [17]. As the flow-pass depended upon stratification where the enhanced stratification sustains longer circulation [1,18] and further precipitation at low-salinity shallow ocean [17].

1.1. Implications of Sr and Mg Isotopes

Four stable isotopes of Sr in nature are ⁸⁴Sr, ⁸⁶Sr, ⁸⁷Sr and ⁸⁸Sr. Part of the ⁸⁷Sr is radiogenic, formed by β⁻ decay of ⁸⁷Rb and accumulated with time in rocks. The ⁸⁷Sr/⁸⁶Sr ratios are unique tracer for source identification and age determination [19]. Major Sr sources in the ocean include rivers, hydrothermal vents, and carbonate dissolution, where each input showed unique ⁸⁷Sr/⁸⁶Sr [20]. The riverine weathering source provides high ⁸⁷Sr/⁸⁶Sr, average 0.7119, compared with modern ocean of 0.7092. Mean ⁸⁷Sr/⁸⁶Sr in the hydrothermal fluids is 0.7036, exchanged with ambient rocks in mid-ocean ridges. The marine carbonate dissolution could supply similar old oceanic ⁸⁷Sr/⁸⁶Sr (~0.7087) and mixed subsequently with the modern seawater [21]. It is evident that vigorous hydrothermal activities have led to low ⁸⁷Sr/⁸⁶Sr in the past [22,23]. The sedimentary ⁸⁷Sr/⁸⁶Sr could provide stratigraphic correlation or to estimate each of multiple sources contributions. Particularly, the ⁸⁷Sr/⁸⁶Sr ratio won’t be affected by chemical changes, evaporation, or bio-synthesis, and thus can reflect faithfully of different mixing scenarios.

The Mg on Earth distributes widely among various reservoirs. It has three natural stable isotopes, ²⁴Mg, ²⁵Mg and ²⁶Mg, with relative abundance of 78.99, 10.00 and 11.01%, respectively [24]. Part of the ²⁶Mg is radiogenic, produced by ²⁶Al decay, half-life of 0.73 Ma [25], produced only during the early solar system formation. The mass differences between Mg isotopes is rather large, e.g., ~8% for ²⁴Mg and ²⁶Mg, and could cause significant fractionation in geological processes. Previous techniques, however, suffered its poor precision for detecting Mg isotopic variation in geological specimens. Recent advance in mass spectrometry, especially in multi-collector inductively coupled plasma mass spectrometry (MC-ICP-MS), has improved >10 times precision (2SD) [26,27,28,29,30]. Later many studies have focused to develop the Mg tracer and to build up Mg isotopes data banks. These geochemical studies include in meteorites and various geological reservoirs [28,30,31,32,33], in partial melting at subduction zone, in weathering processes and other planetary systems [27,28,30,31,34,35,36,37], material recycles in Earth systems [29,37,38] and fractionation in various experimental conditions [39,40,41,42]. However, only scarce results on the Mg isotopic ratio is available in cap dolostone.

Silicate weathering and carbonate precipitation would increase the δ²⁶Mg value [32], but it immune from low-grade metamorphism or alteration [36,43], where δ²⁶Mg is defined as the permils deviation from a standard material. The δ²⁶Mg is heavier in dolomite than limestone, but light Mg isotope is enriched in silicate. This isotopic system may provide weathering intensity information in SBE environments (e.g., [18,44]); however, only scarce data is available. The δ²⁶Mg is unique for tracing enhanced continental weathering if combined with Sr isotopes [32,45].

1.2. Previous Leaching Studies in Literature

Previously available sequential-extraction methods were summarized in

Table 1. The available carbonate sequential-extraction methods using dilute acetic acids.

Reference	Concentrarion of Acetic Acid (v/v)	Mineral Phase	Remarks
This study	0.25%, 1.0% and 1.5%	Dolomite	3-step-extraction
Bailey [46]	20%	Calcite	Dissolve 10% sample in every step
Ling et al. [50]	1 M (~5%)	Dolomite	50 °C for 15 h
Zhao et al. [48]	0.5 M (~2.5%)	Dolomite	Room temp. for 4 h
	3.5 M (~17.5%)		60 °C for 24 h
Li et al. [26]	0.2%/1%	Dolomite	First step for ~30–40% sample
	0.2%/2.5%		Second step for ~30–40% sample
Liu et al. [49]	0.25%–10%	Dolomite	15-step extraction
Wang et al. [47]	1 M (~5%)	Dolomite	For trace elements
	0.1 M (~0.5%), 4 h		Wash out first 30%, then take supernatant

. Bailey et al. [46] used dilute acetic acid (10% v/v) to isolate impurities in carbonates, evaluated in every ~10% weight loss and found the first 30% fractions contained radiogenic Sr. Latter researches either to discard the first ~30% directly [26,47] or performed the 2-step extraction to isolate calcite in the later 70% portions [48]. However, too strong of acetic acid were used to prevent dissolution of other impurities [48]. These procedures were applied for Sr isotopic compositions in dolostone [26]. Later, a more tedious 15-step extraction was performed and found dolomite signatures after >80% dissolution [49]. This confirms the environmental information was stored in the last 20% un-dissolved fractions and proposed possibilities for separating calcite and dolomite using dilute acids in short reaction time.

1.3. Aims of This Study

It is necessary to establish new procedures for further dolostone researches due to unresolved altered calcite (i.e., diagenetic or secondary) and silicate impurities issues. Goals needed to be achieved are: (1) to prevent silicate dissolution, as theirs Sr and Mg isotopes may differ largely in specimens; (2) to separate calcite from dolomite, most of dolostone specimens contain ~50% CaCO₃ (

Table 2. The cap dolostone specimens from the Shiyu section were previously used by Wang et al. [47]. In this paper, the names of SY-01 to SY-26 represents the distance from the bottom to the top of Lantian Formation that each specimen located, roughly 47~98% dolomite, 0.07~50.7% calcite and 1.3~9.6% quartz were estimated based on Raman and XRD results.

Sample	Height (m)	Raman	XRD
			Dolomite (%)	Calcite (%)	Quartz (%)
SY-26	5.5	-	-	-	-
SY-25	5.3	Carbonate	-	-	-
SY-24	5.2	Carbonate	-	-	-
SY-23	4.7	Carbonate	85.74	9.3	4.96
SY-22	4.5	Carbonate	46.62	50.71	2.67
SY-21	4	Carbonate	-	-	-
SY-20	3.8	Carbonate	95.36	0.18	4.46
SY-19	3.6	-	89.1	3.63	7.27
SY-17	3.4	Carbonate	71.26	25.27	3.47
SY-16	3.3	Carbonate	53.91	43.48	2.61
SY-15	3.2	Carbonate	76.4	13.97	9.64
SY-14	3	Carbonate	90.67	0.77	8.56
SY-13	2.7	Carbonate	95.49	1.04	3.47
SY-12	2.5	Carbonate	89.02	9.67	1.313
SY-11	2.2	-	94.92	3.6	1.478
SY-10	1.9	Carbonate	98.25	0.21	1.534
SY-9	1.7	-	-	-	-
SY-8	1.4	Carbonate	-	-	-
SY-7	1.3	Carbonate	-	-	-
SY-6	1	Carbonate	95.89	0.07	4.04
SY-5	0.8	Carbonate	-	-	-
SY-4	0.6	Non-carbonate	-	-	-
SY-3	0.4	Non-carbonate	-	-	-
SY-2	0.2	Carbonate	-	-	-
SY-1	0.1	Non-carbonate	-	-	-

). As calcite dissolved faster in dilute acids (Zhao et al., 2009), one should be able to separate the two phases using dilute acetic acid in a short reaction time; (3) to obtain the least altered dolomite, it is rather heterogeneous in compositions (Gammon et al., 2012); and finally (4) to apply multiple diagenetic proxies, of Sr and Mg isotopes, Mg/Ca, Mn/Sr and Sr/Ca for oceanic mixing reconstruction after Snowball Earth.

2. Materials and Analytical Methods

2.1. Geological Background

In the Shiyu section area, the stratigraphy consists from the base to the top: the Leigongwu Formation that contains a Marinoan diamictite, the Lantian Formation that is principally composed of shale with beds of limestone and dolostone and a cap dolsotone at the base, and the Piyuancun Formation that contains siliceous rocks (Figure 1). The Lantian Formation is subdivided into Member I, II, III and IV. The formation is correlated with the Dousantuo Formation in China [47,51]. Member I is a ~5.5 m-thick dolostone that is attributed a cap dolostone. Member II is composed of a lower part that consist of 2.0 m-thick grayish siltstone and 15.0 m-thick black shales with micro-fossils horizon of the Lantian Biota [47], and its upper part of 22 m-thick black shales with interbedded black dolostone. Member III contains 2.0 m-thick dolostone at the base, and 12 m-thick thin-bedded gray limestone and argillaceous limestone at the top. Member IV contains 20 m-thick black shales. In this paper, 26 specimens from the Member I of the Lantian Formation at Shiyu (29°57.09′ N, 118°02.274′ E), southern Anhui province, China (Figure 1) were selected for geochemistry.

To Develop New Sequential-Extraction Procedures in Cap Dolostone

Proper leaching procedures will be evaluated and tested thoroughly in laboratory, then applied for separation of pristine dolomite in natural cap dolostone. Down-core specimens from the Lantian Formation were previously studied for stable isotopes and geochemistry [47]. Four selected specimens were used for developing series leaching procedures in this study. Individual collecting solutions will be examined in details of major/trace elements, ⁸⁷Sr/⁸⁶Sr and Mg isotopes, as well as the residues that were checked in mineralogy, and weight loss before final digestion for chemical and isotopic measurements. The results will be used to delineate factors that affecting fine-scale chemical and isotopic compositions in solutions. Careful examinations will be conducted after each leaching solutions and compared with the mineralogical results. Previous geochemical discussion with the extraction outcomes has been initiated [18]. However, this study aims to understand mineralogical parameters with the geochemical results of Sr and Mg isotopes, Mg/Ca, Mn/Sr and Sr/Ca in leaching solutions (Figure 2, the experimental flow chart).

Raman spectrometry was performed to identify carbonate minerals, as well as XRD and SEM on micro-phases of impurities. Each specimen will be conducted for series sequential-extraction processes, where major/trace elements in supernatant were determined by ICP-OES. Finally, multiple proxies of Sr and Mg isotopes, Mg/Ca, Mn/Sr and Sr/Ca in down-core specimens were conducted for environmental reconstruction after SBE deglaciation.

In our experiments, specimens with less veins were selected for obtaining the least altered dolomite in the final solutions. Operational procedures include sliced rocks, grinded fresh surface with corundum and supersonic bath, micro-drill powders (~0.5 g) to avoid veins, grains or altered parts, grounded by agate mortar to prevent uneven size, shape (e.g., flakes), XRD powders analyses (~0.5 g), and obtain extraction solutions using dilute acetic acids of 0.25, 1.0 and 1.5 v/v%, respectively.

2.2. Analytical Methods

2.2.1. Mineralogical Analysis

Mineralogy of the cap dolostone specimens that were used in Wang et al. [47]. Raman spectrometry (RS) is a simple and non-destructive mineralogical method and will be applied to scan the specimens before each leaching procedure in this study. Sliced specimen was shaped to fit a flat plate in a Renishaw instrument at Center for Micro/Nano Science and Technology (CMNST), National Cheng Kung University (NCKU). It was equipped with a laser (514 nm/633 nm/785 nm) source, frequency range 0 to 4000 cm⁻¹, a minimum laser spot diameter ~1 μm.

XRD were carried out at Micro and Nano Mineral Science Lab, NCKU, BRUKER, D8 Advance instrument, Cu Kα radiation equipped with Sol-X and Scintillation Counter, operating at 40 kV and 40 mA. Approximately ~0.5 g powder was loaded and recorded 2θ of 10 to 60°, scan step 0.03° for 1 s, and ~35 min in each measurement. Minerals were identified from peaks and their relative intensity. The referenced minerals were collected from Inorganic Crystal Structure Database (ICSD). The major mineral phases found in Shiyu were dolomite with minor quartz. TOPAS software was used to quantify each mineralogy, simulation was less than 25%, without modified coefficient. Based on the XRD patterns, the mineral compositions can be calculated by TOPAS.

SEM analyses were carried out by field-emission scanning electron microscope (FE-SEM, FEI-QUANTA 250) equipped with energy dispersive spectroscopy (EDS) at Micro and Nano Mineral Science Lab.

2.2.2. Geochemical Analysis

All chemical analyses were operated in class-10 working benches, inside a class-10000 clean room at Earth Dynamic System Research Center (EDSRC). Ultrapure water was deionized and purified by Milli-Q system (~18.2 MΩ). Reagents of HCl and HNO₃ acids were GR grade, further purified by double sub-boiling distilled system. Bottles, centrifuge tubes, columns, PFA vails, and other experimental apparatus were cleaned by dilute acid before used.

Approximate 0.05 g of powders were taken to 15 mL centrifuge tube and washed with 6 mL Milli-Q to remove soluble phases, then treated with 3-step of 0.25, 1 and 1.5% (v/v) acetic acids. In each step, solutions were supersonic bath for 15 min and centrifuged at 7500× g rpm, 10 min at room temperature, the supernatants were then carefully dried at 85 °C and re-dissolved in 0.3 N HNO₃ for isotopic analysis. Major/trace elements (i.e., Ca, Mg, Fe, Mn, Sr, Al, Si, K, Ba and Ti) were determined by ICP-OES at EDSRC. Average analytical uncertainty for most elements is better than 5%.

2.2.3. The Sr and Mg Isotopic Analyses

The Sr ion-exchange separation procedures were followed that of Liu et al. [52]. Briefly, acid-clean polypropylene columns with inner diameter of 5–5.5 mm were loaded with 0.25 mL Eichrom Sr Spec resins (50–100 mesh). 250 ng Sr of each extraction step and 450 ng Sr of IAPSO were taken for separation. The recovery after column chemistry was measured by ICP-OES, ~100%. After separation, the Sr were diluted to 150 ng g⁻¹ in 0.3 N HNO₃. ⁸⁷Sr/⁸⁶Sr was determined using MC-ICP-MS, Neptune at EDSRC. Amplifier gains were calibrated daily, and each standard or sample analysis consisted of baseline measurement and forty-eight (6 blocks/8 cycles) 2.097 s of integration. Moreover, ⁸³Kr and ⁸⁵Rb were determined to evaluate and correct the ⁸⁶Kr contribution on ⁸⁶Sr (⁸³Kr/⁸⁶Kr = 0.664740) and ⁸⁷Rb on ⁸⁷Sr (⁸⁷Rb/⁸⁵Rb = 0.385617). These isobaric interferences were corrected by exponential fractionation law. The stability and quality were evaluated using NIST SRM 987 standard (prepared from strontium carbonate powder, NIST, USA) and Sr High-Purity standard. The preferred ⁸⁷Sr/⁸⁶Sr, 0.710245, of SRM987 from long-term TIMS, was used to correct for the daily offset.

The Mg separation procedures were defined by Wombacher et al. [53]). Acid-clean polypropylene column with inner diameter of 5–5.5 mm were loaded with 1 mL cation exchange resins, Biorad AG50W-X8 (200–400 mesh), to separate potential interference of Ca, Fe, Mn, Be, and Ti. Wombacher et al. [53] reported incompletely Mg separation in high Ca/Mg samples, e.g., JCP-1 standard, and it required additional column for Mg purification. Approximate 5 μg Mg of specimen were taken and measured by ICP-OES to check its chemical yields. Average recovery was nearly 100% (mean 101.2% ± 5%), but was easily affected by matrix. Once element/Mg reached ~0.03, it caused significant mass bias shifted, from 0.08 to 0.47‰ by Be, Ca or Ti [53]. These potential interferences were monitoring using ICP-MS to make sure the ratio is <0.03. After separation, the liquids were diluted, 50 ng g⁻¹ Mg in 0.1 N HNO₃, for MC-ICPMS (Neptune Plus) at the Institute of Earth Sciences, Academia Sinica (IESAS). An Aridus II desolvator was used, operated at 140 °C and ~5.35 (L/min) Ar-gas flow. No N₂ gas was added to reduce ¹²C¹⁴N⁺. It is difficult to separate ¹²C¹⁴N⁺ from ²⁶Mg (m/Δm = 1270) under medium resolution. A Jet-cone and X-cone was used as sampler and skimmer, respectively, to achieve high Mg sensitivity. Amplifier gains were calibrated daily, where each analysis consisted of baseline and sixty (6 blocks/10 cycles) 2.097 s integration. Blank was corrected by the beginning and end of each analysis. Standard-sample bracketing (SSB) was used to correct mass bias drift using DSM3 standard, provided by Dead Sea Magnesium Ltd., Israel [34]. Mg isotopic results were presented as the delta notation in permils, as δ²⁶Mg_DSM3 relative to DSM3 (δ²⁶Mg and δ²⁵Mg). Uranium oxide ratio, UO/U, was used to monitor oxides interference.

3. Results

3.1. Mineralogical Compositions in the Cap Dolostone

The RS bands of calcite and dolomite are different due to their structure variations [54]. Selected RS patterns in the Shiyu specimens, at 4000–0 cm⁻¹ region, are summarized in Appendix C and Figure 1A. The peak intensity in calcite occurred at ~1090, 290, 170 and 720 cm⁻¹, in strong contrast to that of dolomite at ~180 or 300 cm⁻¹. However, we could not separate easily of dolomite from calcite, because of high background and poor signal in instruments. Fortunately, the 1090 cm⁻¹ band related with the CO₃²⁻ functional group, can be used to recognize carbonates. Our specimens composed mainly carbonates, except SY-01, 03 and 04 (see Appendix A). Most SY specimens, e.g., SY-10, 12, 16, 15, 16, 18, 19, 20, 21, 22 and 25, shown significant bands centered at 1350 and 1600 cm⁻¹. SY-01 and 02 were mainly composed of quartz at ~463 cm⁻¹, comparable with pure quartz (~467 cm⁻¹). Similarly, SY-04 shown no carbonate bands, identified as non-carbonate. On the other hand, the SY-01, 03 and 04 are unlike carbonates, also no enough powders can be extracted by hand-held micro-drill, possible non-carbonate specimens (Table 2).

The XRD patterns display major minerals in specimens are summarized using peak positions and relative intensities (Figure 3). We found dolomite is the pre-predominate mineral and mixed with minor calcite, quartz and chlorite, similar with the observations in Ediacaran carbonates in South China [55]. Due to lack of chlorite standard in the lab, we simulated compositions without this trace mineral. All obtained compositions are summarized in Table 2, roughly 47~98% dolomite, 0.07~50.7% calcite and 1.3~9.6% quartz. All our specimens consist >90% carbonates, confirmed the RS results, with no correlation of minerals and depth. Our estimated relative calcite and dolomite contents are rather different from the Zhao and Zheng [56]. The latter used CO₂ volume derived by selective acid extraction, contained ~26 to 48% dolomites. This inconsistent results emphasize difficulties associated dolomite and calcite separation, reclaimed the importance of mineralogical analysis prior to the leaching experiments. With knowledge on mineralogical, one can choose more appropriate method for chemical analysis. We conclude that too complicated procedures won’t suit for dolomite dissolution.

The SEM with backscattered image (BSI) of SY-04 and SY-14 are shown in Figure 4. The SY-04 contains mainly non-carbonate, quartz and fluorite (Figure 4a) and should not be applied for leaching treatments. SY-14 is composed of dolomite, calcite and pyrite, clear calcite veins (light gray) across the entire specimen and interlarded with dolomite (dark gray) (Figure 4b). It might be secondary calcite, not precipitated directly and contained no pristine signals. However, this is in strong contrast to the XRD results (

Table 3. The procedures of the newly developed 3-step extraction procedures performed in this study for separating calcite and impurities in the dolostone specimens.

Step	Reagent	Reaction Condition	Target Phase
Wash	6 mL Mili-Q water	10 min supersonic bath	Water soluble phase and ion-exchange site
Step 1	6 mL 0.25% HAc	15 min supersonic bath	Calcite and altered dolomite
Step 2	6 mL 1% HAc	15 min supersonic bath	(Calcite) and dolomite
Step 3	6 mL 1.5% HAc	14 min supersonic bath	Dolomite
Residue	3 mL HNO3 and 3 mL HF	48 h at 100 °C	Silicate phase

) and implies heterogeneous distribution is an issue in the tests. The elemental mappings are shown in the Figure 4, where alumino-silicate were found in SY-14, could be chlorite shown in the XRD pattern. This emphasizes that careful sequential-extraction should avoid clay dissolved in dilute acetic acid. On the other hand, the calcite interlarded with dolomite could not be avoided, but can be washed out by sequential-extraction to prevent contamination.

3.2. Geochemical Results in the Leaching Experiments

Four specimens, SY-8, 14, 16, and 22, were selected to compare in details of the solution and the residue after each leaching. SY-8 and 14 are composed >90% dolomite, but SY-16 and 22 contain 50% dolomite (Figure 4). The results will be used to understand mineralogy impacts in dissolution. The amount dissolved in each step will be calculated by the weight loss and the (Mg + Ca) content in solutions. The initial and the end weight, the (Mg + Ca) in solutions, as well as the dissolution in each step can be calculated based on ((Mg + Ca) _{each step})/((Mg + Ca) _sum), ((Mg) _{each step})/(Mg) _sum) and ((Sr) _{each step})/(Sr) _sum) results.

The estimated loss in each step are summarized in Appendix A and plotted in Figure 5. Overall, the average 1st leaching dissolved 21.3–32.7%, the 2nd step dissolved 29.5–41.5% and the 3rd step dissolved 20.8–29.4%. As SY-14 contains quartz, its leaching sum is much less than others.

4. Discussion

4.1. Evaluation of the New Sequential-Extraction Results

Previously, Bailey et al. [46] used dilute acetic acid (10% v/v) to isolate impurities, evaluated in every ~10% weight loss and found the first 30% fraction contained radiogenic Sr. Later either to discard the first ~30% directly [26,47] or performed 2-step extraction to isolate calcite fraction [48]. The latter, however, used too strong acetic acid to prevent dissolution of others. Li et al. [26] performed 2-step leaching for obtaining Sr isotopic compositions in dolostone. A more tedious 15-step extraction was performed and found signatures only after >80% dissolution [49]. It is evident that important information store in the last 20% un-dissolved fraction and separation of calcite/dolomite using dilute acid in a short reaction time. Dolostone is rather heterogeneous [7] and its ⁸⁷Sr/⁸⁶Sr may vary largely, visible veins should be avoided. The altered parts may dissolve faster in weak acids due to linkage [49]. Our new procedures basically modified from Li et al. [26] and Zhao et al. [48], but have added extra caution for isolating pristine dolomite in the last digestion.

Four specimens, SY-8, 14, 16, and 22, were selected to compare in details of the solution and the residue after each leaching. SY-8 and 14 are composed >90% dolomite, but SY-16 and 22 contain 50% dolomite. The results will be used to understand mineralogy impacts in dissolution. The amount dissolved in each step will be calculated by the weight loss and the (Mg + Ca) content in solutions. The initial and the end weight, the (Mg + Ca) in solutions, as well as the dissolution in each step can be calculated based on ((Mg + Ca) _{each step})/((Mg + Ca) _sum), ((Mg) _{each step})/(Mg) _sum) and ((Sr) _{each step})/(Sr) _sum) results.

The estimated loss in each step are summarized in the Figure 5. Overall, the average 1st leaching dissolved 21.3–32.7%, the 2nd step dissolved 29.5–41.5% and the 3rd step dissolved 20.8–29.4%. As SY-14 contains quartz, its leaching sum is much less than others. Wen et al. [57] reported different specimens dissolved of <1~>10% in carbonate leaching yields, emphasizing potential matrix influence. In the new method, >80% specimens were dissolved, solved the previous unstable yield problems. The sum of dissolved Sr in each step is closed to the total dissolved after the residues disregarded (Figure 5). Silicate contains much less Sr and its Sr contribution in the leaching solution should reflect mainly of dissolved carbonate. In contrast, Mg should indicate either calcite or dolomite dissolution. SY-16 and 22 showed major Mg at steps 2 and 3, implying more representative of dolomite dissolution results. The related geochemical proxies in each leaching step (Figure 5) can be used to recognize possible dissolved minerals. Using these chemical results, we discuss calcite/dolomite separation in the Cap-dolostone selected. SY-08 and 14, composed ~90% dolomite, theirs Mg/Ca, Sr/Ca and Mn/Sr varied rather insignificantly in different leaching solution. In contrast, Mg/Ca and Sr/Ca changed largely and converged at the 3rd step in SY-16 and 22. These Mg/Ca values indicated significant secondary calcite mixed with dolomite. The calcite fraction in SY-16 and 22 showed low Mn/Sr and high Sr/Ca. On the other hand, Si, Ti, Al and Fe are rich in silicates, but trace in carbonate, should be sensitive proxy of detecting any silicate dissolution. The residues provide possible end-member silicates, except of SiF4 escaped during HF and HNO₃ digestion. The residues were increased gradually from step 1 to 3 due to silicate dissolution. Combined with isotopic values in each step (Figure 6), the dissolved fraction can be understood clearly. The ⁸⁷Sr/⁸⁶Sr in residues explain sensitively of silicate dissolution, the highest Ti and ⁸⁷Sr/⁸⁶Sr occurred in this fraction. Similarly, δ²⁶Mg reflected mainly the calcite and dolomite. It is major in dolomite and less prone to silicate dissolution. Low Mg found in the 3rd step solutions, but high in the residues (Figure 6a). It is critical to isolate pristine dolomite from other contaminations for environmental reconstruction. The 1st step fraction should be eliminated, because its low Mg/Ca detected. The 3rd step provides low ⁸⁷Sr/⁸⁶Sr with high Mg/Ca, proper for in detailed discussion of diagenesis-free environmental changes in the past.

4.2. Separation of Non-Dolomite Phases

Mineralogical composition is complicate and diverse in dolostone, where leaching results may affect by mineral matrix. Possible end-member compositions contain altered calcite in early stage and the silicate impurities in later dissolutions. These were closely checked by physical/chemical proxy in evaluation. Particularly, the 3rd-step results were examined in detailed, including C and O isotopes presented in Wang et al. [47] and the new proxies obtained in this study. All chemical and isotopic results obtained n this study are summarized in Appendix B. Mineral content was plotted with geochemical proxies of dolomite. The Mg/Ca vs. the dolomite content plot showed moderate correlation (R² = 0.65, Figure 7), but implies no 100% dissolution in SY-16 and 22. Moderate correlation between C and O isotopes was reported in Wang et al. [47]. Also depleted C and O isotopes in vein calcites than dolomite reported, as well as the low δ²⁶Mg in calcite than dolomites [56]. However, theirs Mg contains are too low to cause detectable δ²⁶Mg shifted, estimated only 3% assuming of 26 and 74 mol% dolomite and calcite, respectively [36]. The dolostone specimens composed of ~50% dolomite, even in SY-16 and 22. The calcite contributed Mg is estimated to be <1~0.1% in the 3rd solutions, confirmed by the detected small δ²⁶Mg changes. On the other hand, the calcite Sr in the 1st solutions could be up to 45% and its ⁸⁷Sr/⁸⁶Sr in each step is much higher (Figure 7).

There is no correlation observed between ⁸⁷Sr/⁸⁶Sr and the dolomite content (Figure 7). Silicate affect ⁸⁷Sr/⁸⁶Sr largely if any clay dissolved in dilute acetic acids. The Si/(Mg + Ca) vs. Al/(Mg + Ca) plot is a sensitive proxy for silicate contribution. Its excellent correlation (R² = 0.94) implies either similar sources or Si-Al bearing minerals dissolution (e.g., chlorite). The Si and Fe, Mn vs. Mg plots have shown low correlation (Figure 8), suggesting negligible dissolution of quartz, abundant in dolostone. Silicate dissolution provides Fe, Mn and Mg in the solutions. There is moderate correlation in the plots, R² = 0.3~0.42, but might cause by minor contents in dolomite. Low correlation detected in the ⁸⁷Sr/⁸⁶Sr, δ²⁶Mg vs. Si plots suggest minor silicate dissolution have occurred and supports the similar source argument (Figure 8).

4.3. Diagenesis Artifacts Evaluation

The new extraction procedures prevent mixture of secondary calcite and silicate impurity. Their Sr and Mg isotopes in the 3rd solutions reflect faithfully the pristine isotopic compositions in dolomite. However, it is important to confirm if the signatures are pristine dolomite, as carbonates are susceptible to post-depositional processes. Previously, Mg/Ca, Mn/Sr, δ¹⁸O and ⁸⁷Sr/⁸⁶Sr proxies were used for alteration evaluation, to check if Mn/Sr < 2 or δ ¹⁸O > −10‰ [58,59]. Available C and O isotopes in Wang et al. [47], however, were bulk dolostone. Proxies of Sr/Ca, Mn/Sr, ⁸⁷Sr/⁸⁶Sr and δ²⁶Mg will need further in-depth discussion. It should find no correlation among the multi-proxies if no post-diagenetic artifacts or synchronized signals. These were abnormally high Mn/Sr (41~202) and ⁸⁷Sr/⁸⁶Sr, 0.70983~0.71359, reported in cap dolostone, suggested due to post depositional alteration [58]. Certain chemicals were sensitive to alteration, such as Mn incorporated or Sr released [45,58]. Since limited Sr/Mn < 10 were used in limestone/pore water evaluation, there must be various ratios existed in different nature calcite or dolomite. High Mn/Sr of 25~85 were reported in cap dolostone, certain portions may cause by alteration. Accumulated Mn and Fe in seawater, may cause high Mn/Sr, were reported [49], assumed exchange coefficient, 5~70, Kd_Mn = (Mn/Ca)_dolomite/(Mn/Ca)_fluid [60]. Alternatively, high Mn/Sr ratio could reflect high Mn in previous seawater. The moderate correlations among Mn/(Ca + Mg), Mn/Sr and Si/(Ca + Mg) were observed in Figure 9, part Mn attributed from clays. Although, the Sr isotopes vs. Mn/Sr and Sr/Ca plots show only weak correlation (Figure 9). Their ⁸⁷Sr/⁸⁶Sr in the 3rd solution, 0.70983~0.71359, is higher than the reported values of 15-step extraction experiments, ~0.707–0.709 [18,49]. In contrast, our results are similar with Namibia, 0.70876–0.71815 (Halverson et al., 2007), North China, 0.71195–0.71364 [61], and South China, 0.70986–0.71308 [62]. Previous hypothesis of enhanced silicate weathering to cause similar ⁸⁷Sr/⁸⁶Sr, due to intense continental weathering during Marinoan deglaciation [11]. The radiogenic Sr isotopes could not be artifacts, as no correlation with other proxies were observed. Owing to high abundance of Mg in dolomite, as well as its high thermodynamic stability, δ²⁶Mg does not easy influenced by post-depositional alteration and is a robust tracer for tracing the origin [36]. Also, the low correlation with the Mn/Sr vs. δ²⁶Mg plot reflect possible seawater source.

4.4. Implications for the Neoproterozoic Deglaciation

There is a medium correlation, R² = 0.48, in the ⁸⁷Sr/⁸⁶Sr and δ²⁶Mg plot of Shiyu specimens if excluded SY-15 of low Mg content (Figure 10). These increased Sr vs. Mg isotopes correlation could link with enhanced continental weathering (e.g., Tipper et al. [63]; Halverson et al. [45]) described in PW scenario [18]. Mg in dolomite is less prone to post-depositional diagenesis and δ²⁶Mg is excellent tracer for reconstructing of weathering changes in the past [36]. The temporal variation of the cap dolostone after the 3-step procedures have shown interesting distributions (Figure 10).

The Mg and Sr Isotopes Co-Variations in Shiyu

Previous REEs study suggested influences of terrigenous and glacial water in Cap-dolostone [55]. The Sr and Mg isotopic variations recorded possible regional enhanced continental weathering events. Weathering would happen at mid-depth and more intense weathering at the beginning and the end of deglaciation. In this study, both high isotopic compositions of Sr and Mg were detected at the low depth, affected by terrigenous and glacial water. Under low sea-level conditions, carbonate would precipitate from poor to non-mixed regions; however, subsequent sea-level rise would cause decreased weathering, lead to depleted in Sr and Mg isotopes. On the other hand, ice uncovered could result to intense weathering and to cause high Sr and Mg isotopes in dolostone. However, the Mg isotopes would be affected by other factors, such as temperature, rate of precipitation and continental weathering. Comprehensive studies are needed to decipher the degree of paleo-environmental changes during the Marinoan deglaciation.

The influence of secondary precipitation on δ²⁶Mg in dolomite is rather limited. The Ca/Sr vs. ⁸⁷Sr/⁸⁶Sr and the Mg/Ca vs. δ²⁶Mg shown no significant correlation in the plots (Figure 11), suggesting silicate, limestone, and dolomite are potential sources in the cap dolostone. However, limestone and dolomite presented distinct Mg isotopic compositions and showed low correlation in the 1/Mg and δ²⁶Mg plot in rivers [64].

5. Conclusions

The post-Marinoan Cap-dolostone collected from the Lantain Formation are not composed of pure dolomite (mostly >90% dolomite), but also with minor calcite, quartz, chlorite and pyrite, based on the results of petrological, mineralogical and chemical compositions. Bulk Cap-dolostone dissolution by strong acids can’t isolate pristine dolomite compositions in specimen. Instead, a newly established 3-step extraction was performed to achieve high purity dolomite isolation. The method has been initiated and tested successfully using 3-step dilute acetic acids. The calcite was separated from the dolomite during the first two steps and the third step was designed for impurities separation. The pristine dolomite contains low ⁸⁷Sr/⁸⁶Sr and high Mg/Ca.

The Al and Si/(Ca + Mg) plot serves as proxy for clay dissolution; however, the Mn/Sr could be resulted partly from high Mn seawater after de-glaciation. The high Sr isotopes in carbonates could be linked with intensive continental weathering during deglaciation. Interestingly, Mg isotopes in pristine dolomite records critical information on dolostone precipitation processes.

The influence of secondary precipitation on δ²⁶Mg in dolomite is limited. The Sr/Ca vs. ⁸⁷Sr/⁸⁶Sr and the Mg/Ca vs. δ²⁶Mg plots have suggested Sr and Mg silicate, limestone, and dolomite sources in the dolomite (Figure 10). Limestone and dolomite presented distinct Mg isotopic compositions, agreed with the low correlation reported in the 1/Mg and δ²⁶Mg plot in rivers (Figure 11).

The observed Sr and Mg isotopes correlation (R² = 0.48, n = 23) in dolomite explains the changes of the continental weathering and the seawater source mixing processes. The temporal changes in these isotopes could be used to constrain the weathering intensity and the oceanic conditions after SBE deglaciation in Neoproterozoic.