Introduction

Molar tooth carbonates (MTCs), or carbonate rocks containing molar tooth structures (MTSs), occur mostly if not exclusively in successions deposited in subtidal environments before 750 Ma (refs 1, 2). The formation of MTCs requires the generation of cracks within unconsolidated sediments, followed by the rapid infilling of such cracks with early diagenetic calcispar before sediment compaction. The formation of molar tooth (MT) cracks have been variously related to subaqueous syneresis3, gas bubble expansion resulting from CH4, H2S or CO2 degassing2,4,5,6 and seismic activities7,8. The disappearance of MTCs at around 750 Ma has been related to the rise of animals5,7,8,9, a drop in calcite saturation of seawater1,2,10 or an increase in the concentrations of calcite precipitation inhibitors such as Fe2+, Mg2+, SO42– or PO4 (refs 1, 2, 3, 11).

To illuminate the origin of MTCs, we measured the Mg, S and C isotopic compositions of MTCs from the early Neoproterozoic (1,000–750 Ma) Wanlong Formation in southern Jilin Province of North China (Supplementary Fig. 1). In the Wanlong Formation, MTSs are abundant within the thick-bedded argillaceous lime mudstone that is intercalated with the finely laminated limestone (Supplementary Note 1 and Supplementary Figs 2 and 3a). Sedimentological evidence, including the predominance of parallel bedding and the lack of subaerial exposure structures, indicates that MTCs in the Wanlong Formation was deposited below fair-weather wave base12. S isotopic data indicate that MT microspar was precipitated within microbial sulphate reduction (MSR) zone and Mg isotopic data suggest that microspar precipitation predated the dolomitization of host rock. We propose that MT microspar was precipitated in the sediment column where MSR and methanogenesis occur simultaneously underneath sulphidic seawaters and where the production of CH4 from methyl sulphides and the inhibition of CH4 oxidation by methanethiol allowed CH4 to build up in the sediments.

Results

Petrographic observations of the MTCs

MTSs are normally oriented vertically or obliquely with respect to bedding planes and show clear cross-cutting relationships with each other (Supplementary Fig. 3b–d). MT cracks are filled with microcrystalline calcite crystals (MT microspars) ranging from 10 to 20 μm in size (Supplementary Fig. 3e,f). The argillaceous host rocks (with an average siliciclastic content of 33.4 wt%, Supplementary Table 5) are partially dolomitized (Supplementary Fig. 3g,h).

Isotopic compositions of the MTCs

Sulphur isotopic values of carbonate-associated sulphate (CAS) extracted from MT microspars (δ34SMT: 31.9–42.8‰) are higher than those of CAS from calcareous host rock (δ34SHR: 19.1–27.6‰; Fig. 1a, Supplementary Fig. 4 and Supplementary Table 2). Mg isotopic compositions of MT microspars (δ26MgMT) is around –3.3‰ (relative to DSM3), ∼1.6‰ lower than those of the host rock (δ26MgHR; Fig. 1b, Supplementary Fig. 5 and Supplementary Table 1). C isotopes of MT microspars (δ13CMT) are systematically heavier than host rock (δ13CHR) by 0.5–1‰ (Fig. 1c and Supplementary Table 3).

Figure 1: Isotopic compositions of MTCs.
figure 1

(a) S isotopic compositions of CAS from MT calcispars (MT) and host rocks (HR). (b) Cross-plot of Mg/Ca (molar ratio) versus δ26Mg. The argillaceous host rock has higher Mg/Ca ratios and is enriched in 26Mg than MT calcispars. (c) C isotopic compositions of MT calcispars (MT) versus host rock (HR) of four samples (A, B, C and D).

Full size image

Discussion

δ34SHR of the Wanlong carbonates is within the range of sulphur isotopic compositions of Neoproterozoic CAS13. The greater values of δ34SMT indicate that MT microspar was precipitated in the sulphate reduction zone in the sediment column, where 32S is preferentially removed from the porewater sulphate pool by sulphate reduction microbes14 (Supplementary Note 2). MT microspar precipitation in the MSR zone is also consistent with generally lower CAS concentrations in MT microspar than in host rock (Supplementary Fig. 6 and Supplementary Table 4).

δ26MgMT is related to the Mg isotopic composition of porewater (δ26Mgpw), from which MT microspar precipitates, and the relationship can be expressed as follows:

where Δcal is the fractionation associated with inorganic precipitation of low-Mg calcite and can be set at 2.2–2.7‰ (refs 15, 16). Thus, δ26Mgpw is estimated to be between –0.6 and –1.1‰, within the range of seawater compositions in the past 70 million years17,18. Greater δ26MgHR values might be attributed to the partial dolomitization of host rock, because dolostone is systematically heavier than limestone in Mg isotopes19,20. On the other hand, as dolomite and other authigenic Ca carbonate formed in the sediment column would preferentially scavenge 24Mg from porewater21,22 (Supplementary Note 3), δ26Mgpw would increase as dolomitization proceeds. It is estimated that 10–25 wt% of carbonate in the host rock of the Wanlong Formation is dolomitized (Fig. 1b), meaning δ26Mgpw would increase by ∼2‰ (Supplementary Fig. 7). Had MT microspar in the Wanlong Formation precipitated after host rock dolomitization, seawater Mg isotopic composition would have to be between –2.6 and –3.1‰, which is even lower than the influx from carbonate weathering (–2.25‰; that is, the lower bound of riverine input)18. Thus, MT microspar precipitation must predate host rock dolomitization. This inference is also consistent with the petrographic observation that MT structures are often ptygmatically folded and sometimes brittly fractured1, suggesting that MT microspar was precipitated before host rock cementation. In this light, it is possible that the exclusive occurrence of MT structures in argillaceous carbonates1 may be related to clay minerals, which tend to delay host rock cementation23.

Thus, sedimentary evidence, S isotopes and Mg isotopes indicate that MT microspar precipitation must occur in unconsolidated sediments, within the MSR zone and before dolomitization. To generate MT structures, cracks must develop in unlithified sediments and gas expansion is a plausible mechanism to generate such cracks2,4. Here we explore the nature of the gases and the unique Proterozoic environments conducive for gas bubble formation in sediments.

During the early to middle Proterozoic, atmospheric oxygen level was extremely low (<1% present atmospheric level) and the deep ocean remained anoxic and sulphidic in places24,25,26,27. Sulphidic conditions were particularly common in Proterozoic continental margins25,26,28 and perhaps in epicratonal environments as well29,30. Although euxinia may have extended over <10% of global seafloor in mid-Proterozoic according to some estimates27, sulphidic waters might have had profound impacts on the Proterozoic Earth system. We propose that methyl sulphides might have been produced in significant quantities in sulphidic marine environments. Methyl sulphides are a group of volatile organic sulphur compounds, including dimethyl sulphide (CH3SCH3) and methanethiol (CH3SH), which are produced in modern marine and freshwater environments. Methyl sulphides can be produced either by the degradation of dimethylsulphoniopropionate in the surface ocean31 or by anaerobic methylation of hydrogen sulphide in sulphidic sediments32. Therefore, it is expected that the production of methyl sulphides would be enhanced in Proterozoic sulphidic marine environments, both in the water column and within sediments.

As volatile gases, methyl sulphides produced in water column tend to readily emit to atmosphere, but those generated within the sediments can serve as a non-competitive substrate for methanogens33,34,35. As sulphur-reducing microbes cannot use methyl sulphides but methanogens can, MSR and methanogenesis can co-occur simultaneously within sediments where methyl sulphides are present36, resulting in the convergence of the MSR and methanogenesis zones. In addition, anaerobic oxidation of methane (AOM) is inhibited by methyl sulphides such as methanethiol. With methane oxidation inhibited, CH4 can accumulate in sediments in significant quantity37, in sharp contrast to modern marine sediments, where the MSR zone lies invariably above the methanogenesis zone38, with intensive AOM at the base of MSR zone consuming most CH4 and consequently modern marine CH4 discharge accounting for only 2% of the global flux39 (Fig. 2).

Figure 2: Schematic diagram showing geochemical reactions within marine sediments bathed beneath ferruginous and sulphidic seawaters.
figure 2

(a) Under ferruginous conditions, the MSR and methanogenesis zones are separated. Methanogens are outcompeted by sulphate-reducing bacteria, if both use competitive substrates (CH2O)n. Within the MSR zone, reaction between H2S and Fe2+ precipitates pyrite and generates H+, which lowers porewater pH. Most CH4 produced within methanogenesis zone is oxidized by sulphate at the base of MSR zone where AOM occurs. Thus, there is little benthic CH4 flux from marine sediments. (b) Under sulphidic conditions, methyl sulphides are produced within both water column and sediments. In sediments, methyl sulphides serve as a non-competitive substrate for methanogens, allowing MSR and methanogenesis to take place concurrently in the MSR-methanogenesis convergence zone. H2S and Fe2O3 react to produce pyrite and generate OH–, thus favouring CaCO3 precipitation. AOM is prohibited by methanethiol, allowing CH4 accumulation in sediments and significant benthic CH4 fluxes into atmosphere.

Full size image

We propose that the accumulation of the insoluble gas CH4 in the convergence zone provided a physical mechanism to generate cracks in unconsolidated sediments2,4. Furthermore, the geochemistry within the convergence zone where MSR and methanogenesis overlap could have facilitated the precipitation of calcite to fill such cracks. With the generally low concentrations of Fe2+ in sulphidic porewaters, pyrite formation would involve the reaction between H2S and Fe2O3. In fact, the host rock of MTCs in the Wanlong Formation contains an average of 0.42 wt% of pyrites (Supplementary Table 5). The overall reactions for pyrite formation fueled by MSR and methanogenesis using methanethiol and dimethyl sulphide can be described as follows:

These reactions generate OH− and , which elevate pH, increase porewater alkalinity and favour CaCO3 precipitation40,41,42. In addition, the dearth of calcite inhibitors such as Fe2+ and SO42– in sulphidic sediments would also promote rapid precipitation of CaCO3 (ref. 11).

δ13CMT of the Wanlong carbonate is only slightly greater than δ13CHR by 0.5–1‰ (Fig. 1c), similar to previous studies showing that MT microspar and host carbonate rock have nearly indistinguishable δ13C values5,6,43. To assess the extent of carbon isotope variation between MT microspar and host rock, we consider a simple model where methanogenesis produces sufficient CH4 to produce cracks that are immediately filled with MT microspar. To generate cracks by CH4 accumulation, gas pressure must be balanced with the hydrostatic pressure, which is dependent on water depth. Our calculation shows that methanogenesis alone does not generate sufficient bicarbonate (and MT microspar) to fill the cracks that would be created at reasonable water depths by the amount of CH4 it produces. Thus, MT microspar precipitation was probably supplemented by porewater bicarbonate (which would be isotopically similar to seawater bicarbonate and to δ13Chost) and bicarbonate derived from sulphate reduction (Supplementary Note 4). To simplify our calculation, we consider the simplest situation in which bicarbonate derived from methanogenesis was entirely used in MT microspar precipitation, with additionally needed alkalinity coming from porewater (that is, a binary mixing model). Assuming that δ13C of methyl sulphides and carbon isotope fractionation during methanogenesis are –30‰ and –60‰ (ref. 44), respectively, mass balance consideration requires that δ13C of derived from methanogenesis be +150‰ based on equations (3) and (4). Our calculation shows that MT microspar precipitation at 100 m water depth would be ∼1‰ heavier than host rock (black solid line in Fig. 3) and methanogenesis-derived only accounts for <1% of MT microspar precipitation. Smaller isotopic difference between MT microspars and host rock would be expected if MSR-derived is involved (dashed lines in Fig. 3).

Figure 3: Geochemical model showing the carbon isotopic difference between MT microspar and host rock formed at different water depths.
figure 3

First, the amount of CH4 required to produce a unit volume of cracks at ambient pressure and temperature was estimated. The co-production of related to CH4 generation was then estimated and assumed to have been used fully for MT microspar precipitation. The methanogenic was inadequate to precipitate enough MT microspar to fill a unit volume and the shortage was made up by (1) pore water (black solid line); (2) from MSR () and methanogenesis () with a molar ratio of 1:1, and the remaining shortage fulfilled by pore water (red dotted line); or (3) from MSR and methanogenesis with a molar ratio of 2:1, and the remaining shortage fulfilled by pore water (blue dashed line). Porewater was assumed to have a δ13C value similar to that of host rock (that is, 1‰). The δ13C value of methanogenic was estimated at +150‰, given a δ13C value of methyl sulphides at –30‰, a fractionation between CH4 and methyl sulphides at –60‰ and the production of 3/4 mole of CH4 and 1/4 mole of from each mole of methyl sulphides (equations (2) and (3)). The δ13C value of MSR was assumed to be –30‰. The x axis represents water depths and the y axis indicates the isotopic difference between MT microspar and host rock (δ13CMT-δ13CHR).

Full size image

The disappearance of MTCs is coincident with the elevation of atmospheric oxygen levels at ∼750 Ma (refs 1, 45), suggesting a possible causal relationship. A direct consequence of ocean oxygenation and ventilation is the reduction of the areal coverage of euxinic waters and decrease in methyl sulphide production, which in turn would result in the spatial separation of the MSR and methanogenesis zones in sediments. As such, most CH4 was consumed at the base of MSR zone by AOM, preventing crack formation by CH4 accumulation. Furthermore, pyrite formation in ferruginous sediments through reaction with Fe2+ would generate protons, lowering porewater pH and favouring CaCO3 dissolution46,47. All these secular changes associated with atmospheric and oceanic oxygenation may have contributed to the disappearance of MT structure around 750 Ma.

CH4 accumulation in sediments also implies benthic CH4 discharge from marine sediments. The environmental impacts of benthic CH4 fluxes on the Proterozoic Earth system could potentially be profound. First, enhanced CH4 discharge would contribute to the persistently low atmospheric O2 levels in Proterozoic25,48. Second, strong benthic CH4 fluxes from continental margins would have contributed to the maintenance of an ice-free Earth in the middle Proterozoic. Finally, a significant reduction of CH4 discharge associated with the 750 Ma oxygenation event might have triggered the Neoproterozoic global glaciations. Thus, the Neoproterozoic oxygenation event may have had an impact on the secular distribution of sedimentary structures such as MT structure and the global climate system, as well as the rise of animals45.

Methods

Mg isotope analysis

Rock samples were split into two parts using a rock saw. A highly polished slab was prepared from one split, while a mirrored thin section was made from the counterpart. Sample powers were drilled from the polished slab using a hand-held micro-drill. The sampling procedure was guided by the petrographic observation of the corresponding thin sections. For Mg isotopic analysis, about 10–30 mg of powder was dissolved in 0.5 N acetic acid in a 15-ml centrifuge tube. Tubes were placed in an ultrasonic bath for 30 min, to allow complete dissolution of carbonate components. After centrifuging, supernatant was collected for column chemistry and elemental composition analysis.

Mg was purified using cation-exchange chromatography. The detailed procedure of column chemistry was reported in Shen, et al.49 and Huang et al.20 Mg was purified in two steps. Column 1 was designated to separate Mg from Ca. Mg was eluted by 4 ml of 10 N HCl, whereas Ca was retained in resin. Column 2 was used to separate Mg from all other elements. Na, Al, Fe and K were sequentially eluted using 1 N HCl, 1 N HNO3+0.5 N HF and 1 N HNO3, whereas Mg was collected using 5 ml of 2 N HNO3. To obtain a pure fraction of Mg, sample solutions passed through column 1 twice, followed by three passes through column 2. After column chemistry, Ca/Mg, Na/Mg, Al/Mg and Fe/Mg ratios were <0.05 and the Mg recovery rate was better than 99%.

Mg isotopic ratios were measured on a Thermo Scientific Neptune Plus high-resolution multicollector inductively coupled plasma mass spectrometry at the Isotope Laboratory in China University of Geosciences, Beijing. The standard-sample bracketing method was used to correct the instrumental mass bias and drift. An in-house solution (FZT) was used as the working standard. Analyses were performed in the low-resolution mode, simultaneously measuring 26Mg, 25Mg and 24Mg isotopes. Mg isotope ratios are reported by the delta notation as ‰ deviation relative to the DSM3 standard50:

where x refers to 25 or 26.

The internal precision was determined on the basis of ≥3 repeated runs of the same sample solution during a single analytical session and is better than ±0.10‰ (2 s.d.). The accuracy is determined by the measurements of synthetic solution (GSB-Mg) and USGS basalt standards (BCR-2). Multiple analyses of the synthetic solution (GSB-Mg) yield δ26Mg values ranging from –2.07 to –2.04‰, which is consistent with the preferred value of –2.05±0.05‰ (2σ). δ26Mg of BCR-2 is –0.17±0.06‰ (2σ), consistent with the published values51,52,53,54.

Sulphur isotope analysis

Traditional CAS extraction procedure typically requires >20 g of carbonate powder. Thus, it is impossible to collect enough sample powder from MT microspar without contamination from the host rock using the traditional method, because MT cracks are typically a few millimetres in width. To analyse CAS of MT microspar, we devised a new extraction procedure that only requires ∼1 g of carbonate powders for each sample. The validity of the new procedure was verified by analysing the same carbonate sample by using both the traditional and new procedures. Powders were carefully drilled from MT microspar only to a shallow depth so as to avoid the potential contamination from host rock. Often, multiple MT cracks in a polished slab were drilled to collect enough powder for CAS extraction. Sample powder was placed in a 50-ml centrifuge tube and were treated with 10% NaCl solution for 24 h to dissolve non-CAS sulphate. After supernatant removal, residues were washed with deionized water for three times. The above cleaning procedures were repeated at least three times, to ensure complete removal of non-CAS sulphate. The cleaned sample powder was dissolved in 40 ml of 3 N HCl. After 1 h of reaction, reaction tubes were centrifuged and the supernatants were collected. About 1–2 mg of nano-SiO2 was added into the centrifuge tube and then 10 ml of saturated BaCl2 was added to precipitate sulphate as barite. The use of nano-SiO2 was to facilitate barite collection. Barite precipitation was allowed to proceed for 48 h. After centrifuging, barite precipitate was washed by DI water for three times, to remove residual HCl, and then dried in an oven.

Sulphur isotopic compositions were measured at Indiana University on a Finnigan Delta V advantage gas source mass spectrometry fitted with a peripheral Costech elemental analyser for on-line sample combustion. Sulphur isotope compositions are reported as ‰ deviation from V-CDT, δ34S=(Rsample/RV-CDT−1) × 1,000, where R is the ratio of 34S/32S. Analytical error is ±0.1‰ (1σ) as determined from repeated analyses of samples and laboratory standards. The analytical results were calibrated using the standard NBS-127 (20.3‰) and three internal standards: a silver sulphide (ERE-Ag2S: −4.3‰), a chalcopyrite (EMR-CP: +0.9‰) and a barite (PQB2: +40.5‰).

Inductively coupled plasma optical emission spectrometer analysis

Elemental compositions were determined at Peking University on a Spectro Blue Sop inductively coupled plasma optical emission spectrometer fitted with a Water Cross-flow nebulizer. All analyses were calibrated by a series of gravimetric standards with different concentrations (ranging from 0.1 to 10 p.p.m.) that were run before sample measurements and between every 20 samples. The external reproducibility for the major and minor elements (Na, Mg, Al, K, Ca, Fe, Mn, Sr and S) is ±2%.

Additional information

How to cite this article: Shen, B. et al. Molar tooth carbonates and benthic methane fluxes in Proterozoic oceans. Nat. Commun. 7:10317 doi: 10.1038/ncomms10317 (2016).