|Home | About | Journals | Submit | Contact Us | Français|
The silicon isotope composition of chert has recently been used to study the historic evolution of the global ocean. It has been suggested that Precambrian cherts have much higher δ30Si values than Phanerozoic cherts do and that the former show an increasing trend from 3.5 to 0.85Ga, reflecting a decrease in ocean temperatures. However, cherts have various origins, and their isotopic compositions might be reset by metamorphic fluid circulation; thus, different types of cherts should be distinguished. Here, we present a new set of δ30Si data for cherts from early and middle Proterozoic carbonate rocks from Northern China. We found that cherts of 1.355–1.325Ga show a peak range of 2.2–3.9‰. Based on these results, we propose that from the Archean to the middle Proterozoic, there was a drastic decrease in silicon content and an increase in the δ30Si value in ocean water due to a temperature decrease and biological activity increase. After that period, the silicon content of the ocean was limited to a low level by a high degree of biological absorption, and their δ30Si values varied in a small range around a significantly lower value.
Chert is a sedimentary rock composed mainly of microcrystalline quartz. It occurs in sedimentary strata from the early Archean to the present. Its chemical and isotopic characteristics have been widely investigated to study the conditions of its formation. Recently, Si isotopes in chert have been used as an important tracer of environmental conditions in the global ocean1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18. Song and Ding (1990) suggested distinguishing sedimentary facies of chert with silicon isotope compositions1. Ding et al. (1996) indicated that different genetic types of cherts have different silicon isotope characters2. They discussed variations in silicon content and silicon isotope composition in marine water according to the data from chert formed in shallow marine environments. According to the silicon isotope variation of chert, Robert and Chaussidon (2006) developed a temperature-evolution curve for the Precambrian ocean3. It has been suggested that Precambrian cherts have much higher δ30Si values than Phanerozoic cherts and that the former show a generally increasing trend from 3.5 to 0.85Ga, thus reflecting a decrease in seawater temperature3. However, these statements have been challenged because cherts can have various origins and their isotopic compositions might have been reset by metamorphic fluid circulation4,5,6,7,8,9. Thus, different types of cherts need to be considered separately4,9,10,11,12 because their isotope composition might record various distinct processes13,14,15,16,17,18. In addition, the factors that affect silicon isotope composition of chert have been investigated in detail13,14,15,16,17,18. It was found that peritidal cherts are enriched in 30Si, but that basinal cherts, which are associated with banded iron formations (BIF), are depleted in 30Si; this difference is attributed to Si having been derived from hydrothermal sources in the BIFs10. Now, it is widely accepted that the formation of chert is a complicated process and that the Si isotope composition of chert is dependent on the relative contributions of various Si sources and the effects of different forming processes9,10,11,12,13,14,15,16,17,18. However, the relative contributions of the various Si sources to the global ocean vary throughout geological history. Thus, specific types of chert might be representative for reconstructing the Si isotope composition of the ocean for different geological periods. For example, during the Archean period, sea-floor weathering and submarine hydrothermal fluids dominated the Si input to the ocean, and continent inputs were negligible; thus, the Si isotope compositions of quartz bands in BIF are likely the best approach for tracing the Si isotope compositions of the ocean. In contrast, in the Proterozoic and Phanerozoic periods, the Si input from the continent to the ocean became dominant, and the Si input from sea-floor weathering and submarine hydrothermal fluids became less prevalent; thus, peritidal chert may be more significant in tracing the Si isotope composition of the ocean. To date, many studies have investigated the silicon isotope composition of silica in BIF to examine the environmental conditions of the Archean ocean2,3,4,5,6,7,8,9,11,12,13,14. However, studies specifically targeting the silicon isotope composition of cherts formed in shallow marine environment are relatively rare10. More systematic and detailed Si isotope studies on this aspect must be undertaken to reconstruct the silicon content and silicon isotope composition in the ancient ocean and understand the historic evolution of the global ocean.
We collected a suite of samples of carbonate rocks containing chert bands and nodules from the early Proterozoic Hutuo Group19 (2.35–2.20Ga) and the middle Proterozoic Changcheng and Jixian Systems20,21 (1.63–1.20Ga) in Northern China (Fig. 1). These cherts were selected because they showed good preservation of their original structure and composition (Figs 2 and and3);3); i.e., no obvious effects of metamorphism or weathering were found in the selected samples. The Si and O isotopic compositions of the chert and the C and O isotopic compositions of dolomite were studied systematically.
The Si and O isotope compositions of the cherts and the C and O isotope compositions of dolomites obtained in this study are listed in Table 1 and shown in Fig. 4. All isotope compositions are reported in delta-notation relative to a standard material, i.e.,
where A represents the isotope, R represents the isotope ratio, the subscript Sa refers to the sample and St refers to the standard. For the Si isotope compositions, A means 30Si, R means 30Si/28Si, St means NBS-28. For the O isotope composition, A means 18O, R means 18O/16O, and St can be either V-SMOW (for chert and dolomite) or V-PDB (for dolomite). For the C isotope composition, A means 13C, R means 13C/12C, and St means V-PDB.
The obtained δ13CV-PDB values of dolomites range from −2.5‰ to 1.9‰, with an average of −0.36±0.82 (1SD) ‰. The obtained δ18OV-PDB values of dolomites range from −12.9‰ to −3.5‰, with an average of −7.11±2.15 (1SD) ‰. These values are similar to those reported in previous studies22,23, reflecting their normal sedimentary origin in a shallow marine environment. The δ18OV-PDB values indicate that the dolomite had been influenced to some degree by diagenesis.
The δ18OV-SMOW values of the chert samples range from 14.3‰ to 27.8‰, with an average of 23.0±3.5 (1SD) ‰. These results also show that they were formed in a shallow marine environment but might have been slightly affected by diagenesis.
The δ30Si values of the cherts collected in this study range from 0.1‰ to 3.9‰. Among them, the early Proterozoic cherts show lower δ30Si values, ranging from 0.1‰ to 1.3‰, with an average of 0.75±0.43‰ (1SD), compared to middle Proterozoic cherts ranging from 0.5‰ to 3.9‰, with an average of 2.22±0.74‰ (1SD). Furthermore, a peak range in δ30Si [2.2–3.9‰, averaging 3.12±0.56‰ (1SD)] is observed in the cherts of 1.325–1.355Ga (Fig. 4). In addition, the δ30Si variation of chert can also be observed on a millimetre scale. For example, 4 chert bands in sample JX-26 and 5 chert bands in sample JX-27 show δ30Si variations of 2.3~3.9‰ and 3.1~3.6‰, respectively (Table 1).
The increasing trend of δ30Si from the early Proterozoic to middle Proterozoic is consistent with the trends reported by previous studies2,3,5,6,7,8,9,10. However, together with the new data provided in this study, the middle Proterozoic δ30Si peak becomes a prominent feature in the Si isotope record of Precambrian cherts, which implies major environmental variations in the ancient ocean (Fig. 4).
Combining data obtained using SiF4 and MC-ICP-MS methods in this study and previous studies10,24,25,26,27,28,29,30, the δ30Si variation trend from the late Archean to present is plotted for chert that formed in shallow marine environments (Fig. 5). The upper limit of δ30Si values of chert increases gradually from 1.8‰ at 2.53Ga to 3.9‰ at 1.335Ga and then decreases drastically to 2.0‰ at 1.104Ga. After 1.104Ga, the upper limit of δ30Si values for chert fluctuates between 1.5‰ and 2.5‰.
Chert is normally formed by the recrystallization of a precipitated amorphous silica precursor in the diagenetic process. To test the possibility of using O and Si isotope compositions of chert to trace those of contemporary marine water, the relationship between the δ18O and δ30Si values of chert and those of the amorphous silica precursor must be evaluated first.
It is known that the O isotope composition of silica would be reduced during diagenesis due to two factors. First, the chert nodular and band in the limestone are considered to form in the groundwater of mixed meteoric-marine coastal systems during diagenesis31. Due to the involvement of meteoric water, the O isotope composition of groundwater would be lighter than that of contemporary marine water, causing a reduction in the δ18O value of chert to some extent. Second, as the diagenetic temperature is normally higher than that of marine water, the O isotope fractionation between silica and water at the diagenetic stage would be smaller than that in the precipitation stage, which would cause a δ18O reduction in the chert. As shown by the microscopic and SEM examinations (Fig. 3), the slight reduction in the δ18O value of chert indicates that the studied cherts were formed in the diagenetic process and their δ18O values cannot represent those of the amorphous silica precursor.
In contrast to the O isotope composition, the silicon isotope composition would not change during early diagenesis for following reasons: (1) Chert bands and nodules are commonly confined in layers of sedimentary strata, which indicate that silica does not move over long distances during diagenesis. (2) The amorphous silica precursor is the dominant form of silica in dolomite rocks and the Si content in groundwater is rather limited. (3) δ30Si variations at the millimetre scale or even the micrometre scale3,8,13,14,32,33 are preserved in cherts, which rules out the possibility of the large-scale mixing of silicon during diagenesis. Thus, the δ30Si value of chert can be used as a representative of the amorphous silica precursor to trace the silicon isotope composition of contemporary seawater5.
In Fig. 5, the inferred variation trend in δ30SiSW (δ30Si value of seawater) is shown. The δ30SiSW is calculated from the equation δ30SiSW=δ30SiCh−Δ30SiCh-SW, where δ30SiCh is the δ30Si value of chert and Δ30SiCh-SW is the relative Si isotope enrichment of chert to seawater. According to the theory of isotope fractionation, Δ30SiCh-SW is temperature dependent. Thus, to determine a proper Δ30SiCh-SW value, precipitation temperatures of the amorphous silica precursor of chert should be evaluated beforehand.
Based on the O isotope composition of chert, Robert and Chaussidon (2006)3 suggested the seawater temperatures were 70°C at 3.5Ga and 35°C at 0.85Ga. However, the inferred high Paleoarchean temperatures are controversial34,35,36,37,38, partly because δ18O determinations of the Paleoarchean temperature rely on the assumption that δ18O of the Archean ocean was similar to that of an ice-free modern ocean. However, it has been suggested by a number of researchers that the δ18O value of the global ocean could have varied significantly over time35,36,37,38. Recently Hren et al.34 studied the δ18O and δD values of the 3.42Ga Buck Reef Chert rocks in South Africa and found that the chert with the highest δ18O was formed in equilibrium with waters below 40°C. Blake et al. (2010) studied the oxygen isotope composition of phosphate in 3.2–3.5Ga sediments of the Barberton Greenstone Belt and found that the phosphate with the highest δ18O value was formed in seawater at a temperature range from 26°C to 35°C39. According to these results, a temperature range of 35°C–40°C is more acceptable for the Archean ocean.
For the temperature of the Proterozoic ocean, fewer results have been reported. Based on the O isotope composition of chert, the temperatures in the Proterozoic ocean are estimated in the range of 35–60°C3. According to the δ18O values of chert and dolomite obtained in this study, the diagenetic temperature can be estimated. Assuming that diagenetic water has a δ18O value of −10‰, and using the O isotope fractionation equation (1000lnαDol-H2O=3.20×106T−2−2.0) of Northrop and Clayton (1966)40, we obtained a diagenetic temperature range of 25~56°C (averaging 36±7°C) for dolomite in the early Proterozoic and a diagenetic temperature range of 12~50°C (averaging 26±9°C) for dolomite in the middle Proterozoic. Assuming the diagenetic water still has a δ18O value of −10‰, and using the O isotope fractionation equation (1000lnαChert-H2O=3.09×106T−2−3.29) of Knauth & Epstein (1975)41, we obtained a diagenetic temperature range of 19~43°C (averaging 34±7°C) for chert in the early Proterozoic and 1~63°C (averaging 17±15°C) for chert in the middle Proterozoic. The calculated average diagenetic temperature (34°C) for early Proterozoic chert is slightly lower than that (36°C) for early Proterozoic dolomite, and the calculated average diagenetic temperature (17°C) for middle Proterozoic chert is significantly lower than that (26°C) for middle Proterozoic dolomite. These observations may be caused by a difference between chert and dolomite during O isotope exchange process with diagenetic solution. The cherts are more resistant to O isotope exchange than dolomite in the diagenetic process. Thus, the calculated diagenetic temperature for dolomite may be more representative. Based on these considerations and assuming the diagenetic temperature is a little higher than the sedimentary temperature on the sea floor, we estimate the ocean temperature as 30°C in the early Proterozoic and 20°C in the middle Proterozoic.
Concerning Δ30SiCh-SW, there has been a number of investigations on Si isotope fractionation during abiotic silica precipitation15,16,17,42,43,44,45. Early experimental studies on abiotic solid–fluid silicon isotope fractionation yielded Δ30Sisolid–fluid values ranging from −2.0‰ to −1.0‰42,44. Geilert et al. (2014) performed seeded silica precipitation experiments using flow-through reactors in the 10–60°C temperature range to quantify the silicon isotope fractionations during controlled precipitation of amorphous silica from a flowing aqueous solution15. The obtained Δ30SiSilica-solution values were −2.1‰ at 10°C, −1.2‰ at 20°C, −1.0‰ at 30°C, −0.5‰ at 40°C, 0.1‰ at 50°C, and 0.2‰ at 60°C. These results can be used to calculate δ30Si values of ocean water in different geological periods.
Assuming that the temperature of the ocean is 40°C, 30°C and 20°C in the Archean, early Proterozoic and since the middle Proterozoic, respectively, and Δ30SiCh-SW are −0.5‰, −1.0‰ and −1.2‰ in the Archean, early Proterozoic and since the middle Proterozoic, respectively, δ30Si values of ocean water are calculated from δ30Si values of chert (Fig. 5).
Figure 5 shows that the upper limit of inferred δ30SiNBS-28 values in ocean water increases gradually from 2.8‰ at 2.53Ga to 5.1‰ at 1.335Ga and then decreases drastically to 3.2‰ at 1.104Ga. After 1.104Ga, the upper limit of inferred δ30SiNBS-28 values in ocean water fluctuates between 2.7‰ and 3.7‰.
As shown in Fig. 6, during the Archean period, the input sources of dissolved Si to the ocean are submarine hydrothermal fluid and sea-floor weathering, and the output paths are chemical precipitation (to form C cherts) and silicification of the precursor sediments or rocks (to form S cherts)4,5,6. The Si concentration in ocean water remains in its saturated concentration at a given temperature, but the δ30SiSW increases gradually due to Si isotope fractionation between dissolved Si and precipitated SiO2. When a steady state is reached, δ30SiOut (the average δ30Si value of all output Si) will be equivalent to δ30SiIn (the average δ30Si value of input Si), and δ30SiSW will be equal to (δ30SiOut -Δ30SiOut -SW), where Δ30SiOut -SW is the relative silicon isotope enrichment of the output Si to the ocean water (δ30SiOut - δ30SiSW). Because the average δ30Si value of Si in the submarine hydrothermal fluid and Si from sea-floor weathering is ~−0.3‰ and Δ30SiOut-SW is ~−0.5‰ at a temperature of 40°C15, the δ30SiSW value of that period is approximately 0.2‰.
The Si cycle in the modern ocean (Fig. 6) is quite different46. First, dissolved Si from the continents (in rivers47 and groundwater) has become a dominant input source (6.4Tmol Si/a) to ocean Si, and the Si input from submarine hydrothermal fluid (0.6Tmol Si/a) and sea-floor weathering (1.9Tmol Si/a) has become less significant46. δ30Siin is calculated as ~0.78‰ using the equation
In the equation, f represents the relative fraction of each Si source and the subscripts Cont, SFW and SHF indicate continent, sea-floor weathering and submarine hydrothermal fluid, respectively.
Second, the biological absorption of Si has become a dominant path for Si output from the ocean and Si contents in modern ocean water (0.05mg/L~0.2mg/L for shallow seawater and 0.3mg/L~3.5mg/L for deep seawater) are 2 orders of magnitude lower than those of the saturated concentration in seawater2. At steady state, the amount and δ30Si value of output Si from ocean water would be equal to those of input Si. Thus, δ30SiOut would also be ~0.78‰ at present.
The silicon isotope fractionations of diatoms-seawater and sponges-seawater have been experimentally studied. The determined Δ30SiDiatom-SW is commonly −1.0~−1.1‰48,49,50, but Δ30SiSponge-SW varies from −1.1‰ to −3.7‰51,52. Because diatoms are much more abundant in the ocean than sponges, we assume Δ30SiOut-SW in the modern ocean is ~−1.2‰. From the above estimation, the δ30SiSW value of the modern ocean can be calculated as ~1.98‰, which is very close to the value (1.9‰) of surface ocean water inferred previously53.
Many papers have reported the δ30Si values of cherts formed during the Proterozoic1,2,3,8,9,10,24,25,26,32, but no model of the Si cycle in the Proterozoic ocean has been presented. Here, we present a conceptual model for the Si cycle of the Proterozoic ocean based on known and inferred boundary conditions (Fig. 6). Similar to the conditions in the Archean ocean, submarine hydrothermal fluid and sea-floor weathering are still important input sources of dissolved Si to the Proterozoic ocean, but the amounts of these inputs decreased as the hydrothermal activity and ocean temperature decreased from their Archean to Proterozoic values. Further, the input of dissolved Si from the continents became significant as supercontinents appeared in the early Proterozoic. For the output of dissolved Si from the ocean in the Proterozoic Eon54, chemical precipitation was still a major pathway, but biological absorption may have also played a significant role.
The Si concentration in ocean water remains in its saturated concentration at a given temperature (30°C for the early Proterozoic and 20°C for the middle and late Proterozoic). When a steady state is reached, δ30SiOut will be equivalent to δ30SiIn, and δ30SiSW will be equal to (δ30SiOut −Δ30SiOut -SW), where Δ30SiOut -SW is −1.0‰ for the early Proterozoic and −1.2‰ for the middle and late Proterozoic. The estimated δ30SiSW value of that period would be approximately 1.47‰ (Fig. 6).
From the discussion above, extreme values of δ30Si for chert and δ30Si for seawater cannot be explained for an ocean at steady state conditions. Thus, this peak in δ30Si indicates an extraordinary period at non-steady state suggesting a scenario at a transition stage.
From the Archean to present, there should be a transition period of the Si cycle in the ocean. In that period, the DSi in ocean water is reduced by approximately 2 orders of magnitude below that of the saturated concentration in seawater2, and the δ30SiSW value first rises from ~0.2‰ to ~5.1‰ and then decreases to ~1.98‰. The rise of δ30SiSW is caused by Rayleigh fractionation when SiO2 precipitates from ocean water (Fig. 7). One mechanism that causes the DSi reduction in ocean water should be a decrease in ocean temperature. As temperature decreases, the saturated Si concentration in ocean water would be reduced, causing additional SiO2 precipitation. The saturated SiO2 concentration in ocean water is ~363.5mg/L, ~221.2mg/L and ~178.9mg/L at 40°C, 30°C and 20°C, respectively55. When the temperature of seawater decreases from 40°C to 20°C, the fraction of dissolved Si remaining in the seawater (f) will be reduced to ~0.687. In the Rayleigh fractionation process, it will cause an increase of ~0.2‰ in δ30SiSW. It seems that the decrease in seawater temperature alone cannot explain the significant increase in the δ30SiSW value, and other mechanisms should be considered. Another mechanism causing a DSi decrease in ocean water is the increase of Si absorption activities by biological species, which can reduce the DSi in ocean water 2 orders of magnitude lower than that of the saturated concentration. In the Rayleigh fractionation process, the combined effect of these two types of mechanisms can cause δ30SiSW to increase ~4.0‰ when f is reduced to 0.01. It is known that diatoms and radiolarians are Si-fixing organisms that were active in the Phanerozoic46; sponges were active in the Phanerozoic46 and the later Proterozoic54. Assuming their appearance is the start of a drastic decrease of DSi in ocean water, we should observe a δ30SiSW peak value in the later Proterozoic or early Phanerozoic. However, according to the data here, the δ30SiSW peak appeared in the middle Proterozoic (1.325~1.355Ga) instead, which indicates the drastic decrease in ocean water DSi happened prior to 1.355Ga.
It is known that microbes were the dominant biological species in the Precambrian. Stromatolites are found in early Archean strata from 3.5Ga56 and are very well developed in Proterozoic strata19,20,21,22,23,57. REE and C, O, Nd isotope compositions have been used to study the formation conditions of stromatolite-bearing sediments, particularly the effect of biological activities22,23,56,57,58. The early and middle Proterozoic chert-bearing dolomites investigated in this study are all rich in stromatolites, showing a close correlation between silica precipitation and biological activities. Moreover, macroscopic eukaryotic fossils were recently discovered in the 1.56Ga Gaoyuzhuang Formation in the Yanshan area of Northern China59. If some Proterozoic species are capable of absorbing or precipitating Si from ocean water, the Si content in the Proterozoic ocean water would be drastically reduced causing δ30SiSW to rise significantly. Thus, the high peak in δ30SiSW values in the middle Proterozoic ocean water may reflect a drastic reduction in Si content caused by a rapid increase in biological activity in the ocean. After that peak period, the DSi reduction rate in ocean water decreased gradually, and the δ30SiSW value decreases to a significantly lower value at steady state.
For Si and O isotope analysis, ~100mg of a chert band or nodule was selected from a polished section of each specimen. The sample was crushed and ground to a powder of −200 mesh. Then, the sample powder was reacted with 6N HCl in Teflon beakers to dissolve small amounts of carbonate. The remainder was washed at least in triplicate with Milli-Q water. Then, the remainder was transferred to a Pt crucible, dried at 105°C in an oven and then calcined at 1000°C in a muffle furnace to remove organic C impurities.
Oxygen isotope analyses were carried out using the BrF5 method (Clayton and Mayeda, 1963)60, and silicon isotope analyses were carried out using the SiF4 method (Ding, 2004)61. Approximately 10mg of pretreated chert was placed in a Ni reactor in a metal vacuum line and reacted with BrF5 at a temperature of approximately 500°C to produce gaseous O2 and SiF4.
The O2 gas was separated from SiF4, BrF5 and BrF3 by evaporating at liquid nitrogen temperature. Then, O2 gas was converted to CO2 by reacting with a carbon rod at 700°C. Finally, the CO2 gas was collected for O2 isotope measurement.
SiF4 was separated from the BrF5 and BrF3 by evaporating at dry ice-acetone temperature. The separated SiF4 was purified further by passing it through a Cu tube containing pure Zn particles at a temperature of 60°C. This procedure removed trace amounts of the remaining active F-bearing compounds (BrF5 and BrF3). Then, the purified SiF4 was collected for silicon isotope measurement.
The isotopic measurements were carried out with a MAT-253 mass spectrometer.
For O2 isotope measurement, the NIST Standard Reference Material for O isotopes (NBS-28) was used directly as the working standard in this study. The precision of the O isotope measurement is better than±0.2‰ (2σ). The O isotope compositions of all samples are reported as δ18O values relative to the V-SMOW standard.
For Si isotope measurements, international reference material for Si isotopes (NBS-28) and two Chinese national standards for Si isotopes (GBW04421 and 04422) were used as working standards in this study. The long-term reproducibility of the silicon isotope measurements is better than±0.1‰ (2σ). The silicon isotope compositions of all samples are reported as δ30Si values relative to the NBS-28 standard.
The continuous-flow isotope-ratio mass spectrometric method was used for C and O isotope analysis of dolomite62. The system consists of a Thermo-Finnigan GasBench II equipped with a CTC Combi-Pal auto-sampler and linked to a Finnigan MAT 253 mass spectrometer.
Approximately 100mg of dolomite was taken from the specimen and ground to a powder of −200 mesh. Approximately 100μg of dolomite powder was loaded manually into a 12ml round-bottomed borosilicate exetainer and sealed using butyl rubber septa. Four national reference materials (GBW04405, GBW04406, GBW04416 and GBW04417) were routinely loaded. The exetainers were automatically flushed with grade 5 He by penetrating the septa using a double-hole needle at a flow rate of 100mL/min. Then, 4–6 drops of phosphoric acid were deposited in each exetainer. The exetainers were placed onto an aluminium tray and kept at 72°C for 24h. Subsequently, the sample gas was introduced into the mass spectrometer through the standard 100μL sample loop, CO2 was separated from other components using a gas chromatographic column heated to 70°C, and the peak corresponding to this CO2 was passed via an open split to the mass spectrometer.
The calculated external precision is typically±0.2‰ (2σ) for δ13C and δ18O. The C isotopic compositions of the dolomite samples are reported as δ13C values relative to the V-PDB standard. The O isotopic compositions of the dolomite samples are reported as δ18O values relative to V-PDB and V-SMOW standards.
How to cite this article: Ding, T. P. et al. The δ30Si peak value discovered in middle Proterozoic chert and its implication for environmental variations in the ancient ocean. Sci. Rep. 7, 44000; doi: 10.1038/srep44000 (2017).
Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
We thank H. K. Li, T. R. Song and J. S. Wu for their kind help in the geological investigation and sample collection. This work was supported by the National Natural Science Foundation of China (Item No. 41273016).
The authors declare no competing financial interests.
Author Contributions T. P. Ding designed the research and wrote the manuscript. J. F. Gao, S. H. Tian and Y. Zhao are responsible for sample collection and treatment. D. F. Wan and C. F. Fan completed the analytical work for the Si, O and C isotopes. J. X. Zhou completed the SEM analysis of chert.