The H2/CH4 ratio during serpentinization cannot reliably identify biological signatures

Serpentinization potentially contributes to the origin and evolution of life during early history of the Earth. Serpentinization produces molecular hydrogen (H2) that can be utilized by microorganisms to gain metabolic energy. Methane can be formed through reactions between molecular hydrogen and oxidized carbon (e.g., carbon dioxide) or through biotic processes. A simple criterion, the H2/CH4 ratio, has been proposed to differentiate abiotic from biotic methane, with values approximately larger than 40 for abiotic methane and values of <40 for biotic methane. The definition of the criterion was based on two serpentinization experiments at 200 °C and 0.3 kbar. However, it is not clear whether the criterion is applicable at a wider range of temperatures. In this study, we performed sixteen experiments at 311–500 °C and 3.0 kbar using natural ground peridotite. Our results demonstrate that the H2/CH4 ratios strongly depend on temperature. At 311 °C and 3.0 kbar, the H2/CH4 ratios ranged from 58 to 2,120, much greater than the critical value of 40. By contrast, at 400–500 °C, the H2/CH4 ratios were much lower, ranging from 0.1 to 8.2. The results of this study suggest that the H2/CH4 ratios cannot reliably discriminate abiotic from biotic methane.

Serpentinization produces molecular hydrogen (H 2 ), resulting from the oxidation of ferrous iron in olivine and pyroxene to ferric iron (Reaction (1)). Abiotic methane (CH 4 ) can be derived from reactions between H 2 and oxidized carbon (e.g., carbon dioxide) through Fischer-Tropsch type (FTT) synthesis (Reaction (2)). Molecular hydrogen and methane support microbial communities in hydrothermal fields [20][21][22][23][24][25][26][27] . Methane may be produced biologically by methanogenic archaea 28 . The identification of abiotic and biotic methane is essential to understand ultramafic ecosystems, which potentially contribute to the origin and evolution of life during early history of Earth and possibly other terrestrial planets.

Results
The H 2 /CH 4 ratios. Molecular hydrogen, methane, ethane, and propane were formed. At 311 °C and 3.0 kbar, the H 2 /CH 4 ratios ranged from 58 to 2,120, much higher than the critical value of 40 ( Fig. 1a,b). The ratios increased as a function of time, implying that rates of H 2 production are faster than the rates of CH 4 formation. In experiments using peridotite with initial grain sizes < 30 μ m, the H 2 /CH 4 ratios varied from 58 to 91. By contrast, for those using larger grain sizes (100-177 μ m), the H 2 /CH 4 ratios were much higher, from 360 to 2,120. At 400-500 °C and 3.0 kbar, the H 2 /CH 4 ratios decreased greatly, 0.1-8.2 (Fig. 1c). In experiments at 500 °C and 3.0 kbar using peridotite with initial grain sizes of < 30 μ m, the H 2 /CH 4 ratios increased during the first 20 days to a maximum value and then decreased slightly during the subsequent 16 days. This decrease suggests an increase of CH 4 production ( Table 1). A similar trend was also observed at 400 °C and 3.0 kbar with grain sizes of 42-59 μm, whereas it was not detected in experiments with larger grain sizes. Solid products. At 311 °C and 3.0 kbar, the major secondary hydrous mineral was fibrous chrysotile (Fig. 2a), whereas tabular shaped lizardite formed at 400 °C and 3.0 kbar (Fig. 2b). Serpentine was identified based on infrared spectra with stretching modes at 954 and 1087 cm −1 for the Si-O group and a stretching vibration at 3686 cm −1 for the -OH group (Fig. 2d) [37][38][39] . Chemical compositions of secondary minerals in HR61 were provided in an experimental study 40 , consistent with compositions of serpentine 41 . At 500 °C and 3.0 kbar, the secondary hydrous minerals produced were talc and lizardite. Talc is characterized by a stretching mode at 671 cm −1 for Si-O-Mg and a stretching vibration at 3677 cm −1 for the -OH group (Fig. 2d) 42 .

Discussion
The hydrocarbons produced in this study are probably abiotic, supported by the following evidence. First, blank experiments were performed at 311-500 °C and 3.0 kbar using peridotite loaded without any fluid. The quantities of H 2 and hydrocarbons were below the detection limit of gas chromatograph after 27 days of reaction time. It suggests that hydrocarbons were not released from the decomposition of organic matter and long-chain hydrocarbons in peridotite 43,44 . Otherwise, it would result in highly elevated hydrocarbons. Moreover, the log of the n-alkane concentrations is linearly correlated with the carbon numbers ( Fig. 3), which is consistent with the Schulz-Flory distribution predicted for FTT synthesis 31 . All these indicate that hydrocarbons were formed through reactions between H 2 and dissolved carbon dioxide from the atmosphere in the starting fluid. A plot of H 2 /CH 4 ratios as a function of temperature is illustrated in Fig. 4, showing that the H 2 /CH 4 ratios greatly depend on temperature. They reached their maximum values at ~300 °C, from 58 to 4,000 ( Fig. 4a) 31,45 . By contrast, the values were much lower at 400-500 °C, much less than 40 ( Fig. 4), resulting from the dramatic decrease in H 2 production and increase in CH 4 formation. The decrease in H 2 production may be induced by very slow rates of olivine serpentinization at temperatures higher than 350 °C [46][47][48] , supported by infrared spectra of solid products with a sharp peak centered at 503 cm −1 for the Mg-O group of olivine and a weak band at 3677 cm −1 for the -OH group of talc (Fig. 2d). It suggests that H 2 is mostly derived from orthopyroxene alteration. As indicated by experimental studies, the quantities of H 2 produced during orthopyroxene alteration at > 350 °C were one to two orders of magnitude less than those formed after olivine serpentinization at 300 °C 31,32 . Consequently, H 2 production at 400-500 °C decreases greatly. By contrast, CH 4 concentrations increased at higher temperatures (Table 1), which possibly results from sufficient Fe-Ni alloys that highly enhance CH 4 production 29 .
Initial grain sizes of peridotite greatly influence the production of H 2 and CH 4 , and the H 2 /CH 4 ratios. Smaller grain sizes result in larger quantities of H 2 and CH 4 ( Table 1). Grain sizes exert a strong influence on serpentinization rates, with smaller grain sizes for faster rates 48 . For experiments with the same run durations, peridotite with smaller grain sizes has larger reaction extents 48 . As suggested by an experimental study, the production of H 2 showed a positive correlation with reaction extents of serpentinized peridotite 34 , and consequently smaller grain sizes result in more H 2 . Larger reaction extents possibly lead to the formation of more catalytic minerals (e.g., Fe-Ni alloys), which could greatly enhance CH 4 production 29 .
Run durations have great effects on H 2 /CH 4 ratios (Fig. 1). At 311 °C and 3.0 kbar, the H 2 /CH 4 ratios increased with longer time, implying that rates of H 2 production are faster than rates of CH 4 formation. By contrast, for experiments at 400-500 °C with smaller grain sizes (e.g., < 30 and 42-59 μ m), the H 2 /CH 4 ratios first increased to a maximum value, and then they decreased slightly during the subsequent reaction time (Fig. 1c). It implies that rates of CH 4 production were slow at the onset of reactions, possibly resulting from insufficient catalytic minerals (e.g., Fe-Ni alloys). When reactions proceeded, more catalytic minerals formed, which promote CH 4 production,    Fluid compositions (e.g., dissolved silica) may dramatically influence the H 2 /CH 4 ratios. As indicated by an experimental study, basalt alteration at 300 °C produced H 2 concentrations approximately two orders of magnitude less than those after peridotite serpentinization, resulting in very low H 2 /CH 4 ratios, 0.04 49 . Consistently, fluids recharged from basalt-hosted hydrothermal fields have much lower H 2 /CH 4 ratios than those from peridotite-hosted hydrothermal fields 50 . It is possibly because basalt alteration releases one to two orders of magnitude more dissolved silica into hydrothermal fluids 49 . Silica impedes the production of magnetite 51 , and consequently H 2 production decreases greatly 52 . By contrast, for experiments at 400-500 °C, differences in H 2 between basalt and peridotite hydration are much less significant 49,53 , leading to comparable H 2 /CH 4 ratios (Fig. 4, Table 1).
As discussed above, the H 2 /CH 4 ratios during serpentinization can be greatly influenced by many factors, including temperature, initial grain sizes of peridotite, run durations, and the dissolved silica in hydrothermal fluids. The H 2 /CH 4 ratios of < 40 can be achieved at temperatures higher than 350 °C or in the presence of silica, which may not necessarily represent biological signatures. In hydrothermal fields, peridotite commonly experiences a retrograde metamorphism, and serpentinization may occur at a wide range of temperatures 5 . It indicates that the production of H 2 in hydrothermal fields can be greatly influenced by temperature. Additionally, high-temperature reactions (aside from serpentinization), microbial oxidation and sulphate reduction possibly affect H 2 production 54 , and consequently the H 2 /CH 4 ratios may be modified. All these indicate that the H 2 /CH 4 ratios cannot reliably identify abiotic and biotic methane.
Interestingly, methane produced in this study has δ 13 C values larger than − 30‰ (referenced to Pee Dee Belemnite, Table 1), consistent with isotopic compositions of abiotic methane 55 . By contrast, methane synthesized in the presence of elevated Fe-Ni alloys has very depleted δ 13 C values, much lower than − 30‰ 29,30 . Iron-Ni alloys are accessory minerals in serpentinites, typically less than 0.5%. Therefore, experiments conducted using elevated Fe-Ni alloy may not represent natural hydrothermal systems. As reported in an experimental study, δ 13 C values of methane greater than − 30‰ was detected in one experiment, whereas in the other experiment under the same condition, methane had δ 13 C values lower than − 30‰ 56 . In particular, the δ 13 C values of methane became more depleted with longer time 56 . Therefore, it is not clear whether stable isotopes of carbon can effectively identify abiotic and biotic methane.
All experiments were conducted in the high-pressure and high-temperature laboratory at Guangzhou Institute of Geochemistry, Chinese Academy of Sciences. Experimental procedures were essentially the same as those described in another experimental study 40 . The reactants and starting fluid were sealed into gold capsules, which were placed into the end of hydrothermal vessels, followed with a filler rod. After heating, the vessels were quenched to room temperature in cold water within 10 min.
The gas components in the gold capsules were analysed using an Agilent 7890A gas chromatograph at the State Key Laboratory of Organic Geochemistry, Guangzhou Institute of Geochemistry. The gold capsule was placed in a vacuum glass piercer, which was connected to a Toepler pump and a volume-calibrated glass pipe through vacuum line. The gold capsule was pierced by a steel needle in vacuum (with a pressure of less than 1 × 10 −2 Pa), and all of gas components were concentrated by a Toepler pump into the volume-calibrated pipe. The hydrocarbons were quantified using an external standard with an accuracy of less than 0.5%. The detailed analysis procedures have been reported in previous studies 40,59,60 .
After gas chromatography analyses, the remaining gas in the vacuum glass piercer and glass pipe, with an amount about 80% of the initial value, was taken with a syringe for gas chromatography-isotope ratio mass spectrometry analyses. The carbon isotope value of CO 2 reference gas was calibrated by NBS 22 oil as a reference using element analysis, combined with isotope ratio mass spectrum. Carbon isotope values of methane were calculated with CO 2 as a reference gas that was automatically loaded into the system at the beginning and the end of each analysis.
The surface morphology of solid products was characterized with a Zeiss Ultra 55 Field emission gun scanning electron microscope at Second Institute of Oceanography, State Oceanic Administration of China. Fourier transformed infrared spectroscopy analyses were performed using a Bruker Vector 33 FTIR spectrometer at Analytical and Testing Center of South China University of Technology. Infrared spectra were obtained at wavenumbers from 400 to 4000 cm −1 at a resolution of 4 cm −1 with 32 scans for each spectrum. The KBr pellets were prepared by mixing around 1 mg of sample powder with 200 mg of KBr.