Introduction

H2S is a harmful gas usually generated in deep marine carbonate gas reservoirs1,2,3,4,5,6, which poses significant challenges for the safe production of natural gas. High concentrations of H2S in gas reservoirs are mainly produced by thermochemical sulfate reduction (TSR)1,7,8,9,10,11, during which sulfate is reduced by organic matters and/or hydrocarbons to H2S and CO2. In general, hydrocarbons with higher carbon number will react more readily with sulfate12,13,14. The reaction equations can be expressed as follows:

$${\text{SO}}_{{4}}^{{{2} - }} + {\text{CH}}_{{4}} \to {\text{CO}}_{{3}}^{{{2} - }} + {\text{H}}_{{2}} {\text{S}} + {\text{2H}}_{{2}} {\text{O}}$$
$${\text{3SO}}_{{4}}^{{{2} - }} + {\text{4C}}_{{2}} {\text{H}}_{{6}} \to {\text{CO}}_{{3}}^{{{2} - }} + {\text{3H}}_{{2}} {\text{S}} + {\text{4CH}}_{{4}} + {\text{CO}}_{{2}} + {\text{H}}_{{2}} {\text{O}}$$
$${\text{3SO}}_{{4}}^{{{2} - }} + {\text{2C}}_{{3}} {\text{H}}_{{8}} \to {\text{3CO}}_{{3}}^{{{2} - }} + {\text{3H}}_{{2}} {\text{S}} + {\text{2CH}}_{{4}} + {\text{CO}}_{{2}} + {\text{H}}_{{2}} {\text{O}}$$
$${\text{3SO}}_{{4}}^{{{2} - }} + {\text{4C}}_{{4}} {\text{H}}_{{{1}0}} + {\text{H}}_{{2}} {\text{O}} \to {\text{5CO}}_{{3}}^{{{2} - }} + {\text{5H}}_{{2}} {\text{S}} + {\text{8CH}}_{{4}} + {\text{3CO}}_{{2}}$$

The effect of TSR has been studied extensively by characterising natural gas components, stable isotopes such as carbon, hydrogen, and sulfur, as well as inclusions and sulfur-bearing minerals9,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30. Cai et al.23,24 concluded that CH4 can be oxidised when the dryness coefficient is greater than 0.97 and presented fractionation equations for TSR in the presence of methane and ethane27, 28. Machel et al.4 proposed an initiation temperature of 100–140 °C for TSR, and Worden et al.31 thought that the initiation temperature for TSR was 140 °C. Amrani et al.29 and Meshoulam et al.32systematically investigated the isotopic composition of sulfur compounds during TSR and proposed that the composition of sulfur isotopes may reflect the degree of TSR. Hao et al.1, Liu et al.20,21, and Cai et al.27 characterised geochemical properties of marine gas in the Sichuan Basin, China, and found that heavy hydrocarbons were preferentially involved in the TSR reaction compared with CH4, which can increase the dryness coefficient of natural gas. TSR was also found to gradually cause partial carbon isotope reversal of the positive alkane gas series20. Liu et al.21 discovered that hydrogen isotope fractionation for CH4 generated by TSR via hydrogen isotope exchange between water and hydrocarbons was greater than that for CH4 directly generated from kerogen. Although a carbon isotope fractionation model for alkane gas in H2S-bearing gas reservoirs, such as the commonly observed pattern of the carbon isotope series of CH4 and C2H6 that changes from positive to reversed and then back to positive1, 20, 33, has been established, the hydrogen isotope fractionation of alkane gas has not been explored by laboratory pyrolysis. Liu et al.34 proposed that the rock salt and/or brine may play an important role for the occurrence of TSR based on the formation mechanism of H2S-enriched gas reservoirs in the Sichuan Basin, China. Cross et al.35 reported, based on laboratory simulations, that temperature was the key factor for the occurrence of TSR, whereas pressure had a minor impact. Pan et al.12 conducted a high-temperature, long-time, step-by-step thermal cracking simulation with organic matter and Fe2O3, MgSO4, and a mixture of both. They found that CH4 was produced by TSR in the presence of heavy hydrocarbons and that isotope fractionation became more pronounced with increasing carbon number. Based on a thermal cracking simulation, Zhang et al.36,37 proposed that the initiation temperature for TSR was affected by the chemical composition of crude oil, and low molar ratios of water and MgSO4 were favorable to the reaction. MgSO4 affects pH, which in turn increases the concentration of the active sulfate species HSO4–38,39. TSR eventually oxidised CH4 into carbon dioxide19,20,21. Zhao et al.40 experimentally simulated the TSR reaction with crude oil, bitumen, and different kinds of kerogen under presence of anhydrite and MgSO4, and found that TSR preferentially modified low molecular hydrocarbons, and type III kerogen was least reactive during the TSR process.

However, most of these laboratory simulations of TSR have focused on the chemical composition, and carbon isotope composition of the pyrolysis12, 36,37,38,39,40,41. The effects of oil cracking and TSR on the hydrogen isotopic composition of alkane gas have rarely been studied42. In addition, participation of water in the reaction and involvement of hydrogen from other sources, e.g. water, can result in a hydrogen isotope fractionation of alkane gas that may differ from its carbon isotope fractionation21.

In this paper, crude oil, nonane (C9H20, C9), and methylnaphthalene (C11H10, MN) were used to simulate TSR alteration at different thermal stages. The chemical and carbon isotopic composition of gaseous products from thermal cracking and TSR of crude oil were analysed to elucidate the impact of TSR. In addition, this is the first study that considered the impact of the presence of water during TSR alteration on the hydrogen isotope fractionation of alkane gas. Therefore, the effects of thermal cracking and TSR on the fractionation of carbon and hydrogen isotopes of alkane gas were studied to explore the isotopic evolution of carbon and hydrogen caused by TSR, aiming to provide experimental evidence to better understand the isotopic composition and variation of gases in H2S-bearing reservoirs.

Methods

Samples

Crude oil sample used in the simulation experiments were collected from the TK772 well within the Ordovician Yingshan Formation (O1−2y) in the No.7 District of the Tahe Oilfield, Tarim Basin, western China, at a depth of 5,557.5–5,591.5 m. The content of saturated hydrocarbons, aromatic hydrocarbons, non-hydrocarbons, asphaltene, and sulfur of the crude oil were measured as 25.28%, 28.66%, 12.23%, 17.26%, and 2.46%, respectively, with a carbon isotope composition of − 32.8‰ (the hydrogen isotopic composition was not measured). The purities of C9 (C9H20, nonane) and MN (C11H10, methylnaphthalene) were 99.99%. The reaction reagents and experimental buffer in this paper were manufactured by the Aladdin Reagent Shanghai Co., Ltd.

Experimental procedures and product analyses

It has been reported that the most suitable temperature for TSR of crude oil in a closed system is approximately 360 °C and that heating time and pH are important factors affecting the occurrence of TSR10,36. To simulate the evolution of reaction products during thermal cracking and TSR, as well as the change in isotopic composition caused by TSR, crude oil samples were heated at a temperature of 360 °C for 4, 10, 24, 40, 72, and 219 h. For each condition, the experiment was conducted with crude oil only, and a mixture of crude oil, buffer and MgSO4. In the pyrolysis system, we added a mixed salt solution (including 5.61 g MgCl2, 10.01 g NaCl and 100 ml distilled water, 350 mg solution per gold capsule) to the gold capsule. A mixture of silica and talc (1:1, about 60 mg per gold capsule) was used as a buffer to maintain a relatively stable pH of approximately 3–5. To determine the effects of temperature on TSR, experiments were also conducted using a solution of crude oil, buffer, and MgSO4 at temperatures of 350 °C and 370 °C. For each experiment, about 10 mg of crude oil, MgSO4 solution, and buffer were combined in a gold capsule (60 mm in length, 6 mm in diameter) under argon gas. The gold capsules were placed in different autoclaves that were connected to each other. The pressure of the experimental system was maintained at 50 MPa to simulate the conditions under which most TSR occurs. After the reaction temperature was increased from room temperature (18 °C) to 200 °C at a rate of 20 °C/h, it was kept constant at 200 °C for half an hour to stabilise the system temperature. The temperature was then increased to the target temperature at a rate of 20 °C/h, and held constant for the pre-set reaction time. The gold capsules were then taken out of the autoclaves and quenched. Fluctuations of temperature and pressure in the autoclaves were 0.5 °C and 1 MPa, respectively. To compare and analyse the effects of TSR, experiments were also conducted on nonane (C9) and methylnaphthalene (MN) at 360 °C following the same procedures. The amount of C9 and MN added in the experiments were both about 10 mg.

The chemical compositions of gas products were analysed using an Agilent 6890 N gas chromatograph equipped with a PoraPLOT Q column (30 mm × 0.25 mm × 0.25 μm). Helium was used as the carrier gas. The oven temperature was programmed as follows: an initial temperature of 50 °C for 2 min, raised to 180 °C at the rate of 4 °C/min, and held constant at 180 °C for 15 min. Deviation between replicate analyses was less than 1%. The stable carbon isotopic composition was analysed with an Isochrom II GC-IRMS isotope ratio mass spectrometer equipped with a  PoraPLOT Q column. Helium was used as the carrier gas. The oven temperature program was started at 50 °C, held for 3 min, raised to 150 °C at a rate of 15 °C/min, and held at 150 °C for 8 min. Each sample was analysed twice with a deviation between the analyses of less than 0.3‰. Stable hydrogen isotope analyses were conducted with a Finningan DeltaPlus XL isotope ratio mass spectrometer equipped with a Dim-type column. The oven temperature was programmed as follows: an initial temperature of 40 °C held for 8 min, raised to 110 °C at a rate of 3 °C/min, then held at 110 °C for 2 min. Each analysis was conducted twice, with a deviation between the analyses of less than 4‰. Detailed results are listed in Tables 1 and 2.

Table 1 Yields and chemical compositions of main products of the thermal cracking simulation experiments.
Full size table
Table 2 Carbon and hydrogen isotope compositions of the products of the thermal cracking simulation experiments.
Full size table

Results

Yields of reaction products under different conditions

The yields of all gaseous alkanes produced from thermal cracking of crude oil at 360 °C increased with longer heating time, with CH4 exhibited the highest yields of 1.00–8.09 ml/g. The yields of H2S, CO2 and H2 also increased with increasing reaction time, but no H2S was detected in the first 40 h (Table 1). During thermal cracking of a mixture of crude oil, MgSO4, and buffer at 360 °C (M15–M20), yields of CH4 and C2H6 increased at first and then decreased as the reaction continued. The yield of CH4 was in the range of 9.94–40.52 ml/g. The yields of C3H8, C4H10, and other heavy gaseous hydrocarbon products gradually decreased with longer heating time. The yield of H2S increased at first and then decreased in the range of 47.92–378.63 ml/g. The CO2 yield increased from 41.38 to 190.38 ml/g with time. The yield of H2 decreased from 0.23 to 0.12 ml/g (Table 1) with increasing heating time. These results suggest that the presence of sulfate significantly increased the yields of H2S and CO2.

Compositions of main reaction products under different conditions

During thermal cracking of crude oil at 360 °C, the relative content of alkane gases generally increased with longer heating time, while some decreased as the heating time was increased to 219 h (Table 1). The relative content of CH4 in the gaseous reaction products ranged from 39.84 to 48.02%, while that of CO2 exhibited a decreasing trend with increasing heating time. Only trace amounts of H2S were detected after a heating time of 72 h and 219 h (Table 1). Similarly, it was found that during thermal cracking of a mixture of crude oil, MgSO4, and buffer at 360 °C (M15–M20), the relative content of CH4 and C2H6 first increased and then decreased as the reaction continued. The relative content of CH4 varied between 6.44% and 13.54%. Concentration of heavy hydrocarbon gases, including C3H8 and C4H10, decreased with longer heating time. The relative content of H2S and CO2 remained high during the entire reaction and increased with increased heating time. In all cases, the relative content of H2S and CO2 produced from thermal cracking of crude oil increased because of TSR. Thermal cracking (M46–M55) of C9 and MN in the presence of sulfate also produced high amounts of CO2 and low amounts of alkane gases (Table 1). To facilitate a comparison with other research and gas geochemical characteristics under actual geological conditions, we modelled the Easy%Ro (Table 1).

Isotopic composition of main reaction products produced under different conditions

Carbon and hydrogen isotopic compositions gaseous reaction products resulting from thermal cracking of various hydrocarbons under a range of experimental conditions are summarised in Table 2 above.

Discussion

Yields and relative contents of TRS reaction products

Yield of CH4, CO2, and H2S from direct thermal cracking of crude oil are significantly lower than those resulting from cracking of a crude oil and MgSO4 solution under similar experimental conditions (Table 1). Increased yields of gaseous products might be attributed to the involvement of sulfur in the reaction, which may also trigger TSR37,38, converting heavier hydrocarbons into CH4, CO2, and H2S12. Since the activity energy for alteration of hydrocarbons TSR is lower than that for thermal cracking38, gas yields from the crude oil and MgSO4 solution is higher than those for thermal cracking of crude oil alone under the same pyrolysis conditions (temperature and time). Compared to the increase of CH4 yield during thermal cracking of crude oil, the CH4 yield during thermal cracking of crude oil and MgSO4 rapidly increased in the first 72 h at 360 °C, and then slightly decreased (Fig. 1a), which may be related to the oxidation of methane to H2S and CO2 during TSR12,27. The H2 yield from thermal cracking of crude oil gradually increased with time up to 219 h, but the H2 yield from crude oil and MgSO4 decreased from 0.23 ml/g to 0.15 ml/g until 72 h, and remained almost constant at less than 0.15 ml/g as the reaction continued (Fig. 1b). This observation suggests that TSR may have a very limited effect on H2 formation during pyrolysis. During the reaction of crude oil and MgSO4, the yield of H2S and CO2 increased rapidly before until 72 h (Fig. 1c, d), after which the CO2 yield remained almost constant as the H2S yield decreased slightly. Because the presence of a MgSO4 solution introduces sulfur and oxygen into the pyrolysis system, the yield of H2S and CO2 increased12,37. After 72 h, the yield of CH4 and H2S slightly decreased (Fig. 1a, d), indicating that the TSR process consumed these components to some extent and led to the production of CO2 and sulfur27,40,43. The dryness coefficient (C1/C1–5) of the gas formed by thermal cracking of crude oil is low, while the dryness coefficient of gas formed during pyrolysis with TSR is higher, accompanied by an increased CH4 content (Table 1), which further indicates that oxidation of heavy hydrocarbons by TSR will generate CH412,40,43.

Figure 1
figure 1

Variations of the yield of major products from thermal cracking of crude oil only (light blue line) and those in the presence of MgSO4 (red line) at the temperature of 360 °C: CH4 (a), H2 (b), CO2 (c), and H2S (d).

Full size image

Because the source gas of CO2 and H2S is produced during the TSR process, the ratios of CH4/CO2 and (CO2 + H2S)/(CO2 + H2S + ∑C1−5) were used to investigate variations of the chemical composition of gas altered by TSR. As shown in Fig. 2, ratios of (CO2 + H2S)/(CO2 + H2S + ∑C1−5) sharply decrease with increasing CH4/CO2 ratios. The ratio of CH4/CO2 for crude oil, C9, and MN with MgSO4 is less than 1.0, while the ratio of CH4/CO2 for thermal cracking of crude oil is above 1.0 (Fig. 2a). In contrast to the ratios of CH4/CO2 and (CO2 + H2S)/(CO2 + H2S + ∑C1−5) for crude oil, the same ratios for C9 and MN with MgSO4 are less than 1.0, and the ratio of (CO2 + H2S)/(CO2 + H2S + ∑C1−5) decreases as the CH4/CO2 ratio increases. Compared to the wide range of ratios for pyrolysis of crude oil and MgSO4, the ratio of CH4/CO2 and (CO2 + H2S)/(CO2 + H2S + ∑C1−5) for C9 or MN with MgSO4 shows a smaller range (Fig. 2b). These variations of CH4/CO2 and (CO2 + H2S)/(CO2 + H2S + ∑C1−5) caused by thermal cracking of crude oil and different degrees of TSR alteration during pyrolysis are similar to those caused by thermal cracking and TSR alteration of natural gas in gas reservoirs1, 20,21. Thermal cracking of crude oil produced more CH4, while TSR increased the yield of CO2 and H2S. The content of CH4 in alkane gas further increased with increased thermal cracking. In contrast, the CO2 and H2S contents varied at different stages of TSR. Based on the properties of H2S-bearing natural gas and previous simulation results1,6,12,23,40, variations in gas produced by TSR pyrolysis are similar to those of H2S-bearing natural gas altered by TSR.

Figure 2
figure 2

The plot of CH4/CO2 versus (CO2 + H2S)/(CO2 + H2S + ∑C1−5) during thermal cracking of crude oil and TSR.

Full size image

Fractionation characteristics of carbon and hydrogen isotopes during TSR

The carbon isotopic composition of produced CO2 from crude oil, C9, and MN with MgSO4 at different temperature is negatively correlated with heating time. The δ13CCO2 value gradually decreased from − 2.1‰ (C9, 360 °C, 4 h) to − 28.5‰ (crude oil, 360 °C, 219 h) (Table 2, Fig. 3). The carbon isotopic composition of CO2 from crude oil, C9, and MN with MgSO4 became significantly lighter over time until about 20 h heating time. After that, the δ13CCO2 value for crude oil and C9 with MgSO4 remained mostly constant, which is similar to what happens during thermal cracking of crude oil. The decrease of δ13CCO2 values during the first 20 h can be attributed to a greater fractionation of carbon isotopes at the onset of TSR12. After 20 h, the carbon isotope fractionation gradually reached a balance between CO2 and CH4. Although the δ13CCO2 value for CO2 produced from MN with MgSO4 decreased with heating time, CO2 was relatively more enriched in 13C when compared to crude oil, C9, and MgSO4. In addition to having a heavy carbon isotopic composition, MN was also the most easily oxidised component during TSR. Due to the addition of MgSO4 to MN, MN was quickly oxidised by TSR and converted to gas with CO2 as the main component. Concentrations of heavy hydrocarbon gases were below the detection limit of the instruments. In summary, the transformation of crude oil, C9, and MN by TSR converts 12C-rich hydrocarbons to 12C-enriched CO2, which might be converted to 12C-rich calcite and precipitated in gas reservoirs9,21.

Figure 3
figure 3

The plot of reaction time versus δ13CCO2 of gaseous products from simulation experiments of different mixtures at various temperatures.

Full size image

The δ13C value of gas produced from crude oil, C9, and MN with MgSO4 at different temperature and heating times is shown in Fig. 4. With increasing carbon numbers, the carbon isotopes of alkane gas from thermal cracking of crude oil at a pyrolysis temperature of 360 °C became heavier in the order of δ13C1 < δ13C2 < δ13C3.

Figure 4
figure 4

The δ13C values of products from simulation experiments with crude oil at 360 °C (a), a mixture of crude oil and MgSO4 at temperatures of 350 °C (b), 360 °C (c), 370 °C (d), and a mixture of C9 and MgSO4 at 360 °C (e).

Full size image

The δ13C1 values gradually increased over a relatively large range, while both δ13C2 and δ13C3 increased in a more narrow range as heating time increased (Fig. 4a). The carbon isotopic composition of alkane gas produced from crude oil and MN with MgSO4 also became heavier with increasing carbon number in the order of δ13C1 < δ13C2 < δ13C3. However, at the same temperature, with the addition of MgSO4 solution, δ13C1 gradually increased in a relatively narrow range, while both δ13C2 and δ13C3 showed larger increase as the reaction proceeded (Fig. 4b–d). It is obvious that the δ13C1 of alkane gas produced in the presence of a MgSO4 solution is higher than that produced during comparative experiments without MgSO4. In a single thermal system, the δ13C of alkane gas from thermal cracking of crude oil gradually becomes higher with increasing carbon number, and is linearly correlated with the reciprocal of the carbon number (1/n)44. The reduction in the variation of δ13C1 values produced by TSR alteration indicates that these hydrocarbons are rapidly oxidised to CH4 with a δ13C1 value similar to the source material. The TSR process leads to a 13C increase in CH4, producing a carbon isotopic composition more similar to that of crude oil (− 32.8‰). In contrast, CO2 becomes gradually enriched in 12C due to equilibrium fractionation of carbon isotopes between CO2 and CH4. Overall, the variation of δ13C1 is significantly smaller in the presence of MgSO4, during TSR of crude oil, similar to that of H2S-bearing alkane gas in the Sichuan Basin, where natural gas shows a heavy carbon isotopic composition of CH420. The carbon isotope composition of CO2formed in the presence of TSR is lighter than that formed in the absence of TSR, because TSR will oxidise a large portion of the hydrocarbons, leading to more intense δ13CCO2 fractionation. Therefore, δ13CCO2 values related to crude oil cracking are relatively heavier, while δ13CCO2 will be relatively enriched in 12C during pyrolysis (with TSR) in the presence of a MgSO4 solution (Table 2).

Our experiments show for the first time that the carbon isotopic composition of gas produced from C9 with MgSO4 became partially reversed to δ13C1 > δ13C2 < δ13C3 after a heating time of 24 h. The CH4 produced from C9 with MgSO4 shows an extremely small variation in δ13C1 values. The variation of δ13C2 became larger than that of δ13C3 with longer reaction time (Fig. 4e). This partial reversal of the carbon isotope series of alkane gas is similar to that of H2S-bearing alkane gas in the Sichuan Basin, Ordos Basin, and other locations1,16,23. Therefore, it is likely that the partial reversal of the carbon isotope series of alkane gas in H2S-bearing natural gas reservoirs happens when light hydrocarbons are altered by TSR. The isotopic composition of reaction products of pyrolysis with MN could not be detected due to the low content of alkane gases.

Figure 5 shows the variation of δ2H–Cn in alkane gases produced from crude oil at 360 °C, from a mix of crude oil and MgSO4 at 350, 360, and 370 °C, as well as from a mix of C9 and MgSO4 at 360 °C. Similar to the δ13C values, the hydrogen isotopic composition of alkane products from crude oil and a mixture of crude oil and MgSO4 became heavier with increasing carbon number under all conditions, i.e. δ2H–C1 < δ2H–C2 < δ2H–C3. The δ2H–C1 values gradually increased in a relatively large range, and both δ2H–C2 and δ2H–C3 increased in a narrower range as heating time increased (Fig. 5a). The δ2H–C1 value for a mix of crude oil and MgSO4 gradually increased over time, but in a relatively small range for each temperature. The δ2H–C2 and δ2H–C3 values also increased with heating time, but in a much larger range compared with that of δ2H–C1 (Fig. 5b–d). The variation of δ2H–C2 is slightly larger than that of δ2H–C3. The δ2H–C1 values remained almost constant during the TSR alteration of C9, while the δ2H–C2 values show much larger variations. The δ2H–C3 show the largest variations, but a reversed trend of the hydrogen isotopic composition of alkane gas was not observed during pyrolysis (Fig. 5e). These results indicate that the TSR can reduce the variation of δ2H–C1, possibly due to the similar hydrogen isotopic composition of the reaction product (C1 gas) and that of the precursor and/or the involvement of hydrogen derived from water21. The δ2H–Cn fractionation could not be accurately calculated because of a lack of the hydrogen isotopic composition of H2 and H2S. However, TSR of crude oil in the presence of MgSO4 greatly reduced the variation of δ2H–C1, which is consistent with the characteristics of H2S-bearing alkane gas in the reservoirs of the Sichuan Basin, China45. In general, hydrogen isotope fractionation during pyrolysis system may be more complicated than fractionation of carbon isotopes, because hydrogen contributes both to formation of alkane gas and H2, while also providing hydrogen for the formation of H2S. More importantly, the presence of water may provide an important hydrogen during pyrolysis21.

Figure 5
figure 5

The δ2H–Cn of gaseous products from simulation experiments with crude oil at 360 °C (a), a mixture of crude oil and MgSO4 at the temperature 350 °C (b), 360 °C (c), and 370 °C (d), as well as a mixture of C9 and MgSO4 at 360 °C (e).

Full size image

These results suggest that TSR can alter both the carbon and hydrogen isotopic composition of gaseous alkane products, and even lead to a reversed trend of the carbon isotope series of CH4 and C2H6 for a mix of C9 and MgSO4 as the source material. This can result in similar isotopic compositions to those of H2S-bearing natural gas45. Liu et al.20 investigated carbonate gas systems in the Sichuan Basin of China and found different amounts of acid gases (CO2 and H2S) in all marine strata. Both positive carbon isotope series and partially reversed sequence of alkanes were found. The positive carbon isotope series formed during production of sour gases was caused by TSR alteration that preferentially reacted with 12C-bearing heavy hydrocarbons. Therefore, the carbon isotope composition of residual heavy gases became heavier, and the partially-reversed carbon isotope sequence was again converted to be positive series.

In this study, we found that the TSR process played a significant role in the pyrolysis of C9 and MN with a MgSO4 solution, which lead to production of mainly H2S and CO2. TSR is more sensitive to MN, which makes it possible that MN can be completely oxidised into CO2 and H2S (Table 1). The smallest variation of both carbon and hydrogen isotopic compositions was observed for CH4, and was closest to that of the precursor. The heavier carbon isotopic composition of the precursor and C2H6 produced from thermal cracking initially led to a reversed carbon isotope series of the alkane gas (Table 2, M-48). However, the temperature of our experiments (360 °C) was much higher than that in natural geological environments. Therefore, heavy hydrocarbons enriched in 12C tended to become unstable, and were subject to thermal cracking, which led to isotope fractionation. 13C was enriched in the residual heavy hydrocarbons, and the effect of fractionation was larger than that for CH4 during TSR, changing the carbon isotope sequence back to positive.

Gas souring index (GSI) and carbon and hydrogen isotope fractionation of alkane gas

Because H2S can be produced during TSR, the gas souring index (GSI), i.e., H2S/(H2S + ∑C1−5), has been used as an indicator for the occurrence and degree of TSR46. The variation of the GSI at different stages of TSR can be established by statistical analysis of H2S-bearing natural gas samples45. To reproduce the variation of molecular and isotopic compositions of H2S-bearing natural gas during TSR, the GSI and carbon and hydrogen isotope fractionation mechanism of alkane gas were studied during pyrolysis at different temperatures with various heating times.

During thermal cracking of crude oil, a very small amount of H2S with minor CO2 was produced, indicating that no TSR occurred during direct thermal cracking of crude oil. In contrast, the CO2 yield gradually increased with a larger GSI (> 0.6) for TSR involving a mixture of crude oil and MgSO4 solution, and increased rapidly with further increasing GSI (Fig. 6), which is similar to the relationship between GSI and CO2 content of H2S-bearing natural gas20. The δ13CCO2 values remained nearly constant but gradually decreased with increasing GSI during thermal cracking due to the presence of TSR (Fig. 7). This phenomenon can be attributed to TSR, which preferentially incorporates 12C from hydrocarbons into CO2 leading to 12C-enrichment with increasing TSR intensity44.

Figure 6
figure 6

The plot of H2S/(H2S + ∑C1−5) versus CO2 of simulation products.

Full size image
Figure 7
figure 7

The plot of H2S/(H2S + ∑C1−5) versus δ13CCO2 of simulation products.

Full size image

Figure 8 shows the relationship between the gas souring index and δ13C1, δ13C2, and δ13C3. The direct thermal cracking of crude oil produced only a small amount of H2S due to low sulfur content of the crude oil, resulting in a GSI of less than 0.1. The GSI significantly increased with the addition of MgSO4 solution to the crude oil to above 0.6 under different conditions. Meanwhile, the carbon isotopic composition of alkane gas became larger with increasing TSR intensity, i.e., longer reaction time (Fig. 8). The carbon isotopic composition of the alkane gas revealed that the δ13C1 variation of CH4 produced in the presence of MgSO4 was much lower than that of C2H6. The carbon isotope composition changed more significantly with the increasing GSI and longer heating time. Therefore, during TSR, the variation of δ13C1 became smaller, compared with that of δ13C2 and δ13C3. The δ13C1–δ13C2 difference became higher with increasing GSI, which is completely different from that of natural gas formed by direct thermal cracking of crude oil. During thermal cracking of crude oil, the carbon isotope composition of gaseous products gradually becomes more similar with increasing carbon number45.

Figure 8
figure 8

The plot of H2S/(H2S + ∑C1−5) versus δ13C1 (a), δ13C2 (b), δ13C3 (c), δ13C1–δ13C2 (d) of alkane gas products from simulation experiments under different conditions.

Full size image

The δ2H–Cn values of alkane gas show a variation similar to the δ13Cn values, with increasing GSI during the TSR process. The range of δ2H–C1 values is fairly low, while that of δ2H–C2 is higher, especially for the TSR of C9 whose δ2H–C1 remained almost constant (Fig. 9). Similar to the carbon isotopic composition, the δ2H–C1–δ2H–C2 difference increased as the gas souring index became larger, indicating that TSR can lower δ2H–C1 values and increase those of δ2H–C2, producing an abnormal hydrogen isotope composition of alkane gas, which is only characteristic of thermal cracking of crude oil. Both δ13C1–δ13C2 and δ2H–C1–δ2H–C2 became lower with increasing TSR intensity. However, in the case of oil cracking, δ13C1–δ13C2 and δ2H–C1–δ2H–C2 show the opposite trend with increasing TSR (Fig. 10). This observation may be caused by the small variation of the δ13C1 and δ2H–C1 produced by TSR, which is similar to the isotopic composition of the precursor. The carbon and hydrogen isotopic composition of C2H6 was more heavily affected by the thermal cracking than oxidisation by TSR. Therefore, the variation of its carbon and hydrogen isotope compositions is higher. In addition, the difference in the carbon and hydrogen isotopic composition between CH4 and C2H6 gradually increased with a larger GSI.

Figure 9
figure 9

The plot of H2S/(H2S + ∑C1−5) versus δ2H–C1 (a), δ2H–C2 (b), δ2H–C3 (c) and δ2H–C1 – δ2H–C2 (d) of alkane gas products from simulation experiments under different conditions.

Full size image
Figure 10
figure 10

The plot of δ13C1– δ13C2 versus δ2H–C1– δ2H–C2 of alkane gas products from simulation experiments under different conditions.

Full size image

Although TSR alteration decreased the variation of both carbon and hydrogen isotope compositions of CH4, as compared with products from direct thermal cracking of crude oil, δ13C2 became larger with increasing δ13C1. The δ13C2 versus δ13C1 plot (Fig. 11a), and δ13C2 versus δ13C3 plot (Fig. 11b), showed positive correlations, with the latter having a higher correlation. The δ13C1 and δ13CCO2 values are negatively correlated (Fig. 11c), possibly because the carbon isotopic composition of CO2 produced by TSR is enriched in 12C34. For CH4, C2H6, and C3H8, δ2H–C1 and δ2H–C2 are positively correlated with temperature and heating time. δ2H–C2 values increase with increasing δ2H–C1 values. The positive correlation between δ2H–C2 and δ2H–C3 is even more significant (Fig. 12). Based on these results, it can be concluded that the oxidation of hydrocarbons by TSR is accompanied by thermal cracking of crude oil, resulting in distinct patterns among the δ2H values of alkane gases. Although TSR effect can lead to smaller variations of carbon and hydrogen isotope compositions of CH4 compared with C2H6 and C3H8, δ2H–C1 values increased with increasing δ13C1 values (Fig. 13), suggesting that direct thermal cracking of crude oil also produced CH4. However, the CH4 produced during pyrolysis of C9 should be primarily produced by TSR.

Figure 11
figure 11

The plot of δ13C1 versus δ13C2 (a), δ13C2 and δ13C3 (b) and δ13C1 and δ13CCO2 (c) of alkane gas products from simulation experiments under different conditions.

Full size image
Figure 12
figure 12

The plot of δ2H–C1 versus δ2H–C2 (a) and δ2H–C2 versus δ2H–C3 (b) of alkane gas products from simulation experiments under different conditions.

Full size image
Figure 13
figure 13

The plot of δ13C1 versus δ2H–C1 of alkane gas products from simulation experiments under different conditions.

Full size image

All TSR simulation experiments of crude oil, C9 and MN under different conditions suggest that the TSR alteration produces CH4 with similar carbon and hydrogen isotopic compositions to those of its precursor and reduced variations of isotopic compositions. The difference between CH4 and C2H6 in carbon isotope and hydrogen isotopic composition increases with increasing TSR intensity, while the carbon and hydrogen isotopic composition of C2H6 and C3H8 became heavier and shows smaller differences with increasing temperature, similar to the results from thermal cracking of crude oil. The production of alkane gas with similar chemical and isotope compositions to H2S-bearing natural gas during these experiments, suggest that TSR can alter the carbon and hydrogen isotopic composition of alkane gas20,21. In addition, a partially reversed carbon isotope series of alkane gas (δ13C1 > δ13C2 < δ13C3) was observed, which further confirms the above conclusion. 12C-enriched CO2 was mainly produced from the oxidation of hydrocarbons by TSR. However, dissolution of CO2 and precipitation of carbonate minerals in aqueous fluids can complicate the carbon isotopic composition in the marine carbonate reservoir45.

Conclusions

Pyrolysis of TSR was carried out using different organic matter (crude oil, nonane and methylnaphthalene), and the characteristics of carbon and hydrogen isotopes, as well as the composition and yields of the reaction products were analysed. The following main conclusions can be drawn:

  • (1) The carbon and hydrogen isotopic composition of alkane gas generally becomes heavier with increasing carbon number, i.e., δ13C1 < δ13C2 < δ13C3 and δ2H–C1 < δ2H–C2 < δ2H–C3. At the same temperature, the carbon and hydrogen isotopic composition of CH4 gradually became larger with longer reaction time. The carbon and hydrogen isotopic composition of C2H6 and C3H8 also became heavier as the reaction continued, but the variation was significantly larger than that for CH4. The variation of the carbon and hydrogen isotopic composition of C2H6 is higher than that of C3H8. Rapid oxidation of source material by TSR produced CH4 with a small variation in carbon and hydrogen isotopic values, revealing that this process can alter the carbon and hydrogen isotopic composition of CH4, making it similar to those of the original precursor material.

  • (2) The partially reversed carbon isotope series observed in alkane gas produced from a mixture of C9 and MgSO4 indicates that TSR can cause abnormal isotope series of alkane gas in natural gas. As the reaction continued, 13C became enriched in residual heavy hydrocarbon gas, which altered the commonly observed order of δ13C1 > δ13C2 to a positive carbon isotope series (δ13C1 < δ13C2). For the first time, we confirmed the ability of TSR to alter the isotopic composition of alkanes, causing isotope reversal during TSR pyrolysis.

  • (3) Under conditions of TSR, hydrogen isotopes of alkane gases form a positive isotope series (δ2H–C1 < δ2H–C2 < δ2H–C3). Isotope fractionation is large, especially for ethane and propane, presumably due to the low molecular weight of hydrogen.

  • (4) During pyrolysis, oil cracking and TSR result in different δ13C1–δ13C2 and δ2H–C1–δ2H–C2 evolutionary trends. While δ13C1–δ13C2 and δ2H–C1–δ2H–C2 became smaller with increasing intensity of TSR, δ13C1–δ13C2 and δ2H–C1–δ2H–C2 became larger with increasing intensity of oil cracking.