5.2. The changing of carbon isotopes in kerogen, bitumen, expelled oil and gaseous hydrocarbons during thermal evolution stages
During thermal evolution, kerogen, expelled oil, bitumen, and gaseous products undergo a series of thermal cracking processes. Among them, the original kerogen (kerogen0) and original bitumen (bitumen0) exhibited a reactive nature (Tissot et al., 1978; Jarvie et al., 2007). When Ts = 335–575 °C, expelled oil, residual bitumen, and gaseous hydrocarbons of C2+ could not only be seen as products but also as reactants participating in the pyrolysis process. From the perspective of pyrolysis, only methane constantly remains a product during the entire thermal evolution process. Therefore, ignoring the intermediate reaction during the entire thermal evolution process, the corresponding reaction formula is as follows: kerogen0 (original) + bitumen0 (original) → kerogenr (residual kerogen) + expelled oil (generated) + bitumenn+r(generated + residual) + C2+ (generated + residual) + CH4 (generated) . In this reaction formula, the thermal evolution pathways were analyzed based on the corresponding change characteristics of carbon isotopes. Based on the principle of chemical kinetics, with changes in the coal-forming environment, organic matter type, maturity, etc., the carbon isotope composition of gases could also be changed. Simultaneously, a series of processes could produce intermediates in the cracking of kerogen, residual bitumen, or liquid hydrocarbons, and thus would result in the lighting of δ13C1~5, i.e., during thermal evolution of organic matter into hydrocarbons, the12C and 13C were primarily enriched in the former and latter generated products, respectively (Kinnon et al., 2010; Papendick et al., 2011; Golding et al., 2013).
In addition, the experimental error range of ±0.5‰ for carbon isotope analysis was also considered in this analysis (Wu et al., 2018). Overall, the δ13Ckerogen0 < δ13C kerogenr and δ13Cbitumen0 < δ13C bitumen0+r, and the former degree of fractionation (δ13C kerogenr13Ckerogen0 = 0.1–0.8, the mean value = 0.5) was less than the latter degree of fractionation (δ13C bitumen0+r13Cbitumen0 = 0.1–1.3, the mean value = 1.11), implying that kerogen and bitumen as reactants in the thermal evolution process would get heavy at their corresponding carbon isotope. In contrast, the degree of heaviness determined the strength of the reaction (Mahlstedt and Horsfield, 2012). Influenced by thermal action, the carbon isotope of kerogen tended to become heavier in general. There was no significant difference in the weight gained between the lower maturity-maturity stage and the higher-maturity-post-maturity stage, indicating that the reaction strength of kerogen did not change much during the entire thermal evolution process. The main sources of bitumen were the residual bitumen and bitumen generated from kerogen by pyrolysis (Jarvie et al., 2007). At the lower to higher maturity stage, the carbon isotope of bitumen became heavier, suggesting that thermal evolution had a greater impact. Similarly, the original and generated bitumen were constantly parted during thermal evolution. At the post-maturity stage, the carbon isotope of the residual bitumen was fundamentally the same as that in the original bitumen, indicating that the amount of bitumen generated from kerogen was almost exhausted and reached a certain equilibrium. The remaining bitumen was that which was generated from kerogen cracking, which could also be proved by their contents, as shown in Fig. 3.
The main formation path of expelled oil was the cracking of bitumen, and the carbon isotopic composition of the expelled oil was essentially unchanged and was lighter than that in kerogen and bitumen. The yield of expelled oil was lower at the lower maturity stage and was almost in equilibrium at the higher maturity-post-maturity stage. Therefore, it can be seen that the expelled oil, as both product and reactant, was in a state of “supply and demand balance” after the maturity stage; indirectly, it showed that at the lower maturity stage, the reaction equation was kerogen → expelled oil + gaseous hydrocarbons, and dominated by the production of gaseous hydrocarbons (Tissot et al., 1978; Jarvie et al., 2007). After the maturity stage, it was dominated by the production of expelled oil. At this stage, the residual bitumen decreased continuously, indicating that the generated and increased expelled oil or gaseous hydrocarbons were further cracked into gaseous hydrocarbons to ensure the equilibrium state of the expelled oil (Castelli et al., 1990; Pepper and Corvi, 1995; Jarvie et al., 2007; Zheng et al., 2011; Qin et al., 2014; Jiang et al., 2016).
Gaseous hydrocarbons can be generated either by kerogen or by further pyrolysis of intermediate bitumen or expelled oil. Therefore, the difference in the carbon isotopes of gaseous hydrocarbons was restricted by the dynamic fractionation effect of the above two reactions. In other words, their respective isotope dynamic fractionation was closely related to the hydrocarbon formation mechanism of organic matter at different thermal evolution stages. For example, methane was the lightest carbon isotope in this study. Its generation pathway can be divided into 1) direct generation from kerogen, 2) generation from the cracking of bitumen, 3) generation from the cracking of expelled oil, and 4) generation from the secondary cracking of C2+gases (Pepper and Corvi, 1995; Hill et al., 2003; Jarvie et al., 2007; Behar et al., 2008).
In general, the carbon isotope of methane showed a particularly good linear correlation with Ro , and the correlation equation was δ13C1 = 4.632 Ro - 43.493 with a correlation coefficient (R 2) of 0.9142, indicating that the carbon isotope of 13C was enriched continuously with thermal evolution. The corresponding enrichment degree of δ13C1 reached 9.7‰. The considerable carbon isotope dynamic fractionation also indicated that it was the most sensitive to maturity; therefore, it could be used for gas source correlation due to its reliability. We observed that atTS ≤400℃, i.e., the low maturity-maturity stage, the enrichment degree of δ13C1 was lower at 2.6‰. However, at TS ≥400 ℃, namely the high-post-maturity stage, the enrichment degree of δ13C1 increased rapidly to 9.7‰. Altogether, the results indicate the lower maturity stage was primarily controlled by kerogen and bitumen, and the main controlling factor was expelled oil and C2+ gases after the maturity stage. It was further proved that the formation mechanism of methane significantly differed before and after 400 °C. Moreover, according to previous studies, the secondary cracking of gaseous hydrocarbons was generally in the order of higher carbon number to lower carbon number (Behar et al., 1992; Jarvie et al., 2007; Sun et al., 2019b). When gaseous hydrocarbons with higher carbon numbers are cracked to gaseous hydrocarbons with lower carbon numbers, the carbon isotope of gaseous hydrocarbons with this carbon number deviates from the normal evolutionary track. According to the variation in the values of δ13Cn13Cn-1(n ≥ 2) with increasing Ro (Table 3), it can also be seen that the different corresponding difference values first decreased and then increased. The Ro corresponding to the lowest point was successively advanced; the larger the value of n , the smaller was the Ro at the low point of the difference value. For example, the values of δ13C4–δ13C3demonstrated little change at Ro = 1.09% and 1.65%, those of δ13C3–δ13C2showed a significant change between Ro = 1.65% and 1.93%, while those of δ13C2–δ13C1showed a significant change after Ro = 2.3%. In the lower evolution stage, Ro < 1.65%, the cracking of gaseous hydrocarbons with higher carbon numbers was not significant. In the higher evolution stage, after Ro > 1.65%, the gaseous hydrocarbons with a higher carbon number would crack into gaseous hydrocarbons with a lower carbon number successively, resulting in an increase in the difference value in the carbon isotope. Therefore, it was further explained that the formation mechanism of methane was different at different thermal evolution stages. Among them, the reactions corresponded to reaction processes 1 and 2 at the lower maturity stage, and reaction 2 was dominant. Reaction processes 1, 2, and 3 occurred at the maturity-higher maturity stage, and reaction 3 was dominant. At the post-maturity stage, reaction processes 1, 2, 3, and 4 all occurred, with reaction 4 being dominant.
Table 3 shows the variation in the values of δ13Cn–δ13Cn-1(n ≥ 2) with increasing Ro