The Dynamic Evolution Model of the Chemical and Carbon Isotopic Composition of C1–3 during the Hydrocarbon Generation Process

A new approach is presented in this paper for the dynamic modeling of the chemical and isotopic evolution of C1–3 during the hydrocarbon generation process. Based on systematic data obtained from published papers for the pyrolysis of various hydrocarbon sources (type I kerogen/source rock, type II kerogen/source rock, type III kerogen/source rock, crude oil, and asphalt, etc.), the empirical evolution framework of the chemical and isotopic composition of C1–3 during the hydrocarbon generation process was built. Although the empirical framework was built only by fitting a large amount of pyrolysis data, the chemical and isotopic composition of C1–3 derived from the pyrolysis experiments all follow evolution laws, convincing us that it is applicable to the thermal evolution process of various hydrocarbon sources. Based on the simplified formula of the isotopic composition of mixed natural gas at different maturities (δ13Cmixed), δ13Cmixed = X×niA×δ13CiA+Y×niB×δ13CiBX×niA+Y×niB, it can be derived that the cumulative isotopic composition of alkane generated in a certain maturity interval can be expressed by the integral of the product of the instantaneous isotopic composition and instantaneous yield at a certain maturity point, and then divided by the cumulative yield of alkane generated in the corresponding maturity interval. Thus, the cumulative isotopic composition (A(X)), cumulative yield (B(X)), instantaneous isotope (C(X)), and instantaneous yield (D(x)) in the dynamic model, comply with the following formula during the maturity interval of (X0~X). A(X) = ∫X0XCX×DXdxB(X), where A(X) and B(X) can be obtained by the fitting of pyrolysis data, and D(x) can also be obtained from the derivation of B(X). The dynamic model was applied on the pyrolysis data of Pingliang Shale to illustrate the quantitative evolution of the cumulative yield, instantaneous yield, cumulative isotope, and instantaneous isotope of C1–3 with increasing maturity. The dynamic model can quantify the yield of methane, ethane, and propane, as well as δ13C1, δ13C2, and δ13C3, respectively, during the hydrocarbon generation process. This model is of great significance for evaluating the natural gas resources of hydrocarbon source rock of different maturities and for identifying the origin and evolutionary process of hydrocarbons by chemical and isotopic data. Moreover, this model provides an approach to study the dynamic evolution of the isotope series of C1–3 (including reversed isotopic series), which is promising for revealing the mechanism responsible for isotopic reversal when combined with post-generation studies.


Introduction
The chemical and isotopic composition of light hydrocarbons have long been used to study the origins, generation, accumulation, and degradation process of hydrocarbons in the past decades [1][2][3][4][5][6][7][8][9]; however, with the large-scale exploration, development, and study of the intensive isotopic geochemistry of natural gas resources in the high evolution strata of superimposed basins, it is increasingly difficult to clearly identify the origin and evolutionary process of hydrocarbons by the chemical and isotopic characteristics of alkanes.This is because there are many factors and geochemical processes responsible for the chemical and isotopic compositions of hydrocarbons in geological conditions, such as the inheritance of isotopic signatures from precursor organics, kinetic isotope fractionations, and equilibrium isotope fractionations, and the mixing of hydrocarbons from different origins and of different maturities.Combined, these factors can readily lead to the overlapping of the bulk isotopic signatures of hydrocarbons of different origins and history, obscuring their origins and evolutionary processes.The mechanism responsible for the evolution of the isotopic composition of alkanes, e.g., the cause of the reversed alkane δ 13 C values in many natural gas plays, is still controversial and remains an open question [10][11][12][13].
The problems in the application of traditional natural gas isotope geochemistry can be attributed to the following two aspects: (a) for the current isotope analysis technology of monomer hydrocarbon, the isotopic information of the internal functional groups of the monomer hydrocarbon disappears in the process of oxidizing the target component to CO 2 or reducing it to H 2 and, accordingly, the intra-molecular isotope distribution information which can disclose the formation and evolution process of natural gas is not fully obtained [7,14,15]; (b) the dynamic evolution laws of the isotope composition of monomer hydrocarbon and corresponding mechanism during the formation and evolution process of natural gas are still in dispute [6,12,16,17].Efforts to explore the intra-molecular isotope distribution of hydrocarbons, including position-specific isotope and clumped isotope analyses, have been made in the past decades [7,14,[18][19][20].This isotopic information with higher dimensions is expected to provide unique constraints on their formation and migration-degradation processes; however, the theoretical studies and applications of intra-molecular isotope distribution are still in their infancy and cannot be widely used as yet.The traditional monomer hydrocarbon isotope is still the main research method of natural gas isotope geochemistry; accordingly, research on the dynamic evolution of the chemical and isotopic composition of alkanes still needs to be strengthened, especially regarding dynamic evolution during the hydrocarbon generation process.
Efforts to study the dynamic evolution of the chemical and isotopic composition of alkanes (especially C 1-3 ) during the hydrocarbon generation process have been made in the past decades.Many empirical models have been put forward to study the compositional and isotopic variation of natural gas [1,[21][22][23]; however, these empirical models were developed based on the statistical analysis of massive data, and the mechanisms responsible for the compositional and isotopic variation of alkanes were not fully understood.The Rayleigh model is the first semi-quantitative model to study the compositional and isotopic variation of alkanes [24][25][26], but the Rayleigh model cannot exactly study the compositional and isotopic variation of alkanes.The obvious defects of the Rayleigh model are as follows [3]: one is the assumption that methane and the higher hydrocarbons can be modeled using a single extent of reaction parameter; the other is the assumption that the fractionation factor of each first-order reaction is constant.The hydrocarbon generation kinetic model based on various pyrolysis experiments is widely applied to the quantitative study of the yield evolution of methane and total gaseous hydrocarbon (C 1 -C 5 ),while the carbon isotope kinetic model based on various pyrolysis experiments is widely applied to the quantitative study of the isotopic evolution of methane during the hydrocarbon generation process [3,[27][28][29][30][31][32].The hydrocarbon generation kinetic model and carbon isotope kinetic model are only applicable to the first-order reaction in the hydrocarbon generation process.The yield of total gaseous hydrocarbon (C 1-5 ) can also reach a plateau in hydrocarbon generation simulation experiments, so C 1-5 can be considered as a whole to calculate its yield evolution using hydrocarbon generation simulation data.However, the yield of individual heavy hydrocarbon components initially increases and then decreases after reaching maximum yield in hydrocarbon generation simulation experiments, indicating that the heavy hydrocarbon components have undergone pyrolysis.In addition, the heavy hydrocarbon components may involve a Fischer-Tropsch reaction in the high evolution stage, so individual heavy hydrocarbon components cannot be regarded as a first-order reaction and are not suitable for the hydrocarbon generation kinetic model and carbon isotope kinetic model.Tang et al. (2000) predicted the evolution of the cumulative yield of methane, instantaneous yield of methane, instantaneous δ 13 C 1 , and cumulative δ 13 C 1 by the hydrocarbon generation kinetic model and carbon isotope kinetic model, based on the pyrolysis of n-octadecane [3].Shuai et al. (2005Shuai et al. ( , 2006) ) carried out the hydrocarbon generation kinetic and carbon isotope kinetic studies of ethane by dividing the evolution process of ethane into two stages [33,34]: the generation-dominated stage and the cracking-dominated stage.However, each stage simultaneously involves the generation and cracking of ethane, and there may be ethane generated by the Fischer-Tropsch reaction in the cracking-dominated stage.To sum up, there is no valid model to quantitatively study the chemical and isotopic variation of C 1-3 during the hydrocarbon generation process to date, making it very difficult to identify the origin, source, and resources of natural gas by chemical and isotopic data.
To address this limitation, a quantitative model was built to illustrate the chemical and isotopic variation of C 1-3 during the process of hydrocarbon generation.
The objective of this work is multifold: (1) Summarize the universal evolution laws of the chemical and isotopic composition of C 1-3 in the hydrocarbon generation process on the basis of fitting systematic data obtained from published papers on the pyrolysis of various hydrocarbon sources; (2) Introduce the theoretical approach and detailed steps to build the dynamic model of the chemical and isotopic evolution of C 1-3 during the hydrocarbon generation process; (3) Illustrate the application of the dynamic model to the pyrolysis experiment of the Ordovician Pingliang Shale from the Ordos Basin, China, to study the chemical and isotopic evolution of C 1-3 during the hydrocarbon generation process.

Results and Discussion
The quantitative model was applied on the anhydrous-confined pyrolysis data of the kerogen extracted from the Ordovician Pingliang Shale in the Ordos Basin, China [31], to study the chemical and isotopic evolution of C 1 -C 3 during the hydrocarbon generation process.Detailed information regarding the Ordovician Pingliang Shale is as follows: marine shale, type II kerogen, TOC 18.1%, vitrinite reflectance (R O ) 0.7%.The δ 13 C of the isolated kerogen is 30.1‰.The detailed procedures of the anhydrous-confined pyrolysis experiment are as follows: the kerogen samples were introduced into gold tubes and were sealed under argon; the tubes containing fresh kerogen were loaded into a stainless vessel; pyrolysis was performed under a constant pressure of 50 MPa, with pyrolysis temperatures ranging from 250 to 600 • C at heating rates of 2 • C/h [35].The simulation temperature was converted to "Easy Ro" to represent the evolutionary maturity of the Pingliang Shale [36].
The cumulative isotope function (A(X)) and cumulative yield function (B(X)) of kerogen from the Pingliang Shale are obtained by the data fitting of confined pyrolysis experiments (Figures 1 and 2).The instantaneous yield function (D(x)) can be obtained by taking the derivative of the cumulative yield function (B(X)); the instantaneous isotope function (C(x)) can be calculated by Formula (6) (Figures 1 and 2).The unit of the instantaneous yield function [mg/(g•TOC•Ro)] is different from the unit of the cumulative yield function [mg/(g•TOC)], so the charts for the instantaneous and cumulative yields were plotted using the dual-axis method, and there is no direct comparability between the value of the instantaneous and cumulative yields.

Dynamic Evolution of C1 Yield during Hydrocarbon Generation Process
The cumulative yield of C1 originating from the pyrolysis of the Pingliang Shale increases continuously with rising maturity and reaches a plateau at the end of the methane generation process, which is consistent with the result that the instantaneous yield of C1 always stays positive during the methane generation process.The cumulative yield of C1 increases quite mildly at two ends of the curve and increases quite sharply in the middle of the curve, which is consistent with the result that the instantaneous yield of C1 initially increases and then decreases with raising maturity, reaching maximum instantaneous yield at a certain intermediate maturity (Easy Ro approx.2.4%).
Methane is widespread in various stages of the hydrocarbon generation evolution of the Pingliang Shale: in the immature stage, low-mature stage, mature stage, high-mature stage, and over-mature stage.At the immature stage, low mature stage and mature stage, liquid hydrocarbons dominate the hydrocarbon generation product of the sapropelic

Dynamic Evolution of C1 Yield during Hydrocarbon Generation Process
The cumulative yield of C1 originating from the pyrolysis of the Pingliang Shale increases continuously with rising maturity and reaches a plateau at the end of the methane generation process, which is consistent with the result that the instantaneous yield of C1 always stays positive during the methane generation process.The cumulative yield of C1 increases quite mildly at two ends of the curve and increases quite sharply in the middle of the curve, which is consistent with the result that the instantaneous yield of C1 initially increases and then decreases with raising maturity, reaching maximum instantaneous yield at a certain intermediate maturity (Easy Ro approx.2.4%).
Methane is widespread in various stages of the hydrocarbon generation evolution of the Pingliang Shale: in the immature stage, low-mature stage, mature stage, high-mature stage, and over-mature stage.At the immature stage, low mature stage and mature stage, liquid hydrocarbons dominate the hydrocarbon generation product of the sapropelic

Dynamic Chemical Evolution of C 1-3 during Hydrocarbon Generation Process 2.1.1. Dynamic Evolution of C 1 Yield during Hydrocarbon Generation Process
The cumulative yield of C 1 originating from the pyrolysis of the Pingliang Shale increases continuously with rising maturity and reaches a plateau at the end of the methane generation process, which is consistent with the result that the instantaneous yield of C 1 always stays positive during the methane generation process.The cumulative yield of C 1 increases quite mildly at two ends of the curve and increases quite sharply in the middle of the curve, which is consistent with the result that the instantaneous yield of C 1 initially increases and then decreases with raising maturity, reaching maximum instantaneous yield at a certain intermediate maturity (Easy Ro approx.2.4%).
Methane is widespread in various stages of the hydrocarbon generation evolution of the Pingliang Shale: in the immature stage, low-mature stage, mature stage, highmature stage, and over-mature stage.At the immature stage, low mature stage and mature stage, liquid hydrocarbons dominate the hydrocarbon generation product of the sapropelic Pingliang Shale.The gaseous hydrocarbons originating mainly from the degradation of the aliphatic side chain only account for a very low proportion of the total hydrocarbons, so the instantaneous and cumulative yield of C 1 is quite low at the beginning of the hydrocarbon generation process.At the high-mature stage, liquid hydrocarbons start to be extensively cracked, and crude oil cracking gas becomes the dominant origin of gaseous hydrocarbon, while the proportion of kerogen cracking gas decreases quickly.Since the hydrocarbongenerating intensity of crude oil cracking is 3-4 times that of kerogen cracking [37,38], the instantaneous yield of C 1 increases sharply and reaches maximum at Easy Ro approx.2.4% in this period.At the over-mature stage, liquid hydrocarbons are mostly cracked and the gas generation potential of kerogen is also nearly exhausted, so the instantaneous yield of C 1 decreases sharply at the end of the over-mature stage [38].At Easy Ro approx.4.5%, the cumulative yield of C 1 reaches a plateau as the instantaneous yield of C 1 approaches zero, implying that the hydrocarbon generation potential of the Pingliang Shale is already exhausted at Easy Ro approx.4.5%.However, the instantaneous yield of C 1 has never become negative during the whole hydrocarbon generation evolution stage, indicating that there is no cracking of methane.Previous research on the calculation of the reaction activation energy and pyrolysis experiments of methane have also proven that methane is so thermodynamically stable that it can hardly be pyrolyzed under geological conditions and hydrocarbon generation simulation experiments [38][39][40].Previous calculations have shown that the activation energy required for methane cracking is 441 kJ/mol, higher than that of ethane and propane [39,40].Moreover, it has been verified by extensive pyrolysis experiments on methane that the threshold temperature of methane pyrolysis is 1100 • C at atmospheric pressure (equivalent Easy Ro approx.5.0%) [38].The maximum temperature of the pyrolysis experiment on the Pingliang Shale is 600 • C, far lower than the threshold temperature of methane pyrolysis.It follows that the instantaneous yield of C 1 is promising for predicting the maturity deadline of hydrocarbon generation, which has practical significance for defining the upper limit of natural gas exploration.

Dynamic Evolution of C 2-3 Yield during Hydrocarbon Generation Process
The variation trends of the cumulative yields C 2 and C 3 are identical: the cumulative yields initially increase and then decrease with increasing thermal maturity, reaching maximum yields at a certain intermediate maturity.C 2 and C 3 reach maximum cumulative yield at Easy Ro approx.2.0% and 1.7%, respectively (Figure 1).At the same time, the instantaneous yield of C 2 and C 3 shifts from positive to negative values at Easy Ro 2.0% and 1.7%, respectively, which indicates the cracking threshold maturity of C 2 and C 3 .Actually, both the generation and cracking of C 2 and C 3 are present before and after the threshold maturity; C 2 and C 3 shift from the generation-dominated stage to the cracking-dominated stage at Easy Ro 2.0% and 1.7%, respectively.For simplicity, it is assumed in this study that there is only generation of C 2 and C 3 before the threshold maturity (Easy Ro 2.0% and 1.7%, respectively), and there is only cracking of C 2 and C 3 after the threshold maturity.
Propane starts cracking at Easy Ro 1.7% and ends cracking at Easy Ro 3.5%, while ethane starts cracking at Easy Ro 2.0% and ends cracking at Easy Ro 3.7%.Previous studies have shown that the activation energy required for ethane cracking is higher than that of propane [38]; therefore, the later cracking of ethane can probably be attributed to its higher molecular stability.The variation trends of the instantaneous yields of C 2 and C 3 are also identical.During the generation stage (instantaneous yield > 0), the instantaneous yield initially increases and then decreases with raising maturity, reaching maximum instantaneous yield at a certain intermediate maturity.The increase of the instantaneous yield of C 2 and C 3 at the early-generation stage can mainly be attributed to the result that C 2 and C 3 originating from kerogen cracking and crude oil cracking begins to be extensively generated at the mature and high-mature stages.The decrease of the instantaneous yield of C 2 and C 3 at the late-generation stage can mainly be attributed to the result that kerogen cracking and crude oil cracking are more inclined to generate C 1 rather than C 2 and C 3 at the high-mature stage.During the consumption stage (instantaneous yield < 0), the instantaneous consumption rate initially increases and then decreases with raising maturity, reaching maximum instantaneous consumption rate at a certain intermediate maturity.The increase of the instantaneous consumption rate at the early-consumption stage is owing to the initiation of the extensive cracking of C 2 and C 3 at the high-mature stage and overmature stage.The decrease of the instantaneous consumption rate at the late-consumption stage is owing to the result that the cumulative yields of C 2 and C 3 are mostly cracked and cracking potential of C 2 and C 3 is nearly exhausted.The cumulative and instantaneous δ 13 C 1 from the pyrolysis of the Pingliang Shale all present identical variation patterns: they shift negatively at the early hydrocarbon generation stage, and then shift positively until the end of methane generation.The variation trend of cumulative δ 13 C 1 depends on the relative isotopic composition values between cumulative δ 13 C 1 and instantaneous δ 13 C 1 : when instantaneous δ 13 C 1 is lighter than cumulative δ 13 C 1 (Easy Ro < 1.1%), the cumulative δ 13 C 1 shifts negatively; when instantaneous δ 13 C 1 is heavier than cumulative δ 13 C 1 (Easy Ro > 1.1%), the cumulative δ 13 C 1 shifts positively.The rollover of cumulative and instantaneous δ 13 C 1 have already be verified by the study of Tang et al. (2000), which used the carbon isotope kinetic model (Figure 3a) [3].Moreover, the rollover of cumulative δ 13 C 1 from the pyrolysis of the Pingliang Shale was also observed in many pyrolysis experiments [26,31,[41][42][43][44][45] (Figure 4).and C3 at the late-generation stage can mainly be attributed to the result that kerogen cracking and crude oil cracking are more inclined to generate C1 rather than C2 and C3 at the highmature stage.During the consumption stage (instantaneous yield < 0), the instantaneous consumption rate initially increases and then decreases with raising maturity, reaching maximum instantaneous consumption rate at a certain intermediate maturity.The increase of the instantaneous consumption rate at the early-consumption stage is owing to the initiation of the extensive cracking of C2 and C3 at the high-mature stage and over-mature stage.The decrease of the instantaneous consumption rate at the late-consumption stage is owing to the result that the cumulative yields of C2 and C3 are mostly cracked and cracking potential of C2 and C3 is nearly exhausted.

Dynamic Isotopic Evolution of δ 13 C1 during Hydrocarbon Generation Process
The cumulative and instantaneous δ 13 C1 from the pyrolysis of the Pingliang Shale all present identical variation patterns: they shift negatively at the early hydrocarbon generation stage, and then shift positively until the end of methane generation.The variation trend of cumulative δ 13 C1 depends on the relative isotopic composition values between cumulative δ 13 C1 and instantaneous δ 13 C1: when instantaneous δ 13 C1 is lighter than cumulative δ 13 C1 (Easy Ro < 1.1%), the cumulative δ 13 C1 shifts negatively; when instantaneous δ 13 C1 is heavier than cumulative δ 13 C1 (Easy Ro > 1.1%), the cumulative δ 13 C1 shifts positively.The rollover of cumulative and instantaneous δ 13 C1 have already be verified by the study of Tang et al. (2000), which used the carbon isotope kinetic model (Figure 3a) [3].Moreover, the rollover of cumulative δ 13 C1 from the pyrolysis of the Pingliang Shale was also observed in many pyrolysis experiments [26,31,[41][42][43][44][45] (Figure 4).The mechanism responsible for the isotope rollover of δ 13 C1 involves the presence of two or more precursors for methane generation with different isotopic compositions, and the mixing of methane originating from different precursors, which leads to the isotope rollover of C1 [3].Heteroatoms such as O and S are widely present in organic matter, and alkyl carbons connected to these heteroatoms are isotopically heavier.Due to the instability of the C-O or C-S bonds, these isotopically heavier alkyl groups were preferentially decomposed to form isotopically heavier methane.When the alkyl carbons connected to The calculated evolution trend of instantaneous and cumulative δ 13 C 1 [3].subfigure (a) represents predicted isotopic composition versus temperature for instantaneous and cumulative methane generated from pyrolysis of n-octadecane; subfigure (b) represents Calculated isotopic composition versus gas yield for representative geological and laboratory heating rates.
The mechanism responsible for the isotope rollover of δ 13 C 1 involves the presence of two or more precursors for methane generation with different isotopic compositions, and the mixing of methane originating from different precursors, which leads to the isotope rollover of C 1 [3].Heteroatoms such as O and S are widely present in organic matter, and alkyl carbons connected to these heteroatoms are isotopically heavier.Due to the instability of the C-O or C-S bonds, these isotopically heavier alkyl groups were preferentially decomposed to form isotopically heavier methane.When the alkyl carbons connected to heteroatoms were almost cracked, the isotopically lighter methane originating from the cracking of more tightly-bound C-C bonds was extensively generated.The initial trend of decreasing δ 13 C 1 values can be explained by the mixing of isotopically lighter methane.For alkyl groups connected with the C-C bond, 12 C-enriched alkyl groups were preferentially decomposed to form isotopically lighter methane.With increasing evolution degree, 13 Cenriched alkyl groups were gradually decomposed to form isotopically heavier methane, resulting in the increase of δ 13 C 1 .However, the rollover of δ 13 C 1 was also observed in many pyrolysis experiments on pure n-alkanes, which are free of heteroatoms, indicating that this theory cannot cover all of the δ 13 C 1 rollover occurrences.There are two sets of methane precursors with significantly different activation energies in the hydrocarbon source, and the precursor with the lower activation energy typically has a higher isotope fractionation factor and tends to generate isotopically lighter methane, while the precursor with the higher activation energy typically has a lower isotope fractionation factor and tends to generate isotopically heavier methane [3].Since the precursor with the lower activation energy was preferentially decomposed to form isotopically heavier methane, the variation trends of δ 13 C 1 initially decrease with the mixing of isotopically lighter methane originating from the decomposition of the precursor with the higher activation.When the precursor with the lower activation was almost decomposed, the precursor with the higher activation was gradually decomposed to form isotopically heavier methane, leading to the increase of δ 13 C 1 [6].
heteroatoms were almost cracked, the isotopically lighter methane originating from the cracking of more tightly-bound C-C bonds was extensively generated.The initial trend of decreasing δ 13 C1 values can be explained by the mixing of isotopically lighter methane.For alkyl groups connected with the C-C bond, 12 C-enriched alkyl groups were preferentially decomposed to form isotopically lighter methane.With increasing evolution degree, 13 C-enriched alkyl groups were gradually decomposed to form isotopically heavier methane, resulting in the increase of δ 13 C1.However, the rollover of δ 13 C1 was also observed in many pyrolysis experiments on pure n-alkanes, which are free of heteroatoms, indicating that this theory cannot cover all of the δ 13 C1 rollover occurrences.(b) There are two sets of methane precursors with significantly different activation energies in the hydrocarbon source, and the precursor with the lower activation energy typically has a higher isotope fractionation factor and tends to generate isotopically lighter methane, while the precursor with the higher activation energy typically has a lower isotope fractionation factor and tends to generate isotopically heavier methane [3].Since the precursor with the lower activation energy was preferentially decomposed to form isotopically heavier methane, the variation trends of δ 13 C1 initially decrease with the mixing of isotopically lighter methane originating from the decomposition of the precursor with the higher activation.When the precursor with the lower activation was almost decomposed, the precursor with the higher activation was gradually decomposed to form isotopically heavier methane, leading to the increase of δ 13 C1 [6].A sudden and sharp increase of instantaneousδ 13 C 1 was observed at the end of the methane generation stage, but the cumulative δ 13 C 1 shows an only very gentle increase at this stage.The sharply increasing instantaneous δ 13 C 1 was also observed when the methane conversion rate approached 1.0 in the study of Tang et al. (2000) (Figure 3b) [3].In addition, Liu and Xu (1999) also documented the sudden and sharp increase of δ 13 C 1 in 106 coal-type gas samples from 10 basins in China at the high-mature and over-mature stages, and a two-stage model of carbon isotopic fractionation was built to explain the fractionation mechanism of δ 13 C 1 [23].As mentioned in Section 2.1.1,methane is mainly generated from the cracking of kerogen, and the instantaneous yield is very limited at the end of the methane generation stage; moreover, there is no cracking of methane in this period.Accordingly, the sudden and sharp increase of instantaneous δ 13 C 1 may be attributed to the sudden transformation of the precursor with the higher activation energy and lower isotope fractionation factor.Since previous normal methane precursors have been mostly cracked at the end of the methane generation stage, the continuously advancing evolutionary process may suddenly result in the cracking of the unconventional precursor with the higher activation energy, leading to the sudden and sharp increase of instantaneous δ 13 C 1 .As the instantaneous C 1 yield is very limited in this period, the sharp increase of instantaneous δ 13 C 1 did not lead to the obvious increase of cumulative δ 13 C 1 .

Dynamic Isotopic Evolution of δ 13 C 2-3 during Hydrocarbon Generation Process
The cumulative δ 13 C 2 /δ 13 C 3 and instantaneous δ 13 C 2 /δ 13 C 3 from the pyrolysis of the Pingliang Shale all present an identical variation pattern: a cumulative δ 13 C 2 /δ 13 C 3 increase continuously with increasing maturity.The instantaneous δ 13 C 2 /δ 13 C 3 present a two-stage pattern (Figure 2).The instantaneous δ 13 C 2 /δ 13 C 3 in the generation stage (instantaneous yield > 0) mainly represents the isotope of instantaneously generated C 2 /C 3 , while the instantaneous δ 13 C 2 /δ 13 C 3 in the consumption stage (instantaneous yield < 0) mainly represents the isotope of instantaneously cracked C 2 /C 3 .The instantaneously generated δ 13 C 2 /δ 13 C 3 increase with raising maturity until the end of C 2 /C 3 generation, which is the result of the preferential generation of 12 C-enriched C 2 /C 3 .The instantaneously consumed δ 13 C 2 /δ 13 C 3 increase with raising maturity until the exhaustion of C 2 /C 3 , which can all be attributed to the preferential decomposition of 12 C-enriched C 2 /C 3 .This is similar to the variation trend of instantaneous δ 13 C 1 at the end of the methane generation stage.The instantaneously generated δ 13 C 2 /δ 13 C 3 also present a sudden and sharp increase at the end of the C 2 /C 3 generation stage, which may also be attributed to the cracking of the unconventional precursor with the higher activation energy.
The variation trend of cumulativeδ 13 C 2 /δ 13 C 3 was controlled by the relative isotopic composition values between cumulativeδ 13 C 2 /δ 13 C 3 and instantaneously generated δ 13 C 2 /δ 13 C 3 in the generation stage, as well as the relative isotopic composition values between cumulative δ 13 C 2 /δ 13 C 3 and instantaneously consumed δ 13 C 2 /δ 13 C 3 in the consumption stage.In the generation stage, the instantaneously generated δ 13 C 2 /δ 13 C 3 is heavier than the corresponding cumulative δ 13 C 2 /δ 13 C 3 ; accordingly, the cumulative δ 13 C 2 /δ 13 C 3 increase continuously with raising maturity.In the consumption stage, the instantaneously consumed δ 13 C 2 /δ 13 C 3 is lighter than the corresponding cumulative δ 13 C 2 /δ 13 C 3 ; accordingly, the cumulative δ 13 C 2 /δ 13 C 3 still increase continuously with raising maturity.As commonly acknowledged, the bond energy of 12 C- 12 C is lighter than that of 12 C-13 C and 12 C-13 C, so 12 C-enriched C 2 /C 3 should preferentially be decomposed during the consumption stage.As a result, the δ 13 C of residual δ 13 C 2 /δ 13 C 3 (cumulative δ 13 C 2 /δ 13 C 3 ) should be heavier than instantaneously consumed δ 13 C 2 /δ 13 C 3 during the consumption stage.This is the case observed in most of the consumption stages for C 2 /C 3 but, at the end of the consumption stage, the instantaneously consumed δ 13 C 2 /δ 13 C 3 reversed to be more positive than cumulative δ 13 C 2 /δ 13 C 3 (yellow labelled area in Figure 2b,c), which is unreasonable according to the bond energy theory.The reversal of instantaneously consumed δ 13 C 2 /δ 13 C 3 and cumulative δ 13 C 2 /δ 13 C 3 happened when the cumulative C 2 and C 3 have almost been decomposed (Easy Ro > 3.5% and 3.2%, respectively).As shown in Figure 2, the reversal of instantaneously consumed 13 C 2 /δ 13 C 3 and cumulative δ 13 C 2 /δ 13 C 3 can mainly be ascribed to the sudden and sharp increase of instantaneously consumed 13 C 2 /δ 13 C 3 , which is similar to the sudden and sharp increase of instantaneously generated δ 13 C 1 .Accordingly, it can be speculated that the sudden and sharp increase of instantaneously consumed 13 C 2 /δ 13 C 3 may also be attributed to the sudden transformation of an unconventional precursor, e.g., propane analogues with unconventionally heavier intra-molecular isotope distribution and higher activation energy, when common precursors have been mostly cracked.

The Universal Chemical and Isotopic Evolution Laws of C 1-3 in Hydrocarbon Generation Simulation Experiments
The universal evolution laws of hydrocarbon composition and isotopes were summarized by the data fitting of massive hydrocarbon generation simulation experiments involving various hydrocarbon sources, obtained from the available literature, including type I kerogen/source rock, type II kerogen/source rock, type III kerogen/source rock, crude oil, and asphalt, etc. [26,28,[30][31][32][33][46][47][48][49][50][51][52][53][54][55][56][57][58] (Figure 4).For the evolution laws of C 1-3 yields, C 1 yields increase continuously with rising maturity, while C 2 and C 3 yields initially increase and then decrease with increasing thermal maturity, reaching maximum yields at a certain intermediate maturity.As regards the evolution laws of δ 13 C 1-3 , δ 13 C 1 presents obvious isotope rollover, initially shifting negatively and then shifting positively with increasing maturity; δ 13 C 2 and δ 13 C 3 in certain simulation experiments also present isotope rollover with increasing maturity, but the variation amplitude is much lower than that of δ 13 C 1 .Overall, δ 13 C 2 and δ 13 C 3 generally shift positively with increasing maturity (Figure 4).
Since the samples adopted in the hydrocarbon generation simulation experiments are of different organic matter type and different maturities, the yields and isotope compositions of C 1-3 may vary greatly from each other; however, the yields and isotope compositions of C 1-3 present universal variation trends.Although these universal laws are empirical models summarized from massive data statistics (only the variation of C 1 yields and δ 13 C 1 values are verified by hydrocarbon generation kinetics and isotope kinetics), the massive hydrocarbon generation simulation data of various hydrocarbon sources all comply with these universal laws, convincing us that these universal laws are applicable to the hydrocarbon generation process of various organic matter.

Calculation of Isotopic Composition of Mixed Natural Gas from Different Maturities (δ 13 C mixed )
The carbon isotope of mixed natural gas at different maturities can be formulated by Equations ( 1) and ( 2), according to the definition of the natural gas isotope.
where R denotes the isotope ratio of 13 C/ 12 C of the PDB standard.
A and B represent different natural gas end members; X and Y represent the proportion of A and B in mixed natural gas, respectively; n i represents the concentration of alkane component i in natural gas; δ 13 C i represents the carbon isotope of component i; ∆ = δ 13 C/1000 + 1; R represents the 13 C/ 12 C ratio of the PDB standard.However, too many parameters in Equation (2) make it too complex to calculate δ 13 C mixed .
Xia et al. (1998) put forward a simplified formula to study the influence of mixture on natural gas isotopes (Equation ( 3)) [59].Although the simplified formula has been widely applied to the isotopic study of natural gas, some researchers are still skeptical of this formula and insist that it is only applicable to limited proportions of end members and may lead to obvious mistakes beyond limited proportions.The comparative study of the definition formula of mixed natural gas (Equation ( 2)) and the simplified formula (Equation (3)) were conducted under different proportions (A:B = 1:9999~9999:1) to verify the validity of the simplified formula.To strengthen reliability and persuasion, the scope between δ 13 C iA and δ 13 C iB should cover the variation range of δ 13 C i in geological and experimental conditions.In addition, the difference between δ 13 C iA and δ 13 C iB should be big enough.Accordingly, we assign δ 13 C iA and δ 13 C iB to be −10‰ and −50‰, respectively.As shown in Table 1, the δ 13 C mixed values calculated according to the simplified formula are very close to those calculated according to the definition formula under obviously different proportions, with only a minor difference on the fifth significant digit.To date, the precision of one digit after the decimal point for δ 13 C is sufficient to meet the requirement of the isotope study of natural gas.So, the simplified formula of δ 13 C mixed can be widely applied to the isotopic study of natural gas under diverse proportions of different end members.

Figure 1 .
Figure 1.The instantaneous and cumulative yield of C1-3 from the pyrolysis of the Pingliang Shale at different maturities, squares in subfigure (a) represents cumulative and instantaneous yield of methane, circulars in subfigure (b) represents cumulative and instantaneous yield of ethane, and triangles in subfigure (c) represents cumulative and instantaneous yield of propane.

Figure 2 .
Figure 2. The instantaneous and cumulative isotope of C1-3 from the pyrolysis of the Pingliang Shale at different maturities, squares in subfigure (a) represents cumulative and instantaneous isotope of methane, circulars in subfigure (b) represents cumulative and instantaneous isotope of ethane, and triangles in subfigure (c) represents cumulative and instantaneous isotope of propane.

Figure 1 . 15 Figure 1 .
Figure 1.The instantaneous and cumulative yield of C 1-3 from the pyrolysis of the Pingliang Shale at different maturities, squares in subfigure (a) represents cumulative and instantaneous yield of methane, circulars in subfigure (b) represents cumulative and instantaneous yield of ethane, and triangles in subfigure (c) represents cumulative and instantaneous yield of propane.

Figure 2 .
Figure 2. The instantaneous and cumulative isotope of C1-3 from the pyrolysis of the Pingliang Shale at different maturities, squares in subfigure (a) represents cumulative and instantaneous isotope of methane, circulars in subfigure (b) represents cumulative and instantaneous isotope of ethane, and triangles in subfigure (c) represents cumulative and instantaneous isotope of propane.

Figure 2 .
Figure 2. The instantaneous and cumulative isotope of C 1-3 from the pyrolysis of the Pingliang Shale at different maturities, squares in subfigure (a) represents cumulative and instantaneous isotope of methane, circulars in subfigure (b) represents cumulative and instantaneous isotope of ethane, and triangles in subfigure (c) represents cumulative and instantaneous isotope of propane.

Figure 3 .
Figure 3.The calculated evolution trend of instantaneous and cumulative δ 13 C1 [3].subfigure (a) represents predicted isotopic composition versus temperature for instantaneous and cumulative methane generated from pyrolysis of n-octadecane; subfigure (b) represents Calculated isotopic composition versus gas yield for representative geological and laboratory heating rates.

Figure 3 .
Figure 3.The calculated evolution trend of instantaneous and cumulative δ 13 C 1[3].subfigure (a) represents predicted isotopic composition versus temperature for instantaneous and cumulative methane generated from pyrolysis of n-octadecane; subfigure (b) represents Calculated isotopic composition versus gas yield for representative geological and laboratory heating rates.

Figure 4 .
Figure 4.The universal evolution laws summarized from massive hydrocarbon generation simulation experiments.Subfigure (a) represents cumulative yield and isotopic composition of methane generated from pyrolysis experiments, subfigure (b) represents cumulative yield and isotopic composition of ethane generated from pyrolysis experiments, subfigure (c) represents cumulative yield and isotopic composition of propane generated from pyrolysis experiments.

Figure 4 .
Figure 4.The universal evolution laws summarized from massive hydrocarbon generation simulation experiments [26,28,30-33,46-58].Subfigure (a) represents cumulative yield and isotopic composition of methane generated from pyrolysis experiments, subfigure (b) represents cumulative yield and isotopic composition of ethane generated from pyrolysis experiments, subfigure (c) represents cumulative yield and isotopic composition of propane generated from pyrolysis experiments.

Table 1 .
The δ13C mixed value of mixed natural gas with different proportions of end members A and B.

Proportion of End Member B (%) δ 13 C mixed according to Definition Formula (‰) δ 13 C mixed according to Simplified Formula (‰)
Instantaneous yield: Yield of certain alkane gas at certain single maturity point or single point of geological time;(3) Cumulative isotope: Isotopic composition of alkane generated in certain maturity interval or geological period; (4) Instantaneous isotope: Isotopic composition of alkane generated at certain single maturity point or single point of geological time.