The study area is located in Cretaceous Qing 1 Member in northern Songliao Basin where typical continental shale oil is developed. In recent years, breakthroughs have been made successively, and a national demonstration area of shale oil development was established, showing promising prospects ^[3]. In this study, 42 samples of Qing 1 Member shale were selected, and 13 samples with R_o of 0.83%-1.65%, Type I and II₁ OM, TOC of 1.41%-3.75%, and S₁ of 1.21-4.39 mg/g were selected for kerogen enrichment. The mineral is dominated by clay with average content of 42.68% and quartz with average content of 23.15%, followed by pyrite and calcite with contents of less than 10% (Table 1). The TOC of kerogen is 37.15%-59.84%. Previous study shows that most OM in the shale are adsorbed on the clay mineral surface and filled in the pores, and the organic-clay complex has the stable properties ^[2,35]. This possibly causes the undesirable purity of enriched kerogen samples.

Low temperature nitrogen adsorption experiments show that the shale pore specific surface area is 3.48-34.65 m²/g, with an average of 25.50 m²/g, and it is positively correlated to the clay mineral content. The specific surface area of kerogen is 0.55-74.92 m²/g, with an average of 28.86 m²/g. The specific surface area of kerogen increases with increase of R_o. When R_o is greater than 1.3%, the organic pore scale increases accompanied by light hydrocarbon cracking ^[36], and the specific surface area of kerogen is larger than that of shale.

No. 2, 8, 10 and 13 shale samples were selected to separate the pure clay by hydrostatic water sedimentation. The mineral components and specific surface area parameters of clay samples are listed in Table 2. The separated clay samples have the clay content higher than 60%, followed by quartz with the content of about 1/4. Clay is dominated by illite/smectite formation, with a small amount of chlorite. The specific surface area of clay sample is 36.33-47.23 m²/g, which is higher than that of corresponding shale, indicating that clay pores mainly contribute to shale pore specific surface area.

A total of 104 experiments of swelling oil were carried out with 13 kerogen samples at various maturity stage and 8 organic solvent samples. The experiments were carried out with following steps. (1) Pick a dry and clean glass tube with open ends. (2) Stack and compact the high-purity quartz cotton at the tube bottom, use a spoon to send the kerogen sample to the tube bottom, and prevent kerogen adherence on the inner wall of the tube. Use a glass rod to tamp the kerogen sample and centrifuge it for 20 min at a speed of 8000 r/min. After centrifugation, measure the kerogen height with electronic vernier caliper for several times, and calculate the average and record it as H₁. (3) The tube is immersed vertically in the organic solvent, and when bubbles are observed, knock the tube wall slightly to escape bubbles. Referring to previous learning ^{[19,37⇓⇓⇓ -41]}, the immersion duration is set as 24 h. Centrifugation is operated again after immersion, and the kerogen height is recorded as H₂. The centrifugation speed is set here by checking when the change of the kerogen mass is stable.

The kerogen swelling ratio is calculated with the change of kerogen height before and after being immersed in solvent (q_V):

The organic solvents of N-octane, xylene, 2-methylthiophene, cyclohexanone, n-octanol, pyridine, ethanol and formic acid were selected, including alkane, benzene, ketone, alcohol, acid and non-hydrocarbon components. The solubility of saturated, aromatic and non-hydrocarbon asphaltene reported previously ^[34] is covered in this study.

The experiments of adsorption-swelling kerogen are operated with above procedures. A quartz tube with kerogen at the bottom is immersed vertically in an organic solvent and is placed in ovens at various temperatures for 24 h, and then, the quartz tube is taken out and is centrifuged at room temperature.

Experiments of oil adsorption by clay were carried out in Shimadzu UV-2450 ultraviolet spectrophotometer with 4 clay samples as the adsorbent, toluene and asphaltene as the adsorbate, and n-heptane as the solvent. The asphaltene is acquired from shale oil with R_o of about 0.9% of Qing 1 Member of Well A and prepared by washing with n-heptane, precipitation and centrifugation. The adsorption experiments are carried out as following steps. (1) Formulate the solutions of various concentrations, and operate full-wave band scanning to determine a characteristic peak of an adsorbate in a solvent. (2) Formulate a solution of a given concentration, scan a characteristic peak to determine the relation between the concentration and absorbance (ABS), and establish an absorbance-concentration calibration equation. (3) Add a small amount of clay into the solution, and operate ultrasonic oscillation during the immersion during of 24 h to ensure full adsorption. (4) When absorption is finished, take the supernatant to measure the absorbance, and calculate the solution concentration after absorption according to the calibration equation. Calculate the adsorbing amount q_c according to the difference in supernatant concentrations before and after adsorption:

Rock-Eval 6 instrument was used to carry out step-by-step Rock-Eval experiments with shale and the separated clay samples. The temperature step was controlled with the IFPEN method ^[42]. According to JIANG et al. ^[14], LI et al. ^[43] and our study ^[44], the hydrocarbon amount released in Rock-Eval below 450 °C is approximately equal to the amount of rock extracts (chloroform bitumen "A"), though kerogen cracking already occurs at 450 °C. Because the IFPEN method sets the temperature of pyrolysis hydrocarbon release as 350 °C, the pyrolysis hydrocarbon release amount below 450 °C is set as the total shale oil amount here.

SEM observation is carried out in ZEISS Crossbeam 550 FE-ESEM and EDAX energy spectrum. Before experiment, both sealed and conventionally cored samples were cut to the size of 1.0 cm×1.0 cm×0.2 cm and treated with argon ion polishing and plating. Observation of sealed coring samples requires quick preparation to reduce fluid loss. Shale oil occurrence is observed directly under low voltage and low current conditions and at 20 °C.

Fig. 2 shows the swelling ratio (q_V) of 13 kerogen samples with maturity of 0.83%-1.65% in different solvents in Qing 1 Member of the study area. With increase of organic solvent solubility parameter (δ_l) value, q_V of each kerogen type increases firstly and then decreases, showing Gaussian distribution. According to the theory of "compatibility with similar components" in polymers, q_V reaches a maximum when the solubility parameter is about 20 (J/cm³)^0.5. The swelling capacity of kerogen with various maturities to each hydrocarbon component generally shows following characteristics. First, more aromatic hydrocarbons are swollen and dissolved in kerogen than the saturated hydrocarbons, e.g. more xylenes are swollen than n-octane. Second, among aromatic hydrocarbons, hydrocarbons with hetero atoms such as oxygen, nitrogen and sulfur are more likely to be retained in kerogen, e.g. more pyridine and dimethyl thiophene are be retained than xylene. Besides, naphthenic hydrocarbons are more likely to be retained than the normal alkanes with the same carbon number ^[34]. Kerogen swells and retains most non-hydrocarbon asphaltene components, followed by aromatic hydrocarbons, and least saturated hydrocarbons.

For the single organic solvent, q_V of kerogen decreases as the maturity increases, indicating weakening swelling oil capacity. This is mainly due to the kerogen chemical structure. Previous X-ray diffraction analysis revealed that the most obvious structural unit of solid bitumen and liquid hydrocarbon during swelling oil is the amorphous carbon (γ band) composed of aliphatic carbon with the spacing reduced gradually and arrangement more compact. The liquid hydrocarbon hardly enters the micro-crystalline formed by tightly stacked aromatic carbons which are hardly affected by swelling ^[47]. Thus, the kerogen with the higher ration of aliphatic carbon absorbs and swells more hydrocarbons. With increase of maturity, the unstable structures in kerogen such as branched chain/side chain gradually fall off, the aliphatic carbon content decreases, aromatization is enhanced, the crosslinking density increases, the structure is more rigid^[48], and the oil swelling capacity is weakened (Fig. 2).

Based on the swelling ratio of kerogen adsorbing and swelling various organic solvents, the parameters K and d in the optimization model are fitted according to Eq. (4). The object of the kerogen adsorption-swelling experiment is simple organic solvent (Fig. 2). The properties of those single solvents and shale oil differ greatly, thus, the composition and density of crude oil should be considered to evaluate actual adsorption-swelling shale oil amount of kerogen. The shale oil component and density can be predicted according to the maturity of the study area (Fig. 3). The solubility parameters of shale oil with various maturities can be obtained by weighted average of component family proportions and solubility parameters according to Eq. (5). Based on this, K of various maturity kerogen is coupled to get the absorbing-swelling shale oil amount by kerogen according to Eqs. (2) and (3).

Taking 13 samples in this study as an example, when R_o is 0.83%-1.65%, the adsorption-swelling oil capacity (or sorption oil capacity) of kerogen is about 50-250 mg/g. According to the swelling oil experiments of kerogen simulated thermally in gold tubes and that enriched in naturally evolved samples in Fig. 4, the swelling oil amount of kerogen decreases with increase of R_o within a wide range. For kerogen R_o < 0.6%, Ballice and Larsen ^[49] carried out swelling experiments of oil shale from Goynuk area of Turkey by thermal simulation to various maturity and derived average q_V of oil shale close to 1.5.Here, the relation between the oil density and R_o (Fig. 3) is used to convert them swelling oil amount to 500 mg/g. According to the swelling oil ratio of kerogen with R_o of 0.47%-0.92% in naturally evolved samples from Kelemen et al. ^[19], the swelling oil amount is estimated as about 250-550 mg/g. There are differences in swelling oil amount of kerogen obtained in different study, and the swelling oil amount of kerogen reported by Guangzhou Geochemical Research Institute of Chinese Academy of Sciences also varies. For example, WEI et al.^[39] conducted thermal simulation on kerogen from lower sub-member of Shahejie Formation 3^rd Member of Jiyang Depression in golden tube, and found that the swelling oil amount of kerogen with R_o of 0.6%-1.0% is about 50-120 mg/g. SUN et al. ^[41] proposed that for kerogen in the upper sub-member of Shahejie Formation 4^th Member of Jiyang Depression, when R_o is 0.7%-1.25%, the swelling oil amount is about 100-300 mg/g, which is close to the results in this study. Beside the maturity, the swelling oil capacity of kerogen is related to its type and thermal evolution. Walters et al. ^[50] reported that the swelling ratio of Type II kerogen is higher than that of Type III kerogen.The swelling ratio of kerogen in naturally evolved samples is generally higher than that in thermally simulated samples. We suggest that this is possibly due to the kerogen chemical structure. Geological evolution of kerogen mostly occurs below 200 °C, and thermal simulation is carried out above 300 °C and even at 600 °C. High temperature possibly changes the physicochemical properties of kerogen and causes chemical reaction that does not occur at geologic temperature less than 200 °C ^[51]. Secondary cracking of hydrocarbon is more significant in thermal simulation experiment in closed system, and a large amount of coke is generated during coking reaction ^[52], which causes more concentrated aromatic ring structure and higher cross-linking density, resulting in the swelling capacity possibly lower than that of naturally evolved sample. According to the measured swelling oil amount of 13 naturally evolved kerogen, the relationship between the swelling oil amount and R_o was fitted (Fig. 4).

The kerogen swelling oil experiments above were conducted under conditions of abundant oil supply. In the course of geological evolution, the hydrocarbon generated by kerogen in the early stage of low maturity or immature stage cannot realize full saturation. The actual swelling oil amount of kerogen should be determined by considering the effects of hydrocarbon generation. Previously, hydrocarbon generation thermal simulation experiments on dark black shale (TOC=11.02%, Type I, R_o=0.54%) in the study area was completed in Guangzhou Institute of Geochemistry, Chinese Academy of Sciences, and the hydrocarbon generation curve of kerogen was established, as shown in black dashed line in Fig. 4. When R_o is less than 0.8%, hydrocarbon generated by kerogen cannot realize full saturation, and its actual adsorption- swelling oil amount is equal to the oil generation amount M_g. When R_o is equal to 0.8%, the hydrocarbon generation amount of kerogen approximates to the adsorption-swelling oil amount. When R_o is greater than 0.8%, the oil generation amount M_g of kerogen is larger than the adsorption-swelling oil amount (Fig. 4). Thus, the actual adsorption-swelling oil amount of kerogen is expressed as:

The accuracy of M_k quantification in this study is possibly affected by the low purity of the enriched kerogen samples (the kerogen TOC is basically less than 60%, as listed in Table 1). Currently, there is controversy on effect of the kerogen purity on its swelling amount. Ballice ^[53] carried out swelling oil experiments with oil shale from Goynuk area in Turkey, and found that after removing the minerals from shale, the swelling amount of some samples increased. Nevertheless, Larsen et al. ^[37] suggested that inorganic minerals have little effects on shale swelling, even if the mineral content in kerogen reaches 40%. In addition, due to the difference in solubility parameters of various solvents, the solvents with a medium-high solubility parameter reach the equilibrium in a short time, and those with the low solubility parameters take a long time ^[54]. In this study, the duration for immersing the solvent in kerogen is set as 24 h, which is based on previous learning, and it possibly increases the uncertainty of quantification of swelling oil.

The above experiments of kerogen adsorption-swelling oil and clay adsorbed oil were completed under ambient conditions (25 °C, and 0.1 MPa). The effects of temperature and pressure should be considered in in-situ conditions.

(1) To reveal the effect mechanism of temperature on adsorption-swelling oil amount of kerogen, experiments of absorption-swelling oil were carried out at different temperatures with No.7 and No.10 kerogen samples. Change in volume of kerogen and solvent due to thermal effects should be considered, and there is a large difficulty of evaluating the kerogen swelling ratio by height method in Eq. (8). Here, the mass method was applied, the mass of kerogen after adsorption-swelling experiments at different temperatures is recorded (Fig. 11a). The results of both samples show that the kerogen mass after swelling oil at 60 °C and 80 °C is similar to that at room temperature. According to Ertas et al. ^[38], the kerogen swelling ratio only changes slightly at 30-150 °C, which indicates little effect of temperature on swelling. This is due to that temperature has little effect on solubility parameters, the solubility parameters only change 1-2 (J/cm³)^0.5 from room temperature to 400 °C ^[34]. The effect of temperature on oil adsorption by clay is referred to the curve of adsorbed oil capacity of clay at different temperatures established by Tian Shansi ^[62] (Fig. 11b).

(2) For the effect of pressure, Pathak et al. ^[27] carried out molecular simulation and revealed that there is basically no difference in kerogen at normal pressure and 30 MPa, indicating that pressure has little effect on shale oil adsorption ^[62⇓-64].

The effect of temperature on clay adsorption oil is mainly considered here. The adsorbed oil capacity of clay is coupled with the current geothermal gradient (4.2 °C/100 m) in the study area, and the adsorbed oil capacity of clay with various maturity are established, as illustrated in Fig. 12.

Based on the above model and the effect of temperature and pressure on adsorption oil, the total adsorbed oil amount of Qing 1 Member shale with R_o of 0.85%-1.65% in 6 wells in the study area is estimated. Specific surface area of clay pores in shale is calculated by subtracting the shale specific surface area by the kerogen specific surface area, and the former is calculated according to formula for conversion of porosity-specific surface area ^[23], and kerogen specific surface area is estimated with maturity.

As shown in Fig. 13a, the predicted total adsorbed oil amount is 0.62-10.68 mg/g, with an average of 3.48 mg/g, and it decreases as the shale maturity increases. When the geothermal effect is neglected, the predicted total adsorbed oil amount of shale is 1.17-11.33 mg/g, with an average of 4.33 mg/g, which is basically consistent with the evaluation results of 42 cores. Taking Well F as an example, Fig. 13b shows the vertical prediction results of adsorption-swelling oil amounts of kerogen, adsorbed oil amount of clay pores and shale total adsorbed oil amount respectively. When R_o is 1.55%-1.65%, the estimated total adsorbed oil amount of shale under the formation conditions of this well ranges from 1.40 mg/g to 4.13 mg/g, with an average of 2.47 mg/g. From the perspective of vertical prediction results, although the total adsorbed oil content evaluated at room temperature (Fig. 13b, green line) is very close to the measured value (semi filled scatter), it is different from those obtained at room temperature. When considering the influence of formation temperature, the reduction effect of temperature on the amount of adsorbed oil by clay is obvious, leading to that the adsorbed oil from shale still dominated by organic matter.

Previous evaluation of absorbed oil amount of shale is mostly based on understanding obtained in experiments under conditions of room temperature and pressure ^[5,12,15,26]. As shown in Fig. 14, the laboratory test at room temperature often overestimates the adsorbed oil amount of shale about 8%-22%, with an average of about 15%, and the higher shale maturity causes a larger deviation. Therefore, the effects of shale maturity and formation temperature should be considered in evaluating the adsorbed oil amount of high maturity shale.

It should be noted that the adsorbed oil amount measured and predicted by rock pyrolysis here mainly refers to the heavy hydrocarbons with relatively high boiling point. The research object is conventionally cored shale samples where a large amount of light components are lost. Some proposed that the adsorbed oil is not lost ^[9]. Nevertheless, authors found that the pyrolysis analysis parameter of S₂ peak shows a decreasing trend in some sealed coring shale exposed in the open environment, indicating loss of some adsorbed oil. In addition, the adsorbed oil on pores and swelling oil in kerogen skeleton are quantified uniformly there, but their occurrence mechanisms are different. Especially for the high maturity shale, adsorption capacity of the organic pores is possibly dominant. The evaluation model and chart for adsorbed oil presented here and prediction of adsorbed oil amount of shale under conditions of in-situ temperature and pressure with molecular simulation method provide references for the industry.