Adsorption of water in organic matter, clay minerals and deep gas-bearing reservoir: Insights from experimentation and molecular simulation

Xianwei Heng; Xijian Li; Jinlei Fu; Xidong Du; Tengfei Wu

doi:10.25259/AJC_1261_2025

View/Download PDF

Buy Reprints

PDF

Translate this page into:

Original Article

ARTICLE IN PRESS

doi:

10.25259/AJC_1261_2025

Adsorption of water in organic matter, clay minerals and deep gas-bearing reservoir: Insights from experimentation and molecular simulation

Xianwei Heng^a,b, Xijian Li^a,*, Jinlei Fu^b, Xidong Du^c,d,e,, Tengfei Wu^f

aCollege of Resources and Environmental Engineering, Guizhou University, Guiyang, Guizhou, PR China

bGuizhou Mine Safety Scientific Research Institute Co., Ltd., Guiyang, Guizhou, PR China

cFaculty of Land Resources Engineering, Kunming University of Science and Technology, Kunming, Yunnan, PR China

dKey Laboratory of Petroleum Resources Exploration and Evaluation, Lanzhou, Gansu, PR China

eNorthwest Institute of Eco-Environment and Resources, Chinese Academy of Sciences, Lanzhou, Gansu, PR China

fHenan Energy Group Research Institute Co., Ltd., Zhengzhou, Henan, PR China

*Corresponding author: E-mail address: xjli1@gzu.edu.cn (X. Li)

Received: 2025-10-21, Accepted: 2026-01-19, Epub ahead of print: 2026-04-06,

Licence

This is an open-access article distributed under the terms of the Creative Commons Attribution-Non Commercial-Share Alike 4.0 License, which allows others to remix, transform, and build upon the work non-commercially, as long as the author is credited and the new creations are licensed under the identical terms.

Abstract

In this research, the isothermal curves for water molecule adsorption in gas-bearing reservoir were obtained through gravimetric adsorption experiment. The adsorption performances of water molecule on the three individual components of organic matter, illite and montmorillonite were analyzed through molecular simulation method. Then, combining experiment and molecular simulation results, the adsorption mechanism of water on gas-bearing reservoir was discussed. The findings reveal that at the small loading range, primary adsorption regulated through surface functional groups predominates during the adsorption process. At the large loading range, secondary adsorption determined through water cluster formation within inner pore space plays a crucial role. Under high pressure, inner pore structure possesses an obviously stronger effect on water adsorption than surface functional groups. The effect for water on gas transport is primarily attributed to the formation of water cluster during secondary adsorption. For organic matter, water molecule adsorption only appears in the area near the two sides of pore wall. For clay minerals, water molecule adsorption occurs both in the central area and two sides of nanopore. The uptake for water molecule on organic matter is significantly lower than those on clay minerals. Clay minerals dominate water adsorption on gas-bearing reservoir. The mode for water adsorption on gas-bearing reservoir differs from those on organic matter and clay minerals. Water molecules retained on clay minerals are more difficult to desorb, which has the more obvious effect on the pore connectivity of gas-bearing reservoir and the ultimate gas production behavior.

Keywords

Adsorption

Clay minerals

Gas-bearing reservoir

Organic matter

Water

Show Related Articles from PubMed

1. Introduction

With the continuous deterioration of global ecological environment, the need of clean fossil fuels such as natural gas resources is constantly increasing. Presently, deep gas-bearing reservoir has received many attentions from scholars and engineers [1,2]. Because of the extremely low permeability for gas-bearing reservoir, the traditional depressurization mining method is difficult to obtain stable and efficient gas recovery. The use of water-based hydraulic fracturing technology to fracture gas-bearing reservoir and increase reservoir permeability has become the key to improve gas recovery [3]. However, when the fracturing process is completed, most of the injected fluid remains at the underground reservoir [4]. Through blocking the inner pore, the fluid is able to prevent the movement of gas and ultimately affect gas well production behavior [5]. Thereby, for better applying and optimizing the technology of hydraulic fracturing into gas-bearing reservoir, researching water-reservoir interaction has great significance.

The composition of gas-bearing reservoir is very complex, and two major components of gas-bearing reservoir are organic matter and inorganic minerals. Through investigating the physical-chemical characteristics for gas-bearing reservoir components, it is discovered that clay minerals within inorganic minerals and organic matter all have rich pore structure and offer abundant adsorption space for fluid [6]. Importantly, because of the possession of polar groups and exchangeable cations, clay minerals are hydrophilic and can effectively promote water adsorption [7]. Hence, the interaction process of water-reservoir requires investigating the adsorption performances for water in organic matter and inorganic minerals, especially the water-clay minerals interaction. Meanwhile, according to the sedimentary environment and original material sources, deep gas-bearing reservoir types can be divided into continental reservoir, marine reservoir and marine-continental transitional reservoir. In marine reservoir, there are many brittle mineral components such as quartz and feldspar [8]. In continental reservoir, it is mainly rich in clay minerals, concentrated at 55% to 68%, due to the dilution effect of continental debris [9]. In view of the hydrophilic characteristics of clay minerals, understanding the water-continental reservoir interaction is more urgent when adopting water-based fracturing technology in continental gas formation.

Regarding the adsorption process for water in clay minerals and gas-bearing reservoir, many researches have conducted some efforts through the methods of experiment and molecular simulation [4,5,7,10-18]. In terms of experiment research, Duan and Li [10] investigated water vapor adsorption performance on marine reservoir of Sichuan Basin of China and found that monolayer adsorption quantity improves with clay mineral content. Sang et al. [11,12] studied the process for water vapor adsorption on marine reservoirs collected from Illinois basin of United States and Sichuan Basin of China under dynamic water vapor condition. Their results show that pore structure and surface chemistry can both affect reservoir hydration. Tang et al. [4] measured the isothermal ad/desorption curves of water vapor under different temperature on marine Marcellus reservoir and reported that the similar hysteresis loops are appeared between desorption and adsorption isotherms at different temperatures. Chen et al. [5] discussed the interaction mechanism of water on marine reservoir sampled from Sichuan Basin of China. Their findings indicate that the surface chemistry characteristics control the cluster mechanism of adsorption process of water. Zolfaghari et al. [14] researched the interaction of reservoir collected from the Horn River Basin of Canada with water and observed that the higher the content of clay minerals, the faster the adsorption rate and the stronger the adsorption ability. In terms of molecular simulation research, Li et al. [15] analyzed water adsorption on uneven organic matter surface and pointed out rough surface is conducive to water adsorption. Zhang et al. [16] investigated the impact of pore structure and chemical structure on water adsorption in organic matter and discovered that the narrow ultramicropores within hydrophilic oxygen-containing structures are the primary adsorption centers. Zhang et al. [17] researched the occurrence of water on illite pore structure. They reported that the process for water adsorption on illite can be divided into three different stages. Liu et al. [18] studied the migration process of water in clay nanopore and indicated that the migration speed of water is related to temperature and pressure. As mentioned in the above literature, although water adsorption on deep gas-bearing reservoir has received lots of attentions, the published data mainly focuses on the hydration on marine reservoir and the studies about water adsorption process on continental reservoir are very lacking. Meanwhile, there is still dearth of the explanation about the occurrence mechanism for water in continental reservoir from molecular level. Considering that the main adsorption space is provide by clay minerals and organic matter, for better understanding water adsorption mechanism on continental reservoir, studying and comparing the adsorption behaviors for water on organic matter and different clay minerals by the combined methods of physical experiment and molecular simulation is necessary.

For this study, we firstly measured the desorption and adsorption isotherms on continental reservoir for water vapor by the gravimetric adsorption apparatus. Through using three different adsorption models to match experimental data, water adsorption process and characteristics in continental reservoir were investigated. Then, the method of molecular simulation was adopted for constructing the models of organic matter (kerogen) and two different clay minerals (montmorillonite and illite). The performances for water adsorption in organic matter and two different clay minerals were analyzed and compared from molecular level. Lastly, combing experiment research and molecular simulation results, the underlying mechanism for water adsorption on continental reservoir was discussed. The gained results will provide theoretical support for water-based fracturing technology application into deep gas-bearing reservoir.

2. Materials and Methods

2.1. Geological setting

The adopted continental Yanchang sample was collected from Ordos Basin in Yanchang County, Yan’an City, Shaanxi Province, China. The Yanchang Formation was deposited during the Late Triassic period. Because of the Indosinian movement effect, the Ordos Basin transformed from being part of the Paleozoic Dahuabei Basin into an inland depression lake basin.

During the depositional period of the Yanchang Formation, the climate was humid and warm, predominantly fostering the development of delta-semi-deep lake deposit. The mudstone exhibits a high boron content, with a wide range of sedimentary water salinity, peaking at approximately 12%. The lake water is significantly influenced by volcanic hydrothermal activity, suggesting that the organic-rich mudstone has a hydrothermal origin. The warm and humid climate, coupled with hydrothermal activity in the lake area, facilitates the flourishing of organisms and the accumulation of organic matter in the lake basin, laying the foundation for the formation of organic-rich mudstone. The reduced conditions created by the deep water are crucial for the preservation of organic matter, while the water salinity has a limited impact on the enrichment of organic matter. The burial depth of adopted Yanchang sample is 1600 m.

2.2. Characterization of sample

2.2.1. Surface physical properties characterization

The surface morphology and roughness characteristics of sample were observed through Bruker Dimension icon/Dimension icon XR atomic force microscope (AFM). The shape, size and connectivity of pore were observed through Hitachi TM-300 scanning electron microscopy (SEM).

2.2.2. Surface chemical properties characterization

Unlike the nonpolar adsorption molecules (CO₂, N₂ and CH₄), the polar water molecule can be preferential trapped on some functional groups on matrix surface [5]. In this research, the chemical properties of matrix surface of sample were characterized using Fourier transform infrared (FTIR) spectrometer (Nicolet iS50, Thermo Scientific).

2.2.3. Pore structure measurement

The methods of adsorption of CO₂ and N₂ were utilized to measure sample pore structure. Because of the effect of molecular sieving, CO₂ can entry some smaller micropores, while N₂ entry is prevented completely or kinetically controlled [19]. Moreover, the measurement temperature of N₂ adsorption of 77 K is extremely low than that of CO₂ adsorption of 273 K. The higher temperature can offers more thermal energy to CO₂ molecule to transport in constricted pore structure [19]. Therefore, adopting the method of CO₂ adsorption to analyze micropore information is more accurate. The mesopore and macropore information of sample was measured through N₂ adsorption, and the micropore information was measured through CO₂ adsorption. The Micromeritics ASAP-2020M volumetric adsorption system was utilized to perform N₂ adsorption and CO₂ adsorption experiments. According to the obtained adsorption experiment data, pore structure information can be analyzed.

2.3. Adsorption experiment of water vapor

Presently, gravimetric method and volumetric method belong to the main methods for conducting adsorption measurement. Compared to volumetric method, the sensitivity of gravimetric method to measurement environment is lower, and the testing precision of gravimetric method is higher [20]. Hence, for this research, the adsorption experiment for water vapor in sample was conducted through the gravimetric apparatus (IGA-100 B, U.K.). Before the experiment, the 170 mg sample was put into the sample tank and outgassed for 12 h at 10^-5 Pa and 378 K. The adsorption temperatures are 298 K, 313 K and 328 K. The relative pressure for adsorption (P/P₀) ranges from 0 to 0.9 with the interval of 0.05. At each pressure point, the equilibrium time is 6 h.

2.4. Theory for water adsorption

2.4.1. Model for adsorption

The obtained adsorption isotherm is beneficial to evaluating the adsorption ability of adsorbent and understanding of pore structure information and surface characteristics of adsorbent [20]. Meanwhile, the adsorption isotherms also can reflect the interaction mechanism of adsorbent and adsorbate. At present, many adsorption models have been proposed to deal with isothermal adsorption curve of water [10]. In this study, three different adsorption models for Dent model, D’Arcy and Watt model and Modified Dent model were applied for matching gained isothermal curves of water adsorption.

Dent model assumes that primary binding sites and secondary binding sites exist on adsorbent surface [5]. The primary binding sites possess the greater binding energy. The secondary binding sites possess the lower binding energy. Dent model can be described through Eq. (1) [10].

(1)

q = \frac{q_{0} k_{1} P}{(1 - k_{2} P) (1 - k_{2} P + k_{1} P)}

where q represents the uptake at pressure P; q₀ is the saturated amount of monolayer adsorption; k₁ represents the constant in relation to primary binding sites energy; k₂ represents the constant in relation to secondary binding sites energy.

When utilizing Dent model, the adsorption quantity on primary binding sites (q₁) can be calculated by Eq. (2), and the adsorption quantity on secondary binding sites (q₂) can be gained by Eq. (3).

(2)

q_{1} = \frac{q_{0} k_{1} P}{1 - k_{2} P + k_{1} P}

(3)

q_{2} = \frac{q_{0} k_{1} k_{2} P^{2}}{(1 - k_{2} P) (1 - k_{2} P + k_{1} P)}

Modified Dent model also assumes that two kinds for adsorption sites exist on adsorbent surface. The parameter α in Modified Dent model indicates that α site can be occupied through one water molecule [10]. The form of Modified Dent model is given through Eq. (4) [10].

(4)

q = \frac{q_{0} k_{1} P^{α}}{(1 - k_{2} P^{α}) (1 - k_{2} P^{α} + k_{1} P^{α})}

where α reflects the heterogeneity for adsorption system.

When utilizing Modified Dent model, the value of q₁ can be obtained by Eq. (5), and Eq. (6) can be used to acquire q₂ value.

(5)

q_{1} = \frac{q_{0} k_{1} P^{α}}{1 - k_{2} P^{α} + k_{1} P^{α}}

(6)

q_{2} = \frac{q_{0} k_{1} k_{2} P^{2 α}}{(1 - k_{2} P^{α}) (1 - k_{2} P^{α} + k_{1} P^{α})}

D’Arcy and Watt model combines Langmuir model and original Dubinin and Sierpinski model and is according to the statistical adsorption thermodynamics analysis under multisite substrate [10]. Eq. (7) displays D’Arcy and Watt model [10]. In Eq. (7), the left term and the right term can be utilized to obtain primary and secondary adsorption amounts, respectively.

(7)

q = \frac{S_{0} K P}{1 + K P} + \frac{s k P}{1 - k P}

where S₀ and K are the density and interaction parameter on primary binding sites, respectively; s and k are the density and interaction parameter on secondary binding sites, respectively.

After fitting the adsorption isotherm with a model, the parameter values in the model can be obtained. Then, by substituting these parameter values into the corresponding formulas, the primary adsorption amount and secondary adsorption amount can be obtained. However, it should be noted that the adsorption model may not uniquely separate specific chemical sites.

When adopting above three adsorption models to match the gained adsorption isotherms, the fitting accuracy was estimated through the four parameters of root-mean-square error (RMSE), determination coefficient (R²), average relative deviation (ARE) and sum of squares of error (SSE). R² value, ARE value, SSE value and RMSE value can be calculated through Eq. (8), Eq. (9), Eq. (10) and Eq. (11), respectively.

(8)

R^{2} = 1 - \frac{\sum_{i = 1}^{n} {(q_{\exp} - q_{f i t})}^{2}}{\sum_{i = 1}^{n} {(q_{\exp} - {\bar{q}}_{\exp})}^{2}}

(9)

A R E = \frac{100}{n} {\sum_{i = 1}^{n} (\frac{q_{f i t} - q_{\exp}}{q_{\exp}})}_{i}

(10)

S S E = \sum_{i = 1}^{n} {(q_{\exp} - q_{f i t})}^{2}

(11)

R M S E = {(\frac{1}{n} \sum_{i = 1}^{n} (q_{\exp} - q_{f i t}))}^{1 / 2}

where n represents the data point number; q_exp represents the experiment value; q_fit represents the fitting value; and ${\bar{q}}_{\exp}$ represents the mean of experiment value.

2.4.2. Affinity for adsorption

Under low pressure region, the uptake for adsorption always displays a linear increase trend with enhancing pressure. In this region, the adsorption isotherm can be matched using Henry’s equation, and the density for adsorbate is very small on adsorbent surface. Hence, in this region, the adsorbate will be preferentially and freely trapped on high-binding energy adsorption center, and the interaction for adsorbate-adsorbate is very weak and can be ignored [19]. Accordingly, it can be inferred that the adsorption process is controlled through adsorbate-adsorbent surface interaction in this region and the Henry’s coefficient (K_H) can be reasonably employed to assess the adsorption affinity. The smaller the Henry’s coefficient, the weaker the interaction for adsorbate-adsorbent surface, and the lower the adsorption affinity.

In this low pressure region, Eq. (12) can represent the relationship between the uptake and equilibrium pressure [20].

(12)

ln (q / P) = A_{0} + A_{1} q + A_{2} q^{2} + \cdot \cdot \cdot

where A₂, A₁ and A₀ represent the virial coefficients. K_H can be obtained through A₀, and K_H=exp(A₀).

In this low pressure condition, high-terms in Eq. (12) will be negligible, and Eq. (13) can replace Eq. (12).

(13)

ln (P / q) = - A_{0} - A_{1} q

According to the isothermal curves for water adsorption, through fitting the linear region for ln(P/q) vs. q, A₀ value will be gained. Then, the value for K_H can be calculated.

2.5. Molecular Simulation

2.5.1. Model construction

2.5.1.1. Clay minerals model

Generally, the clay minerals within reservoir matrix include kaolinite, chlorite, illite and montmorillonite. Due to the differences in physical-chemical characteristics for these clay minerals, kaolinite, chlorite and illite are taken as non-swelling clay components, and montmorillonite is representative for swelling clay component [21]. For better investigating the effect for clay minerals effect on water adsorption behavior in continental reservoir, illite and montmorillonite were selected as research subjects. Therefore, the pore models based on two types of clay minerals of montmorillonite and illite were established. The crystal cell data of montmorillonite and illite comes from the American Mineralogical Crystal Database [22].

The set of potential functions used to describe the interactions between molecular systems are defined as molecule force field. Molecular force field includes uniform valence force field, polymer one force field, clay force field, universal force field, dreding force field and condensed phase optimized molecular potential (COMPASS) force field suitable for atomic simulation research. This simulation study uses the COMPASS force field. COMPASS force field covers most covalent bonding molecular systems, including organic compounds, inorganic materials, and metal materials. COMPASS force field is constructed through a rigorous parameterization process, with parameters derived from advanced first principles calculations, which can effectively solve the compatibility issues of simulation models [23]. The crystal cells of illite and montmorillonite and the various parameters for the selected montmorillonite and illite crystal cells are shown in Supplementary Materials.

Supplementary Materials

Download

The cleavage planes of montmorillonite and illite are both (001), stacked along the Z-axis. Therefore, a model is constructed by cutting along the cleavage plane (001) to match the physical properties of the minerals as closely as possible [18]. Two mineral pore wall models, shown in Figures 1(a,b), can be obtained by expanding the cut unit cell along the x, y, and z directions.

(a) Illite pore wall model, (b) montmorillonite pore wall model, and (c) organic matter pore wall model.

2.5.1.2. Organic matter model

In this study, the kerogen was adopted to represent reservoir organic matter. The I-A type kerogen molecule established by Ungerer et al. [24] was used to construct a reservoir matrix model, with the molecular formula of C₂₅₁H₃₈₅N₇O₁₃S₃. The molecular structure of I-A type kerogen is displayed in Supplementary Materials.

At 298 K, 8 geometrically optimized kerogen molecules were placed in a periodic simulation box. The initial density and cutoff distance are 0.6 g/cm³ and 15.5 Å, respectively. Eward and atomic methods were utilized to determine electrostatic and vdW interactions, respectively. Nosé-Hoover and Berendsen methods were utilized for temperature and pressure controls, respectively. Firstly, 10 annealing cycles were performed with the simulation temperature increasing from 300 K to 800 K and then decreasing to 300 K within each cycle. Each cycle lasted for 1000 ps, with a total simulation time of 10 ns. Then, 800, 600, 400, and 300 K simulations were conducted successively under the isothermal isobaric ensemble at 20 MPa, with each cycle lasting for 2000 ps, to ensure that the relaxation of configuration energy and system density reach equilibrium. Finally, a 1 ns molecular dynamics simulation was conducted using the NVT ensemble at 298 K to ensure density convergence and energy balance, resulting in the organic matter pore wall model as shown in Figure 1(c), with 1.02±0.01 g/cm³ stable density. The density value is consistent with previous research result (0.95-1.05 g/cm³) [25], confirming the rationality and reliability of the constructed organic matter pore wall model.

The narrow slit nanopore models of montmorillonite and illite were constructed using Materials Studio software based on the expanded mineral pore wall model. Organic matter narrow slit nanopore model was assembled using matrix pore wall constructed from kerogen macromolecule. The dimensions of three models at three spatial directions (origin O, corresponding to lengths OA, OB, OC in each direction) are depicted in Figure 2. In order to simulate real geological conditions and physical properties, the models constructed have periodic boundaries at X direction, Y direction and Z direction. The pore size was set to 1.5 nm. Through performing water adsorption simulations on constructed organic matter and clay minerals nanopores, the differences in water adsorption on inorganic minerals and organic matter was analyzed and elucidated.

(a) Organic matter pore model, (b) illite pore model, and (c) montmorillonite pore model.

2.5.2. Simulation details

The Grand Canonical Monte Carlo (GCMC) method was utilized for simulating water adsorption properties in montmorillonite, illite and organic matter (kerogen) at 328 K and 0-25 MPa. Simulations were conducted using Materials Studio software, with a fixed pressure task selected from the Sorption module. The production step and equilibrium step were set to 2×10⁶ and 1×10⁶, respectively. GCMC method was utilized for simulating the distribution of water molecule under different pressures. The relationship between pressure and fugacity was converted by NIST. The set pressure points are 1 MPa, 5 MPa, 10 MPa, 15 MPa, 20 MPa and 25 MPa. The fugacity and fugacity coefficients of water at these pressure points at 328 K are given in Supplementary Materials.

3. Results and Discussion

3.1. Surface morphology information

The obtained SEM images of sample are reported in Figure 3. The nanopore structure of sample shows the typical silt morphology property. Meanwhile, the slit pores have the multi-scale distribution feature. The two-dimensional and three-dimensional rough surface properties of sample through Atomic Force Microscope test is depicted in Supplementary Materials.

(a-f) SEM images of sample (red circles show the typical pore types).

3.2. Pore structure information

3.2.1. Micropore information

The isothermal curve for CO₂ adsorption under the temperature of 273 K is depicted in Figure 4(a). The micropore size distribution for sample is elucidated in Figure 4(b). Apparently, the micropore size distribution is segmented and mainly distributed in 0.45-0.55 nm, 0.56-0.78 nm and 0.79-0.85 nm. Especially at 0.54 nm, 0.62 nm and 0.84 nm, there are lots of micropores. The obtained micropore structure parameter is listed in Table 1.

(a) Isothermal adsorption curve of CO2, (b) micropore size distribution, (c) N2 isothermal desorption and adsorption curves and (d) macropore and mesopore size distribution.

Table 1. Micropore structure parameter.

Sample	V_CO2 (cm³/g)	S_CO2 (m²/g)
Yanchang	0.00105	8.4510

3.2.2. Mesopore and macropore information

The isothermal curves for N₂ adsorption and desorption are plotted in Figure 4(c). There is an obvious hysteresis loop between desorption isotherm and adsorption isotherm. According to the IUPAC classification, the gained hysteresis loop is classified as H3 type [26]. The H3 hysteresis loop indicates that the pores are the plate- or slit-shaped, which is agreeing with the observation result as shown in Figure 3. When P/P₀ is low, the isothermal curve for N₂ adsorption increases quickly, which is caused through the prominent micropore filling effect [19]. Once P/P₀ is close to 1.0, the isothermal curve for N₂ adsorption enhances rapidly without saturation, indicating that the filling in the inner pore system is incomplete.

The macropore and mesopore size distribution is detailed in Figure 4(d). The macropore and mesopore size distribution can be divided into two continuous sections of 1.50-10.00 nm and 10.01-62.0 nm. At 4.2 nm and 12.5 nm, there are lots of mesopores.

The gained macropore and mesopore structure parameter is summarized Table 2. The total pore volume (V_t) is 0.0145 cm³/g. The specific surface area of BET (S_BET) is 8.6604 m²/g. The micropore volume (V_mic) is 0.0013 cm³/g. The mesopore volume (V_mes) is 0.0123 cm³/g. The macropore volume (V_mac) is 0.0009 cm³/g. Evidently, the value of V_mes is greater than those of V_mic and V_mac, illustrating that V_t and S_BET are mainly contributed through mesopore. In addition, V_CO2 value is obviously lower than V_mes value, and the pore volume of sample is mainly contributed through mesopore, agreeing with the finding of Li et al [27,28].

Table 2. Macropore and mesopore structure parameter.

Sample	V_t (cm³/g)	V_mac(cm³/g)	V_mes (cm³/g)	V_mic(cm³/g)	S_BET (m²/g)
Yanchang	0.0145	0.0009	0.0123	0.0013	8.6604

3.3. Surface chemical properties

The obtained FTIR spectra of sample are described in Figure 5(a). These spectra are divided into four main parts of specifically surface functional groups of 3600-3000 cm^-1 (hydroxyl functional groups), 3000-2800 cm^-1 (aliphatic functional groups), 1800-1000 cm^-1 (oxygen-containing functional groups) and 910-675 cm^-1 (aromatic functional groups). Figures 5(b-e) display the four FTIR peak fitting curves, and Table 3 lists the detailed fitting results. The accuracy of all fitting results is higher than 99.99%, illustrating the excellent fitting effect.

(a) FTIR spectra of sample and (b-e) FTIR spectra fitting curve.

Table 3. Results of functional group peak fitting.

Wave range (cm^-1)	Pisition (cm^-1)	Height	Peak area	Type	Total peak area
3600-3000	3605.05516	0.03812	4.41912	Hydrogen bonding between molecules	43.4200
	3492.37389	0.09644	16.57163	Phenol/alcohol/carboxylic acid –OH or hydrogen bonding between molecules
	3396.61444	0.05211	6.22426
	3278.2629	0.05959	16.20494	Hydrogen bonds formed by amino groups
3000-2800	2989.5955	0.00871	0.7141	Methyl asymmetric stretching vibration	1.0405
	2922.52104	0.00238	0.05772	Asymmetric stretching vibration of methylene groups
	2874.71475	0.00305	0.23545	Asymmetric stretching vibration of methylene groups
	2852.25501	0.00156	0.03325	Methylene symmetric stretching vibration
1800-1000	1637.67662	0.0447	5.34428	Aldehydes, ketones, esters, quinones C=O stretching vibration	61.41707
	1624.20224	0.02663	1.2151	C=C stretching vibration of aromatic hydrocarbons
	1443.7788	0.07397	13.97518	Hydroxyl (-OH) in-plane bending vibration
	1161.3633	0.06134	17.96992	Phenol, alcohol, ether, ester carbon oxygen bond
	1091.73992	0.03726	1.70968	Phenol, alcohol, ether, ester carbon oxygen bond
	1067.7012	0.11267	21.20291	Phenol, alcohol, ether, ester carbon oxygen bond
910-675	875.53694	0.04245	2.2696	Out of plane deformation vibration of CH in a benzene ring replaced by a single H	9.54007
	800.38456	0.04088	0.68971	Out of plane deformation vibration of CH in benzene rings with three adjacent H substituted
	780.90239	0.08957	5.06007
	705.75664	0.02449	1.52069	Out of plane deformation vibration of CH in five adjacent H substituted benzene rings

It is manifest that the adsorption peak areas are controlled through oxygenated functional groups, followed by hydroxyl, aromatic and aliphatic functional groups in sequence. The oxygen-containing functional groups within the spectra range for 1800-1000 cm^-1 mainly contains aldehydes, ketones, esters, C=C stretching vibration for aromatic hydrocarbons, quinones C=O stretching vibration, hydroxyl (-OH) in-plane bending vibration, and phenol, ether, alcohol, ester carbon oxygen bond. The hydroxyl functional groups within the spectra range for 3600-3000 cm^-1 mainly contains hydrogen bonding between molecules, phenol/alcohol/carboxylic acid –OH and hydrogen bonds formed by amino groups. The aromatic functional groups within the spectra range for 910-675 cm^-1 mainly contains many out of plane deformation vibrations as listed in Table 3. The aliphatic functional groups within the spectra range for 3000-2800 cm^-1 mainly contains methyl asymmetric stretching vibration, asymmetric stretching vibration for methylene groups and methylene symmetric stretching vibration.

3.4. Adsorption performance for water on sample

3.4.1. Isothermal desorption and adsorption curves

The measured isothermal desorption and adsorption curves on sample for water vapor as a function of P/P₀ are illustrated in Figure 6(a). Noticeably, with enhancing P/P₀, water vapor adsorption quantity continuously improves. Meanwhile, as given in Figure 6(a), water vapor adsorption quantity improves with the increase in temperature. When P/P₀ is constant, the higher the temperature, the larger the pressure P. Although improving temperature decreases the adsorption capacity, increasing pressure promotes the adsorption. Thereby, the combined effect gives rise to the positive correlation between the obtained adsorption quantity and experiment temperature. By comparing the adsorption capacity of water in continental reservoir in this study and the adsorption capacity of water in marine reservoir in previous study [10], it can be found that the adsorption amount of water in this study in slightly higher than that in previous study, which is due to the higher clay minerals content of adopted sample.

Adsorption and desorption isotherms of water vapor as a function of relative vapor pressure (a) and vapor pressure (b).

For water vapor, the isothermal desorption and adsorption curves as a function of P are displayed in Figure 6(b). A noticeable hysteresis loop between desorption and adsorption isotherms exists, and the smaller hysteresis loop corresponds to the lower temperature. Gregg and Sing [29] pointed out that the formation of hysteresis loop is due to the expansion or adsorption in the micropore. The bigger hysteresis loop manifests that the swelling effect at high temperature is more remarkable. Meanwhile, as given in Figure 6(b), increasing temperature will decrease the uptake for adsorption. When temperature is improved, the molecule motion is accelerated, and the adsorbate-matrix surface interaction is weakend [13]. As a result, with the increase in temperature, the adsorbate molecule is more unstable and the possibility of desorption of adsorbed molecule is greater, leading to the decrease in adsorption quantity.

The isothermal adsorption curves for water vapor are all type II isotherm. According to the findings of Duan et al. [30], the isothermal curves for CH₄ and CO₂ adsorption both are type I isotherm. The different types of isotherms illustrates that the mechanisms for CO₂ and CH₄ adsorption are different from that for water vapor adsorption. For water vapor, the whole adsorption process does not belong to monolayer adsorption. When pressure approaches saturated vapor pressure, the unrestricted monolayer-multilayer adsorption brings about the type II isotherm [26]. The rapid adsorption in the low pressure region for water vapor corresponds to the monolayer adsorption. The first inflection point for the isotherm indicates monolayer adsorption completion for water vapor. The gentle part in the middle of water vapor isotherm denotes the beginning of multilayer adsorption. The quick growth of adsorption quantity in the later stage is because of the infinite increase in thickness of multilayer adsorption.

The fitting results for gained adsorption isotherms adopting Dent model, Modified Dent model and D’Arcy and Watt model are illustrated in Figure 7. The obtained fitting parameters for Dent model, Modified model and D’Arcy and Watt model are given in Table 4.

Fitting of adsorption isotherms of water vapor using different models.

Table 4. Matching parameters for dent model, modified dent model and D’Arcy and watt model.

Dent model	Model coefficients				Goodness of fit indexes
Dent model	q₀ (mmol/g)		k₁ (kPa^-1)	k₂ (kPa^-1)	R²	ARE	SSE	RMSE
298 K	0.2097		5.2965	0.3353	0.9975	2.3900	0.0013	0.0828
313 K	0.2050		3.3547	0.1966	0.9954	3.9027	0.0033	0.1057
328 K	0.2035		2.1095	0.1157	0.9918	5.5033	0.0064	0.1256
Modified dent model	Model coefficients				Goodness of fit indexes
Modified dent model	q₀ (mmol/g)	k₁ (kPa^-1)	α	k₂ (kPa^-1)	R²	ARE	SSE	RMSE
298 K	0.2349	6.8370	1.3096	0.2532	0.9984	2.4127	0.0008	0.0740
313 K	0.2473	3.4653	1.4885	0.0954	0.9976	3.2521	0.0016	0.0881
328 K	0.2609	1.4621	1.6927	0.0287	0.9962	4.4181	0.0028	0.1016
D’Arcy and Watt model	Model coefficients				Goodness of fit indexes
D’Arcy and Watt model	S₀ (mmol/g)	K (kPa^-1)	s (mmol/g)	k (kPa^-1)	R²	ARE	SSE	RMSE
298 K	0.3177	2.2870	0.1099	0.3812	0.9995	1.4703	2.2288*10^-4	0.0522
313 K	0.3579	1.0760	0.0977	0.2217	0.9996	1.5797	2.5597*10^-4	0.0543
328 K	0.4140	0.5145	0.0780	0.1329	0.9997	1.4787	2.2287*10^-4	0.0497

For D’Arcy and Watt model, R² value is the biggest, and the values for ARE, RMSE and SSE is the smallest. According to the obtained results of R², ARE, SSE and RMSE, it can be discovered that D’Arcy and Watt model is the most suitable. Moreover, the residual plots for three models are described in Supplementary Materials, which also illustrates that the fitting effect is the best for D’Arcy and Watt model.

3.4.2. Behaviors for primary and secondary adsorption

Figure 8(a) displays the isothermal adsorption curves on primary and secondary binding sites for water vapor. The adsorption isotherms on primary binding sites differ from those on secondary binding sites, exhibiting that the modes for two adsorptions are different.

Adsorption quantity for water vapor as a function of vapor pressure on primary binding sites and secondary binding sites (a) and three binding sites (b).

At the initial stage, there is a rapid increase in adsorption quantity on primary binding sites, while there is a very slow increase for the uptake on secondary binding sites. When the pressure is relatively low, water molecule number is small and the interaction for water molecule-water molecule is very weak. Hence, at the initial phase, the adsorption process is governed through adsorbate-adsorbent interaction force, and water molecule is preferentially captured at high-binding energy centers. Chen et al. [5] have demonstrated that the initially injected water molecule will be captured in specific centers rather than be indiscriminately retained on the whole matrix surface. Meanwhile, as reported in the published literature, the functional groups on matrix surface, especially oxygen-containing functional groups, possess the higher affinity and are served as bindings sites for primary adsorption [31]. According to the FTIR measurement results in Section 4.3, abundant functional groups on matrix surface containing oxygenated functional groups are discovered. –OH and –COOH groups are taken as H-bond acceptors or donors. The only oxygenated groups can capture injected water molecules through H-bond acceptor interaction. It is clear that sample possesses the surface chemical properties for forming primary binding sites. In the initial phase, water adsorption process is dominated through primary adsorption.

As the pressure further rises, the rate for primary adsorption is becoming slower. On the contrary, the uptake of secondary adsorption increases rapidly, and there is an increasing adsorption rate on secondary binding sites. Evidently, with improving pressure, the effect of primary adsorption is quickly weakened, while the influence of secondary adsorption is significantly enhanced. The gradually weakening primary adsorption process is resulted from the heterogeneity of physical and chemical properties of matrix surface. Through studying the contact angle at matrix surface for water molecule, it is discovered that three types for binding centers, including strongly hydrophobic binding centers, weakly hydrophobic binding centers and hydrophilic binding centers, exist on matrix surface [32]. With improving pressure, the favorable primary binding sites are rapidly occupied, and the water molecule can only be adsorbed at weaker primary binding centers, bringing about the less obvious primary adsorption. Furthermore, Bahadur et al. [33] pointed out that if primary binding sites are occupied, the preferentially adsorbed water molecule can be served as the nucleation center for subsequent adsorption or as the secondary binding site. Meanwhile, the formation of water cluster has been found on pre-adsorbed water molecule [34]. Thereby, with improving pressure, more clusters appear around primary binding centers based on hydrogen bonding, leading to the increasingly significant secondary adsorption.

At the high pressure stage, the primary adsorption quantity does not enhance, and secondary adsorption uptake continues to enhance. Accordingly, at the high pressure stage, adsorption process is completely determined through secondary adsorption. As pressure improves, many smaller clusters aggregate with surrounding clusters, forming more ordered three-dimensional cluster system. Meanwhile, with pressure improves, the captured water molecule through cluster structure and multilayer adsorption is able to be retained in inner pore space by capillary condensation phenomenon. The emergence of capillary condensation phenomenon results in the rapid improvement for adsorption quantity on secondary binding sites at the high pressure condition. However, when these large-sized water cluster structures are appeared on the pore throat, the entry or exit of other molecules will be greatly restricted. Through studying the influence of cluster system on the connectivity of pore structure, Brennan et al. [35] reported that the distribution for pore size will be narrowed and the bigger pore structure will be converted into the smaller pore structure through the gradually growing cluster network, resulting in the decrease in pore connectivity. The reduced pore connectivity will have adverse effect on fluid transport process. Hence, the effect of water on fluid diffusion and transport is mainly owing to the formation of large water cluster network during secondary adsorption process. When extracting natural gas resource or implementing CO₂ geological sequestration, secondary adsorption requires more attention.

As elucidated in Figure 8(a), at the initial stage, secondary adsorption uptake is lower. However, as pressure continues to increase, secondary adsorption uptake begins to exceed the uptake of primary adsorption. Ultimately, primary adsorption uptake is manifestly lower than secondary adsorption uptake. The functional groups on matrix surface mainly affect primary adsorption, and secondary adsorption mainly occurs through the formation of cluster network within the inner pore space. The significantly higher secondary adsorption quantity indicates that inner pore structure possesses an obviously stronger effect on water adsorption than surface functional groups in high pressure condition. However, for water adsorption in marine reservoir, primary adsorption capacity is always greater than secondary adsorption capacity, illuminating that surface functional groups possess a more evident influence on water adsorption in marine reservoir than pore structure [10].

Figure 8(b) compares the isotherms for three types for adsorption centers. The primary adsorption isothermal curve is type I isotherm. Generally, Langmuir model can match the type I isotherm well. For the derivation process of Langmuir model, the important assumptions are to ignore the interaction between adsorbate molecules. Thereby, it is evident that the interaction of adsorbed water molecule on primary binding sites with surrounding adsorbed water molecule is very weak. Considering that functional groups on matrix surface dominate the process for primary adsorption, it is concluded that functional groups at matrix surface is very dispersed. Meanwhile, type III isotherm corresponds to the isothermal curve for secondary adsorption. The weaker interaction between adsorbate-adsorbent and the formation of cluster lead to type III isotherm on secondary binding sites.

When injected water molecule is adsorbed on naturally occurring carbon-containing material, inner pore structure, functional groups on matrix surface and pressure determine isotherm shape [34]. In the low pressure phase, primary adsorption predominates, and the isotherms for total adsorption are close to those for primary adsorption. With improving pressure, the gradually slow primary adsorption process and the gradually accelerated secondary adsorption process give rise to the straight section in the middle of total adsorption isotherm. At high pressure, the sharp improvement for total adsorption is mainly provided through secondary adsorption, and consequently the total isothermal adsorption curve is close to isothermal curve for secondary adsorption.

3.4.3. Affinity for water adsorption

The result for Henry’s coefficient (K_H) at three types of adsorption sites is listed in Table 5. With improving temperature, K_H values on three types of adsorption sites all reduce, indicating that the adsorption affinity is weakened under high temperature condition. Accordingly, the adsorbed water molecule is more likely to escape from the adsorption sites at high temperature, causing the decrease in adsorption amount.

Table 5. Henry’s coefficient for water vapor.

Fluid	T ( K)	K_H (mmol/g/kPa)
Fluid	T ( K)	Primary binding sites	Secondary binding sites	Total binding sites
Water vapor	298	0.7924	0.0420	1.1753
	313	0.4108	0.0217	0.5861
	328	0.2231	0.0104	0.3203

As presented in Table 5, total binding sites has the biggest K_H value, while secondary binding sites possesses the lowest K_H value. The K_H value of primary binding sites is significantly greater than that of secondary binding sites. Hence, primary binding sites have stronger affinity under low pressure condition and primary binding sites can preferentially capture water molecule in the initial stage for adsorption process.

3.5. Adsorption performance for water on organic matter and clay minerals

3.5.1. Density distribution of water on organic matter and clay minerals

The aggregation state for water molecule on montmorillonite, illite and organic matter simulated using GCMC method at 20 MPa is shown in Figure 9. It can be observed that, for the narrow slit nanopore of organic matter, there is a small amount of water molecule adsorbed near the pore wall, and almost no water molecule appears at the center of pore. On the contrary, for clay minerals nanopore, there is a large amount of water molecule adsorbed, and water molecules significantly aggregate at the surface of pore wall, forming the water molecule layer. Meanwhile, water molecule is mainly distributed at both sides of the narrow slit nanopore of clay minerals, and there is also less distribution at the center of nanopore. The closer to the pore wall, the stronger the adsorption affinity, which is more conducive to water molecule adsorption. By analyzing the aggregation state for water molecule, it can be concluded that there are significant differences between the adsorption processes in clay minerals and organic matter for water molecule. The uptake of clay minerals for water molecule is much larger than that of organic matter.

The aggregation state of water molecule on (a) organic matter, (b) illite and (c) montmorillonite.

Based on the relative concentration distribution of water molecule, the adsorption density distribution curves of water molecule in organic matter, illite and montmorillonite were obtained, as shown in Figure 10. The gray area in Figure 10(a) is the matrix part of organic matter, the brown area in Figures 10(b,c) is the matrix part of two clay minerals, and the light white area is the narrow slit nanopore part of organic matter and two clay minerals. It can be seen that, under different pressures, there is no significant change in the adsorption density of water molecule on organic matter and two clay minerals. This is because during the adsorption process of water molecule, hydrogen bonding forces are formed. Compared with van der Waals forces and electrostatic forces, hydrogen bonding forces are stronger and are not greatly affected by pressure. By analyzing the density distribution curves, it can be also found that, the adsorption capacity of organic matter for water molecule is very weak, and the density distribution in organic matter is apparently different from those in clay minerals. For organic matter, water molecule is mainly adsorbed in the area near the wall of pore surface, and no adsorption occurs in the central area for narrow slit nanopore. For the clay minerals of illite and montmorillonite, in addition to the two sides of the pore wall, the adsorption process also appears in the central area of narrow slit nanopore. The adsorption density for water molecule in montmorillonite is the highest in the area near the pore wall, with a value of 1360 kg/m³, and is the lowest in the central area of the narrow nanopore, with a value of 530 kg/m³. The adsorption density for water molecule in illite is the highest in the area near the pore wall, with a value of 1340 kg/m³, and is the lowest in the central area of the narrow slit nanopore, with a value of 510 kg/m³. The adsorption density for water molecule in organic matter is the highest in the area near the pore wall, with a value of 40 kg/m³. Because there is no water molecule retained in the middle of the organic matter nanopore, the lowest adsorption density in organic matter is 0 kg/m³ in the middle of nanopore. Evidently, the adsorption density for water molecule in organic matter is significantly lower than those in clay minerals. Under the same condition, the adsorption density for water molecule is as follows: montmorillonite > illite > organic matter.

Density curve for water molecule adsorption in (a) organic matter, (b) illite and (c) montmorillonite.

The density field of water molecule adsorbed on organic matter and clay minerals at 1.5 nm pore and 10 MPa pressure is shown in Figure 11. The red dots in density field represent water molecule density state distribution, that is, the probability of water molecule appearing at specific position. The number of red dots in organic matter is significantly smaller than those in clay minerals, indicating that clay minerals have a significantly stronger water absorption capacity than organic matter, consistent with the phenomenon in the adsorption density distribution. Meanwhile, as depicted in Figures 11(b,c), the number of red dots in montmorillonite is higher than that in illite, which reveals that the probability density distribution of water molecule in montmorillonite is larger and the adsorption capacity of montmorillonite is greater for water molecule. The probability density distribution of water molecule is also as follows: montmorillonite > illite> organic matter.

Probability density distribution of water molecule on (a) organic matter, (b) illite and (c) montmorillonite.

3.5.2. Adsorption heat of water on organic matter and clay minerals

During the molecular simulation process, the result of average adsorption density was adopted to represent adsorption isotherm of water molecule. The obtained isotherms on organic matter and clay minerals for water molecule adsorption are described in Figure 12. It is apparent that the adsorption isotherms on single organic matter and clay minerals belong to type I isotherm. For the type I isotherm, many adsorption models are suitable for the matching process. In this study, Langmuir model and Freundlich model were adopted to match the adsorption isotherms for water molecule on organic matter and clay minerals. Eq. (14) and Eq. (15) give the forms of Langmuir model and Freundlich model, respectively. The fitting results of two adsorption models are depicted in Figure 12. The obtained matching parameters for Langmuir model and Freundlich model are listed in Table 6.

(14)

q = \frac{q_{m} b P}{1 + b P}

(15)

q = a P^{1 / n}

Adsorption isotherms of water molecule on organic matter, illite and montmorillonite.

Table 6. Matching parameters for Langmuir model and Freundlich model.

Model	Adsorbent	Model coefficients		Goodness of fit indexes
Model	Adsorbent	q₀ (mmol/cm³)	b (MPa-1)	R²	ARE	SSE	RMSE
Langmuir	Kerogen	1.4999	28.1669	0.9998	0.4407	0.0004	0.0811
	Illite	71.0472	116.1993	0.9999	0.0882	0.0333	0.2500
	Montmorillonite	72.6574	127.0984	0.9999	0.2212	0.2713	0.4006
Model	Adsorbent	Model coefficients		Goodness of fit indexes
Model	Adsorbent	a (mmol/cm³/MPa)	n	R²	ARE	SSE	RMSE
Freundlich	Kerogen	1.4511	87.0783	0.9999	0.3128	0.0002	0.0683
	Illite	70.5589	444.4362	0.9999	0.0507	0.0087	0.1897
	Montmorillonite	72.2031	489.6172	0.9999	0.2032	0.2423	0.3884

where q_m represents the single layer saturated adsorption capacity; b represents the Langmuir coefficient; and a and n both are the constants.

From Table 6, it can be found that when adopting Langmuir model and Freundlich model, the values of R² are basically the same. However, the values of ARE, SSE and RMSE when using Freundlich model are all smaller than those when applying Langmuir model, which reveals that Freundlich model has the better fitting effect. Therefore, the obtained adsorption isotherms of water molecule on organic matter and clay minerals were dealt with Freundlich model.

Compared to previous study about water adsorption on organic matter models with the pore sizes of 1 nm and 3 nm [15], it can be found that the adsorption amount of water in 1.5 nm organic matter nanopore in this study and the adsorption amount of water in previous study are in an order of magnitude, and the adsorption amount of water in 1.5 nm nanopore is lower than that in 1 nm nanopore and larger than that in 3 nm nanopore. This comparison result not only verifies the reliability of the constructed model and the simulated results, but also indicates that reducing pore size is beneficial for water adsorption.

Meanwhile, through comparing the shapes of adsorption isotherms of water molecule on continental reservoir and single organic matter and clay minerals, it is found that the adsorption shapes of these isotherms are different. The adsorption isotherm on continental reservoir possesses the shape of type II isotherm, while the adsorption isotherms on single organic matter, illite and montmorillonite possess the shape of type I isotherm. This phenomenon reveals that the composition of continental reservoir is extremely complex and the composition of continental reservoir is not the simple superposition of clay minerals components and organic matter. The adsorption behavior for water on continental reservoir needs consider complex factors rather than just the components of single clay minerals and organic matter.

In addition, by the comparison of adsorption ability of water molecule on continental reservoir and single organic matter and clay minerals, it is discovered that the adsorption capacity for water molecule on continental reservoir is obviously lower than those on clay minerals, and the adsorption quantity for water molecule on continental reservoir is slightly lower than that on organic matter. Thereby, although continental reservoir is rich in clay minerals, the ability for continental reservoir for water adsorption is still significantly smaller than that for clay minerals. This finding suggests that the combination and superposition of many components can block clay minerals pore system, reduce clay minerals expansion, and consequently weaken clay minerals adsorption ability within reservoir.

The gained adsorption heat in organic matter and clay minerals for water molecule is plotted in Figure 13. There are significant differences in the adsorption heat in three components of organic matter, illie and montmorillonite. The adsorption heat is the highest in montmorillonite, while the adsorption heat in organic matter is the lowest. The higher adsorption heat illustrates the stronger interaction for adsorbate-adsorbent [20]. Hence, the adsorption affinity in organic matter for water molecule is the lowest, and the interaction for water molecule-organic matter is the weakest. Accordingly, the water molecule adsorbed on organic matter is more easily desorbed, and it can be inferred that the reservoir with great organic matter content is not conducive to water retention and is more suitable for the application of hydraulic fracturing technology. In contrary, clay minerals have the stronger interaction with water molecule, especially the swelling clay mineral of montmorillonite. For reservoir with high content of clay minerals, the application of hydraulic fracturing technology needs to be cautious. Moreover, there is no significant change in adsorption heat under 1 MPa and subsequent pressure conditions, further demonstrating that the adsorption of water molecule in the three components is insensitive to pressure change.

Adsorption heat of water molecule on organic matter and clay minerals.

3.5.3. Energy distribution of water on organic matter and clay minerals

Figure 14 presents the probability distribution of adsorption potential energy for water molecule on slit pore models of organic matter, illite and montmorillonite, where x-axis represents energy and y-axis represents Poisson distribution of potential energy. The potential energy has the negative value, which illuminates that water molecule adsorption in the narrow slit pores of organic matter, illite and montmorillonite is a spontaneous process. The bigger the potential energy, the more unstable the adsorbates, and the more likely they is to escape [20]. For organic matter, illite and montmorillonite, the potential energy distributions of water molecule at two pressures of 20 MPa and 25 MPa basically overlop. The change in pressure does not alter the distribution of adsorption potential energy, suggesting the stability of organic matter and clay minerals towards pressure during water adsorption process.

Poisson distribution of energy of water molecule on (a) organic matter, (b) illite and (c) montmorillonite.

As displayed in Figure 14(a), there are three distribution peaks for adsorption potential energy of water molecule on organic matter. For organic matter, the energy corresponding to the adsorption peak on the far left is -20.02 KJ/mol, and the energy corresponding to the adsorption peak on the far right is -8.25 KJ/mol. For illite, there are also three distribution peaks for the adsorption potential energy. The energy on illite corresponding to the adsorption peak on the far left is -74.15 KJ/mol, and the energy on illite corresponding to the adsorption peak on the far right is -18.05 KJ/mol. Unlike organic matter and illite, the adsorption potential energy on montmorillonite has four distribution peaks, with the leftmost adsorption peak corresponding to an energy of -77.65 KJ/mol and the rightmost adsorption peak corresponding to an energy of -18.45 KJ/mol. The reason for the occurrence of multimodal distribution of adsorption potential energy for water molecule is that on the uncontaminated matrix surface, water molecules preferentially bind to surface functional groups, forming a set of water molecule adsorption sites that provide binding sites for the next set of water molecules. From organic matter to illite and then to montmorillonite, the adsorption peaks on the leftmost and rightmost sides of the adsorption potential energy both shift to the left, and there is a decrease in adsorption potential energy, revealing that the adsorption stability is gradually increasing. Therefore, the adsorption stability in three components for water molecule is as follows: organic matter < illite < montmorillonite. Meanwhile, the adsorption potential energies on illite and montmorillonite are significantly more negative than that on organic matter, which also reflects that water molecule adsorption on organic matter is the most unstable.

The more negative adsorption potential energy and the higher adsorption heat not only favor the adsorption process but also facilitate the formation of efficient packing [20]. Due to the more negative adsorption potential energy and the higher adsorption heat for water molecule on clay minerals, it can be inferred that adsorbed water molecule may form a more stable rearrangement on clay surface. The stable and efficient packing restricts the mobility for adsorbed molecules, suggesting that the adsorbed water molecule on clay surface may be more difficult to desorb. Moreover, clusters are easily formed between water molecules, and the size of clusters can continue to enlarge. Based on the less negative adsorption potential energy and the smaller adsorption heat, the adsorbed water molecules may not form effective stacking and stable cluster systems on the surface of organic matter, resulting in greater fluidity of the adsorbed molecule and the limited impact of adsorbed water molecule on pore connectivity of organic matter. By contrast, on clay minerals surface, water molecule is firmly adsorbed, facilitating the formation of structured and three-dimensional cluster network with subsequently injected water molecule. Hence, in the clay minerals, the adsorbed water molecule is likely to have adverse effect on the pore structure, which worsens the pore connectivity and affects the production behavior of shale gas.

When fracturing fluid is injected into deep reservoirs, water molecules mainly exist in liquid form. But at higher temperatures, a small amount of water molecules will exist in water vapor form. Therefore, it is necessary to conduct adsorption experiments on water vapor and molecular simulation studies on water. The results of joint experimental and molecular simulation studies indicate that the adsorption of water in deep reservoirs is influenced by both surface functional groups and pore structure.

Apparently, through this study, it can be found that deep reservoirs with higher clay minerals content are more likely to adsorb water molecules. These adsorbed water molecules can reduce the pore connectivity and affect gas production behavior. Hence, water-based hydraulic fracturing technology is not very suitable for reservoirs with high clay minerals content. Meanwhile, it should be noted that during the molecular simulation process, the constructed model indeed has some simplifications, such as the use of homogeneous surface, single pore size and ideal slit pore and the neglect of multi-mineral interface. These simplifications of constructed model have brought certain limitations to this study. In future study, more factors will be considered to better reveal the adsorption mechanism of water in deep gas-bearing reservoirs.

4. Conclusions

In this research, the isothermal behaviors for water molecule adsorption were investigated by gravimetric adsorption experiment. The method of molecular simulation was adopted to analyze the performances for water molecule adsorption in three individual components of organic matter, illite and montmorillonite. The gained main findings are as follows:

The surface functional groups and pore structure both controls the water molecule adsorption process. The surface functional groups mainly affect primary adsorption, and pore structure mainly affects secondary adsorption.

Under small pressure, primary adsorption predominates. Under high pressure region, the adsorption process is mainly contributed by secondary adsorption. The influence of surface functional groups on the uptake for water molecule is lower than that of inner pore structure.

Water molecule adsorption on organic matter only appears in the area near the two sides of pore wall, while water molecule adsorption occurs both in the central area and the two sides of nanopore of clay minerals.

The modes for water adsorption on gas-bearing reservoir and individual components are different. Studying the characteristics for water adsorption needs consider complex factors rather than just the components of clay minerals and organic matter.

The uptake for water molecule on organic matter is significantly lower than those on clay minerals. Water adsorption ability is mainly contributed through clay minerals.

Within clay minerals, the water molecule adsorbed is more difficult to desorb. The impact for water molecules adsorbed in clay minerals on pore connectivity and gas production behavior is more pronounced.

Acknowledgment

This work was financially supported by the Guizhou Provincial Basic Research Program (Natural Science) (QKHJC ZK[20 2 3] General 199), the Guizhou Provincial Science and Technology Support Plan Project (Qiankehe Support [2024] General 025), the Guizhou Provincial Science and Technology Program (QKHFQ [20 2 4] 006), Yunnan Fundamental Research Projects (Grant No. 202401AT070406) and the Open Fund of Key Laboratory of Petroleum Resources Exploration and Evaluation, Gansu Province (Grant No. KLPREEGS-2024-04).

CRediT authorship contribution statement

Xianwei Heng: Writing-original draft. Xijian Li: Supervision. Jinlei Fu: Funding acquisition. Xidong Du: Conceptualization. Tengfei Wu: Resources.

Declaration of competing interest

There are no conflicts of interest.

Data availability

The data applied to support the results in this study are available from the corresponding author upon request.

Declaration of generative AI and AI-assisted technologies in the writing process

The authors confirm that there was no use of artificial intelligence (AI)-assisted technology for assisting in the writing or editing of the manuscript and no images were manipulated using AI.

Supplementary data

Supplementary material to this article can be found online at https://dx.doi.org/10.25259/AJC_1261_2026.

References

Shen W., Li X., Ma T., Cai J., Lu X., Zhou S.. High-pressure methane adsorption behavior on deep shales: Experiments and modeling. Physics of Fluids. 2021;33:063103. https://doi.org/10.1063/5.0054486
[Google Scholar]
Rezlerova E., Brennan J.K., Lisal M.. Methane and carbon dioxide in dual-porosity organic matter: Molecular simulations of adsorption and diffusion. AIChE Journal. 2021;67:e16655. https://doi.org/10.1002/aic.16655
[Google Scholar]
Ghanbari E., Dehghanpour H.. The fate of fracturing water: A field and simulation study. Fuel. 2016;163:282-294. https://doi.org/10.1016/j.fuel.2015.09.040
[Google Scholar]
Tang X., Ripepi N., Valentine K.A., Keles C., Long T., Gonciaruk A.. Water vapor sorption on marcellus shale: Measurement, modeling and thermodynamic analysis. Fuel. 2017;209:606-614. https://doi.org/10.1016/j.fuel.2017.07.062
[Google Scholar]
Chen J., Gai H., Xiao Q.. Effects of composition and temperature on water sorption in overmature Wufeng-Longmaxi shales. International Journal of Coal Geology. 2021;234:103673. https://doi.org/10.1016/j.coal.2020.103673
[Google Scholar]
Liu F.Y., Ellett K., Xiao Y.T., Rupp J.A.. Assessing the feasibility of CO2 storage in the new albany shale (devonian-mississippian) with potential enhanced gas recovery using reservoir simulation. International Journal of Greenhouse Gas Control. 2013;17:111-126. https://doi.org/10.1016/j.ijggc.2013.04.018
[Google Scholar]
Feng D., Li X., Wang X., Li J., Sun F., Sun Z., Zhang T., Li P., Chen Y., Zhang X.. Water adsorption and its impact on the pore structure characteristics of shale clay. Applied Clay Science. 2018;155:126-138. https://doi.org/10.1016/j.clay.2018.01.017.
[Google Scholar]
Guan Q.Z., Dong D.Z., Wang S.F., Huang J.L., Zhang C.C.. Analyses on differences of microstructure between marine and lacustrine facies shale reservoirs. Natural Gas Geoscience. 2016;27:524-531. https://doi.org/10.11764/j.issn.1672-1926.2016.03.0524
[Google Scholar]
Guo X.S., Hu D.F., Liu R.B., Wei X.F., Wei F.B.. Geological conditions and exploration potential of Permian marine-continent transitional facies shale gas in the Sichuan Basin. Natural Gas Industry. 2018;38:11-18. https://doi.org/10.3787/j.issn.1000-0976.2018.10.002.
[Google Scholar]
Duan S., Li G.. Equilibrium and kinetics of water vapor adsorption on shale. Journal of Energy Resources Technology. 2018;140:122001-1-122001-10. https://doi.org/10.1115/1.4040530
[Google Scholar]
Sang G., Liu S., Elsworth D.. Water vapor sorption properties of illinois shales under dynamic water vapor conditions: Experimentation and modeling. Water Resources Research. 2019;55:7212-7228. https://doi.org/10.1029/2019wr024992.
[Google Scholar]
Sang G., Liu S., Elsworth D., Yang Y., Fan L.. Evaluation and modeling of water vapor sorption and transport in nanoporous shale. International Journal of Coal Geology. 2020;228:103553. https://doi.org/10.1016/j.coal.2020.103553.
[Google Scholar]
Wang T., Tian S., Li G., Sheng M., Ren W., Liu Q., Tan Y., Zhang P.. Experimental study of water vapor adsorption behaviors on shale. Fuel. 2019;248:168-177. https://doi.org/10.1016/j.fuel.2019.03.029.
[Google Scholar]
Zolfaghari A., Dehghanpour H., Holyk J.. Water sorption behaviour of gas shales: I. Role of clays. International Journal of Coal Geology. 2017;179:130-138. https://doi.org/10.1016/j.coal.2017.05.008.
[Google Scholar]
Li L., Du X., Zhou J., Wang G., Zhang D., Duan M., Yin H., Heng X., Fu J., Wang H.. Water adsorption and transport in shale nanopore: The impact of three-dimensional surface roughness. Physics of Fluids. 2025;37:082131. https://doi.org/10.1063/5.0282876.
[Google Scholar]
Zhang C., Yao Y., Swennen R., Wu H.. Combined effects of the chemical structure and nanopore development on water vapor/liquid adsorption in shale kerogen. Colloids and Surfaces A: Physicochemical and Engineering Aspects. 2022;653:129920. https://doi.org/10.1016/j.colsurfa.2022.129920.
[Google Scholar]
Zhang W., Mo F., Qi Z., Huang X., Yue P., Liu Q., Wan Q., Fu W.. Imbibition mechanism of water in illite nanopores of deep shales by molecular simulation. Journal of Molecular Liquids. 2024;414:126232. https://doi.org/10.1016/j.molliq.2024.126232.
[Google Scholar]
Liu J., Wen Y., Jiang L., Yuan H., Liu Z., Shui Z., Wang B., Yan X.. Molecular dynamics simulation of micro mechanisms of gas and water migration and diffusion in shale reservoirs. Colloids and Surfaces A: Physicochemical and Engineering Aspects. 2024;682:132865. https://doi.org/10.1016/j.colsurfa.2023.132865.
[Google Scholar]
Du X., Cheng Y., Liu Z., Yin H., Wu T., Huo L., Shu C.. CO2 and CH4 adsorption on different rank coals: A thermodynamics study of surface potential, gibbs free energy change and entropy loss. Fuel. 2021;283:118886. https://doi.org/10.1016/j.fuel.2020.118886.
[Google Scholar]
Du X., Pang D., Cheng Y., Zhao Y., Hou Z., Liu Z., Wu T., Shu C.. Adsorption of CH₄, N₂, CO₂, and their mixture on montmorillonite with implications for enhanced hydrocarbon extraction by gas injection. Applied Clay Science. 2021;210:106160. https://doi.org/10.1016/j.clay.2021.106160
[Google Scholar]
Hatch C.D., Wiese J.S., Crane C.C., Harris K.J., Kloss H.G., Baltrusaitis J.. Water adsorption on clay minerals as a function of relative humidity: Application of BET and freundlich adsorption models. Langmuir. 2012;28:1790-1803. https://doi.org/10.1021/la2042873
[Google Scholar]
Shi K.Y., Chen J.Q., Pang X.Q., Jiang F.J., Hui S.S., Zhao Z.C., Chen D., Cong Q., Wang T., Xiao H.Y., Yang X.B., Wang Y.Y.. Wettability of different clay mineral surfaces in shale: Implications from molecular dynamics simulations. Petroleum Science. 2023;20:689-704. https://doi.org/10.1016/j.petsci.2023.02.001.
[Google Scholar]
Li W., Pang X., Snape C., Zhang B., Zheng D., Zhang X.. Molecular simulation study on methane adsorption capacity and mechanism in clay minerals: Effect of clay type, pressure, and water saturation in shales. Energy & Fuels. 2019;33:765-778. https://doi.org/10.1021/acs.energyfuels.8b03462.
[Google Scholar]
Ungerer P., Collell J., Yiannourakou M.. Molecular modeling of the volumetric and thermodynamic properties of kerogen: Influence of organic type and maturity. Energy & Fuels. 2015;29:91-105. https://doi.org/10.1021/ef502154k
[Google Scholar]
Sun Z., Li X., Liu W., Zhang T., He M., Nasrabadi H.. Molecular dynamics of methane flow behavior through realistic organic nanopores under geologic shale condition: Pore size and kerogen types. Chemical Engineering Journal. 2020;398:124341. https://doi.org/10.1016/j.cej.2020.124341.
[Google Scholar]
Thommes M., Kaneko K., Neimark AV., Olivier J.P., Rodriguez-Reinoso F., Rouquerol J., Sing K.S.W.. Physisorption of gases, with special reference to the evaluation of surface area and pore size distribution (IUPAC Technical Report) Pure and Applied Chemistry. 2015;87:1051-1069.
[Google Scholar]
Li H., He S., Radwan A.E., Xie J., Qin Q.. Quantitative analysis of pore complexity in lacustrine organic-rich shale and comparison to marine shale: Insights from experimental tests and fractal theory. Energy & Fuels. 2024;38:16171-16188. https://doi.org/10.1021/acs.energyfuels.4c03095.
[Google Scholar]
Li H., Xu Z., Gao X., Xie J., Qin Q., Wang H., He S.. Pore structure evolution and geological controls in lacustrine shale systems with implications for marine shale reservoir characterization. Scientific Reports. 2025;15:17702. https://doi.org/10.1038/s41598-025-02415-y.
[Google Scholar]
Gregg S.J., Sing K.S.W., Salzberg H.W.. Adsorption surface area and porosity. Journal of the Electrochemical Society. 1967;114:279Ca. https://doi.org/10.1149/1.2426447.
[Google Scholar]
Duan S., Gu M., Du X.D., Xian X.F.. Adsorption equilibrium of CO₂ and CH₄ and their mixture on sichuan basin shale. Energy Fuels. 2016;30:2248-2256. https://doi.org/10.1021/acs.energyfuels.5b02088.
[Google Scholar]
Fletcher A.J., Uygur Y., Thomas K.M.. Role of surface functional groups in the adsorption kinetics of water vapor on microporous activated carbons. The Journal of Physical Chemistry C. 2007;111:8349-8359. https://doi.org/10.1021/jp070815v.
[Google Scholar]
Gutierrez-Rodriguez J.A., Purcell R.J., Aplan F.F.. Estimating the hydrophobicity of coal. Colloids and Surfaces. 1984;12:1-25. https://doi.org/10.1016/0166-6622(84)80086-4.
[Google Scholar]
Bahadur J., Contescu C.I., Rai D.K., Gallego N.C., Melnichenko Y.B.. Clustering of water molecules in ultramicroporous carbon: In-situ small-angle neutron scattering. Carbon. 2017;111:681-688. https://doi.org/10.1016/j.carbon.2016.10.040.
[Google Scholar]
Svabova M., Weishauptova Z., Pribyl O.. Water vapour adsorption on coal. Fuel. 2011;90:1892-1899. https://doi.org/10.1016/j.fuel.2011.01.005
[Google Scholar]
Brennan J.K., Thomson K.T., Gubbins K.E.. Adsorption of water in activated carbons: Effects of pore blocking and connectivity. Langmuir. 2002;18:5438-5447. https://doi.org/10.1021/la0118560.
[Google Scholar]

Show Sections

[1] Shen W., Li X., Ma T., Cai J., Lu X., Zhou S.. High-pressure methane adsorption behavior on deep shales: Experiments and modeling. Physics of Fluids. 2021;33:063103. https://doi.org/10.1063/5.0054486
[Google Scholar]

[2] Rezlerova E., Brennan J.K., Lisal M.. Methane and carbon dioxide in dual-porosity organic matter: Molecular simulations of adsorption and diffusion. AIChE Journal. 2021;67:e16655. https://doi.org/10.1002/aic.16655
[Google Scholar]

[3] Ghanbari E., Dehghanpour H.. The fate of fracturing water: A field and simulation study. Fuel. 2016;163:282-294. https://doi.org/10.1016/j.fuel.2015.09.040
[Google Scholar]

[4] Tang X., Ripepi N., Valentine K.A., Keles C., Long T., Gonciaruk A.. Water vapor sorption on marcellus shale: Measurement, modeling and thermodynamic analysis. Fuel. 2017;209:606-614. https://doi.org/10.1016/j.fuel.2017.07.062
[Google Scholar]

[5] Chen J., Gai H., Xiao Q.. Effects of composition and temperature on water sorption in overmature Wufeng-Longmaxi shales. International Journal of Coal Geology. 2021;234:103673. https://doi.org/10.1016/j.coal.2020.103673
[Google Scholar]

[6] Liu F.Y., Ellett K., Xiao Y.T., Rupp J.A.. Assessing the feasibility of CO2 storage in the new albany shale (devonian-mississippian) with potential enhanced gas recovery using reservoir simulation. International Journal of Greenhouse Gas Control. 2013;17:111-126. https://doi.org/10.1016/j.ijggc.2013.04.018
[Google Scholar]

[7] Feng D., Li X., Wang X., Li J., Sun F., Sun Z., Zhang T., Li P., Chen Y., Zhang X.. Water adsorption and its impact on the pore structure characteristics of shale clay. Applied Clay Science. 2018;155:126-138. https://doi.org/10.1016/j.clay.2018.01.017.
[Google Scholar]

[8] Guan Q.Z., Dong D.Z., Wang S.F., Huang J.L., Zhang C.C.. Analyses on differences of microstructure between marine and lacustrine facies shale reservoirs. Natural Gas Geoscience. 2016;27:524-531. https://doi.org/10.11764/j.issn.1672-1926.2016.03.0524
[Google Scholar]

[9] Guo X.S., Hu D.F., Liu R.B., Wei X.F., Wei F.B.. Geological conditions and exploration potential of Permian marine-continent transitional facies shale gas in the Sichuan Basin. Natural Gas Industry. 2018;38:11-18. https://doi.org/10.3787/j.issn.1000-0976.2018.10.002.
[Google Scholar]

[10] Duan S., Li G.. Equilibrium and kinetics of water vapor adsorption on shale. Journal of Energy Resources Technology. 2018;140:122001-1-122001-10. https://doi.org/10.1115/1.4040530
[Google Scholar]

[11] Sang G., Liu S., Elsworth D.. Water vapor sorption properties of illinois shales under dynamic water vapor conditions: Experimentation and modeling. Water Resources Research. 2019;55:7212-7228. https://doi.org/10.1029/2019wr024992.
[Google Scholar]

[12] Sang G., Liu S., Elsworth D., Yang Y., Fan L.. Evaluation and modeling of water vapor sorption and transport in nanoporous shale. International Journal of Coal Geology. 2020;228:103553. https://doi.org/10.1016/j.coal.2020.103553.
[Google Scholar]

[13] Wang T., Tian S., Li G., Sheng M., Ren W., Liu Q., Tan Y., Zhang P.. Experimental study of water vapor adsorption behaviors on shale. Fuel. 2019;248:168-177. https://doi.org/10.1016/j.fuel.2019.03.029.
[Google Scholar]

[14] Zolfaghari A., Dehghanpour H., Holyk J.. Water sorption behaviour of gas shales: I. Role of clays. International Journal of Coal Geology. 2017;179:130-138. https://doi.org/10.1016/j.coal.2017.05.008.
[Google Scholar]

[15] Li L., Du X., Zhou J., Wang G., Zhang D., Duan M., Yin H., Heng X., Fu J., Wang H.. Water adsorption and transport in shale nanopore: The impact of three-dimensional surface roughness. Physics of Fluids. 2025;37:082131. https://doi.org/10.1063/5.0282876.
[Google Scholar]

[16] Zhang C., Yao Y., Swennen R., Wu H.. Combined effects of the chemical structure and nanopore development on water vapor/liquid adsorption in shale kerogen. Colloids and Surfaces A: Physicochemical and Engineering Aspects. 2022;653:129920. https://doi.org/10.1016/j.colsurfa.2022.129920.
[Google Scholar]

[17] Zhang W., Mo F., Qi Z., Huang X., Yue P., Liu Q., Wan Q., Fu W.. Imbibition mechanism of water in illite nanopores of deep shales by molecular simulation. Journal of Molecular Liquids. 2024;414:126232. https://doi.org/10.1016/j.molliq.2024.126232.
[Google Scholar]

[18] Liu J., Wen Y., Jiang L., Yuan H., Liu Z., Shui Z., Wang B., Yan X.. Molecular dynamics simulation of micro mechanisms of gas and water migration and diffusion in shale reservoirs. Colloids and Surfaces A: Physicochemical and Engineering Aspects. 2024;682:132865. https://doi.org/10.1016/j.colsurfa.2023.132865.
[Google Scholar]

[19] Du X., Cheng Y., Liu Z., Yin H., Wu T., Huo L., Shu C.. CO2 and CH4 adsorption on different rank coals: A thermodynamics study of surface potential, gibbs free energy change and entropy loss. Fuel. 2021;283:118886. https://doi.org/10.1016/j.fuel.2020.118886.
[Google Scholar]

[20] Du X., Pang D., Cheng Y., Zhao Y., Hou Z., Liu Z., Wu T., Shu C.. Adsorption of CH₄, N₂, CO₂, and their mixture on montmorillonite with implications for enhanced hydrocarbon extraction by gas injection. Applied Clay Science. 2021;210:106160. https://doi.org/10.1016/j.clay.2021.106160
[Google Scholar]

[21] Hatch C.D., Wiese J.S., Crane C.C., Harris K.J., Kloss H.G., Baltrusaitis J.. Water adsorption on clay minerals as a function of relative humidity: Application of BET and freundlich adsorption models. Langmuir. 2012;28:1790-1803. https://doi.org/10.1021/la2042873
[Google Scholar]

[22] Shi K.Y., Chen J.Q., Pang X.Q., Jiang F.J., Hui S.S., Zhao Z.C., Chen D., Cong Q., Wang T., Xiao H.Y., Yang X.B., Wang Y.Y.. Wettability of different clay mineral surfaces in shale: Implications from molecular dynamics simulations. Petroleum Science. 2023;20:689-704. https://doi.org/10.1016/j.petsci.2023.02.001.
[Google Scholar]

[23] Li W., Pang X., Snape C., Zhang B., Zheng D., Zhang X.. Molecular simulation study on methane adsorption capacity and mechanism in clay minerals: Effect of clay type, pressure, and water saturation in shales. Energy & Fuels. 2019;33:765-778. https://doi.org/10.1021/acs.energyfuels.8b03462.
[Google Scholar]

[24] Ungerer P., Collell J., Yiannourakou M.. Molecular modeling of the volumetric and thermodynamic properties of kerogen: Influence of organic type and maturity. Energy & Fuels. 2015;29:91-105. https://doi.org/10.1021/ef502154k
[Google Scholar]

[25] Sun Z., Li X., Liu W., Zhang T., He M., Nasrabadi H.. Molecular dynamics of methane flow behavior through realistic organic nanopores under geologic shale condition: Pore size and kerogen types. Chemical Engineering Journal. 2020;398:124341. https://doi.org/10.1016/j.cej.2020.124341.
[Google Scholar]

[26] Thommes M., Kaneko K., Neimark AV., Olivier J.P., Rodriguez-Reinoso F., Rouquerol J., Sing K.S.W.. Physisorption of gases, with special reference to the evaluation of surface area and pore size distribution (IUPAC Technical Report) Pure and Applied Chemistry. 2015;87:1051-1069.
[Google Scholar]

[27] Li H., He S., Radwan A.E., Xie J., Qin Q.. Quantitative analysis of pore complexity in lacustrine organic-rich shale and comparison to marine shale: Insights from experimental tests and fractal theory. Energy & Fuels. 2024;38:16171-16188. https://doi.org/10.1021/acs.energyfuels.4c03095.
[Google Scholar]

[28] Li H., Xu Z., Gao X., Xie J., Qin Q., Wang H., He S.. Pore structure evolution and geological controls in lacustrine shale systems with implications for marine shale reservoir characterization. Scientific Reports. 2025;15:17702. https://doi.org/10.1038/s41598-025-02415-y.
[Google Scholar]

[29] Gregg S.J., Sing K.S.W., Salzberg H.W.. Adsorption surface area and porosity. Journal of the Electrochemical Society. 1967;114:279Ca. https://doi.org/10.1149/1.2426447.
[Google Scholar]

[30] Duan S., Gu M., Du X.D., Xian X.F.. Adsorption equilibrium of CO₂ and CH₄ and their mixture on sichuan basin shale. Energy Fuels. 2016;30:2248-2256. https://doi.org/10.1021/acs.energyfuels.5b02088.
[Google Scholar]

[31] Fletcher A.J., Uygur Y., Thomas K.M.. Role of surface functional groups in the adsorption kinetics of water vapor on microporous activated carbons. The Journal of Physical Chemistry C. 2007;111:8349-8359. https://doi.org/10.1021/jp070815v.
[Google Scholar]

[32] Gutierrez-Rodriguez J.A., Purcell R.J., Aplan F.F.. Estimating the hydrophobicity of coal. Colloids and Surfaces. 1984;12:1-25. https://doi.org/10.1016/0166-6622(84)80086-4.
[Google Scholar]

[33] Bahadur J., Contescu C.I., Rai D.K., Gallego N.C., Melnichenko Y.B.. Clustering of water molecules in ultramicroporous carbon: In-situ small-angle neutron scattering. Carbon. 2017;111:681-688. https://doi.org/10.1016/j.carbon.2016.10.040.
[Google Scholar]

[34] Svabova M., Weishauptova Z., Pribyl O.. Water vapour adsorption on coal. Fuel. 2011;90:1892-1899. https://doi.org/10.1016/j.fuel.2011.01.005
[Google Scholar]

[35] Brennan J.K., Thomson K.T., Gubbins K.E.. Adsorption of water in activated carbons: Effects of pore blocking and connectivity. Langmuir. 2002;18:5438-5447. https://doi.org/10.1021/la0118560.
[Google Scholar]

Adsorption of water in organic matter, clay minerals and deep gas-bearing reservoir: Insights from experimentation and molecular simulation

Abstract

Keywords

Adsorption

Clay minerals

Gas-bearing reservoir

Organic matter

Water

1. Introduction

2. Materials and Methods

2.1. Geological setting

2.2. Characterization of sample

2.2.1. Surface physical properties characterization

2.2.2. Surface chemical properties characterization

2.2.3. Pore structure measurement

2.3. Adsorption experiment of water vapor

2.4. Theory for water adsorption

2.4.1. Model for adsorption

2.4.2. Affinity for adsorption

2.5. Molecular Simulation

2.5.1. Model construction

2.5.1.1. Clay minerals model

2.5.1.2. Organic matter model

2.5.2. Simulation details

3. Results and Discussion

3.1. Surface morphology information

3.2. Pore structure information

3.2.1. Micropore information

3.2.2. Mesopore and macropore information

3.3. Surface chemical properties

3.4. Adsorption performance for water on sample

3.4.1. Isothermal desorption and adsorption curves

3.4.2. Behaviors for primary and secondary adsorption

3.4.3. Affinity for water adsorption

3.5. Adsorption performance for water on organic matter and clay minerals

3.5.1. Density distribution of water on organic matter and clay minerals

3.5.2. Adsorption heat of water on organic matter and clay minerals

3.5.3. Energy distribution of water on organic matter and clay minerals

4. Conclusions

Acknowledgment

CRediT authorship contribution statement

Declaration of competing interest

Data availability

Declaration of generative AI and AI-assisted technologies in the writing process

Supplementary data

References

Suggested read for related articles: