Study on Shale Adsorption Equation Based on Monolayer Adsorption, Multilayer Adsorption, and Capillary Condensation
Shale gas is an effective gas resource all over the world. The evaluation of pore structure plays a critical role in exploring shale gas efficiently. Nitrogen adsorption experiment is one of the significant approaches to analyze pore size structure of shale. Shale is extremely heterogeneous due to component diversity and structure complexity. Therefore, adsorption isotherms for homogeneous adsorbents and empirical isotherms may not apply to shale. The shape of adsorption-desorption curve indicates that nitrogen adsorption on shale includes monolayer adsorption, multilayer adsorption, and capillary condensation. Usually, Langmuir isotherm is a monolayer adsorption model for ideal interfaces; BET (Brunauer, Emmett, Teller) adsorption isotherm is a multilayer adsorption model based on specific assumptions; Freundlich isotherm is an empirical equation widely applied in liquid phase adsorption. In this study, a new nitrogen adsorption isotherm is applied to simultaneously depict monolayer adsorption, multilayer adsorption, and capillary condensation, which provides more real and accurate representation of nitrogen adsorption on shale. In addition, parameters are discussed in relation to heat of adsorption which is relevant to the shape of the adsorption isotherm curve. The curve fitting results indicate that our new nitrogen adsorption isotherm can appropriately describe the whole process of nitrogen adsorption on shale.
Shale gas has attracted much attention in United States, China, Canada, and so forth, because of the gas storage mechanism and recovery potential of shale gas reservoirs [1, 2]. To investigate gas adsorption capacity and pore size distribution of shale rocks, high-pressure methane adsorption and low-pressure nitrogen or carbon dioxide adsorption experiments are conducted, respectively. Many researches have been done to find and modify adsorption equations suitable for describing methane adsorption. Considering methane adsorption as monolayer adsorption, Langmuir equation, L-F (Langmuir-Freundlich) equation, and M-L (modified Langmuir) equation are successfully applied to evaluate methane adsorption [3–5]. Furthermore, D-R (Dubinin-Radushkevich) equation, D-A (Dubinin-Astakhov) equation, and S-D-R (supercritical Dubinin-Radushkevich) equation are also used with consideration of methane adsorption as micropore filling [6–8]. For carbon dioxide adsorption, to take into account the monolayer adsorption property, both Langmuir equation and L-F equation are applied to depict variations of the adsorption capacity with pressure [9–11]. On the contrary, it is hard to find an equation to depict low-pressure nitrogen adsorption because of the complicated adsorption mechanism. On the basis of BDDT (Brunauer-Deming-Deming-Teller) adsorption isotherm classification, nitrogen adsorption belongs to type IV, which indicates that it includes three processes: monolayer adsorption, multilayer adsorption, and capillary condensation. Unfortunately, the majority of adsorption equations are developed based on only one kind of adsorption mechanism, and they can be categorized into three aspects: monolayer adsorption, multilayer adsorption, and micropore filling.
In terms of monolayer adsorption, a widely accepted one is Langmuir adsorption equation which assumed only one type of adsorption sites on the surface of adsorbent [12, 13]. When extending the Langmuir equation for gas-liquid-phase adsorption studies, two types of sites are considered and the relationship between equilibrium concentration and amount of adsorbate is obtained [14–19]. Because the Langmuir equation describes adsorption on homogeneous surface, Gaussian energy distribution is used to adjust monolayer adsorption theory to heterogeneous surface [20–22]. To study multicomponent, monolayer adsorption of multicomponent gas, the assumption that the saturated amount of adsorption for each component is equal based on Langmuir equation was derived [23, 24].
In the aspect of multilayer adsorption, BET (Brunauer, Emmett, and Teller) equation is the most popular one, and it proposes a multilayer adsorption model which assumes that the interaction on adsorbent surface is much larger than that between neighboring adsorbate molecules [25–27]. The theory is appropriate for adsorption on solid surfaces with homogeneous chemical properties, which is frequently applied to calculate specific surface area for porous media. To extend BET equation to multicomponent adsorption, three kinds of n-component BET equations were proposed considering that adsorbed layers have evaporation-condensation characters for liquid mixture, supposing that the adsorbed layer of gas mixture is an ideal solution according to statistic thermodynamics and assuming gas mixture is immiscible liquid [28–31].
Micropore filling is also a common adsorption mechanism, which is introduced on the basis of Polanyi adsorption potential theory . According to thermodynamics, adsorption potential is transferring unit mass of adsorbate from gas phase to adsorbent surface. On account of thermodynamics, D-R and D-A equations were generated [33–36]. Changing micropore filling to surface coverage and keeping feature of Gaussian distribution of energy, D-R-K (Dubinin-Radushkevich-Kaganer) equation was built [37, 38]. For micropore filling on nonregular porous media, D-R equation was modified by fractal dimension function . Furthermore, for supercritical fluid adsorption, S-D-R equation was built .
In fact, most adsorbents are heterogeneous porous media. Combination of adsorption equations is a solution to build equation for heterogeneous adsorbent. In studies of methane adsorption, empirical Freundlich equation was combined with Langmuir equation to obtain the L-F equation which is widely used in depicting CBM (coal bed methane) adsorption successfully [40, 41]. Moreover, heterogeneity of adsorbent surface has been taken into account, and an adsorption equation was built to express the relationship between equilibrium concentration and mass of adsorbate by combining Freundlich adsorption isotherm with Langmuir adsorption isotherm [42, 43].
It is clear that these adsorption equations only focus on one adsorption mechanism and cannot be applied to interpret monolayer adsorption, multilayer adsorption, and capillary condensation simultaneously. In addition, it is well known that shale consists of clay minerals (kaolinite, illite, chlorite, etc.), detrital minerals (quartz, feldspar, etc.), and some characteristic minerals (such as pyrite) [44–48], each with its specific adsorption property. On the other hand, pore size distribution in shale is irregular , which results in an uneven distribution of adsorption potential. Compared with homogeneous materials used in other interfacial phenomenon studies, shale is an extremely heterogeneous adsorbent. However, most of current adsorption models assume that adsorbent is homogenous. Hence, our research is aiming at building a new adsorption equation for shale which enables us to depict complex adsorption including monolayer adsorption, multilayer adsorption, and capillary condensation.
Adsorption and desorption data of shale measured by nitrogen at low temperature (77 K) is a fundamental method to analyze pore structure of shale. Samples were collected from Yanchang formation (Triassic, Ordos), Pingliang formation (Ordovician, Ordos), Wulalik formation (Ordovician, Ordos), Xujiahe formation (Triassic, Sichuan), Niutitang formation (Cambrian, Sichuan), and Doushantuo formation (Ediacaran, Sichuan). Properties of samples are described in Table 1. All samples were ground to pass a sieve size of 60 mesh (250 μm). For outgassing, the pulverized samples were dried and vacuumized at 80°C for 12 hours.
The apparatus used for nitrogen adsorption experiment is Quadrasorb SI surface area and pore size analyzer (manufactured by Quantachrome in USA) which is provided by State Key Laboratory of Oil & Gas Reservoir Geology and Exploitation (China). There are four stations of the experimental instrument. The lower limitation of specific surface area is 0.01 m2/g for nitrogen. In the aspect of pore size distribution analysis, the minimum pore volume is 0.0001 cc/g (STP), and the pore size range is 0.35~400 nm. In our experiment, nitrogen is used as adsorbate. Measurement is conducted at temperature 77 K, and the minimum is 0.001.
All the experimental data are prepared to analyze nitrogen adsorption processes and determine the value of parameters in shale adsorption isotherm.
3. Adsorption Characteristics of Nitrogen Adsorbed on Shale
3.1. Adsorption Processes
According to BDDT (Brunauer-Deming-Deming-Teller) adsorption isotherm classification , nitrogen adsorption isotherms of shale belong to type IV (Figure 1), which indicates that adsorption on shale can be divided into three stages: monolayer adsorption, multilayer adsorption, and capillary condensation [51–54]. The three stages can be specifically expressed as follows: most adsorption isotherms of shale have an inflection point at low relative pressure, which refers to saturated adsorbed content in monolayer adsorption regime. Before this point, only monolayer adsorption takes place. As relative pressure increases, the thickness of adsorbed layers gradually increases and multilayer adsorption occurs. When relative pressure reaches initial capillary condensation pressure (usually around 0.4 ), adsorption-desorption curve forms hysteresis loop, which demonstrates that capillary condensation exists in the process of nitrogen adsorbed on shale.
3.2. Adsorption Equations and Adsorption Processes
As mentioned above, shale adsorption includes processes of monolayer adsorption, multilayer adsorption, and capillary condensation. Therefore, the generated new shale adsorption isotherm would be capable of depicting all features of these processes.
In terms of monolayer adsorption, Langmuir equation will be an appropriate choice. Langmuir built an adsorption model with the following assumptions: () surface of adsorbent has one type of adsorption sites and one site can accommodate only one adsorbate molecule or atom; () the surface is homogeneous and there is no lateral interaction between adsorbate molecules; () adsorption reaches dynamic equilibrium . Based on these assumptions, the adsorption isotherm can be given as
All terms used in the equations are defined in the nomenclature section.
For multilayer adsorption, BET adsorption isotherm is a representation of multilayer adsorption model generated by Brunauer, Emmett, and Teller, which assumes the interaction between adsorbate and adsorbent surface is much larger than that between neighboring molecules. The theory is appropriate for adsorption on surface of solid with homogeneous chemical properties, which is frequently applied to calculate specific surface area for porous media. The BET equation can be expressed as 
Capillary condensation is a process where gas phase transforms into liquid phase. Thus, an adsorption equation applicable to describe liquid adsorption is suitable for this adsorption stage. Among investigated adsorption equations, Freundlich adsorption isotherm is an empirical equation describing equilibrium concentration of solute in solution with respect to concentration of solute adsorbed on the surface of solvent. The adsorption equation is 
Figure 2 points out that if we only apply Langmuir isotherm to shale adsorption in low relative pressure section (before the inflection point where monolayer adsorption switches to multilayer adsorption), Langmuir isotherm can properly match experimental data, which indicates that Langmuir isotherm is suitable for monolayer adsorption in shale and then justifies the analytical result that monolayer adsorption takes place in the process of nitrogen adsorption at low temperature for shale.
As a normal method to acquire surface area of shale, multipoint BET method testifies that BET equation can be applied to describe adsorption on shale at certain conditions (usually relative pressure below 0.4 ). From curve fitting result (Figure 3), BET is appropriate for low and medium relative pressure sections, which illustrates that BET adsorption isotherm can depict experimental data before the presence of capillary condensation. This also reveals that multilayer adsorption exists in the process of nitrogen adsorption isotherm for shale.
As shown in Figure 4, Freundlich isotherm fits the medium-high relative pressure section of nitrogen adsorption on shale, especially relative pressure section after occurrence of capillary condensation.
On behalf of potentials of the three equations representing adsorption in different relative pressure sections, the new adsorption equation for shale needs to contain features of Langmuir isotherm, BET isotherm, and Freundlich isotherm.
BET and Freundlich adsorption isotherms can be changed to functions which consider relative pressure as an independent variable. Thereafter, Langmuir adsorption isotherm is a case of BET adsorption isotherm, in which the pressure is much lower than saturated vapor pressure.
Rearranging BET adsorption isotherm equation one gets
Substituting into (3) gives
Freundlich adsorption isotherm describes the relationship between pressure and mass of adsorbate adsorbed on surface of adsorbent per unit of mass. In order to express adsorption capacity in same dimension, (5) is converted to
Under experimental conditions, saturated vapor pressure and density of adsorbate are constants. Therefore, setting , then (6) becomes
From (1) and (2), the coefficient and exponent of pressure in Langmuir and BET adsorption isotherm () correspond to physical and chemical parameters of monolayer and multilayer adsorption. We apply the function with a form of BET adsorption isotherm and combine Freundlich adsorption isotherm which can describe characteristic of liquid adsorption to build up shale adsorption isotherm. The coefficient and exponent of relative pressure are variable fitting parameters in the new shale adsorption isotherm expressed as follows:A, B are undetermined coefficients; M, N, K are undetermined exponent.
5.1. Physical and Chemical Meaning of Variables in Shale Adsorption Isotherm Equation
From above discussion, coefficient and exponent in the new shale adsorption isotherm equation are related to physical and chemical meanings of coefficients and exponents of Langmuir, BET, and Freundlich adsorption isotherms.
Variable A in (8) can be given as
Thus, is related to maximum amount of monolayer adsorption (), adsorption heat according to formula of which will be detailed below, saturated vapor pressure at experimental temperature (), and exponent .
Variable in (8) is
is related to heat of adsorption and experimental temperature ().
Compared with exponents of relative pressure in (8), coefficients , , and are relevant to experimental temperature, which indicates that the enthalpy represents the strength of adsorption effect.
After clarifying the physical and chemical meaning of variables in (8), the range of these variables should be determined.
In (9), , , are all positives, and then A should be positive ().
In BET theory, it is assumed that the strength of interaction between adsorbate at first adsorbed layer and adsorbent is much bigger than the strength between adsorbates at subsequent layers. Thus, set heat of adsorption between subsequent adsorbates as . Then, represents the heat of adsorption between adsorbate at first layer and adsorbent, and . Thus, must be larger than 0, and then we can obtain .
In terms of adsorption isotherm system:
At given temperature, based on ideal gas law, can be obtained. Thus, in the whole system, change in adsorption enthalpy is equal to change in internal energy; namely, . The adsorption is a process of heat release, , so and all variables , , are positive.
For each shale sample, these parameters can be calculated by the new shale adsorption isotherm based on experimental data.
5.2. Specialization of Shale Adsorption Isotherm
When ; , (8) can be simplified to
At the right hand side of equation, multiply both numerator and denominator by saturated vapor pressure as
The term can be written as . Equation (15) can be converted to
By multiplying density of adsorbate () on both sides of (7) one gets
5.3. Application of New Adsorption Equation
According to the new shale adsorption isotherm equation and the range of variables, we applied Matlab to perform curve fitting of relative pressure versus amount of adsorption for 80 shale samples. The results of 6 samples are selected randomly and displayed in Figure 5, and the other fitting results are shown in Table 1. It appears that average value of is 0.9782, maximum value of is 0.9999, minimum value of is 0.7652, and the percentage of shale samples for which the value of is larger than 0.9 is 96.25%. It indicates that the new generated shale adsorption isotherm can represent a complete process of adsorption including monolayer adsorption, multilayer adsorption, and capillary condensation processes compared with Langmuir, BET, and Freundlich adsorption isotherms individually. In particular, it demonstrates better performance on depicting nitrogen adsorption isotherm at low temperature.
5.4. Coefficient B and the Shape of Adsorption Curve
Taking into account the physical and chemical meaning of coefficient in the new generated shale adsorption isotherm, it represents coefficient in BET isotherm (11). According to the research by Kondou et al. , the value of in BET isotherm is related to heat of adsorption. The value of is bigger, and the heat of adsorption is larger, which indicates that strength of interaction for adsorption is larger and adsorption curve increases more rapidly in low-pressure section shown in Figure 6(a). Focusing on the value of in shale adsorption isotherm and the shape of adsorption isotherm curve, we figure out that the curve becomes gradually convex as the value of increases, as shown in Figure 6(b). This illustrates that the generated shale adsorption isotherm can express the difference of heat of adsorption released between different shale samples. Furthermore, the changes in the shape of the curve are relevant to changes in heat of adsorption.
() The new shale adsorption isotherm is built up based on Langmuir adsorption isotherm, BET isotherm, and Freundlich isotherm, which can offer description for shale adsorption isotherm including monolayer adsorption, multilayer adsorption, and capillary condensation processes. The new shale adsorption isotherm can be converted to Langmuir, BET, and Freundlich isotherms by giving certain values to variables.
() The variables in new shale adsorption isotherm are related to coefficients and exponents in Langmuir, BET, and Freundlich adsorption isotherms. The physical and chemical meanings of parameters are figured out and ranges for each parameter are determined, which is used to restrict value of variables in the adsorption isotherm when doing regression analysis to match data from shale samples adsorption experiment.
() Based on new shale adsorption isotherm and variable range, curve fitting of relative pressure versus amount of adsorption has been performed. The adsorption isotherms with ability to illustrate the process of monolayer adsorption, multilayer adsorption, and capillary condensation for 80 shale samples from Ordos Basin and Sichuan Basin are obtained. The results of curve fitting are highly accurate.
() Variable in shale adsorption isotherm is related to shape of adsorption curve due to adsorption heat. Variables in shale adsorption isotherm are related to shape of adsorption curve and parameter of heat of adsorption. Adsorption isotherm curve becomes gradually convex as the value of increases. Computation according to physical and chemical meaning of coefficient B on heat of adsorption between adsorbate at first layer and adsorbent demonstrates that diversity exists among the heat of adsorption from different shale samples.
|:||Adsorption volume (cc/g)|
|:||Equilibrium pressure (MPa)|
|:||Saturated adsorption volume (cc/g)|
|:||Saturated vapor pressure at certain temperature (MPa)|
|:||Adsorption mass (g/g)|
|Pr:||The ration of equilibrium pressure and saturated vapor pressure ()|
|:||Adsorption enthalpy (J)|
|:||Molar gas constant (J/(mol·K))|
|:||Gas density (g/ml)|
|:||Heat of adsorption between adsorbate at first layer and adsorbent (kJ/mol)|
|:||Heat of adsorption between adsorbate at layer and adsorbate at layer () (kJ/mol).|
Conflicts of Interest
The authors declare that they have no conflicts of interest.
The experimental data support provided by State Key Laboratory of Oil and Gas Reservoir Geology and Exploration is gratefully acknowledged. The authors also thank Dr. Christine Ehlig-Economides from University of Houston for her guidance and assistance.
J. A. V. Butler and C. Ockrent, “Studies in electrocapillarity. III,” The Journal of Physical Chemistry, vol. 34, no. 12, pp. 2841–2859, 1930.View at: Google Scholar
E. Besedovß, D. Bobok, and E. Besedová, “Single-component and binary adsorption equilibrium of 1,2 dichloroethane and 1,2 dichloropropane on activated carbon,” Petroleum & Coal, vol. 47, no. 2, pp. 47–54, 2005.View at: Google Scholar
J. S. Jain and V. L. Snoeyink, “Adsorption from bisolute systems on active carbon,” Journal of the Water Pollution Control Federation, vol. 45, no. 12, pp. 2463–2479, 1973.View at: Google Scholar
J. Toth, “State equation of the solid-gas interface layers,” Acta Chimica Academiae Scientiarum Hungaricae, vol. 69, pp. 311–328, 1971.View at: Google Scholar
D. M. Ruthven, Principles of adsorption and adsorption processes, John Wiley & Sons, 1984.
G. Tiren, “The comparison of different BET-type mixed adsorption equations with some experimental results,” Journal of Chemical Industry and Engineering (China), vol. 1, pp. 80–84, 1984.View at: Google Scholar
G. Tiren, “The extended BET adsortion theory in mixture solution,” CHemistry, vol. 9, pp. 1–7, 1984.View at: Google Scholar
M. M. Dubinin and L. Radushkevich, “Equation of the characteristic curve of activated charcoal,” Proceedings of the Academy of Sciences USSR, vol. 55, no. 2, pp. 331–333, 1947.View at: Google Scholar
W. F. L. Brown, “The joule-thermal effect and the equation of state of gases with non-polar molecules,” Russian Journal of Physical Chemistry, vol. 30, pp. 691–704, 1959.View at: Google Scholar
D. Avnir and M. Jaroniec, “An isotherm equation for adsorption on fractal surfaces of heterogeneous porous materials,” Langmuir, vol. 5, no. 6, pp. 1431–1433, 1989.View at: Google Scholar
L. Hong et al., “Experimental study of mineral composition and brittle characteristics in longmaxi formation of lower silurian, southeast chongqing,” Science Technology and Engineering, vol. 13, pp. 8567–8571, 2013.View at: Google Scholar
T. Li, H. Tian, J. Chen, and L. Cheng, “Application of low pressure gas adsorption to the characterization of pore size distribution of shales: An example from Southeastern Chongqing area, China,” Journal of Natural Gas Geoscience, vol. 1, no. 3, pp. 221–230, 2016.View at: Publisher Site | Google Scholar
S.-B. Chen, Y.-M. Zhu, H.-Y. Wang, H.-L. Liu, W. Wei, and J.-H. Fang, “Structure characteristics and accumulation significance of nanopores in Longmaxi shale gas reservoir in the southern Sichuan Basin,” Journal of the China Coal Society, vol. 37, no. 3, pp. 438–444, 2012.View at: Google Scholar
S. Kondou, T. Ishikawa, and I. Abe, “Adsorption Science (Chinese version),” pp. 37–39, Chemical Industry Press, Beijing, China, 2006.View at: Google Scholar
H. Freundlich, “Over the adsorption in solution,” The Journal of Physical Chemistry, vol. 57, Article ID 385471, pp. 1100–1107, 1906.View at: Google Scholar
S. Kondou, T. Ishikawa, and I. Abe, Adsorption Science (Chinese version), p. 36, Chemical Industry Press, Beijing, China, 2006.
S. Kondou, T. Ishikawa, and I. Abe, Adsorption Science (Chinese version), p. 42, Chemical Industry Press, Beijing, China, 2006.
S. Kondou, T. Ishikawa, and I. Abe, Adsorption Science (Chinese version), p. 44, Chemical Industry Press, Beijing, China, 2006.