Graphitic Slit Pores Research Articles

Effects of Surface Heterogeneity on the Adsorption of CO2in Microporous Carbons

Carbon capture combined with utilization and storage has the potential to serve as a near-term option for CO(2) emissions reduction. CO(2) capture by carbon-based sorbents and CO(2) storage in geologic formations such as coal and shale both require a thorough understanding of the CO(2) adsorption properties in microporous carbon-based materials. Complex pore structures for natural organic materials, such as coal and gas shale, in addition to general carbon-based porous materials are modeled as a collection of independent, noninterconnected, functionalized graphitic slit pores with surface heterogeneities. Electronic structure calculations coupled with van der Waals-inclusive corrections have been performed to investigate the electronic properties of functionalized graphitic surfaces. With Bader charge analysis, electronic structure calculations can provide the initial framework comprising both the geometry and corresponding charge information required to carry out statistical modeling. Grand canonical Monte Carlo simulations were carried out to determine the adsorption isotherms for a given adsorbent-adsorbate interaction at temperature/pressure conditions relevant to carbon capture applications to focus on the effect of the surface functionalities. On the basis of the current work, oxygen-containing functional groups were predicted to enhance CO(2) adsorption in microporous carbon materials in the absence of water vapor, and the hydrated graphite was found to hinder CO(2) adsorption.

Coarse-grained potential models for structural prediction of carbon dioxide (CO2) in confined environments

In this paper, we propose coarse-grained single-site (CGSS), wall-CO(2), and CO(2)-CO(2) interaction potential models to study the structure of carbon dioxide under confinement. The CGSS potentials are used in an empirical potential based quasi-continuum theory, EQT, to compute the center-of-mass density and potential profiles of CO(2) confined inside different size graphite slit pores. Results obtained from EQT are compared with those obtained from all-atom molecular dynamics (AA-MD) simulations, and are found to be in good agreement with each other. Though these CGSS interaction potentials are primarily developed and parameterized for EQT, they are also used to perform coarse-grained molecular dynamics (CG-MD) simulations. The results obtained from CG-MD simulations are also found to be in reasonable agreement with AA-MD simulation results.

Simulation of gas adsorption on a surface and in slit pores with grand canonical and canonical kinetic Monte Carlo methods

We present for the first time in the literature a new scheme of kinetic Monte Carlo method applied on a grand canonical ensemble, which we call hereafter GC-kMC. It was shown recently that the kinetic Monte Carlo (kMC) scheme is a very effective tool for the analysis of equilibrium systems. It had been applied in a canonical ensemble to describe vapor-liquid equilibrium of argon over a wide range of temperatures, gas adsorption on a graphite open surface and in graphitic slit pores. However, in spite of the conformity of canonical and grand canonical ensembles, the latter is more relevant in the correct description of open systems; for example, the hysteresis loop observed in adsorption of gases in pores under sub-critical conditions can only be described with a grand canonical ensemble. Therefore, the present paper is aimed at an extension of the kMC to open systems. The developed GC-kMC was proved to be consistent with the results obtained with the canonical kMC (C-kMC) for argon adsorption on a graphite surface at 77 K and in graphitic slit pores at 87.3 K. We showed that in slit micropores the hexagonal packing in the layers adjacent to the pore walls is observed at high loadings even at temperatures above the triple point of the bulk phase. The potential and applicability of the GC-kMC are further shown with the correct description of the heat of adsorption and the pressure tensor of the adsorbed phase.

Packing effects on argon and methanol adsorption inside graphitic cylindrical and slit pores: A GCMC simulation study

Using Grand Canonical Monte Carlo simulation, we have studied the effects of confinement on argon and methanol adsorption in graphitic cylindrical and slit pores. Linear chain, zigzag and incomplete helical packing are observed for argon adsorption in cylindrical pores. However, for methanol adsorption different features appear because the electrostatic interactions favour configurations that maximize the hydrogen bonding among methanol molecules. We have found zigzag chains with hydrogen-bonded structures for methanol adsorption in cylindrical and slit pores. To investigate how dense the adsorbed phase is and how many molecules could be packed per unit physical volume of the solid, we consider two different definitions of pore density; one based on the physical volume and the other on the accessible volume. That based on accessible volume gives a measure of the fluid density, while that based on the physical volume gives a measure of how much adsorbate can be stored per unit volume of the adsorbent. It is found that the adsorbate is denser in cylindrical pores, but that slit pores can pack more molecules per unit solid volume. We also discuss the effects on the isosteric heat of argon and methanol of pore size, pore geometry and loading.

Phase Transitions of Water in Graphite and Mica Pores

We report all-atom molecular dynamics simulations of water confined in graphite and mica slit pores of variable size ranging from 10 to 60 A. For each pore size, we demonstrate that the confinement...

Electrolyte Effects in a Model System for Mesoporous Carbon Electrodes

In this paper, a variety of alkali halide aqueous electrolyte solutions in contact with planar graphite slit pores are simulated using classical molecular dynamics. Size trends in structure and transport properties are examined by varying the choice of ions. The intermediate atomic weight ions within each group are found to diffuse faster than the larger or smaller ions. System dynamics are driven by changes in water hydration behavior and, specifically, by variations in the number of hydrogen bonds per water molecule. Both the cation and the anion sequences demonstrate that confinement effects can significantly alter the expected trends of alkali halide electrolytes.

Evaluation of the interaction potentials for methane adsorption on graphite and in graphitic slit pores

This paper compares the performance of the Buckingham Exponential-6 and Lennard-Jones potential models in the description of bulk phase and adsorption properties of methane on graphitic surfaces and pores. The solid-fluid potential used in the choice of the LJ model is Steele 10-4-3 equation and for the Exp-6 model, the Crowell and Chang equation which has been rarely used in the literature. From an extensive computer study using Grand Canonical Monte Carlo simulation, the two potential models perform almost equally well in the bulk fluid behavior except at extremely high density, where the LJ model is better. For adsorption on surface, the Exp-6 performs better in the correct description of the experimental Henry constant. However, both potential models describe well the isotherm outside the Henry law region. Under supercritical conditions, the same behavior is seen in the Henry law region, but the opposite is observed at extremely high pressures. For adsorption in slit pores, significant difference is seen at low pressure region for all pore sizes examined. In this region, the Exp-6 always predicts a higher capacity than the LJ model. In the smallest pore size examined (0.65 nm), the LJ model predicts a higher capacity than the Exp-6 with approximately 4% difference at higher pressures. However, this behavior is not seen in the other pore sizes. The comparison shows that the Exp-6 can describe experimental adsorption data, albeit only, better than the LJ.

On the influence of heterogeneity of graphene sheets in the determination of the pore size distribution of activated carbons

We model carbon-based microporous materials, such as activated carbons, taking into account surface defects in the form of geometrical rugosity in the inner surface of each graphitic slit pore. The used model is a simplified variation of the randomly etched graphite (REG) pore model (Seaton et al., Langmuir 13:1199–1204, 1997).

Pore-Width-Dependent Preferential Interaction of sp2Carbon Atoms in Cyclohexene with Graphitic Slit Pores by GCMC Simulation

The adsorption of cyclohexene with two sp2and four sp3carbon atoms in graphitic slit pores was studied by performing grand canonical Monte Carlo simulation. The molecular arrangement of the cyclohexene on the graphitic carbon wall depends on the pore width. The distribution peak of the sp2carbon is closer to the pore wall than that of the sp3carbon except for the pore width of 0.7 nm, even though the Lennard-Jones size of the sp2carbon is larger than that of the sp3carbon. Thus, the difference in the interactions of the sp2and sp3carbon atoms of cyclohexene with the carbon pore walls is clearly observed in this study. The preferential interaction of sp2carbon gives rise to a slight tilting of the cyclohexene molecule against the graphitic wall. This is suggestive of a π-π interaction between the sp2carbon in the cyclohexene molecule and graphitic carbon.

A novel and consistent method (TriPOD) to characterize an arbitrary porous solid for its accessible volume, accessible geometrical surface area and accessible pore size

We present an improved Monte Carlo integration method to calculate the accessible pore size distribution of a porous solid having known configuration of solid atoms. The pore size distribution obtained with the present method is consistent with the accessible volume and the accessible geometric surface area presented in previous publications (Do and Do, in J. Colloid Interface Sci. 316(2):317–330, 2007; Do et al. in Adsorpt. J., 2010). The accessible volume, accessible geometrical surface area and the pore size distribution method construct an unambiguous and robust single framework to characterize porous solids. This framework is based on the derivation of the space accessible to the center of mass of a probe molecule. The accessible pore size presented is an absolute quantity in the sense that a zero value is possible. We present the entire framework of this characterization method and compare the improved method with the one presented previously for a set of porous solids such as graphitic slit pores, defective slit pores, bundle of carbon nanotubes, zeolite and some metal organic frameworks.

Equilibrium behaviour of a novel gas separation process, with application to carbon capture

A novel gas separation process is described and analysed in the context of carbon capture. It involves a highly selective absorbent fluid below its saturation pressure pre-mixed with the gas to be separated and adsorbed into a porous solid. This fluid mixture simultaneously forms gas-like and liquid-like regions within the porous solid depending on the pore size. The gas component to be separated is selectively absorbed by the liquid-like regions, and a novel ‘pressure-swing wetting layer absorption’ process is used to recover the absorbed gas. This work examines the equilibrium behaviour of this process in the context of carbon capture using the density functional theory (DFT) of classical fluids. The DFT model employed represents the porous solid in terms of ideal graphitic slit-pores, and a ternary fluid model is calibrated to represent mixtures of tetrahydrofuran (the absorbent fluid), carbon dioxide and nitrogen. Under the conditions investigated here we find that the equilibrium behaviour of this system is superior to the analogous pressure-swing adsorption process without solvent. This result motivates further experimental and dynamical process modelling studies of this system.

On the Existence of Negative Excess Isotherms for Argon Adsorption on Graphite Surfaces and in Graphitic Pores under Supercritical Conditions at Pressures up to 10,000 atm

In this paper, we consider in detail the computer simulation of argon adsorption on a graphite surface and inside graphitic slit pores under supercritical conditions. Experimental results in the literature for graphitic adsorbents show that excess isotherms pass through a maximum and then become negative at high pressures (even for adsorption on open surfaces) when a helium void volume is used in the calculation of the excess amount. Here we show that, by using the appropriate accessible volume (which is smaller than the helium void volume), the excess isotherms still have a maximum but are always positive. The existence and the magnitude of this maximum is because the rate of change of the adsorbed density is equal to that of the bulk gas, which has a large change in bulk gas density for a small variation in pressure for temperatures not far above the critical point. However for temperatures far above T(c), this change in the bulk gas density is no longer significant and the maximum in the surface excess density becomes less pronounced and even disappears at high enough temperatures. The positivity of the adsorption excess persists for all pressures up to 10,000 atm for adsorption on surfaces and in slit pores of all sizes. For adsorption on a surface, the surface excess density eventually reaches a plateau at high pressures as expected, because the change in the adsorbed phase is comparable to that of the bulk gas. Positive excess lends support to our physical argument that the adsorbed phase is denser than the bulk gas, and this is logical as the forces exerted by the pore walls should aid to the compression of the adsorbed phase.

Molecular simulation of collection of methane isotherms in carbon material using all-atom and united atom models

In this study the performances of all-atom (AA) and united atom (UA) models (fitted and unfitted) of the methane molecule are evaluated for the description of adsorption isotherms on graphite surfaces and in a collection of graphitic slit pores. We simulated collections of isotherms using the grand canonical Monte Carlo method with an all-atom model and adjusted united atom models in graphene layers made up of discrete atoms of carbon. The collections of isotherms are used to determine the pore size distribution (PSD) of four activated carbon samples. We also investigated the sensitivity of the system to the cutoff and solid–fluid standard parameterization. It was found that the simulated AA model isotherm shape on the graphite surface is much more similar to the experimental data than the UA model isotherm. The cutoff had little influence on isotherm and different solid–fluid standard parameterizations change the PSD. We also found that despite presenting similar fitting with the experimental isotherms, the models presented distinct PSDs. The unfitted united atom model (UA1) suggested less plausible PSDs, while the all-atom (AA) and the fitted united atom model (UA2) model provided apparently more realistic estimates of the internal structure of microporous carbons.

A new method to determine pore size and its volume distribution of porous solids having known atomistic configuration

A simple method, based on Monte Carlo integration, is presented to derive pore size and its volume distribution for porous solids having known configuration of solid atoms. Because pores do not have any particular shape, it is important that we define the pore size in an unambiguous manner and the volume associated with each pore size. The void volume that we adopt is the one that is accessible to the center of mass of the probe particle. We test this new method with porous solids having well defined pores such as graphitic slit pores and carbon nanotubes, and then apply it to obtain the pore volume distribution of complex solids such as disordered solids, rectangular pores, defected graphitic pores, metal organic framework and zeolite.

Survey of classical density functionals for modelling hydrogen physisorption at77K

This work surveys techniques based on classical density functionals for modeling the quantum dispersion of physisorbed hydrogen at 77K. Two such techniques are examined in detail. The first is based on the "open ring approximation" (ORA) of Broukhno et al., and it is compared with a technique based on the semiclassical approximation of Feynman and Hibbs (FH). For both techniques, a standard classical density functional is used to model hydrogen molecule-hydrogen molecule (i.e., excess) interactions. The three-dimensional (3D) quantum harmonic oscillator (QHO) system and a model of molecular hydrogen adsorption into a graphitic slit pore at 77K are used as benchmarks. Density functional results are compared with path-integral Monte Carlo simulations and with exact solutions for the 3D QHO system. It is found that neither of the density functional treatments are entirely satisfactory. However, for hydrogen physisorption studies at 77K the ORA based technique is generally superior to the FH based technique due to a fortunate cancellation of errors in the density functionals used. But, if more accurate excess functionals are used, the FH technique would be superior.

Effect of pore wall model on prediction of diffusion coefficients for graphitic slit pores

The effect of the pore wall model on the self-diffusion coefficient and transport diffusivity predicted for methane in graphitic slit pores by equilibrium molecular dynamics (EMD) and non-equilibrium MD (NEMD) is investigated. Three pore wall models are compared--a structured wall and a smooth (specular) wall, both with a thermostat applied to the fluid to maintain the desired temperature, and a structured wall combined with the diffuse thermalizing scattering algorithm of MacElroy and Boyle (Chem. Eng. J., 1999, 74, 85). Pore sizes ranging between 7 and 35 angstroms and five pressures in the range of 1-40 bar are considered. The diffuse thermalizing wall yields incorrect self-diffusion coefficients and transport diffusivities for the graphitic slit pore model and should not be used. Surprisingly, the smooth specular wall gives self-diffusion coefficients inline with those obtained using the structured wall, indicating that this computationally much faster wall can be used for studying this phenomenon provided the fluid-wall interactions are somewhat weaker than the fluid-fluid interactions. The structured wall is required, however, if the transport diffusivity is of interest.

Evaluation of the adsorption free energy of light guest molecules in nanoporous host structures

We present a simple method for estimation of the adsorption free energy of lightweight featureless molecules in nanoporous structures. The partition function of the system is computed quantum mechanically within the ideal gas approximation, and leads directly to the free adsorption energy. Storage capacities can be estimated by combining the ideal gas results with the real gas equation of state. Two possible implementations of the approach are discussed. The assumptions and potential sources of errors in the calculations are analyzed. The performance of the method is illustrated for graphitic slit pores and solid C60.

Molecular simulation of carbon dioxide adsorption in chemically and structurally heterogeneous porous carbons

AbstractCapture of carbon dioxide from fossil fuel power plants via adsorption and sequestration of carbon dioxide in unmineable coal seams are achievable near‐term methods of reducing atmospheric emissions of this greenhouse gas. To investigate the influence of surface heterogeneity upon predicted adsorption behavior in activated carbons and coal, isotherms were generated via grand canonical Monte Carlo simulation for CO2 adsorption in slit‐shaped pores with underlying graphitic structure and several variations of chemical heterogeneity (oxygen and hydrogen content), pore width, and surface functional group orientation. Adsorption generally increased with increasing surface oxygen content, although exceptions to this trend were observed on structurally heterogeneous surfaces with holes or furrows that yield strongly adsorbing preferred binding sites. Among the heterogeneous pore structures investigated, those with coal‐like surfaces adsorbed carbon dioxide more strongly than planar, homogeneous graphitic slit pores of comparable width. Electrostatic adsorbate–adsorbent interactions significantly influenced adsorption onto model surfaces. © 2006 American Institute of Chemical Engineers Environ Prog, 25: 343–354, 2006

Adsorption, structure and dynamics of fluids in ordered and disordered models of porous carbons

Grand Canonical Monte Carlo and Molecular Dynamics simulations are used to investigate the adsorption and dynamics of argon in ordered and disordered models of porous carbons. The ordered porous carbon (model A) is a regular slit pore made up of graphene sheets. The disordered porous carbon (model B) is a structural model that reproduces the morphological (pore shape) and topological (pore connectivity) disorders of saccharose-based porous carbons. Three pore widths, H = 7, 11, and 15 Å, are selected for model A; they correspond to the smaller, mean, and larger pore sizes of model B. The filling pressures for the graphite slit pores are lower than those for the disordered porous carbon. It is also found that model A is not able to capture the behaviour of the isosteric heat of adsorption of model B. For all pressures, the confined phase in model A is composed of well-defined layers, which crystallize into hexagonal 2D crystals after complete filling of the pores. In contrast, the structure of argon in the disordered porous carbon remains liquid-like overall. It is also found that the slit pore model cannot reproduce the dynamics of argon in the disordered porous carbon. While the self-diffusivity of argon in model A decreases with increasing loading, it exhibits a maximum for model B. Such a non-monotonic behaviour of the self-diffusivity for the disordered porous carbon can be explained by the surface (energetic) heterogeneities of the material.

Correct Procedures for the Calculation of Heats of Adsorption for Heterogeneous Adsorbents from Molecular Simulation

Several procedures for calculating the heat of adsorption from Monte Carlo simulations for a heterogeneous adsorbent are presented. Simulations have been performed to generate isotherms for nitrogen at 77 K and methane at 273.15 K in graphitic slit pores of various widths. The procedures were then applied to calculate the heat of adsorption of an activated carbon with an arbitrary pore size distribution. The consistency of the different procedures shows them to be correct in calculating interaction energy contributions to the heat of adsorption. The currently favored procedure for this type of calculation, from the literature, is shown to be incorrect and in serious error when calculating the heat of adsorption of activated carbon.