Search

The high conductivity of nanoprous carbons has a significant effect on adsorption of polar molecules; however, the mechanisms underlying this effect are not well-understood, and this effect is generally not considered in adsorption modeling. To investigate the impact of high host conductivity on the adsorption properties of vertical stacks of the graphene sheet, phenol vapor has been chosen as a simple polar adsorptive for our case studies. The influence of high surface conductivity of graphene is taken into account during Grand Canonical Monte Carlo (GCMC) simulations by considering the resulting image charges, and using the corresponding analytical solution to compute the electrostatic energy of the polar molecules confined between two infinite parallel conducting planes. We find that the consideration of the high conductivity affects the atomic configuration of the adsorbed molecules, based on our results for two different pore widths of 1 nm and 0.65 nm. Allowing for the high conductivity results in more stable energy levels, greater heat of adsorption and smaller distances between phenol molecules and the graphene sheet. There results are shown to be in accord with electronic density functional theory calculations, and literature experimental data. The contributions of different guest–guest (phenol–phenol) and guest-host (phenol and graphene) interactions between molecules are studied.

This paper dwells into the diffusional transport of water inside a pelleted zeolites wherein the transport behavior in binder phase as well as in the zeolite phase is explored. It is noted that both the phases contributes towards the separation process. A new variant of bi-disperse diffusion model is proposed for the water transport inside the adsorbent pellets. The equilibrium uptake of water is separately encountered in the binder and zeolite phases using a dual site Henry-Langmuir isotherm and a thermodynamically consistent correlation is developed for the boundary layer concentration of the zeolite phase with respect to the concentration within the binder phase. Adsorption kinetics measurements are reported for various adsorbent sizes and process temperatures, and their influence on transport diffusivity is explored. The proposed model prediction is in good agreement with the experimental observation in which the pellet diffusion controls the overall transport behaviour.

We extend the Maxwell-Stefan (M-S) formulation of irreversible thermodynamics to multicomponent transport in mixed-matrix membranes (MMMs), using a simulation-based rigorous modeling approach (SMA) through finite-element method (FEM) solution of the three-dimensional (3-d) transport problem in full-scale MMMs. In the new approach, we generalize the dual-mode/partial immobilization (DM/PI) theory for the local permeability in glassy polymers to describe multicomponent permeation in pure glassy polymer membranes and MMMs, by reformulating the M-S constitutive equations in the Onsager formalism considering concentration-dependent transport diffusivities and non-uniform concentration gradients across the MMM. In this way, the new M-S formulation explicitly considers effects of intrinsic MMM features such as finite filler particle size and isotherm nonlinearity in the MMM constituent phases, as well as mixture-related effects, such as competitive adsorption and friction amongst permeants, on the calculation of the mixture fluxes (permeabilities). This is achieved without introduction of empirical fitting parameters in the MMM permeability calculation and only requiring single-gas experimental or simulation-based adsorption and permeation data on the individual MMM materials to predict the mixture perm-selectivity in the MMM as a whole. Further, we validate the new approach by using available experimental permeation data for the separation of an equimolar binary mixture of propylene ( C 3 H 6 ) and propane ( C 3 H 8 ) in ZIF‑8/PIM-6FDA-OH MMMs, with the rigorous simulation results showing very good agreement with both experimental single and mixed-gas permeabilities and perm-selectivities.

This work investigates the effect of condensable Hg vapour on the transport of N-2 gas across cobalt oxide silica (CoOxSi) membranes. Experimental results suggest that Hg significantly affects N-2 permeation at 100 and 200 degrees C, though this effect is negligible at 300 degrees C. This effect was found to have a correlation with Hg adsorption on CoOxSi xerogels. In order to understand the Hg effect in the transport phenomena of N-2 permeation, the oscillator model was used to model gas transport through pores with different sizes. By including effective medium theory (EMT), the oscillator model fitted well the experimental results and gave good prediction of mass transfer in ultra-microporous materials with a tri-modal pore size distribution, such as silica membranes. It is postulated that Hg seeks lower level potentials in micro-pores, and therefore Hg molecules tend to block small pores (2.5-4 angstrom from 2.9 angstrom), or reduce the average pore size of larger pores (6.7-7.8 angstrom and 12-14 angstrom). Although N-2 permeation decreased with the presence of Hg, it did not decrease when the Hg load was increased by a factor of ten; this strongly suggests the adsorption of Hg molecules in the smaller pores (2.5-4.0 angstrom), or along the pore wall for the larger pore ranges (6.7-7.8 angstrom and 12-14 angstrom). (C) 2018 Elsevier Ltd. All rights reserved.

We present a new theory to estimate fluxes and effective transport conductances of binary mixtures through a membrane comprising a nonuniform porous medium with both pore size and pore length distributions, using the Onsager formulation at the single pore level. The theory defines a conductance of each species that is dependent on the concentration gradients of the various species, and on using effective medium theory determines the fluxes and concentration profiles self-consistently in the porous medium. The transport of CH4/H2 mixtures in a silica membrane having a known pore size distribution is examined using this theory, and the results compared with those from rigorous simulations, showing good agreement. It is found that an optimal network coordination number exists at which species fluxes are a maximum, due to the opposing effects of increasing porosity and mean pore length with increase in coordination number. Further, network fluxes decrease with increase in pore dispersion, indicating that uniform pore size is optimal. A species and pressure-dependent optimal temperature is also predicted, due to the competing effects of increase in diffusivity and decrease in adsorption on increasing temperature. It is seen that the CH4 selectivity is very sensitive to temperature, and undergoes a cross-over, with the membrane being more selective to CH4 at low temperature and to H2 at high temperature. In general, the selectivity is very sensitive to the distribution of pore volume, and for bimodal pore networks, undergoes a sharp transition at the percolation threshold, when the smaller pore size is impermeable to the larger species, CH4. The approach offers a convenient adaption of effective medium theory to multicomponent systems with nonlinear isotherms, overcoming drawbacks of existing theory.

We investigate the structure of polyimide (PI) at the surface of a silicalite zeolite (MFI), as part of a model hybrid organic-inorganic mixed matrix membrane system, through equilibrium molecular dynamics simulations. Furthermore, we report a comparison of the adsorption and transport characteristics of pure components CO2 and CH4 in PI, MFI, and PI-MFI composite membranes. It is seen that incorporation of MFI zeolite into PI results in the formation of densified polymer layers (rigidified region) near the surface, having thickness around 1.2 nm, before bulklike behavior of the polymer is attained, contrary to empirical fits suggesting the existence of an approximately 1 μm thick interface between the polymer and filler. This region offers an extra resistance to gas diffusion especially for the gas with a larger kinetic diameter, CH4, thus improving the CO2/CH4 kinetic selectivity in the PI-MFI composite membrane. Furthermore, we find that the kinetic selectivity of CO2 over CH4 in the rigidified region increases with temperature and that additivity of transport resistances in MFI, interfacial layer, and bulklike region of the polymer satisfactorily explains transport behavior in the composite sandwich investigated. The gas adsorption isotherms are extracted considering the dynamics and structural transitions in the PI and PI-MFI composite upon gas adsorption, and it is seen that the rigidified layer affects the gas adsorption in the polymer in the PI-MFI hybrid system. A significant increase in CO2/CH4 selectivity as well as gas permeability is observed in the PI-MFI composite membrane compared to that in the pure PI polymer membrane, which is correlated with the high selectivity of the rigidified interfacial layer in the polymer. Thus, while enhancing transport resistance, the rigidified layer is beneficial to membrane selectivity, leading to improved performance based on the Robeson upper bound plot for polymers.

Grand Canonical Monte Carlo (GCMC) simulations have been conducted to investigate the interaction between phenol and a commercial activated carbon cloth KIP1200 in aqueous solution. The nanostructure of KIP1200 has been extracted using Hybrid Reverse Monte Carlo (HRMC) simulations based on experimental measurements of XRD-based structure factor, and shows the presence of pore spaces inaccessible to both phenol and nitrogen. The impact of empirical force fields (EDIP and Erhart-Albe) on the structural parameters and adsorption capacity, as well as pore accessibility is investigated via molecular simulations and materials informatics employing lattice and Voronoi decomposition algorithms. While the Erhart-Albe force field leads to more amorphous structures and inaccessible pores, the EDIP force field creates more lamellar structures with sharp edges and narrow pore width. The molecular state of adsorbed phenol in these structures is markedly different, with the narrower pore spaces of the EDIP-based structure providing stronger interaction between phenol molecules and carbon structure than for water. The calculated X-ray intensities and adsorption isotherms for the simulation-based structure are in good agreement with corresponding experimental results. The simulated solvation state of phenol has been verified by experiments, and demonstrates the existence of inter-layer spaces inaccessible to phenol for both force fields.

The paper deals with the understanding of the CO2 capacity of adsorption through the anhydrite cement compound, i.e., CaO·SiO2 structure. The Wollastonite structure is prepared with various CaO/SiO2 ratios for the atomistic simulations. Subsequently, simulations of binary adsorption of H2O and CO2 have been performed to arrive at C–S–H structure. The CaO/SiO2 ratio is varied from 0.66 to less than 1.5 in the present study. The interlayer spacing between CaO sheets is considered as 13.5 A. The impact of the electrostatic charges has been explored using available CLAY_FF and CSH_FF potentials. The outcomes have been successfully examined and compared with experimental results. To understand the capacity adsorption of CO2 and selectivity effect on C–S–H forming, numerical experiments have been performed using CO2 concentrations from 10 ppm up to 1,000,000 ppm at 20 °C. These numerical experiments allow understanding of the selectivity of H2O and CO2 in binary mixture adsorption, and the capacity of CO2 adsorption in the Wollastonite atomic structure. According to the physisorption simulations, the CO2 adsorption capacity depends not only on CO2 concentration but also the Wollastonite structure. The possible CaCO3 precipitation and C–S–H structure are addressed.

We use grand canonical Monte Carlo simulation to investigate the adsorption of a CO2/N2 mixture in neutral and charged (7, 7) carbon nanotube (CNT) arrays. It is found that both the adsorption of CO2, and the CO2/N2 selectivity are either enhanced or reduced when the charges are positive or negative. The CO2/N2 selectivity in a CNT bundle carrying +0.05e charge with intertube distance of 0.335 nm exceeds 1000 for pressures up to 15 bar, which is remarkably high. It is seen that strong electrostatic interactions from neighbouring CNTs enhance the adsorption of CO2 over N2, and while the adsorption of CO2 has complex dependence on intertube distance, the CO2/N2 selectivity decreases with intertube spacing. We propose a quantitative performance coefficient as an aid to assessing the efficiency of CNT bundles to separate CO2 from flue gas, and show that a +0.05e charged bundle with intertube distance of 0.335 nm provides the best performance. Further, it is found that water vapor in flue gas imposes negligible effect on the adsorption of CO2 and its selectivity over N2 in the neutral and positively charged (7, 7) CNT bundles, but dramatically reduces the adsorption of CO2 and N2 in negatively charged bundles.

The adsorption of CO2 and H2O on divacanacy (DV) defected graphene cluster, and its bilayer counterpart is investigated using first-principles calculations. Both single and bilayer DV graphene cluster, are functionalised with H and F atoms. On these sheets the gas molecules are physisorbed, and the divacancy defect effectively improves the adsorption of CO2, while fluorination enhances the hydrophobicity of the graphene cluster. Among the convex and concave curvature regions induced due to the DV defect, the adsorption of the gas molecules on the concave meniscus is more favourable. Fluorine termination induces 73% reduction in Henry law constants for H2O, while for the CO2 molecule it increases by 8%, which indicates the DV defective sheet is a better candidate for CO2 capture compared to the STW defective sheet. Besides, both AA and AB divacant defect bilayer sheets are equally stable, wherein AA stacking results in a cavity between the sheets, while in AB stacking, the layers slide one over the other. Nevertheless, both these bilayer sheets are comparatively stabler than the monolayer. However, intercalation of lithium decreases the interlayer separation, particularly in AA stacking, which enhances the CO2 adsorption, but in the Bernal stacking enhances it hydrophobicity.

We develop a new approach for modeling titanium carbide derived-carbon (TiC-CDC) systems with residual titanium by the generation of modified atomistic structures based on a silicon carbide derived-carbon (SiC-CDC) model and the application of weighted combinations of these structures. In our approach, the original SiC-CDC structure is modified by (i) removing carbon, (ii) adding carbon and (iii) adding titanium. The new atomic scale carbide-derived carbon (CDC) structures are investigated using classical molecular dynamics simulations, and their pure CO2 adsorption isotherms are calculated using grand canonical Monte Carlo simulations. The system of TiC-CDC with residual titanium is modeled as weighted combinations of pure carbon CDC structures, CDC structures with titanium and a TiC crystalline structure. Our modeling is able to produce both structural properties and adsorption isotherms in accordance with experimental data. The fraction of different models in the systems successfully reflects the structural differences in various experimental TiC-CDC samples. The modeling also suggests that in partially etched TiC-CDC systems, the titanium that may be accessible to CO2 gas at the transitional interface may provide significant interaction sites for CO2 and may lead to more efficient overall gas adsorption.

The molecular orientation of permeating molecules crucially affects the gas transfer behavior in kinetic molecular sieving, particularly if the molecules are linear; however, this effect has been largely overlooked in the literature. To investigate this effect here we induce specific unfavourable molecular orientation of several linear gases based on adsorption effects in a hollow fiber NaA zeolite membrane, by anchoring a 3-dimensional alkali lignin molecule on the hydroxyls of the membrane surface. The results indicated that although the permeance was systemically reduced for all the gases investigated due to unfavourable orientation, the ideal selectivity for H2 relative to the linear molecules was consistently enhanced due to the presence of the alkali lignin molecule. On the other hand the permeation behavior for the spherically symmetric CH4 molecule was unaffected by the alkali lignin molecule, with the H2/CH4 unchanged and nearly constant over the operating temperature range. In addition to the defect healing, the enhancement in selectivity for H2 over the linear gases is strongly attributed to the normal orientation of the linear molecule relative to the pore axis on entry, induced by the adsorption force of the alkali lignin, leading to a repulsive potential at the pore mouth of the zeolite membrane. To validate the induced orientation effect of the linear molecules, ab initio method was employed to estimate the interaction energies between the pore mouth and linear molecules oriented parallel and normal to the pore axis. The computational results showed that the interaction energies were consistently higher for each specie when oriented normal to the pore axis (i.e. parallel to the pore cross-section), so that the tilted molecular orientation experienced stronger energy barriers; this is in accord with the experimental observation of increased selectivity for the more weakly-adsorbed H2 over the linear molecules. These findings can provide scientific guidance and facilitate new solutions to extend the separation boundaries of nanoporous membrane among gases.

International audience; We report the effect of fluorine doping on hydrophobicity of microporous silicon carbide-derived carbon (SiCDC), exploring water vapor adsorption in virgin and fluorinated samples experimentally and using semiempirical isotherm models. The surface chemistry of the virgin carbon and samples fluorinated to three different levels is characterized by X-ray photoelectron spectroscopy and 19F nuclear magnetic resonance analysis techniques. Besides having different textural features such as surface area and pore size distribution, the virgin and fluorinated SiCDCs show different surface chemistries in terms of the nature and quantity of the C–F bonds. The comparison of the characterization results of the samples and model parameters is used as the methodology to understand the water adsorption mechanism in virgin and fluorinated SiCDCs. We demonstrate that with increasing fluorination level the hydrophobic character of the low and medium fluorinated SiCDC samples remains almost constant while the highly fluorine-doped SiCDC is less hydrophobic. We identify a pore accessibility problem for argon in the fluorine-doped SiCDCs, demonstrating that fluorine functionalization blocks the carbon pores for the nonpolar argon while allowing polar H2O molecules to form clusters that enter the internal volume of the micropores. These findings provide important new understanding of the mechanism of water adsorption in microporous carbons and their fluorinated counterparts.

We have used nonequilibrium molecular dynamics simulations to investigate the transport diffusion of methane, at 300 K and pressures of up to 15 bar, in 30 nm-long (10, 10) carbon nanotubes (CNTs) held between two flanges mounted at the ends to represent the surface layers of an embedding matrix material. Strong interfacial resistance to the entry and exit of molecules is found in the 30 nm-long CNTs, which reduces their permeability by more than 2 orders of magnitude. Increasing the adsorption affinity and surface area of the flange reduces the interfacial resistance and consequently enhances the methane diffusivity in CNT membranes. Curved streamlines near the flange surface make a significant contribution to the permeability, even when the adsorption on the matrix surface is negligible. We propose a model to calculate the separate components of the interfacial resistance, the flange resistance, which increases with increase in the membrane porosity, and the entrance-exit resistance, which is independent of the membrane porosity. While the flange resistance accounts for the reduction of interfacial resistance with decrease in the membrane porosity, the entrance-exit resistance is responsible for the reduction of interfacial resistance with increase in the flange adsorption affinity. The flange resistivity demonstrates a complex dependency on the flange adsorption affinity, which is attributed to the competition between the enhanced adsorption and the enhanced migration time of the molecules on the flange. It is concluded that the embedding matrix adsorption affinity and membrane porosity separately play critical roles in determining the interfacial resistance and permeability in CNT membranes. Our simulation results can help reduce the interfacial resistance and improve the permeance in CNT membranes by appropriate choice of intertube spacing and flange material and are readily applied to all nanoporous membranes with a passive matrix.

We explore the effect of fluorine doping on hydrophobicity of nanoporous silicon carbide-derived carbon (SiCDC), and investigate the underlying barriers for adsorption and diffusion of water vapor and CO2 in the fluorinated and nonfluorinated structures. We develop atomistic models of fluorine-doped SiCDC at three different levels of fluorination, based on a hybrid reverse Monte Carlo constructed model of SiCDC, and develop a novel first-principles force field for the simulation of adsorption and transport of water and CO2 in the fluorine-doped carbon materials. We demonstrate an apparent dual effect of fluorination, showing that while fluorination generates more hydrophilic carbon surfaces, they actually act as more hydrophobic structures due to enhanced energy barriers in the disordered network of microporous carbon. While an increase in adsorption energy and in water uptake is seen for fluorine-doped carbon, large internal free energy barriers as well as the results of MD simulations demonstrate that the increased adsorption is kinetically limited and not experimentally observable on practical time scales. We show that an increase in apparent hydrophobicity due to fluorination is mediated by larger free energy barriers arising from stronger binding of fluid molecules inside the pore network, as opposed to repulsion or steric hindrance to the diffusion of molecules through narrow pore entries. For carbon dioxide, adsorption enthalpies and activation energy barriers are both decreased on fluorination, indicating weakened solid-fluid binding energies in the fluorinated systems.

We report a comparison of the adsorption of CH4 and CO2/CH4 mixtures of different composition in three different types of nanoporous carbons including carbon nanotubes, and activated carbon fibre (ACF-15) and silicon carbide derived carbon (SiC-DC) having distinctly different disordered structures, using Monte Carlo simulation. CO2 is represented as a linear molecule, and both the united-atom and full-atom models are investigated for CH4. It is found that the united-atom model of CH4 overestimates the adsorption of CH4 in all these adsorbents compared to the 5-site model, as a consequence of the enhanced 1-site CH4-adsorbent potential energy. Moreover, the selectivities of the nanoporous carbons for CO2 relative to CH4 calculated using the 1-site CH4 model are underestimated compared to those from the 5-site model, at pressures up to 3.0 MPa. However, differences in the structural disorder of porous carbon models have little impact on CO2 selectivity. Our simulations reveal that the selectivity of an adsorbent for a particular species is strongly dependant on adsorbate-adsorbate interaction effects, comprising the adsorbate-adsorbate potential interactions and an adsorbate sieving effect. As a balance between the confinement and adsorbate-adsorbate effects, it is found that increasing the concentration of CO2 in the gas phase increases the selectivity of (10, 10) CNT dramatically, while having negligible impact on the selectivities in amorphous carbons. Further, it is shown that increasing the temperature reduces the performance of all the carbons in separating CO2, and that an isolated (7,7) CNT has the best performance for CO2/CH4 separation in comparison to the disordered nanoporous carbons investigated. (C) 2014 Elsevier Ltd. All rights reserved.

We investigate the transport diffusion of methane at 300 K and pressures of up to 15 bar in dry and wetted carbon nanotubes (CNTs) having diameters ranging from 0.95 to 2.034 nm using nonequilibrium molecular dynamics (NEMD) simulation. Because of their strong hydrogen bonding, preadsorbed water molecules transport in the form of clusters and block the diffusion of methane, reducing the Onsager coefficient of methane dramatically compared to that in dry CNTs. The reduction in the methane Onsager coefficient is greater in narrower CNTs or at higher water densities. Because the diameter of the water clusters is almost invariant with water density, the Onsager coefficient of water in the (10, 10) CNT increases linearly with water density. It is further found that whereas decreasing the CNT diameter from 2.034 to 0.95 nm enhances the Onsager coefficient of pure methane by about 1 order of magnitude, the Onsager coefficient of water is almost independent of the CNT diameter at a water density of 0.05 g/cm3. We...

The adsorption of pure H2O, CO2/CH4, and H2O/CO2/CH4 mixtures in carbon nanotubes (CNTs) ranging from (6,6) to (15,15), and in a reconstructed model of a silicon carbide derived carbon (SiC-DC), was studied using Monte Carlo simulations, at 300 K. We also investigated the effect of preadsorbed water and the choice of water model including SPC, SPC/E, TIP3P, TIP4P, and TIP5P, on the adsorption of CO2/CH4 and H2O/CO2/CH4 mixtures in the CNTs and SiC-DC. Using the SPC model, our simulations reveal that, below saturation pressure, the adsorption of water is negligible in CNTs with diameters ranging from 0.81 to 2.03 nm, and in the SiC-DC. Water fills the CNTs and the SiC-DC model suddenly when the pressure reaches a critical value that is above the saturation pressure, and exhibits nonmonotonic variation with CNT diameter. It was also found that the small amount of water adsorbed from the saturated H2O/CO2/CH4 mixture has little impact on the adsorption of the components CO2 and CH4. When water is preadsorbed...

We investigate the effect of pore confinement and molecular geometry on the adsorption and self-diffusion of H2O, CO2, Ar, CH4, C3H6, SF6 and C5H12, in a realistic model of nanoporous silicon carbide derived carbon (SiC-DC), constructed using hybrid reverse Monte Carlo simulation. Adsorption isotherms, adsorbate–adsorbate and adsorbate–adsorbent contributions to the isosteric heat of adsorption are determined to study the effect of pore confinement, microporosity and molecular geometry on adsorption of these gases. We describe the cooperative effect of pore confinement and hydrogen bonding on the formation of water clusters and anomalous adsorption behaviour of water compared with non-polar gases. We find that, in contrast to literature results based on the slit-pore model, pore-filling does not occur below the saturation pressure in hydrophobic amorphous carbon materials such as SiC-DC and activated carbon fibre. We also compare self-diffusivities and activation energy barriers of water and non-polar gas...

We investigate macroscopic uptake kinetics of CH4 in silicon carbide-derived carbon (SiC-DC). Ultra-microprosity in SiC-DC is found based on CO2 adsorption at 273 K, but which has poor accessibility to Ar at 87 K. The adsorption kinetics of CH4 is found to follow a bidisperse pore structure model, considering relatively rapid particle scale diffusion in large micropores, and a much slower local grain (or microparticle) scale diffusion in ultra-micropores. The grain scale activation energies are comparable with values for carbon molecular sieves, and consistent with values expected for the size range of the ultra-micropores, while the activation energies for transport in the larger particle scale micropores are comparable to those for conventional activated carbons. The particle scale diffusivities compare well with the results of equilibrium molecular dynamics simulations using a hybrid reverse Monte Carlo simulation constructed model of SiC-DC, with similar activation energy. On the other hand microscopic quasi-elastic neutron scattering measurements are found to probe only short-range barriers with lower activation energy. It is anticipated that ultra-micropores will not make a significant contribution to the transport in any membrane or adsorption-based process based on SiC-DC, due to the extremely slow transport in these ultra-micropores and their small pore volume.

Ordered structures such as zeolites play an important role in a myriad of technological applications, from catalysis to adsorptive separations. Moreover, in recent years, a new class of zeolites with a “hierarchical” structure has been the subject of intense research, due to its potential to dramatically improve yield in those processes where the critical dimensions of the pore and the fluid are similar. In these hierarchical structures, a mesoporous network is purposefully embedded into the microporous framework; the presence of mesopores in the interior of microporous particles may significantly improve their transport properties. We analyze here our recent experiments on mass transfer in hierarchical CHA-like zeolite, microporous SAPO-34, having embedded mesopores as transport pathways providing access to intracrystalline micropores. We use effective medium theory (EMT), where no topological configuration of the phases is assumed other than they mix in an isotropic manner. Our results suggests that, wh...

We investigate gas adsorption and transport in a disordered microporous silica membrane having mean pore diameter 1.5 nm, coated on a porous tubular asymmetric support. The adsorption isotherms are found to be Langmuirian, with equilibrium constants that are accurately predicted for nonpolar gases, considering Lennard–Jones (LJ) interactions with a single layer of oxygen atoms on the pore surface. For the polar gas, CO2, the hydroxyls groups on the pore walls strongly increase the affinity with the pore walls, and a superposition of the LJ potential and an empirically represented electrostatic interaction is found to be adequate in correlating the Langmuirian equilibrium constant. The gas transport in the microporous silica layer is investigated using effective medium theory, with single pore transport represented by combination of pore mouth and internal pore diffusion resistances. Good agreement is observed for all the gases using different coordination numbers, indicating that the essential features of the transport in the silica micropores are captured in the approach. It is found that the overall transport resistance is dominated by the pore mouth barrier; however, the internal diffusion resistance in the relatively smaller pores is significant, especially for weakly adsorbed gases at higher temperature. In addition, the dependence of the pore mouth barrier coefficient on temperature and diffusing species are in good agreement with predictions of transition-state theory, with larger more strongly adsorbed molecules having higher activation energy. The proposed methodology is validated against experiment by comparison of the predicted flux for different gases in the supported membrane at various feed pressures in the low pressure range of 200–400 kPa, using the parameters obtained at 200 kPa.

We investigate the accessibility of Ti3SiC2-derived carbons (Ti3SiC2-DCs) synthesized non-isothermally using a temperature ramp. The microstructure of the Ti3SiC2-DCs is characterized using TEM, XRD, Raman spectroscopy and gas adsorption. For the characterization by gas adsorption, we adopt our Finite Wall Thickness (FWT) model to invert Ar adsorption isotherms at 87 K to obtain pore size and pore wall thickness distributions of the Ti3SiC2-DCs. Accordingly, we identify a pore accessibility problem in the Ti 3SiC2-DCs, as reported for Ti3SiC 2-DCs prepared at 1073 K in our previous work. A striking feature is that Ti3SiC2-DC prepared at the slowest ramping rate (2 K/min) has a very narrow pore size distribution, while the Ti 3SiC2-DCs synthesized at higher ramping rates (5 and 15 K/min) have much broader pore size distributions centered around 5.2 A. A significant amount of previously unreported ultra-microporosity is observed based on low pressure CO2 adsorption at 273 K. Our results indicate that slow ramping rate could potentially be utilized for fine control of the ultra-microporosity of carbide-derived carbons. Finally, we have found that fast ramping rate above 5 K/min leads to subtle changes in microstructure, with long and periodic graphitic multilayers having some large pores formed in between.

Strong interfacial interactions between the filler(s) and chosen polymer in mixed matrix membranes (MMMs) is a crucial factor in obtaining high gas separation efficiency. However, it is challenging to obtain excellent filler/polymer contact simply by direct incorporation of micron-sized metal-organic frameworks (MOFs) into a polymer matrix without modification. In this study micron-sized ZIF-67 particles were coated with a thin layer of 3 different ionic liquids (ILs) and characterized for the change in adsorption properties. The coated ZIF-67 particles were then incorporated into 6FDA-durene polymer at different loadings to fabricate MMMs for gas separation. Playing the role of interfacial binder, all ILs effectively enhanced the polymer/ZIF adhesion, minimizing the formation of non-selective interfacial defects, which was evidenced by the scanning electron microscopy (SEM) as well as by the focus ion beam scanning electron microscopy (FIB-SEM), leading to an increment in CO2/N2, CO2/CH4 and C3H6/C3H8 selectivity. The successful combination of ZIF-67/IL is proposed as a potential method to overcome the interfacial issues in MMMs, particularly in the application of larger micron-sized fillers.

The widespread developments in the use of nanomaterials in catalysis, adsorption, and nanofluidics present significant new challenges in achieving optimal adsorbed fluid flow characteristics. Here we demonstrate, using molecular dynamics simulations of nanoconfined fluids, that at nanoscales, fluid-solid friction is not restricted to a sharp interface as is commonly assumed; instead it is distributed over the whole adsorbed fluid phase, and is strongest in an interfacial region that is not negligible in comparison to the system size. Our simulations yield position-dependent dynamical fluid-solid friction coefficients, and lead to a modification of conventional hydrodynamics, incorporating distributed momentum loss in the fluid due to fluid-solid interaction. The results demonstrate that the usual concepts of slip length or interfacial friction coefficient are meaningful only for uniform fluids, and lose their significance for adsorbates in nanospaces, which are intrinsically inhomogeneous. We show that static friction coefficients, based on equilibrium density distributions, follow the same spatial dependence as the dynamical coefficients. These results open up possibilities for tailoring nanomaterials and surfaces to engineer low friction pathways for adsorbed fluid flow by tuning the potential energy landscape.

The adsorption of carbon dioxide in the presence of water in single-walled carbon nanotubes is studied using Monte Carlo simulation, at 300, 325, and 350 K. We also investigate the influence of the diameter and chirality of the nanotubes on the adsorption isotherms of CO2. It is observed that increasing the nanotube diameter from 1.36 nm (10, 10) to 2.03 nm (15, 15) leads to enhanced CO2 capacity, while change in chirality has little effect on the adsorption capacity of carbon nanotubes. Our results show that the influence of preadsorbed water on CO2 adsorption is dependent on both the effects of excluded volume and H2O–CO2 interactions. The maximum adsorbed amount of CO2 decreases linearly with the loading of water, and drops more rapidly in narrower nanotubes. The structure of water in hydrophobic nanopores is in the form of hydrogen-bonded clusters, and its adsorption does not affect the arrangement and orientation of CO2 molecules (i.e., it does not affect the mechanism of CO2 adsorption). The average...

We explore the temperature dependence of the self-, corrected-, and transport-diffusivities of CO2, CH4, and N2 in a polyethylene (PE) polymer membrane through equilibrium molecular dynamics simulations. We also investigate the morphology of the polymer membrane based on the intermolecular radial distribution function, free volume, and pore size distribution analysis. The results indicate the existence of 1.5–3 A diameter pores in the PE membrane, and with the increase in the temperature, the polymer swells linearly with changing slope at 450 K in the absence of gas and exponentially in the presence of gas. The gas adsorption isotherms extracted via a two-step methodology, considering the dynamics and structural transitions in the polymer matrix upon gas adsorption, were fitted using a “two-mode sorption” model. Our results suggest that CO2 adsorbs strongly, whereas N2 shows weak adsorption in PE. The results demonstrate that CO2 is more soluble, whereas N2 is least soluble. Further, it is found that an i...

International audience; We investigate the effect of fluorine-functionalisation of nanoporous silicon carbide-derived carbon (SiCDC), on its structural as well as adsorption properties, with the aim of assessing its suitability for CO2 capture. The morphology and structure of samples fluorinated to three different F/C ratios are characterized by several analysis techniques, as well as gas adsorption. Fluorination is seen to have a stabilising effect on SiCDC, and stronger C–F bonds are formed at high levels of fluorination. Further, increasing fluorination level leads to a decrease of specific surface area and total pore volume, consistent with recent simulation results from this laboratory. Fluorination has little effect on the ultra-microporosity at low levels of fluorination, but leads to significant decrease at high levels of fluorination. Sub-atmospheric pressure CO2 adsorption kinetics is interpreted using a bidisperse pore structure model, considering particle scale diffusion in large micropores and local grain scale diffusion with an interfacial barrier in ultra-micropores. The comparison of the CO2 uptake-time curves for the fluorinated and non-fluorinated samples shows slightly slower uptake with increasing fluorination level, largely due to decrease in pore volume and surface area. The isosteric heat of adsorption and activation energy barriers both decrease mildly with increased level of fluorination.

We determine the pore size distribution for five activated carbons (comprising carbide derived as well as commercial activated carbon samples) by the interpretation of experimental small angle neutron scattering (SANS) intensity profiles, based on the primary assumption of an infinitely dilute solution of hollow spherical particles. The interpretation yields the pore size distribution of the carbon samples that have predominantly micropore populations (size 20 A, which are not present in the PSD results based on adsorption isotherms of either Ar at 87 K or CO2 273 K. This inaccessible porosity becomes accessible to CO2 and Ar on heat treatment, leading to increase in the adsorption based pore volume. However, the surface area does not commensurately increase, indicating the inaccessible microporosity to predominantly comprise surface defects and roughness that are removed on heat treatment or activation. This finding sheds the light onto the evolution of porosity of activated carbons during gasification or post synthesis-treatment.

We report a comparison of the adsorption and transport characteristics of one-site and five-site molecular models of methane in silica nanopores, using grand canonical Monte Carlo and equilibrium molecular dynamics simulations. It is found that while the two models show similar effective molecular sizes, based on similar high-pressure densities in the bulk and nanopore fluids, the conventional parameters of the one-site model yield somewhat stronger intermolecular and pore wall interaction. This leads to higher densities in the bulk and adsorbed fluids at intermediate pressures, for the one-site model. However, the self- and collective-diffusion coefficients are similar for the two models for most nanopores, except at low densities in large mesopores. In this case, the five-site model shows slightly larger low-density diffusivity, due to its weaker interaction with the pore surface. On the basis of comparison with molecular dynamics simulations for the five-site model fluid, the predictive ability of our recent frictional theory of transport in nanopores is confirmed over a wide range of densities and pore diameters, using only the low-density diffusivity from a single simulation. Exceptions are found in the region of the critical point where the correlation length of the fluid diverges and when intermolecular interactions become significant in narrow nanopores where the fluid is nearly one-dimensional. In such cases, the local average density model used to estimate local transport properties becomes inaccurate.

The entry of water into hydrophobic nanospaces is critical to a variety of biological processes and in nanotechnologies for desalination and separation by nanofluidic devices. Idealized models of hydrophobic carbons have hitherto been used in simulations to investigate the anomalous adsorption of water, but the answer to the difficult question of how water enters such spaces has remained elusive. Here we show that while water entry does not occur in idealized independent carbon slit pores it is observed in realistic models of carbons having connected pore spaces. Good agreement with experimental water adsorption data is obtained for a realistic atomistic model of a disordered hydrophobic activated carbon fiber. Upon analyzing the adsorption in this atomistic model, and in another comprising adjacent slit pores connected by a small window in the separating wall, we find that the critical factor governing the behavior is the formation of sufficiently large and stable water clusters at windows or nanospaces connecting small and large pores. When this window size is large enough for the formation of a stable water cluster, condensation in a small pore induces the filling of empty large connected pores. This unique feature is not observed for nonpolar or weak polar gases (e.g., Ar or N-2) at subcritical conditions and explains why the Kelvin equation fails to estimate the condensation pressure for water.

Experiments of H2O and N2 adsorption in silicon carbide-derived carbon (SiC-CDC) are reported, showing only weak adsorption of N2, with H2O adsorption being highly kinetically restricted due to the strong hydrophobicity of this material. Such results suggest that SiC-CDC is an attractive option for adsorptive CO2, as our prior experiments have shown much more significant adsorption of this gas. This is confirmed by grand canonical Monte Carlo simulations, which show that in a 2-stage adsorption system the CO2 mole fraction can be increased from 15% (typical of flue gas) to 80–99.6% for the pore sizes that comprise the SiC-CDC.

Using a one-site spherical CH4 molecular model, we predict permeabilities in carbon slit pores of width 0.65–0.75 nm, based on adsorption isotherms obtained from grand canonical Monte Carlo (GCMC) simulation and diffusivities from equilibrium molecular dynamics (EMD). In the range of temperature of 298–318 K and pressure of 0.01–80 bar, the permeability within the slit pore is more strongly influenced by adsorption isotherms than by diffusion. It is found that the permeabilities predicted by the slit pore model far exceed those experimentally measured in practical membrane systems, consistent with our recent observations based on simulations of carbon dioxide permeability in slit pores. These results suggest that the pore entrance resistance (or pore mouth barrier), not considered in simulations of transport in semi-infinite slit pores, may be a dominating factor in nano-porous membrane transport.

We investigate the effect of heat treatment on the structure of carbide-derived carbons (CDC) prepared by chlorination from nanosized βSiC particles and on their methane as well as hydrogen storage and delivery performance. Pore size and pore wall thickness distributions of the CDCs are obtained from interpretation of argon adsorption data using the finite wall thickness (FWT) model. The adequacy of the FWT model for adsorption modeling in the SiC-CDC samples is demonstrated by satisfactory prediction of subatmospheric and high pressure adsorption isotherms of CO2 and CH4 at 313 and 333 K. From the characterization results, it is observed that the SiC-CDC particles are predominantly amorphous with slight graphitization of the external surface. The degree of graphitization is more pronounced in the sample prepared at 1000 °C and increases slowly with heat treatment time. During this time the accessibility of methane molecules is found to increase, as a result of short-range ordering and opening up of pore ...

A one-site spherical CO2 model and five three-site linear CO2 models are compared for adsorption in carbon slit pores, using grand canonical Monte Carlo (GCMC) simulation. A three-site CO2 model validated with bulk density experimental data is used to predict adsorption isotherms and diffusivities in the slit pore with the width from 0.65 nm to 0.75 nm. In the range of temperature from 298 K to 318 K and pressure from 0.01 bar to 20 bar permeability is calculated from the GCMC adsorption isotherms and the collective diffusivity obtained from equilibrium molecular dynamics (EMD). The permeability within the pore widths exceeds by three orders of magnitude that of reported macroscopic measurements. The simulation permeability obtained in this study is comparable to that derived from modified Knudsen diffusion with an energy barrier of 3.5–5 kJ/mol. The simulation results give an upper bound of the permeability for an ideal carbon membrane without pore mouth resistance and complex pore network.

Catalytic degradation of high density polyethylene (HDPE) using silica-alumina has been investigated in a thermogravimetric analyser, and the degradation kinetics determined using a population balance model recently developed in our laboratory. The incorporation of multisite adsorption into the model greatly improved the fit to experimental data. It is proposed that both thermal and catalytic cracking occur simultaneously, effectively through a two-step process: cracking of the large initial polymer molecules dominated by the catalyst with an activation energy of approximately 174 kJ/mol, followed by further breakage strongly influenced by thermal cracking with an activation energy of approximately 256 kJ/mol, so that it is the thermal degradation that is especially responsible for over-cracking and formation of gaseous products. In addition, it is found that the pre-exponential factor has a linear dependence on the catalyst loading. The breakage kernel used in the model allows for random scission, with mid-point being the most probable, so that the product distribution does not comprise a single peak moving smoothly through time-but peaks form at several discrete sizes. The model can predict product distributions at various conditions; however, as the model does not incorporate any specific mechanisms for adsorption and reaction, more direct investigation of the product distributions is also needed. This is of industrial importance as these products are economically attractive for the production of liquid fuels. The required reaction time can be predicted for a specific product distribution.

We investigate the characterization and adsorption modeling of Ti3SiC2-derived carbons (Ti3SiC2-DCs) prepared by chlorination at three temperatures (600, 800, and 1000 degrees C) as well as samples heat-treated at 1100 degrees C for one and three day periods. The modeling exploits our recent finite Wall thickness model-based density functional theory approach (Nguyen, T. X.; Bhatia, S. K. Langmuir 2004, 20, 3532), Utilizing experimental adsorption isotherms OF Ar at 87 K for the characterization. In general, characterization results show that the carbon wall structure of carbide-derived carbons (CDCs) is precursor-inherited, and Ti3SiC2-DCs have predominantly graphitic walls whose helium density is close to the true density of graphite (2.27 g/cm(3)), while SiC-derived carbons (SiC-DCs) have predominantly diamond-like walls whose helium density approaches the true density of diamond (3.52 g/cm(3)). This precursor-inherited character is also seen to give rise to two distinct evolutions of the Pore Structure of Ti3SiC2-DCs and SiC-DCs under chlorination temperature or post-heat-treatment conditions. In Particular, both pore enlargement and opening are observed for Ti3SiC2-DCs, but only pore opening, especially in the region of smallest pores (

We use path integral simulations to investigate the separation of H 2 and HD, as well as H 2 and D 2 , in carbon slit pores of various sizes using a new improved set of carbon-hydrogen interaction parameters determined in our laboratory. As expected, the selectivity of HD over H 2 is lower than that of D 2 over H 2 at the same conditions due to the smaller mass of HD molecules and hence larger quantum effects. In the pressure range of 0.1-10.0 bar, the selectivity is not sensitive to the pressure at the temperature of 77 K. At 40 K the selectivity shows a positive relation to the adsorbed phase density. We also report an unusual crossover effect in which the selectivity in a pore of width 0.85 nm exceeds that in a smaller pore (0.69 nm) at high densities due to enhanced quantum confinement effects when a second layer forms in the larger pore. The optimal pore widths for HD/H 2 separations were identified to be 0.56-0.57 nm, with operating pressures of 10.0 and 0.1 bar for the two pore sizes, respectively. We also simulate equilibrium separation in the commercial Takeda 3 A carbon molecular sieve, based on a slit-like pore model with a distribution of pore sizes, but find only modest equilibrium selectivity for HD over H 2 . It is suggested that while quantum effects are small within the pore bodies, narrow pore entrances must lead to significant quantum effects on the dynamics in order to explain literature data of faster uptake of D 2 compared to H 2 at 77 K in this material. Thus, kinetic molecular sieving at narrow necks, for which this material is well established, maybe a more attractive option than equilibrium separation. Alternatively, materials with controlled smaller pore sizes are needed for more efficient equilibrium HD/H 2 separation. The ideal adsorption solution theory (IAST) is also examined for prediction of the binary hydrogen isotope mixture isotherms in the presence of quantum effects and is found to match simulations at all operating conditions investigated.

We present a review of our recent studies oil the accessibility of simple gases (Ar, N-2, CH4 and CO2) in disordered microporous carbons using transition state theory (TST) and molecular simulation techniques. A realistic carbon model rather than the slit-pore approximation is utilised, providing more accurate understanding, of complex adsorption equilibrium and dynamics behaviour at the molecular level in porous carbons, especially kinetic restriction of adsorbate molecules through highly constricted pore mouths of coals and molecular sieve carbons (MSC). This kinetic restriction leads to a molecular sieving effect which plays a vital role in gas separation using the MSCs. In particular, the realistic carbon model of a saccharose char used in a recent study was obtained by hybrid reverse Monte Carlo simulation. The time of adsorption or desorption of the single gas molecule between two neighbouring pores through a highly constricted window of the realistic saccharose char model was determined using TST. Finally, the validation of TST calculated results of adsorption and desorption times against experimental measurements as well as molecular dynamics Simulation is also presented in this article. (C) 2009 Curtin University of Technology and John Wiley & Sons, Ltd.

Two Australian coals were heat-treated, and the accessibility of the pore space to CH4 and CO2 was investigated. Samples heat-treated at 573 and 673 K exhibit larger adsorption/desorption hysteresis and smaller surface areas (measured by CO2 adsorption at 273 K) than untreated samples. For samples heat-treated at 773 K, however, the surface area increased by 50% and the hysteresis was lower, compared to untreated samples. These results demonstrate that volatile hydrocarbons at pore mouths are the cause of energy barriers that prevent adsorbing molecules from passing through. A conceptual model is proposed to illustrate changes in activation energy at constricted pore mouths. Also, the results suggest that both adsorption and desorption isotherms should be measured to determine kinetically inaccessible pore spaces in order to correctly estimate CH4 recovery and CO2 storage capacity. The results have importance to the problem of estimating CH4 recovery and CO2 storage capacity for CO2 geosequestration as pa...

We deduce a new value for the potential well depth for the C-H2 interaction on the basis of experimental validations of isotherms of H2 and D2 predicted using independently characterized microstructural parameters. We use two carbons, one an activated carbon fiber whose structure has been recently characterized by us (Nguyen, T. X.; cohaut, N.; Bae, J.-S.; Bhatia, S. K. Langmuir 2008, 24, 7912) using hybrid reverse Monte Carlo simulation (HRMC) and the other the commercial Takeda 3A carbon molecular sieve whose pore size distribution is determined here from the 273 K CO2 adsorption isotherm. The conventional grand canonical Monte Carlo simulation technique incorporating a semiclassical Feynman and Hibbs (FH) potential approximation (FHGCMC) as well as path integral Monte Carlo calculations is employed to determine theoretical adsorption isotherms. It is found that curvature enhances the well depth for the LJ C-H2 interaction by a factor of 1.134 over that for a flat graphite surface, consistent with our recent study (Nguyen, T. X.; cohaut, N.; Bae, J.-S.; Bhatia, S. K. Langmuir 2008, 24, 7912). A value of the C-C well depth of 37.26 K, used for estimating the C-H2 well depth in conjunction with the Berthelot rules, with the Steele C-C well depth used for interaction with heavier gases (Ar, CO2 and CH4), leads to excellent agreement with experimental isotherms in all cases.

We present characterization results of silicon carbide-derived carbons (Si-CDCs) prepared from both nano- and micron-sized betaSiC particles by oxidation in pure chlorine atmosphere at various synthesis temperatures (600-1000 degrees C). Subsequently, the adsorption modeling study of simple gases (CH4 and CO2) in these Si-CDC samples for a wide range of pressures and temperatures using our Finite Wall Thickness model [Nguyen, T. X.; Bhatia, S. K. Langmuir 2004, 20, 3532] was also carried out. In general, characterization results showed that the core of Si-CDC particles contains predominantly amorphous material while minor graphitization was also observed on the surface of these particles for all the investigated synthesis temperatures (600-1000 degrees C). Furthermore, postsynthetic heat treatment at 1000 degrees C for 3 days, as well as particle size of precursor (betaSiC) were shown to have slight impact on the graphitization. In spite of the highly disordered nature of Si-CDC samples, the adsorption modeling results revealed that the Finite Wall Thickness model provides reasonably good prediction of experimental adsorption data of CO2 and CH4 in all the investigated Si-CDC samples at the temperatures of 273 K, 313 K, and 333 K for a wide range of pressure up to 200 bar. Furthermore, the impact of the difference in molecular size and geometry between analysis and probing gases on the prediction of the experimental adsorption isotherm in a disordered carbon using the slit-pore model is also found. Finally, the correlation between compressibility of the Si-CDC samples under high pressure adsorption and their synthesis temperature was deduced from the adsorption modeling.

The carbon slit pore has been investigated as a medium for low temperature hydrogen isotope mixture separation. It is shown that the path integral formalism with Silvera–Goldman potential provides the most accurate results for the adsorption simulations. At 40 K, an operating pressure of 1.0 bar and carbon slit width of 0.56 nm is found to be optimal for efficient quantum separation. Simulations for practical separation processes using a 99.95% H2 and 0.05% D2 bulk composition H2/D2 mixture, using the optimal carbon slit pore at 40 K and 1.0 bar, demonstrate that the deuterium mole fraction increases from 0.05 to 50.3% after three separation steps. The results show highly efficient equilibrium separation of binary H2/D2 mixtures in narrow carbon slit pores, indicating the potential of molecular sieving carbon materials for hydrogen isotope separations.

Recently (Nguyen and Bhatia, J. Phys. Chem. C 111:2212-2222, 2007) we have proposed a new algorithm utilising cluster analysis principles to determine pore network accessibility of a disordered material. The algorithm was applied to determine pore accessibility of the reconstructed molecular structure of a saccharose char, obtained in our recent work using hybrid reverse Monte Carlo simulation (Nguyen et al., Mol. Simul. 32:567-577, 2006). The method also identifies kinetically closed pores not accessed by adsorbate molecules at low temperature, when their low kinetic energy cannot overcome the potential barrier at the mouths of pores that can otherwise accommodate them. In the current work, the results are validated by transition state theory calculations for N-2 and Ar adsorption, showing that N-2 can equilibrate in narrow micropores at practical time scales at 300 K, but not at 77 K. Large differences between time scales for micropore entry and exit are predicted at low temperature for N-2, the latter being smaller by over three orders of magnitude. For N-2 at 77 K the time constant for pore entry exceeds 3 hr., while for exit it is 134 days. At 300 K these values are smaller than 1 mu s, indicating good accessibility at this temperature. These results are verified by molecular dynamics simulations, which reveal that while N-2 molecules enter and leave all pores frequently at 300 K, entry and exit events for apparently inaccessible pores are absent at 77 K. For Ar at 87 K better accessibility is evident for the saccharose char compared to N-2 at 77 K. This finding is now experimentally shown in this work by comparison of pore size distributions obtained from experimental nitrogen adsorption isotherms of nitrogen and argon at 77 K and 87 K.

The idea that increasing the enthalpy of adsorption increases the adsorptive capacity of carbon and makes it a better storage material for hydrogen is examined here considering the entire adsorption-desorption cycle. Structural modifications of carbon are examined to reveal the complex relationships between the enthalpy of adsorption, the pore volume, and the amount of hydrogen delivered over the course of a single cycle. The results provide an understanding of the connection between enthalpy and effective storage capacity in carbon materials and serve as a guide toward the search for an adsorbent which satisfies the DOE targets. Extensive GCMC simulations show that carbons having single graphene walls are optimal for hydrogen storage and that attempts to increase the enthalpy of adsorption either by increasing the wall thickness or by decreasing the pore size are detrimental to adsorptive capacity over a complete cycle from charging to exhaustion. It is found that carbon nanotubes display the same trend as slit pore carbons. The search for an adsorbent suitable for hydrogen storage should be aimed at the discovery of an entirely new high-capacity adsorbent with an enthalpy of adsorption of 15 kJ/mol, intermediate between that of carbon (4-6 kJ/mol) and metal hydrides (30-75 kJ/mol).

We propose a new algorithm based on application of cluster analysis to group adsorbate molecules of a highly dense adsorbed phase in the atomistic structural model of a disordered material into connected and disconnected clusters, through which pore network connectivity of the material is identified. Our proposed algorithm is then validated using a synthetic pore structure, as well as the reconstructed structure of a saccharose char obtained in our recent work using hybrid reverse Monte Carlo simulation.1 The algorithm also identifies kinetically closed pores in the latter structural model that are not accessed by adsorbate molecules at low temperature, at which their kinetic energy cannot overcome potential barriers at the mouths of pores that can otherwise accommodate them. The results are validated by transition state theory calculations for N2 and Ar adsorption, showing that N2 can equilibrate in narrow micropores at practical time scales at 300 K, but not at 77 K. Large differences between time scale...

This paper presents a detailed study of the air reactivity of petroleum cokes measured at temperatures between 400 and 600 °C using a combination of characterization techniques and reactivity measurements. The microstructure of the coke was found to comprise an essentially inaccessible pore system at low temperatures of 77-273 K used in characterization, and it is more accessible to oxygen at higher temperatures of about 773 K used in oxidation. The correlation of reactivity data using the random pore model suggests that the true micropore area is significantly larger than that measured using physical gas adsorption methods. The difference in surface area can be attributed to the low kinetic energy of gas molecules at the lower temperatures of characterization; as a result, they are unable to overcome the pore mouth energy barrier. By examining the variation of coke structure with burnoff level, we find that most of the internal reaction occurs in pores in the narrow pore width range 1-2 nm. For pores greater than 2 nm, however, the surface areas remains essentially constant with burnoff level. The apparent activation energy of the coke-air reaction derived from the extracted rate constants falls in the range 145-160 kJ/mol.

Adsorption isotherms of methane and carbon dioxide on two kinds of Australian coals have been measured at three temperatures up to pressures of 20 MPa. The adsorption behavior is described by three isotherm equations: extended three-parameter, Langmuir, and Toth. Among these, the Toth equation is found to be the most suitable, yielding the most realistic values of pore volume of the coals and the adsorbed phase density. Also, the surface area of coals obtained from CO2 adsorption at 273 K is found to be the meaningful parameter which captures the CO2 adsorption capacity. A maximum in the excess amount adsorbed of each gas appears at a lower pressure with a decrease in temperature. For carbon dioxide, after the appearance of the maximum, an inflection point in the excess amount adsorbed is observed close to the critical density at each temperature, indicating that the decrease in the gas-phase density change with pressure influences the behavior of the excess amount adsorbed. In the context of CO2 sequest...

We present results of the reconstruction of a saccharose-based activated carbon (CS1000a) using hybrid reverse Monte Carlo (HRMC) simulation, recently proposed by Opletal et al. [1]. Interaction between carbon atoms in the simulation is modeled by an environment dependent interaction potential (EDIP) [2,3]. The reconstructed structure shows predominance of sp2 over sp3 bonding, while a significant proportion of sp hybrid bonding is also observed. We also calculated a ring distribution and geometrical pore size distribution of the model developed. The latter is compared with that obtained from argon adsorption at 87 K using our recently proposed characterization procedure [4], the finite wall thickness (FWT) model. Further, we determine self-diffusivities of argon and nitrogen in the constructed carbon as functions of loading. It is found that while there is a maximum in the diffusivity with respect to loading, as previously observed by Pikunic et al. [5], diffusivities in the present work are 10 times lar...

The microstructural variation of Norit R1 Extra activated carbon, progressively heated at 1373 K, was explored in terms of pore size and pore wall thickness distributions, for various periods of heating time, determined by argon adsorption at 87 K, both using an infinite as well as and finite wall thickness model. The latter approach has recently been developed in our laboratory and has been applied to several virgin carbons. The current results show significant variations in small pore size regions (