Eduardo D. Glandt

Storage of natural gas by adsorption on activated carbon

Abstract Natural gas may be stored by liquefaction, compression, or adsorption. For use as a transportation fuel, liquefaction is impractical and compression requires high pressure (20 MPa) and an expensive multi-stage compression facility. At relatively low pressure (3–4 MPa) achievable by single-stage compression, adsorbed natural gas (ANG) has nearly the capacity of compressed natural gas (CNG). Monte Carlo calculations were performed to simulate the adsorption of natural gas on activated carbon. The model is pure methane intercalated between parallel planes of graphite at a slit of width 11.4 A, optimized for ANG storage. The simulations predict that the maximum delivered energy density of ANG is 0.25 for monolithic carbon and 0.17 for pelletized carbon, compared to 0.29 for CNG and 1.0 for gasoline.

Monte Carlo simulation of multicomponent equilibria in a semigrand canonical ensemble

A general formalism and methodology are presented for the Monte Carlo simulation of equilibria in multicomponent systems, and are applied to the study of phase equilibrium in a model binary mixture, and to phase and chemical equilibrium in the ternary mixture Br2−Cl2−BrCl. Very good agreement with available experimental data is obtained. The formalism is based upon an ensemble in which the ratios of all fugacities to a reference fugacity are imposed. Simulations of mixtures performed in this ensemble fluctuate in composition while keeping constant the total number of particles. The reference fugacity is computed by integration of simulation averages of the composition for varying values of the fugacity ratios. Several features of the approach are: (1) it can be readily applied to mixtures of any number of components, even to polydisperse systems, with little additional computational effort; (2) the chemical potential of only one species at one state need be evaluated by conventional techniques, regardless...

Distribution functions for fluids in random media

A random medium is considered, composed of identifiable interactive sites or obstacles equilibrated at a high temperature and then quenched rapidly to form a rigid structure, statistically homogeneous on all but molecular length scales. The equilibrium statistical mechanics of a fluid contained inside this quenched medium is discussed. Various particle-particle and particle-obstacle correlation functions, which differ from the corresponding functions for a fully equilibrated binary mixture, are defined through an averaging process over the static ensemble of obstacle configurations and application of topological reduction techniques. The Ornstein-Zernike equations also differ from their equilibrium counterparts.

Aggregation and percolation in a system of adhesive spheres

We have studied the aggregation and percolation behavior of a fluid of adhesive spheres by means of Monte Carlo simulations. Cluster statistics have been obtained as a function of temperature and density and compared with the virial results of Post and Glandt. The determination of the percolation threshold from computer simulations is discussed. The percolation threshold of the adhesive sphere system has been obtained as a function of temperature and compared with the Percus–Yevick solution of Chiew and Glandt. Agreement is good except where the Percus–Yevick solution would be expected to fail on physical grounds.

Statistical thermodynamics of polydisperse fluids

The statistical mechanical formalisms for dealing with mixtures containing an infinite number of components are discussed. The partition function and the thermodynamic properties can be easily generalized from the corresponding results for pure substances when the polydisperse fluid is studied in a four‐dimensional space. The chemical potential field acts along the additional coordinate: the composition axis. A perturbation theory for narrow distributions is developed. A small parameter measuring the width of the chemical potential function is used within a semigrand ensemble. An equivalent, hybrid expansion expresses the results in terms of the variance of the composition distribution.

Effective conductivity of dispersions: The effect of resistance at the particle surfaces

Abstract The effective thermal or electrical conductivity of a suspension of monodisperse spherical particles embedded in a matrix of different conductivity is computed for the case when there exists a finite resistance at the particle-matrix interface. The calculation is carried out to second order in the volume fraction θ of the particle phase for a dispersion having the structure of equilibrium hard spheres. Both the first-order (Maxwell) and the second-order approximations are found to represent experimental data reasonably well for dispersions of low to moderate concentration. To first order in θ, the calculation of the conductivity of a dispersion with surface resistance is equivalent to the calculation of the conductivity of another dispersion with no surface resistance, but with a different particle conductivity. In the case of polydisperse systems, the effective conductivity is found to be sensitive to the particle size distribution.

The effect of structure on the conductivity of a dispersion

Abstract The equilibrium hard-sphere fluid is used to model the structure of dispersions of identical impenetrable spheres within a matrix. Pair-correlation functions adjusted to Monte Carlo simulation results, and reported in the literature, are used to compute the contribution of pairs of spheres to the effective thermal (or electrical) conductivity of the dispersions. An improved form of Maxwells equation is proposed, which is correct to order φ2, where φ is the volume fraction of the dispersed phase. A comparison with experimental measurements shows good agreement over a wide range of conditions. The approach fails for highly concentrated dispersions of very conducting spheres. Alternative models are discussed which are appropriate in this limit.

Model microporous carbons : microstructure, surface polarity and gas adsorption

Abstract A new molecular model for activated carbons is presented, which consists of a collection of randomly oriented carbon crystallites. The carbon displays energetic heterogeneity as a result of its complex microstructure and surface polarity. Adsorption isotherms and isosteric heats of adsorption for methane, ethane and their binary mixtures were estimated by grand-canonical Monte Carlo simulations. Simulations of the adsorption of water vapor were performed to adjust the surface polarity of the adsorbent.

Fluids in equilibrium with disordered porous materials. Integral equation theory

The Ornstein–Zernike equations previously introduced by two of the authors for a fluid in equilibrium within a quenched disordered matrix are solved numerically within the Percus–Yevick approximation. The structure of a fluid of hard spheres is reported for two types of microporous matrices: a sintered‐type structure of mutually penetrable (randomly placed) obstacles or sites, and a packed bed of quenched hard spheres. The integral‐equation results agree well with Monte Carlo simulation data also reported here. For the case of point obstacles, when both models coincide, the structure of the fluid is found to be insensitive to obstacle concentration. The structure of a hard‐sphere fluid in a bed of other quenched hard spheres is found to be significantly different from that of the equilibrium binary mixture of the two types of particles. Pressures and bulk‐pore partition coefficients are reported for beds of randomly placed obstacles. A sparse‐matrix approximation is presented and compared with the full so...

Clustering and percolation in multicomponent systems of randomly centered and permeable spheres

For a multicomponent system of particles in equilibrium, an exact integral equation is derived for the pair connectedness function (which measures the probability that two particles, with centers a distance r apart, are connected). The pair connectedness function, mean cluster size, and percolation thresholds for mixtures of randomly centered (noninteracting) spheres and permeable spheres are then obtained analytically in the Percus–Yevick (PY) approximation. (The permeable‐sphere model provides a one‐parameter bridge from randomly centered sphere mixtures to hard‐sphere mixtures.) For this family of models, connectedness is defined by particle overlap. It is found that, within the PY approximation, a multicomponent mixture of randomly centered spheres percolates at ξ3=π/6∑iρiR3i =1/2, independent of the concentration and size distributions of the particles. For the permeable‐sphere case the percolation threshold depends on the relative densities, size, and interparticle permeability among the species. Th...