The adsorption of water and water/X mixtures on model activated carbons will be studied using Grand Canonical Monte Carlo techniques with biased sampling as needed. Such simulations are very demanding because of the hydrogen-bonding between the various species; such bonds are relatively long lived, and long simulations are needed to ensure ergodic behavior. Large systems are also needed to avoid size effects. This project will study pure water on an activated carbon surface, in which oxygenated surface sites are attached to the surface and can H-bond to the H atom in the water molecules. The effect of variation in the density of sites, species of sites, and geometric arrangement of sites on the surface will be studied. In particular the effect of these variables on the absorption, filling mechanism, and heats of adsorption will be of interest. This will be followed by studies of water/X mixtures on such surfaces. Initially, X will be a hydrocarbon, methane or benzene. The pore filling mechanism for the hydrocarbon and water are expected to be completely different, so that variation of the site parameters is likely to have a decisive effect on the selectivity of the pore for one of the components, and selectivity inversion is likely to be seen. Later studies will include the case where X is acetic acid. Both benzene and acetic acid are of practical interest, since they are EPA priority pollutants. The selectivity mechanism for acetic acid will be subtle and difficult to predict in advance of the simulations, since both water and acetic acid can bond to the surface. Parallel codes will be developed for these simulation on the IBM SP/2 at Cornell.
