No abstract

Density functional theory ͑DFT͒ calculations are used to characterize the interaction of mercury with copper, nickel, palladium, platinum, silver, and gold surfaces. Mercury binds relatively strongly to all the metal surfaces studied, with binding energies up to ϳ1 eV for Pt and Pd. DFT calculations underestimate the energy of adsorption with respect to available experimental data. Plane-wave DFT results using the local density approximation and the Perdew-Wang 1991 and Perdew-Burke-Ernzerhof parametrizations of the generalized gradient approximation indicate that binding of mercury at hollow sites is preferred over binding at top or bridge sites. The interaction with mercury in order of increasing reactivity over the six metals studied is Ag Ͻ AuϽ CuϽ NiϽ PtϽ Pd. Binding is stronger on the ͑001͒ faces of the metal surfaces, where mercury is situated in fourfold hollow sites as opposed to the threefold hollow sites on ͑111͒ faces. In general, mercury adsorption leads to decreases in the work function; adsorbate-induced work function changes are particularly dramatic on Pt.

A new theoretical method was developed to compute the Henry's law constant for gas absorption in a solvent through strong nonphysical interactions. The new method was created by expanding the test particle insertion method typically applied to physisorbing systems to account for the strong intermolecular interactions present in chemisorbing systems. By using an ab initio (AI)-based Boltzmann-averaged potential to model the interaction between CO2 and the tetra-n-butylphosphonium acetate ([P4444][CH3COO]) ionic liquid, the total Henrys's law constant at 298 K was computed to be 0.011 to 0.039 bar, reasonably comparable to the experimental value of 0.18 bar measured in this work. Three different AI potentials were used to verify the applicability of this approach. In contrast, when a classical force field (FF) was used to describe the interaction between CO2 and [P4444][CH3COO], the Henry's law constant was computed to be 27 bar, significantly larger than the experimental value. The classical FF underestimates the CO2-[P4444][CH3COO] interaction compared with the AI calculations, which in turn leads to the smaller simulated CO2 solubility. Simulations further indicate that the CO2 interaction with the [CH3COO](-) anion is much stronger than with the [P4444](+) cation. This result strongly suggests that the large CO2 solubility in [P4444][CH3COO] is due to the strong CO2-[CH3COO](-) interaction.

The performance of [emim][CH(3)COO] ionic liquid (IL) to separate mixtures of CO(2) and H(2) is studied using both classical and ab initio simulation methods and experiments. Simulations show that H(2) solubility and permeability in [emim][CH(3)COO] are quite low with Henry's law constants about 1 × 10(4) bar and permeabilities in the range 29-79 barrer at 313-373 K. In the case of CO(2) absorption in [emim][CH(3)COO], ab initio molecular dynamics simulations predict two types of CO(2) absorption states. In type I state, CO(2) molecules interact with the [CH(3)COO](-) anion through strong complexation leading to high CO(2) solubility. The C atom of CO(2) is located close to the O atoms of the [CH(3)COO](-) anion with an average distance of about 1.61 Å. The CO(2) bond angle (θ(OCO)) is about 138°, significantly perturbed from that of an isolated linear CO(2). In type II state, the CO(2) molecule maintains a linear configuration and is located at larger separations (>2.2 Å) from the [CH(3)COO](-) anion. The weaker interaction of CO(2) with the [CH(3)COO](-) anion in type II state is similar to the one observed when CO(2) absorbs in [bmim][PF(6)]. Simulations further demonstrate that the [emim](+) cation competes with CO(2) to interact with the [CH(3)COO](-) anion. The predicted high CO(2) permeability and low H(2) permeability in [emim][CH(3)COO] are also verified by our experiments. The experimental CO(2) permeability in [emim][CH(3)COO] is in the range of 1325-3701 barrer, and high experimental CO(2)/H(2) permeability selectivities of 21-37 at 313-373 K are observed. We propose that by replacing [emim](+) cation with 1-butyl-1-methylpyrrolidinium ([PY(14)](+)) further enhancement of CO(2) solubility in [PY(14)][CH(3)COO] IL will be obtained as well as good performance to separate CO(2) and H(2).

Mixed matrix membranes are being studied for their potential use in post-combustion carbon capture on the premise that they could dramatically lower costs relative to mature technologies available today.

Slab and cluster models are used to study H 2 desorption from a single dimer of the Si(100)-2 × 1 surface. The cluster models are constructed using geometries obtained from slab-model optimizations. The largest cluster model considered, Si 89 H 62 , contains eight surface dimers and gives reaction and activation energies for desorption nearly identical with the slab-model values when the same electronic structure method is used. The barrier for H 2 desorption, calculated using the Si 89 H 62 cluster model and the Becke3LYP functional, is 64.3 kcal/mol. When this result is corrected for the effects of basis set expansion and vibrational zero-point energy correction, the barrier decreases to about 61.0 kcal/mol, which is only 4.0 kcal/mol greater than the observed desorption barrier.

No abstract

CO2 and H2 solubilities, CO2/H2 solubility selectivities, CO2 diffusivities, and solvent viscosities in 27 commercially available physical solvents at 298 K were calculated from molecular simulations using the CHARMM36 all-atom force field for most solvents, and the simulation results were compared with available experimental data. The van der Waals radius parameters for solvents were slightly tuned to reproduce the experimental solvent density. The simulated CO2 solubilities are comparable with the experimental data, with an average absolute difference of 28%. For the homologous compounds containing the −(OCH2CH2)– repeat unit, both simulated and experimental data show that CO2 solubility decreases when the number of repeat units is increased; CO2 solubilities in these homologous compounds exhibit almost a perfect positive linear correlation with the solvent free-volume fractions. The simulated H2 solubilities and CO2/H2 solubility selectivities are also comparable with the experimental data, with differences of 22% and 17%, respectively. The H2 solubilities in all solvents studied in this work correlate very well with the solvent free-volume fractions, exhibiting a positive linear correlation coefficient of 0.84. Additionally, simulations show that CO2 solubility decreases when the temperature is increased. In contrast, H2 solubility increases at elevated temperature, which is partly due to the increased solvent free-volume fraction at elevated temperature. Finally, although the viscosity difference tends to be large (30%–246%) between simulation and experiment, both simulated and experimental data exhibit a similar solvent viscosity trend. Furthermore, simulations show that CO2 diffusivities in solvents are very strongly correlated with the solvent viscosities and the relationship between them is given by D CO2 = (2.6 ± 0.3) × 10–9/ηsolvent 0.59 ± 0.03.