Date of Award

2012

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Chemistry

First Advisor

Jerome Delhommelle

Abstract

The experimental determination of the critical point of metals is inherently connected with difficulties of measuring large temperatures and pressures. On the other hand, that of straight alkane chains larger than ten carbon atoms and branched chains larger than eight atoms is burdensome due to decomposition of the molecules while in the vapor phase. Therefore molecular simulations stand as an appropriate alternative for determining the critical point and phase coexistence properties of these fluids. Most simulation methods that are apt for this task rely on particles insertion and deletion moves that can significantly complicate the simulation, since high density in the liquid phases restrains this type of move. In our simulation this issue is avoided by combining the Wang-Landau (WL) sampling method in the NPT ensemble with efficient techniques for simulating dense liquid regions, such as configurational bias Monte Carlo (CBMC) and hybrid Monte Carlo (HMC). We simulate the liquid-vapor equilibria curves of copper and branched alkanes, isobutane and isopentane, by combining the HMC technique with the WL sampling. The standard boiling points of n-alkanes such as eicosane, tetracosane and triacontane are simulated by implementing the CBMC technique in the WL method in order to sample efficiently the various conformations of the long chain molecules. For copper we obtain a critical temperature Tc=5695 ± 50 K, critical pressure Pc=1141 ± 100 bar and critical density Ρc=1.80 ± 0.03 g/cm3. All of these values lie within the range of experimental data. The vapor-liquid equilibria curves and critical points of the branched alkanes are in excellent agreement with experimental data as well as simulation results using the Gibbs ensemble Monte Carlo method. The boiling points of the above listed n-alkanes also show good match with experimental data, with deviations in the range of 2 to 3 %. The WL simulations in the NPT ensemble is a simple and robust method for establishing vapor-liquid phase diagrams, as a single simulation run is necessary to evaluate the properties at a given temperature and for a large variety of substances. Its reliability is proven by the good agreement between experimental and simulation results.

Included in

Chemistry Commons

Share

COinS