Viscosity and Critical Properties of n-Decane, n-Pentadecane and n-Eicosane using Molecular Simulation

Abstract

Studies involving heavy hydrocarbons have been steadily increasing due to the increase in the extraction of shale or heavy oil of which heavy hydrocarbons are a major constituent. Experimental studies are difficult to conduct on species found in heavy oil because of thermal instability that results in their decomposition at high temperatures. This is especially true in the case of determining critical properties. Additionally, extraction of pure species from heavy oil is difficult due to the high energy requirements. Since they are extrapolative in nature, existing correlations can be used but with a certain degree of inaccuracy. This study uses molecular simulation, mainly molecular dynamics, to predict viscosity, saturated liquid and vapor densities, and critical properties of three pure heavy n-alkanes, namely, n-decane, n-pentadecane and n-eicosane. The viscosity was predicted as a function of temperature using equilibrium molecular dynamics with the Green-Kubo relations and the AMBER force-field. Comparison with experimental data showed that the AMBER force-field predicts reasonably the viscosity of smaller molecules at higher temperatures. The percentage deviation for n-decane was 40% at 300 K and 25% at 500 K. On the other hand, for n-pentadecane and n-eicosane, the percentage deviations were 34.7% at 550K and 76.2% at 600 K, respectively. For the critical properties, which were determined from the simulated saturated liquid and vapor densities, AMBER, COMPASS and TraPPE force fields were employed. For all three force fields, the methodology of volume-expansion molecular dynamics was used, while the Gibbs ensemble Monte Carlo technique was used with the AMBER force field. The results were compared with experimental data, as well as with those from several correlations and equations-of-state. The results show that the TraPPE force field is the most accurate in predicting critical properties (less than 10% deviation), followed by AMBER (less than 20% deviation, except for the critical pressures for n-eicosane) and COMPASS (up to 24% deviation). Finally, a comparison was made for the needed computational time; COMPASS required the largest execution time, followed by AMBER and TraPPE.