MedeA® GIBBS: Liquid - Vapor Pressure Curve of Methane

In modeling this system we are limited to a rather small number of particles, typically a few hundred to keep computational efforts within limits. In MedeA® GIBBS we therefore define two boxes containing the liquid and the gas phase respectively. Technically this means that we squeeze many more particles into a volume unit for the liquid phase than we do for the gas phase resulting in a higher starting density of articles in the box representing the liquid.

During the computation, we allow for random moves of the rotations and moves between the boxes. For each move into a new configuration a “penalty” function is calculated using forcefields and over the duration of the simulation a minimum energy configuration is established.

In this final or equilibrium configuration, particles are distributed over both the liquid and the gas phase, provided that the initial temperature lies in the range of liquid-vapor coexistence.

Now, let us get started with the MedeA® GIBBS code. In the rest of this document you will learn how to: - Assign a forcefield to a molecule using the MedeA® interface - Extract experimental liquid-vapor equilibrium data from the NIST webbook - Run MedeA® GIBBS to calculate the liquid-vapor curve of methane

Outline of procedure

• Build methane (CH4) molecule and assign a forcefield
• Get experimental data from NIST webbook
• Set up the conditions of the run in the GIBBS interface
• Analyze the result

... more details in the attached pdf files

Analyze the result

After completion, the calculated equilibrium properties are summarized in the file job.out. Similar to all other MedeA® jobs, you can access the output files through View and Control Jobs in the Job Control menu.

For a liquid vapor equilibrium calculation, job.out gives both initial and calculated values of the properties of the two phases. In our calculations we assigned phase 1 and phase 2 to be the liquid phase and gas phase respective. Now we are interested in the vapor pressure as a function of temperature, and thus need the calculated pressures of phase 2. In job.out, scroll to the

Calculated Results for Phase 2

The calculated vapor pressure is listed in job.out for each of the temperatures. The pressures are given in units of bar. The calculated values (converted in MPa) are plotted below together with the experimental values from the NIST webbook. The agreement is very good, but we were using excellent initial conditions: the experimental values.

Sensitivity analysis

Initial pressure:

To check the sensitivity of the computed equilibrium pressure on the initial density we rerun with initial pressures for the liquid/vapor phase randomly deviating up to 26 % from the experimental equilibrium pressures. We change the initial density of the vapor phase from {0.042, 0.144,0.372,0.806,1.582,3.027,7.803} to {0.05,0.12,0.4,0.9,2,2.5,7} mol/L respectively.

In addition we reduce the number of Monte Carlo steps per temperature to 100000. Apart from that we start with identical parameters to the first set of calculations. The resulting vapor pressure curve is shown above. Even though quite large deviations from experimental values of the initial pressures of the two phases were introduced, the calculated vapor pressure curve is in good agreement with experiment. Not surprising, the most difficult temperature to obtain good computed values for methane is 190 K since this is close to the critical point of methane.

Full Version (pdf): 

MedeA-Gibbs Liquid Vapor Pressure Curve of Methane.pdf