2.25. MedeA Viscosity: Reliable Momentum Transport Properties from Classical Simulations

download:pdf

Key Benefits of MedeA Viscosity

  • Automatic analysis including the fitting of results
  • Quick and easy validation based on graphs, reported fitting errors, and access to all intermediate results through the convenient web interface
  • Integrated with MedeA Forcefields for advanced forcefield handling and assignment

2.25.1. Equilibrium Molecular Dynamics (EMD)

The viscosity of a material can be determined using the equilibrium molecular dynamics method, also known as the Green–Kubo Method, which calculates integrates the autocorrelation function of the shear (off-diagonal) components of the stress tensor. The method is only applicable for homogeneous configurations, i.e., no defects, interfaces, multiple phases, etc. The required length of the simulation depends on the viscosity: higher viscosities require longer simulation times. Good approximations can be obtained for many organic systems, such as hydrocarbons, alcohols, ethers, esters, etc. A significant advantage is that only moderate system sizes are required.

../../_images/image00123.png

  • Method: Green–Kubo.

  • Time: Duration of the simulation run.

  • Time Step: Time step size employed in solving the equations of motion.

  • Sampling: Number of samples employed in performing averaging. This parameter does not affect dynamics.

  • Trajectory: Number of trajectory frames saved during the molecular dynamics calculation. This parameter does not affect dynamics.

  • Number of bins in autocorrelation functions: The default is 2000.

  • The spacing of points in the autocorrelation function: The default is 1 steps.

  • Multiplier for sampling time in successive autocorrelation functions: The default is 2.

  • icon-checkboxoff Also calculate thermal conductivity: An option to also compute the autocorrelation function of the heat flux and estimate thermal conductivity from the same trajectory.

    Note

    This option is only visible if you also have the MedeA Thermal Conductivity license.

    Hint

    This option reduces computational expense when you would like to calculate both the viscosity and thermal conductivity in a single stage!

After completing a simulation, results are written to Job.out. For example:

../../_images/image00221.png

Additionally, visualization of the autocorrelation function and the viscosity integral are available in the file {stage_id}_SSt_decay_average.gif. For example:

../../_images/image00319.png

2.25.2. Reverse Non-Equilibrium Molecular Dynamics Method (RNEMD)

The reverse non-equilibrium molecular dynamics method, also known as the Müller-Plathe [1] method, induces a flow gradient (shear velocity or momentum flux) by performing momentum swaps between positive and negative shearing atoms and monitoring the resulting velocity profile. Viscosity is then calculated using the slope of the induced shear velocity profile.

This method applies to all systems but requires elongated cells in the direction of flow. Higher viscosities, which arise from stronger interactions, require correspondingly longer cells. The effect of the cell cross-section and length should ideally be examined, and the momentum transfer rate may need to be optimized, requiring some user intervention. Zero shear rate viscosities may be obtained by performing several simulations, followed by extrapolation.

../../_images/image00415.png

The required parameters are:

  • Method: Muller-Plathe NEMD.
  • Direction of layers: Direction in which the shear velocity gradient should be established.
  • Number of layers: Divides the above direction into this many layers to calculate the shear velocity profile.
  • Buffer layers: Number of layers on and near the boundary that are not used in the final analysis.
  • Velocity sampling rate: Sample shear velocity profile every this many steps. The default value of 1000 can usually be accepted.
  • Group of atoms to swap: Defaults to all.
  • Momentum swapping frequency: Defaults to 1000.
  • No. of momenta swapped: Defaults to 1.

After completing a simulation, results are written to Job.out. For example:

../../_images/image00514.png

Additionally, visualization of the shear velocity profiles is available as {stage_id}_average_velocity_profile.png. For example:

../../_images/image00616.png
[1]Florian Müller-Plathe and Dirk Reith, “Cause and Effect Reversed in Non-Equilibrium Molecular Dynamics: an Easy Route to Transport Coefficients”, Computational and Theoretical Polymer Science 9, no. 3 (1999): 203-209.
download:pdf