Ab initio molecular dynamics

A key feature of Qcore is the capability to perform efficient classical and quantized path-integral molecular dynamics (MD) simulations using on-the-fly electronic structure potential energy surfaces, including GFN-xTB, DFT and EMFT.

The minimal input for running an ab initio molecular dynamics simulation looks like this:

aimd(
  structure(molecule = water)
  n_steps = 1000
  time_step = 0.5 fs
  xtb()
)

This input corresponds to an MD trajectory simulation of water starting from a built-in geometry, with an integration step size of 0.5 fs and a total number of 1,000 integration steps (which gives a total simulation time of 0.5 ps). The energy and forces are evaluated at GFN-xTB level of theory. The initial velocities are set to zero by default, and no thermostat is applied.

Additional simulation options/subcommands can be found in the user manual. Here, we show another example to illustrate options/subcommands for running a thermalization trajectory using the Andersen thermostat:

aimd(
  structure(molecule = water)
  n_steps = 1000
  time_step = 0.5 fs
  gradient( xtb() )

  velocities(
    sampling = thermal
    temperature = 300 kelvin
  )
  thermostat(
    type = andersen
    temperature = 300 kelvin
  )
)

Here, we generate the initial velocities from Maxwell-Boltzmann distribution at 300 K, by specifying sampling = thermal in the velocities() subcommand. We apply the Andersen thermostat to the trajectory with target temperature of 300 Kelvin.