[QE-users] Atomic velocities units

[valeria arteaga muñiz]

Hello,

Thanks Kazume and Lorenzo for your time.
I think having them printed would be great for users, in the meantime I will calculate them with what you gave me.
And Kazume, I took a look to the module you mentioned (thanks for that, it was really useful) and I could see that the starting velocities therefore follow a Maxwell-Boltzmann distribution. That means that the system is already thermalized? I know this is not a technical question, but I was wondering so since when I run a MD simulation in pw.x in the past without inputting velocities, the variation in temperature was high at the beginning of the simulation.

Thanks,
Valeria

From: Kazume NISHIDATE<mailto:nisidate at iwate-u.ac.jp>
Sent: Monday, August 9, 2021 7:02 PM
To: Quantum ESPRESSO users Forum<mailto:users at lists.quantum-espresso.org>
Subject: Re: [QE-users] Atomic velocities units

Hello Valeria:

As the manual says that the 'dt’ is
  "time step for molecular dynamics, in Rydberg atomic units
  (1 a.u.=4.8378 * 10^-17 s : beware, the CP code uses
   Hartree atomic units, half that much!!!)”
  https://www.quantum-espresso.org/Doc/INPUT_PW.html

If you did not supply the initial velocities, the initial velocities
at the specified temperature are given by (dynamics_module.f90):

  ! ... N.B. velocities must in a.u. units of alat and are zero
  ! ...      for fixed ions
  !
         vel(:,na) = DBLE( if_pos(:,na) ) * &
                       gauss_dist( 0.D0, sigma, 3 ) / alat

So the unit is a.u. divided by 'alat'.

And the kinetic energy 'ekin' is calculated by (dynamics_module.f90):
         ekin  = ekin + 0.5D0 * mass(na) * &
                        ( vel(1,na)**2 + vel(2,na)**2 + vel(3,na)**2 )

then it is rescaled as
     ekin = ekin*alat**2
to recover the atomic unit in the above vel**2 operation.


> Another question; is there a way to grep or look for the velocities
>  in some kind of output file from a pw.x run?

Possible. We can extract the atomic positions at each MD time steps
from the output file. The velocities can be directly calculated by
(r(t2)-r(t1))/dt.

But I think QE may save the velocities at the final MD step in the
working dir to make it possible the restart run. To check this
you can calculate some short time MD run and investigate the directory
if there is the final velocity data.



> 2021/08/10 8:43、valeria arteaga muñiz <valitzelar at hotmail.com>のメール:
>
> Hello Lorenzo,
>
> Thank you again. I don’t know how could I fix that bug since the version of QE I am currently using is installed on a supercomputer, and I don’t think I have access to the main directories.
> Another question; is there a way to grep or look for the velocities in some kind of output file from a pw.x run?
>

best regards
kazume NISHIDATE
敬具　西館数芽

nisidate at iwate-u.ac.jp
kazume.nishidate at gmail.com










_______________________________________________
Quantum ESPRESSO is supported by MaX (www.max-centre.eu<http://www.max-centre.eu>)
users mailing list users at lists.quantum-espresso.org
https://lists.quantum-espresso.org/mailman/listinfo/users

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20210810/525ae4fd/attachment.html>