Can i-PI project out the rotations and translations of an isolated molecule within a MD simulation box?

[Arpan Kundu]

Dear all,

I would like to perform PIMD simulations of an isolated molecule within a periodic box. I'll use a plane-wave based ab-initio code as a driver. However, if we cannot reset translations and rotations of the center of mass (COM) iteratively, MD will not properly sample the vibrational degrees of freedom. Is there any tool in i-PI so that I can iteratively reset COM velocities and angular velocities to zero during the progress of the simulation?

Thanks,

Arpan

[Michele Ceriotti]

You can fix the center of mass motion by including

 <motion mode='dynamics'>

      <fixcom> True </fixcom>

  ....

in the input file. 

There is no method implemented to constraint rotations, because it's not a well defined problem when you have multiple beads - do you zero the angular momentum of the centroid or that of the entire ring polymer?

On the other hand, I don't understand your statement that MD will not sample the vibrational dof properly. It will, and you can always post-process the trajectory to remove unwanted motion, if you need to and you know what you're doing.

Best

M

[Arpan Kundu]

Dear Prof. Ceriotti,

Thank you very much for telling me how to fix the center of mass motion. I am very new to PIMD simulations and I didn't know that for multiple beads constraining the rotation is not a defined problem. However, I would like to use PIMD mostly to compare the vibrational properties of molecules. In my previous experience with classical MD of small isolated molecules (eg one N2 molecule in a box), I saw the molecule is rotating very fast. Consequently, while the rotational temperature is way higher, the vibrational temperature is way lower than the target temperature. Moreover, in the VDOS calculated from the velocity autocorrelation, I found a peak ~70-100 cm-1 at ~1000K and this is probably not real since rotational wavenumbers are usually in the microwave region. This problem wasn't severe if I constrain the rotation. 

I haven't started PIMD simulations yet, however, I am trying to imagine what could be the effect of uncontrolled rotations of the molecules in the PIMD trajectory. As you suggested, projecting out the rotation by post-processing is, probably, the way to go. 

Thanks,

Arpan

[Michele Ceriotti]

I suggest that before you start running simulations you study a bit the theory of what you're going to do. 

You need to understand the difference between PIMD, RPMD, CMD, and then how rotations and vibrations are coupled for gas-phase molecules.

Check out at the very least 

http://doi.org/10.1146/annurev-physchem-040412-110122

http://doi.org/10.1063/1.4883861

http://doi.org/10.1063/1.3125009

[Arpan Kundu]

Dear Prof. Ceriotti,

Thank you very much for the valuable suggestion and for sending the literature. I will study them to get familiar with these methods.

Regards,

Arpan