[CPMD-list] Test of NPT CPMD simulation

[junwang]

Dear All:

Now I am trying to understand how to implement a reasonable NPT CPMD simulation. So I took one hydrogen molecule in a cubic box as the simplest test example. Following the turtorial of CPMD (day2.pdf). In order to start a CPMD simulation I followd the instructions to optimize wavefunctions first(same as inp1). Then I added several lines into inp2 to start a NPT CPMD simulation with temperature controlled by rescaling velocities. The modified input file is shown below:
&CPMD
  MIRROR
  MOLECULAR DYNAMICS
  RESTART WAVEFUNCTION COORDINATES CELL
  EMASS
  200
  TIMESTEP
    2
  TEMPCONTROL IONS
    300  30
  PARRINELLO-RAHMAN NPT
  MAXSTEP
   10000
  TRAJECTORY SAMPLE
   5000
&END
&SYSTEM
  SYMMETRY
    1
  CELL
   8.0  1.0    1.0   0.0 0.0 0.0
  CUTOFF
   30.
  PRESSURE
   0.0
  ISOTROPIC CELL
&END
&DFT
  FUNCTIONAL LDA
&END
&ATOMS
*H_MT_LDA.psp
   LMAX=S
  2
     0.00000     0.00000     0.00000
     0.80000     0.80000     0.80000
&END

But I found several problems when I checked the output file:
1. The symmetry reported in the SUPERCELL section is TRICLINIC instead of original SIMPLE CUBIC. I don't know why this change happened because the original cell and positions of two hydrogen atoms will not change during the previous wavefunction optimization step.

************************** SUPERCELL ***************************
 SYMMETRY:                                              TRICLINIC

2.  The cell size does not change at all during the simulation although the total stress tensor keeps changing. The cell temperature is also zero all the time. It seemed that the barostat does not adjust the cell length at all. I don't know the reason.

 ****************************************************************
 *                      AVERAGED QUANTITIES                     *
 ****************************************************************
                           MEAN VALUE <x>   DEVIATION <x^2>-<x>^2
 ELECTRON KINETIC ENERGY    0.455337E-04             0.148296E-04
 IONIC TEMPERATURE                302.42                    14.65
 CELL TEMPERATURE                   0.00                     0.00
 DENSITY FUNCTIONAL ENERGY     -1.131362             0.477349E-03
 CLASSICAL ENERGY              -1.129926             0.466355E-03
 CONSERVED ENERGY              -1.129880             0.466823E-03
 NOSE ENERGY ELECTRONS          0.000000              0.00000
 NOSE ENERGY IONS               0.000000              0.00000
 ION DISPLACEMENT            1.04871                 0.669774
 CELL VOLUME                 512.000                  0.00000

   CELL PARAMETERS:
                         8.00000000     0.00000000     0.00000000
                         0.00000000     8.00000000     0.00000000
                         0.00000000     0.00000000     8.00000000

 ****************************************************************
   TOTAL STRESS (kB)
             -29.22596123       -1.07079616        0.70578807
              -1.07079616      -38.53390080       -0.88527846
               0.70578807       -0.88527846      -30.41372020
 ****************************************************************

Does anybody know the possible reasons? 

BTY, I am also interested to know the meaning of fictitious cell mass and how to set it to an appropriate value during NPT simulation. I have some experience on classical MD simulation. In classical MD simulation we don't have to consider the mass of a cell even in NPT simulation since the cell serves only as a boundary for us to calculate the volume and apply the periodic boundary condition. So we are only interested in cell shape and cell size and adjust them accordingly in a NPT simulation. It seemed that CPMD treats the cell as the additional degree of freedom and assign the cell a fictitious mass and solve the special equation of motion for the cell during a NPT simulation. For electrons the situation is the same. A fictitious mass is assigned to the electrons and the equations of motion for electrons are solved together with the equations of motion for ions(nuclus). A larger electronic mass will allow us to use a larger timestep but also brings the stronger coupling be!
 tween electrons and ions. That's the balance we have to choose for electronic mass. What is the effect of cell mass on a NPT simulation then?



Best regards,
Sincerely,
Jun WANG

Department of Chemistry
University of Nebraska-Lincoln
536 Hamilton Hall
Lincoln, NE 68588-0304
U.S.A.
E-mail:junwang at unlserve.unl.edu
       junwang at bigred.unl.edu