[QE-users] Cell temperature in NPT simulations using cp.x
Jie Peng
jiepeng at umd.edu
Tue Apr 24 17:12:27 CEST 2018
Dear all:
I have been running Molecular dynamics simulations on NPT assemble of HfS2.
The starting configuration has already been relaxed by consecutive steps of
electron relaxation, ion relaxation, and cell relaxation. As CPMD in
quantum espresso uses Gamma point only wavefunction sampling, an accurate
calculation of structure and energy requires usage of a large supercell.
Therefore I carried out NPT simulations of a 2*2*2 and 3*3*3 HfS2
supercell, containing 24 and 81 atoms respectively. The input file and
results are attached below:
*For 2*2*2 supercell*
*&control*
* calculation='vc-cp',*
* restart_mode='restart',*
* tstress = .true.*
* tprnfor = .true.*
* ndr=54*
* ndw=55*
* nstep=40000*
* dt=10*
* etot_conv_thr=1e-6*
* forc_conv_thr=1e-3*
* ekin_conv_thr=1e-5*
* wf_collect=.true.*
* prefix='HfS2',*
* pseudo_dir='/home/jpeng/HfS2/potential'*
* outdir='./tmp/',*
* /*
* &system*
* ibrav= 4,*
* a=7.3058*
* c=11.3088*
* nat=24, ntyp= 2,*
* ecutwfc =50*
* vdw_corr='DFT-D',*
* ! lspinorb=.true.*
* ! noncolin=.true.*
* ! ecutrho=300*
* ! nbnd=14*
*! occupations='smearing'*
*! smearing='gaussian'*
*! degauss=0.01*
* ! nspin=2*
* ! starting_magnetization(1)=0.1*
*/*
* &electrons*
* electron_dynamics='verlet'*
*! electron_velocities='zero'*
* emass=400*
* emass_cutoff=1*
*/*
* &ions*
* ion_dynamics = 'verlet'*
* ion_temperature='nose'*
*! ion_damping=0.1*
* tempw=300*
* fnosep=6*
*! ion_nstepe=10*
* /*
* &cell*
* cell_dynamics = 'pr'*
*/*
*ATOMIC_SPECIES*
* Hf 178.49 Hf.pbe-mt_fhi.UPF*
* S 32.065 S.pbe-mt_fhi.UPF*
*ATOMIC_POSITIONS angstrom*
*(atom position list)*
*For 3*3*3 supercell*
*&control*
* calculation='vc-cp',*
* restart_mode='restart',*
* tstress = .true.*
* tprnfor = .true.*
* ndr=54*
* ndw=55*
* nstep=40000*
* dt=10*
* etot_conv_thr=1e-6*
* forc_conv_thr=1e-3*
* ekin_conv_thr=1e-5*
* wf_collect=.true.*
* prefix='HfS2',*
* pseudo_dir='/home/jpeng/HfS2/potential'*
* outdir='./tmp/',*
* /*
* &system*
* ibrav= 4,*
* a=10.9587*
* c=16.9632*
* nat=81, ntyp= 2,*
* ecutwfc =50*
* vdw_corr='DFT-D',*
* ! lspinorb=.true.*
* ! noncolin=.true.*
* ! ecutrho=300*
* ! nbnd=14*
*! occupations='smearing'*
*! smearing='gaussian'*
*! degauss=0.01*
* ! nspin=2*
* ! starting_magnetization(1)=0.1*
*/*
* &electrons*
* electron_dynamics='verlet'*
*! electron_velocities='zero'*
* emass=800*
* emass_cutoff=1*
*/*
* &ions*
* ion_dynamics = 'verlet'*
* ion_temperature='nose'*
*! ion_damping=0.1*
* tempw=300*
* fnosep=6*
*! ion_nstepe=10*
* /*
* &cell*
* cell_dynamics = 'pr'*
*/*
*ATOMIC_SPECIES*
* Hf 178.49 Hf.pbe-mt_fhi.UPF*
* S 32.065 S.pbe-mt_fhi.UPF*
*ATOMIC_POSITIONS angstrom*
*(atom position list)*
The temperature is set to 300K, the system is connected to a Nose-Hoover
thermostate to equilibrate. The pressure is set to 0.
I attache the results of temph and tempp printed out in the output files
below. The inset is a zoom in of the period I think the system has reached
equilibrium.
For 2*2*2 supercell
For a 3*3*3 supercell
There are lots of thoughts and doubts here:
1. Has the system reached equilibrium? I consider the system to be in
equilibrium as I see a stable variation of ionic temperature around the
desired temperature 300K, is it enough to support the conclusion that the
system is in equilibrium? I know that the total energy, kinetic energy of
the system are also oscillating around a constant value. What other
behaviors or quantities can further consolidate this conclusion?
2. What is the temph, or the fictitious cell temperature? I have read the
paper Bernasconi, M., G. L. Chiarotti, P. Focher, S. Scandolo, E. Tosatti,
and M. Parrinello. "First-principle-constant pressure molecular
dynamics." Journal
of Physics and Chemistry of Solids *56, no. 3-4 (1995): 501-505, *on
which I think the variable cell CPMD in quantum espresso is built on. There
is no fictitious cell temperature in the paper, but I infer that would be
the temperature corresponding to the kinetic energy of the cell, which is
sth like 1/2*KB*T=1/2*W*Tr(dh'/dt*dh/dt) where W is the fictitious cell
mass, h is the cell vectors. Is it correct? What is the unit for the
fictitious cell temperature? How come it can be as high as, for instance
8000 for a 3*3*3* supercell NPT simulation?
3. How to reduce variations of the temperature? Right now I have ionic
temperature varying between 150K to 550 K for a 2*2*2 supercell while 180K
to 420K for a 3*3*3 supercell. How do I reduce the temperature variations?
Any general suggestions?
4. The initial "jump" in cell and ionic temperature. In both NPT
simulations for 2*2*2 and 3*3*3 supercell, there is a huge jump in the
temperature values at the initial stage. However, I take a look at the
animation of the atom positions and cell shapes during the entire NPT
simulation, there is no significant change of lattice structure at the
initial state. This observation leads me to believe that it is kind of like
a pulse of kinetic energy injected into the system that quickly get
distributed among atoms equally. But is it normal for the ionic temperature
to reach as high as 5000K?
I know I have a long list of questions here. I would greatly appreciate if
you can provide any suggestions to me!!
Thank you very much in advance.
Best
--
------------------------------------------------------------------------------------------------------------------------
Jie Peng
PhD student
2134 Glenn Martin Hall, Mechanical Engineering, University of Maryland
College Park, Maryland, USA
Phone:(+1) 240-495-9445
Email: jiepeng at umd.edu
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20180424/35932401/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image.png
Type: image/png
Size: 14505 bytes
Desc: not available
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20180424/35932401/attachment.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image.png
Type: image/png
Size: 18208 bytes
Desc: not available
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20180424/35932401/attachment-0001.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image.png
Type: image/png
Size: 14265 bytes
Desc: not available
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20180424/35932401/attachment-0002.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image.png
Type: image/png
Size: 18166 bytes
Desc: not available
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20180424/35932401/attachment-0003.png>
More information about the users
mailing list