Nose thermostat total energy is rising

[comorado]

Hello,

I am running MD simulation of super cell (384) of Be atoms using Nose thermostat. I start with crystal structure. I discovered that my total energy is rising for the length of calculation (approx. by 5eV) as well as potential energy (SP) of the Nose thermostat (5eV). Also my atoms start to drift with the constant speed. I've been running the calculation for the long time (5000 steps each step 2 fs), and never got to stabilize the energy of Nose thermostat. I've tried to fix one of atom position and do the same run. Energy still is rising but not as much, center of mass of the system is oscillating with the frequency of about 600 steps. I was wondering if it is possible at all to change the behavior of Nose thermostat so it does not dump energy into the system indefinitely.

Thank you very much.

Egor Klevak
PhD candidate
University of Washington

[admin]

Display your POSCAR and INCAR.

[comorado]

INCAR

System = Be 6x4x4 repetition of orthogonal cell

PREC = Medium

ISMEAR = 1 ! changed to gaussian smearing
!SIGMA = 0.05
ISIF = 0
SIGMA = 0.172
NCORE = 8 ! number of cores

IALGO = 48 ! RMM-DIIS algorithm for electrons
LREAL = A ! evaluate projection operators in real space
NELMIN = 4 ! do a minimum of four electronic steps
BMIX = 2.0 ! mixing parameter
MAXMIX = 50 ! keep dielectric function between ionic movements

LPLANE = .TRUE.
NPAR = 64
LSCALU = .FALSE.
NSIM = 8

ISYM = 0
IBRION = 0
!NELMIN = 4
ALGO = Very Fast
ENCUT = 120
EDIFF = 1E-5

NSW = 600
SMASS = 1.00
POTIM = 2.00
TEBEG = 1000

ISTART = 1
ICHARG = 1

Poscar is rather large will post only part of it this is with one atom fixed in the postion
POSCAR

Be
13.744800
1.000000000000000 0.000000000000000 0.000000000000000
0.000000000000000 1.154698504161574 0.000000000000000
0.000000000000000 0.000000000000000 1.042197776613701
Be
384
Selective dynamics
Direct
0.055560000000000 0.166666666666667 0.062500000000000 F F F
0.138890000000000 0.041670000000000 0.062500000000000 T T T
0.055560000000000 0.083330000000000 0.187500000000000 T T T
0.138890000000000 0.208330000000000 0.187500000000000 T T T
0.222220000000000 0.166666666666667 0.062500000000000 T T T
0.305560000000000 0.041670000000000 0.062500000000000 T T T
0.222220000000000 0.083330000000000 0.187500000000000 T T T
0.305560000000000 0.208330000000000 0.187500000000000 T T T
0.388890000000000 0.166666666666667 0.062500000000000 T T T
0.472220000000000 0.041670000000000 0.062500000000000 T T T
0.388890000000000 0.083330000000000 0.187500000000000 T T T
0.472220000000000 0.208330000000000 0.187500000000000 T T T
0.055560000000000 0.416670000000000 0.062500000000000 T T T
0.138890000000000 0.291670000000000 0.062500000000000 T T T
0.055560000000000 0.333333333333333 0.187500000000000 T T T
0.138890000000000 0.458330000000000 0.187500000000000 T T T
0.222220000000000 0.416670000000000 0.062500000000000 T T T
0.305560000000000 0.291670000000000 0.062500000000000 T T T
0.222220000000000 0.333333333333333 0.187500000000000 T T T
0.305560000000000 0.458330000000000 0.187500000000000 T T T
0.388890000000000 0.416670000000000 0.062500000000000 T T T
0.472220000000000 0.291670000000000 0.062500000000000 T T T
0.388890000000000 0.333333333333333 0.187500000000000 T T T
0.472220000000000 0.458330000000000 0.187500000000000 T T T
0.055560000000000 0.166666666666667 0.312500000000000 T T T
0.138890000000000 0.041670000000000 0.312500000000000 T T T
0.055560000000000 0.083330000000000 0.437500000000000 T T T
0.138890000000000 0.208330000000000 0.437500000000000 T T T
0.222220000000000 0.166666666666667 0.312500000000000 T T T
0.305560000000000 0.041670000000000 0.312500000000000 T T T
0.222220000000000 0.083330000000000 0.437500000000000 T T T
0.305560000000000 0.208330000000000 0.437500000000000 T T T
0.388890000000000 0.166666666666667 0.312500000000000 T T T
0.472220000000000 0.041670000000000 0.312500000000000 T T T
0.388890000000000 0.083330000000000 0.437500000000000 T T T
0.472220000000000 0.208330000000000 0.437500000000000 T T T
0.055560000000000 0.416670000000000 0.312500000000000 T T T
0.138890000000000 0.291670000000000 0.312500000000000 T T T

---------------------------------------------

[admin]

Your system should be 6x4x4 repetition of the orthogonal cell (96 atoms).
The dimensions of the cell (7.6x9.3x8.0) and the atomic coordinates you attached
indicate that in your cell there are sites for 62 atoms. In the POSCAR
you define 384 atoms. Too much atoms in the cell corresponds to extremely
high pressure. Starting MD with such a cell leads to (numerical) explosion
with rapidly increasing temperature. Is this what you observe?

[comorado]

[quote="admin"]Your system should be 6x4x4 repetition of the orthogonal cell (96 atoms).
The dimensions of the cell (7.6x9.3x8.0)

Orthogonal cell for Be has 4 atoms, thus 6x4x4x4 = 386 atoms

I did not provided the full POSCAR because of the limit of characters one can have in the message on the forum.

The temperature equilibrates to the right value fairly fast and oscillates around it. The problem is the energy of the NOSE thermostat which never stops growing, additionally I have a drift of atoms in the system.

[admin]

Input parameters are OK. Already within the first picosecond of the simulation
you should get a variation of temperature and energy aroud some averaged value.
When you use ISTART/ICHARG=1
you should pay attention which WAVECAR/CHG are taken
as a start for the MD run. Also check velocities in POSCAR.