comb/comb3 thermostat does not work?

Hi, all

A quick question: pair style comb within nvt ensemble does not control the temperature. The simulation input is from the example folder (1500atom HfO2 case). Only tiny modifications.

The temperature keeps going up when using fix nvt. The temperature seems out of controll. I also tested the fix nve option and found that the log file is almost identical to the log of nvt ensemble. I suspect that the thermostat did not work.

My lammps version is 22Feb2018, compiled on linux 2.6.32. I hope the temperature fluctuate around the set value. (The initial velocity was set to a higher level just for test. ) Anything wrong in my input? The original input only has quite a few steps so I extended it to 2000 to test the thermostat. If needed, I can provide another comb3 simulation of pure water which is also facing the same problem.

Appreciate your help.

Luo

Here is the input: (ffield.comb is in the example/comb/ from lammps pakage )

##--------------input script----------------------------

units metal
atom_style charge
dimension 3
boundary p p p

read_data data.m-HfO2
#read_data data.t-HfO2
#read_data data.c-HfO2

mass 1 178.0
group type1 type 1
compute charge1 type1 property/atom q
compute q1 type1 reduce ave c_charge1
mass 2 16.00
group type2 type 2
compute charge2 type2 property/atom q
compute q2 type2 reduce ave c_charge2

pair_style comb
pair_coeff * * ffield.comb Hf O

neighbor 0.5 bin
neigh_modify every 10 delay 10 check yes

timestep 0.00020

thermo_style custom step temp etotal pe evdwl ecoul press vol lx ly lz xz c_q1 c_q2
thermo_modify norm yes

fix 1 all nvt temp 300.0 300.0 100.0
thermo 1

only modified these following lines from the example.

velocity all create 400.1 2398378 # 400K for test
run 1
fix 2 all qeq/comb 50 0.002 file fq.out # Nevery=50 for test
run 2000

#-------------few lines of data file: data.m-HfO2----------------------

1500 atoms
0 bonds
0 angles
0 dihedrals
0 impropers

2 atom types
0 bond types
0 angle types
0 dihedral types
0 improper types

0.0000000E+00 25.6420000000000 xlo xhi
0.0000000E+00 25.9570000000000 ylo yhi
0.0000000E+00 26.4845000000000 zlo zhi
0.0000000 -4.4669100 0.0000000 xy xz yz

Atoms

1 1 3.60000000 1.23385 0.22323 1.10705
2 1 3.60000000 -1.68054 2.81893 1.54140
3 1 3.60000000 -1.23385 -0.22323 -1.10705
4 1 3.60000000 0.78716 2.37247 3.75550
5 2 -1.80000000 0.07843 1.77546 1.78506
6 2 -1.80000000 -0.52512 4.37116 0.86339
7 2 -1.80000000 -0.07843 -1.77546 -1.78506
8 2 -1.80000000 -0.36826 0.82024 4.43351
9 2 -1.80000000 1.86781 3.94027 2.54781
10 2 -1.80000000 -2.31450 1.34457 0.10064
11 2 -1.80000000 -1.86781 -3.94027 -2.54781
12 2 -1.80000000 1.42112 -1.34457 5.19626
13 1 3.60000000 0.34047 0.22323 6.40395
14 1 3.60000000 -2.57392 2.81893 6.83830
15 1 3.60000000 -2.12723 -0.22323 4.18985
16 1 3.60000000 -0.10622 2.37247 9.05240
17 2 -1.80000000 -0.81495 1.77546 7.08196
18 2 -1.80000000 -1.41850 4.37116 6.16029
19 2 -1.80000000 -0.97181 -1.77546 3.51184

1491 1 3.60000000 15.70622 20.54237 20.08055
1492 1 3.60000000 17.72723 23.13807 24.94310
1493 2 -1.80000000 17.01850 22.54106 22.97266
1494 2 -1.80000000 16.41495 25.13676 22.05099
1495 2 -1.80000000 16.86164 18.99014 19.40254
1496 2 -1.80000000 16.57181 21.58584 25.62111
1497 2 -1.80000000 18.80788 24.70587 23.73541
1498 2 -1.80000000 14.62557 22.11017 21.28824
1499 2 -1.80000000 15.07227 16.82533 18.63979
1500 2 -1.80000000 18.36119 19.42103 26.38386

#-------------few lines of nvt.log--------------------

step temp totalEng PotEng …

1996 567.05406 -9.5651066 -9.6383552 3.0021992 -12.640554 -73714.101 17627.802 25.642 25.957 26.4845 -4.46691 3.3516661 -1.6758331
1997 567.26897 -9.5679657 -9.6412421 2.9993359 -12.640578 -73930.914 17627.802 25.642 25.957 26.4845 -4.46691 3.3516661 -1.6758331
1998 567.47909 -9.5671472 -9.6404507 3.0001521 -12.640603 -73749.118 17627.802 25.642 25.957 26.4845 -4.46691 3.3516661 -1.6758331
1999 567.68743 -9.5669257 -9.6402562 3.0003728 -12.640629 -73573.282 17627.802 25.642 25.957 26.4845 -4.46691 3.3516661 -1.6758331
2000 567.91019 -9.5686838 -9.642043 2.9986132 -12.640656 -73784.661 17627.802 25.642 25.957 26.4845 -4.46691 3.3513216 -1.6756608
2001 568.3258 -9.5707245 -9.6441374 2.9950956 -12.639233 -74176.304 17627.802 25.642 25.957 26.4845 -4.46691 3.3513216 -1.6756608
Loop time of 69.9351 on 8 procs for 2000 steps with 1500 atoms

#-------------few lines of nve.log--------------------

1996 567.17801 -9.5652134 -9.638478 3.0019796 -12.640458 -73759.62 17627.802 25.642 25.957 26.4845 -4.46691 3.3516525 -1.6758263
1997 567.38697 -9.568078 -9.6413697 2.9991113 -12.640481 -73978.366 17627.802 25.642 25.957 26.4845 -4.46691 3.3516525 -1.6758263
1998 567.58836 -9.5679222 -9.6412398 2.9992657 -12.640506 -73854.982 17627.802 25.642 25.957 26.4845 -4.46691 3.3516525 -1.6758263
1999 567.7853 -9.5668565 -9.6401996 3.0003317 -12.640531 -73561.401 17627.802 25.642 25.957 26.4845 -4.46691 3.3516525 -1.6758263
2000 567.99805 -9.5673814 -9.6407519 2.9998063 -12.640558 -73561.256 17627.802 25.642 25.957 26.4845 -4.46691 3.3513215 -1.6756607
2001 568.39966 -9.5708739 -9.6442963 2.9949391 -12.639235 -74190.638 17627.802 25.642 25.957 26.4845 -4.46691 3.3513215 -1.6756607

Hi Luo,

The Tdamp parameter of fix nvt is way too big. It should be equal to about 100 timesteps, yours is several orders of magnitude higher, so no wonder the thermostat works poorly. The higher the Tdamp parameter, the longer it takes to achieve desired temperature.
Additionally, a run with only 2000 steps is not enough, especially for such a small timestep.

Michal