[lammps-users] Modified SW potential for Germanium

Hi all,

When I used the original SW potential for germanium, I got thermal conductivity of germainum about 3 times higher than experimental data at 300K.
When I used the modified SW potential from Zi Jian et al. Physics Review B 41(18), 1990, I simply fixed NPT and the volume of the system shrunk dramatically, from 179991 to 128213 while I expected the volume to expand a little bit.
Below is my files for the modified SW potential. Anyone could give me any suggestions? Thanks a lot.

Ge.sw:
Ge Ge Ge 1.93 2.181 1.80 8.37 1.2 -0.333333333333
6.768571075 0.616574299 4.0 0.0 0.0

INPUT fileļ¼š

atom_style atomic
units metal
dimension 3
boundary p p p
lattice diamond 5.64613
region box block 0 10 0 10 0 10
create_box 1 box
create_atoms 1 box
mass 1 72.64
pair_style sw
pair_coeff * * Ge.sw Ge
timestep 0.001
thermo 100

velocity all create 300 8746369 rot yes dist gaussian

fix NPT all npt temp 300 300 0.01 iso 0.0 0.0 0.1 drag 0.2
thermo_style custom step temp press vol
run 5000

log.lammps:
LAMMPS (21 Jul 2010)

atom_style atomic
units metal
dimension 3
boundary p p p
lattice diamond 5.64613
Lattice spacing in x,y,z = 5.64613 5.64613 5.64613
region box block 0 10 0 10 0 10
create_box 1 box
Created orthogonal box = (0 0 0) to (56.4613 56.4613 56.4613)
1 by 1 by 1 processor grid
create_atoms 1 box
Created 8000 atoms
mass 1 72.64
pair_style sw
pair_coeff * * Ge.sw Ge
timestep 0.001
thermo 100

velocity all create 300 8746369 rot yes dist gaussian

fix NPT all npt temp 300 300 0.01 iso 0.0 0.0 0.1 drag 0.2
thermo_style custom step temp press vol
run 5000
Memory usage per processor = 3.58011 Mbytes
Step Temp Press Volume
0 300 13226.263 179991.76
100 231.64939 -245.4156 181822.09
200 294.72366 -23.70232 179130.13
300 395.19915 -62.131671 174150.6
400 534.61021 -135.90145 166008.23
500 613.06629 -140.86055 157396.15
600 638.61054 -89.03642 149675.53
700 569.09468 -71.476272 143708.63
800 505.94911 -32.771998 139131.75
900 444.72706 -15.820057 135919.04
1000 406.05802 -4.595655 133641.61
1100 368.63888 0.64898138 132112.43
1200 348.9221 9.6140193 131005.54
1300 333.34873 4.6069046 130243.36
1400 324.44239 4.3980742 129732.99
1500 315.08547 1.510367 129415.74
1600 312.37419 1.8282896 129127.45
1700 307.10587 1.4928076 128925.18
1800 303.60216 1.4364289 128764.65
1900 302.66219 -3.2145116 128635.44
2000 303.08576 -5.342349 128558.15
2100 302.31333 1.6799552 128520.55
2200 300.90196 5.3982331 128499.28
2300 299.13194 -5.4892974 128466.09
2400 304.16606 -12.316301 128421.26
2500 300.83486 9.6020228 128371.91
2600 298.07815 11.340175 128340.67
2700 305.36257 -21.935451 128313.43
2800 300.81923 9.1833047 128293.9
2900 295.70154 27.904347 128302.7
3000 303.90131 -44.247171 128287.19
3100 300.37157 14.207614 128267.68
3200 297.80482 39.913662 128247.82
3300 306.28692 -81.211011 128263.77
3400 296.9978 65.604496 128211.67
3500 303.66717 15.258883 128197.61
3600 297.61202 -118.65753 128236.57
3700 303.24135 189.18132 128225.94
3800 299.73241 -161.67803 128220.35
3900 303.67796 22.259061 128197.92
4000 297.64737 190.07714 128196.83
4100 295.72924 -405.84336 128209.11
4200 303.01913 500.17018 128183.26
4300 297.57016 -384.52662 128207.16
4400 303.01119 75.887905 128197.95
4500 302.43886 409.24169 128176.56
4600 299.45105 -878.65758 128210.63
4700 301.91118 1158.8592 128151.89
4800 295.24188 -1135.9436 128244.9
4900 301.09536 776.75207 128156.52
5000 299.04652 -203.96205 128213.89
Loop time of 796.853 on 1 procs for 5000 steps with 8000 atoms
Pair time () = 791.239 (99.2955) Neigh time () = 1.04982 (0.131746)
Comm time () = 0.476744 (0.0598284) Outpt time () = 0.0021956 (0.000275534)
Other time (%) = 4.08537 (0.512688)
Nlocal: 8000 ave 8000 max 8000 min
Histogram: 1 0 0 0 0 0 0 0 0 0
Nghost: 7106 ave 7106 max 7106 min
Histogram: 1 0 0 0 0 0 0 0 0 0
Neighs: 0 ave 0 max 0 min
Histogram: 1 0 0 0 0 0 0 0 0 0
FullNghs: 450032 ave 450032 max 450032 min
Histogram: 1 0 0 0 0 0 0 0 0 0
Total # of neighbors = 450032
Ave neighs/atom = 56.254
Neighbor list builds = 38
Dangerous builds = 0

Zhiting Tian

Once you are equlibrated, I would switch to NPT with a bigger
damping constant and no drag term. Then run it longer.
If the system pressure oscillates around 0.0, then that
is the volume the potential is telling the system it wants.

Steve