[lammps-users] Question on the airebo potential

Hi,
I found some intersting phenomenon when i ran the attached input file.
(1) i got different results of the pressure when I ran it with different processors.
(2)when i change the temperature from 100 to 8000,it ran normally in the first dozens steps,but suddenly broke down,I think if the input file can ran in the first steps,then the structure will be more reasonable in the later steps(so-called relaxation).but it was collapsed.why?
the varied part of the below input file
velocity all create 8000.0 123456789 dist gaussian units box

set fixes

fix 1 all nve
fix 3 all temp/berendsen 8000.0 8000.0 0.005
(3) in the manual,about the pair_style airebo command
in the part of Mixing, shift, table, tail correction, restart, rRESPA info:
it said that:
This pair style can only be used via the pair keyword of the run_style respa command. It does not support the inner, middle, outer keywords.
run_style verlet
I don’t really understand this sentence. i think it means that the pair style airebo can used with run style velert or run style respa . but when used “run style respa”, it can only used with this:run_style respa 4 2 2 2 pair 3
is it right?
(4) when i simulat the carbon atom deposited to the surface of diamond substrate,after one carbon atom collided with the surface, the TotEng changed slightly .but at some point, the sign of TotEng altered from negative(-14800.763) to positive (3796.5833). runing the same script with tersoff potential gives good results.

Memory usage per processor = 4.84544 Mbytes
Step Temp TotEng 1 2 3
5000 60.860035 -14881.773 0 86.040086 100.90086
5001 66.026167 -14880.186 1.5853404 84.997155 100.42895
5002 65.977553 -14880.185 1.5853404 84.864724 100.40142
5003 66.399081 -14880.187 1.5853404 85.725622 100.87953
5004 67.194631 -14880.191 1.5853404 87.391881 101.74731
5005 68.145307 -14880.196 1.5853404 89.43203 102.74348
5006 68.978727 -14880.201 1.5853404 91.300175 103.55044
5007 69.449054 -14880.204 1.5853404 92.490369 103.89259
5008 69.403714 -14880.204 1.5853404 92.677998 103.6077
5009 68.824348 -14880.202 1.5853404 91.809944 102.68797
5010 67.37605 -14880.325 1.5853404 90.106701 100
5011 66.202113 -14880.32 1.5853404 87.983981 98.43951
5012 65.121222 -14880.316 1.5853404 85.931716 97.084176
5013 64.356026 -14880.313 1.5853404 84.388052 96.200343
5014 64.051755 -14880.313 1.5853404 83.642639 95.958531
6140 159.85955 -14905.957 12.593407 271.62393 100
6141 157.54337 -14905.945 12.411606 266.15693 99.820319
6142 154.63812 -14905.924 12.186182 258.8125 99.974331
6143 152.36709 -14905.901 12.022112 252.54754 100.40865
6144 151.98003 -14905.892 12.035858 250.36738 100.98874
6145 153.94913 -14905.908 12.269455 253.55405 101.56151
6146 157.50589 -14905.934 12.630689 260.7453 102.01279
6147 161.23162 -14905.954 12.964448 269.03718 102.30337
6148 163.90759 -14905.964 13.165271 275.50936 102.47398
6149 164.90777 -14905.971 13.195271 278.3747 102.62047
6150 162.91137 -14906.259 13.04409 276.55517 100
6151 159.99278 -14906.237 12.733574 269.78984 100.492
6152 156.28054 -14906.205 12.388028 260.30057 101.35612
6153 153.82578 -14906.136 12.236699 252.40644 102.49821
6154 154.72514 -14906.137 12.63704 249.10155 103.76389
6155 161.35709 -14906.049 13.87447 255.22689 104.98582
6156 193.92405 -14903.321 15.908557 340.20289 106.02505
6157 1215.1891 -14800.763 17.464387 3796.5833 106.79162
6158 1248485.6 341462.56 18.099759 4249192.2 107.24818
6159 5187583.9 1443518.6 18.571688 17656916 107.40505
6160 5187623.2 1443536 17.999158 17657065 100
6161 5187591.6 1443536.7 14.639981 17656998 100.02011
6162 5187575.2 1443548.5 10.860501 17656987 100.47089
6163 5187515.3 1443538.8 16.911502 17656709 101.3205
6164 5187528.4 1443544.3 46.106761 17656399 102.47777
6165 5187666.1 1443549.7 76.519532 17656498 103.81565
6166 5187642.5 1443548.1 73.588754 17656451 105.18951
6167 5187626.2 1443544.9 71.463677 17656420 106.47568
ERROR: Lost atoms: original 2188 current 2185

I don’t know the reason.is there some bug in the potentail of airebo ?

1 proc:
Memory usage per processor = 2.92432 Mbytes
Step Temp E_pair E_mol TotEng Press
0 100 -7372.0547 0 -7359.1416 -14454.715
1 100.01122 -7372.0562 0 -7359.1416 -14454.52
2 99.957079 -7372.0491 0 -7359.1415 -14450.372
3 99.83914 -7372.0334 0 -7359.1411 -14442.25
4 99.659094 -7372.0092 0 -7359.1402 -14430.145
5 99.418748 -7371.9767 0 -7359.1386 -14414.058
6 99.120018 -7371.9358 0 -7359.1363 -14394.001
、、
24988 100.09174 -7361.9232 0 -7348.9982 -8685.3775
24989 100.17706 -7361.9346 0 -7348.9987 -8689.2019
24990 100.2662 -7361.9469 0 -7348.9994 -8693.143
24991 100.35877 -7361.9598 0 -7349.0003 -8697.1852
24992 100.443 -7361.3313 0 -7348.361 -8383.1503
24993 100.52173 -7361.3428 0 -7348.3624 -8386.5716
24994 100.60722 -7361.3555 0 -7348.364 -8390.2394
24995 100.69903 -7361.3692 0 -7348.3658 -8394.1414
24996 100.79675 -7361.3839 0 -7348.3679 -8398.2645
24997 100.89994 -7361.3996 0 -7348.3703 -8402.5944
24998 101.00814 -7361.4162 0 -7348.373 -8407.1156
24999 101.1209 -7361.4337 0 -7348.3759 -8411.812
25000 101.23775 -7361.4521 0 -7348.3792 -8416.6667

10 proce
Memory usage per processor = 2.40168 Mbytes
Step Temp E_pair E_mol TotEng Press
0 100 -7374.1602 0 -7361.2471 -15493.962
1 100.06852 -7374.1692 0 -7361.2473 -15496.052
2 100.1123 -7374.1752 0 -7361.2476 -15495.81
3 100.13187 -7374.178 0 -7361.2479 -15493.228
4 100.12782 -7374.1778 0 -7361.2482 -15488.3
5 100.10079 -7374.1746 0 -7361.2485 -15481.031
6 100.05149 -7374.1684 0 -7361.2486 -15471.428
、、
24988 100.77077 -7373.0768 0 -7360.0642 -15465.183
24989 100.68996 -7373.0682 0 -7360.066 -15453.979
24990 100.60772 -7373.0592 0 -7360.0676 -15442.455
24991 100.52428 -7373.0498 0 -7360.069 -15430.627
24992 100.43989 -7373.0401 0 -7360.0702 -15418.514
24993 100.35477 -7373.03 0 -7360.0711 -15406.132
24994 100.26918 -7373.0197 0 -7360.0718 -15393.501
24995 100.18336 -7373.0091 0 -7360.0723 -15380.641
24996 100.09753 -7372.9982 0 -7360.0725 -15367.57
24997 100.01196 -7372.9872 0 -7360.0726 -15354.309
24998 99.926856 -7372.976 0 -7360.0724 -15340.879
24999 99.842472 -7372.9647 0 -7360.072 -15327.301
25000 99.759035 -7372.9533 0 -7360.0713 -15313.597

20 proce
Memory usage per processor = 2.38213 Mbytes
Step Temp E_pair E_mol TotEng Press
0 100 -7373.8312 0 -7360.9181 -15331.579
1 99.839833 -7373.8101 0 -7360.9177 -15324.394
2 99.654588 -7373.7852 0 -7360.9168 -15314.694
3 99.445383 -7373.7568 0 -7360.9153 -15302.498
4 99.213384 -7373.7247 0 -7360.9132 -15287.832
5 98.959799 -7373.6892 0 -7360.9105 -15270.726
6 98.685874 -7373.6504 0 -7360.907 -15251.218
、、、
24988 100.89482 -7371.7629 0 -7358.7343 -14466.346
24989 100.92196 -7371.7688 0 -7358.7367 -14476.699
24990 100.94664 -7371.7745 0 -7358.7392 -14486.975
24991 100.96877 -7371.7799 0 -7358.7417 -14497.158
24992 100.98828 -7371.785 0 -7358.7443 -14507.232
24993 101.0051 -7371.7899 0 -7358.747 -14517.18
24994 101.01915 -7371.7944 0 -7358.7497 -14526.987
24995 101.03039 -7371.7985 0 -7358.7524 -14536.637
24996 101.03874 -7371.8024 0 -7358.7551 -14546.116
24997 101.04416 -7371.8058 0 -7358.7579 -14555.409
24998 101.04659 -7371.8089 0 -7358.7606 -14564.502
24999 101.04599 -7371.8116 0 -7358.7634 -14573.38
25000 101.04232 -7371.8138 0 -7358.7662 -14582.03

the input file

#melt test
dimension 3
boundary p p p
atom_style atomic
units metal
lattice diamond 3.57
region box block 0 5 0 5 0 5
create_box 1 box
create_atoms 1 box
mass 1 12.011
neighbor 0.5 bin
#define pair_styles
pair_style airebo 3.0 0 0
#pair_style tersoff
pair_coeff * * CH.airebo C
#pair_coeff * * SiC.tersoff C
#set initial velocity
velocity all create 100.0 123456789 dist gaussian units box

set fixes

fix 1 all nve
fix 3 all temp/berendsen 100.0 100.0 0.005
thermo 1
timestep 0.0001
dump 1 all xyz 100 c100.xyz
run 25000
write_restart heat_1.restart

Comments below.

Steve

2010/9/7 黄德明 <[email protected]>:

Hi,
   I found some intersting phenomenon when i ran the attached input file.
(1) i got different results of the pressure when I ran it with different
processors.

Please post a simple, small example of this.

(2)when i change the temperature from 100 to 8000,it ran normally in the
first dozens steps,but suddenly broke down,I think if the input file can ran
in the first steps,then the structure will be more reasonable in the later
steps(so-called relaxation).but it was collapsed.why?
the varied part of the below input file
      velocity all create 8000.0 123456789 dist gaussian units box
     # set fixes
     fix 1 all nve
     fix 3 all temp/berendsen 8000.0 8000.0 0.005

This temp/b command has way too short a time constant.
If your system isn't stable with NVE, then it is unlikely to
work with NVT.

(3) in the manual,about the pair_style airebo command
in the part of Mixing, shift, table, tail correction, restart, rRESPA info:
it said that:
This pair style can only be used via the pair keyword of the run_style respa
command. It does not support the inner, middle, outer keywords.
run_style verlet
I don't really understand this sentence. i think it means that the pair
style airebo can used with run style velert or run style respa . but when
used "run style respa", it can only used with this:run_style respa 4 2 2 2
pair 3
is it right?

Yes, you can run it with rRESPA as you say, but if you don't have
any other interactions (bonds, long-range, etc) in your system,
what's the point? There is only a single interaction being computed.

(4) when i simulat the carbon atom deposited to the surface of diamond
substrate,after one carbon atom collided with the surface, the TotEng
changed slightly .but at some point, the sign of
TotEng altered from negative(-14800.763) to positive (3796.5833). runing
the same script with tersoff potential gives good results.

Well, since your temperature went from 60 degrees to 10^6, I wouldn't
conclude anything from the energy, except that your model blew up.