Unknown error using COMB potential

Dear Lammps Users,

I attempted to heat SiO crystal using the COMB potential. However, the job was unable to run and forced closed LAMMPS before a single step could be run. log.lammps showed no error, it simply does not run. The same file has been used to run on the reparameterized Tersoff potential, which I adapted to the COMB potential with the fix qeq/comb. I attached my input file below, kindly advise whether there was something I did not do correctly to implement the COMB potential.

COMB potential


dimension 3
boundary p p p

units metal
atom_style charge

Create geometry

read_data silica.dat


pair_style comb
pair_coeff * * ffield.comb Si O

Initialise velocities

compute new all temp
compute 1 all pe/atom
compute 2 all ke/atom
compute 3 all stress/atom pair

velocity all create 300.0 2345346 dist gaussian


timestep 0.0005
thermo 1000

#Dump format: dump ID group-ID style N file args
dump 1 all custom 10000 *.dump x y z
dump_modify 1 sort id

fix 1 all nvt temp 300.0 5000.0 1.0

#Perform charge equilibration
fix 3 all qeq/comb 10 0.0001 file fq.out

run 100000

Yeo Jingjie
Ph.D. Student
School of Mechanical and Aerospace Engineering
Nanyang Technological University, Singapore

Hi Yingjie,

I ran your script with an example alpha cristobalite structure and it was successful. I suggest you check if you have the latest version of LAMMPS, and also check your silica.dat structure file.

Please note: COMB potentials are usually ran with a 0.2 fs timestep; 0.5 fs is too big. Also, “fix 1 all nvt temp 300.0 5000.0 1.0” raises temperature from 300 to 5000 in 1 ps, which is too fast.

Please let me know if you still have troubles in running COMB.


Hi Dr. Shan,

Thank you very much for the comments and recommendations. I have looked through my silica.dat file and am not sure where the anomaly lies. I have used a beta cristobalite structure from Wyckoff, if you could kindly take a look at my attached structure file and let me know whether there was indeed something faulty that I did not notice.

Also I noticed during the run, the following message was output before it stopped:

Pair COMB:
generating Coulomb integral lookup table …
will not apply over-coordination correction …
Setting up run …

Yeo Jingjie
Ph.D. Student
School of Mechanical and Aerospace Engineering
Nanyang Technological University, Singapore

silica.dat (67.6 KB)

The structure file looks alright, but try visualize the snapshot to verify everything is correct. Also, try with the structure file and the script provided in the example dir before you try your silica.dat.

That message is normal. However I am surprised no other warning/error message occured before it stopped.