ERROR: All pair coeffs are not set

LAMMPS LAMMPS Beginners
1

Dear LAMMPS users,
I try to simulate interactions of C60 with S3N4 lattice.
Got ERROR: All pair coeffs are not set (…/pair_tersoff.cpp:386)
The relevant part of the input looks as :

mass 1 28.0855 # Si atoms has mass 9.012182
mass 2 14.0067 # N atoms has mass 14.0067
mass 3 28.0855 # Si atoms has mass 9.012182
mass 4 14.0067 # N atoms has mass 14.0067
mass 5 12.011 # C atoms has mass 12.011

#-------------- Define Interatomic Potential -------------------
pair_style hybrid/overlay tersoff tersoff/zbl airebo/morse 3.0
pair_coeff * * tersoff SiN.tersoff Si N Si N NULL
pair_coeff * * tersoff BNC.tersoff NULL N NULL NULL C
pair_coeff * * tersoff BNC.tersoff NULL NULL NULL N C
pair_coeff * * tersoff/zbl SiC.tersoff.zbl Si NULL NULL NULL C
pair_coeff * * tersoff/zbl SiC.tersoff.zbl NULL NULL Si NULL C
pair_coeff * * airebo/morse CH.airebo-m NULL NULL NULL NULL C

#------------- Run the simulation for equilibration -------------
group Si3N4 region Si3N4
group addSi3N4 region addSi3N4
group Si3N4all union Si3N4 addSi3N4
velocity Si3N4all create 300.0 12345 dist gaussian
fix nptSi Si3N4all npt temp 300.0 300.0 (100.0*dt) x 0.0 0.0 (100.0dt) y 0.0 0.0 $(100.0dt) # .5 x 0.0 0.0 .5 y 0.0 0.0 .5

read_data C60.dat add append group impactor
velocity impactor create 1530.0 4928459 rot no dist gaussian
timestep 0.001
run 20000
…
The output of the ERROR region looks as:

mass 1 28.0855 # Si atoms has mass 9.012182
mass 2 14.0067 # N atoms has mass 14.0067
mass 3 28.0855 # Si atoms has mass 9.012182
mass 4 14.0067 # N atoms has mass 14.0067
mass 5 12.011 # C atoms has mass 12.011

#-------------- Define Interatomic Potential -------------------
pair_style hybrid/overlay tersoff tersoff/zbl airebo/morse 3.0
pair_coeff * * tersoff SiN.tersoff Si N Si N NULL
pair_coeff * * tersoff BNC.tersoff NULL N NULL NULL C
Reading tersoff potential file BNC.tersoff with DATE: 2013-03-21
pair_coeff * * tersoff BNC.tersoff NULL NULL NULL N C
Reading tersoff potential file BNC.tersoff with DATE: 2013-03-21
pair_coeff * * tersoff/zbl SiC.tersoff.zbl Si NULL NULL NULL C
Reading tersoff/zbl potential file SiC.tersoff.zbl with DATE: 2009-04-15
pair_coeff * * tersoff/zbl SiC.tersoff.zbl NULL NULL Si NULL C
Reading tersoff/zbl potential file SiC.tersoff.zbl with DATE: 2009-04-15
pair_coeff * * airebo/morse CH.airebo-m NULL NULL NULL NULL C
Reading airebo/morse potential file CH.airebo-m with DATE: 2016-03-15

#------------- Run the simulation for equilibration -------------
group Si3N4 region Si3N4
744372 atoms in group Si3N4
group addSi3N4 region addSi3N4
8424 atoms in group addSi3N4
group Si3N4all union Si3N4 addSi3N4
744372 atoms in group Si3N4all
velocity Si3N4all create 300.0 12345 dist gaussian
fix nptSi Si3N4all npt temp 300.0 300.0 (100.0*dt) x 0.0 0.0 (100.0dt) y 0.0 0.0 (100.0*dt) # .5 x 0.0 0.0 .5 y 0.0 0.0 .5 fix nptSi Si3N4all npt temp 300.0 300.0 0.10000000000000000555 x 0.0 0.0 (100.0dt) y 0.0 0.0 (100.0*dt) fix nptSi Si3N4all npt temp 300.0 300.0 0.10000000000000000555 x 0.0 0.0 0.10000000000000000555 y 0.0 0.0 (100.0*dt)
fix nptSi Si3N4all npt temp 300.0 300.0 0.10000000000000000555 x 0.0 0.0 0.10000000000000000555 y 0.0 0.0 0.10000000000000000555

read_data C60.dat add append group impactor
Reading data file …
orthogonal box = (-195.33785 -192.38662 -100.00000) to (195.33785 192.38662 100.00000)
4 by 2 by 2 MPI processor grid
reading atoms …
60 atoms
read_data CPU = 0.026 seconds
velocity impactor create 1530.0 4928459 rot no dist gaussian

timestep 0.001
run 20000
ERROR: All pair coeffs are not set (…/pair_tersoff.cpp:386)
Last command: run 20000

Job for Si3N4 itself (atoms 1 to 4, where atoms 1 and 2 is the original lattice with delete_atoms region and 3 and 4 replace deleted atoms of the original lattice) is completed OK
Jobes for C60 interacting with other lattices as Be, Al, etc are also run OK
“C60.dat” data file is :
C60 impact Au LAMMPS data file

C60 Data file for LAMMPS

	60  atoms

	1  atom types

-35.67 65.67 xlo xhi
-35.67 65.67 ylo yhi
-100.00 100.00 zlo zhi

Masses

	5     12.011   # C60		(Nose_Hoover)

Atoms

1 5 3.45100000 0.68500000 7.00000000
2 5 3.45100000 -0.68500000 7.00000000
…
60 5 -2.28000000 -2.58100000 6.27600000

Highly appreciate any suggestions
Victor

2

You are treating MANYBODY potentials like pairwise additive potentials. That is incorrect and cannot work.

When using a hybrid pair style to combine different potential styles, you can can only apply this to manybody interactions when each domain has a single manybody potential and all atom types of that domain are contained and not shared with any other domain. For the mixed interactions between the domains you must use a pairwise additive potential.

3

Thank you very much Axel. I’ve tried it now and it works exactly as you said. Unfortunately, this makes life a bit more complicated due to the need to find an appropriate pairwise one :). Regards, Victor

4

Hybrid potentials are always only the second best solution. The preferred option would be to have the entire system represented by the same potential function. Also, there are different levels of combinations, Tersoff and AIREBO are similar in what data they use and how they compute interactions. Combining multiple EAM/MEAM potentials is much more problematic because of the splitting of the computation of the embedding term (and hence very inconsistent interactions) and even worse is if one subsystem uses a potential with charge equilibration like ReaxFF or COMB.

One final comment, you should be using pair style hybrid not hybrid/overlay. The latter is meant for when you want to add two or more potential functions for the same atoms (it will become functionally equivalent to hybrid, if there are not overlaps, but it could lead to undesired double counting of interactions otherwise).

1 Like