ERROR on proc 0: Cannot open COMB3 lib.comb3 file

Dear all,

I've tried pair_style comb3 for the ineraction between carbon and hydroxy in graphene oxide, but an error occured (cannot open comb3 lib.comb3 file). I read the manul (The COMB3 library file cannot be opened), but i don't know where the library file is and wonder if there are something wrong in LAMMPS installation. Part of my input is as followed.

Hope and thanks for your reply!

# Initialzation

units metal
atom_style full
boundary p p p

# Force fields

pair_style hybrid lj/cut/coul/long 10 12 comb3 polar_off
bond_style harmonic
angle_style harmonic
read_data waterGO-$rnm.data
kspace_style pppm 1.0e-6

pair_coeff * * comb3 ffield.comb3 C NULL H NULL NULL O NULL NULL # GO comb3
... ...

Regards,

Chen Bo

Dear all,

I've tried pair_style comb3 for the ineraction between carbon and hydroxy
in graphene oxide, but an error occured (cannot open comb3 lib.comb3 file).
I read the manul (The COMB3 library file cannot be opened), but i don't
know where the library file is and wonder if there are something wrong in
LAMMPS installation. Part of my input is as followed.

​your e-mail two serious problems that make it very difficult to provide an
answer

- you don't state which exact version of LAMMPS you are using and what
platform you are running on
- you provide an incomplete input. how do you know that the problem is not
cause by anything you have omitted?

also, you obviously have not spent much time reading the documentation,
because the documentation for the pair style COMB3, already gives the
answer:

For style *comb3*, in addition to ffield.comb3, a special parameter file,
*lib.comb3*, that is exclusively used for C/O/H systems, will be
automatically loaded if carbon atom is detected in LAMMPS input structure.
This file must be in your working directory or in the directory pointed to
by the environment variable LAMMPS_POTENTIALS, as described on the
*pair_coeff* <http://lammps.sandia.gov/doc/pair_coeff.html> command doc
page.

more importantly, what you describe you are doing sounds like a bad model.
it sounds you want to use COMB3 like a pairwise additive potential, which
cannot work (that is, the simulation will run, but you get bogus results).

axel.

Dear Axel,

Thank you for your advice! I’ve really ignored the important information in manual. However, I encounter another problem after I moved the files to my working directory. I have tested the case in three different devices,

  1. LAMMPS( 31, Oct, 2015) in Windows-64bit, as long as i set up to run the case, a pop-up window appears that shows lmp_serial.exe has stop working and it is forced to be terminated.

  2. LAMMPS(29, Mar, 2015) in Window-32bit, it can set up to run but the thermo is weird.

Step Temp E_pair E_mol TotEng Press
0 274.50995 -1.#IND 106087.82 -1.#IND 1.#QNAN
Error on proc 0: Out of range atoms - cannot compute pppm

3)LAMMPS (15 May 2015) in Linux HPC-cluster, it is also forced to be terminated and the output and error files is enclosed.

My last question, I want to simulate that water go through three graphene oxide (graphene with hydroxy) layers. The graphene oxide atoms interact with each other via an comb3 potential. i want to use a REBO potential at first, but it can’t present the interaction with O atoms. And the water atoms interact with each other via a lj/cut/coul/long potential, the same as water/GO interaction. i don’t understand why it cannot work. Is comb3 potential improper in GO? or the hybrid pattern?

My complete input is as followed.

Initialzation

units metal
atom_style full
boundary p p p

Force fields

pair_style hybrid lj/cut/coul/long 10 12 comb3 polar_off
bond_style harmonic
angle_style harmonic
read_data waterGO-$rnm.data
kspace_style pppm 1.0e-6

set type 2 charge -1.0 # Cl charge
set type 4 charge 0.41 # Hw charge
set type 5 charge 1.0 # Na charge
set type 7 charge -0.82 # Ow charge

pair_coeff * * comb3 ffield.comb3 C NULL H NULL NULL O NULL NULL # GO comb3
pair_coeff 2 2 lj/cut/coul/long 0.00464007 4.446297312 # Cl unit parameter
pair_coeff 4 4 lj/cut/coul/long 0 0 # Hw 23.06
pair_coeff 5 5 lj/cut/coul/long 0.069693966 1.897438886 # Na
pair_coeff 7 7 lj/cut/coul/long 0.006739654 3.165520088 # Ow
pair_coeff 8 8 lj/cut/coul/long 0.681613182 2.542 # Pt

pair_coeff 1 2 lj/cut/coul/long 0.005457119 4.031673056 # C & Cl
pair_coeff 1 4 lj/cut/coul/long 0 0 # C & Hw
pair_coeff 1 5 lj/cut/coul/long 0.021149436 2.757243843 # C & Na
pair_coeff 1 7 lj/cut/coul/long 0.004545944 3.19 # C & Ow
pair_coeff 1 8 lj/cut/coul/long 0.066140915 3.0795244 # C & Pt

pair_coeff 2 3 lj/cut/coul/long 0 0 # Cl & H
pair_coeff 2 4 lj/cut/coul/long 0 0 # Cl & Hw
pair_coeff 2 5 lj/cut/coul/long 0.017982904 3.1718695 # Cl & Na
pair_coeff 2 6 lj/cut/coul/long 0.006773296 3.653041092 # Cl & O
pair_coeff 2 7 lj/cut/coul/long 0.005592179 3.8059087 # Cl & Ow
pair_coeff 2 8 lj/cut/coul/long 0.056238177 3.494148656 # Cl & Pt

pair_coeff 3 45 lj/cut/coul/long 0 0 # H & Hw Na
pair_coeff 3 7
8 lj/cut/coul/long 0 0 # H & Ow Pt

pair_coeff 4 * lj/cut/coul/long 0 0 # H & other

pair_coeff 5 6 lj/cut/coul/long 0.026250367 2.378611879 # Na & O
pair_coeff 5 7 lj/cut/coul/long 0.021672868 2.531479487 # Na & Ow
pair_coeff 5 8 lj/cut/coul/long 0.217954871 2.219719443 # Na & Pt

pair_coeff 6 7 lj/cut/coul/long 0.008163127 3.01265428 # O & Ow
pair_coeff 6 8 lj/cut/coul/long 0.082093122 2.700892436 # O & Pt

pair_coeff 7 8 lj/cut/coul/long 0.067777849 2.853760044 # Ow & Pt

Settings

neighbor 2.0 bin
neigh_modify delay 5

velocity sys create ${temp1} 123456
velocity fix set 0 0 0
velocity Pt set 0 0 0
timestep 0.001
write_dump all atom dump.setup_Watergraphene_rnm_{temp2}K.lammpstrj

Run for equlibrium

Fix

fix fixedatom fix setforce 0 0 0
fix Piston1 Pt1 rigid group 1 Pt1
fix Piston1 Pt2 rigid group 1 Pt2
fix 1 sys nvt temp {temp1} {temp1} 100 drag 0.2
fix p1 Pt1 aveforce NULL NULL 0.0004154
fix p2 Pt2 aveforce NULL NULL -0.0004154

Output

thermo 1000
restart 50000 restartWatergraphene_rnm_{temp2}K.*

Dump

dump 1 all atom 10 dump_Watergraphene_rnm_equil_{temp2}K.lammpstrj # 10000

Run

run 10000#0

error.22868 (1.68 KB)

output.22868 (6.76 KB)

Dear Axel,

Thank you for your advice! I've really ignored the important information
in manual. However, I encounter another problem after I moved the files to
my working directory. I have tested the case in three different devices,

1) LAMMPS( 31, Oct, 2015) in Windows-64bit, as long as i set up to run the
case, a pop-up window appears that shows lmp_serial.exe has stop working
and it is forced to be terminated.

​hard to tell. could be the same problem as in 2), but manifesting
differently.

2) LAMMPS(29, Mar, 2015) in Window-32bit, it can set up to run but the
thermo is weird.

Step Temp E_pair E_mol TotEng Press
0 274.50995 -1.#IND 106087.82 -1.#IND 1.#QNAN
Error on proc 0: Out of range atoms - cannot compute pppm

this is ​usually ​cause by a bad combination of potentials or potential
parameters or a bad initial system geometry and/or topology.

3)LAMMPS (15 May 2015) in Linux HPC-cluster, it is also forced to be
terminated and the output and error files is enclosed.

​see 1).

My last question, I want to simulate that water go through three graphene
oxide (graphene with hydroxy) layers. The graphene oxide atoms interact
with each other via an comb3 potential. i want to use a REBO potential at
first, but it can't present the interaction with O atoms. And the water
atoms interact with each other via a lj/cut/coul/long potential, the same
as water/GO interaction. i don't understand why it cannot work. Is comb3
potential improper in GO? or the hybrid pattern?

you will be best off to ​find a person that understands MD *and* LAMMPS and
can explain this to you and work with you to systematically build your
skills until you understand what is required to do such a complex system
correctly. what you are doing is rather complex and ambitious, yet you seem
to be lacking the suitable training for it. this looks like there are very
fundamental problems in deciding what kind of force field to use and why
and you need to understand your choices. the fact that you want to mix and
match potentials in as ignorant a fashion as you currently do, demonstrates
that you haven't understood what you are trying to do and need to revisit
and review and relearn. i can explain as much as i want, yet you won't
understand without sufficient background knowledge.
i already gave the key information in my previous response, yet you failed
to recognize is. there is little i can do via e-mail outside of
recommending to get a competent and experienced collaborator, and spend
more times learning the basics of setting up simulations by trying to set
up simpler systems with less complex force field inputs first.

axel.