pair_style hybrid reax/c l

LAMMPS LAMMPS Mailing List Mirror
1

Dear LAMMPS users and developers

The version of lammps is LAMMPS (11 Jan 2013)
I want to use hybrid two reax force field in my simulation

units real
dimension 3
atom_style charge
boundary p p f

read_data data.14
mass 3 58.693
mass 1 12.010
mass 2 1.008
mass 4 15.999

pair_style hybrid reax/c reax/c lmp_control
pair_coeff * * reax/c 1 ffield.test O Ni Null
pair_coeff * * reax/c 2 ffield.reax.FC C Ni

The error is "error opening the force filed file! terminating..."
The error should be come from the incorrect input line about pair_style hybrid or pair_coeff

Would anybody tell me the origin for this kind error?

Thanks

Best regards

Wang

2

Dear LAMMPS users and developers

The version of lammps is LAMMPS (11 Jan 2013)

I want to use hybrid two reax force field in my simulation

[...]

pair_style hybrid reax/c reax/c lmp_control

pair_coeff * * reax/c 1 ffield.test O Ni Null

pair_coeff * * reax/c 2 ffield.reax.FC C Ni

this is wrong in several ways:
- for reaxFF there are very few cases, where using pair style hybrid
makes sense. in general, you should look for a force field file that
has been parameterized specificall for your system at hand, and then
no hybrid is needed
- when using hybrid pair styles with multiple manybody potential
instances, there must no overlap between atom type assignments.
manybody potentials (like reaxff) are not pair-wise additive and thus
mixed terms must be done with a simple potential
- in your pair coeff statements you have three atom types in the first
and two atom types in the second input. that doesn't make sense. also
the first atom type is assigned to two different elements. that is
nonsense, too.
- your LAMMPS version is far too old. it neither supports having
multiple instances of the same pair style nor does your the reax/c
pair style in your version support pair_style hybrid.
- the syntax of the pair_style command is wrong, too.

The error is “error opening the force filed file! terminating...”

The error should be come from the incorrect input line about pair_style
hybrid or pair_coeff

the *real* problem is that you are trying something that is a *very*
bad idea. you cannot "merge" multiple reaxff force field files this
way. chances are, that they are not even consistent with each other.
find a proper force field file that supports all atom types your
system has and that has been parameterize for the compounds and
environments you want to study and also upgrade your LAMMPS version.

axel.

3

This comes when there is problem with parameters file. Check your parameters file

4

This comes when there is problem with parameters file. Check your parameters
file

not in this case. the LAMMPS version is too old and doesn't support
multiple instances of the same pair style for manybody potentials,
thus it want to read the "1" and "2" numbers as force field files.
however, there are so many other things wrong with that input, it is
pointless to try and address this specific error.

axel.

5

Dear,axel.
Thank you for your suggestion. I have been updated the version of lammps The version is LAMMPS (7 Dec 2015) now.

I want to simulate the CO and H2 on nickel.
However, the CO cannot adsorb on the nickel based on the reaxff forcefiled which provide by the ref.( Tavazza, F.,Senftle, T. P.,Zou, C.,Becker, C. A.&van Duin, A. C. T. Molecular Dynamics Investigation of the Effects of Tip–Substrate Interactions during Nanoindentation. The Journal of Physical Chemistry C 119, 13580-13589 (2015).) May be this is not a good idea to merge two reaxff.
However, I cannot find other reasonable forcefields to describe the interaction between NI and O.

Hence, I want merge the two exist reaxff about (Ni,C,H)(ref. Mueller, J. E.,van Duin, A. C. T.&Goddard, W. A. The Journal of Physical Chemistry C 114, 4939-4949 (2010).) And reaxff about (Ni O) (I have been test oxygen atom on the nickel.)

pair_style hybrid/overlay reax/c lmp_control reax/c lmp_control
pair_coeff * * reax/c 1 ffield.reax.nich C H Ni NULL
pair_coeff * * reax/c 2 ffield.reax.FC NULL NULL Ni O
fix 2 all qeq/reax 1 0.0 10.0 1e-6 reax/c

Now the error is "ERROR: No pair reax/c for fix qeq/reax (../fix_qeq_reax.cpp:165)"

Regards
Wang.q

6

Dear,axel.
Thank you for your suggestion. I have been updated the version of lammps The version is LAMMPS (7 Dec 2015) now.

that is *still* an outdated version. there have been two (2!) stable
releases since then. but then again, the main issue is that you are
ignoring the most important advices from my e-mail.

I want to simulate the CO and H2 on nickel.
However, the CO cannot adsorb on the nickel based on the reaxff forcefiled which provide by the ref.( Tavazza, F.,Senftle, T. P.,Zou, C.,Becker, C. A.&van Duin, A. C. T. Molecular Dynamics Investigation of the Effects of Tip–Substrate Interactions during Nanoindentation. The Journal of Physical Chemistry C 119, 13580-13589 (2015).) May be this is not a good idea to merge two reaxff.

apparently, your attention is very selective. i already *explicitly*
stated, that what you are trying to do is a very, *very* bad idea.
there are no "maybe"s there. how much more drastic should i get?

However, I cannot find other reasonable forcefields to describe the interaction between NI and O.

you are looking for the interaction of Ni and CO and not Ni and O.
that is a very different thing. you seem to be thinking that
classical/empirical force fields can be employed similar to
pseudoptentials in plane-wave DFT. they cannot. as i already stated,
your approach is wrong at a very fundamental level.

Hence, I want merge the two exist reaxff about (Ni,C,H)(ref. Mueller, J. E.,van Duin, A. C. T.&Goddard, W. A. The Journal of Physical Chemistry C 114, 4939-4949 (2010).) And reaxff about (Ni O) (I have been test oxygen atom on the nickel.)

pair_style hybrid/overlay reax/c lmp_control reax/c lmp_control
pair_coeff * * reax/c 1 ffield.reax.nich C H Ni NULL
pair_coeff * * reax/c 2 ffield.reax.FC NULL NULL Ni O
fix 2 all qeq/reax 1 0.0 10.0 1e-6 reax/c

Now the error is "ERROR: No pair reax/c for fix qeq/reax (../fix_qeq_reax.cpp:165)"

i already explained what the constraints for mechanical coupling of
manybody force fields are. your setup violates that and is flat out
wrong as all Ni-Ni (and Ni-Ni-Ni and Ni-Ni-Ni-Ni) interactions are
computed twice, but others, e.g. O-C, Ni-O-C, Ni-C-O and many others
are missing.

your main problem seems to be that you are confusing a syntactically
correct LAMMPS input with a meaningful LAMMPS input. you may be able
to get past syntax restrictions, but with your approach for this kind
of system you will *NEVER* get a meaningful simulation.

either you get *one* set of parameters that can correctly cover *all*
Ni, CO and H2 together at the intended conditions, or you'll have to
resolve to other, less empirical, methods.

axel.