Hello, I am Manuel Monge Palacios, a postdoctoral researcher at the University of Columbia-Missouri.
I have one question related to the “fix neb” command in LAMMPS. Can I use that command to optimize (and in the best case, also to characterize) the saddle points of the reactions described by ReaxFF for the H2-O2 system? The way to use the command is very well explained in the manual, but nothing is said about the use of the command when using ReaxFF. I would appreciate your help.
Thank you very much for your time and attention
Manuel Monge-Palacios
Postdoctoral Fellow
Department of Theoretical Chemistry
University of Missouri-Columbia
Hello, I am Manuel Monge Palacios, a postdoctoral researcher at the
University of Columbia-Missouri.
I have one question related to the "fix neb" command in LAMMPS. Can I use
that command to optimize (and in the best case, also to characterize) the
saddle points of the reactions described by ReaxFF for the H2-O2 system? The
way to use the command is very well explained in the manual, but nothing is
said about the use of the command when using ReaxFF. I would appreciate your
help.
the NEB method doesn't care how you compute the forces. it is nothing
more than a set of minimizations with some additional forces added.
hence no mentioning of any specific model or potential in the
documentation.
whether you'll get meaningful results would depend mostly on whether
the method you choose (ReaxFF) with the settings/parameters can
represent the system and the required properties in question. for that
you'd have to study the published literature documenting the
capabilities and expected errors and shortfalls of the parameter set
you want to use and ReaxFF in general for the problem at hand. there
is no question that ReaxFF parameters cannot be used in as general a
fashion as methods/basis sets in quantum chemistry programs, as they
are usually tuned for specific environments and systems.
> Hello, I am Manuel Monge Palacios, a postdoctoral researcher at the
> University of Columbia-Missouri.
> I have one question related to the "fix neb" command in LAMMPS. Can I use
> that command to optimize (and in the best case, also to characterize) the
> saddle points of the reactions described by ReaxFF for the H2-O2 system?
The
> way to use the command is very well explained in the manual, but nothing
is
> said about the use of the command when using ReaxFF. I would appreciate
your
> help.
the NEB method doesn't care how you compute the forces. it is nothing
more than a set of minimizations with some additional forces added.
hence no mentioning of any specific model or potential in the
documentation.
whether you'll get meaningful results would depend mostly on whether
the method you choose (ReaxFF) with the settings/parameters can
represent the system and the required properties in question. for that
you'd have to study the published literature documenting the
capabilities and expected errors and shortfalls of the parameter set
you want to use and ReaxFF in general for the problem at hand. there
is no question that ReaxFF parameters cannot be used in as general a
fashion as methods/basis sets in quantum chemistry programs, as they
are usually tuned for specific environments and systems.
But that's of course the case for each and every potential available in
LAMMPS.
I am also so interested in using this command with ReaxFF in order to find a saddle point of a reaction. The manual is not enough clear to me and that might be due to my lack of knowledge.
Assume we detected a reaction in a system of N molecules and we were able to identify the reactive encounters as well. In order to use NEB do we have to isolate the reactive encounters and take the geometries for them?
In other words how can we use NEB and what does the data-file look like?
We are interested in finding the SP for a reaction, should we separate the reactive encounters from the other non Reactive molecules?
I do appreciate if you elaborate on this more.
I am also so interested in using this command with ReaxFF in order to find a
saddle point of a reaction. The manual is not enough clear to me and that
might be due to my lack of knowledge.
please note that the LAMMPS manual is meant to describe how particular
commands work (i.e. syntax and semantics and which features/settings
are activated or disabled or modified). it is *not* its purpose to
explain the underlying method as well or provide you with a customized
tutorial for your specific kind of simulation. thus before using a
particular command/method, you will have to familiarize yourself with
the fundamental methodology and theoretical derivation (usually from
either text books or publications or both).
Assume we detected a reaction in a system of N molecules and we were able to
identify the reactive encounters as well. In order to use NEB do we have to
isolate the reactive encounters and take the geometries for them?
there is a fundamental issue here: NEB is a 0K method (based on
minimization), while you are running MD at finite temperature (based
on time integration).
In other words how can we use NEB and what does the data-file look like?
this is all explained in the documentation for the neb command (which
works in conjunction with fix neb) and there are also examples
provided that you can study and compare to the documentation. it is
highly recommended to work on simple problems and only if you are
confident that you have understood how they work, move on to more
complex systems.
We are interested in finding the SP for a reaction, should we separate the
reactive encounters from the other non Reactive molecules?
that is really a question about your research and not so much about
how NEB is implemented in LAMMPS. if you understand the NEB method
well enough, you'll be able to answer the question yourself (and it is
really something that you have to choose and not somebody else for
you). there is also the question as to how well ReaxFF can describe
the energy of the entire reaction pathway altogether and for your
particular system of interest, for which you'll have to study the
relevant literature as well.
The short answer to the first question is 'Yes.' You can use the
LAMMPS neb command to find saddle points for collections of atoms
interacting with the ReaxFF potential. This works best for small
clusters. As you add more atoms, the complexity of the potential
energy surface increases dramatically. In order to get useful
results, a lot of human-based assistance is required. It's a good idea
to start with a simple system, like CH3 + CH4 -> CH4 + CH3.
The short answer to the second question is 'No.' There is currently no
post-processing script for taking a LAMMPS trajectory generated at
1000 K with thousands of atoms, during which the reaction A+B -> C+D
occurred, and generating LAMMPS input to a 20 atom NEB calculation
with end points A+B and C+D, that relaxes to a minimum energy path
passing through the correct lowest energy (AB)* transition state.
The short answer to the first question is 'Yes.' You can use the
LAMMPS neb command to find saddle points for collections of atoms
interacting with the ReaxFF potential. This works best for small
clusters. As you add more atoms, the complexity of the potential
energy surface increases dramatically. In order to get useful
results, a lot of human-based assistance is required. It's a good idea
to start with a simple system, like CH3 + CH4 -> CH4 + CH3.
The short answer to the second question is 'No.' There is currently no
post-processing script for taking a LAMMPS trajectory generated at
1000 K with thousands of atoms, during which the reaction A+B -> C+D
occurred, and generating LAMMPS input to a 20 atom NEB calculation
with end points A+B and C+D, that relaxes to a minimum energy path
passing through the correct lowest energy (AB)* transition state.
To some degree, this is what TAD does, which is implemented in LAMMPS. But
I don't think it will work for gas-phase reactions, because NEB will have
major difficulties to find saddle points between two states that don't only
differ by bonding state, but are also translated/rotated/underwent
conformational changes in between, as is typically the case in gas-phase
systems. In general, you want that the transition invoves the translation
of a small amount of atoms, with the rest of the system remaing stationary
(for example, at their lattice positions).