The potential coefficients for any potentials

Dear all,
I’m developing the molecular model of the EPON 862-DETDA crosslinked epoxy system. wherein, the EPON 862 is Di-glycidyl ether of Bisphenol-F, and the crosslinking agent DETDA is Dietheylene Toluene Diamine.
I have one problems: How I can find the potential coefficients for them with any potentials (Compass, or OPLS…)? any potential coefficients are ok for full atoms, united atoms,…
Thanks in advance,
yancheng

Dear all,
            I'm developing the molecular model of the EPON 862-DETDA
crosslinked epoxy system. wherein, the EPON 862 is Di-glycidyl ether of
Bisphenol-F, and the crosslinking agent DETDA is Dietheylene Toluene
Diamine.
             I have one problems: How I can find the potential coefficients
for them with any potentials (Compass, or OPLS...)? any potential
coefficients are ok for full atoms, united atoms,...

you need to search the literature! not MD software or mailing lists.
there are different force fields with different levels of accuracy and
transferability for different uses and also special purpose
parameterizations for specific molecules or functional units.

you need to know what kind of studies you want to do
and what level of detail and accuracy you require to get
meaningful results (keeping in mind that time and length
scales can matter as much as direct accuracy of the model)

based on that, you can look for a specific set of parameters
or a "generalized" force field that can handle different unusual
compounds to some degree. please also note, that some
force fields are only available with a license of the corresponding
commercial software. LAMMPS only offers the MD engine,
i.e. the facility to run your simulation once you have found
all the parameters you need. for some force fields and
unusual atom types or fragments, you may need to first
learn how to derive such parameters yourself. most force
fields have a well documented "recipe" for adding your
own parameters so that they are consistent with the
remaining force field. it is highly recommended to resist
the urge to mix and match different force fields, as many
classical force fields depend to a significant degree on
error cancellation, and that is not working well with
parameters mixed from different force fields, as the
parameterization strategy differs.

axel.

Thank you very much for your detailed response.
It seems that I still need to do many research.
Best wishes,
yancheng

"there are different force fields with different levels of accuracy
and transferability for different uses and also special purpose
parameterizations for specific molecules or functional units."

Just to feed my knowledge,
For a specific structure of one material, are the values of force
field parameters generally different for different temperatures, for
example 0K, 10K, 300K, 3000K?
For different structures of one material, for example crystal and
amorphous silica, are the values of force field parameters generally
different?
if you don't mind I ask.
Thank you.

L.Zhang

2012/1/10 Long Zhang <[email protected]>:

"there are different force fields with different levels of accuracy
and transferability for different uses and also special purpose
parameterizations for specific molecules or functional units."

Just to feed my knowledge,
For a specific structure of one material, are the values of force
field parameters generally different for different temperatures, for
example 0K, 10K, 300K, 3000K?

yes. particularly for pairwise additive potentials.
those contain the not pairwise additive parts
as "effective" components.

that doesn't keep some people from
using the same parameters over a large
range of temperatures anyway, one just
has to be *very* careful with the interpretation
of the results. the farther you get away from
the conditions a parameter set has been
derived for, the less accurate it gets and
this may not be a linear or small deviation.
however, certain trends may still be very
well represented.

For different structures of one material, for example crystal and
amorphous silica, are the values of force field parameters generally
different?

depends on the system, the formulation of the
force field and how the material is parameterized.
there can even be a difference and thus lack of
transferability between surface and bulk atoms.

there is no general answer to this.
otherwise a lot of people doing ab initio
MD simulations would be wasting their time. :wink:

if you don't mind I ask.

no.

axel.