I'd be very interested in knowing the charges on individual atoms in the
alkane-thiol chains if you have that data. I'm modeling a similar system
sorry, but no. not sure if that would help regardless.
for force fields, you have to follow a specific recipe
to determine charges and they have to be adapted
to polarization by solvent and other effects. so something
like mulliken charges (which are "free" with conventional
quantum chemistry calculations) will usually not do.
most of the time some electrostatic potential fit, often
with additional constraints, is required.
with CPMD calculations you use a plane wave basis
set, and as a consquence getting any local atom property
would require to first do a projection on a local basis
set and then perform the traditional quantum chemistry
procedures (with the restrictions that this projection is
typically incomplete and that one can only project on a
minimal basis set). if a reduced representation of the
electrostatic environment is required, then most commonly
a localization is performed and the centers of those
localized orbitals are used (so-called wannier centers).
using published OPLS/Charm parameters, but I suspect in reality the charges
in this system are different from those parameterized for isolated
hmmm... not sure that this would matter that much.
in my understanding the key problem with thiols on gold
are the S-Au and Au-Au interactions. the S-Au interactions
are so strong, that they can pull Au atoms out of a flat surface
leaving ad atoms and vacancies and on top of that, this is a
dynamical process (we can barely see it on the DFT time scales)
i.e., there are constantly atoms popping out and going back,
so that the binding on the surface keep changing; although
one sees domains with preferred orientation in the experiment.
please take my statements with a grain of salt, i have only
limited experience with au-thiol interactions. i came into
doing those calculations, since i had the experience in
running such a complicated and demanding calculation.
in general, there is an insane amount of literature on these
kind of systems, and a lot of it is not consistent. specifically
quantum chemistry calculations (be it DFT or other methods)
are always very restricted by the assumptions made about
the inputs. our study from 2-3 years ago, is in so far different,
as we went back and forth between experiment and simulations
multiple times and were feeding the fitting of the structures
with initial guesses from simulations and vice versa. still
the time scales are restricted and there is still a sufficient
amount of speculation left.