Dear all,
I try to use the Al2O3 vashishta potential combined with zbl potential to simulate the irradiation progress. Here is one question:how to make sure these two parameters - inner and outer?
Thanks for your help.
Best wishes！
yinbo
Dear all,
I try to use the Al2O3 vashishta potential combined with zbl potential to simulate the irradiation progress. Here is one question:how to make sure these two parameters - inner and outer?
Possible solutions:
1 - read/apply carefully the documentation
2 - vary the parameters and observe the differences (calculated forces, etc.)
3 - read the source code
Not sure what you are asking. Are you saying you will use
pair hybrid to superpose the two potentials? That is not
really the proper way to do it. It you look at the pair tersoff/zbl
page you will see a merged single pair style has been
created which correctly connects the ZBL formula
with the pairwise portion of the Tersoff formulas.
And the ZBL params (including the cutoff) for making
the connection are encoded in the potential file.
Something similar has not been done for the Vashishita
potentials, but someone could code up a pair vashishita/zbl.
Steve
Not sure what you are asking. Are you saying you will use
pair hybrid to superpose the two potentials? That is not
really the proper way to do it. It you look at the pair tersoff/zbl
page you will see a merged single pair style has been
created which correctly connects the ZBL formula
with the pairwise portion of the Tersoff formulas.
And the ZBL params (including the cutoff) for making
the connection are encoded in the potential file.Something similar has not been done for the Vashishita
potentials, but someone could code up a pair vashishita/zbl.
Even if it is difficult to know exactly what people are doing in their papers (a lot of technical detail omitted), it seems to me quite common in the irradiation litterature to merge a material-specific potential (like Stillinger-Weber or an EAM) with a ZBL potential for short inter-atomic distances. And when authors state using Lammps, it is most probably done via the “overlay”.
I always thought the tersoff/zbl as an built-in potential for irradiation/high energy simulation, but without expecting any difference with the overlay. Am I wrong?
Julien
Steve
[...]
Even if it is difficult to know exactly what people are doing in their
papers (a lot of technical detail omitted), it seems to me quite common in
the irradiation litterature to merge a material-specific potential (like
Stillinger-Weber or an EAM) with a ZBL potential for short inter-atomic
distances. And when authors state using Lammps, it is most probably done
via the "overlay".I always thought the tersoff/zbl as an built-in potential for
irradiation/high energy simulation, but without expecting any difference
with the overlay. Am I wrong?
yes, you are wrong. in tersoff/zbl the repulsive pairwise interaction is
*replaced* with a ZBL based function. hybrid/overlay *adds* the ZBL
function to the existing repulsive function. please check out
pair_tersoff_zbl.h and you should see.
axel.
Axel is correct about tersoff/zbl. However it
may be also possible to do something reasonable
with pair hybrid/overlay and just superposing
ZBL over some other pair style. So long as
the cutoff for ZBL is short, the tersoff pair part
(for example) will probably be much smaller
and therefore the ZBL term will dominate when
2 atoms (very rarely) approach closely, which is the
motivation for using ZBL.
I just don’t know what that cutoff or other
ZBL params should be for that to work well.
Steve
Axel is correct about tersoff/zbl. However it
may be also possible to do something reasonable
with pair hybrid/overlay and just superposing
ZBL over some other pair style. So long as
the cutoff for ZBL is short, the tersoff pair part
(for example) will probably be much smaller
and therefore the ZBL term will dominate when
2 atoms (very rarely) approach closely, which is the
motivation for using ZBL.
...or one can just create a tabulated potential that either subtracts out
the repulsive part from the original potential so that ZBL can be added to
it or which is the difference between the original pairwise potential and
ZBL and then uses this table as overlay. i recall that this latter option
was discussed a couple of times on this list. it works even for most
manybody potentials, since their (atom core) repulsive part is usually
pairwise additive.
axel
Thanks! Very interesting comments!
The tabulated solution is indeed the easiest way to obtain kind of a hybrid/switch function.
Julien
Axel is correct about tersoff/zbl. However it
may be also possible to do something reasonable
with pair hybrid/overlay and just superposing
ZBL over some other pair style. So long as
the cutoff for ZBL is short, the tersoff pair part
(for example) will probably be much smaller
and therefore the ZBL term will dominate when
2 atoms (very rarely) approach closely, which is the
motivation for using ZBL.I just don’t know what that cutoff or other
ZBL params should be for that to work well.
As far as I remember (not a too recent article thought: Nordlund et al., Nucl. Instrum. Methods Phys. Res B, 2006), it was acknowledged to be largely determined subjectively, and that greater care should be taken since it could have a significant influence. Not sure someone looked at that…
Julien
Steve