Dear Sapna Bondwal,
pw.x (almost**) always describes systems that are periodic so if you
desire to approximately describe an isolated cluster you need to
surround it by some vacuum space so that periodically repeated images
interact as little as possible.
let's say you want to describe SiH4, let's say SiH bondlength is about
1.5 AA, so the molecule is about spherical with a diameter twice as
much, 3 AA... add 6-7 AA of vacuum to separate it from its periodic
image and you get ~20 a.u.: a SC cell of that size should be reasonable.
the easiest way to specify the coordinates is to choose ATOMIC_POSITIONS
in angstrom or bohr so you don't have to bother about rescaling them
when you test convergence w.r.t. cell size... something that you should
always do to make sure your cluster is really isolated.
**Once you are at ease with this simple calculations for isolated
molecules you can explore the
effect of the assume_isolated flag (there are various options that
correct for the spurious interactions among periodical images with
various degrees of sophistication)
stefano
On 11/10/2015 14:30, sapna bondwal wrote:
Hello,
I am working on Silicon nanoclusters. A nanocluster is not a periodic,
so one needs to create artificial periodicity using supercell. I
choose say, ibrave = 1 i.e cubic lattice, cell dimensions= 20.00.
Could someone guide me how to create a supercell using these
parameters? How to give coordinated of a cluster for example Si5 or
SiH4? How to specify the vacuum space, Cell length? Could someone give
me clear idea about my query? What is the importance of supercell
approach? I request someone to give a simple illustartion so that I am
able to appreciate the supercell approach in a better way.
Thank You in advance.
--
Sapna Bondwal
Research Scholar
Theoretical Chemistry Lab
IIT ROORKEE
_______________________________________________
Pw_forum mailing list
[email protected]
http://pwscf.org/mailman/listinfo/pw_forum
_______________________________________________
Pw_forum mailing list
[email protected]
http://pwscf.org/mailman/listinfo/pw_forum
