Re: [Towhee-users] Beginner`s question on Towhee

[Marcus M. <mar...@us...>]

On Mon, Dec 22, 2008 at 7:35 AM, Babak Fazel <bk...@ya...> wrote:
> I have run towhee code with the following input file, for an NPT simulation
> of methane. Yet my results look to be significantly different from the
> corresponding experimental values. Any suggestions on the things which I
> might have missed?

Sure.  You missed performing any volume moves and therefore your
simulation is going to be at the vapor volume that you started it at.

> initlattice
> 'simple cubic'
> initmol
> 500
> inix iniy iniz
> 8    8    8
> hmatrix
> 270.0d0 0.0d0 0.0d0
> 0.0d0 270.0d0 0.0d0
> 0.0d0 0.0d0 270.0d0

That initial box size for 500 molecules is certainly a vapor.

> pm1boxcbswap
> 0.0d0
>           pm1cbswmt
>           1.0d0
> pmavb1
> 1.0d0
>           pmavb1in
>           0.5d0
>           pmavb1mt
>           1.0d0
>           pmavb1ct
>           1.0d0
>           avb1rad
>           20.0d0

By setting pmavb1 to a value of 1.0 you are only performing
Aggregation-volume bias moves of type 1.  While this is a very nice
move for sampling clusters in a vapor, it is not doing anything to
sample the volume changes.

A more typical single-box NPT simulation would consist of about 1%
volume moves, perhaps 5% of each of the aggregation volume bias moves,
and the remainder broken up between CBMC, translation, and rotation
moves.  As you are simulating a monatomic molecule in this particular
case there is no need for CBMC or rotation moves.

Marcus
-- 
Marcus G. Martin
Director, Useful Bias Incorporated
88 Martinez Road
Edgewood NM 87015-8222
ph. (505) 286-4457
www.usefulbias.com
www.photobirder.com