[Wien] ETA in PORT

[L. D. Marks]

On Tue, 14 Sep 2004, Jack London wrote:

> Dear Laurence
>
> I have problems about ETA in Port.
>
> Firstly, In the old version of Wien2K, ETA is ignored
> by PORT. But in the new version, ETA is bond order
> parameter. I don't know why you can do that? Is eta
> not important parameter in the old version?

	In the old version ETA was not used in the code -- read but then
ignored. ETA was (still is) used in NEWT.

>
> Secondly, based on UG, Do you means if I constraint an
> atom(DELTA(i)=0.0), I also should set ETA
> approximately to the bond order (or left at 1.0) NOT
> 0.0? Right?

	Yes, set it to one, 0.0 could create some other problems. (I think
I trapped these but cannot remember and don't have the code here.)

>
> At last, If I set ETA=1.0, do you mean Mini programme
> can find its band order?

	No, it does not find the bond order. Inside the code, the gradient
used is
	G_i = - M_i * (Fx,Fy,Fz)
where M_i is the multiplicity of atom "i", and Fx, Fy, Fz are the forces.
The diagonal elements of the initial Hessian approximation are set to
	H_ii = Constant * M_i * B_i
where B_i is the "Bond Order" term, and the "Constant" has been adjusted
so it is about right. The off-diagonal elements are set to zero.

If you happen to know ahead of time that a certain atom is poorly bonded
to the rest so it's second derivative will be smaller, you can input this
information in by having B_i smaller for that atom; similarly for an atom
which is more stongly bonded which will have a larger second derivative.
In general it is probably wisest to set this term to 1.0 -- the code will
automatically adjust the value.

Of course the "best" thing to do would be to set the initial Hessian (in
.minrestart) to something derived from an approximate pairwise bonding
model. This would save some CPU time/iterations, although someone would
have to write the code to do this. (It would not have to be that accurate,
something as simple as a spring model would probably work.)

>
> Thank you very much!
>
> Best Regards
>
> Zhang
>
>
>
> __________________________________
> Do you Yahoo!?
> Yahoo! Mail - 50x more storage than other providers!
> http://promotions.yahoo.com/new_mail
>
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
>

-----------------------------------------------
Laurence Marks
Department of Materials Science and Engineering
MSE Rm 2036 Cook Hall
2225 N Campus Drive
Northwestern University
Evanston, IL 60201, USA
Tel: (847) 491-3996 Fax: (847) 491-7820
mailto:L-marks at northwestern.edu
http://www.numis.northwestern.edu
-----------------------------------------------