The ABINIT Users Mailing List ( CLOSED )

["ÁõÔ¥¶«" <liuyudong97@mails.tsinghua.edu.cn>]

Dear Steve:
Very,very glad to get your letter! 
Your help is desperately needed by me! I am in honor to be the second to use 
the
constraint equation. Hahah

I calculate the migration of interstitial atom from (000) to (1/2 1/2 1/2) 
along 
[111] direction. So I wish to relax the interstitial atom at different (111) 
cross
plane with the different sum___X(i)+Y(i)+Z(i)=0.1,0.2,...0.5. something like a
series of  C.T. photoes. Then I link the stable sites at the different C.T.
photoes to get the migration path. My crystal cell is rhombohedral, so its
symmetry should let the atoms near interistial atom to also move inside a 
cross
plane like (111). 

Your explaination tells me how to move the atom from (000) site to (100) site
along x axils, that is very usful, and I will try it.  However in my case, my
diffusion path is from (000) to body center ,then to (100). Thus, I perfer to 
know
how to contraint the atom in a cross plane like (111). 
Now, can you tell me how to perform such kind of constraint?

yours

Yudong Liu



>From: Steven Erwin <erwin@dave.nrl.navy.mil>
>Reply-To: forum@abinit.org
>To: forum@abinit.org
>Subject: Re: [abinit-forum] how to use constrait equation in relaxing the
atom?
>
>Dear Yudong Liu,
> 
> Since I contributed the geometry constraint subroutine (and may be the 
> only user to date!), I will try to answer your question.
> 
> The constraints are implemented in a fairly general -- although possibly 
> confusing -- manner.  The input variable 'wtatcon' (WeighTs for AToms in 
> CONstraint equations) tells the whole story:
> 
> --
> "Gives the weights determining how the motion of atoms is constrained 
> during structural optimization or molecular dynamics (see nconeq, 
> natcon, and iatcon). For each of the nconeq independent constraint 
> equations, wtatcon is a 3*natcon array giving weights, W[I], for the x, 
> y, and z components of each of the atoms (labeled by I) in the list of 
> indices iatcon. Prior to taking an atomic step, the calculated forces, 
> F[I], are replaced by projected forces, F'[I], which satisfy the set of 
> constraint equations
> 
> Sum[mu=x,y,z; I=1,natcon]: W[mu,I] * F'[mu,I] = 0 for each of the nconeq 
> arrays W[I]."
> --
> 
> Your example -- relaxing an atom in a fixed plane -- is especially easy. 
>   Use the following tokens/values to allow all atoms to relax while 
> keeping  the z-coordinate of atom #16 frozen at its initial value:
> 
> nconeq 1 natcon 1 iatcon 16 wtatcon 0 0 1.
> 
> A slightly more complicated example would be the following.  Suppose you 
>   have a surface in which two nominally equivalent atoms (say, #15 and 
> #16) MIGHT have a broken-symmetry ground state in which their heights 
> are different.  You could, of course, just choose your surface unit cell 
> large enough to allow this, perturb the heights of these two atoms, and 
> then relax everything.  But a more systematic approach would be to 
> calculate the total energy as a function of their height difference. 
> You can do this by the following setting:
> 
> nconeq 1 natcon 2 iatcon 15 16 wtatcon 0 0 +1  0 0 -1,
> 
> which enforces the constraint z(atom15)-z(atom16)=constant during a 
> geometry relaxation.
> 
> I hope this gives you enough of an answer to make some progress.  Please 
> let me know if you need more information.  From the generality of your 
> question, I guess the documentation wasn't very helpful to you.  This 
> makes me think I should re-write the 'wtatcon' entry!
> 
> Regards,
> 
> Steve Erwin
> 
> 
> liuyudong97@mails.tsinghua.edu.cn wrote:
> > Dear all:
> > Abinit seems to permit the relax a atom in a fixed plane by employing the
structural optimization command------nconeq, natcon, iatcon, and wtatcon.I 
guess
that equations' formulasim possible is xi+yi=0.5 or sth.
> > 
> >  Now,My question is: 
> > 
> > How to use these command (or constrait equation) exactly? Can an expert 
> > show
me an example?
> > your help will be greatly appreciated!
> > 
> > by
> > yours
> > Yudong Liu
> > 
> 
> -- 
> Steven C. Erwin, Code 6390                    erwin@dave.nrl.navy.mil
> Center for Computational Materials Science           tel 202/404-8630
> Naval Research Laboratory, Washington DC 20375       fax 202/404-7546
> +++++++++++++++  http://cst-www.nrl.navy.mil/~erwin  ++++++++++++++++
> 
> 
>