ABINIT Discussion Forums

Dear all,

I wonder if it is currently possible to run structural optimization (ionmov) but keep some of the ions fixed?
There are many reasons people do it including coupling to MD (effectively prescribing the boundary conditions in terms of ions positions)
as well as some surface energy calculations of the slab with bottom ions fixed during optimization.
I know that this option is implemented in VASP, since the book i was looking for give some examples calculated via this code.

Do you think it is relatively straight forward to modify Abinit code to implement this?
I would assume one could add a couple of lines in those BFGS procedures and actually keep positions of certain ions (given by their index in the rang of [0;natom-1]) fixed, whereas let the others to move towards the minimum. If one is to do so, which files should be actually modified?

A related question is, can Abinit be compiled as a library and used from the "driver"?

Kind regards,
Denis.

Hello Denis,


You should look at the natfix/iatfix variables for fixing atoms as well as natfixx, natfixy, natfixz and iatfixx, iatfixy and iatfixz for fixing atoms only in a given direction (or two if you wish).


David

thanks a lot, David! I just started to play with Abinit and missed those variables. That's exactly what i was looking for.