> Dear all,
> I have something unclear with lattice parameter of GGA Siesta calculation
> on
> Body centered Cr.
> Calculation without SpinPolarized gives  lattice parameter a=2.89
> Angstrom.
> The experimental value is a=2.88.
> Same calculation with  "SpinPolarized= True" gives a=3.94  Angstrom which
> is
> very differnt from 2.88.
> Could anyone tell me where i am wrong?  Thank you

You want a wild guess, without giving any details
on what you are doing? Well, one remark is that the "true" Cr
has a complicated magnetic structure, a simplest aproximation
to which is the antiferromagnetic one which you probably did not
consider. Although it is of course your good right to calculate it
ferromagnetic, and the difference in volume could not be that big.
Another speculation is that the 3p semicore which you probably
neglect may have a big effect on getting a good lattice constant,
and in the magnetic case you might have got it good by error cacellation.
Also, the pseudpot of Cr might be tricky.
And the number of k-points is demanding for metals. Do you have > 1 ?
In any case, don't be fascinated by the value
of equilirium lattice constant only; it is worth having a look
at, say, density of states you get (and compare it with previous
Good luck

Andrei Postnikov

Reply via email to