[Wien] Questions on high spin and low spin FCC-Fe

Stefaan Cottenier Stefaan.Cottenier at fys.kuleuven.ac.be
Fri Aug 22 08:15:29 CEST 2003


> I am trying to calculate the properties of FCC-Fe with high spin and low
> spin. But I don't know how to describe the high-spin and low-spin
> configurations of FCC-Fe in WIEN2K.  I would be most grateful if someone
> could tell me how to do it.

When you start with any case.inst, you actually choose a starting point in the 
configuration space. The scf-cycle guides you then from that (more or less 
random) starting point to a self-consistent configuration (density). If there 
are several local minima (such as for high-spin and low-spin), you might end 
up in the other minimum than you wanted. Choosing another case.inst as you 
suggest, means that you choose another starting point, and you can *hope* that 
this will bring you to the other solution. But it remains uncontrollable trial 
and error.

The recipy given by Vladimir Timochevski is the synthesis of such a trial and 
error: by experience, he can identify starting configurations (in this case 
controlled by the lattice constant and not by case.inst) that lead to the 
desired solution.

There is a more systematic approach possible also: Fixed Spin Moment (FSM) 
calculations. These allow you to fix right at the start the spin moment you 
want to obtain. For a given lattice constant, you can then systematically 
sample a region of spin moments (e.g. from 0 to 5 in steps of 0.5 or 0.25), 
and plot the total energy as a function of spin moment. In the case of fcc Fe, 
you will see two energy minima: the low-spin and high-spin. You can repeat 
this process for other lattice constants, to obtain a 3D plot of the total 
energy as a function of lattice constant and spin moment: this allows you to 
obtain for both solution their respective lattice constant and value of the 
moment.

If you later want to do a regular calculation to obtain one of those 
solutions, you can start from the density in the desired minimum.

In order to use FSM, you have to use the script 'runfsm_lapw -m x.x' (where 
x.x indicates the desired moment, e.g. 2.5).

Stefaan




More information about the Wien mailing list