How to incorporate Atomic Site Occupancy ?

[Phani Kanth]

Hi All,

I am working with LiMnO2 material for calculating the lithium intercalation potential. The ICSD structure has partial site occupancies for both Li and Mn. The atomic positions are given as follows:
Li1 Li1+ 0.25 0.25 0.114 0.025 Uiso 0.98
Mn1 Mn3+ 0.25 0.25 0.114 0.025 Uiso 0.02
Mn2 Mn3+ 0.25 0.25 0.6345 0.0128 Uiso 0.984
Li2 Li1+ 0.25 0.25 0.6345 0.0128 Uiso 0.016
O1 O2- 0.25 0.75 0.1418 0.0179 Uiso 1
O2 O2- 0.25 0.75 0.6011 0.0165 Uiso 1
the heading for each column is as mentioned below:
_atom_site_label
_atom_site_type_symbol
_atom_site_fract_x
_atom_site_fract_y
_atom_site_fract_z
_atom_site_U_iso_or_equiv
_atom_site_thermal_displace_type
_atom_site_occupancy

As you see in the last column, the site occupancies are not equal to 1, for Li and Mn.
Could someone help me, how to incorporate site occupancies for atoms while relaxing the structure?

Thanks in Advance,
Phani Kanth

[forsdan]

One suggestion is to use a sufficently large cell where the site fraction can be fulfilled and subsequently assume som periodic arrangement for the exchanged Li and Mn atoms. Naively you can then also perform calculations for several different "randomized" configurations to see if there is any energy gain.

Otherwise the coherent potential approximation (CPA) or the KKR method could be worth looking up references on, in order to deal with even more random configurations. But these methods are quite far outside my area of expertise...

Best regards,
/Dan

<span class='smallblacktext'>[ Edited Thu Jun 03 2010, 08:12PM ]</span>