Dear

Bhrigumoni Deka

Did you perform the optimization calculation for every LixCoO2 compound?

It is impossible if you get no difference in the total energy of crystal while occupancy of Li in LixCoO2 compound.

How are you modeling fractional occupancy? In an x-ray structure it is common to report that a position has fractional occupancy, (1/n) which implies that 1/n-th of such positions are occupied in an ensemble of unit cells. At the microscopic level, however, a given position is either occupied or it is not, and that is generally true in a computational model as well. If a position is half occupied, one must model a super-cell by combining two adjacent unit cells, then fill the position in one half of the super-cell and leave it empty in the other half. This can impart more long-range order into the model than is actually present in the real material, but it's generally a good first-order approximation.

I suggest you double-check that you are using fractional atomic occupancy -- not all input file converters will include fractional atomic occupation, some just set it to "1". In your positions block you should see something like:

Li 0.125 0.125 0.125 MIXTURE:( 1 0.35 )

i.e. a "mixed" atom has been defined, which is 0.35 of a lithium atom.

This will use the Virtual Crystal Approximation (VCA) to compute the energy; as Karl said, it is usually much better to use a supercell and do explicit vacancies.

All the best,

Phil Hasnip

(CASTEP developer)