[Wien] :ENE units

[Paula Harkins]

>Hi,
>
>I'm working on spin-polarised calculations of rock salt structured 
>Chromium Nitride, and I have calculated the density of states for 
>non-magnetic, ferromagnetic, and antiferromagnetic spin systems. When I 
>analyse the total energy (:ENE) of these systems, I find a large 
>discrepancy with accepted values. The Wien97 values are an order of 
>magnitude greater than those in my reference paper.
>
>I suspect this may be due to my lack of understanding of what the total 
>energy refers to, i.e. is it per formula unit, per unit cell,  relative to 
>volume of unit cell determined by the lattice parameters, or some other 
>factor I haven't considered.
>
>The non-magnetic and ferromagnetic CrN unit cells are cubic, space group 
>Fm-3m, with a = 4.1480 Angstroms, hence my case.struct file has two 
>inequivalent atoms listed.
>
>The antiferromagnetic unit cell is orthorhombic, with a = 5.757, b = 
>2.964, c = 4.134 Angstroms. To properly state a magnetic unit cell with 
>spin-up and  spin-up chromium atoms, I have constructed a supercell with 8 
>inequivalent atoms listed, and transform this with the space group P1.
>
>Therefore, when I obtain my total energy values, I have divided the AF 
>value by 4 to arrive at a value per formula unit (as there are 4x as many 
>atoms in the AF unit cell - is this acceptable?
>
>My problem is that the energy difference between the non-magnetic and AF 
>systems is of the order of hundreds of mRyds, when it should be tens of 
>mRyds. Am I doing something foolish (probably), or is this a natural 
>phenomenon when doing an AF calculation with Wien97? I cannot find another 
>way to input my AFM system in the case.struct file which works.
>
>I should mention that the density of states obtained is pretty much an 
>exact match to the accepted DOS, so obviously my AFM parameters are correct.
>
>Hopefully someone can help, thanks in advance for any help,

>Dave Eustace