Cell volume

[rghomari]

Hi everybody,

I’m a Phd student and i start working on vasp and I’m locating a few problems. FIirstly, I read some article and including the cell volume (V in A/fu). but when i I shall not fall the same values of cell volume by applying this formula a^3. c/a. SQRT(3)/2. what's the difference between the units(A AND A/fu)?
I'm working on hexagonal structure (a=b #c and ß=120). I’wd like to use the cell volume in my POSCAR. But i found that the KPOINTS in my OUTCAR is very large points 205. what you think it's logical.

POSCAR:
System=surface1
-67
3.64 0.0 0.0
0.0 3.64 0.0
0.0 0.0 5.90
1 2
DIRECT
0.0 0.0 0.0
0.66 0.33 0.33
0.33 0.66 0.66

I will greatlly appreciate if some can help me.

cheers

[forsdan]

1. fu = 'formula unit' represents the kinds and numbers of atoms in the smallest representative unit, so the volume Ã…^3/fu would represent the volume of your primitive cell. If you have more than one primitive cell in your supercell you will instead deal in terms of a volume (Ã…^3) which is the number of primitive cells multiplied with the volume/primitive cell. The value for the volume you supply in the POSCAR file (measured in Ã…^3) will be the total volume for the supercell regardsless of the number of primitive cells you have.

2. Regarding your geometry. I might be missing something but I can't see how your lattice vectors together with the direct coordinates makes up a hexagonal structure. The struture I see seems to have very few symmtries which might explain the large number of k-points. I can't see anything that resembles a hexagonal pattern, but I havn't worked with that particular structure to a large extent. If the c-axis is oriented in the (0,0,1) direction wouldn't the (0,0,0) species make up a honeycomb lattice plane? Are you sure your geometry is set up correctly? Which source do you use for the structure? Any particular paper?

Anyone else that have had more experience with this structure might have more qualified comments to add.

Best regards,
/Dan Fors



<span class='smallblacktext'>[ Edited Fri Jul 18 2008, 03:00PM ]</span>

[rghomari]

Hello.
thank you for your amswer.
my reference is: J.Appl.Cryst. (2008) 41, 68-82.
thanks

[forsdan]

Well, judging from Fig. 1 in the reference you gave:

The Mg atoms can be described according to a hexagonal lattice with the lattice vectors:

(a, 0, 0), (a/2,sqrt(3)*a/2,0) and (0,0,c). With a = 3.64 Ã…, c = 5.90 this translated into:

POSCAR:
System=surface1
-67
3.64 0.0 0.0
1.82 3.15233247 0.0
0.0 0.0 5.90
1 2

Since the Mg atoms should be "sandwitched" between the Cl atoms, the Cl atoms in direct coordinates will sit on (1/3,1/3, q) and (2/3,2/3,1-q) where q is an internal parameter equal the deviation from the hexagonal plane. According to Chemical Physics Letters 350, p543-550 (2001) q will be given by 0.25. Thus I would say that your POSCAR should read:

POSCAR:
System=surface1
-67
3.64 0.0 0.0
1.82 3.15233247 0.0
0.0 0.0 5.90
1 2
DIRECT
0.0 0.0 0.0
0.333333333 0.333333333 0.25
0.666666666 0.666666666 0.75

However, this is just my interpretation. I might be wrong. Please verify it.

Best regards,
/Dan Fors




<span class='smallblacktext'>[ Edited Tue Jul 22 2008, 01:36PM ]</span>

[rghomari]

Thanks for your answer,i appreciated.
i tried again my calcul with a new POSCAR, but my KPONTS is still large; 366 kPONTS. you think thats is normal.

[home79]

I apologize that i am a newbie and not familiar with cell volume optimization. Currently, i am working on ZnO with similar hexagonal structure (a=b, c lattice). I would like to know, if i build a supercell structure, to optimize the cell volume, for this structure of mine, do I need to constrain the a and b lattice (since ZnO has lattice a=b) expanding or contracting by the same amount? Or should i release all a , b, c lattice optimzation freely?