Hi,
I am trying to optimize the lattice parameters for lithum (BCC). I did it in two ways:
1. using a primitive unit cell (1 atom at 0,0,0 with rprim =(-0.5 0.5 0.5, 0.5 -0.5 0.5, 0.5 0.5 -0.5)) and
2. using a non-primitive unit-cell (2 atoms at 0,0,0 and 0.5, 0.5, 0.5 with rprim =(1 0 0, 0 1 0, 0 0 1))
The input files are attached and an extract from the output files is given below. Although the calculated lattice parameters are almost the same in both the cases, the energy in the case 2 is exactly twice that of case 1. For some specific reason I want to use non-primitive unit-cells.
Can some one point out what changes should I make in the input to get the right energy?
Thanks,
Syed
output when using primitive unit cell
acell 6.3940787146E+00 6.3940787146E+00 6.3940787146E+00 Bohr
amu 6.94100000E+00
diemac 1.20000000E+01
dilatmx 1.50000000E+00
ecut 6.00000000E+00 Hartree
ecutsm 5.00000000E-01 Hartree
etotal -2.9403809117E-01
output when using non-primitive unit cell
acell 6.3905321431E+00 6.3905321431E+00 6.3905321431E+00 Bohr
amu 6.94100000E+00
diemac 1.20000000E+01
dilatmx 1.50000000E+00
ecut 6.00000000E+00 Hartree
ecutsm 5.00000000E-01 Hartree
etotal -5.8831139468E-01
non-primitive unit cell
Moderator: bguster
non-primitive unit cell
- Attachments
-
li-non-prim.in- (1.29 KiB) Downloaded 236 times
-
- li-prim.in
- (1.26 KiB) Downloaded 252 times
-
david.waroquiers
- Posts: 138
- Joined: Sat Aug 15, 2009 12:45 am
Re: non-primitive unit cell
Hello,
The fact that you get twice the energy of the system with 1 atom (primitive cell) for the system with 2 atoms (non-primitive cell) is normal. The total energy (its absolute value) that you get out of your calculation is not physically meaningful as is. It will strongly depend on the pseudopotentials used for example (also on the number of electrons that are treated as valence). On the other hand, total energy differences have a physical meaning. For example, you can get the binding energy of a solid by substracting the energy of an isolated atom from the energy of the bulk system (if this system has one atom). These total energy differences can be compared from one pseudopotential to another.
I hope it helps you
David
The fact that you get twice the energy of the system with 1 atom (primitive cell) for the system with 2 atoms (non-primitive cell) is normal. The total energy (its absolute value) that you get out of your calculation is not physically meaningful as is. It will strongly depend on the pseudopotentials used for example (also on the number of electrons that are treated as valence). On the other hand, total energy differences have a physical meaning. For example, you can get the binding energy of a solid by substracting the energy of an isolated atom from the energy of the bulk system (if this system has one atom). These total energy differences can be compared from one pseudopotential to another.
I hope it helps you
David
Re: non-primitive unit cell
Hi,
Yes, I understand that the absolute value of energy is not important. However, this is my concern. If one compares the two input files all the parameters are same except for natom, typat, xred and rprim. I run the two simulations with the same pseudo potential. So, i expected that the energies should be identical. I am planning to calculate elastic constants using non-primitive unit-cells, so I want to make sure that I am not making any mistake in the first step itself.
Syed
Yes, I understand that the absolute value of energy is not important. However, this is my concern. If one compares the two input files all the parameters are same except for natom, typat, xred and rprim. I run the two simulations with the same pseudo potential. So, i expected that the energies should be identical. I am planning to calculate elastic constants using non-primitive unit-cells, so I want to make sure that I am not making any mistake in the first step itself.
Syed