Interatomic potential LCAO using GGA

[starvis]

Can someone tell me where possible cause of the problem ?

I am using home made code where LCAO (using LDA and GGA) implemented for calculation of electronic structure. Calculating structure for Cr (BCC) gives 1 Ang off the potential minimum (3.6 A should be 2.88 A). I checked k point mesh, primitive vectors, core charge (18), and atomic positions. If I calculate for SC I am getting the right potential minimum at 2.88.

Thank you so much for your help in advance.

 

[cgk]

You have a self-made DFT code and you don't even know where to look for the problem?...

 

[starvis]

Instead of asking this question I would prefer some help. If you are not able to
give reasonable suggestions please do not bother, learn and study more.

 

[cgk]

In order to give some helpful suggestions you need to provide information about your problem. From your problem description it is impossible to see what you are doing, and thus also what you are doing wrong.

In particular, you did not provide information on some *very* obvious questions:
0. What kind of basis set/methods do you use? Plane waves? Plane waves + pseudopotentials? Projector augmented plane waves? Gaussians? Something else? How do you calculate the terms in your DFT (e.g. Coulomb, core repulsion, pseudo-potentials, functional)? Are you sure these terms work?

1. Can you calculate the correct energies/wave functions/lattice constants for simpler systems? E.g., does your calculation work for a H solid or jellium? (This would allow you to exclude problems with the core representation and various basis set issues)

2. Does your program work for Coulomb-only "DFT" (i.e., the zero functional).

3. Do you know what the GGA/LDA result is supposed to be for Cr solid? I.e., did you calculate it with the same functional with a working code, or is the value you expect an experimental number?

In your OP you effectively said: "I have a home made program and it does not do what I expect." What kind of answer do you expect to that?

 

[starvis]

Thank you so much for your constructive reply:

0. I am using LCAO with Gaussian basis set solving KS equations in self consistent manner.

1. I have not calculated it yet, just finished Cr. The terms are calculated using functionals.

2.Have not check it yet, thank you so much I will check it.

3. I have checked the potential using vasp, qespresso and experiment. In vasp and qe i calculated using pseudopotentials.