Tuesday, June 3, 2014

LDA vs. GGA

Dear Diary,

Ok, doing some stuff with ABINIT and mastering their loop stuff.  I'm glad the first programming language I learned was Fortran because ABINIT is based on it.  Oh my, what's this?  In the tutorial it reads:

We will use the Perdew-Burke-Ernzerhof functional, proposed in Phys. Rev. Lett. 77, 3865 (1996)

Hmm...I wonder if that's the same Burke.  I'll have to ask Kieron when I get back.  It sure does take a bit longer to run than LDA...but the numbers get slightly better.  For the bond length, that is.  LDA does better on the atomization energy.

Do not forget that the typical accuracy of LDA and GGA varies with the class of materials studied...

[Editor's note:  By slightly, we mean a few fractions of a Hartree which is a big unit.  A small increase in a big number is a big number.  Idiot.]

2 comments:

  1. LDA should not be getting better energies than GGA...

    ReplyDelete
  2. For full disclosure, it was

    "In GGA, we are within 2% of the experimental bond length, but 5% of the experimental atomisation energy. In LDA, we were within 4% of the experimental bond length, and within 2% of the experimental atomisation energy."

    It has been changed in the text.

    ReplyDelete