Page 1 of 1

tpss through libxc convergence

Posted: Mon Jul 12, 2010 10:32 pm
by boz
Hello,

Does anyone have experience with libxc plugin? I have encountered some problem with convergence of tpss functional through libxc. I have tested several atoms ( H, O, C) and molecules (H2, O2), and graphite. The cell for a single atom or molecule is a 10 by 10.1 by 10.2 angstrom box. Abinit does pretty well when converging the H, H2, O2, and C. But for O atom and graphite ( using experimental lattice constants), they are never converged within 1000 steps. The pseudo potential I am used is the FHI GGA type. Does anyone can help?