about the tsmear
Posted: Tue Sep 13, 2011 10:45 am
Hellow,
When i do the calculations about the structural optimized or single point energy calculation, I use the tsmear = 0.01 and occopt = 7 which is recommended for d-band metals. And the -kT*entropy = -2.70988931698273E-03 Hartree(73.74 meV). Is it reasonable to reduce the tsmear to get the -kT*entropy less than 1 meV? I used rather lower tsmear = 0.0006 and get the -kT*entropy = -1.17412789876294E-05 Hartree(0.3195 meV). When I do the ecut or k point sample convergence, is it reasonable to get the energy convergence to 1 meV (it is reasonable convergence to 1 meV in VASP code)
When i do the calculations about the structural optimized or single point energy calculation, I use the tsmear = 0.01 and occopt = 7 which is recommended for d-band metals. And the -kT*entropy = -2.70988931698273E-03 Hartree(73.74 meV). Is it reasonable to reduce the tsmear to get the -kT*entropy less than 1 meV? I used rather lower tsmear = 0.0006 and get the -kT*entropy = -1.17412789876294E-05 Hartree(0.3195 meV). When I do the ecut or k point sample convergence, is it reasonable to get the energy convergence to 1 meV (it is reasonable convergence to 1 meV in VASP code)