ABINIT Discussion Forums

Hello,

I have a general question about the quality of the calculations of the thermal properties (free energy, entropy, specific heat, and internal energy) out of the anaddb tool. According to the tutorial they are calculated using the harmonic approximation. I want to double check that they are calculated via the method described in the article by Lee and Gonze, PRB 51, 13 (1995).
Also what are the quality of the calculations?

What would be good anaddb input values to use for quality calculations of the following variables: ng2qpt, nchan, nwchan, and thmtol, or any other important variables I am missing here. How dependent on the qpoint grid (ngqpt) initially calculated in abinit are these properties?

From my understanding of the harmonic approximation, the only values needed from the ddb's are the phonon density of states. Once those are obtained we can simply calculate the thermal properties at a given temperature. Would it be better to just take the density of states out of anaddb (using prtdos 1) and calculate the thermal properties on my own?

Thank you for your time!