As suggested by VASP, the Materials Project uses the Li_sv PAW potential. The ENMAX for this potential is 499.034 eV. However, the documentation states that the 520 eV cutoff was chosen since it corresponds to 1.3 times the maximum ENMAX of the pseudopotentials used. That doesn’t seem to be the case for elements involving Li, as the 520 eV cutoff is well below 1.3*499 eV = 649 eV. I believe this is the only element where the choice of pseudopotential results in an ENMAX greater than 520 eV. I wonder then how relevant the unphysical Pulay stresses are during the structural optimizations of Li-containing compounds as a result of this. I don’t have a major interest in Li-containing compounds so have not personally investigated it.
You’re correct that the latest Li_sv PAW potential has 499 as ENMAX, but the version that’s been used in all of our calculations (i. e. the Jan 2001 version) is from an older set of PSPs, which has an ENMAX of ~270. I’m not sure why this was chosen initially or when it was integrated into the official VASP potential distro, but we’re very likely to update both this and the ENMAX parameter for a database rebuild using a metaGGA in the next couple of years.
Oh, I see. Thanks for clarifying that. I didn’t check the older PAW potentials – what a difference. Regardless, thanks for the helpful reply (and a metaGGA rebuild would be interesting!).