Hey @Minju,
- You can still use the legacy pymatgen API and
get_cohesive_energyfunction with an older API key (should be 16 characters rather than the 21 characters in the new API). You can check your old API key here - I pulled down the PBE GGA atomic energies from our legacy API here, units are eV/atom. Keep in mind that these energies cannot be used to calculate cohesive energies for solids with r2SCAN total energies on MP:
{'H': -1.11587797, 'He': 0.01038154, 'Li': -0.29720495, 'Be': -0.01764556, 'B': -0.26858985, 'C': -1.25998247, 'N': -3.11173655, 'O': -1.53343974, 'F': -0.50059069, 'Ne': -0.01160895, 'Na': -0.22867367, 'Mg': -0.09551372, 'Al': -0.21633877, 'Si': -0.8265492, 'P': -1.88794092, 'S': -0.89072159, 'Cl': -0.25877019, 'Ar': -0.04921952, 'K': -0.22735189, 'Ca': -0.09267815, 'Sc': -1.97271039, 'Ti': -2.20322453, 'V': -3.66390549, 'Cr': -5.59832953, 'Mn': -5.32596236, 'Fe': -3.41439808, 'Co': -1.9470445, 'Ni': -0.91354767, 'Cu': -0.60077969, 'Zn': -0.16418411, 'Ga': -0.32917553, 'Ge': -0.87843304, 'As': -1.68376491, 'Se': -0.76921751, 'Br': -0.22119763, 'Kr': -0.03314942, 'Rb': -0.18694104, 'Sr': -0.068011, 'Y': -2.17417537, 'Zr': -2.04279634, 'Nb': -3.13299031, 'Mo': -4.60247889, 'Tc': -3.36167381, 'Ru': -2.14870945, 'Rh': -1.35815518, 'Pd': -1.4765485, 'Ag': -0.33901873, 'Cd': -0.16683316, 'In': -0.35508443, 'Sn': -0.83605507, 'Sb': -1.41061736, 'Te': -0.65688183, 'I': -0.18905393, 'Xe': -0.00984012, 'Cs': -0.13555854, 'Ba': -0.03393307, 'La': -0.67461885, 'Ce': -1.18331004, 'Pr': -0.19451181, 'Nd': -0.20166733, 'Pm': -0.20603743, 'Sm': -0.21234603, 'Eu': -8.36550995, 'Gd': -10.27279609, 'Tb': -0.24654003, 'Dy': -0.24630703, 'Ho': -0.25218271, 'Er': -0.28255396, 'Tm': -0.25582842, 'Yb': -0.06421108, 'Lu': -0.26442917, 'Hf': -3.18577473, 'Ta': -3.42503432, 'W': -4.65706744, 'Re': -4.70114419, 'Os': -2.89756089, 'Ir': -1.62318678, 'Pt': -0.53597051, 'Au': -0.28808586, 'Hg': -0.12456359, 'Tl': -0.31762167, 'Pb': -0.77557173, 'Bi': -1.32632748, 'Ac': -0.16706447, 'Th': -0.75124164, 'U': -4.80121024, 'Np': -7.53922056, 'Pu': -10.71752857}
- The difference you observe in Quantum Espresso and VASP total energies is because they are both pseudopotential codes. Absolute total energies have no meaning in a pseudopotential calculation because the vacuum level (the Kohn-Sham potential at a point infinitely far from the solid) is poorly-defined
Energy differences calculated with the same pseudopotential are physically meaningful, so you’d just need to do the same calculation in VASP with the same pseudpotentials MP uses to get compatible total energies
Hope that helps, let me know if I missed anything