Ground-state cohesive energy calculated using pymatgen is very different from the one reported in reference.

[hongyi-zhao]

According to the tutorial here, I tried to calculate the ground-state cohesive energy of Au using pymatgen as follows:

In [24]: from mp_api.client import MPRester
    ...: 
    ...: m = MPRester()
    ...: 
    ...: # Retrieve data for gold (Au)
    ...: entry = m.get_entries('Au')[0]
    ...: bulk_energy = entry.energy
    ...: num_atoms = entry.composition.num_atoms  # Get the total number of atoms
    ...: 
    ...: # Retrieve the energy data for isolated atoms of the elements in the compound
    ...: elemental_entries = [m.get_entries(element.symbol)[0] for element in entry.composition.elements]
    ...: isolated_atoms_energy = sum([e.energy * entry.composition[e.composition.elements[0].symbol] for e in elemental_entries])
    ...: 
    ...: # Calculate the cohesive energy and average it per atom (unit: eV/atom)
    ...: cohesive_energy_per_atom = (bulk_energy - isolated_atoms_energy) / num_atoms
    ...: print("Cohesive Energy per Atom (eV/atom):", cohesive_energy_per_atom)
Retrieving ThermoDoc documents: 100%|█████████████████████████████████████████████████████████████████████████████████████████████| 10/10 [00:00<00:00, 199728.76it/s]
Retrieving ThermoDoc documents: 100%|█████████████████████████████████████████████████████████████████████████████████████████████| 10/10 [00:00<00:00, 220752.84it/s]
Cohesive Energy per Atom (eV/atom): 50.582851645

But this result differs greatly from the one reported in the reference mentioned in the above tutorial, as shown below:

Any hints or explanations for this discrepancy will be helpful.

Regards,
Zhao