Electrode objects with negative voltage

acarnevali · September 27, 2022, 2:33pm

Hi all!
While browsing the Battery Explorer I stumbled across some electrodes with a negative calculated average voltage. I was wondering what was going on with these entries: is there an error in the calculations upstream? It seems difficult to me that TM oxides could have redox potentials below the one of Li+/Li, provided that this is the actual zero value.

A couple examples I found: [mp-1176726_Li], [mp-246_Na]

jshen · September 27, 2022, 5:14pm

Hi @acarnevali, it is possible within our present definition of intercalation electrodes to see negative energies since we are only looking at pairs in the topotactic subset of materials:

github.com

materialsproject/pymatgen/blob/d7e9b7a7986f06601136190398d2dc2ecb7ca85d/pymatgen/apps/battery/insertion_battery.py#L104-L112


      
          # create voltage pairs
          _vpairs: tuple[AbstractVoltagePair, ...] = tuple(
              InsertionVoltagePair.from_entries(
                  _stable_entries[i],
                  _stable_entries[i + 1],
                  working_ion_entry,
              )
              for i in range(len(_stable_entries) - 1)
          )

If you are seeing negative voltages that you don’t expect this could be due to that particular chemical subspace (Fe2O3 + Li_x for example) being undersampled. Or it could be due to the fact that the Tolerances for structure matching were too high and a higher voltage (lower Li/ no Li) structure was missed.

If you grab the structures, from the entries you can use the following code to structure match them together and play around with the tolerances:

github.com

materialsproject/emmet/blob/a7924eb3b241b59b0fc0a9f128aa305295e03453/emmet-core/emmet/core/structure_group.py#L168-L213


      
          results = []
          sm = StructureMatcher(
              comparator=ElementComparator(),
              primitive_cell=True,
              ignored_species=[ignored_specie],
              ltol=ltol,
              stol=stol,
              angle_tol=angle_tol,
          )
          
          
# Add a framework field to each entry's data attribute
          for ient in entries:
              ient.data["framework"] = _get_framework(
                  ient.composition.reduced_formula, ignored_specie
              )
          
          
# split into groups for each framework, must sort before grouping
          entries.sort(key=lambda x: x.data["framework"])
          framework_groups = groupby(entries, key=lambda x: x.data["framework"])

This file has been truncated. show original

However, for the Fe2O3 material, there does seem to be some rendering problems:

Just gonna ping some people here about it
@mkhorton
@munrojm
@tschaume