Graphene wetting benchmark

[dwl38]

Pre-review checklist for PR author

PR author must check the checkboxes below when creating the PR.

• I've confirmed the contribution guidelines.

Summary

Added a new benchmark based on the adsorption energy curves of a single water molecule (at various orientations) on a sheet of graphene (under various strain conditions), which is useful for understanding nanoscale wetting. Reference calculations based on PBE functional, calculated using FHI-aims on "intermediate" settings.

Three metrics of equal weight:

1. MAE of all single-point calculations (across all orientations, strains, and distances)
2. MAE of binding energies (across all orientations and strains), by comparing fitted adsorption energy curves
3. MAE of binding lengths (across all orientations and strains), from the same method

Linked issue

Resolves #292

Progress

• Calculations
• Analysis
• Application
• Documentation

Testing

Benchmark tested on all currently-implemented models, i.e.:

• mace-mp-0a
• mace-mp-0b3
• mace-mpa-0
• mace-omat-0
• mace-matpes-r2scan
• orb-v3-consv-inf-omat
• pet-mad

with no issues. At current time of writing, mace-mp-0a performs the best on all metrics, whereas pet-mad completely fails to produce physically-reasonable adsorption energy curves.

New decorators/callbacks

Added a new "plot_from_scatter" callback, which implements almost identical functionality to "struct_from_scatter" except that it renders a Plotly Graph object instead.

[dwl38]

Database file:
graphene_wetting_under_strain.zip