-
Notifications
You must be signed in to change notification settings - Fork 35
Open
Labels
new benchmarkProposals and suggestions for new benchmarksProposals and suggestions for new benchmarks
Description
Summary
This issue proposes adding a benchmark for evaluating machine learning interatomic potentials on solid-state electrolyte systems using long-timescale molecular dynamics.
- Systems: 49 SSE structures, comprising Li, Cs, Cu, Na, and O-containing ionic conductors
- Simulation length: 1 ns MD trajectories per system
- Temperature range: 300–1300 K
- Evaluation metric: Pairwise RDFs resolved per element combination, scored via the minimum overlap score (range 0–1, where 1.0 = perfect agreement with reference)
Interactive features
Scatter plots.
Category
I suggest the creation of the electrolytes category.
Data availability
López, C., Rurali, R. & Cazorla, C. How Concerted Are Ionic Hops in Inorganic Solid-State Electrolytes? J. Am. Chem. Soc. 146, 8269–8279 (2024).
Computational cost
Expensive. The benchmark consists of 1 ns simulations with a timestep of 0.5 fs for each model and system.
Reactions are currently unavailable
Metadata
Metadata
Assignees
Labels
new benchmarkProposals and suggestions for new benchmarksProposals and suggestions for new benchmarks