Skip to content
04 / energy
Energy

Electrolyte chemistry, screened on real physics.

Redox windows, solubility, and stability for Li-ion and Na-ion electrolytes — computed with quantum methods and neural-network potentials. A flexible toolkit, not a funnel: screen one candidate in depth or sweep a library for the voltage stability window, on the same engine, before a cell is ever built.

predict_redox_potentialpredict_solubilityrun_qm_calculationoptimize_geometry_nnpcompute_energypredict_bde
One instruction

What it looks like in practice.

any MCP-compatible assistant
you Screen these solvent candidates for oxidative stability above 4.5 V and report solubility.
optimized geometries with optimize_geometry_nnp (AIMNet2)
computed oxidation potentials via predict_redox_potential
flagged 2 solvents stable past 4.7 V vs Li/Li⁺
reported solubility with predict_solubility (temperature-aware)
Benchmarked

The accuracy, stated plainly.

Redox
oxidation /
reduction windows
NNP
geometry
optimization
BDE
bond dissociation
energies
T-aware
solubility with
temperature

Redox windows on demand

Oxidation and reduction potentials for electrolyte solvents and salts — the voltage stability window that decides whether a chemistry survives in a cell.

Temperature-aware solubility

A solubility model pretrained on AqSolDB and fine-tuned on BigSolDB with a temperature feature — solubility as it actually behaves across an operating range.

Bond dissociation energies

BDE prediction for degradation and stability questions, with a pre-trained fallback (alfabet) when a custom estimate is not needed.

NNP-accelerated geometry

Neural-network potentials (AIMNet2, MACE, ANI-2x) give fast, accurate geometries so a screen of many candidates stays tractable.

Research preview

Screen the voltage window before you build the cell.

NovoMCP is open to a small group of PIs, postdocs, and research engineers. Tell us what you are working on.