Solve your hardest materials problems
Pair state of the art AI, machine learning, simulation, and lab management with your own lab data to answer in seconds what used to take months.
PHIN is accelerating catalyst development 15x with PNNL.
Read why battery, catalyst, and mining companies choose PHIN →Unify digital and physical testing with PHIN to accelerate R&D.
PHIN eliminates the serendipity of materials development by merging physical experiments, digital experiments, machine learning, and AI to predict behavior and performance across chemistries, structures, surfaces, and interfaces.
Explore solutions →
Energy Storage
Which chemistries and interfaces can improve capacity, lifetime, safety, and cost?
Quantum calculations resolve diffusion barriers. Machine-learned potentials scale them to electrolyte and SEI dynamics. Surrogates sweep thousands of chemistries in minutes.
Cycling and impedance data calibrate digital experiments and ML surrogates, grounding both in cells you have actually built and tested.
Identify a promising chemistry with a synthesis route with fewer exploratory experiments.
Catalysis
Which catalytic materials can improve activity and selectivity while reducing cost and scarcity?
Quantum calculations predict adsorption energies. Machine-learned potentials scale them to determine activity and identify reaction pathways. Surrogates search entire design spaces to optimize activity and selectivity across substitutions.
Measured turnover, selectivity, and stability calibrate the digital experiments and surrogates, grounding each round in your bench results.
Catalyst synthesis and validation is focused on the paths most likely to work.
Separations
Which materials, conditions, and additives work together to separate target species with greater selectivity and recovery?
Quantum calculations set binding and solvation energetics. Machine-learned dynamics predicts transport through the membrane. Screening sweeps structural libraries against your target species.
Permeability and rejection measurements calibrate the digital experiments and surrogates, grounding predictions in membranes you have actually tested.
Identify promising recovery and purification paths before scaling exploratory laboratory campaigns.
Semiconductors
How do defects, surfaces, and process materials effect device performance and manufacturing yield?
Electronic-structure calculations predict defect levels and band alignment. Machine-learned potentials calculate thermal transport at device scale. Surrogates map that physics to process conditions.
Process and yield data calibrates the digital experiments and surrogates against your line, not an idealized one.
Narrow material and process choices earlier with evidence grounded in atomic-scale behavior.