Technology

One platform, from molecule to market

Molecular simulation, machine learning, process engineering, and techno-economics usually live in isolation. Planck connects them into one automated workflow.

The gap we close

Why existing computational screening falls short

Properties ≠ performance

Existing tools predict adsorption under idealized conditions, not survival through heat buildup, slow kinetics, and regeneration in a real cycle.

Viability checked too late

Water stability, thermal resilience, and supply-chain risk are usually checked after expensive campaigns. We check them first.

One method doesn't fit all physics

CO₂ capture is classical; hydrogen at metal sites is quantum. Most platforms apply a single method everywhere, sacrificing either accuracy or compute.

Flexibility is ignored

Many top materials breathe and flex under gas loading, transforming their performance. Nearly all screening tools pretend frameworks are rigid. Ours doesn't.

Limited to known materials

Database screening can only rank what has already been synthesized. The optimal material for your process may not exist yet, so we generate it.

Brute force doesn't scale

When accurate simulation is expensive, evaluating 100,000+ candidates is infeasible. A search that learns where to spend compute is missing, so we built one.

The screening pipeline

100,000 candidates in. ~50 proven winners out.

Each stage eliminates candidates before the next, more expensive one. Total compute drops by orders of magnitude.

01 · Structural & supply-chain filtering ~100,000
Pore size, surface area, element availability
02 · Stability screening (machine learning) ~30,000
Water, thermal, and chemical stability
03 · Gas adsorption simulation ~5,000
Adaptive method selection per problem physics
04 · Structural flexibility check ~3,000
Dynamic response under gas loading
05 · Process simulation ~500
Full industrial process optimization per material
06 · Cost & economic analysis ~50
Levelized cost, supply chain, sensitivity
07 · Decision dashboard Final shortlist
Interactive comparison for stakeholders
Not a list of lab curiosities: a ranked shortlist scored on purity, recovery, energy per tonne, and cost, delivered in an interactive dashboard with full sensitivity analysis.

Preview the decision dashboard on real data

Beyond brute force

When screening isn't enough, we search smarter and invent.

Adaptive optimization

Adaptive Bayesian optimization

Rather than simulating everything, our Bayesian optimization learns which material characteristics predict performance and picks the next candidate to evaluate. Expensive simulations concentrate on the most promising regions of the space.

Generative AI

Materials that don't exist yet

The optimal material for your process may not exist yet. Our generative module (diffusion models over framework building blocks) proposes new structures for specific targets. Every AI-generated candidate enters the same screening pipeline.

Why Planck

One workflow, every capability

From molecular simulation to techno-economics, validated stage by stage over years.

  • Molecular simulation of gas adsorption
  • Stability prediction: water, thermal, chemical
  • Quantum-accuracy methods for complex systems
  • Automated structural flexibility assessment
  • Integrated industrial process simulation
  • Techno-economic analysis
  • Generative AI for novel materials
  • Adaptive Bayesian optimization
The materials

Metal-organic frameworks: crystalline sponges, designed atom by atom

MOFs are advanced porous materials (100,000+ reported structures) whose surface area and chemistry can be tuned to a specific gas and process. One gram can hide the surface area of a football pitch.

<1 GJ/t

Target CO₂ capture energy

Versus 3.5-4 GJ per tonne for today's amine-based plants: about a 70% reduction.

3× lower

Storage pressure

Adsorbent-based storage reaches the same capacity at a fraction of the pressure, cutting tank, compressor, and infrastructure cost.

TRL 6

Platform maturity

Applied in paid industrial work and validated against laboratory measurements.

See the pipeline run on your problem.

Bring a feed composition and a performance target, and we'll show you what the joint space holds.

Contact us