Defect-aware inputs for digital twins

Put real material physics into every simulation.

MatProps delivers defect-aware material properties in the form of Transport Physics-Informed Neural Networks (PINNs). With real trasport physics laying the foundation of every Transport PINN, our models are built to extrapolate to conditions outside of what you can measure. You can use our flagship models directly, or we can train a Transport PINN exclusively for your team, calibrated to your measured data. Use transport values that represtent your real material, rather than the value of a pristine crystal found in a database.

Defective lattice / real material
Crystal lattice with highlighted defect sites

Your material is not the pristine crystal in a database.

Explore our defect-aware property predictions

How our models are delivered

Use the models wherever your work already happens.

Every model, simulation, and Transport PINN on MatProps is available two ways: interactively through the MatProps web application, or programmatically through an open API. Both serve the same models and return the same values.

The missing layer in smart manufacturing

Engineering teams spend millions building precise simulations on imprecise material data.

MatProps closes that gap — from the ground state up. We focus on continual improvement of Material Property simulations, with the goal of giving you the most accurate values possible. Your team focuses on interative engineering design while we focus on the materials physics. We relentlessly identify defects that matter, capture the physics, and deliver the values you need in your simulations, digital twins, and modeling.

The length-scale problem

From atomic physics to continuum engineering.

MatProps operates across the Ångström-to-micron length scale, bridging quantum mechanics and continuum simulation. Real trasport physics lays the foundation of every Transport PINN and is calibrated against real measured data.
If you don't have measured data, you can use our flagship models directly. If you do have measured data, our PINN models allow you to blend your internally measured data with fundamental transport physics, ab initio simulation, and the vast amount of published research. Measure data where you can and let fundamental transport physics extrapolate to conditions outside of your measurement capabilities.

Solutions, not software

Solutions built around your materials.

MatProps delivers the material properties your engineering work requires, tailored to the materials you actually use. Matprops is not another specialized software package to operate. Provide the information already available in specification sheets, incoming quality records, or metrology reports through a straightforward form. We run the model on your behalf. Since it is grounded in real trasport physics our models are built to extrapolate to conditions outside of your measurements.

We handle the physics and compute. You stay focused on engineering. To demonstrate MatProps capabilities, explore our Si Thermal Conductivity Transport PINN below.

Material inputs

  • TemperatureK
  • Sample sizeµm
  • Grain sizeµm
  • Boron concentrationcm⁻³
  • Phosphorus concentrationcm⁻³
  • More defect typesIn development

Have a model you think is useful? Allow us to be your distibution partner.

Make academic models usable—and valuable—beyond the lab.

MatProps builds infrastructure for academics to distribute their models and simulations to industry at scale. We help you host, package, and deliver physics that engineering workflows can actually use — and that researchers can license.

pip3 install matprops-sdk
matprops init my-sim
cd my-sim
matprops validate
matprops run --input example.input.json
matprops login
matprops publish
Model licensing + royalty program
Model licensing + royalty program

Host and distribute your simulation packages on MatProps. Reach industrial users, retain ownership of your IP, and earn royalties when your models power production workflows.