AI workflows for materials R&D

From quantum materials to reliable chips.

Matter42 turns messy spectroscopy, microscopy, documents, and simulation outputs into an agentic research workspace for 2D materials characterization and manufacturing decisions.

Read the docs

Sample workspace

WS2 PFIB Raman map

parsed

spatial mapx=-7.1, y=4.8 um

Explore

Cluster

Estimate

E2g linewidthagent tool

Defect density map ready

885 valid Raman pixels

Suggested next step

Compare defect type classification against the clean interior mask before making process conclusions.

Platform

One workspace for characterization, context, and simulation.

Matter42 gives materials teams a shared place to upload multimodal evidence, ask reproducible questions, and keep the scientific trail attached to the project instead of scattered across notebooks and one-off scripts.

Upload research data

Bring Raman, PL, microscopy, tabular measurements, and papers into one project workspace.

Multimodal intake

Ask scientific questions

The agent keeps project context, cites files, and calls domain tools instead of leaving analysis in a generic chat.

papermaprun

Context preserved

Map defect populations

Explore spectra, cluster regions, estimate defect density, and classify likely defect families.

Spatial maps

Connect to process knobs

Use simulation and structured outputs to reason from growth conditions to measurable material quality.

growthquality

Process signal

Workflow

A cleaner loop from raw data to next experiment.

See the research workflow

Collect

Create a project and upload the raw files that define a sample, experiment, or growth run.

Raman mapPL filepaper

Analyze

Let the agent parse data, inspect maps, run defect tools, and return figures with structured outputs.

parseclusterestimate

Decide

Compare regions, document caveats, and turn characterization into the next process experiment.

comparecitenext run

Agent analysis

From raw spectra to defensible defect maps.

Upload Raman or PL data and get interactive maps, spectral inversion, defect estimates, and caveats that stay tied to the evidence instead of disappearing into a one-off notebook.

Raman pixels

885

Mean density

1.38%

Correction

10.0 cm^-1

WS2 PFIB Raman map

Defect density estimation

Raman inversion

Loading Plotly figure...

Hover the maps to inspect local linewidth and density estimates. The table keeps calibration status visible.

Tool output

Mean E2g FWHM: 9.2 cm^-1. Best estimate: 1.4% S->O substitution. Valid Raman pixels: 885.

Defect typeEst.Status

S->O substitution1.37%interpolated

W vacancy5.05%interpolated

S vacancy6.08%interpolated

S->C substitution4.00%above calibration

Vision

Matter42 proposes a transformative AI‑driven framework to accelerate the discovery, characterization, and manufacturing

As traditional silicon reaches its limits, complex new architectures face critical production bottlenecks. Matter42 solves this by integrating Agentic AI, multimodal analytics, and physics‑ based models into a unified, closed‑loop manufacturing workflow.

MetricGlobal Semiconductor Industry2D Materials Niche (Graphene/TMDs)

Current Scale$0.8 – $1.0T~$1.02B (< 0.2% share)

Projected (5-10y)$1.0T+$1.3 – $1.5B

Primary DriversAI Infrastructure & ElectrificationTransition-Metal Dichalcogenides

Team

Built by scientists and operators who understand experimental complexity.

Matthias Kling, PhD

Co-Founder

Physicist with 20+ years expertise in ultrafast electronics, nanophotonics, and quantum physics. Experience in leading $100M+ science programs. Passionate about advancing the frontiers of technology with AI for solving urgent grand problems.

Thomas Linker, PhD

Co-Founder

Physicist with 8+ years expertise in multiscale simulations and ML for atomic scale control and characterization of materials. Experience in leading combined experimental and theoretical science campaigns on complex systems.

Kamila Stepniowska

Founding Team Member

Evangelist in AI and data science, 10+ years experience in operations. Experience in building partnerships in tech industry & open source community. Partnered with Intel, Loreal, United Nations, Orange.

Founding experience: In Browser AI, Ginger Tech, GC

Supported by

Backed by a community of builders, researchers, and company creators.

Let's talk

Have a dataset, workflow, or materials problem in mind?

We can help you map a first experiment, evaluate the current tool fit, or discuss how Matter42 could support your research team.

Explore docs

AI workflows for materials R&D

From quantum materials to reliable chips.

Matter42 turns messy spectroscopy, microscopy, documents, and simulation outputs into an agentic research workspace for 2D materials characterization and manufacturing decisions.

Read the docs

Sample workspace

WS2 PFIB Raman map

parsed

spatial mapx=-7.1, y=4.8 um

Explore

Cluster

Estimate

E2g linewidthagent tool

Defect density map ready

885 valid Raman pixels

Suggested next step

Compare defect type classification against the clean interior mask before making process conclusions.

Platform

One workspace for characterization, context, and simulation.

Upload research data

Bring Raman, PL, microscopy, tabular measurements, and papers into one project workspace.

Multimodal intake

Ask scientific questions

The agent keeps project context, cites files, and calls domain tools instead of leaving analysis in a generic chat.

papermaprun

Context preserved

Map defect populations

Explore spectra, cluster regions, estimate defect density, and classify likely defect families.

Spatial maps

Connect to process knobs

Use simulation and structured outputs to reason from growth conditions to measurable material quality.

growthquality

Process signal

Workflow

A cleaner loop from raw data to next experiment.

See the research workflow

Collect

Create a project and upload the raw files that define a sample, experiment, or growth run.

Raman mapPL filepaper

Analyze

Let the agent parse data, inspect maps, run defect tools, and return figures with structured outputs.

parseclusterestimate

Decide

Compare regions, document caveats, and turn characterization into the next process experiment.

comparecitenext run

Agent analysis

From raw spectra to defensible defect maps.

Upload Raman or PL data and get interactive maps, spectral inversion, defect estimates, and caveats that stay tied to the evidence instead of disappearing into a one-off notebook.

Raman pixels

885

Mean density

1.38%

Correction

10.0 cm^-1

WS2 PFIB Raman map

Defect density estimation

Raman inversion

Loading Plotly figure...

Hover the maps to inspect local linewidth and density estimates. The table keeps calibration status visible.

Tool output

Mean E2g FWHM: 9.2 cm^-1. Best estimate: 1.4% S->O substitution. Valid Raman pixels: 885.

Defect typeEst.Status

S->O substitution1.37%interpolated

W vacancy5.05%interpolated

S vacancy6.08%interpolated

S->C substitution4.00%above calibration

Vision

Matter42 proposes a transformative AI‑driven framework to accelerate the discovery, characterization, and manufacturing

MetricGlobal Semiconductor Industry2D Materials Niche (Graphene/TMDs)

Current Scale$0.8 – $1.0T~$1.02B (< 0.2% share)

Projected (5-10y)$1.0T+$1.3 – $1.5B

Primary DriversAI Infrastructure & ElectrificationTransition-Metal Dichalcogenides

Team

Built by scientists and operators who understand experimental complexity.

Matthias Kling, PhD

Co-Founder

Thomas Linker, PhD

Co-Founder

Kamila Stepniowska

Founding Team Member

Founding experience: In Browser AI, Ginger Tech, GC

Supported by

Backed by a community of builders, researchers, and company creators.

Let's talk

Have a dataset, workflow, or materials problem in mind?

We can help you map a first experiment, evaluate the current tool fit, or discuss how Matter42 could support your research team.

Explore docs