Near-infrared reflectance spectroscopy is the gold standard for material identification in recycling, and Matoha puts it in your hand for the first time at sorting-floor speed and price.
nm wavelength range
for polymer identification
Full scan + AI classification
per sample
AI runs entirely on hardware
No internet required
Polymer-level identification
accuracy rate
Near-infrared light (Matoha devices operate at 1,550–1,950 nm) interacts with the chemical bonds in a material, C-H, O-H, N-H, and C=O bonds each absorb NIR energy at specific, predictable wavelengths.
The resulting absorption spectrum is unique to each polymer or fibre blend, as distinctive as a fingerprint. Matoha's AI model compares this spectrum against a library of thousands of reference scans to return a classification in under a second.
A tungsten-halogen or LED light source emits near-infrared light across 1,550–1,950 nm. The light penetrates the material surface and interacts with the molecular structure beneath.
The material's chemical bonds absorb specific wavelengths and reflect the rest. The reflected spectrum is captured by a photodetector in milliseconds.
The on-device AI model runs a nearest-neighbour classification against thousands of reference spectra. Blend percentages are calculated from the spectral weighting.
Material composition and blend percentages are displayed on-screen, transmitted to the app, and sent to the cloud, all within under 1 second of pressing the button.
Every major automated recycling system in the world uses NIR. Matoha brings the same proven technology to handheld, benchtop, and field use cases.
TOMRA, Steinert, BHS Sorting and every major conveyor NIR sorter in the world uses the same NIR principle. Matoha brings it to handheld and benchtop form.
NIR spectroscopy is the reference method in ISO textile composition standards and ASTM plastic identification procedures worldwide.
Unlike chemical analysis or burn tests, NIR leaves the sample completely intact. Scan the same item hundreds of times with identical results.
NIR reads molecular chemistry, not surface colour. Black polyester and white polyester return identical spectra, colour cannot fool the sensor.
Press the sensor window against the material and press the button. No cutting, dissolving, burning, or lab preparation of any kind.
Unlike RFID labels or visual inspection, NIR measures actual polymer percentages, 67% cotton / 33% polyester, not just "cotton blend".
Lab NIR, conveyor sorters, visual grading, and burn tests each have a role. Here's where Matoha fits, and where it doesn't.
10+ tonnes/hr on a conveyor? You need a TOMRA-class system. Matoha's S-Bench is designed for bench-level and small-facility throughput, not full-conveyor automation.
If you need regulatory-grade chemical composition certificates, a full-range lab spectrometer with trained chemists is required.
Highly contaminated samples or unusual polymer composites may require custom material training or specialist analysis.
The verdict: FTIR offers a broader spectral range and is essential for unknown-substance analysis. For known polymer families in recycling streams, NIR is faster, more robust, portable, and 10–50× cheaper, which is why every industrial sorter in the world uses NIR, not FTIR.
📖 Further Reading
Talk to our team and see the NIR identification process in action, textiles, plastics, or carpet, so you can see exactly how it works before you decide.
Talk to Our Team