Technology
Two breakthrough technologies. One chip
Computational spectroscopy and monolithic integration combine to deliver broadband spectral insights with high accuracy, solid-state reliability, and scalable manufacturing at a fraction of the size and cost of traditional bulky-optics solutions. Built on years of pioneering research.
Computational spectroscopy
A broadband spectral fingerprint at once
ChipSense™ computational spectroscopy architecture captures 64 broad overlapping spectral bands in a single shot using only 16 photodetectors. The resulting feature-rich spectral fingerprints enable AI/ML-based prediction models with accuracies comparable to 1 nm resolution laboratory spectrometers.
Monolithic integration
Solid-State Reliability. Scalable Manufacturing.
Monolithic wafer-level integration delivers high SNR, high responsivity, outstanding repeatability, and solid-state robustness against vibration, shock, and temperature variations, while enabling scalable, cost-efficient manufacturing.
InGaAs photodetectors and optical filters are integrated directly at wafer level, eliminating filter alignment and assembly steps. This enables high manufacturing yield, consistent device performance, and reliable high-volume production at competitive cost.
Technology Considerations
Designing for optimal performance
ChipSense™ is designed for compact, high-performance spectral sensing across a wide range of applications. As with any optical sensing technology, achievable performance depends on the complete system design and application conditions.
Target concentration
As a rule of thumb, ChipSense™ is best suited for applications where the target compounds are present at concentrations typically above 0.1% (1,000 ppm). Higher sensitivity may be achievable depending on the optical configuration, measurement conditions, and application-specific AI/ML model.
Working distance
ChipSense™ is primarily designed for contact and close-proximity measurements, typically within 5 cm of the target. Longer working distances are possible but generally require increased illumination power, controlled ambient lighting, dedicated optical and mechanical design, and application-specific optimization.
Application optimization
Performance depends on factors including the target material, optical path, illumination, measurement geometry, and AI/ML model development. MantiSpectra works closely with customers and ecosystem partners to assess application feasibility and optimize system performance.
Intellectual property
Rooted in Research. Built for Impact.
MantiSpectra builds on years of pioneering research, a strong intellectual property portfolio, and scientific publications that underpin ChipSense™ technology.
Patents
- A multi-pixel spectral sensorUS11747203B2EP3935356A1
- Modular housing for a spectrometerNL2029194B1US12163836B2
- Spectral sensor system for harsh environmentsNL2032862B1EP4577825A1US20260071952A1
- Model calibration transfer frameworkN2039365
- Integrated spectrometer and opto-mechanical sensorEP3400427B1US10732035B2
Publications
- Integrated near-infrared spectral sensingNature Communications · 2021 — opens in a new tab
- Handheld NIR spectral sensor module based on a fully-integrated detector arrayMDPI Sensors · 2022 — opens in a new tab
- On-site illicit-drug detection with an integrated near-infrared spectral sensor: a proof of conceptTalanta · 2022 — opens in a new tab
- SLURP! Spectroscopy of Liquids Using Robot Pre-Touch SensingIEEE ICRA · 2023 — opens in a new tab
- Spectral sensing in the near-infrared and chemometrics for flour quality multi-analysisSPIE · 2024 — opens in a new tab
- Advancing Food Multi-analysis with Pocket-Sized Multi-spectral Sensors and ChemometricsSpringer — opens in a new tab
- Spectral Sensing Using a Handheld NIR Module Based on a Fully Integrated Sensor ChipSpectroscopy — opens in a new tab
Get started
Let's build the future of material intelligence
Get in touch to explore how ChipSense™ can enable your next distinctive solution.