# A miniaturized optical sensor for particulate matter detection

Abstract : We present an innovative optical particulate matter sensor. This optical sensor ‘on-a-chip’ combines a visible fibered light source and a custom-made CMOS image sensor chip. By illuminating a single particle in an air channel, we can record the light scattering signature on the photodiode matrix. A piece realized in 3D printing achieves fiber alignment and an efficient stray light protection. A specific scattering pattern occurs from the interaction of light with a single particle. Unlike traditional optical PM sensors based on a single photodiode detection, we measure a lens-free projection of the scattering signature on the nearby image sensor (1.5mm projection distance). This allows us to count particles and determine their size and refractive index. These parameters are retrieved through image processing and by comparison with a radiometric model that calculates the projection of a Lorenz-Mie’s scattering pattern. We describe the sensing technique, the architecture and fabrication of this sensor as well as the characterization results, which are in good agreement with our theory-based predictions. In particular, we show that it is possible to differentiate calibrated particulates of different sizes (monodisperse polystyrene-latex spheres). The sensor is sensitive enough to detect single particle and smallest than 1$\mu$m.
Keywords :
Document type :
Conference papers
Domain :

Cited literature [15 references]

https://hal-cea.archives-ouvertes.fr/cea-02432952
Contributor : Gabriel Jobert <>
Submitted on : Wednesday, January 8, 2020 - 5:14:18 PM
Last modification on : Wednesday, July 8, 2020 - 12:43:19 PM
Long-term archiving on: : Friday, April 10, 2020 - 12:06:16 AM

### Files

ProcSPIE_MF_hal.pdf
Files produced by the author(s)

### Citation

Maryse Fournier, Pierre Barritault, Gabriel Jobert, Adrien Marchant, Salim Boutami, et al.. A miniaturized optical sensor for particulate matter detection. Photonics West 2020, SPIE, Feb 2020, San Francisco, United States. ⟨10.1117/12.2546128⟩. ⟨cea-02432952⟩

Record views