Deciphering multispectral optical signals exposed on a capacitive CMOS platform

BACKGROUND

Photodetectors that convert incident light into electrical signals are essential in many applications like motion detectors, radiation detectors, IR and UV sensors, and certain image sensors. Color filters or scintillator materials can be used to make them sensitive to light of certain wavelengths. The read-out electronics usually contains amplifiers and a CMOS microcontroller (to process, analyze/interpret and/or transmit the signals). Ideally all components are monolithically integrated in the same low-power low-cost CMOS technology. However, this requires adding additional manufacturing steps to the CMOS process and bringing new CMOS-incompatible materials into the manufacturing cleanroom. If possible/allowed at all this would drastically increase the manufacturing cost and limit the versatility of the solution. Avoiding this dilemma requires a disruptive breakthrough. Recently, F.P. Widdershoven et al. have developed a CMOS Pixelated Capacitive sensor array for biosensing application.¹ We are exploring similar sensors to detect wide spectral-range light. Our aim is to develop a smart CMOS-based  hyperspectral (ultraviolet, visible, infrared) image sensor with embedded real-time data analysis and interpretation by novel machine learning algorithms.

RESEARCH QUESTION

Can we decipher multispectral optical signals detected by  a CMOS Pixelated Capacitive sensor at real-time?

BACKGROUND

Photodetectors that convert incident light into electrical signals are essential in many applications like motion detectors, radiation detectors, IR and UV sensors, and certain image sensors. Color filters or scintillator materials can be used to make them sensitive to light of certain wavelengths. The read-out electronics usually contains amplifiers and a CMOS microcontroller (to process, analyze/interpret and/or transmit the signals). Ideally all components are monolithically integrated in the same low-power low-cost CMOS technology. However, this requires adding additional manufacturing steps to the CMOS process and bringing new CMOS-incompatible materials into the manufacturing cleanroom. If possible/allowed at all this would drastically increase the manufacturing cost and limit the versatility of the solution. Avoiding this dilemma requires a disruptive breakthrough. Recently, F.P. Widdershoven et al. have developed a CMOS Pixelated Capacitive sensor array for biosensing application.¹ We are exploring similar sensors to detect wide spectral-range light. Our aim is to develop a smart CMOS-based  hyperspectral (ultraviolet, visible, infrared) image sensor with embedded real-time data analysis and interpretation by novel machine learning algorithms.

RESEARCH QUESTION

Can we decipher multispectral optical signals detected by  a CMOS Pixelated Capacitive sensor at real-time?

Contact person: Suman Kundu; S.Kundu@tudelft.nl

Assignment

OPPORTUNITIES

•       Work on an innovative multidisciplinary research topic.  

•       Work closely with PhDs and postdocs and get exposure to an academic research environment.

•       Gain expertise in analytical and experimental techniques.

•       Learn to give quality presentations and scientific writing to give value to your work.

Requirements

YOUR TASKS

•       Carry out a literature review on existing photoactive materials and CMOS based photodetectors.

•       Working with an optical measurement setup and acquire real-time sensor data with a microcontroller-based read-out system.

•       Establishing different machine learning algorithms to decode and predict multispectral optical signals.

Contact

Last modified: 2023-11-24