Low-light imaging applications in astronomy, fluorescence microscopy, and quantum sensing, among others, require photodetectors with single-photon detection and photon-counting capabilities. Existing ultra-low noise detectors, such as SPADs (Single Photon Avalanche Diodes), EMCCDs (Electron Multiplying Charge Coupled Devices), and APDs (Avalanche Photodiodes), have disadvantages including high dark count and after-pulsing, excess noise, high voltage operation, and low fill factor. These fundamental limitations significantly impact image quality and scalability, encouraging the pursuit of alternative solutions, particularly in CMOS technology.
CMOS Image Sensors (CIS) have evolved significantly over the past twenty years, replacing CCD technology in most imaging applications due to advantages in power efficiency, readout circuit integration, and scalability. Early CIS designs were limited by high noise and low sensitivity; however, innovations such as pinned photodiodes (PPD), Correlated Double Sampling (CDS), and advanced pixel architecture have enabled CIS to rival and surpass CCD performance in many domains, notably in achieving lower noise levels. Today, CIS technology is the backbone of modern imaging, from consumer cameras to scientific instruments.
This presentation explores advancements in CIS designed to achieve readout noise as low as 0.3 , a critical threshold for quantum-limited detection. We review the state-of-the-art device and circuit techniques in CMOS Active Pixel Sensor (APS) design, including high in-pixel conversion gain, and strategies to reduce the dominant flicker noise from front-end readout circuits.
30/1/2026 11:00 - 12:00
ESAT Aula L
