PhD Thesis Defense: Xin Yue

For info on how to attend this videoconference, email xin.yue.th@dartmouth.edu.

"Design of a burst mode ultra high-speed low-noise CMOS image sensor"

Abstract

Ultra-high-speed (UHS) image sensors have been extensively used in various fields such as medical, scientific, and industrial applications to visualize and understand UHS phenomena. Recently, several published studies have successfully achieved frame rates of up to millions of frames per second (Mfps) for these specialized sensors. However, these studies have primarily relied on advanced processes like 130nm backside illumination (BSI) or customized processes to meet the specific design requirements of UHS image sensors. Therefore, there is a general interest in reducing image sensor fabrication costs and improving process compatibility.

This thesis presents an ultra-high-speed high conversion-gain CMOS image sensor (CIS) based on charge-sweep transfer gates in a standard 180nm CIS process. By optimizing the photodiode geometry and utilizing charge-sweep transfer gates, the proposed pixels achieve charge transfer time of less than 1 0ns without process modification. Additionally, the gate structure significantly reduces the floating diffusion capacitance, thus increasing the conversion gain. To demonstrate the effectiveness of the proposed design, a few pixels were modeled and simulated in TCAD. Finally, a proof-of­concept CMOS image sensor was designed, taped out and characterized.

This thesis covers the development and characterization of the burst mode UHS high conversion-gain image sensor and emphasizes the reduction of charge transfer time, improvement of pixel conversion gain in a standard process. The projected performance of this pixel enables the burst mode image sensor to run at 20 Mfps with better than state-­of-the-art noise ( <<8.4e-), which shows great potential in the cost-sensitive niche market.

Thesis Committee

• Eric Fossum (Chair)
• Jifeng Liu
• Kofi Odame
• Rihito Kuroda (Tohoku University)
• Zhehui Wang (Los Alamos National Lab)

Contact

For more information, contact Theresa Fuller at theresa.d.fuller@dartmouth.edu.