Bachelor's DegreesBachelor of ArtsBachelor of EngineeringPartner School Dual-DegreeUndergraduate Admissions
Doctoral DegreesDoctor of PhilosophyPhD Innovation ProgramDoctor of Medicine-PhDGraduate Admissions
All Thayer Events
PhD Thesis Proposal: Xin Yue
11:00am - 12:30pm ET
Rm 202, Cummings Hall/Online
For Zoom link contact firstname.lastname@example.org.
"Design of a burst-mode global-shutter ultra-high-speed high-conversion-gain CMOS image sensor"
Ultra-high-speed (UHS) image sensors are widely used in medical, scientific, and industrial applications to visualize and elucidate UHS phenomena. In recent years, a few published researches pushed the frame rate of these specialized image sensors to the range of millions of frames per second (Mfps). Most of these researches either used an advanced process, such as a 130nm backside illumination (BSI), or a process that was customized to meet the design requirement of the UHS image sensors. It is of general interest to lower image sensor fabrication costs and increase process compatibility.
This thesis describes an ultra-high-speed high-conversion-gain CMOS image sensor (CIS) based on sequential transfer gates in a standard 180nm CIS process. With sequential transfer gates and optimized photodiode geometry shape, photoelectrons in the proposed pixel can achieve less than 12ns charge transfer time without any process modification. The sequential transfer gate structure has also been proven to effectively reduce the floating diffusion capacitance and increase the conversion gain of the pixel. In this study, a pixel was modeled using the sequential transfer gates and simulated in TCAD, and a proof-of-concept CMOS image sensor was simulated in Cadence Virtuoso and taped out.
This thesis covers the development and characterization of the burst-mode ultra-high-speed high-conversion-gain image sensor and emphasizes the reduction of charge transfer time, improvement of pixel conversion gain, and reduction of charge transfer inefficiency. The projected performance of this pixel enables the burst-mode image sensor to run at least 20Mfps with better than state-of-the-art noise (<<8e-), which shows great potential in this cost-sensitive niche market.
- Prof. Eric Fossum( Chair)
- Prof. Jifeng Liu
- Prof. Kofi Odame
- Prof. Rihito Kuroda
For more information, contact Theresa Fuller at email@example.com.