< Back to previous page

Project

Single-photon detection based on charge multiplication for timeresolved imaging

A single-photon avalanche diode (SPAD) is a semiconductor device capable of detecting individual light particles, or photons, with high precision in time. The mobile and automotive industries increasingly adopt SPADs for depth-sensing systems based on the direct time-of-flight method, by which the distance to objects can be measured. This technique relies on the precise extraction of the arrival time of back-scattered photons from an illuminated scene. Spatial resolving capabilities can be attained by integrating the individual detectors into arrays. Furthermore, depth-sensing systems generally perform best for near-infrared (NIR) and short-wave infrared illumination wavelengths.

The fabrication of SPADs with silicon manufacturing methods has many advantages, including the availability of well-established processing technologies and the ability to achieve a low detector noise. Moreover, silicon single-photon diodes have the potential to be miniaturized and integrated into large-format arrays. Unfortunately, silicon exhibits a low absorption coefficient for infrared radiation. Therefore, extensive optimization of the SPAD structure is required to extend the sensitivity towards longer wavelengths and improve compatibility with NIR time-of-flight applications. Although many advances have been made in recent years, the design of NIR-sensitive silicon SPADs remains challenging. For example, attempts to increase the long-wavelength detection efficiency typically result in the degradation of the timing resolution or operating voltage. The constraints are even more pronounced when the devices are to be integrated into arrays, for which uniformity, energy efficiency, and scalability are also essential.

In this research, near-infrared enhanced silicon SPADs are studied, with a focus on the time-of-flight application. Existing detector solutions are analyzed, design constraints of typical devices are identified, and new SPAD architectures are proposed. A significant part of the investigation is based on the characterization of fabricated detectors. In addition, the work involves the numerical analysis of devices using existing and custom-made simulation tools. No actual large-format arrays are developed as part of this project. However, the eventual integration of such sensors is considered.

The outline and results of the research are as follows. Firstly, a standard planar NIR-enhanced SPAD is investigated, and the performance constraints of the device are identified. Secondly, the shortcomings of existing technologies are addressed with the development of proof-of-concept planar charge-focusing SPADs, which achieve a remarkable NIR detection efficiency and temporal response. Finally, two pathways for scaled SPADs are examined. The first approach involves the continued development of the charge-focusing architecture with smaller sizes and reduced operating voltages. The second direction concerns a novel non-planar pillar SPAD technology by which some of the typical constraints of planar devices are circumvented. Due to promising simulated performance and compatibility with arrays, the scaled technologies show great potential to advance the state-of-the-art of NIR-enhanced silicon SPADs for direct time-of-flight imaging.

Date:18 Oct 2017 →  7 Jun 2022
Keywords:single-photon imaging, CMOS, time-of-flight measurement, SPAD, image sensor
Disciplines:Design theories and methods, Semiconductor devices, nanoelectronics and technology
Project type:PhD project