It was tough to write after a few years away from my Ph.D research, but finally my book chapter on "Single-Photon Avalanche Diodes in CMOS for Optical Communications" is now online through InTechOpen, with the print copy soon to be in Edinburgh University Library. It is part of a collection of papers collated in the book: "Optical Communication Technology", edited by Pedro Pinho, ISBN 978-953-51-3418-3... It is open access and free to download (below)... It is also CC-BY if anyone compains of difficulties in using figures from papers, indeed CC-BY is far easier a route for figure reuse than the restrictive terms of Elsivier etc.
The chapter mostly includes material and discussion from my thesis, but acts as a way of getting that material more visible, and while my thesis is available online, this hopefully will have more reach.
I've included discussion of the 'internal photoelectric effect', as there are old-school "vacume tube era" academics that are in the entirely mistaken opinion that it is, and I quote, "irrelevant" for silicon and semiconductor detection of light. What these academics fail to understand is that the phrase, 'internal photoelectric effect', is the de-facto nomenclature of the field to describe a photon of sufficient energy, promoting an electron from the valence band to the conduction band in a semiconductor. Indeed,, I've previously blogged about the use of this phrase in a) solid-state physics textbooks, b) engineering and cmos imaging textbooks, c) academmic papers over multiple decades and over the entire globe, d) legal patents, white papers and websites from international companies such as Sony, and e) even undergraduate notes from prestidious red-brick universities.
It really is a laughable position, principally as through my historical literature review for avalanche multiplication detectors, I've read many many journal papers which include the phrase. Perhaps if the academic conserned (I won't name names) is such a fan of Shockley and has used the p-n junction and transistors throughout their career, they care to read some of Shockley and Bell Labs original papers from the late 1940s and 1950s. And yes the phrase is there, and you can go back further to Nix in 1932. It is so common infact that I'll drop a few more, Christensen in 1978, Antoncik and Gaur in 1978 and Tada in 2005, Dirini in 2014. At the time, he forcefully told me a) I was wrong, b) that the photoeffect was "irrelevant" for silicon as the electron is not liberated from the surface (external version of the effect) and c) that I, and I quote, "shouldn't dig myself a hole". To retort here, perhaps the academic conserned should not cite wikipedia as it is certainly not academically rigorous (now updated, although not by myself BTW) and should not cite "a physics book" with no title, no author and no year. His behaviour that day was beyond unprofessional.
The bulk of the chapter aims to discuss three architectures for integrated CMOS SPAD receivers for visible optical communications (VLC). These are digital parallel summation of per-SPAD multi-bit counters, the XOR or NAND/NOR summation of SPAD pulses and the current-mode summation of current pulses. The chip developed during my Ph.D used an amalgamation of two of these approaches, however it is clear that there is some optimum between XOR summation and counter parallel summation based upon area usage, power and of course the pulse density at the final XOR gate in a large summation tree.
The current summation approach is interesting as they use the SPAD discharge current to develop a SPAD voltage of the order of 2.5V, but then use it to switch a 100uA current through one of two arms. Once dropped over the external load resistance in a similar manner to Si-PMs, the SPAD voltage of 2.5V becomes very small indeed. Unsuprisingly, it seems a shame to include an inherent signal attenuation just to read the SPAD signal out.
I'm in the process of writing another InTechOpen chapter, this time on the early historical development of avalanche mode devices such as avalanche photodiodes (APDs) and single-photon avalanche diodes (SPADs). This is because some of the existing historiccal SPAD development papers are either very poor (Renker 2005) or stop in the mid -1970's (Cova 1996). My chapter is covering the period of 1901 to 1969, covers descovery of photo-condustive effects, use of silicon, p-n junctions, Zener and avalanche breakdown mechanisms, Microplasmas as noise sources in Bell Lab's early transistors and the beginnings of single-photon and ionising radiation counting detectors.