Single-Photon Avalanche Diodes (SPADs) in modern small dimension CMOS processes give a singular advantage for optical communications. With further development they allow the gain required in traditional optical recevers to be shifted away from electrical circuits and to be placed within the optical sensing diode itself. This technique allows gains >10,000 to be implemented with the diode rather than using a large and often power-hungry TIA amplifier. SPADs have a number of issues that remain to be resolved, however they fundamentally allow the detection of single-photons, often with detection efficiencies in the 45% region and array fill factors now reaching the 70% region.
One would ask the question, can arrays of these devices be used for optical communications, and could they be a route to lower optical powers throughout our communications networks.
My work has demonstrated that they are feasible optical receivers, however a number of properties require further work before a high speed receiver would be suitable for modern requirements. We have therefore shown feasibility, discussed the limitations, the trends in the devices that would help propogate the technology and suggested design methods to abate some of the short term issues with the concept.
However, despite working at lower optical powers, the overall power per received bit will be difficult to scale, and will not be competetive with mature photodiode and trans-impedance amplifier solutions. With this in mind SPADs for communications, especially for more advanced transmission protocols such as OFDM or QPAM, may only be useful in niche, low-light or long distance communications.
Visible Light Communications for home/office use and communications in harsh industrial enviromnents are being considered for these devices at present.
I would point the reader to a) my Ph.D thesis and b) to the VLC work being caried out by Harald Haas' research group within IDCOM at Edinburgh University. Interested readers should also take a look at the work being carried out with SPAD-based systems for biomedical and 3D camera applications in Robert Henderson's group within IMNS at Edinburgh University.