Our increasing technological sophistication relies on our understanding of fundamental principles and enabling technologies. One such technology is Single Photon Avalanche Detectors (SPAD). These semiconductor diodes are commonly used in a wide range of applications ranging from fluorescence microscopy to quantum communication. We rely on our understanding of detectors and the fundamental physics behind them to make accurate measurements. This is why we decided to investigate methods to characterize these SPADs. In particular, we studied a common source of error called afterpulsing where the detector can click – emit an output signal corresponding to one photon – more than once per incident photon.
Afterpulsing is a well known defect in SPADs and different theoretical models exist to explain this behaviour. In the past, several different theories have attributed the cause of afterpulses to different fundamental properties of semiconductor junctions. Each of these theories has been published along with some experimental evidence to support it. Naturally, there must be a simple answer as to why all these different theoretical models based on fundamentally different physical principles were able to find supporting experimental evidence. Our investigations reveal this underlying truth.
We tested several individual detectors from each of 3 common manufacturers, observed their afterpulsing behaviour and compared this to the three most popular theoretical models explaining their behaviour. Given the vast amount of conflicting literature, we did not expect to find a single theoretical model that would best fit all our data. Naturally, what we expected to find is that certain theoretical models are best suited to certain manufacturers or certain types of detectors. Instead, we find that the best fit model for various samples of otherwise identical detectors can be widely different. The paper, published in Scientific Reports, details our findings and poses the question – If the behaviour of afterpulses is indeed linked to some fundamental principle behind these SPADs then why is there a clear and significant individual piece to piece variation?
The most likely explanation is the presence of some factor like associated electronics, batch to batch variation/manufacturing errors that cause this drastic difference. Which would also imply that the behaviour of afterpulses cannot be explained based only on simple fundamental properties of semiconductor physics. At this stage, you are probably wondering why earlier studies did not come to a similar conclusion. The answer is that almost all such studies were based on just a few very similar detectors. We used improved experimental techniques to allow us to study a large number of detectors and uncover a scientific puzzle that was so easily overlooked.