part of your problem is going to be having the photodetector in the path of the photon. also, the energy level of the photon is very important. with a given detector, you may be able to detect a gamma or X-ray photon, but not one in the visible spectrum. one application where the ability to count single photons is important is in fusion research, where you want to be able to count how many fusion reactions take place. this is done by counting the gamma and xray photons emitted by a target in the path of the high energy neutrons from the reaction. many detectors in use will detect such photons, but the photons from visible light may not be energetic enough to be detected.
this is how people testing the "cold fusion" apparatus back in the 80's knew there wasn't any fusion taking place. the D2O in the apparatus should have emitted neutrons when the deuterium nuclei fused into helium. but without neutron emission, 1)the reaction is not sustainable, even if it does occur on a small scale, and 2) a byproduct of fusion reactions is neutrons, because some of the reactions will create 3He with the loss of a neutron and another byproduct of the reaction could be 3H (tritium), which will then fuse with 2H (deuterium) to form 4He, with, again, the emission of a neutron.
however, fusion in a Farnsworth-Hirsch fusor emits a LOT of neutrons, indicating there is fusion taking place in the reaction.