PMT's are pretty amazing. They detect single photons, have very low noise, and have incredible internal gain (about 500,000 at 900V), so the tiny, tiny energy from that single photon (Planck's law) can actually be measured.
Mine are 2" units made by Hamamatsu. Their web site has a fair amount of information.
They have 27% quantum efficiency. This means that instead of a photon being nnn Joules, I get about a quarter of that.
Planck's constant (h) = 6.624E-34 Joule Second
At 555 nm (green light 5.4E14 Hz ) a single photon has an energy of about 3.6E-19 Joule..So even after the PMT bumps that up by half a million, the pulse still only has an energy of 0.18 picoJoule...
Stating it another way, the output of the tube is typically 50 Amps/lumen
Here is the data sheet for the R1306-01 (pdf from scanned image of fax)
Here is Richard Hull's preamp design and assembly notes (hullpmt.pdf)
For detecting particles, you use a scintillator that converts the particle into a flash of light, which the PMT detects
pmt.htm - 24 Sep 99, Jim Lux