Assessment of Photoelectron Number at Very Low Light Intensities

Pulse height distributions are often used to estimate the mean photoelectron number in a photomultiplier. For large scale photoelectron number this method is viable, provided the correct statistics for the tube are used and due allowance is made for fluctuations in the dynode multiplication process. For very low light levels, however, the distributions often become too narrow to be reliable and an alternative method would be a helpful supplement. This paper considers the procedure of accurately measuring the number of zero responses in a distribution and using this to obtain the photoelectron number. The results of this method are compared with the more tradition approach, and also with the output of a high sensitivity tube (RCA C70133B) where the photoelectron peaks are clearly visible