> controls the brightness of the image). In the Bio-Rad, the gain controls
> the voltage applied to the PMTs. And as far as I understood from
>  this discussion different PMTs can reach maximum sensitivity at different
> input voltage (gain). In this case, what I expect from an "enhanced PMT" is
> one that reaches the highest sensitivity at a lower voltage input. Otherwise
> I have to increase the gain and, as we all know, above a certain gain the
> S/N ratio starts to decrease.
>
I thought that the enhanced PMT provides a higher probability of
photon detection  by causing the photons to hit the photoemissive
surface more than once. This should increase the quantum
efficiency. After that, the PMT is like any other, as far as electron
multiplication and all the other good things that turn that
photoelectron into a pulse which is counted.

A "better" PMT may also produce less thermal noise (ie, false counts).

> applied to the PMT. And if an "enhanced PMT" needs a gain setting of 650
> to produce the same brightness that an old PMT could reach at 300 (in a scale
> from 0 to 1000) then I'm afraid they're quite useless. So probably I should
> rephrase my initial question to the following : can we expect an "enhanced PMT"
> to have a sensitivity increased by a factor of 2 (or more) at similar
> voltage inputs
> when compared to a "standard PMT" ?
>
If you are photon counting, then it should not matter what the gain
is, as long as you get a pulse when the photon hits which can be
discriminated from the background noise. Once you have enough PMT
voltage to detect your photons, you dont want more because that
increases the chance that a thermal electron that comes from a dynode
will generate a count.

If you are in the analog mode, it is probably because you have so much
light that you cannot photon-count, and then what matters is that the
other noise sources (electronic) are low. A tube which needs more
voltage to produce a given photocurrent may also produce less dark
current, even at the higher voltage. So again, the question is
whether the new PMT ("gain" 650) produces a noisier image than the old
PMT ("gain" 300).

BTW, all the comments on S/N ratio of photon counting assume that the
photon counter only responds to true photon "hits", with some quantum
efficiency. Is it really the case that the photon counting options on
our systems, which dont use cooled PMTs (at least the MRC-1024
doesnt), have negligible thermally generated counts?

--aryeh
Aryeh Weiss                          | email: [log in to unmask]
Department of Electronics            | URL:   http://optics.jct.ac.il/~aryeh
Jerusalem College of Technology      | phone: 972-2-6751146
POB 16031                            | FAX:   972-2-6422075
Jerusalem, Israel                    | ham radio: 4X1PB/KA1PB