2
$\begingroup$

I have an old Hamamatsu PMT, and I have a scintillator that can detect neutrons, gammas, alphas. But no matter what source I use, I'm only seeing counts at the first few channels (up to about the 15th channel) and then no counts for any of the other channels. I've been using this PMT since a while ago, and previously it would show counts (for the same source and same conditions) upto channel 700 even. What could be an issue to investigate? I only have one PMT.

$\endgroup$
  • 1
    $\begingroup$ You've likely either lost optical coupling or lost a lot of gain in the PMT. You can dismount the PMT and text it in isolation, but the PMtT may be toast. Did you make mistake with it (HV on while exposed to room light, knocked around or other stuff)? $\endgroup$ – dmckee --- ex-moderator kitten Jul 28 '17 at 21:31
  • $\begingroup$ @dmckee thanks for your reply! The PMT is quite old, and I have been using it non-stop for about 2 months, and for sure have bumped it on a table once or twice, and had once accidentally exposed it to room light without turning off the voltage. $\endgroup$ – Betsy Jul 28 '17 at 21:43
5
$\begingroup$

When you expose a PMT with full light, you can seriously affect its gain. Years ago I was involved in a project where a PMT had its gain stabilized by monitoring the output due to a (dim) LED that flashed at 1 kHz. We noticed that the voltage needed to maintain constant output increased over time; when we reduced the intensity and frequency of the LED pulse, the "gain loss curve" slope decreased significantly.

The mechanism for this was believed to be electron bombardment of the dynodes - they cause changes to the surface of the dynodes, reducing the electron multiplying qualities.

The other thing that might have happened - you could have created a microcrack when you "knocked it". In that case, air enters the enclosure and this would show up as a discoloration of the PMT surface (for bialkali cathodes, it would go from light brown to clear).

The third thing that can sometimes cause gain loss (but this is usually a slower effect) is leakage current across the bulb (yes, I had to measure that to believe it). With the cathode at - 1500 V, mounted to a scintillator in a grounded enclosure, coupled with grease, we were getting fA of current through the envelope. This apparently caused ion migration (Na+ from the glass) that "poisoned" the cathode, resulting in a reduced gain. Adding a guard ring (at HV) made the problem go away.

And of course, it's possible that knocking the PMT+scintillator assembly would have knocked the scintillator off the PMT. If the light is no longer coupled efficiently, the peak will essentially disappear.

Three final ideas - HV supply is fried, or the preamp is. Or the preamp power is off... unlikely, but I thought I had to mention it.

It's hard to know which of these is your problem, but I hope it helps with your troubleshooting.

See also this earlier answer about PMT troubleshooting.

$\endgroup$
  • $\begingroup$ Thanks for your detailed answer! The coupling issue I can eliminate, because the PMT and scintillator come as separates, and I manually attach them for every use using optical grease. $\endgroup$ – Betsy Jul 28 '17 at 22:24
  • 1
    $\begingroup$ You're welcome! If you ever reach a conclusion let me know! $\endgroup$ – Floris Jul 28 '17 at 23:03
  • 1
    $\begingroup$ My HV was fried, so I replaced it again, and used a plastic proton-recoil detector to obtain the spectra from Gd-148. Having 1024 channels, the 40th channel is at 3200 KeV. Does that sound reasonable or should I still suspect my PMT? $\endgroup$ – Betsy Jul 31 '17 at 15:35
  • 1
    $\begingroup$ Glad I had listed "fried HV" as an "it's possible"... I don't have a good feeling for the scaling on your system. Obviously if you increase HV, you will change the scale of your spectrum. What is appropriate depends on a LOT of things, and I am not familiar enough with your components to comment on that. Sorry. $\endgroup$ – Floris Jul 31 '17 at 15:42

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.