Channel electron multiplier as an electron source

[Christoffer B]

TL;DR Summary

Channel electron multiplier as an electron source

Hi!

I'm playing around with a type of electron (and by electron/molecule interactions: ion) source that was briefly touched upon in the 1960's.

The basic idea is to use a channel electron multiplier "channeltron" with the anode removed; similar to a single MCP channel. The burst of electrons out will then be observable as a pulse on the output end of the channel, similar to extracting the dynode signal from a PMT.

The initiator of the burst can be anything from a light source producing photoelectrons, to a beta source like Sr-90. I went for the former.

I have modified a commercial channeltron by removing its anode cap, and provided negative bias at the cone, but no pulses are observed.

The channel has a built-in resistor to make sure the electrons don't reach zero potential within the channel, before they hit the anode. In my setup this can be shorted if needed.

However both with Rint shorted or not, I see no output pulses, regardless of light strenght (a quartz window provides light to the cone)

Please see attached image and schematic of the setup. I am unsure if it's an electronics problem or a fundamental issue.

Pressure is ~5E-7 Torr, bias is approx. -3.5kV.

The signal has been observed both directly on oscilloscope, through a preamplifier, and through a pulse inverter and preamplifier. same disappointing result.

Thank you for the interest!

Chris

 

[tech99]

Chris
It looks as if the device is de-mountable, so have you the necessary high vacuum?
Also are you using very short pulses. You will need to arrange the output circuit/load resistor so high speed is obtainable.
Does the light pulse fall on a photo emitter?

 

[Christoffer B]

Thank you for your reply!

I should have explained, one of the flanges goes to a turbomolecular pump and ion gauge - working pressure is about 10E-6-7 mbar.

The plan is not to use pulses but to use so low light flux that the photoelectrons generated are in the range of 100's /second. The emissive surface is the entrance cone of the multiplier.

I have tested this previously with an unmodified CEM, showing decent control over count rate by adjusting light. This one however is a different model, and modified as mentioned. Perhaps it is treated to avoid photoelectron generation, as that would just be noise in many applications.

I was also concerned photoelectrons would be repulsed by the large negative voltage, and be attracted to the nearby chamber wall out of the CEM, instead of into it, so I tried with reverse polarity; ground at the cone and +4kV at the output. No difference.