I have no idea how I found them. I remember browsing the Mouser catalog. At some point between finding screw terminal blocks and supercapacitors, they just caught my attention. Neon Lamps! They just grabbed my attention.
It’s coming back to me now. I was looking at gas discharge tubes for some pretty straight-forward oscillators. With a gas discharge tube, the voltage will reach a certain point before the tube starts to conduct. The GDTs at Mouser were too expensive for just a single-use application. This is where Neon lamps come in.
Neon lamps are a type of gas discharge tube. Two electrodes enclosed in a glass tube filled with a low-pressure Neon gas.
Now, this is where it gets really interesting. It might be possible to use these lamps as photon sensors. Depending on the material of the electrodes, a photon of a high enough frequency striking the cathode (the electrode of negative DC potential) of the tube would cause it to loose some electrons. This is called the Photoelectric Effect. These electrons then add to the electrons already on the cathode. When enough electrons form, they will cause a breakdown in the gas between the electrodes and the gas will glow. To make the lamp a detector, one should be able to hold the voltage between the electrodes just below the lamp’s DC breakdown voltage. Once a photon of high enough frequency strikes the lamp’s cathode and contributes enough electrons to make the voltage between the electrodes high enough, the gas inside the lamp will breakdown, glow, and conduct larger amounts of current, lowering the voltage between the electrodes. When the lamp is glowing, it will not be able to detect photons, so the detection circuitry will need to see the drop in voltage/increase in current of the lamp and shut it off, and then bring it back to the pre-breakdown voltage.
I wish I had more than the time I have spent writing this entry to try this out. I would be curious if anyone knows if have made any wrong assumptions or if they have tried this. This method of photon detection would not be very good compared to semiconductor methods, but it would be pretty cool to see two arrays of neon lamps, one a detector, the other a display of what the detector detected.