It's always nice to see these experiments.

Thanks for showing us your experiment, it's nice to see the sodium-mercury segregation in such a long discharge tube! Have you measured the current fed to the lamp? The overheating of the mercury-rich part of the burner depends on that current and on the voltage drop over that section. Chances are you did not damage it, assuming also you did not run it for hours in that state, which I would certainly not recommend (the SHCP I experimented on was a spare lamp and it survived the treatment).

On the subject of your lamp, it's interesting that GE still produced this T18 tubular variant of their high-voltage 1 kW Lucalox (i.e., the ANSI S52XA-1000 type) so long after they replaced it by the now-standard E25-jacketed version in North America (i.e., the ANSI S52XB-1000 type released in 1978). This upgrade coincided with that of the burner to a 3rd-gen design featuring a crimped amalgam reservoir permitting a stable operation at all positions. I have a T18-jacketed LU1000 from 1978 with that new burner, which was available in the USA upon special request at that time (until 1981), but I wasn't aware the production of that type continued well into the 1990s (and possibly later). What info do you have about that particular lamp?

Sammi - I love doing experiments, just want to do them all the time.

Max - I had been thinking of making a low-pressure tube with neon and mercury for sometime to try to migrate the discharge with DC to show the two spectral colours in one tube. I haven’t tried it due to the chance of dragging mercury vapour through my vacuum equipment by accident. I already did this with iodine in a previous event . I know the process of adding mercury in from a bulb in the tubulation but I can see that going wrong with me easily. Anyway, when I saw your demonstration with the HPS lamp the penny dropped that a long tube HPS lamp would demonstrate this perfectly without the risk of building a mercury filled tube, hence the excitement.

The whole experiment with the HPS lamp probably totalled less than 5 minutes. Noticing the temperature rise actually was the reason for it not running any longer. Unfortunately, I didn’t make any current or voltage measurements and the lamp was under running on the 400W ballast. I will revisit this again at some point with a different lamp at some point though.

I have very little on this lamp. I found it on an eBay listing, at a good price, a number of years ago. I always assumed someone had bought it as a ‘recreational’ grow lamp then discovered it would not work on a normal UK 1kW HPS ballast. It doesn’t seem to have many hours on it either but the getter suggest it has some age. Unfortunately, one of the top spot-welded supports broke free from the frame. I have managed to wedge it at the top of the frame preventing it from rattling around inside. The lamp also incorporates a safety short-circuiting device at the base of the lamp. I believe this is to short the lamp if the discharge occurs in the vacuum jacket preventing complete destruction and socket damage. The date I added 0296 is as printed on the lamp and may be a code rather than a date.

Sodium arc tubes certainly have many uses beside general lighting, intended or not. Years ago, when I was a teenager, I repaired a professional strobe light by replacing the dead xenon lamp with a 400 W HPS burner (a new triac was needed also). It worked like a charm and that sure was a very economical fix, especially since I sourced the arc tube from a dead sodium lamp (still plenty of life left in the pulsed regime though )

On the subject of DC-driven plasma segregation, the effect can also be seen in mixtures of noble gases, preferably with a large difference in atomic weight and ionization energy. At the university we have a long demo tube designed to that end, it is filled with a neon-krypton mixture (krypton being the pumped species due to its lower ionization energy). If making such a tube is not possible, then you can also get creative with a long T8 fluorescent tube. While the segregation between mercury and the argon-krypton buffer does not result in a dramatic change in plasma color (it's there, but not as striking as the orange/blue difference of the sodium-mercury segregation in HPS lamps), difference in UV output will be seen clearly from the tube's fluorescence. The effect can be even stronger in last-gen T5 tubes (as introduced by Philips in 1994) as some of them contain a neon-based buffer, which will also result in a different light color, beside the change in intensity.

Good use of an old HPS arc tube. I have one I keep as it has the dopant that demonstrates fluorescent in the blue to UV. Might not make a good strobe with the persistence but I keep meaning to work out the mechanisms involved. Some say it’s UV induced but I haven’t managed to excite the alumina with an external UV source. I was wondering if it was electron bombardment but I’d need a broken one to put in a vacuum to test that.

Would love to play with gas tubes, but the noble gasses are expensive staring with neon and just increase to very expensive xenon. It’s a pity as there are so many unusual effects with mixes of these.

I know that the red fluorescence excited by blue light is caused by trivalent chromium ions, an impurity that comes from the manufacturing and processing of the ceramic tubes, but I don't know what exactly causes the blue phosphorescence. Since this is not a general occurrence, I can only suppose that this is from a specific dopant or contaminant. A spectral measurement could help identify the species responsible for that light emission.