This is a continuation of a different thread which started with the transcription of Renzo Mondaini's plasma electrolysis experiments and ended up with partial replication on my part (mainly using a cheap low-power high-voltage power supply) as well as doing related experimental research.
The main problem so far has been finding a suitable power supply for experiments at higher currents. It is not easy to find affordable regulated power supplies for these experiments, and high voltages always pose some risks.
Recently I realized however that by using considerably higher electrolyte concentrations than usual (about 15–20% KOH by weight), cathodic plasma electrolysis can start from as low as 40V, with breakdown voltage (the voltage above which current starts decreasing with voltage, signaling the transition from normal electrolysis to plasma electrolysis) around 18–20V. This means that in principle it's possible to use commonly available 0–60V / 5A adjustable bench power supplies for these experiments, at least up to moderate power.
I only have a 31V bench supply at disposal, but due to the floating outputs it is easy to use it in series with another fixed-voltage source. I happened to have a 19V / 4.5A power supply, so I am able to test the 0–50V range up to 4.5A. Below (solid region) is the current-voltage relationship of the cathodic reaction using a 1 mm-thick sharp tungsten electrode. It appears that current reached a minimum at about 45V in this test, and should only increase after that. This minimum is technically called mid-point voltage in related literature.
The non-linear results are similar in general behavior to those reported by others, for example as seen in these graphs below from https://link.springer.com/article/10.1007/s11090-017-9804-z (open access) and https://doi.org/10.1088/0963-0252/26/1/015005 (paywalled), respectively. At least up to 50V the current value reported by my bench power supply appeared to be in agreement with that of a 40A clamp meter, for what it's worth.
Here is a test with the same tungsten cathode, switching between 19V and 50V repeatedly. Gas production seems much greater at 19V where normal electrolysis should be taking place.
An interesting observation with this relatively high-current, low-voltage plasma electrolysis is that by using a tungsten rod at the anode, oxidation and removal of the oxide layer formed by the plasma is so fast that the electrode becomes cleaned up and turns quite sharp in the process. This could be a useful method for obtaining tungsten needles. Other materials tested (so far only nickel, titanium, kanthal A-1) did not seem to behave in the same way during anodic oxidation.
As far as reaction products go, I haven't been able to detect anything unusual using a CMOS "cosmic ray detector" that has been collecting background data for weeks. At the maximum voltage I can use at the moment, electrode wear is rather limited and the reaction can go on almost indefinitely if it wasn't for electrolyte evaporation, and electromagnetic noise from prior testing appears to occur when wear starts getting severe, thus at higher voltages than the mid-point voltage.
The cathode reaction however appears to be subjectively putting proportionally more heat than normal electrolysis does at a much higher input power into the electrolyte, although to be fair I only have a single temperature sensor, so this is far from a reliable indication of excess heat (example: electrolyte temperature during plasma electrolysis of 75 °C, whereas with regular electrolysis it's about 81 °C, but with more than 3 times the total input power. Odd). The plasma reaction also appears to aerosolize large amounts of electrolyte, so this would have to be taken into account when performing evaporation calorimetry.
So is this method better than using high voltage power supplies and standard electrolyte concentrations? On one hand it does allow using more affordable and safer power supplies, but on the other a highly conductive electrolyte is needed and there aren't many other choices available besides KOH and NaOH, which are dangerous to use at high concentrations. K2CO3 could be used, but the maximum conductance (at 34 wt.%) is less than half that of KOH, according to a document I found here: