Plasma electrolysis at lower voltages

I repeated the suggestion by Stevenson using a closer distance between the electrodes and different cathode position so that it would draw material from the anode nickel wire "wool", and results (again using 10% HCl) were interesting.

Unfortunately experiment duration turned out to be limited. I used 24V and the the cathode was a 0.3mm tungsten wire. Nickel would homogeneously deposit on the tungsten wire as a spiky ball, and at the same time a glow (?) plasma would occur on its tips. Such Ni ball appeared to be quite reactive and readily burn in air at room temperature. Unfortunately the anode broke after a few minutes of testing.

(Video focusing improves after minute 0:33)

EDIT: overall, the reaction side effects (noise, light) appear to be considerably weaker than with alkaline electrolytes, even at 62V (the maximum I can currently apply).

EDIT 2: from the video, interestingly it appears that the anode has bright spots too during the reaction. I was focused on the cathode during the test and did not notice this until I reviewed the video.

Stevenson

For what it's worth, after that test I also tried using a 50 pence coin (3:1 CuNi alloy) as the anode, in the same 10% HCl solution, obtaining similar results. The solution acquired a darker yellow-green color.

I wasn't intensely focused on the power supply, but current at about 25–30V was in the order of 1.2A.

I noticed that there is an optimal voltage range for deposition to occur. Below 25V it's slow, and below 22–23V the reaction easily reverts to normal electrolysis although once the glow-deposition-plasma starts voltage can be sometimes decreased down to 17V. In the 25–31V range deposition is faster and the spikes tend to glow by incandescence; the lower end of this range seems optimal for the deposition process. Above 35V and up to 62V the deposited spiky ball formation gets destroyed and gets replaced by denser layer, of slightly larger diameter than the base tungsten wire. At the same time a blue plasma (possibly from chlorine) is developed, with emission of occasionally intense EMI.

I did not use laboratory-grade HCl, but according to the label it's 10% HCl. The bottle is relatively new.

I haven't noticed much from the Cosmic Ray Finder program. I track the number of events using a spreadsheet but there did not seem to be much clearly associated with the testing, and overall the curve shows a rate of 250 events/day as usual (the threshold is the minimum value that does not cause false events). The image on the right shows all the events composited on top of each other since about 9 hours ago.

Another non-obvious way along which a magnet can improve the reaction is by placing it behind the anode. Then, part of the ferromagnetic particles depositing and then falling off the cathode (nickel) will contribute increasing the surface area of the anode, making the cathodic plasma/deposition reaction less likely to fail as the cathode grows.

Quote from Stevenson

It would be interesting to wash the cathodic material, dry it, and gracefully place it on a CD for a couple of days: this material has all the ingredients that typically generate Strange Radiation...

I would need to come up with a good plan to do that properly (e.g. how to dry, how quickly, how to avoid external scratches to the easily scratched CDs/DVDs, etc)

Stevenson

A possible solution would be using an plastic sheet or standard DVD/CD case as an intermediate protection layer, but I do not know if this would affect the results as shown by Parkhomov and coworkers. There does not seem to be much detailed accessible information on the methodology; from my past attempts, both making sure that the disks do not get scratched and keeping track of existing light scratches proved challenging.

For most other researchers who have used CR-39 sheets to observe loosely similar particle tracks&spots, there is usually a need for the sheets to be very close to the cathode during active conditions, and to etch them in a temperature-controlled hot caustic solution.

EDIT: gif animation from the previous video, unrelated to the comment, showing how the reactive spiky surface appeared to burn/combust in air. I tried looking at it closer once and it had a dark green color (presumably from NiO). In later attempts I used Ni25Cu75 alloy and the color was different (blue-gray, similar to how coins made of the same material look when oxidized).

Quote from can

[...] the reactive spiky surface appeared to burn/combust in air. I tried looking at it closer once and it had a dark green color (presumably from NiO)

Here is a photo of that.

When the solution appears to have a large amount of metal chlorides dissolved in it, growth up to a certain size is particularly fast; here with about 30V applied:

Here is another showing how switching between 31V (where deposition takes place) and 62V causes such deposition to get destroyed. Before it does, however, the plasma reaction looks stronger and produces a brighter light. In the final quiet state there appears to be a relatively large amount of electromagnetic noise getting produced, oddly.

I see this electromagnetic noise from the appearance of a large "hole" in my DSL connection spectrum, which gets logged by my DSL router (Fritz!Box 7590); at the same time I get "unrecoverable errors". High-speed DSL connections are sensitive to electromagnetic disturbances.

I used to have these often with the previous high-voltage power supply and cathodic plasma in diluted alkaline solutions at high voltages (800V). When the noise is bad enough, disconnections occur, causing new "synchronization" events. None occurred yet, but nearby electronics may malfunction or cause issues. Though, the webcam CMOS "cosmic ray detector" did not seem to be affected yet.

I haven't tried higher voltages yet; I have about 10V headroom (up to 72V total). Alternatively if the HCl solution could be more concentrated that could possibly help, but chlorine tends to evaporate away from the solution with time and heat (as well as getting sequestered by metal chloride formation), so it will progressively lose potency.

Given past observations, I had a suspicion that adding slight amounts of KOH to the HCl/NiCl2 solution formed would make the plasma brighter, although to keep the solution acidic compensation by the addition of more HCl would be necessary. I added about 1g KOH and 10.5g 10% HCl to the previously used solution.

This was the visible result:

After turning on the first power supply at 31V, I engaged the other for a total of 62V. I then repeatedly switched one of the power supplies on-off. At 31V dendritic growth occurs on the cathode.

As expected, the plasma at high voltages became much brighter, noisier and yellower, but Ni deposition at low voltages became slower and I couldn't obtain large dendritic Ni balls as earlier, not even by manually operating the cathode. Furthermore, the electromagnetic noise at high(er) voltages surprisingly became greatly reduced.

Compare the right portion of the top graph here with the previously posted one:

As for heat, I am not sure. The solution does appear to heat up quickly, but NiCl2 formation should be exothermic and it could be the main reason why this occurs. Much of the nickel deposited on the cathode appears to combust also, and the NiO there would readily turn into NiCl2.

EDIT: I tried looking at the "Cosmic Ray Finder" application events, and although the average seems somewhat elevated today (for now at least), the actual number of events did not increase while I was experimenting, so it could still be part of normal background variation.

Stevenson

Regarding that, here at minute 7:45 in this older ColdFusionNow interview, Pamela Ann Boss said that they tried with nickel chloride as the deposition source, but it did not produce results, even if it produced a dendritic plating:

https://www.coldfusionnow.com/…s-Cold-Fusion-Now-016.mp3

Quote

[...] Using these detectors we showed that the palladium deposit was the source of the tracks; control experiments using either copper chloride or nickel chloride in place of the palladium chloride did not produce tracks in the detectors. And in all three systems the metals plated out in the presence of dissolved (?) deuterium gas, and all three systems formed a dendritic plating. The only thing that was again (?) different is that palladium loads the deuterium, copper and nickel do not.

I don't know however if they examined the dendrites at a microscopic level. It's possible that more than nuclear effects the pits could be due to the very large electric field concentration occurring from the dendrite tips, from which a visible plasma is generated at low voltages. Different materials may lead to differently sized nanofeatures, but this will not be apparent just by eye.

Furthermore, this also means that higher applied voltages would lead to a stronger effective electric field, so samples that do not work during normal electrolysis may show an effect if voltage is increased high enough to cause a visible plasma.

Rob Woudenberg

No, either I missed or ignored it due to Pd and D₂O being used. (EDIT: upon having a look at the paper, it looks like I just forgot about it)

I just had a very quick look at the related paper uploaded by Ahlfors a few comments down the one you linked and the authors do point out that the electric field strength at the tip of the dendrites could be in the order of 10^9 V/cm.

A more practical observation on this point is that in order for a visible plasma to occur in the tests I described in this thread, the breakdown voltage of the chaotic H₂O/H₂ gas layer formed around the cathode has to be overcome. This layer should be in the order of 0.2–0.5 mm thickness with atmospheric pressure or more. In the plasma electrolysis literature, electric field strengths in the order of 10⁵–10⁸ V/m have been sometimes pointed out.

Just wanted to report that the same process using carbon steel (iron) at the anode does not seem to produce large dendritic deposition spheres as with nickel or cupronickel (not even after placing ordinary steel wool there), but appears to visually heat up more, producing stronger incandescence. A significant contribution to this could be due to oxidation.

Adding some alkali impurities (in the form of KOH, which reacted with HCl to form KCl) appeared to help forming dendrites on the cathode also above the surface, likely due to the potassium as a helper.

After this, as the electrolyte kept evaporating, occasionally small explosions could be heard and seen in the dendritic formation under the surface; I think that was due to alkali electrolyte getting segregated in the deposition layer and decomposing to alkali metal similarly to what occurred earlier on in strongly alkaline conditions using sodium or potassium carbonate. This is just my supposition, though.

Alan Smith

From a quick read, it seems mostly concerned with the breakdown of dielectric liquids under the direct application of high voltages (many kV) with moderately sharp electrodes.

I think the circumstances in the experiments described in this thread are different, in that the electrolyte is instead highly conductive—it could be considered a liquid electrode—and the dielectric (insulating) volume subjected to breakdown is the gas layer (normally H₂ and H₂O vapor) formed around the active electrode/cathode with electrolysis at elevated voltage.

Most importantly, in my tests the high electric field allowing a plasma to occur at all despite the low voltages used (30–40V or less) supposedly comes either from nano-sharp dendrites (at low pH and low voltages), and/or from the large amount of positive ions accumulating at the electrolyte-gas layer interface (mostly at high pH).

The physical mechanism according to which actual electric breakdown is initiated could still be similar to what the paper describes, but I have no way for directly verifying it, besides acknowledging that the formation of a gas layer appears to be necessary, similarly to how bubble formation at the tip of the electrode is considered a basic premise there.

For what it's worth, in plasma electrolysis the voltage from which the process starts has been found to have a direct correlation with electrolyte concentration. Among many other parameters, Gupta and Singh made a study with potassium bisulfate (KHSO4), although the authors—like others—associate the effect with primarily electrolyte conductivity and did not investigate the change all the way to saturation concentration.

E.g. see table 1 in http://dx.doi.org/10.1088/0963-0252/26/1/015005 (paywalled) :

If you plot the mid-point voltage V_D against electrolyte concentration using logarithmic scales, it seems as if it could be extrapolated all the way to voltages similar to the ones observed in this thread (factoring out dendrite formation from electrodeposition processes which would not occur with inert anodes like Pt). KHSO4 has a maximum concentration in water of about 3.6M at 20°C and 8.9M at 100 °C, however. Other studies likely exist but I'm not aware of them.

If you instead decrease electrolyte concentration enough, it should take again kilovolts or anyway the expected voltage required for the breakdown of the gaseous layer formed around the cathode (or more likely water itself, since no significant gas layer would be formed at low concentrations).

I think the electrolytic conditions used in these experiments are "special" in that a similarly large concentration of ions close to the cathode—mainly due to the electrolyte boiling action, as well as the applied current—is not very easily achieved in other environments; that does not necessarily mean though that more exotic phenomena are directly involved if the necessary electric field for electrical breakdown is attained anyway.

Since I wanted to make sure of the behavior at high concentration, I made a few other tests up to 31V with a close-to-saturation KOH solution (35g KOH in 25g water, which would be 140g in 100 ml or almost exactly 25 molar concentration) and a flat copper anode. I used a 0.3 mm tungsten cathode as usual.

The minimum breakdown voltage V_B (voltage beyond which current starts dropping and the reaction transitions from normal electrolysis to plasma electrolysis) observed after adding about 16g of KOH or so was 12–13V, but as the electrolyte concentration was increased, this in turn increased to about 16–17V. Interestingly, though, the so-called mid-point voltage V_D (voltage at which plasma electrolysis takes place at the minimum possible current) remained unaffected, with a minimum at 29–30V over a wide range of KOH concentrations.

In the process, the KOH solution turned deep blue, reminiscent of copper sulfate solution. Unexpectedly, copper deposition on the cathode also occurred, which might have been part of the reason for the increase in the breakdown voltage (possibly due to the decrease in current density). Despite the high electrolyte concentration, I haven't noticed any blocking effect as with potassium carbonate, although I was looking at the reaction from afar and did not video the process.

What compound was formed here in the solution? Copper(I) hydroxide is suggested to be an unstable orange-yellow compound that may turn red with impurities, and Copper(II) hydroxide should have negligible solubility in water. It might be something else, possibly some poorly-defined potassium–copper complex.

The near room-temperature, strongly caustic blue solution showed a slowly falling precipitate which reflected light, although the effect ceased after a while:

EDIT: upon looking for more information, it is plausible that the blue compound formed could be potassium cuprate K2CuO4 (or possibly KCuO2), which should be similar in some ways to the dark purple potassium ferrate solution that was previously formed under similar high-concentration conditions using a carbon steel anode.

Quote from Stevenson

BTW, a couple of days ago I made a similar experiment, and I used a webcam with the Cosmic Ray Finder software in order to detect something anomalous. It recorded 3 events while the current was limited to few tens of mA, than when I increased it to 1 A for about 1 second, the webcam stopped working completely (it was permanently damaged)...

I haven't used it yet with the HCl electrolyte (the experiments with HCl generate annoying gases, eject large amounts of metal dust and tend to heavily corrode metals inside and outside the vessel used—at least a semi-sealed is required, so I have avoided putting too much time on them yet), but today I tried putting my webcam very close to the wall of the operating cell using high-concentration K2CO3.

At first it started malfunctioning and I thought I damaged it, but it appears that the slight EMI generated affects its normal operation depending on its position close tothe electrodes. The EMI was significantly stronger with HCl, so it's possible that it could have become actually damaged with it. I would start getting strange images like these:

A related issue I noticed by using the camera at a close distance is that heat/temperature tends to increase the camera's internal noise, and after a while a large number of low-contrast dots start filling the image, with events getting detected at a rate in the tens of thousands/day (the program cannot keep up with this rate and the dead time increases considerably).

Most of the dots are fixed in place, however. These are two different images made by compositing on top of each about 150 "detection events". Note that I'm referring to the low-contrast dots only, not the darker, likely real traces.

I think it's just heat-induced noise in the CMOS sensor, but I haven't investigated this very much in depth. Increasing the camera's temperature using a hair drier did not cause quite the same results, but still caused general noise to increase and false detection events.

Below are graphs made by processing/analyzing the files generated by the program. Before moving the webcam close to the jar I did not notice any correlation with experimental conditions.

Stevenson

I might have already written this elsewhere, but a while back I tried putting two canisters of potassium hydroxide/carbonate close to the webcam, and the average readings (detection events) increased by about 50%, similarly to what the Geiger counter I previously had did. Indeed, at least this no-brand webcam appears to be somewhat sensitive to ionizing radiation, and I think the signal (event detection rate) I'm getting on average could be largely due to background radiation.

Depending on the design, sensitive and poorly-shielded equipment may be destroyed by strong EMI, although it's possible that part of the problem could have been from a small aluminium foil I placed directly in front of the CMOS sensor (mainly for light-tightness), which might have worked as a sort of antenna.

Currently I'm unsure of where to put the CMOS sensor exactly. I haven't noticed any real correlation with experimental changes so far with the Cosmic Ray Finder application using distances in the order of 55–70 cm. I did notice earlier on that noise appeared to increase with time but I didn't realize right away it was due to temperature. The webcam typically heats up somewhat during operation, but remains just warm. The wall of the glass jar during operation was hot to the touch.

In past related tests at higher voltages (500–800V), dilute electrolyte solutions (0.05M or less) and a considerably higher cathode wear rate, which I documented in the previous 'Mondaini' thread, I used to routinely get rather strong EMI that at times even affected my phone line (=> internet connection). With 10% HCl electrolyte I got EMI close to that obtained in those tests, but not quite as strong. I still think this is the most interesting reaction product and it seems related to cathode material vaporization (using HCl solution, from the deposition layer getting continuously destroyed by the plasma reaction, at least at voltages greater than 45–50V. The bulk of the cathode did not seem to be affected in the short term).