Fusione fredda Renzo Mondaini—trascrizione

Today I received some 0.4mm (0.26ga) Kanthal A1 wire for further testing. Out of the box, the first impression was that it's quite stiff compared to 28ga Ni200 wire, but it can still be bent with relative ease (+ some effort).

With cathodic plasma electrolysis operation using just tap water, which was previously shown with Ni wire to cause oxidation, it quickly develops a dark layer, likely from Al2O3. As it does, initially a gray-brown substance gets dissolved in the aqueous solution, possibly some aluminium compound. Applied voltage was 800V, but under operation and locally hot water, it decreased in the 240-300V range (due to the power supply limiting current).

After the initial induction period where the tip becomes black, it seems more difficult than with the 0.3mm Ni wire to start a reaction, but also to melt it quickly when it begins. It does not seem to become incandescent easily either. This change seems proportionally larger than the difference in wire gauge. By bringing the wire closer to the top of the water surface the reaction does become stronger and sometimes white flashes also appear (my idea is that this is due to the high-melting point surface Al2O3 oxide layer).

A possible factor decreasing performance could be the wire stiffness coupled with high electrical resistivity, which could make electrical conduction difficult under my current configuration. After wrapping the wire more tightly around the tungsten rod I used so far (which I also cleaned up a bit with sandpaper), it seemed to improve a bit, but still did not become incandescent as well as with the 0.3 mm Ni wire.

The idea behind getting Kanthal wire was that it could have helped somewhat in initiating a plasma reaction due to the spontaneously formed surface oxide layer (potentially inducing a sort of dielectric barrier discharge-kind of operation), but it's possible that it's having the opposite effect; alternatively it could be that too much energy is being wasted into joule heating as this is after all a material intended for resistors. A second motivation was that Al2O3 seemed to show particularly bright results in different tests.

The area surrounding the cathode tip still seems quite hot after operation on the other hand, although this is just a subjective impression. So, apparent wire temperature might not necessarily be a good indicator of how well the reaction is progressing if not perhaps for theoretical reasons. It certainly does not look as good, though.

In 1-2 weeks I should receive some 28ga titanium wire, also chosen for similar reasons (spontaneously formed, passivating surface oxide layer). Additionally, the melting point is somewhat higher than most common materials. Titanium dioxide however has a lower melting point than aluminium oxide.

I managed to make the kanthal wire work to some extent, but I think the conditions making it work (some %wt. acetone in the aqueous solution) prevent a passivating oxide layer from forming, and as a result a large amount of dark residue forms. The plasma formed does not look that hot either (no bright flashes occurring) Furthermore, cathode wear is high. Even so, the wire could only be immersed by a few millimeters before the plasma reaction stops. These residues appeared to be weakly ferromagnetic, so it's likely mostly iron.

So far (recently, at least) I used relatively large diameter plastic tubes for enclosing the cathode, but it is also possible to use metallic cylinders. Of course, unless openings are made, they will have the disadvantage that inspecting the reaction might be difficult or impossible. I tried using a small steel tube of unknown origin of about 5–6 mm internal diameter and it seemed to help. After usage it appeared oxidized however, and the internal region surrounding the cathodic plasma appeared heavily covered in black residues; the bottom region in contact with the glass jar almost not at all.

Perhaps sputtering as suggested by Alan Smith earlier is really occurring, although I have no particular knowledge on the process. From a quick search it appears that metal ions from the cathode get sputtered on the anode in the process, so it could be a possibility. This is usually accomplished in a vacuum, however.

An interesting idea could then perhaps be using a nickel cathode (which works better anyway) and a surrounding copper tube, which sounds familiar.

EDIT: I tried 28ga Ni200 wire again using just tap water, and it worked much better (as shown in previous posts), but I'm still not sure as for the exact causes. I don't think it's just the difference in wire gauge, as the 1 mm tungsten rod proportionally seems to work better than the 26ga Kanthal A1 wire (although it won't get as incandescent, it can be immersed for about the same or deeper depth despite the larger thickness). This is probably either the oxide layer formed on the Kanthal wire, or the increased resistivity. If so, then the titanium wire I should get soon won't work very well.

EDIT2: with a very diluted (in the order of 1–2% wt) water-ethanol solution performance appears to improve, but it's still somewhat inconsistent, with fast wear compared to using Nickel wire.

I attempted a test using a small copper cylinder with a slit for real-time visual inspection, with an 8 mm internal diameter. The cathode was a 28ga Ni200 wire.

Unfortunately while it helped, it turned out to not work well, until I further completely enclosed it with a plastic tube. It appears that not only local water temperature (necessary for forming an insulating gas/steam jacket around the electrode to initiate the plasma reaction) is important, but also that there is some sort of obstruction in the electrical path between the anode and the cathode. So the previously used plastic tubes had two functions.

This could be mostly due to limitations of my power supply, which limits current and power to roughly 0.2A and 40W respectively. A higher effective electrolyte resistance would decrease current and make the power supply provide higher voltages on average.

After reaction, green and brown residues were left in the electrolyte (just tap water) and the copper tube looked somewhat oxidized.

As mentioned in previous comments, I keep voltage fixed at 750–800V (open circuit voltage; it decreases with load) and regulate the reaction with electrode depth in the aqueous solution. With the plastic tube installed around the slitted copper tube, a larger depth could be used than without.

I find that a slightly acidic solution seems to work better than a slightly alkaline solution. I used 0.025M citric acid, a weak organic acid. At this concentration, solution pH should be 2.4 according to a calculator I found online. Conductivity was clearly increased from a plain tap water solution, but the cathode (0.3 mm Ni200) could be turned incandescent just as well, if perhaps not better.

Deposition from the copper tube mentioned above would also occur. This isn't clear just using tap water since the wire appears to get oxidized, but adding about 2.5 wt.% acetone as done earlier cleans up the surface even when significant temperatures are reached. The photos below show one such wire before and after acetone addition. Ethanol (and possibly similar compounds that I have not tested yet) also works, but acetone does not leave a strong odor in the environment.

As pointed out in other posts, after adding acetone the reaction also seems slightly more intense, although it is still largely dependent on how much one is willing to push the reaction (current density). The right photo above hints that from the bubbly appearance the Ni200 wire must have reached (and probably slightly exceeded) temperatures close to the melting point along its length.

Curbina

I see the arc more as a point-to-point, high-intensity current flow. What I'm doing appears to be glow plasma, with the reaction more or less homogeneously distributed along the immersed electrode surface, at high voltage and relatively low current flow.

Right now I'm mostly occupied with checking out what conditions appear to improve the reaction or making it easier to trigger with my power supply, and haven't been measuring heat at all. Besides, single-point temperature measurements are not very reliable due to thermal stratification, so more complex methods would be needed.

Larger amounts of acetone tend to be kind of hazardous due to high flammability, but even small amounts appear to significantly increase the amount of evolved hydrogen (either directly or indirectly—e.g. from acetone decomposition in the plasma). I'm aware that R.M.Santilli has reported using ethylene glycol (antifreeze), so perhaps an idea along those lines could be adding propylene glycol, a non-toxic antifreeze compound also often used for vaping, etc. However since it would increase the boiling point of water, it might make the reaction more difficult to trigger in my case.

EDIT: glycerol is another possible choice in alternative to propylene glycol, slightly less expensive and non-toxic.

EDIT2: I ordered one liter of each. I plan testing initially something like 5–10 wt%. I typically use about 35–40 ml water, so they should last for many tests.

I mentioned this before, there are a number of references to "carbon-rich liquids" in glow plasma electrolysis in this open access paper: Sen Gupta, S.K. Contact Glow Discharge Electrolysis: A Novel Tool for Manifold Applications. Plasma Chem Plasma Process 37, 897–945 (2017). https://doi.org/10.1007/s11090-017-9804-z

The author refers to their function as H/OH -radical "scavengers". Generally they are ketones like acetone, or monohydric (with a single OH group) alcohols like ethanol or iso-propanol.

Quote

[...] The H₂ generation by CGDE could be further increased through the use of potential H^· scavengers, one H₂ molecule being formed from one H^· radical by interaction with its scavenger. Among the potential H^· scavengers are acetone, methanol, ethanol, iso-propanol, n-propanol, n-butanol, etc.

In another paragraph from the same paper he quotes the observations of another group that used phenol. The mode of operation is more explicitly pointed out, i.e. the compound removes (or scavenges) OH radicals generated from the plasma area and prevents H₂O from forming. Note that CGDE = cathodic glow plasma discharge.

Quote

[...] The rate of reduction was interestingly found enhanced in the presence of phenol etc. They ascribed the observed reduction to the role of H^· radicals produced during anodic CGDE and its enhanced rate in the presence of phenol to the increased availability of H^· since phenol could scavenge OH^· and prevent formation of H₂O by the reaction between H^· and OH^·.

In this other paper (paywalled) from the author of the first one it's suggested that the carbon-bound hydrogen atoms in such organic compounds can react with both the reactive H and OH radicals formed (same mechanism but worded differently). This implies that they are decomposed in the process: Gangal, U., Srivastava, M. & Sen Gupta, S.K. Scavenging Effects of Aliphatic Alcohols and Acetone on H^• Radicals in Anodic Contact Glow Discharge Electrolysis: Determination of the Primary Yield of H^• Radicals. Plasma Chem Plasma Process 30, 299–309 (2010). https://doi.org/10.1007/s11090-010-9216-9

The same author(s) do not report using alcohols like propylene glycol or glycerol, but the proposed mechanism should not exclude them.

I won't be able to measure the exact gas output but I should at least be able to visually observe the extent to which these compounds prevent cathode oxidation when allowing the cathode to become incandescent during plasma electrolysis.

EDIT: as an additional note, Ruggero Santilli (often cited by Curbina) has sometimes mentioned that the main inefficiency source in underwater arcs is from hydrogen and oxygen recombining to water; see for example page 63(79) in : https://thechurchoflife.net/wp…942d08aee8c477fade4da.pdf

Quote

[...] Recall that the primary source of the large glow created by underwater arcs is the recombination of H and O into H2O following its separation. This recombination is the reason for the low efficiency of underwater arcs and consequential lack of industrial development until recently.

By comparison, the PlasmaArcFlow causes the removal of H and O from the arc immediately following their creation, thus preventing their recombination into H2O, with consequential dramatic increase of the efficiency, that is, of the volume of combustible gas produced per Kwh.

The addition of an alcohol or carbonyl compound as mentioned earlier would, incidentally, improve the reaction proposed by Santilli by preventing or mitigating recombination to water.

Curbina

Difficult to tell without more details. The pulsating or oscillating action may be forming voids or gas pockets around the electrodes and plasma discharge could be occurring there if voltage is high enough.

I also suspect that electrode erosion or debris in the electrolyte solution may also potentially be a scavenging source for the reactive H/OH species formed in the plasma. So one has to wonder whether under active conditions the electrolyte solution looks as in the beginning of the experiments. Earlier on I tried adding graphite also for this reason (more in detail, in an attempt to provide a non-metal surface for H atoms to get adsorbed on), but it gets quickly displaced away from the reaction zone.

I made a few tests with a cheap multimeter in series with the electrodes for measuring current during the plasma reaction. Readings are not very accurate, but it appears it's in the order of 60 mA at 400V. I think the general shape of the current-voltage curve is mainly due to HV power supply limitations and not due to the characteristics of actual plasma reaction. For properly checking that out I'd need a better constant-voltage power supply.

Above a certain voltage (when the cathode becomes brightly incandescent) current readings become unstable and the multimeter starts displaying implausibly low or even negative values. Interestingly, with just tap water (with some citric acid residues) this did not happen, but it started again after adding acetone.

This is likely mostly a power supply–multimeter-related artifact during high load conditions or possibly when a large amount of gas surrounds the electrode (which should be occurring with acetone addition), but it could still be regarded as another indicator that the few wt.% acetone indeed has some effect. Of course, this also means that for reliable power measurements better tools are required.

Curbina

Acetone does not dissociate into ions and should not contribute increasing the conductivity of the solution, so I was thinking that perhaps the current instability effect arising after adding it again in small amounts might be due to something else. For instance, if it causes a larger volume of gas to be produced and surround the cathode, it might make current flow choppier and confuse low-cost multimeters.

The larger volume of gas from acetone (or other organic compounds) addition was proposed a few posts above to be due to their "scavenging" effect of the H/OH radicals produced in the plasma region (which would otherwise quickly recombine to water) rather than increased solution conductivity.

Curbina

I probably did not write too clearly what I actually meant. I was mainly referring to the differences between run #2 and #3. In both cases the aqueous solution was the same, but in run #3 2.5 wt.% acetone was added. Run #3 became choppier, while run #2 overall showed stable readings except when voltage was high (dynamically depending on testing conditions; I did not alter voltage directly).

Here is input power (calculated as simply V*I) versus voltage as a standard line graph (incorrect since readings were not taken continuously).

DC current (yellow instrument; it's showing the opposite current sign here. The other instrument is showing voltage÷2) would often drop like this, sometimes more wildly:

Again, I am fully aware that multimeters like these are only meant for stable DC or standard AC, and that one is not supposed to mix together readings from separate instruments updating at a different rate.

Alan Smith

Occasionally I get crackles from nearby loudspeakers, so it is definitely producing intense back EMF at times.

From further testing, the oscillating readings appear to be loosely related to the degree to which electrolyte solution surrounding the electrode is boiling, or electrolyte temperature: readings are more stable under "only" warm conditions and it takes a period of operation before this happens. Increased gas evolution and temperatures from acetone addition would possibly have a magnifying effect.

While it would be very interesting if I truly got significant backcurrents decreasing the average power applied into the plasma reaction, I realize that the multimeter used is inadequate for measuring this reliably.

Before hitting "reply" I tried another test by adding more citric acid in slight but arbitrary amounts and the current instability effect got stronger. This did not increase conductivity so much that turning the cathode incandescent under water became difficult (like what usually occurs with KOH—potassium hydroxide), but gas production likely increased. It also started producing a more noticeable sugary smell from citric acid decomposition.

Unfortunately the reaction became less manageable and the Ni wire appeared to melt more easily as in this brief test where I also put the same multimeter in current reading mode in the background.

EDIT: after adding some more citric acid I tried with both 0.4 mm kanthal wire and the 1.0 mm tungsten rod but it seems more difficult to obtain the same reaction with them even if the load on the HV power supply isn't any lower. It looks like the cathode still has to become brightly incandescent after all, and with thicker wires this is more difficult within the limitations of my power supply. A too large amount of citric acid also starts working against the effect with the Ni wire.

I removed the slitted copper tube and only used the plastic one to try narrowing down the source of the above effect, again using 28ga Ni200 wire.

1. Just 37.4 ml tap water: after allowing water to locally reach high enough temperature and start the reaction, it can be only occasionally reproduced.
2. Added 0.22g citric acid (0.031M concentration): significantly increased current draw, but the effect cannot be reproduced anymore; readings are stable even with the wire close to melting.
3. Added 1.12g acetone (3 wt.% relatively to water): now the unstable current effect is noticeably visible again and the wire can be more easily turned incandescent.

From the above (with the caveat that the very limited equipment used does not allow proper verification), it would appear that acetone has a synergistic effect of some sort.

EDIT: to clarify, in the previous post I reported that the effect seemed larger after adding more citric acid, but there was already acetone in the solution. So I started over to observe the effect of just citric acid, but it did not seem to work on its own.

Acetone addition might seem completely arbitrary, but regarding excess hydrogen evolution, its effect has been described for anodic glow discharge electrolysis—which also works but requires more voltage and energy on average than the cathodic version—in this paywalled paper:

Scavenging Effects of Aliphatic Alcohols and Acetone on H^• Radicals in Anodic Contact Glow Discharge Electrolysis: Determination of the Primary Yield of H^• Radicals

Plasma Chem Plasma Process 30, 299–309 (2010). https://doi.org/10.1007/s11090-010-9216-9

Figure 3 and 4 summarize the results also for the alcohols tested. Excess hydrogen evolution seems to follow a general trend where most of the effect is observed at about 0.5M concentration (in a 0.05M K2SO4 solution and 450V). I generally used 2–3 wt% acetone, which should be around the upper end of this range.

At this point one might wonder what would be the effect of polyalcohols like glycerol or propylene glycol (these are also popular vaping compounds, so they should be cheap and relatively safe to use). Unfortunately it appears that in this paper by Saksono et al. (an indonesian group that made many experiments on glow discharge electrolysis and uploaded most of their papers on ResearchGate) increasing concentrations up to 3% of glycerol might actually have a negative effect, with the no-glycerol solution producing more hydrogen with less energy:

Hydrogen Production System Using Non-Thermal Plasma Electrolysis in Glycerol-KOH Solution

International Journal of Technology.1. 8-15 (2012). 10.14716/ijtech.v3i1.1091.

This makes me wonder about the effect of propylene glycol, an antifreeze compound similar to ethylene glycol. Ruggero Santilli used the latter advantageously in an underwater carbon arc plasma reaction to produce larger amounts of his MagneGas compound. If it has a neutral or negative effect similar to glycerol, then hydrogen evolution might not be what one needs to watch for, in order to produce energetically useful novel hydrogen–carbon clusters.

While looking for information from other groups that might have used a glycerol–water mixture in the same reaction type, I came across this very recently published (September 2020) open access paper mentioning it. Unfortunately they do not report the results as the focus was on the solution conductivity–breakdown voltage relationship (or at least, I cannot find them here):

Contact Glow Discharge Electrolysis: Effect of Electrolyte Conductivity on Discharge Voltage

Catalysts 2020, 10(10), 1104; https://doi.org/10.3390/catal10101104

Quote

[...] With respect to pure aqueous media, the addition of an organic co-solvent leads to a higher faradaic efficiency for the H₂ evolution preceding the glowing discharge. Among different (poly)alcohols, we decided to employ glycerol due to the presence of three alcoholic groups per molecule and the high miscibility in water (this avoids the formation of different phases in the same system). Glycerol/water mixture was considered a meaningful case-study since it mimics the behavior of organic-contaminated wastewater [46,47,48].

Quote

[...] Glycerol was chosen as an organic compound because of its high miscibility with water, its low cost, its ease of used and handling. Additionally, as mentioned before [5,41], the production of H₂ exploiting the non-faradaic effects of CGDE, in H₂O and alcohol solutions, was a promising energetic alternative and glycerol, having three OH alcoholic groups, was of great interest for this.

The paper provides in any case a good general overview of contact glow discharge electrolysis (CGDE).

I got a bottle glycerol and one of propylene glycol and after some testing there was no too unusual effect that I could observe with their addition up to 5%.

• Glycerol is a rather viscous liquid at room temperature; propylene glycol significantly less so but more than water.
• On their own (just tap water and either of these compounds), neither clearly appears to cause a cleaning effect on the cathode like acetone does, nor current instability effects.
• After a period of operation, glycerol makes the aqueous solution turn yellow, which was pointed out by Saksono et al in a previously linked paper, while propylene glycol does not behave the same and the solution remains more or less clear longer.
• I haven't tested much propylene glycol, but the most noticeable effect of glycerol is that at least in low percentages (2.5 wt%) it makes the Ni200 cathode heat up faster and more care needs to be applied in order not to melt it repeatedly. Operation seems sort of unstable; it could be due to the increased viscosity of the solution.
• Larger amounts (>5 wt%) seem counterproductive, but more testing is needed to make sure.

An unrelated but interesting effect I observed is that 38.5ml water + 3.85 wt% acetone + 2 ml of previously prepared 0.1M KOH solution (this should be equivalent to about 0.005M KOH in water) appeared to produce at times rather intense EM noise that could be heard on nearby loudspeakers. This noise was associated with the previously reported current instability effect, although the correlation was not very strong. Unfortunately I could not reproduce this to the same extent this in a second attempt (EM noise still appears but not as strongly). Also just KOH on its own in similar concentrations produces current instability, it turns out.

I eventually also tried measuring current with my 4000-counts clamp meter and it reports several amperes (5–15A) both in DC and AC mode while the plasma reaction is occurring and, importantly, the Ni cathode is incandescent. Current draw seems accurate with normal electrolysis, on the other hand. So apparently it's not just the multimeter that is displaying strange values. Of course, it's to be expected with a widely and randomly changing current and inadequate measuring instrumentation.

EDIT: below in the spoiler tag are notes from the tests made today.

I got to try a short section of 0.25mm tungsten wire from a small broken 12V halogen lamp as suggested earlier on in this thread by Peter (although he probably meant a high-power lamp). It turned out to be relatively brittle when I tried to extend the coil into a straighter wire.

Unfortunately I couldn't obtain large enough amounts for many tests (mainly because the wire broke in several small sections). Once it started getting hot it produced rather strong white light but it also combusted quite rapidly. Given the costs, I'm unsure if it's a suitable material for this type of semi-uncontrolled experiment where I deliberately raise temperatures to high levels.

I could only manage to make a brief video of the intense light strobing effect produced by the last few millimeters of wire. Due to the camera's autoexposure setting, it does not look as bright as it did in person.

Hopefully 28ga titanium wire (0.321 mm) which I should receive next week will work more reliably. It should form a passivating oxide layer, but also have a significantly lower electrical conductivity compared to the 28ga Ni200 wire I have been using so far. If that won't work properly either, then I will be out of other candidate pure materials to test.