How to properly treat potassium-iron oxide catalysts

Quote from Alan Smith

My thought was that the KFe complex would be solid, though I have not checked its water-solubility - and thus can be separated by filtration. If you are concerned about free KOH forming carbonates then most of the unbound material - which would be potassium citrate at this stage-could be removed by washing the solid fraction in distilled water and re-filtering. However, I do suspect that the presence of some small percentage of carbonates in the end product would not be particularly detrimental. Sometimes you have to improvise a bit when exploring the possibilities.

KFeO₂ is a metastable solid compound that only exists under specific conditions. Similarly to KOH, it quickly absorbs moisture and CO₂ from the environment but with the addition that it decomposes in the process.

I don't think it can be filtrated as wetting the compound with water at low temperature immediately decomposes it into a brown mixture of various iron oxides, hydroxides and KOH. Excerpt below from https://doi.org/10.1002/9783527610044.hetcat0163

On the very short term this brown mixture can be reverted to KFeO₂ if water is allowed to evaporate again at elevated temperatures. Actually I have a video where this is in part shown; see the first 30 seconds here (I used a KOH aqueous solution rather than just water):

The carbonates formed with prolonged dry exposure at mild temperature to atmospheric gases tend to have a white (when dry) or initially emerald green look (possibly partially containing Fe in some form although I can't rule out other impurities in the material). They form quicker at lower temperatures; see for instance in the photo below on the cooler edges of the olive-green sample.

From the same source as above:

From what I have observed however, stable carbonate presence is tolerated, if more KFeO₂ can be formed again and if the existing carbonates get dissolved onto a wider area. I have usually accomplished this by adding more KOH solution to the heated surface, temporarily wetting it.

Rob Woudenberg

Thanks for reminding. In some aspects it seems to contain more or less general information on the catalysts, while in others it's more detailed, but except one quick mention in paragraph [0029] it does not discuss in detail the characteristics of what is considered the catalytically-active phase, how it's formed, how it relates to Rydberg K emission according to the published literature, and so on.

What I described in the past few posts is for the most part an alternative, simple method for apparently directly obtaining thin films of such active phase at relatively low temperatures on various steel surfaces (in practice, it's not clear how good the end result is. For instance, it is easy to add excessive amounts of potassium hydroxide and form a potentially undesirable transparent liquid film that stays on top the newly-formed olive-green layer).

After some thought, I believe the process is something along these lines, at least in the ideal scenario:

Below is an gif sped up by 20x of part of the above process (it does not start from a clean Fe surface) from a video I made earlier.

magicsound

For what it's worth, some time ago I tried using the yellow flame of a small lighter, heating a smaller carbon steel blade from the underside, and the process still worked. However, when I tried a blue flame from a gas stove, I couldn't manage to get it working. I haven't done many tests with that however, as I try to avoid making chemical experiments in the kitchen.

As for the temperature, I think the surface can get hotter than the melting point of potassium hydroxide (360 °C), but the green potassium ferrite layer starts getting formed at lower temperatures. At times, especially when excess KOH is used (or added), a transition can be seen where the surface becomes shiny and "wet" despite the quite elevated temperature. I've made a video showing such "wet" surface earlier:

Using excess KOH, it seems that water can get trapped below such liquid layer, and form large bubbles on the surface as it tries to break through it with temperature. Or at least, that's my interpretation.

With lower amounts of KOH on the surface, that does not happen. A few hours ago I started over with cleaned steel blades, and wetting them with a 2.5M KOH solution and applying heat seemed to work just right, although solution molarity could probably be further decreased. The initially formed green layer looked wet, but after a while it became opaque, possibly both from more of the underlying steel surface oxidizing and absorption of environmental gases.

The heating coil is kinked but that it was not intentionally set that way. It used to be straight. The setup was intended to be as in the schematic diagram. It's a rather improvised but relatively safe setup (so far, at least) that I'm running on the desk on my side.

By the way, I'm also running a Geiger counter at about 1 meter distance from the jar, but besides daily variations from presumably Radon progeny, which causes a slowly increasing signal when the windows are closed, no increase was detected so far. I don't think any should have been expected given that it's not continuously running in a hydrogen gas atmosphere, but the catalytic layers formed should be able to also dehydrogenate water to some extent, and the formation process itself should evolve some hydrogen (the [2Fe + 4H₂O → 2FeOOH + 3H₂] step, assuming it's correct).

An interesting (?) although probably expected effect of the procedure is that once the treated surfaces reach a high enough temperature they start "fuming". The fumes form a persistent "fog" that quickly fills the environment and condenses on surfaces leaving short-lived white residues. I think it is some form of water-potassium compound, but I'm not entirely sure. Potassium hydroxide has a boiling point of 1327 °C (wikipedia).

This is easier to observe with wires. The one in the video looks rather hot, but it only looked orange by eye, so probably about 900 °C.

At minute 2:05 here:

(previously posted gif animation also left attached)

I tried a sort of catalyst-ceramic wire-wound resistor hybrid.

I mixed K₂CO₃ and Fe₂O₃ in roughly equal amounts with some water to obtain a thick paste (although perhaps "mud" better describes what it felt like) and applied it inside and outside a newly made, Zn-stripped mild steel coil. I then applied first mild heat to summarily dry the material, then more heat to increase temperatures to orange heat (about 900 °C, although it doesn't look like it in one of the photos). Cooling down the wire revealed a bright green coloration that should indicate the KFeO₂ compound discussed in this thread.

One (probably obvious) advantage of this method is that the heating coil heats up more homogeneously. Without a thermal sink for the wire, some portions always end up heating substantially more than others, causing premature failure.

I tried repeating the process to show it from start to finish. Unfortunately it did not go too well.

• 15g K2CO3 was first dissolved in 10g H2O (not all dissolved) in a plastic container, then 15g Fe2O3 added and the compound stirred
  • Therefore, H2O:K2CO3:Fe2O3 = 2:3:3
  • A thick, highly staining, bloody red cream was obtained
• A coiled resistor wire (steel) was then immersed and covered into such cream (or paste) inside-out and installed in the usual jar
• Heating with only 2 voltage levels (5V and 12V) was difficult due to the relatively high water content. Temperatures have to be increased very slowly or the material will start bubbling from the evaporating water, which it did.
  • As a result, the surface layer obtained was not as nice as it initially looked, and much of the internally applied mixture leaked out.
• In the end I could still obtain an outer green KFeO2 layer quite quickly
  • The surface turns clearer in the process at high temperature (mid to high incandescence)
  • When temperature decreases again, the transient green color becomes observable
• Temperatures in the order of 900 °C at least externally have been achieved
  • Probably higher temperatures were internally reached, which might have contributed to failure later on
• In the end the wire broke, probably because a portion of it got calcined on the underlying glass support from the initially leaked potassium-iron oxide mixture. The wire got stuck to the jar
  • The interior of the ceramic/catalyst resistor was found to be hollow and dark

• I tried restoring the wire function by forcibly (manually) connecting together the broken pieces while applying 12V, which melted a section and restored conduction
  • Heating however wasn't homogeneous
  • A section had a high resistance and was quite hot, probably above 1000 °C. It fumed, possibly from potassium. These fumes left white residues on the pathway to the opening of the the semi-closed jar (see photo)

The ceramic resistor broke again, but the broken pieces could be reattached together and function easily restored, which is an interesting property. The second attempt at this looked better.

It can then be of interest to observe the the decay of the catalytically-active compound KFeO2, which as previously mentioned can be tracked from its color change from green to red. I made a timelapse video of this process:

Possibly due to the dry and hot day, or the amount of potassium deposited on the walls of the jar (which absorbs moisture), or perhaps the high synthesis temperature, the decay was slow compared to previous attempts at forming the material, in particular with KOH: after more than two hours without heating there were still visually detectable amounts of the compound on the surface of the ceramic resistor remains, whereas usually it only takes a couple minutes for the material to turn red and moist. This might possibly also indicate that the material formed was not very catalytically-active, perhaps due to sintering from the high temperatures used during synthesis.

At about 00:22 in the video (1h 29min in real-time) i blew on the jar to see if any faster change could be observed, since the decay was slower than usual. What this accomplished was mostly moisture absorption by the white potassium oxides (or hydroxides) deposited on the jar walls. So it could be just that the decay was slow due to very low local moisture levels.

Not sure if this is still of interest to anybody else, but I made more detailed photos of the process at various stages using another wire and the previously prepared K₂CO₃-Fe₂O₃ mixture.

It seems that after holding the material for a long period at very high temperatures >900 °C there is a further transition which makes the green KFeO₂ material turn darker and become more stable against decomposition in air. This could be a related compound with the formula K₂Fe₂₂O₃₄. After an even more prolonged period the material turns almost black.

Earlier (in a previous attempt), I tried scraping this almost-black layer from a broken catalyst piece, and it seemed to reveal underneath a more familiar olive green color. What this means exactly, it's not clear, but in one of the old papers I linked at the beginning of the thread it was pointed out by the author:

The appearance of this black layer from my testing (mainly in the atmosphere at high temperature) seems related to increased loss of potassium from the catalyst, which forms a slow, thick vapor or "smog" that quickly fills the environment and leaves a short-lived white layer upon condensation on surfaces. I also made a video of that yesterday. The sample shown eventually broke into several pieces and had to be replaced.

SindreZG

A few points to clarify:

1. I haven't really put a lot of money behind these tests.
2. I'm for the most part only exploring (learning about) the processes related to the synthesis of these catalysts, not trying to come up with a commercial product—besides, actual performance testing would need a proper vacuum chamber and system.
3. I have also been exploring different methods for obtaining the catalytically-active materials similar to those reported in the literature, with possible relevance to some of those reported in LENR experiments.
4. It's only recently that I tried putting some time into focused experimentation onto this topic, but I'm not really that much committed to it.

While heating a newly made* catalyst-resistor hybrid from a salvaged (used) coiled wire to fairly high temperatures, I noticed that the interior ceramic body seemed to "outshine" the wire (also coated in the same material). To the right the same sample after cooling down.

I just wanted to point out that in one notable case (fig 12), in completely different experiments by another group, this behavior was suggested to be evidence of an internal [excess] heat source, although later analyses by others suggested this to the result of the emissivity and possibly also the translucency of many ceramic materials like alumina.

The internal temperature here certainly exceeded that of melting point of steel, but apparently conduction can continue for a while through the material even if the internal wire breaks or melts, if the ceramic structure can maintain its integrity. When conduction appears to occur through the catalytic material, current increases with temperature (at least up to a certain level), which is the opposite of what happens with an ordinary resistor.

*This just involves dipping the coiled wire in the previously prepared Fe₂O₃–K₂CO₃ paste, removing the excess and heating it slowly after placing it in the glass support (jar).

Below are just some more photos related to a couple tests I attempted yesterday after crafting another catalyst-ceramic resistor, for personal future reference.

I used 12g K₂CO₃ in 8g H₂O (did not dissolve completely), then added 18g Fe₂O₃. I used soap-making-grade K₂CO₃ (stated to be 99.5% pure, but I think perfumes have been added; it smells nice) and PY42+PR101 red iron oxide pigment, which should be a FeOOH+Fe₂O₃ mixture with probably a few percent of SiO₂ and Al₂O₃ impurities. This produced a very thick red paste that did not seem to mix very well; I should have probably added a bit more water. However, too much water produces a slurry that is difficult to apply. The Fe₂O₃ to K₂CO₃ ratio was 60:40 (as opposed to 50:50 like last time).

The heating core was again composed of an arbitrary-length (about 80 cm) galvanized steel gardening wire that had most of the surface zinc removed in citric acid. I could have used HCl but I did not want to leave Cl residues on the wire. The wire was turned into a roughly 5mm diameter coil before this step.

Under running conditions, the long sample was heated with 11.0V and 5.2A = 57.2W, 2.12 Ohm. These measurements are not very accurate but should be in the right ballpark. I had no control on the heating except for selecting 5/12V (voltage sags under load) or manually PWM'ing the applied power (which was mainly useful in the initial stage where the pellet was in the form of a paste).

Two photo sequences showing the changes from a red paste to a green material (after cooling). Heat was not applied very homogeneously though the sample. The second shorter sample was made out of the first one, when it broke following internal failure (from too high temperatures and/or wire oxidation).

A couple detailed photos of the first sample. Heating causes thermal or oxide expansion and cracking. It seems that this occurs mainly when cooling them down, and a crackling sound can be heard at the same time. Note that it turned green (denoting the catalytically-active potassium-iron oxide compound known as KFeO₂) where temperature was the highest.

After prolonged heating to high temperature eventually most of the sample turns green. This process could be easier in a vacuum, but high temperatures are still required (at least 600 °C homogeneously applied to the sample, after partial reduction) for this to be achieved within reasonable time. Here is the second one that was made.

I allowed it to fail by applying a high power continuously. Some wire sections internally were likely thinning out and kept getting increasingly hotter. A section that heated the most turned black; I'm not sure if from metal vapors or other changes in the mixture composition.

I did not observe visible potassium oxide emission in these tests, which could indicate that last time there was excess potassium.

In another test with a similarly made coil using a different form factor (>10mm diameter; it did not perform well) I used 75:25 Fe₂O₃ to K₂CO₃ mixture and I could still obtain a green layer on the hottest portions. Unfortunately I ended up adding too much water and it did not look nice (photos omitted).

A high-temperature furnace would likely achieve better samples, but on the other hand the method I'm using is relatively quick (except the coil forming part) and cheap, and heats up the material directly. I'm always looking for similarly low-priced and better performing alternatives, though.

In theory, actual usage in industrial reactors at about 650 °C with steam and hydrocarbons should also cause the changes observed, but the green layer formed on the surface (from the accumulation of potassium diffusing out with heating and possibly other processes) is not normally visible due to it being covered in loose carbon residues.

As far as I have been made aware of, heating actual catalyst samples in a rough vacuum at or just below incandescence, even for weeks, does not seem to bring an appreciable shift over time to such green hue, although a darkening (not from reduction or carbon coverage) is easily observed, as also shown in the above photos.

Out of curiosity, I took several detailed photos of a 25% K2CO3 - 75% Fe2O3 resistor while increasing gradually temperatures with an improvised Arduino PWM setup. The sample was actually dipped again in the mixture after calcining it once with it with an uncontrolled temperature ramp.

When it gets close to incandescence it turns quite dark brown.

This change is actually reversible (i.e. it turns red again after cooling, not pictured here). I think this behavior should be consistent with this excerpt from an old paper. The author noted that the change is due to thermally excited states in the material:

https://doi.org/10.1080/01614947408071864

It's probably worth mentioning that it was difficult to obtain the sought-after metastable green compound described earlier on. Some was visible (again on the hottest portions) in the first attempt with an uncontrolled temperature ramp (before dipping it again in the Fe-K oxide mixture), which seemed to heat up more efficiently, but still not very homogeneously. I did not remove the Zn layer in citric acid this time, and this probably caused excess gas evolution, contamination of the material and a bubbly surface appearance. However, the lower potassium percentage very likely had an effect too.

When keeping temperatures within controlled levels the resistor can—not unexpectedly—be heated for quite a while without failure, so in the end I tried playing it and ended up breaking it in several pieces. In the process I could verify that the material at high temperature is indeed "practically metallic conductor" (as per excerpt above) and conduction could be restored by reattaching the broken pieces together almost arbitrarily, heating up to bright incandescence at the junction point in the process (not shown).

I realized that I didn't show yet that the darkening process is reversible, so here is a gif about it. The animation also shows considerable thermal expansion at high temperature. In the background there is an aluminium foil to improve heating efficiency and shield the bottom of the jar (and desk) from the heat, which worked very well. The video has been sped up by a factor of 100.

In some instances the pellet looked almost black, although this isn't very clear from the animation.

In the end some of the green KFeO2 was also formed, but mainly on the bottom portion which was exposed also to reflected infrared heat.

I used an Arduino for PWM'ing the power supply using the green pin on the ATX connector of the computer power supply (not an intended mode of operation). Power was intermittently applied at a rate of 2 Hz at a duty cycle between 10 and 100% at 12V.

Because I used such a low PWM frequency, in the beginning I also noticed something unusual, although probably expected due to the relatively high current (I did not measure it but I think it was in the order of 10A peak). When applying power on/off at a low duty cycle the pellet would slightly rock sideways. Below is a gif in real-time that shows the effect.