How to properly treat potassium-iron oxide catalysts

  • Here is another test with a different powered-catalyst pellet having roughly half the resistance.

    Similar overall results were observed after adding acetone (which caused a rapid decrease in resistance possibly due to hydrogen—although I cannot rule out other causes), but this time the jar vessel, which I had to replace due to breakage, was apparently better sealed and eventually I had to open the lid to let air in. This caused resistance to increase much faster. The final resistance matched the starting value of 2.85 Ohm. The minimum value at 13V was 1.97 Ohm.

  • Not impossible, but you also mention carbon deposition - this could also conduct current so that effectively you have 2 resistances in parallel for part of the wire at least.

    A possibility I was thinking about is that the ceramic pellet might be simply becoming more conductive due to partial reduction to metal. The presence of carbon on the surface would then be coincidental, as possibly under these conditions it would consume ambient oxygen before the same oxygen could re-oxidize the ceramic material. A related consideration is that porous materials made from consecutive oxidation cycles (e.g. Celani wires) may then decrease their resistance not due to hydrogen absorption directly, but due to the layers turning metallic.

    But—before I get asked this question—why should this effect be of any interest? The reason is that Rydberg matter is supposed to be a conductive material, and some may be formed inside the catalyst material as hydrogen gets absorbed. Alternatively, the increased conductivity could be due to ultra-dense hydrogen formation, but it's unlikely that the superconductive states would be present at the temperatures involved here (these states disappear above a certain temperature). So, if any of these could be easily formed and measured with the crude methods used here, it would be very useful and indicate that simpler experiments for their detection may be possible.

    Sveinn Ólafsson has done conductivity experiments from Rydberg matter in a high-vacuum cell and a different setup, but they have not been reported in detail in a full paper yet.

  • After testing with a small magnet pile some fragments of the brittle black/dark portions composing the K-Fe oxide pellets after acetone or ethanol exposure, it turns out that are strongly ferromagnetic. So, the color, and probably the apparent conductivity, might be not (just) because of carbon, but mainly reduction to iron and/or magnetite instead. Or, something else is going on.

    This shouldn't anyway be entirely unexpected as a similar behavior was apparently noted for actual catalysts in this 45 years-old paper which I linked at the beginning of the thread:

    • Official Post

    So, the color, and probably the apparent conductivity, might be not (just) because of carbon, but mainly reduction to iron and/or magnetite instead. Or, something else is going on.

    The conversion of heamatite, Fe3O4 to magnetite Fe2O3 is not difficult and is also readily reversible. One of my own experiments produces large amounts of black nano-magnetite ( one of the purposes of the experiment)_which converts to heamatite in moments when exposed to the air.

  • Alan Smith

    That seems similar to what the excerpt above also is implying. As for the increased conductivity at high temperatures, I'm not sure. The same excerpt mentions it is due to "thermally excited electronic energy levels in the catalyst", but in in the literature it is indeed reported that magnetite has a higher conductivity than hematite, at least in rocks.

  • Do you know any source for potassium doped iron oxide catalysts?
    I understand from A Condensed Excited (Rydberg) Matter: Perspective and Applications that they are designed for styrene production, and probably commercially available.

  • milton

    I do not have direct experience, but some people around have got them from this source:…n-catalyst_876047913.html

    I've also been told that Leif Holmlid has been using Shell 105-equivalent catalysts from BASF, but I'm assuming it's not simple for regular people to purchase limited amounts of such catalysts.

    If possible, you would be better off synthesizing your own from iron oxide and potassium carbonate or hydroxide, and optionally other oxides (e.g. CeO2 and/or Cr2O3). The process is not difficult, provided that you can apply sufficient heat in air.

    I also found that it's possible to directly synthesize the metastable active phase of these catalysts (KFeO2) at much lower temperatures from FeOOH and KOH, even in the form of thin films on steel surfaces, but whether they are catalytically efficient, there are no published studies.

  • I do not have direct experience, but some people around have got them from this source:…n-catalyst_876047913.html

    I was able to purchase 0.5 kg from that source some years ago. I had it analyzed by a very good lab:

    K2O 63.2 %

    FeOx 32.8 %

    CaO 2.8 %

    CeO2 1.2 %

    Mg trace

    Here's the manufacturer's product sheet:


    • Official Post

    Found this description of Shell 105 in a publication that used Shell 105:

    "Many catalysts have been described for this reaction. One catalyst – Shell 105, also used in this experiment – dominated the market for many years, and was the first to include potassium as a promoter for the water-gas reaction. This catalyst is typically 84.3 % iron as Fe2O3 , 2.4 % chromium as Cr2O3 , and 13.3 % potassium as K2CO3 . It has good physical properties and good activity, and it gives fair yields."

    Sourced from here:…/tc30_styrene_english.pdf

  • magicsound, Curbina

    My understanding is that within reasonable margins the detailed composition of these iron oxide catalysts is of relatively limited importance because under active conditions, due to diffusion of potassium to the outside from heat and partial reduction from hydrocarbons (or hydrogen), the surface (from which Rydberg clusters are proposed to be formed) becomes eventually covered with K and Fe oxides in stoichiometric amounts. The additives as other oxides mostly serve to regulate the rate of loss of potassium from the bulk, improve properties like mechanical or reduction resistance and tune selectivity to specific compounds.


    Abstract: The combination of an XPS/UPS surface analysis instrument with a microreactor allowed the investigation of the surface composition of catalysts characterized by varying activities and selectivities. The active surface is a potassium iron oxide with a 1 : 1 atomic ratio of K : Fe, whereby iron is only in its trivalent state. Conversion of oxidic oxygen to OH groups is detrimental to the activity. No significant amount of promotor additives is present in the active surface. [...]

    The pure KFeO2 phase synthesized from stoichiometric K2CO3 and Fe2O3 not only was found to be about as active as the real catalysts, but also on its own to be already able to form K Rydberg states and matter (in a vacuum at elevated temperatures).


    Abstract: The industrial catalyst for high temperature dehydrogenation of ethylbenzene based on iron and potassium oxides undergoes, under reaction conditions, essentially a transformation into magnetite, Fe3O4, and a mixture of ternary oxides containing trivalent iron, viz. K2Fe22O34 and KFeO2. The latter compound constitutes the outside of the catalyst particles and is indeed the catalytically active phase.…of_KFeO2_and_KAlO2_phases


    Abstract: Well‐characterized catalyst model compounds of KAlO2 and KFeO2 are investigated by thermal desorption of potassium from the material. [...] results agree with the data obtained earlier for industrial catalysts for ammonia and styrene production. They are interpreted in terms of the Schottky cycle, which is completed for KAlO2 and fails for KFeO2. This failure indicates a non‐equilibrium desorption process. K Rydberg states are only found to desorb from KFeO2, in agreement with the suggestion that such states in some way are responsible for the catalytic activity.

    Rydberg states from the active catalyst are also emitted by heating at 1 bar in air at operating temperature, but clustering to Rydberg matter is likely to occur mostly in a vacuum.


    Abstract: The direct spectroscopic observation of K* Rydberg states with principal quantum number n = 5 and 6 by anti-Stokes stimulated Raman spectroscopy at a K-promoted iron oxide surface (commercial catalyst for styrene production) proves that such states are formed thermally at surfaces of alkali-promoted heterogeneous catalysts. The K* states can be detected at 1 bar air pressure downward and at normal catalyst operating temperature in a vacuum. They exist in the boundary layer at the surface. Previous reports of the detection of K* Rydberg states from such catalysts using field ionization and laser ionization in a vacuum are thus confirmed. The implications for the reactivity of alkali-promoted catalysts are discussed.

  • It looks like they're nominally composed of 10% K2O, 10% additives and Fe2O3 balance, but possibly not just the composition, also the preparation method will have an effect on their final performance to some extent. The photos in the brochure suggest anyway that they're still extruded pellets calcined at high temperatures and made industrially in bulk amounts.

    While better properties often translate to lower operating costs for the chemical processes where they are used by the tons, I don't think they would necessarily imply better success for ultra-dense hydrogen experiments where just milligrams are typically used.

  • By the way, recently I have had the occasion of testing some of the K-Fe oxide material at high temperature under flowing H2 inside a pencil-sized furnace open to the atmosphere. The flow was such that the H2 combusted near the opening, also catalytically at the iron oxide material put there.

    Under these conditions I haven't been able to observe heat that was not clearly associated with H2-O2 recombination, unfortunately, and the material located more inside the tube did not seem to show increased temperatures from the hydrogen flow (not to any visible extent at least). The tests however were as usual relatively crude—I was just looking for large-scale changes—and the H2 gas from electrolysis was not dried. Moreover, it's unclear whether cluster formation takes places at pressures close to atmospheric under (still) relatively low hydrogen density conditions.

    Below is an example of the observed changes without (left) and with (right) hydrogen flow:

    At higher temperatures the emission of potassium from the material was such that apparently it started producing a faint potassium flame (visible on the left photo below), which was markedly different from that of H2 combustion more often visible at lower temperature and/or by increasing H2 flow (right):

    The heating wire was made of Kanthal A1 and the tube of SS304 alloy. They didn't short-circuit due to the thick oxide layer intentionally formed on their surface.

    EDIT: earlier I also tried to plug the opening with the same catalyst material, but it turned out to almost completely block hydrogen flow, and no additional heating was seen at all from H2 admission.

    The Kanthal wire looked green at low temperature due to the formation on its surface of KFeO2 due to KOH solution application and heat.

    EDIT: here is also a short animation of the previously mentioned apparent potassium flame following hydrogen flow and higher temperatures. I think the volatilized potassium reacted with oxygen more readily than hydrogen would and prevented a "proper" hydrogen flame to form. I could be wrong, though.

  • Thanks - Did you run Holmlid-inspired experiments back then? If so, were there any positive/negative results worth mentioning?

    No, not Holmlid-like testing. I crushed a catalyst pellet and added 9% (wt.) to the Ni+LAH fuel in Glowstick 5.1. That produced a null result (no excess heat or radiation). The data files for that run were not posted but are available if needed.

  • Just curious, what kind of detectors are you using? Im still surprised researchers are trying to detect nuclear reactions using temperature sensors....

    SindreZG Was that directed to me? I have a 50mm NaI 4k channel gamma spectrometer: Spectrum Techniques UCS30. Also an X123 system with CdTe detector, covering roughly 5-100 kev. For neutrons, I built a He3 corona tube detector, from the open-source design of Bob Higgins. I have a 60 mm pancake G-M counter with pulse output for the DAQ system. For the Glowstick experiments I also used a pair of Bubble Technology tubes for thermal neutrons, but those are now well past their short useful life. I also built a Li6I neutron scintillator feeding a NIM rack, not currently working but possibly fixable. And a box of self-developing dental x-ray films.

  • Given past history, I don't expect a reply back for the time being, but for the sake of clarity:


    The very large heat of condensation of hydrogen to the ultra-dense form would not be nuclear, but it would still be considerably larger than ordinary chemical processes. Since I finally had the opportunity to have a very simple and safe system where flowing hydrogen could be continuously admitted through and over the catalytic material at high temperature, that seemed an obvious first test, although I realize there may be difficulties due to the inherent magnetic field and other non-optimal factors like excessive total pressure.



    [...] It is possible to have an energy output by forming H(0) from hydrogen gas. This condensation energy will easily be believed to be non-chemical thus nuclear due to its size (of the order of hundred times larger than normal chemical energy output). It may be a large part of the energy which is considered to be caused by so-called cold fusion, as suggested previously by Winterberg [6,7]. Other nuclear reactions in H(0) may be the main processes considered to be cold fusion, with very little of normal fusion products like 4He and neutrons out.

    That's also the basis for how Randell Mills has been suggesting to harness energy from his Hydrino, even if Holmlid (in the same paper) does not think that the Hydrino concept has been convincingly demonstrated so far.



    [...] Other forms of hydrogen H have been proposed to exist but have not been convincingly observed or deeply studied. The most discussed case may be the hydrinos proposed by R. Mills [4] with very little experimental evidence. The proposed hydrinos have no resemblance to H(0). Further, based on quantum mechanical calculations a form of picometer-sized hydrogen molecule was proposed by Mayer and Reitz [5] to exist at high pressure. These proposed molecules are similar to H(0) in some respects, and may well exist, at least transiently.

    Since again from the same paper it is also suggested that ordinary catalytic reactors may be emitting muons (which sounded promising for the very basic system shown above and more professionally crafted catalytic microreactors), and that muon detection could be a method for detecting H(0) I am now trying simple detection methods in this regard (e.g. CMOS/CCD sensor detection), which may not be selective to the radiation emitted by the material but could potentially give more information in other ways.



    [...] It is expected that catalysts used in many existing large-scale industrial processes form H(0), but that these species has remained unobserved in these processes due to lack of suitable methods for its detection. One possibility could be to measure the muon emission [9] from the catalytic reactors, but it is not known to us if such experiments have been attempted. There are also several problems with such an approach: 1) most muon detectors are not selective and any energetic nuclear particle will give a similar detector response; however muon capture may give characteristic X-ray spectra from negative muons in suitable materials [89], 2) the negative muons will give nuclear reactions in many materials and to some extent even in hydrogen gas, thus few muons will be able to leave the catalytic reactor before they decay with a time constant of 2.20 μs [16].

    Putting a few thousands euro into a suitable photomultiplier tube-based fast custom detection system is not economically feasible for me at the moment, even if it was shown to work for muon detection from H(0) by Holmlid and Ólafsson in a few published papers.