Perreault-Bruce: Science Contrarian review (Step 2)

Reviewer role: judge by experimental evidence, build the strongest honest case, examine how the claims were treated, and apply the same rigour to the inventor. Sources are tagged [read] (project documents or Investigator notes), [web] (URL), [memory] (confidence), [calc] (my arithmetic, .venv python).

Source gap note: none for the text of the two documents (patent US 7,800,286 and the 2-page "Atomic Decay" leaflet were read in full by Step 1b). But neither contains a single measurement [read: deep reads]. Every experimental statement below about this inventor therefore rests on his own later documents, reached only through Investigator snippets and Wayback notes (not downloaded). The Ge cross-section paper cited by the patent (Abbood and Ebrahiem 2020) could not be read: my fetch returned only page metadata [web: https://jih.uobaghdad.edu.iq/index.php/j/article/view/2374/1571]. The cross-section scale used below is therefore an assumption, flagged where used.

Headline evidence situation

• Tests on record of any central claim: none. No measured current, power, voltage, activity or input/output balance appears in the patent, the leaflet, the 2014 report or the 2018 plasma-valve paper [read]. The Investigator found community builds of the later "plasma valve" (Energetic Forum 2017, groups.io 2020) with no measured output seen [read: research notes]. A 1999 tape page says "tests and experiments confirm that radiant energy devices operate on a self-sustaining method", with no data; this is the inventor's/publisher's own statement, which I record but cannot weigh as a test [read: teslatech note].
• So the correct label for most claims is "never properly tested", not "tested and failed". Nobody rejected the work with an experiment either. It was granted a US patent (examiner Ashok Patel, 2010), which does not test anything, and lapsed in 2014 for an unpaid maintenance fee [read; web: Google Patents]. No suppression story is documented; the record shows neglect and a self-publishing inventor. Google's removal notices on one query (Lumen) are unexplained and noted as an anomaly [read: google note], not as evidence of anything.
• Importantly, the claims partly fail on arithmetic that needs no experiment (claims 1, 3, 6). I say so below with the numbers.

Key physics numbers I computed (used below)

All [calc] with standard atomic masses [memory, high; Ge-72 71.9220758 u, He-4 4.0026033 u, Se-75 74.9225234 u, n 1.0086649 u, Se-76 75.9192136 u]:

• Ge-72 + alpha -> Se-75 + n: Q = -6.063 MeV. The reaction is endothermic: it absorbs 6.06 MeV. The patent's "6.06 MeV" is this magnitude (the patent never says it is a Q-value, and never shows the derivation) [read: US7800286 p.5 gap list].
• Lab-frame threshold including recoil: 6.063 x (1 + 4.003/71.922) = 6.40 MeV. Compound nucleus Se-76: neutron separation energy 11.15 MeV, alpha binding 5.09 MeV, consistent.
• Coulomb barrier alpha on Ge (nuclear radius 1.2 fm x (4^(1/3)+72^(1/3)) = 6.9 fm): about 13.4 MeV. All the alphas in the patent (6.0-8.8 MeV) are below it.
• Energy in one Ci (3.7e10 decays/s) of Rn-222 with Po-218 and Po-214 daughters in equilibrium (alphas 5.49 + 6.00 + 7.69 = 19.2 MeV per chain) = 0.114 W. So about 8.8 Ci per watt of alpha power, at 100% conversion.
• 1 Ci of Rn-222 is 6.5 micrograms. 1 g of Rn-222 with its two alpha daughters releases 8.3e9 J (1.99e9 cal, 7.9e6 BTU) in total, over days, not at once.

Corrections to Step 1

1. Claim 1 says "above a 6.06 MeV threshold". 6.06 MeV is the magnitude of the Q-value, not the lab threshold (6.40 MeV) [calc]. The patent's own list says Rn-220 (6.29 MeV) "can also be used" [read: US7800286 p.4-5]; at 6.29 MeV it is below the true 6.40 MeV threshold, so by the patent's own physics it cannot drive the reaction. Po-216 (6.78 MeV) is above. Step 1 repeated the patent's framing without noting this.
2. Claim 1 also says the reaction "releases" electrons. The reaction is endothermic and removes energy from the alphas; electrons freed by alpha slowing in any material (delta rays, secondary emission) are ordinary and need no nuclear reaction. This distinction is the heart of the review.
3. Claim 6 wording in claims.md ("not internally consistent on isotopes") is correct; I add detail: the prose promises stable Se-77 from Ge-72, but equation 1 gives Se-75, which equation 2 says decays to stable As-75 and stops. The Ge-76 -> Ge-77 -> As-77 -> Se-77 branch starts from a different germanium isotope and a neutron that equation 1 does produce but which would have to find a Ge-76 nucleus (7.8% natural abundance [memory, medium]). Mass and charge balance of each single line is correct [read: deep read]. The Be-9 example is genuine physics (the Be + alpha neutron source, Chadwick/Curie era; Q = +5.70 MeV, exothermic) but it is a neutron source, the Q>0 case, and it does not carry over to Ge [calc: Q for Be-9(alpha,n)C-12 is +5.70 MeV, memory high].
4. Investigator lead on Jaynes is confirmed and sharpened. The printed Se-75 rest energy, 70718.29 MeV, divided by 931.494 is 75.919 u, which is the mass of Se-76, not Se-75 (74.9225 u = 69789.9 MeV) [calc]. With the printed numbers the balance is -934 MeV, not -6.06. The "answer" matches the true Q, so the author evidently knew the target number. Also in the Jaynes note "682 MeV" for the As-77 beta is a unit slip for 0.683 MeV [memory, medium]; his alpha-energy table lists Po-218 at 6.11 MeV above the 6.06 line while the patent excludes it at 6.00. Documents on the inventor's own site that disagree with the patent on the same inputs.
5. Step 1 and the Investigator list "radon gas ... 5,944 BTUs per gram" as a unit step to check. Checked: 1.5e9 cal = 5.95e6 BTU (the printed figure is low by 1000, though the per-pound figure 2.70e9 BTU is consistent with 5.95e6 per gram). The comparison line "nuclear fuel only 200,000,000 BTU per pound" is wrong: complete fission of U-235 gives about 3.5e10 BTU/lb [calc, 8.2e13 J/kg], so fission fuel has roughly 13 times more energy per pound than the radon figure, not 13.5 times less. Only pp.1-2 of that document were read by the Investigator [read: wayback-radioionics note].
6. Step 1 claim 3 describes "thermal positive feedback in the cathode". I add that the patent's mechanism, "the heated cathode increases the odds of alpha particles hitting head on" [read: US7800286 p.6], is contradicted by the physics of decay (see claim 3).

Claim 1: Alpha particles above a 6.06 MeV threshold on Ge, Si or PbS cause an "alpha fusion reaction" releasing high-energy electrons

a. Mathematics consistent? The mass balance gives -6.06 MeV, which is correct arithmetic for (alpha,n), but the sign is negative: the reaction needs energy. Calling it "fusion" that "generates" energy is incorrect labelling of a nuclear reaction with positive threshold; the patent never states a Q-value or energy released. Threshold stated as 6.06 MeV should be 6.40 MeV (recoil). Verdict on maths: partially consistent, mislabelled. b. Consistent with experiment? Alpha-induced (alpha,n) reactions exist and are measured (EXFOR data; the Be-9 neutron source is the textbook case) [memory, high]. Alpha bombardment does eject electrons from any surface (delta rays, secondary emission) [memory, high], and that is the real, measured part of the story. But nothing measured shows a net energy release from Ge(alpha,n). Rate estimate [calc, assumption flagged]: a thick-target yield is roughly n x sigma x (path over which the alpha is above threshold). Using Ge atom density 4.4e22 per cm3, an alpha range of about 25 micrometers [memory, medium] and generously sigma = 1 millibarn over a 2 MeV window, yield is about 1e-8 reactions per alpha. An s-wave Coulomb penetration estimate gives 1e-19 to 1e-22 at 6-9 MeV [calc], so the 1 mb figure is probably generous; the Abbood and Ebrahiem 2020 tables would settle this and I could not read them. Either way each reaction costs 6.06 MeV. At 1e-8 per alpha the fraction of alpha energy touched is below 1e-7. c. Tested? No test by anyone. Not dismissed without testing either; it was never put to test. Jaynes (self-hosted) did a mass balance only. Abbood and Ebrahiem (2020) cite the patent and compute alpha cross sections on Ge from evaluated libraries (EXFOR, ENDF, JEFF) [web: Ibn Al-Haitham J. 33(1)]; calculation, not a device test; I cannot say what they conclude about the patent. d. If true: the Q-value of Ge-72(alpha,n) would have to be positive, contradicting the mass table, or a new alpha-driven energy release channel would be needed. Nothing in mainstream physics would otherwise change.

Verdict (split): - Electron emission from alpha-bombarded Ge/Si/PbS as such: Established (ordinary secondary-electron emission), but it is not fusion and not new. - "Alpha fusion" reaction as net energy source: Refuted by calculation (endothermic by 6.06 MeV, true threshold 6.40 MeV, Rn-220 below it, rate ~1e-8 per alpha at the most). I show the calculation above. This is a calculation-based refutation, not a failed experiment.

Strongest honest case for the inventor: the Ge-72 target and the threshold are real nuclear data and the energies of Po-212/214/216 alphas do exceed 6.40 MeV, so a few neutrons and Se-75 would be made. This is checkable with a gamma spectrum (Se-75 emits 136 and 265 keV gammas [memory, high]). Even a null result would be informative and cheap. The inventor's own words call the equations "theoretical examples" [read: US7800286 p.5].

Claim 2: Sealed radon vessel with negative Ge/Si/PbS wire and Pd or carbon collector gives DC at far higher current per density than earlier nuclear batteries

a. Mathematics: no mathematics. "Amperage per density" is never defined [read]. Internally nothing contradicts; nothing is computed. b. Experiment: Alpha-driven electrostatic generators are real and old: Moseley's radium-electrode cell (1913) and Linder 1952 "use of radioactive material for generation of high voltage" are listed by Google Patents as similar prior art [web: Google Patents note], and the prior art are of the "microwatt" class as the patent says. Real, relevant effects: radon progeny (Po-218 ions) are collected on a negatively charged surface, which is how electrostatic radon monitors work [memory, high]; so the patent's "attracting and concentrating radon onto the wire" is more correctly polonium daughter plate-out, and this effect is real. Energy ceiling [calc]: a curie of Rn-222 in equilibrium with its two alpha daughters carries 0.114 W of alpha power; a milliwatt of electrical output needs about 9 mCi at 100% conversion, a watt needs 8.8 Ci. Gas ionisation can show much larger currents (saturation current about 3 mA per Ci [calc, 34 eV per ion pair, memory high]) but that current flows down the applied field of the 1000 V supply, which supplies the work. A 1000 V corona at 0.1 atm (the patent's pressure) is a gas discharge by itself. So a measured current in this device does not by itself show generation; only output power minus input power, with the radon removed as control, shows that. The patent's "high current" statement is exactly what a biased ion chamber would produce. c. Tested? No data anywhere. The Chinese radon-enrichment alpha battery (CN113539542B, PLA Rocket Force University, granted 2024) cites the patent and uses mine-air radon with a photovoltaic conversion [web: Google Patents]; its power figures were not seen. NASA's NTAC patents cite it too; NTAC literature found by the Investigator and my search presents design and calculation results, not laboratory prototype measurements [web: https://ntrs.nasa.gov/citations/20210024712 and search results]. Citation is not corroboration; both sets cite it as prior art. d. If true: only if conversion efficiency were far beyond known alpha-voltaic limits; no change in physics is needed for a microwatt-scale device.

Verdict (split): - A radon-fuelled alpha-electron cell producing microwatt-to-milliwatt DC, using secondary-electron emission and polonium plate-out: Plausible, untested (no test on record for this geometry). - "Much higher power density than prior art" and any watt-scale output: Never properly tested; and bounded by the 0.114 W/Ci budget, so watts need ~9 Ci of radon (about 60 micrograms) with perfect conversion. Not dismissed; the inventor simply offers no number.

Strongest honest case: the geometry (radon progeny plated on a negative alpha-sensitive wire, 2-pi emission into the metal) is sensible, and an electron multiplication stage could in principle raise the current per alpha. The key unknowns (current per alpha, voltage developed) are one afternoon of measurement.

Claim 3: Once primed above 1000 V, the device is "self-generating" through thermal positive feedback

a. Mathematics: none supplied. The feedback logic is internally flawed: radioactive decay rate does not depend on temperature or on the cathode, so heating the cathode cannot raise the "alpha hit rate" [memory, high]. Heating raises thermionic emission only if the heater power is paid by something, and a gas at fixed pressure becomes less dense next to a hot wire. The patent itself says an electronic voltage source should preferably be kept on [read: p.6], which is the opposite of self-generation. Fig. 2 shows loops from the capacitors back to the step-up converter [read]. b. Experiment: no known device converts a fixed-rate radioactive source into more than the source's decay power (0.114 W per Ci here). Thermionic conversion of decay heat is real and limited by Carnot efficiency [memory, high]. c. Tested? Never, by anyone. The 1999 tape page claims "self-sustaining" operation after a negative-electrostatic start, "tested for many years in remote areas" [read: teslatech note], with no data, by the inventor/publisher; I report it and give it the weight of an unmeasured claim. Not dismissed by anyone with a test. d. If true in the strong form (net energy beyond the radon decay power): conservation of energy would fail. If the claim is only that it keeps running on the radon's own decay power after priming, nothing changes.

Verdict (split): - Output beyond the source's decay power: Refuted by calculation (energy conservation, with the 0.114 W/Ci budget). - Running unassisted on radon decay power after priming: Never properly tested. As written, the patent does not decide between these, which makes it close to Not testable as stated until an input/output power log is shown.

Claim 4: Mercury-vapour variant converts HV/RF/AC to DC even without radioactive material

a. Mathematics: none. b. Experiment: mercury-vapour rectifier tubes and diode gas tubes are old, well-measured technology [memory, high]. That a vapour-filled tube conducts one way when the electrodes differ is established. The patent's statement that it is "more efficient than the prior art" is unquantified. c. Tested? The conversion is tested by a century of use. The efficiency claim has never been tested. Perreault's 2018 paper lists related "plasma field discharge" devices (Coler, Hubbard, Britten, Correa, Gray) as inspiration, and gives no measured data [read: wayback-radioionics note]. d. Nothing would change.

Verdict (split): rectification by a vapour tube: Established. Efficiency advantage: Never properly tested. Note that this part of the patent is plain rectification and carries none of the nuclear claims.

Claim 5: Multi-stage chain of indirectly heated radioactive-bearing semiconductive-cathode valves fed from antenna and earth multiplies electron emission "exponentially" and powers lamps like Moray's valve

a. Mathematics: "multiplied exponentially" with no gain per stage, no energy source accounting and no power rating [read: leaflet p.2]. Fig. 3 has four identical stages; "each successive stage the cathode becomes more heated" is asserted, not derived. Without a power source in the drawing it is not a closed energy account. b. Experiment: Moray's 1920s-30s demonstrations are historical claims by Moray and his family (a 4 kW device before witnesses is a family account) [read: teslatech note]; the leaflet's evidence is a 1954 recollection drawing by Carl Betz that says of itself "only correct in a rough detail", and its text on radioactive clay is Betz's own question ("Could it supply needed ions?") [read: leaflet p.1]. I treat these as documented witness-type evidence from interested parties with no instrumented record. Atmospheric energy is a measurable but tiny resource: fair-weather vertical current density is about 2 pA/m2 at roughly 100 V/m [memory, medium], so an antenna with 100 m2 effective collection area would give nanowatts to microwatts, not a bank of lamps [calc]. The inventor's own 2014 report concludes that in Moray's device "it was the circuitry, not the valve, which produced the amount of electrical power reported" [read: wayback-radioionics note], which moves the energy source away from the nuclear material in the cathode; in 2018 he says the plasma valve runs on cosmic-ray atmospheric ions and withholds the proprietary electrodes for a 3% royalty. Withholding the key materials stops independent test, and I note that against the inventor. c. Tested? Moray's device was reportedly examined in the 1930s; I cannot source an independent test report here [memory, low]. No replication of the Perreault multi-stage circuit is on record; builders of the Perreault plasma valve (2017, 2020) post no measurements [read]. Never properly tested. d. If true: a net-energy extraction from ambient fields at kilowatt scale would require an energy source far above the measured atmospheric budget, or the circuit would have to be a resonant energy store releasing previously stored charge.

Verdict (split): - Radioactive material in the cathode raising electron emission stage by stage: Never properly tested; the idea of a secondary-emission cascade is ordinary (photomultiplier dynodes) but each stage needs its energy supplied. - Powering lamps from antenna and earth plus a trace of radioactive material: Never properly tested; the atmospheric budget calculation makes watt-scale output improbable, so the claim is Refuted by calculation for kilowatt-scale lamp banks if the source is only ambient field and a trace radioactive material [calc, memory medium for the atmospheric figures]. - Historical Moray accounts: record kept; unverified single-family source with no instrumented data.

Claim 6: The printed nuclear equations support the device (Ge-72 + alpha -> Se-75 + n, chain ending in stable Se-77)

a. Mathematics: per-line charge and mass balance is correct [read: deep read; checked]. As a chain it is inconsistent: the prose says stable Se-77 from Ge-72; equation 1 gives Se-75; Se-75 electron-captures to stable As-75 (equation 2) and stops; the Ge-76 in equation 3 is a different isotope from the one produced. Energies are absent from every line. b. Experiment: individual decays are correct as I recall them (Ge-77 t1/2 11.3 h, As-77 38.8 h, Q-beta 2.70 and 0.68 MeV) [memory, medium]. Se-75 decays by electron capture (t1/2 120 d) [memory, high]. Nothing in them supports electrical output; beta decays of products add tiny extra energy and also radioactivity, contrary to the claim that a fusion reaction gives non-radioactive products (Se-75 and Ge-77 are radioactive) [read: p.4 text vs equations]. c. Tested? Not applicable; no one has reported the reaction occurring in this device. d. Nothing.

Verdict: Internally inconsistent (prose vs equation chain; energies absent), and the radiation-free claim contradicts its own products. Equations are individually valid nuclear bookkeeping.

How the claims were treated

• No negative experiment exists. No rejection record exists. The patent was granted (examiner cited Ralph 2,926,268 and Anton 2,958,798) and lapsed after the first maintenance period [read; web: Google Patents]. The Investigator found only 11 forward citations, mostly from NASA (nine patents) and one Chinese radon-battery patent, all as background art [web: Google Patents]. No foreign family. My assessment: the physics community has ignored it, not refuted it. Silence is not a test.
• Absence of evidence worth analysing: "Alpha Fusion Energy Corporation" (launch announced Dec 2012) never had live content and no registry entry was found; competing explanations are an unregistered trade name, an abandoned launch, or content withdrawn; the NH Secretary of State QuickStart search would decide [read: wayback and opencorporates notes]. The inventor's later work drifts from radon alpha (2007) to low-energy gamma (2014), atmospheric plasma valve (2018) and electrets (2022): the funded nuclear core idea was not pursued publicly with data. The inventor sells or describes "radon emissive material" in the 2002 book [read: teslatech note].

Applying the same rigour to the inventor

• Unit and arithmetic slips in his orbit: BTU line (x1000), the fission-fuel comparison (off by ~170x), Jaynes' Se-75 mass, "682 MeV".
• Opening paragraphs of the 1998 article closely follow statements attributed to Tesla without quotation marks [read: Investigator note, memory medium]; not to be taken as Perreault's measurements.
• "Self-sustaining" and "exponentially" are never defined; claims withhold the cathode material (2018) and ask for royalties; the patent hedges its own equations as "theoretical examples".
• The one claim format (the patent's single claim) recites apparatus and a "radon gas ... greater than 6.06 MeV" feature, not any measured output [read].

Investigator findings kept in the record (pertinent, whatever the verdict)

• 2002 book "Harnessing Cosmic Energy" (not obtained), 1998 article, 1999 tape page: dated priority for the alpha-reaction idea and the inventor's most detailed pre-patent text.
• 2014 report: energy source shifts to the circuit (weakens claims 3 and 5 by the inventor's own account).
• 2018 plasma valve and battery paper, withheld electrodes, list of related inventors.
• Jaynes calculations (self-hosted, arithmetic does not reproduce).
• NASA NTAC and Choi work: a mainstream-institution "electron avalanche" nuclear battery concept citing the patent, whose published results I found to be design and projection, not hardware measurements. CN113539542B, Moses US20220108814A1, Weber US20100062288A1.
• Community builds (2017, 2020) of the plasma valve with no measurements; Inventors Weekend organisation (2001-2005) and Don Smith 2004 e-mails ("tritium battery", high voltage), which locate Perreault in a circle of self-sustaining-device claims.
• Unread, high value: "Radon Gas is the Ultimate Fuel" pp.3-26, "HF-box-lab-tests.pdf", the 2019 full-disclosure editions, "Ion Valves v1.3". These might contain the first measurements; I did not see them.

What would settle it

1. Bench test: the patent's vessel, radon-filled and a control with the radon removed or replaced by a dummy, same 1000 V priming; log input and output power for days. Radon decays with a 3.8-day half-life; a "self-generating" device should fade at that rate, which is itself a test of source.
2. Gamma spectroscopy of a Ge cathode in the radon vessel for Se-75 (136/265 keV) to measure the actual reaction rate.
3. Read Abbood and Ebrahiem (2020) tables and EXFOR for Ge-72(alpha,n) cross sections at 6.4-8.8 MeV.
4. Obtain "Radon Gas is the Ultimate Fuel" and "HF-box-lab-tests.pdf" for any test data.

Verdict summary