RankingsPrimerArbiterPhysicistMathematicianContrarianQESites logClaimsSourcesResearch notes

Perreault-Bruce: Physicist review (Step 2)

Reviewer role: PhD physicist. Documents: US 7,800,286 B2 "Alpha Fusion Electrical Energy Valve" (Perreault, 2010) and a 2-page leaflet "Direct Converson from Atomic Decay to Electrical Power" (SonoMagnetics Alternative Energy Research Trust). Tags: [read] from the project documents/Step 1 deep reads, [web] with URL, [memory] with confidence, [calc] my own calculation (Python/.venv, 2026-10-04).

Source status: there is no source gap in the text, but neither document contains any measurement (no current, power, voltage output, radon activity, dimensions or test) [read: deep/US7800286 section (c)]. Every verdict below is therefore about a design on paper plus the inventor's assertions.

Mainstream view in brief (for the reader)

Nuclear batteries are real. A radioactive source emits charged particles; a collector gathers them or the ionisation they cause. Alpha, beta and "alphavoltaic/betavoltaic" cells deliver nanowatts to milliwatts [memory, high]. The hard ceiling is the decay power of the source: power out cannot exceed (activity) x (energy per decay). A nuclear reaction triggered by an alpha particle can add to this only if the reaction is exothermic (positive Q) and actually happens often. For the patent's headline reaction neither condition holds, as shown under Claim 1.

Useful definitions: Q-value = rest-mass energy of the reactants minus that of the products (positive = releases energy; negative = must be paid for with kinetic energy). Coulomb barrier = electrostatic repulsion between the positively charged alpha (charge 2) and the nucleus (germanium: charge 32) that a slow alpha must tunnel through. Becquerel (Bq) = decays per second; curie (Ci) = 3.7e10 Bq.

Claims (from claims.md)

  1. Alphas above a 6.06 MeV threshold hitting Ge/Si/PbS cause an "alpha fusion reaction" releasing high-energy electrons.
  2. A radon-filled vessel with negative Ge/Si/PbS corona wire and Pd/carbon collector delivers DC power at far higher current density than earlier nuclear batteries.
  3. Once primed at 1000 V or more, the device is "self-generating" through thermal positive feedback.
  4. A mercury/vapour-filled variant converts high-voltage RF/AC to DC even without radioactive material.
  5. A multi-stage chain of indirectly heated radioactive-bearing semiconductive-cathode valves fed from antenna and earth multiplies emission "exponentially" and lights lamps, as Moray's germanium valve did.
  6. The printed nuclear equations (Ge-72 + alpha -> ... -> stable Se-77) support the device.

Claim 1: alpha "fusion" on germanium above a 6.06 MeV threshold

a. Mathematics. Each printed line balances mass number and charge [read: p.4; deep read]. The set does not hang together: eq.1 gives Se-75 (not "stable Se-77" as the prose says), and eq.3 consumes Ge-76, which nothing in the chain makes [read]. The patent gives no Q-values. Calculation [calc]. With standard atomic masses (Ge-72 71.9220758 u, He-4 4.0026033 u, Se-75 74.9225234 u, n 1.0086649 u [memory, high]; 1 u = 931.494 MeV): Q(Ge-72(alpha,n)Se-75) = -6.063 MeV. The reaction is endothermic: it absorbs 6.06 MeV. The patent's "at least 6.06 MeV" is this Q-value; the true laboratory threshold, including recoil of the compound nucleus, is Q x (1 + 4.0026/71.92) = 6.40 MeV. Step 1's note that Jaynes's printed mass table does not reproduce 6.06 MeV (his Se-75 energy corresponds to Se-76) is correct; his headline result matches the real Q, so the arithmetic error is in the table, not the Q-value [read: Investigator note on Jaynes, which I confirmed by recomputing]. Consequences: (i) the claim that "fusion" releases "an enormous reserve of stored atomic energy" is inverted for the first step. The reaction costs energy. (ii) Rn-220, whose alpha is 6.288 MeV [web: chemlin.org/isotope/radon-220, search result 2026-10-04], is below the 6.40 MeV threshold, though the patent lists it as usable (6.29 MeV, above 6.06 only by ignoring recoil). The patent's ordering of Po-216 (6.78), Po-214 (7.69), Po-212 (8.78) as usable and Po-218 (6.00), Po-210 (5.30) as unusable is right as far as it goes; the quoted alpha energies are the standard values [memory, high; Po-212 about 8.8 MeV confirmed in the same search]. Whole chain: if every link occurred, Ge-72 + He-4 + Ge-76 -> As-75 + Se-77 releases +4.26 MeV net [calc]. That is the largest nuclear energy the chain could add per alpha. Compare the alpha's own decay energy, 6-9 MeV, already released whether or not any reaction follows. So even a perfect chain adds under a factor of about 1.5 to the bare decay energy; it cannot be "high power density" by itself. Also, most of the chain's energy is neutron-capture gamma rays and neutrinos/antineutrinos, not "high-energy electrons"; Se-75 is radioactive (half-life about 120 days [memory, medium]) and emits gamma rays, so the "non-radioactive stable by-products" statement [read: p.4] is wrong for the printed first step.

b. Experimental consistency. Alpha-induced reactions on germanium are measured and used (e.g. Ge-72(alpha,n)Se-75 in-beam spectroscopy; production of Se-73 and Se-75 on Ge targets) [web: https://www.osti.gov/etdeweb/biblio/7250415 and https://www.sciencedirect.com/science/article/abs/pii/S0969806X25009570 , titles and abstracts only]. They need accelerator-energy alphas and are endothermic, consistent with the Q above. For the patent's 6-9 MeV alphas the Coulomb barrier is about 11.5-13.4 MeV (radius parameter 1.4-1.2 fm) [calc]. A point-charge tunnelling factor exp(-0.9895 Z1 Z2 sqrt(A_red/E_MeV)) is about 1e-21 at 6.4 MeV and 8e-19 at 8.8 MeV; for Be-9 + alpha at 5 MeV it is about 3e-3 [calc]. This factor is not a cross section, but it shows the germanium reaction is some 15 orders of magnitude harder to drive than the Be-9 neutron source the patent uses as its model. Alphas also lose energy quickly in solid Ge, so only the first microns of path could be above threshold. The rate is therefore negligible [calc, order-of-magnitude; exact cross section at 8-9 MeV I did not obtain, only existence of data]. The Be-9 equations are real physics: Be-9(alpha,n)C-12 has Q = +5.70 MeV (Jaynes's figure; I agree [memory, high]) and is the old Ra-Be/Po-Be neutron source. Calling it "fusion" is acceptable for Be; for Ge it is not an energy source. Silicon and PbS are claimed as emitters too; the same barrier argument holds, and for them the patent offers no reaction at all. The one robust real effect: alpha particles knock out many secondary electrons (delta rays) from any surface, and any material would do this, Ge having no special role. No "fusion" is needed.

c. Tested? No test of the reaction-for-power idea is on record in the documents. The Investigator found no measurements anywhere in the inventor's material read so far (the 26-page radon paper and HF-box-lab-tests.pdf are still unread; see "Open items"). d. If true: one would need alpha-driven (alpha,n) rates on Ge many orders of magnitude above what Coulomb-barrier physics and measured excitation functions give, and an exoergic path where the standard mass table gives an endoergic one. That would overturn nuclear-reaction theory (well tested to high precision).

Verdict (split): The strong form (the Ge reaction releases energy and drives useful power) is Refuted by calculation (Q = -6.06 MeV; barrier). The weak form (alphas on a metal/semiconductor produce secondary electrons) is Established but is conventional and no new source of energy.

Claim 2: radon vessel delivers far higher current density than earlier nuclear batteries

a. Mathematics. The patent states no power, voltage or current; "amperage per density" is undefined [read]. So there is nothing to check internally, except the energy ceiling below. Energy budget [calc, worked example]. Radon-222 plus its three short-lived alpha-emitting daughters (Po-218, Po-214, and Rn-222 itself: alphas 5.49 + 6.00 + 7.69 = 19.2 MeV per chain; Pb-214/Bi-214 betas ignored) at 1 Ci (3.7e10 Bq): power = 3.7e10 x 19.2 MeV x 1.602e-13 J/MeV = 0.11 W total, all forms. Mass of 1 Ci of Rn-222 is 6.5 micrograms [calc]. If every alpha's energy were converted into ion pairs at 35 eV per pair [memory, high], the saturation ion current is about 3.2 mA. A current of 3 mA at the patent's 1000 V would be 3 W, thirty times the decay power. The excess can only come from the external supply that the patent requires ("a negative charge of one-thousand volts or higher is applied") [read: p.5]. This is the same trap as in an ordinary ionisation chamber: current is large, but the power balance is dominated by the applied voltage; the net electrical power the source can deliver is at most a fraction of 0.11 W per curie. For scale, the patent's vessel is said to be at 1/10 atmosphere filled with radon [read: p.6, claim 1(c)]. 100 cm3 of pure radon at 0.1 atm is 0.09 g, which is about 14,000 Ci and about 1.6 kW of decay heat plus lethal radiation [calc]. A real vessel would use a few curies at most (radon has a 3.82 day half-life [memory, high], so it decays away and must be replenished). b. Consistency. Radon-daughter ions plating onto a negatively charged wire is real and used in radon monitors [memory, medium] , so concentrating activity on the corona wire by high voltage [read: p.5] is plausible. Helium accumulation blocking operation is acknowledged by the inventor [read: p.5]. Existing devices produce microwatts [read: patent's own phrase "millionth of a watt"]; a bound of order 0.1 W per curie, with conversion efficiency well under 100%, means any improvement is bounded by source activity, not by "alpha fusion". The Investigator lists Chinese PLA Rocket Force University CN113539542B (a radon-enrichment alpha nuclear battery) and NASA's Nuclear Thermionic Avalanche Cell patents citing this patent [web: Google Patents, per sites.md]. They show others pursue radon/alpha secondary-electron batteries with ordinary physics; I did not read their measured results (NTRS TM-20220019348 unopened), and they are not evidence for this claim's reaction mechanism. They are related work worth keeping in the record. c. Tested? No test on record. Patent lapsed in 2014 for non-payment of the maintenance fee [web: Google Patents, per Investigator]. The inventor's later public work moves to gamma sources (2014), atmospheric-ion "plasma valve" (2018), electrets (2022), which is the inventor himself shifting the energy source away from the patent's mechanism. No lab data seen; HF-box-lab-tests.pdf not yet read. d. If true: a conventional alpha-volt device with a ceiling of about 0.1 W/Ci would be fine; a device beating that ceiling would need new energy.

Verdict (split): A radon-fed ionisation/secondary-electron cell producing tiny power is Plausible, untested (ordinary physics, no measurement on record, bounded by about 0.1 W per curie). Any "high power density" beyond the radiometric ceiling is Refuted by calculation (energy budget).

Claim 3: self-generating after priming, thermal positive feedback

a. Mathematics. Circuit-level: Fig. 2 shows the HV capacitors feeding back into the step-up converter; the text itself says "it is preferable that an electronic voltage source be used to keep the alpha fusion valve in a constant energized state" [read: p.6]. So the preferred mode is externally powered. No net-energy accounting is given. Thermionic check [calc]. Richardson-Dushman J = A T^2 exp(-W/kT), with A = 120 A/cm2/K2, W about 4.8 eV (typical for germanium; [memory, medium]): J = 2e-33 A/cm2 at 600 K and 8e-17 A/cm2 at 1000 K. The cathode heating available from 0.1 W/Ci of decay power cannot reach emission levels; the "positive feedback" is negligible. Germanium melts at about 940 C [memory, high] and becomes conducting thermally at modest temperature, but the loop gain needed (more reactions per degree) fails because alpha-hit rate does not depend on temperature. b. Self-sustaining output greater than input from a decaying source is excluded by energy conservation: output power <= decay power + any ambient energy actually collected. c. Untested; the 1999 tape page claim of "tests and experiments ... confirm that radiant energy devices operate on a self-sustaining method" is an interested-party statement with no data on the page [web: teslatech.info, per Investigator]. I keep it in the record; it is not a measurement. d. Net energy gain beyond the radioactive and ambient sources would require violating energy conservation.

Verdict: "Self-generating" in the sense of net energy from a decaying source above its decay power: Refuted by calculation. Weak form (a stored charge/primed device that continues to deliver a trickle while the radon lasts): Plausible, untested. The positive-feedback mechanism as described: Refuted by calculation (thermionic estimate).

Claim 4: mercury/vapour variant converts HV RF/AC to DC without radioactivity

a. Described in a few sentences; no equations [read: p.6]. b. Gas/vapour rectification is old, established physics: mercury-arc rectifiers (Cooper Hewitt, 1900s) and gas-discharge diodes convert AC to DC and are used industrially [memory, high]. The asymmetry comes from electrode geometry (corona wire versus large cylinder), as in point-plane corona. So the plain claim is correct, but the claim that it is "more efficient than the prior art" because of "less electrical resistance" is unsupported [read]. c. Tested in history as a rectifier (not this device); no data for the patent's variant. d. Nothing in mainstream physics changes. Verdict: Established as a rectifier concept (not novel); the efficiency superiority is Never properly tested.

Claim 5: multi-stage radioactive-cathode valve chain from antenna and earth, "exponential" emission, Moray lineage

a. Mathematics. None given; gain per stage, couplings and component values are absent [read: deep/Atomic-Decay]. "Multiplied exponentially" with each cathode "more heated than its previous stage" requires each stage to deliver more power than it receives. A passive cascade of transformers and tank circuits cannot do this; gain requires an energy source. The only sources named are the antenna/earth ("radiant energy") and a trace of radioactive material. Energy budget [calc/memory]. The fair-weather atmospheric electric field is about 100 V/m near ground with an air-earth current density of order 2 pA/m2 [memory, medium], so the global dissipation is of order a microwatt per square metre or less. A practical antenna collecting radio-frequency energy gives microwatts to milliwatts (it is how crystal radios work). Powering "a bank of lamps" [read] requires watts to kilowatts. Neither the atmosphere nor a trace radioactive source supplies that. The inventor's own 2014 report concludes of Moray's device that "it was the circuitry, not the valve, which produced the amount of electrical power reported" [web: Wayback radioionics.com, per Investigator] , which relocates the energy question from the nuclear part to the circuit, where it is the unsolved question. b. Moray's devices (1920s-30s demonstrations, claims of kilowatts) lack independent, recorded, controlled measurement as far as I know [memory, medium]; the leaflet's Betz drawing is "drawn from memory" 1954, "only correct in a rough detail", and its radioactive-salt idea is Betz's own question ("Could it supply needed ions?") [read: deep read]. So the leaflet's "as Moray's valve did" is attestation by recollection, not data. Germanium "very very pure" in a 1927 affidavit [read: p.2] is an unusual detail for 1927, since high-purity germanium semiconductors were developed in the 1940s [memory, medium]; the quotation comes from a later book edition, so dating is uncertain. I raise it as a question, not a refutation. The Investigator also found a 1998 article whose first-person cosmic-ray passages track statements attributed to Tesla [memory, medium]; this bears on provenance rather than physics. c. No replication on record. Builders posted on Energetic Forum (2017) and groups.io (2020), no measurements seen [web, per Investigator]. Related lines by others (Coler, Hendershot, Hubbard, Britten, Prentice, Gray, Plauson, listed in the inventor's 2018 paper) are also uncontrolled historical claims; I keep them in the record as the inventor's lineage list. d. Exponential output growth from a passive cascade with a micro-watt source would violate energy conservation.

Verdict: Never properly tested (no neutral controlled test; no data in either document) in the sense of the experiment; Refuted by calculation for the specific statement that lamps are powered through exponential multiplication from antenna/earth plus trace radioactivity (energy budget). The lineage statement (Moray's design) is Not testable as stated (no composition or drawing that is verified).

Claim 6: printed nuclear equations support the device

a. Not internally consistent (see Claim 1): prose says Ge-72 gives stable Se-77, equations show Se-75 and a Ge-76 appearing from nowhere; reaction energies absent; the chain's first step is endothermic [read; calc]. b. Individually each beta decay is known physics: Ge-77 -> As-77 (beta, max about 2.7 MeV, as printed by Jaynes) -> Se-77 [memory, medium]. They are real but require neutrons. c. Untested; only a derivation of the threshold by Jaynes (hosted by the inventor) exists [web: Wayback, per Investigator]. d. Nothing new. Verdict: Internally inconsistent (isotope bookkeeping and the prose conclusion) and the net-support claim Refuted by calculation (Claim 1).

Weighing the Investigator's findings (kept whatever the verdict)

  • Jaynes calculation (inventor-hosted; author unidentified): supports the 6.06 MeV figure as a mass balance; I find the figure correct as the Q-value but it is the cost, not the yield; his alpha-energy table lists decay Q-values, not alpha kinetic energies.
  • Abbood and Ebrahiem (2020), Ibn Al-Haitham J., cite the patent and compute alpha-induced Ge cross sections from evaluated libraries [web]; I did not read the paper. It would give the actual rate. It is the most useful next document.
  • NASA NTAC patents (9 cite US7800286): related secondary-electron-avalanche concept from a mainstream institution; measured outputs not yet checked. Not support for "alpha fusion".
  • CN113539542B (radon battery, PLA Rocket Force University of Engineering): same fuel idea; supports that radon is considered as an alpha-battery fuel.
  • Weber US20100062288A1, Moses US20220108814A1, earlier art (US2517120, US2527945, Linder 1952, Windle 1964): show conventional beta/alpha-voltaic and electron-multiplication art at microwatt level.
  • Historical record: patent lapsed 2014; entities (Nu Energy Horizons, Alpha Fusion Energy Corp, SonoMagnetics Trust, Radioionics) left thin records; OpenCorporates found nothing. The shift of the inventor's own explanation from alpha fusion (2007-2010) to gamma/charge multiplication (2014), plasma valve (2018), electrets (2022) is an anomaly in the record. Competing explanations: the inventor concluded the first mechanism did not give power; the inventor treats these as a continuous research programme; commercial/funding failure. The record found so far cannot decide; the NH register and USPTO file wrapper (11/744,769) might.
  • Perreault Plasma Valve and Battery (2018): no measurements; "proprietary cathode and anode" withheld; Fe/Pb electrodes with atmospheric ions. A battery of dissimilar metals in air is a galvanic cell (an Fe/Pb pair, 0.5 V drawing in the paper), with a real but chemical energy source [memory, medium].
  • Radon-is-the-ultimate-fuel paper: figures "1.5e9 cal/g" -> "5,944 BTU/g" as printed. 1.5e9 cal = 5.95e6 BTU [memory, high] so the printed value is off by a factor of 1000 (kBTU vs BTU), and the stated "13.5 times" comparison depends on it. Decay energy per gram of Rn-222 with its three alpha daughters is 19.2 MeV x 6.022e23/222 = 8.3e9 J/g, about 2.0e9 cal/g [calc], so the printed 1.5e9 cal/g is the right order of magnitude (it is the decay energy of the chain, not a burnable fuel value; the 'ultimate fuel' comparison with fission fuel mixes a quantity that is released over days with one released on demand). The 1 Ci = 6.5 microgram scale above shows the intrinsic limitation: radon fuel decays in days.

Corrections to Step 1

  1. The Investigator noted that the standard Q is -6.06 MeV and the lab threshold about 6.40 MeV: confirmed [calc]. Add: Rn-220 (6.288 MeV) is below the true threshold although the patent lists it as usable.
  2. claims.md item 6 says the chain is inconsistent "on isotopes"; correct, and the first step is also endothermic, which claims.md does not mention.
  3. claims.md item 2 says the millionth-of-a-watt figure is attributed to earlier devices: confirmed [read]. The patent's phrase "extremely low amperage" for earlier art is the inventor's characterisation (applicant's stake) and is not supported with a citation.
  4. The leaflet says the circuit is "a continuation of U.S. Patent ..." (text truncated in the deep read); the continuation relationship is not checked.
  5. The Investigator's Radon paper conversion error (cal to BTU) is correct as a factor-of-1000 slip; whether the rest of the paper depends on it is unknown (only pp.1-2 read).

Open items (not done; do not assume)

  • Read Abbood and Ebrahiem 2020 for the actual Ge-72(alpha,n) cross section near 8-9 MeV; I give only a barrier-penetrability estimate.
  • Read "HF-box-lab-tests.pdf" and the full "Radon Gas" paper (Wayback) for any measured data.
  • USPTO file wrapper for 11/744,769.
  • Web budget: 2 searches used for the Ge cross section and alpha energies; the Jaynes/Rn-220 chain, Moray and atmospheric-field items remain [memory, medium] and unchecked.

Verdict summary

  1. Alpha fusion on Ge: strong form Refuted by calculation (Q -6.06 MeV); weak form (secondary electrons) Established.
  2. Radon cell, high current density: low-power cell Plausible, untested; high power Refuted by calculation (about 0.11 W/Ci ceiling).
  3. Self-generating: Refuted by calculation; trickle after priming Plausible, untested.
  4. Vapour rectifier: Established as concept; superior efficiency Never properly tested.
  5. Moray multi-stage chain: Never properly tested; lamp-power claim Refuted by calculation (energy budget); lineage Not testable as stated.
  6. Printed equations: Internally inconsistent.