Perreault-Bruce: Mathematician review (Step 2)

Reviewer: PhD mathematician role. Sources: facts.json, claims.md, deep/ reads, sites.md "Leads for review", research/ notes (research/files/ is empty; no downloaded documents to read). Tags: [read] document, [web] URL, [memory] recall with confidence, [calc] my own calculation (run in the project venv this session).

Source gap: none for the text (patent and 2-page leaflet read in full). But neither document contains a single measurement, so there is nothing numerical to check except the printed equations, the 6.06 MeV figure, and the circuit logic. All "does it work" questions below are answered by calculation or by absence of data.

Mathematical background you need (short)

• Q-value: for a reaction A + a -> B + b, Q = (sum of rest masses in - sum out) c^2. Q>0 releases energy (exothermic); Q<0 absorbs energy (endothermic) and needs a minimum "threshold" kinetic energy. With projectile mass m_a and target mass M_t, threshold = -Q (1 + m_a/M_t), because some kinetic energy must stay in the centre-of-mass motion.
• Coulomb barrier: an alpha (charge 2) and a nucleus (charge Z) repel; barrier height roughly B = 1.44 MeV·fm × 2Z / R, with R about 1.2 (A_t^(1/3) + 4^(1/3)) fm. Below the barrier, reaction probability falls roughly as exp(-2 pi eta), eta = 2Z e^2/(hbar v) (Gamow factor).
• Delta ray: a fast secondary electron knocked out of an atom by a passing charged particle. Maximum energy given to an electron by a heavy particle of kinetic energy E and mass M is T_max = 4 (m_e/M) E.

Claims and checks

Claim 1: alpha particles above 6.06 MeV cause an "alpha fusion reaction" that releases high-energy electrons

Printed reaction: 32Ge72 + He4 -> 34Se75 + n; "at least 6.06 MeV" needed [read: US7800286 p.4-5, page image].

a. Internal consistency. Mass and charge balance of each printed line is correct (72+4=75+1; 32+2=34) [read, confirmed in deep read]. The 6.06 number is arithmetically right, but it is the NEGATIVE of the Q-value. Using atomic masses Ge-72 71.9220758 u, He-4 4.00260325 u, Se-75 74.9225234 u, n 1.00866492 u [memory, medium; standard AME values]: Q = -6.063 MeV [calc]. So the reaction is endothermic: it consumes 6.06 MeV (and strictly needs 6.40 MeV of alpha energy once recoil is counted [calc]). The patent calls it a "fusion reaction" that "generates" energy; the sign is the opposite of what the narrative needs. Jaynes (inventor-hosted calculation, "a substantial amount of energy") reaches the same 6.06 figure by mass balance per the Investigator; the Investigator's note that the standard Q is -6.06 and threshold about 6.40 is CONFIRMED by my calculation. Whether Jaynes treats 6.06 as a release or a cost I could not read (document not saved); his conclusion as reported is a reading of the sign I cannot reproduce. Second inconsistency, already noted in the deep read and confirmed: prose says Ge-72 yields "stable Se-77", equation 1 gives Se-75 -> As-75 (stable, electron capture), and equation 3 starts from Ge-76, which is not a product of equation 1. The Se-77 end of the chain belongs to neutron capture on Ge-76 (natural abundance about 7.8% [memory, medium]), a different target isotope. A direct route to Se-77 does exist: Ge-74(alpha,n)Se-77, Q = -4.47 MeV, threshold 4.71 MeV [calc]; so the "6.06 MeV threshold" does not even apply to the most abundant germanium isotopes (Ge-74 about 36%, [memory, medium]). The patent's list of "usable" and "unusable" alphas (Po-218 6.00 and Po-210 5.30 "cannot be used") is therefore not a property of germanium; it is a property of one isotope. Delta-ray claim: ordinary alpha ionisation already produces delta rays with no nuclear reaction. T_max = 4 (m_e/M_alpha) E gives 4.8 keV for the 8.78 MeV alpha and 2.9 keV for 5.3 MeV [calc]; these are "fast" but keV, not "high-energy" in the sense of MeV beta particles. The only MeV electrons in the chain are the beta decays of Ge-77 and As-77 (patent eq. 4-5), and those need neutrons.

b. Experiment. Alpha-induced secondary-electron emission and radon-daughter collection on negative electrodes are established; positive Po-218 ions plate out on a negatively charged wire [memory, high]. That supports the patent's "attracting radon onto the corona wire" only for the charged daughters, not for neutral radon. Nuclear reactions of 6-9 MeV alphas on mid-mass nuclei are known to be strongly suppressed (see d below). I know of no measurement of the claimed effect.

c. Tested? No test by anyone on record. Abbood and Ebrahiem 2020 compute cross sections from evaluated libraries but that is calculation (the Investigator's lead); I tried to fetch the full PDF and the fetch returned only navigation text, so I did not read their values [web: https://jih.uobaghdad.edu.iq/index.php/j/article/view/2374/1571, fetch failed].

d. If true: nothing in mainstream physics, as such, but the Coulomb barrier. Ge-72 + alpha barrier is about 13.4 MeV in the centre of mass (14.1 MeV lab) [calc]. A 7.69 MeV alpha (Po-214) has a Gamow factor exp(-2 pi eta) of about 1.5e-20, 8.78 MeV about 2.7e-19 [calc]. Bound on rate: take a very generous cross section of 1 microbarn (1e-30 cm^2), Ge-72 number density about 1.2e22 /cm^3 (Ge is 4.4e22 /cm^3 times 27% Ge-72 [memory, medium]) and only the first 5 micrometres of alpha path being above threshold: probability about 6e-12 per alpha [calc, assumed inputs]. For 1 curie (3.7e10 Bq) of Po-214 alphas that is about 0.2 reactions per second, each yielding no prompt electrons (Se-75 decays by electron capture; Auger electrons of eV-keV) [calc]. The reaction cannot be the source of any current that matters. The Gamow estimate says the true figure is probably far lower.

Verdict claim 1: SPLIT. (i) Energy release from the Ge-72 reaction: Refuted by calculation (Q = -6.06 MeV, endothermic; rate negligible). (ii) Delta-ray electron emission from alpha bombardment: Established, but of keV energy and not requiring any "fusion". (iii) The specific chain and Se-77 end product as printed: Internally inconsistent.

Claim 2: radon-filled vessel with Ge/Si/PbS wire and Pd/carbon collector gives much higher current than earlier nuclear batteries

a. Patent gives no number, only "greater amperage per density" with undefined "amperage per density" [read: p.6]. Check via energy accounting. Alpha power available from radon plus daughters in equilibrium: Rn-222 5.49 + Po-218 6.00 + Po-214 7.69 = 19.2 MeV per chain [memory, high], so 1 Ci gives 3.7e10 x 19.2 MeV x 1.602e-13 J/MeV = 0.114 W of alpha kinetic power at most [calc]. Delivering 1 microwatt needs about 8e5 alphas/s of 7.69 MeV, i.e. about 22 microcuries [calc]. So the prior-art "millionth of a watt" is just what small sources give; larger power needs larger radon activity, not a new reaction. The ceiling for any alpha-voltaic device is 0.114 W per curie before conversion losses, whatever the electrode. Ionisation current is large but is not power: a 7.69 MeV alpha makes about 2.2e5 ion pairs (35 eV per pair [memory, medium]), so 1 Ci of Po-214 alphas gives about 1.3 mA of saturated ion current [calc]. With a 1 kV bias applied, current times voltage is 1.3 W, 11 times the alpha power of the whole chain. That is possible only because the "power" is mostly supplied by the 1 kV source; output measured as I x V at the bias is not generated power. This is a plausible hidden-assumption trap for any test of this device: net power must be (power out) minus (power into the HV supply), measured over the whole loop. The patent's own preferred mode keeps the external supply on [read: p.6]. b. Consistent with Corliss-type direct-charge and alpha-voltaic devices (low power) [read: patent's own citation]. The claim of higher current per density is not supported by any data. c. Tested by no one on record. Builders (Energetic Forum 2017, groups.io) report no measurements seen by the Investigator; absence stated, not weighed as refutation. d. If true at above the 0.114 W/Ci ceiling, energy conservation would fail. Below it, nothing in mainstream physics changes. Verdict claim 2: Never properly tested. Any output above the alpha-power ceiling (0.114 W per curie of radon in equilibrium) would be Refuted by calculation; a microwatt-class output is Plausible, untested and unremarkable.

Claim 3: once primed at 1000 V or more, "self-generating" by thermal positive feedback

a. Mathematics: positive feedback is only self-sustaining if loop gain exceeds 1 and the power budget closes. Heating raises thermionic emission (Richardson law, current ~ T^2 exp(-W/kT), appreciable above about 1000 K [memory, high]). But the heat source is alpha power, at most 0.114 W per curie [calc], in the best case, and the emitted thermionic electrons must be paid for by heat flowing in, so output cannot exceed input heat times Carnot-type efficiency. The patent gives no loop gain, no energy accounting and no stable operating point; the quoted "boiling off additional electrons" is a qualitative description of exponential growth that has no stated limit. The preferred mode keeps the external supply on [read: p.6], which already concedes that it is not self-sustaining in the preferred form. Also the helium build-up statement [read: p.5] says the device chokes itself. b. No experiment; c. none. d. A self-powering loop would violate energy conservation unless the net source is the radon. Verdict claim 3: Refuted by calculation in the strong form (sustained output exceeding the alpha power, or free of external input); the weak form (a device that delivers a small fraction of the alpha power while fed by a priming supply) is Plausible, untested.

Claim 4: mercury (or other) vapour variant converts HV RF/AC into DC even without radioactive material

a. Consistent: a mercury-vapour tube or corona gap is asymmetric and conducts one way (mercury-arc rectifier; corona rectification) [memory, high]. "Less electrical resistance than prior art" is unquantified [read: p.6]. b. Consistent with established rectifier practice. c. Rectification by vapour tubes is well tested historically; the patent's variant has no test on record. d. Nothing. The claim is only rectification (output energy at most equal to input minus losses), so no energy issue arises. Verdict claim 4: Established for rectification in principle (mercury-arc rectifiers), Never properly tested for this specific geometry and for the "more efficient" statement.

Claim 5: multi-stage chain of indirectly heated radioactive-bearing cathode valves, antenna and earth, with "exponentially multiplied" emission, powering lamps as Moray's germanium valve did [read: leaflet pp.1-2]

a. Mathematics: "multiplied exponentially" across stages is stage gain g^n. That is ordinary cascaded amplification, requiring a power source at each stage. The leaflet says the heater battery B and rheostat r "no longer required" because the cathode is heated by the previous stage's signal; so stage k+1 is heated by the output of stage k, and each stage must output more power than it receives. Power gain above 1 per stage needs a supply. If the supply is the antenna and earth, a stage drawing from an antenna is limited by ambient atmospheric power; if the supply is the radioactive "small amount", it is limited by alpha power as above (microwatts for microcuries). Neither source is given a number. In the drawing the valves are drawn as diodes with a third element not connected [read: leaflet Fig. 3, deep read]. "Exponential" with no saturation is mathematically impossible for a bounded source. The Fleming diagram (Fig. 2, Nature 1917) is correct as a detector circuit; calling the heater battery unnecessary is the inventor's statement, with no data. Historical items: Betz's own note is a question ("Could it supply needed ions?"), not an assertion; John Moray's "very very pure germanium mixed with various substances" is not a composition [read: leaflet pp.1-2]. The 2014 report by the inventor concludes that in Moray's device "it was the circuitry, not the valve" gave the power [web: Wayback radioionics.com, per Investigator; I did not re-read it]. That statement shifts the question from nuclear physics to the resonant circuit. b. Resonant tuned circuits do store and build up energy (Q factor), which is established [memory, high]; building up a signal in a tank is not creation of energy. c. No replication on record; the Moray apparatus was reportedly examined and not reproduced [memory, medium; this was not verified on the web in this review]. d. If an antenna and earth supplied kilowatts, atmospheric-electricity energy flow would have to be orders of magnitude above the measured fair-weather air-earth current (about 1e-12 A/m^2 [memory, medium]). Verdict claim 5: SPLIT. Net power gain beyond the source: Refuted by calculation (energy budget, with no source numbers given by the inventor). Resonant buildup and antenna detection: Established. Moray's historical device performance: Never properly tested by neutral parties (as far as the record shows; see Investigator's historical leads).

Claim 6: the printed nuclear equations support the device

a. Covered under claim 1. Additional check of the beryllium example: Be-9(alpha,n)C-12 has Q about +5.70 MeV (exothermic) [memory, high], so this reaction does release energy and it is the classic Ra-Be neutron source; it is the inventor's analogy, but its neutrons and not electrons are the product, and Be-9(n,2n)Be-8 then 2 alpha is also standard (threshold about 1.7 MeV) [memory, medium]. The inventor's choice of Ge is not analogous: Q is negative. The Fermi patent US 2,206,634 is cited for this; I did not verify its content [memory, low]. b. Neutron production by Ra-Be sources is established. Not for Ge. c. Untested for Ge. d. Nothing. Verdict claim 6: Internally inconsistent (isotopes, Se-75 vs Se-77, Ge-76 source) and Refuted by calculation as a support for energy release (Q < 0). The beryllium example by itself is Established physics, but does not support the device.

Investigator findings kept in the record (with my reading)

• Jaynes calculation (inventor-hosted): shows the reasoning the inventor relied on; my result on the sign and on the 6.40 MeV threshold differs (see claim 1). Author identity unknown. Not read directly.
• Abbood and Ebrahiem 2020: independent academic calculation of alpha cross sections on Ge; would settle the rate in claim 1; fetch failed this run, still to read.
• NASA NTAC (Choi et al., nine patents citing US7800286), Moses US20220108814A1, CN113539542B radon battery, Weber US20100062288A1: related lines of work on radiation-driven electron emission and radon-fuelled alpha batteries. They share the secondary-electron idea. Their citation of the patent is not evidence it works (citation lists include prior art listed for context). NTAC measured power not yet checked. These devices rely on known physics (secondary emission, thermionic emission from heated radioisotope), with the same ceiling of source power.
• "Radon Gas is the Ultimate Fuel": the Investigator flags the figure "1.5e9 cal/g translates into 5,944 BTUs per gram". Check: 1.5e9 cal = 6.28e9 J = 5.95e6 BTU, so 5,944 BTU/g is wrong by a factor 1000 if the cal/g is right (1 BTU = 1055 J; 1.5e9 x 4.184 / 1055 = 5.95e6) [calc]. The printed value is off by a factor of 1000 (kilo-BTU vs BTU) unless "cal" there means a different unit. Separately, radon-222 half-life 3.82 days gives alpha energy per gram of the order of 1e9 J? A quick check: 1 g Rn-222 = 2.7e21 atoms; 19.2 MeV per atom in chain = 8.3e9 J [calc] (about 2e9 cal), so 1.5e9 cal/g is the right order if it is the total chain alpha energy of pure radon, but pure radon is released over days, so a gram is an enormous activity (about 160 kCi [calc]) and the energy is released at about 18 kW per gram only at the start. Fine as an order of magnitude; the point is that it is the radioactive decay energy, not "fusion".
• Absence-of-evidence: US-only patent that lapsed unpaid in 2014; Alpha Fusion Energy Corporation not found in registries; "HF-box-lab-tests.pdf" not read. Competing explanations (commercial abandonment, unregistered trade name, withdrawn content, inability to show output) remain open; a lapse is not itself proof of failure, and nothing here decides between them.
• Cosmic-ray paragraphs in the 1998 article resembling Tesla statements (without quotation marks): flagged by Investigator, not evaluated here.

Corrections to Step 1

1. claims.md claim 1 says "above a 6.06 MeV threshold" and describes the reaction as releasing electrons. Correct: 6.06 MeV is the magnitude of a negative Q (cost), true threshold about 6.40 MeV; releasing electrons from this reaction is not supported (Se-75 decays by electron capture) [calc].
2. claims.md claim 6 says the chain "is not internally consistent on isotopes". Confirmed, with more detail: Ge-74(alpha,n)Se-77 is the direct route to Se-77 with Q = -4.47 MeV.
3. Deep read uses "Po-214 7.68 MeV" as printed; the tabulated value is 7.69 MeV [memory, medium]; immaterial.
4. Deep read correctly reports no data; facts.json and claims.md also say so. No other errors found in Step 1. One item I could not verify: the Investigator's 1.5e9 cal/g and BTU conversion (original page not read by me).
5. Web checks: the one fetch (Abbood and Ebrahiem) failed. Mass values and 8.78/7.69/6.78/6.00/5.30 MeV alpha energies are [memory, medium] and are reproduced from standard tables; the 6.06 MeV match with the patent's own number is an internal check on the masses.

Verdict summary

1. Alpha-on-Ge "fusion" releasing electrons: Refuted by calculation (energy release); delta-ray emission Established but keV.
2. Radon-Ge corona device with high current: Never properly tested; above-ceiling output Refuted by calculation.
3. Self-generating by thermal feedback: Refuted by calculation (strong form); weak form Plausible, untested.
4. Vapour AC-to-DC variant: Established (rectification), specific claim never properly tested.
5. Exponential valve cascade: Refuted by calculation (net power); resonance Established; Moray performance never properly tested.
6. Printed nuclear equations: Internally inconsistent.