TernionAG: physicist review (Step 2)

Reviewer role: PhD physicist, mainstream view. Date 2026-10-03. Inputs: claims.md, facts.json, all deep/ reads, key passages of the deep reads of US2089966, GB417501, GB483062 (plate-circuit patent), the Debye/heat-pump note and the essay in the project folder. Tags: [read] from a project document (via the deep reads, which cite page images), [calc] my arithmetic (project venv/numpy), [memory] recalled with confidence, [web] none (see "Web checks").

Source gap repeated: every Kassner/Ternion scientific claim rests on the patentee's own text. There is no lab notebook, independent test, replication or Kassner biography in either folder. Conclusions about tissue, colloid, rubber and preservation effects rest on the patents' assertions plus [memory]. Section C claims rest on the essay and secondary AI notes.

Web checks (budget note)

I tried to retest the medium-confidence [memory] facts the earlier reviews could not verify (tissue/serum dielectric dispersions; non-thermal microwave effects; 1945 electronic vulcanisation). The WebSearch budget for the session was exhausted and the one WebFetch (Wikipedia, dielectric spectroscopy) did not contain the needed detail. The checks FAILED. Those facts stay [memory] with the confidence stated, and no verdict below turns on them alone.

Physics background used (short, for the reader)

• Photon energy E = h f, f = c/lambda. For 18.6 cm, f = 1.61 GHz, E = 6.7e-6 eV; thermal energy kT at 298 K = 0.0257 eV, ratio 3,850. For 30 cm: 6,200. For 90 cm: 18,600. For the 1.28 cm water band (23 GHz): 265 [calc]. An O-H bond is about 4.8 eV [memory, high], 720,000 times a 18.6 cm photon [calc]. A single photon at these wavelengths cannot break or weaken a bond; any non-thermal effect must be collective (orientation, field-driven transport) not photochemical.
• Debye relaxation: water, tau = 8.3 ps, loss peak at f = 1/(2 pi tau) = 19.2 GHz (lambda 1.56 cm) [calc; tau memory, high]. This is a broad "friction" dispersion (Q of order 1), and the patents' own 1.5 cm figure for water matches it [read: US2089966 p.5].
• Conducting liquids (electrolytes, serum): the dominant response at 0.3-2 GHz is ionic conduction. Serum sigma ~1.5 S/m [memory, medium] gives sigma/(omega eps0) = 81 at 333 MHz and 20 at 1.6 GHz [calc]; skin depth 2.3 cm at 333 MHz and 0.94 cm at 1.6 GHz [calc, ignoring the dielectric correction]. Such a medium is strongly damped. Tissue and protein dispersions (beta, delta) are broad, spanning decades of frequency [memory, medium]. A resonance with half-width under 5 percent (Q > 20) in a conducting aqueous liquid would be an extraordinary feature.

Claim-by-claim

Claim 1: dipolar substances have sharp natural frequencies coinciding with anomalous-dispersion bands (electrolytes 2 percent wide, serum near 90 cm, water 1.28 to 1.5 cm)

a. Mathematics: no maths is offered; it is an empirical assertion. The patents themselves separate "friction dispersion" (Debye, broad, water at 1.5 cm) from a "resonance dispersion" in conducting dielectrics "scarcely wider than 2/100 lambda" [read: US2089966 p.5, page image]. The 1.5 cm figure is consistent with Debye theory [calc]. The 1.28 cm figure for "chemically pure water" is labelled the shortest band in serum, which is an odd place for a pure-water number [read: US2089966 p.8]. b. Experimental consistency: the friction-dispersion half is established (water's ~20 GHz peak). The "resonance dispersion" half conflicts with what I know of electrolyte and serum permittivity: smooth, broad dispersions, no 2 percent line near 90 cm [memory, medium; web check failed]. Water dielectric relaxation shows no sharp resonance below THz [memory, high]. The method of detection (lightening of a beam between crossed Nicols while the sample sits in a plate oscillator) is a birefringence readout. The patents never show a spectrum or tuning curve. As Deep read §1.9 notes, the plate oscillator's own size is the only handle, so the sample's "resonance" is never separated from the cavity's resonance. An electric dipole in a damped liquid is not an oscillator with a long ring time (Q = 50 at 333 MHz would mean ringing for about 48 ns [calc]), whereas liquid-phase molecular reorientation occurs in picoseconds. c. Tested: not by anyone on record. The patentee reports it in 1932-37; no independent replication in the folder; Voss 1987 only comments on the patent's foresight [read: Voss]. Dispersion in serum and water has been measured thoroughly by many groups since (Schwan, Gabriel) and, in my memory, no sharp 2 percent resonance has been reported [memory, medium; not verified on the web]. d. If true: dielectric theory of liquids would need an unknown collective ionic resonance with Q of tens in a medium with sigma/(omega eps0) near 80. This would be new physics in condensed matter, not a small amendment. Verdict: split. Friction dispersion of water near 1.5 cm: Established (Debye). The sharp "resonance dispersion" of electrolytes, serum and colloids: Never properly tested (no neutral test on record; not clearly dismissed without test, simply never taken up), with a mainstream expectation against it; plausibility low because strongly damped media do not support Q ~ 50 [calc + memory].

Claim 2: field at a natural frequency permanently alters "energy content" by orienting molecules (double refraction), and a sample frozen while oriented (electret) is "richer in energy" and re-radiates the same wavelength on thawing

a. Maths: internally the patent is a verbal argument. The surface charge range 0.5 to 2.0 x 10^-9 C/cm^2 is given as an order of magnitude only [read: US2089966]. Electrets are real (stored polarisation, e.g. wax or polymer electrets; the patent itself admits they were known). "Energy content" is never defined as a thermodynamic quantity: no measurement, no J/g. b. Experimental consistency: field-induced orientation producing birefringence (Kerr effect) is established; the Kerr effect in water is real but tiny and relaxes in ps [memory, high]. Freezing in an orientation to make an electret is established for ferroelectrics and some polymers, but ice (hexagonal ice Ih) is not a good electret and the stored energy of an electret is the electrostatic energy of the charge/dipole layer, which is minuscule: with surface charge sigma = 2e-9 C/cm^2 = 2e-5 C/m^2, field sigma/eps0 = 2.3e6 V/m and energy density (1/2) eps0 E^2 = 24 J/m^3 [calc]. Releasing it on thawing as radiation "at the same wavelength" has no mechanism and contradicts thermal equilibration on thawing (Debye relaxation in ps) [memory, high]. The "same wavelength" re-emission is a claim a simple detector could test; none is reported. c. Tested: no detector measurement in the patents; no replication on record. d. If true: relaxation and entropy of ice/water would need a long-lived non-equilibrium orientation reservoir carrying radiative memory. A modest electret (stored polarisation) is mainstream; "returning energy as radiation" is not. Verdict: split. Field-induced orientation and electrets: Established. Permanent "energy content" gain and radiative re-emission on thawing: Never properly tested; measured energies would be tiny (tens of J/m^3 at best [calc]).

Claim 3: practical effects (preservation, colloids, vulcanisation/Bakelite/polymerisation without catalyst, seed growth, effects on men and animals)

a. Maths: none; operating figures: 0.1 to 250 W, shelf-life improvements stated without sample size, spoilage criterion, bath temperature or off-resonance and heat-matched controls [read: GB417501, US2089966]. 10 W absorbed in 1 g of water would warm it 2.4 K/s [calc], so heating is the default explanation unless excluded; the patents do not exclude it. b. Experimental consistency: heating and dielectric heating explain vulcanisation, curing, drying and thermosetting resin cure (Bakelite) in conventional RF dielectric heating, and 1945 clippings say Goodrich and Firestone bought electronic-vulcanisation patents [read: clippings; patent identities unknown]. Pasteurisation-like shelf-life gain from heating is mainstream. Non-thermal specific "resonant" effects at microwave and RF frequencies have been repeatedly sought and are controversial and, in my memory, not established for food preservation or macroscopic chemistry [memory, medium; web check failed]. The gold-sol colour change red to orange then flocculation could result from heating or field-driven aggregation (colloid electro-optics) [memory, medium]. The "men and animals" claim is a "may" statement [read: claims.md]. c. Tested: only by the patentee; no controls described. No replication on record. Not dismissed without testing, but never taken up in a documented way (Ternion liquidated in 1937 for lack of funds [read: Ternion AG 2]). d. If true (for the resonant, off-resonance-null form): a new class of narrow-line, non-thermal field-matter coupling for liquids at sub-GHz frequencies, which would revise bioelectromagnetics, where the accepted consensus is that effects below the thermal level need a specified mechanism [memory, medium]. Verdict: split. Heating-driven curing, drying, preservation (thermal): Established (conventional RF dielectric heating). Resonance-specific non-thermal effects with off-resonance null: Never properly tested. Human effects: Not testable as stated (a "may" with no dose or endpoint).

Claim 4: plate pair with thin dipolar layer is a tunable sharp resonator following Bessel-zero theory (radius 10 cm disc at 26.2 cm)

a. Mathematics: GB483062 is correct textbook radial-line theory [read: deep read]. My check: lambda = 2 pi R/x. For R = 10 cm, x = 2.405 gives 26.1 cm [calc], matching the quoted 26.2. But the 1932 figures (10 cm diameter disc, 1 cm gap, lambda 30 cm; 6.2 cm at 18.6; 39 cm at 116) all give lambda/D = 3.0, i.e. x = 2 pi R/lambda = 1.05 [calc], below the lowest Bessel root Kassner lists (1.841). Even the lowest mode needs an effective eps = (1.841/1.047)^2 = 3.1 across the gap; a 1 mm water layer in a 1 cm gap gives an effective eps of only ~1.11 [read: deep read, calc]. GB483062's own 40 cm disc (R = 20 cm, d = 1 cm) resonates at 32.7 cm (x = 3.83) [calc 32.8 cm], while GB417501 says a 39 cm disc, same gap, is at 116 cm; factor 3.5. In GB483062's equations the gap d does not enter the frequency at all, but GB417501 says halving d lets a disc shrink from 39 to 20 cm at fixed wavelength. This is internally inconsistent between the two patents. [read: deep read of Plate-circuit patent; calc] b. Experimental consistency: radial resonators and stripline/disc resonators are standard microwave engineering [memory, high]; the theory agrees with experiment. The 1932 figures do not. c. Tested: the patentee says "confirmed by experiment" without data. No independent measurement. d. If true: nothing, the Bessel theory is mainstream; only the 1932 numbers conflict. Verdict: split. Bessel-mode theory of plate resonators (GB483062): Established. The 1932 tuning numbers (lambda about 3 diameters, gap-dependent tuning): Internally inconsistent with the same author's 1935 theory by a factor of about 3, and not reconciled; I cannot tell whether the cause was feed-lead/valve capacitance, a rule of thumb or an error.

Claim 5: special valves, coaxial-tube resonators, wavemeter and polarimeter; tuning tables 0.30-4.9 m; about 40 percent efficiency of two-valve push-pull circuit

a. Maths: the circuit analysis in GB483062 (Q = pi/delta, N = V^2/2R, eq. (12) energy U) is dimensionally and numerically correct [read: deep read; calc, agrees to 0.15 percent]. Push-pull "doubles energy" by using both half-cycles is trivial. b. Experimental consistency: 1930s UHF triode/magnetron/Barkhausen-Kurz generation at decimetre wavelengths is historically attested [memory, high]. About 40 percent efficiency for a decimetre tube oscillator is high for the period but not impossible for the claimed push-pull class-C figure [memory, low]; no measured data shown. c. Tested: not independently on record. The patent claims are apparatus-only. d. If true: nothing; conventional engineering. Verdict: Plausible, untested (as engineering). The efficiency number is Never properly tested.

Claim 6: hand-held Tesla applicator and later lamps radiating ultra-short waves from 0.5 mm upward for therapy, with efficiency improved by exciting ultra-short oscillations or matching organic absorption bands

a. Maths: none. A spark gap emits broad-band damped waves; "waves from 0.5 mm" is the spark-transmitter limit idea (Glagolewa-Arkadiewa mass radiator, 1923) [memory, medium]. No spectrum, dose or efficiency. b. Experimental consistency: Tesla/diathermy therapy existed; sub-mm radiation from sparks at milliwatt level is real; lamp radiating IR is a known filament. Matching "organic absorption bands" with broad thermal IR is conventional IR therapy. Therapeutic benefit at these doses is not established. c. Tested: no dose or controlled trials in the documents. d. If true: nothing for the emission; therapeutic specificity would need new biophysics. Verdict: Not testable as stated (no dose, spectrum or endpoint); the physical emission is Plausible, untested.

Claim 7: Kassner's work was commercially significant but obscure (1945 vulcanisation patent purchase; Voss 1987)

a. Not a mathematical claim. b. A historical claim: the clippings do not give patent numbers or seller [read]. Voss writes that US 2,089,966 showed Kassner "foresaw" microwave oven and industrial applicators, hedged [read: Voss]. The patent does describe field heating with oscillating fields; but dielectric heating existed from the 1920s-30s (Schereschewsky, short-wave diathermy) [memory, medium], so "foresaw" is a generous reading. c. n/a. d. nothing. Verdict: Plausible, untested: the chain Kassner to Goodrich/Firestone is unproven; Voss's claim is an opinion.

Claim 8: Ternion AG corporate history (founded 18 April 1932, CHF 100,000, research account CHF 1.12 million, liquidation voted 20 Dec 1937, deleted August 1952)

Not physics. Documented by register copies [read: Ternion AG 2/3, Statuten]. The physics-relevant inference: CHF 1.12 million of loans spent on research with no recorded product is consistent with an experimental programme that did not deliver commercially in five years; it is not evidence that the effects are false, nor that they were suppressed. Verdict: Established as history (documents); outside the verdict scale for physics. Kassner is not named in the register papers [read].

Claim 9: Bauch did a Ternion project on UHF during a 1934 GEC secondment

Not physics; one sentence in a 1996 memoriam, GEC stay corroborated by IWM interview, but not the Ternion project [read: claims.md]. Verdict: Never properly tested as history (single uncorroborated sentence). Not a scientific claim.

Claim 10: natural-frequency fields can raise or lower covalent and ionic bond strengths

a. Maths: no derivation anywhere in Ternion text; the essay's own claim. The photon energy argument above gives 3,850 times below kT for 18.6 cm and 720,000 times below the O-H bond energy [calc]. Field strength: to change a bond energy by even 1 percent (0.05 eV) with an electric dipole interaction p E needs E ~ 0.05 eV / (1 e-Angstrom) = 5e8 V/m [calc: 0.05 x 1.6e-19 J / (1.6e-29 C m) = 5e8 V/m], versus the breakdown field of air ~3e6 V/m and fields of order 1e3-1e4 V/m in 10 W laboratory plate oscillators [memory, medium]. Resonance does not change this energy scale; it concentrates energy in the mode but a molecule in liquid water loses coherence within picoseconds. b. Experimental consistency: none found. Known field effects on chemistry (Stark shift, field-assisted dissociation, microwave-assisted synthesis) are explained by heating or by fields of 1e8 V/m and above [memory, medium]. c. Tested: not tested. The idea is not in any Ternion document [read: claims.md]. d. If true: chemical bonding by weak resonant fields would contradict quantum chemistry's energy scales unless a new collective coupling exists. Verdict: Refuted by calculation in the strong form (a 1.6 GHz photon or field of laboratory strength cannot change bond energy; energy scales differ by 3-6 orders [calc]). Weak form (field-driven reorientation, reaction-rate changes through heating): Established / ordinary.

Claim 11: closed electrolysis-and-burning cycle or "chemical heat pump" yields net energy from the environment

a. Maths: I checked the Debye/heat-pump note (secondary, AI-written). Textbook numbers are right: dG = 237.1 kJ/mol, dH = 285.8, T dS = 48.7, E = 1.229 V [calc]. The note's COP = dH/(W_elec + W_res) is bounded by dH/dG = 285.8/237.1 = 1.205 [calc], reached by an ordinary reversible electrolyser at 1.23 V. Any cell between 1.23 and 1.48 V (thermoneutral 2F x 1.481 V = 285.8 kJ [calc]) already has COP above 1 without new physics. So "COP 1.4" and the Moray "COP much greater than 1.2-1.4" exceed the second-law limit, and the note's claim that the Moray gap "no longer requires violating thermodynamics" is false for COP > 1.21. The note's Landauer argument is backwards (Landauer prevents net gain from a demon in a cycle). Its water assignments are off: the 600 GHz "libration" is misassigned (water libration at 400-800 cm^-1 [memory, medium]); its candidate resonances (0.05-0.5 mm) are 25 to 200 times shorter than Kassner's water figures (1.28-1.5 cm) [calc]. Its bench diagnostic (COP_B > 1 as evidence of new physics) is wrong, since a catalyst alone gives COP > 1 below 1.48 V; the proper anomaly is total work input below dG with heat drawn from surroundings measured by isothermal calorimetry [calc]. b. Consistent with: second law, Gibbs energy as state function [memory, high]. Plasma electrolysis claims (Mizuno, Cirillo etc.) are contested and unreplicated by neutral labs [memory, medium]. c. Tested: no experiment run; Ternion never claimed it. d. If true, in the over-unity form: the second law would fail; coupling to a fuel cell would give a perpetual-motion machine of the second kind. Verdict: Refuted by calculation (the closed cycle with W_total < dG). Weak form (COP up to 1.2 by drawing T dS from surroundings): Established, and not novel.

Claim 12: engineered binding energies as weapons; defence interests suppressed Ternion

a. Not mathematical. b. The physical premise is the one refuted in Claim 10; a weapon requires an effect that does not appear to exist at available power. Directed-energy and microwave bio-effects at high power (heating, microwave auditory effect) are real [memory, high] but are thermal or thermoelastic. c. No test; evidence offered is absence (no peer-reviewed papers, Bauch missing from the GEC history); the 1937 liquidation gives an ordinary financial explanation [read: Ternion AG 2]. The essay's claim that American General Electric owns BAE Systems is wrong [read: claims.md; I accept the earlier finding]. d. Nothing in physics changes; this is a historical/political claim. Verdict: Not testable as stated (as suppression: an argument from absence). The physical premise (binding energies weaponised): Refuted by calculation (see Claim 10).

Claim 13: effects hard to replicate because frequencies shift with temperature and pressure; failures dismissed as error

a. Not mathematical. b. It is plausible that a real resonance moves with temperature (the patents name temperature and pressure of the sample as tuning handles [read]); liquid dielectric properties do shift with temperature (water tau changes ~ factor 1.5 from 20 to 37 deg C [memory, medium]). c. No replication attempt of any kind in the folders; so there are no failures to explain. d. Nothing, as a methodological point. But as stated, without a tuning protocol (how to find the resonance, how to hold temperature), it cannot lose; this is an unfalsifiable rescue. Verdict: Not testable as stated (no tuning protocol). Never properly tested.

Summary table

Corrections to Step 1

1. Claim 4 in claims.md says the 10 cm disc at 26.2 cm "follows Bessel-zero theory"; I confirm the arithmetic (R = 10 cm, x = 2.405 gives 26.1 cm [calc]) but the claim text calls the figure "radius 10 cm" while the 1932 patents' "10 cm" is a diameter (lambda = 30 cm = 3 diameters). The 26.2 cm and the 30 cm figures are for different discs and should not be set side by side as if a single disc. Say "radius 10 cm" vs "diameter 10 cm".
2. Claim 4's "factor of about 3": for the 40 cm disc it is 116 vs 32.7 cm, factor 3.5; for the 10 cm discs it is lambda/D = 3.0 versus the lowest listed mode x = 1.841 (disc of diameter 10 cm gives 17 cm in air, 2.4 gives 13 cm, 3.83 gives 8.2 cm [calc]), i.e. about 1.8 to 3.7. Report both ranges instead of "about 3".
3. Claim 1 lists "serum near 90 cm" as an electrolyte resonance; at 333 MHz serum has sigma/(omega eps0) near 81 [calc, sigma 1.5 S/m memory medium], so conduction dominates and a Q of 20-50 is implausible. Step 1 should flag that a pure Debye "friction" band at 1.5 cm is the only well-established feature and that the 1.28 cm "pure water" number is attributed in the patent to serum.
4. Claim 2: the energy-density scale of the electret (24 J/m^3 at 2e-5 C/m^2 [calc]) is missing; it shows "richer in energy" is a very small quantity.
5. Claim 10's Step 1 note gives "about 3,900 times below kT"; my figure is 3,850 for 18.6 cm (the earlier 3,900 is rounding). Also the claim should add the 720,000-fold gap to the O-H bond energy, which is the stronger argument.
6. Claim 11: the heat-pump note's "realistic ceiling COP 1.2 to 1.4" is described in Step 1 as an unsupported assertion; more exactly, 1.2 (1.205) is the textbook limit and 1.4 exceeds it. Also, COP > 1 below 1.48 V is ordinary, so the note's diagnostic is wrong.
7. Web verification status: Step 1 says it was not done; in this step the attempt also failed (search budget exhausted; Wikipedia page lacked tissue-dispersion detail). Medium-confidence [memory] items (serum dispersions, non-thermal microwave effects, Goodrich/Firestone vulcanisation patents) remain unverified and should be retried in the Arbiter step.

Investigator findings (2026-10-04)

Inputs weighed: every entry under "Leads for review" in sites.md (leads 1-29, runs 1-3) and the research/ notes of 2026-10-03 and 2026-10-04 (arnold, depatisnet, heller-pearl-chain, related-work, records-and-rechecks, plus the run 1-2 notes), and the files in research/files/ (Arnold case JSON, US 2,094,602, 2,096,459, 2,415,025, 2,495,429, Google Patents descriptions, run3 Heller/Navy PDFs). My review above (written 2026-10-03 15:18) did not mention any of them, so all are weighed here. Tags: [read] project document or research note; [web] as the Investigator recorded it (I did not repeat the web searches, so the URL is the Investigator's); [calc] my arithmetic (project venv); [memory] with confidence. Findings are leads, not tests. I keep every pertinent one whatever my judgement. Source gap unchanged: still no Kassner lab record, replication or notebook; the new items are patents, case law, trade press and others' experiments.

A. Findings that bear on Claim 1 (sharp natural frequencies)

A1. Kassner's own 1934 French texts equate his "natural frequencies" with Debye relaxation (lead 21). FR 786,294 p.1: the "fréquences propres" are "les fréquences de relaxation dites bandes de dispersion anomales", "en conformité avec la théorie des dipôles de Debye"; FR 788,428 p.1: the first natural frequency is "donnée par le temps de relaxation pour l'orientation des dipôles" [read: research note 2026-10-04-depatisnet.md, pp.1 and 6 only]. Weight: this is the best news for Claim 1. A Debye relaxation frequency is exactly the established feature (water, about 19 GHz, lambda 1.56 cm [calc]). It is a fair and mainstream reading of the "friction dispersion" half. It does not rescue the "hitherto unknown" sharp resonance dispersion of GB 417,501 (width 2 percent), because a single-time-constant Debye loss peak is about 1.1 decades wide at half height [memory, high], i.e. Q of order 1, not 20-50 [calc: Q = 1/0.02 = 50; 5 percent half-width gives Q about 10]. Kassner's French and English texts thus use one word for two different things. I only saw pp.1 and 6 of FR 788,428 (Investigator); I cannot say whether the sharp-resonance language also appears in the French text.

A2. Sharp-resonance claims by others and their tests (lead 27). For: Webb and Booth, Nature 222:1199 (1969); Grundler and Keilmann, Phys. Rev. Lett. 51:1214 (1983), narrow windows near 41.7 GHz in yeast growth. Against: Gandhi et al., Bioelectromagnetics 1:285 (1980), swept 26.5-90 GHz over DNA, RNA, cells, also frozen: "Sharp spectra reported by other workers were not observed", resonances "likely to be in error"; Gos et al., Bioelectromagnetics 18:142 (1997), controlled concurrent repeat at 41.682-41.710 GHz: "no consistently significant differences" [web: Wiley abstract pages, Investigator]. Weight: these are tests of kin claims (narrow biological resonances), not of Kassner's. Frequency regime is different: Kassner's serum band at 90 cm is 0.33 GHz [calc]; Gandhi and Gos worked at 26-90 GHz (41.7 GHz photon 1.7e-4 eV, still 150 times below kT [calc]). The Gandhi frozen-sample result touches Claim 2's frozen-electret idea only loosely (different frequency; no thawing or re-emission measured). They show that sharp-line claims in aqueous biological media have been tested where tested, and failed to reproduce, which supports (does not prove) my mainstream expectation. They do not test the 0.3-4.9 m region. I add no change to the Claim 1 verdict except to record that analogues at mm-wave are "Tested, not supported", while Kassner's own bands remain untested.

A3. Arnold's "critical frequency at the centre of the anomalous-dispersion range" (lead 24). US 2,522,082 (filed 1945) states Kassner's principle in the language of Debye dielectric physics, cites measurement of epsilon(f) to locate the S-shaped step, and the examiner cited Kassner's US 2,089,966 [web: Google Patents, Investigator]. He gets a Debye picture (dipole dragged against neighbours, absorption maximal at the centre). Physics: at the Debye loss peak absorption per unit field is at its maximum, so "activation without destructive heating" at the loss peak is contradictory unless the sample is thin or heat is removed. His surface-versus-interior selectivity needs a penetration depth: in water-like media at 27 MHz with sigma 0.01 S/m the skin depth is about 1 m [calc], so no surface selectivity from skin effect at that frequency; at the 10-20 GHz band it would be millimetres. Arnold lost on "too broad and indefinite" (237 F.2d 584, D.C. Cir. 1956), not on a physics finding [web: Justia]. One result sentence ("excellent results ... alkyd resin monomers"), no data [read: research note]. Weight: an independent, qualified claimant (dielectric chemist, JACS 1936 and J. Electrochem. Soc. papers) of the same principle, and the first record that a mainstream-trained writer put Kassner's idea in textbook Debye terms. It is a claim and a court finding of indefiniteness, not a test; I have not seen the 1955 trial evidence, which the Investigator proposes to obtain (NARA RG 21, RG 276).

A4. Contemporary notice (lead 7, 25). GB 417,501 was abstracted in Chem. Abstracts 1935 and cited in Thorpe, Weissberger and Ellis; Botanical Review 1944 places it beside Benedetti and Liebesny and Wertheim (GB 417,863, 1.8 mm to 120 m, organism-specific frequencies) [web: Internet Archive, Investigator]. All cite claims; none reports a test. Weight: shows contemporaries took the claim seriously enough to cite; no physics weight.

Claim 1 verdict: unchanged (split). Slight refinement: the Debye half is now supported by Kassner's own words as well as by mainstream physics; the sharp half remains "Never properly tested" for his bands, and analogous sharp-resonance claims at other frequencies were tested and not reproduced.

B. Findings that bear on Claims 2 and 3 (non-thermal effects, electrets, preservation, curing)

B1. Heller and Teixeira-Pinto, Nature 183:905 (1959) (lead 13, 23). 27 MHz pulsed field (about 5 microsecond pulses, 80-180 per second), garlic root tips in water between insulated electrodes with an air gap, "No temperature increase in the water was noted", chromosome aberrations "mimicked those produced by ionizing radiation", dependent on frequency, pulse, power, exposure and cell axis. A one-page letter: no cell counts, controls or statistics [read: Investigator note; copy in files/run3/]. Heller's patent US 3,095,359 (cites US 2,089,966): "each organism is responsive to a critical frequency and voltage" [web]. My check of "no heating": fields quoted by the Navy survey are 300-1,200 V/cm (3e4 to 1.2e5 V/m); at duty about 7.5e-4 and water conductivity 1e-4 S/m the heating over 5 min is 0.005 to 0.08 K, but at 0.01 S/m (not a stated value; a tap-water order of magnitude [memory, medium]) it is 0.5 to 8 K [calc], and the field in water would be lower than the quoted value because of the insulation and air gap. So "no temperature increase" is credible only if the medium was low-conductivity, and the letter does not say. Alignment of micrometre particles is a known thing: induced-dipole energy for a 5 micrometre radius sphere in water at 3e4 to 1.2e5 V/m is 6e4 to 1e6 times kT [calc, assuming Clausius-Mossotti factor 0.5], so orientation and pearl chains need no new physics (this is the Navy survey's own mainstream explanation, Maxwell-Wagner interfacial polarisation, Muth 1927, Krasny-Ergen 1936-37 [web: Navy survey]). Chromosome damage by this route is an experimental claim; the Navy survey calls it "a new specific nonthermal effect ... apparently" and "not a thermal one" [web: MRI 1960 survey, Navy contract NObs-77142]. That survey is a government contractor's reading with a stake in setting exposure limits; it records the claim, not a replication. Against: Mittler 1976, no genetic damage in Drosophila at 30 and 150 MHz (title only; full text blocked); different organism, so not a replication. Whether anyone repeated the garlic root-tip result is unknown to me and to the Investigator. Weight: this is the strongest new item for "non-thermal effects at VHF": a published, patented, government-surveyed claim by someone other than Kassner. It supports that frequency-dependent orientation effects are real and that a dispute existed; it does not support Kassner's molecular "natural frequencies" (Heller's frequency dependence is broad dielectric dispersion of particles, not a narrow resonance), and it is untested by replication.

B2. The 1930s "specific effect" dispute (lead 26). Schliephake and Compere (Klin. Wschr. 1933), Archangelsky, Kowarschik 1936; German physiotherapy snippets state no specific short-wave effect on bacteria is proven and that none of the experiments claiming athermic effect proved it [web: snippets only, Investigator; sources not opened]. Weight: Kassner's patents were filed in an open dispute, which I had not recorded. The negative statements are snippets, so I cannot weigh them as tests; I record them beside the positive reports (Liebesny and Wertheim; Benedetti).

B3. 1945 trade-press hint (lead 8). Rubber World, Apr 1945: results "may be connected in some way with electronic power itself" [web: Investigator]. A remark with no data; the Goodrich 1950 account is heating-only. Both are interested-party statements (a trade journal of the industry; Goodrich, buyer and licensor) [see Claim 7]. No change.

B4. Kappe, Pieber, Dallinger, Angew. Chem. 52:1088 (2013), and a published reply (lead 27): non-thermal microwave effects in synthesis remain in dispute [web: Investigator, entries only]. Bears on catalyst-free curing and on Arnold's "selective activation". Consistent with my "controversial, not established" [memory, medium], which is now partly sourced: the dispute is documented to 2013 [web]; I did not read the paper.

Claims 2 and 3 verdicts: unchanged. Added "stays Never properly tested" notes: for Claim 3's resonance-specific non-thermal effect, the new material adds claimants (Liebesny and Wertheim, Benedetti, Arnold, Heller) and one failed test in another organism (Mittler), none a replication of Kassner's examples. For Claim 2, nothing new on electret re-radiation; nobody on the record measured it.

C. Findings that bear on Claim 4 and 5 (plate resonators, apparatus)

C1. New patent family members (leads 4, 20, 22). FR 788,428 (1934; French member of the energy-content patent, same gold-sol, agar, milk, veal "plus de 15 jours" and juice examples; ES 135,727); GB 448,418 and US 2,094,602 (ellipsoidal reflector applicator, liquid circulated through a glass sphere at the second focus); GB 435,643 (plate-resonator light detector, "Divided out of 417,501"); US 2,096,459 (twin of GB 448,419; "about 40 %" again unmeasured) [read/web: Investigator; PDFs in files/]. Weight: no new number that tests my factor-of-3 inconsistency (Claim 4), and the 40 percent efficiency figure is again asserted, not measured (Claim 5). An ellipsoidal reflector focusing quasi-optical waves on a liquid is sound standard optics [memory, high]. Verdicts unchanged. I have not read these in full; the Investigator read only pp.1-2 and claims of FR 786,294, and pp.1 and 6 of FR 788,428.

C2. DE 1,128,930 (applied 14 Apr 1960; Kassner "†, München"; applicant Wilma Hencke geb. Unger) (leads 10, 19): UHF rotating fields from circular plate resonators for particle acceleration, plasma and "Kernfusionsforschung" [web: Google Patents, DEPATISnet]. I read only the description summary in the Investigator's note. Rotating fields from phased circular plate resonators for acceleration are conventional in concept (rotating-field and cyclotron-type resonators [memory, medium]); whether the 1960 text uses the plate-resonator numbers of the 1930s is unknown. It shows Kassner's plate-resonator idea continued for about 30 years (c. 1929 to 1960) in his own hands. No verdict impact; no fusion claim can be assessed from the summary. The Investigator found Kassner's death by May 1962 (dagger convention [memory, high]): the record of the man does not stop in 1937.

C3. Forward citations (lead 12): US 2,495,429 (Spencer, Raytheon, microwave cooking, filed 1945) lists "2,094,602 Kassner" in References cited [read: USPTO scan in files/]; Raytheon's 1954 US 2,833,657 cites US 2,089,966; RCA, Bell, GE and Westinghouse UHF-tube patents (32) cite US 2,109,843 [web: Google Patents]. The examiner or applicant treated Kassner as relevant prior art for UHF apparatus and dielectric heating. Weight: supports the engineering significance (Claim 7), and equally shows that these citing patents frame the invention as heating, not as a non-thermal effect. It does not support non-thermal claims.

D. Findings that bear on Claim 7 (commercial significance) and the 1945 vulcanisation chain

D1. A 1950 B. F. Goodrich Research Center paper (Rubber World, Oct 1950) names Kassner's US 2,089,966 and US 2,109,843 in the electronic-vulcanisation portfolio, says they "touch on this problem somewhat", and describes the work as "about 150 million cycles per second" (lead 1) [web: Internet Archive, Investigator]. This is a statement by an interested party (the licensing company, whose stake favours a heating-only reading). Beside it Kassner states his object "is not to produce a rise of temperature" (GB 417,501). Both are characterizations by interested parties and I weigh neither as neutral. The physics: at 150 MHz in rubber with some loss, dielectric heating is the standard mechanism (conventional RF heating, P = omega eps0 eps'' E^2 [memory, high]); a heating-only explanation for rubber cure is mainstream and would be expected whether or not Kassner meant something else. This does not test Kassner's non-thermal claim.

D2. Chain of ownership (leads 1, 14). H. H. Giodvad Grell (New York) held the Dufour/Leduc option (US Alien Property Custodian Vesting Order 679, 1943); Goodrich and Firestone formed Industry Inventions, Inc. (Ohio, Aug 1945), whose licensing announcements name only "Leduc and Dufour"; in Nov 1945 Grell sold unnamed "patent rights" for Fr. 200,000 to Vibranic Ltd, Glarus, chaired by Max Stoffel; Vibranic's FR 926,351 and CH 250,508 share priority dates with Grell/Richards US 2,415,025 and CH 254,655 (Firestone/Goodrich) [web/read: Investigator; US 2,415,025 in files/]. Weight: the January 1945 release names Kassner, the Industry Inventions announcements do not, and the 1950 paper lists his patents in the portfolio: so the chain Kassner-Goodrich is now partly documented in a Goodrich source, but who sold the Kassner patents and on what terms is still unknown. Absence note (prompt rule): Stoffel founded Ternion, liquidated it, and in 1945 chaired a company buying a heating portfolio from Grell; this is a pattern consistent with continuity of the patent holdings under a new name; evidence for it is the common chairman and domicile; evidence against is that only US 2,558,568 (1946) is assigned to Vibranic among Kassner patents, and nothing shows Kassner patents in the Grell sale. The Vibranic Glarus file would decide.

Claim 7 verdict: changed from "Plausible, untested (chain unproven)" to "Partly documented: Kassner's US 2,089,966 and 2,109,843 are named in a Goodrich portfolio (1950, interested party) and Kassner is cited as prior art by Raytheon; the seller, price and use remain unknown; 'foresaw the microwave oven' is Voss's opinion; the non-thermal significance is not supported by any buyer or citing patent." Reason: new [web] sources document the commercial link and the patent-office recognition.

E. Findings that bear on Claims 8, 9, 12 (history; outside the physics scale)

E1. Dipol A.-G., Arbon, closed in 1933 (lead 15): Thurgauer Zeitung 30 Dec 1933 and two reprints name "die Aufhebung der Dipol A.G." as a cause of Arbon's tax shortfall [web: e-newspaperarchives.ch]. Weight: if the laboratory ceased in late 1933 (about 18 months), the CHF 1.12 million "research account" at end 1936 (Claim 8) was either spent in Arbon in 18 months, or continued elsewhere (London 1934-36, Kassner "residing in England"; GB patents of 1934-36), or booked as a valuation of patents. I cannot decide between them; the 17 Jul 1936 minutes and Dipol's register file would. This slightly weakens the inference in my Claim 8 note ("five years of experimental programme"), which should read "at most 18 months of the Arbon laboratory plus patent work in London". Verdict (as history): unchanged.

E2. Wiegner (leads 16, 28): Dipol's signing chemist, Dr. Alfred Wiegner (b. 1905), ETH dissertation on carbides and silicides under acid attack, diploma on dyes; later textile trade [web: swisscovery, Investigator]. A dielectric or radio specialist he was not; no evidence on what was measured. Slight weight against the laboratory being a specialist dielectric one; no verdict change.

E3. Kassner's later record (leads 5, 10, 11, 19): Phono-Vision folder 1937 (KB Appenzell Ausserrhoden, CMO-59-01-A-13); Kayser visiting card 1938; General Infra Co AG, Chur, 1943-53, "spectral infra-red high-power emitters"; Swiss federal archive file E2001E#1968/78#9311* "Schwarze Liste" 1946-48 (my reading of B.51.322 as Allied blacklist files is [memory, medium]; the title alone does not prove General Infra was listed); police dossiers on "Kassner, Ernst (20.11.1900)" (identity unconfirmed) [web: archives-online, BAR]. Weight for Claim 12 (suppression by defence interests): a commercial blacklist or wartime economic-warfare file is a different thing from suppression of physics, and a German national running a Swiss infra-red emitter company in 1943 fits ordinary wartime trade restrictions. It is an anomaly in the record (the first government security-context file touching the group) and the Investigator rightly says its contents decide. Not weighed as evidence either way. Claim 12 verdicts unchanged: suppression Not testable as stated (argument from absence, now with one security-context file whose contents are unread); physical premise Refuted by calculation (Claim 10). The BAR files stay open as a record to settle explanation 5 (wartime secrecy) in the arbiter's "record goes dark" analysis.

E4. Bauch (lead 6): Wireless World Feb 1953 gives Bauch's Hertz Institute work as supersonics and "a short spell at the G.E.C. Research Laboratories, Wembley"; no Ternion or UHF [web: Investigator]. Silence, not contradiction (memoriam sentence stands as one published source). Claim 9 verdict unchanged.

E5. Other items kept in the record: 1929 "Ernst Kassner" of Elektro-Physikalische Gesellschaft (identity probable, not proven); 1923-28 "Ernst Kassner" adding-machine patents (identity unknown); SPSL file "W Kaeser - E Kassner" at the Bodleian (which Kassner is unknown); NARA matches for Giodvad (Grell) and "Ernst Kassner" in seized German records (contexts unread); BIOS Final Report 866 "High Frequency Heating" (no online copy); Goodrich Series VII (finding aid blocked); Espacenet and ScienceDirect blocked by CAPTCHA. Closed: the 1941 SHAB "Kassner Ultrakurzwellen" hit (false), GB 435,043 is GB 435,643, US 2,190,668 is Llewellyn, the "Vibranic sterilizing canned foods" index line. No physics weight.

F. What I could not do and what would settle things

• I did not repeat the Investigator's web searches; where a statement depends on a snippet or an abstract I say so. The medium-confidence [memory] items still unverified by me: serum dispersion Q and sigma (Claim 1), the non-thermal literature outside A2 and B4, and the Debye-Falkenhagen electrolyte dispersion (the mainstream benchmark for "electrolyte bands" in Claim 1; the Investigator lists it as untried; I recall dilute-electrolyte ionic-atmosphere relaxation in the MHz range and a broad, not 2 percent, width [memory, low]).
• A decisive test for Claim 1 in Kassner's own range: swept permittivity and loss of serum and 0.1 M NaCl from 0.2 to 2 GHz with a coaxial probe at fixed temperature. A Q of 20-50 resonance near 0.33 GHz would appear as a feature a few percent wide; the mainstream expectation [memory, medium] is a smooth curve dominated by sigma/(omega eps0). For Heller-type claims: repeat the 27 MHz garlic root-tip exposure with measured conductivity, thermistor logging and sham-exposed, heat-matched controls, blind scoring.

Verdict changes (this update)

1. Claim 7: Plausible, untested -> Partly documented (Goodrich 1950 portfolio list naming US 2,089,966 and 2,109,843 and Raytheon citations; chain, seller and non-thermal relevance still unknown).
2. Claim 1: unchanged (split), but the Debye half is now confirmed in Kassner's own French text and analogous sharp-resonance claims at 26-90 GHz are on record as Tested, not supported; his own 0.3-4.9 m bands stay Never properly tested.
3. Claims 2, 3, 4, 5, 8-13: unchanged; Claim 8's history note adjusted (Dipol closed in 1933); Claim 13 stays Not testable as stated, though Arnold's protocol (locate the dielectric step, track it with temperature) shows what a tuning protocol would look like, and Kassner gives none.

Investigator findings (2026-10-08)

Inputs weighed: sites.md "Leads for review" 30-40 (follow-up Q03, 2026-10-06; follow-up Q05, 2026-10-07), the notes research/2026-10-06-followup-Q03.md and research/2026-10-07-followup-Q05.md, and followups.json (A03, revised 2026-10-07). Leads 1-29 and the 2026-10-03/04 notes were weighed in the section above and are not repeated. research/files/ holds nothing newer than 2026-10-04. Tags: [read] project document or Investigator note; [web] as the Investigator recorded it, with its URL; [web, rechecked] I re-fetched the page today; [calc] my arithmetic (project venv, rerun today); [memory] with confidence. Findings are leads, not tests. Every pertinent item is kept whatever my judgement of it.

Recomputed numbers (the Investigator's arithmetic, checked)

Single-Debye water (eps_static 78.4, eps_infinity 5.2, tau 8.27 ps): eps = 41.9 - 36.6j at 19.2 GHz, loss tangent 0.87; 34.7 - 35.9j at 23.4 GHz, loss tangent 1.03; loss tangent 0.12 at 2.45 GHz and 0.05 at 1 GHz [calc]. Photon at 23.4 GHz: 9.3 J/mol; at 19.2 GHz 7.7 J/mol; 13.56 MHz 0.0054 J/mol [calc]. 22.235 GHz is 1.348 cm, 5.3 percent from 1.28 cm [calc]. Absorbed power at 19.2 GHz, 10 V/cm rms: 39 W/cm^3, about 9 K/s [calc]. Field energy stored in water at 10 kV/cm: 0.006 J/mol [calc]. All agree with the notes.

A. Bearing on Claim 1 (sharp natural frequencies)

A5. Kassner's own text says absorption is greatest at resonance (lead 36). "The excitation is especially strong and the absorption of energy therefore especially great in the case of resonance"; his Fig. 1 plots absorption beside the dielectric constant [read: US 2,089,966 p.3, via the Q05 note]. Weight: this supports the mainstream reading. At the natural frequency the sample draws the most power, which for liquid water at 19-23 GHz is dielectric heating. It sits uneasily beside his "not to produce a rise of temperature" (GB 417,501 p.6) unless the sample is thin, cooled or pulsed, none of which the patents state. It also means the heat-matched, off-resonance control that Claim 3 lacks matters more, because resonance is by his own account where heating is largest.

A6. A water-filled resonator at water's band has Q near 1 (lead 40). Q = eps'/eps'' = 1.14 at 19.2 GHz and 0.97 at 23.4 GHz [calc]. Weight: this reinforces my point about damped media (Claim 1 b) with the best-characterised case. Water at its "friction" band is one of the lossiest common dielectrics, so a plate resonator whose gap is filled with it cannot be a sharp resonator through the water's contribution. Kassner's thin-layer geometry (a 1 mm layer in a 1 cm gap) lowers the filling factor, so the cavity Q can stay high, but then the water perturbs the resonator weakly, which fits the weak effects and the separate-the-cavity-from-the-sample problem I raised in Claim 1 b. The "lightly loaded tank" premise in the Q05 hypothesis holds only with a low filling factor, and then little energy per mole reaches the water.

A7. The 1.28 cm number (lead 31). The Investigator suggests 1.28 cm "characteristic of water" may come from the water-vapour rotational line (22.235 GHz, 1.348 cm). I add a second reading: 1.28 cm corresponds to 23.4 GHz, tau = 6.8 ps, which is liquid water at a temperature well above 25 C (tau falls with temperature; I recall about 6 to 7 ps near 40 to 45 C [memory, low]); the patent itself says the band shifts with temperature and concentration. Either reading would explain the patent's two water values (1.28 and 1.5 cm) without new physics; neither is checked.

Claim 1 verdict: unchanged (split). Debye friction dispersion of water: Established. Sharp "resonance dispersion": Never properly tested for his 0.3-4.9 m bands; the Q = 1 result for water at its own band removes any role for water's Debye line as a sharp resonance.

B. Bearing on Claims 2 and 3 (non-thermal effects, "richer in energy")

B5. Second Wien effect (lead 37): strong static fields (about 10^8 V/m in bipolar membranes [memory, medium]) orient water and speed its dissociation (Onsager 1934; Mafe 1998; Bui 2021; Yamamoto 2025; Stovall 2026 disputes the model) [web: Investigator, Google snippets, abstracts only; not opened by me]. This is the nearest mainstream relative of "a field orients water and changes how it reacts", a measured effect. It is DC, a rate effect, at fields 10^3 to 10^4 times the 10^4 V/m of a 10 W plate oscillator [memory, medium], and not a resonance. It supports the orientation half of Claim 2 (already Established), not the permanent energy gain.

B6. Kassner's chemistry claim (lead 30): resonant fields "enforce" reactions "which otherwise only proceed under the influence of heat or of catalysts" (GB 417,501 p.4; examples p.5 including acetylene plus water to acetaldehyde) [read via Q03 note]. A rate claim with no yield or control. Mainstream: microwave and RF effects on rates are overwhelmingly thermal or local superheating (Kappe 2013 and the published correspondence show the matter was disputed) [web]. Niether (below) shows that selective heating of a catalyst surface can lower an electrochemical barrier with a 5 C bulk rise; that is a mainstream mechanism for "reaction at low bulk temperature" which a sympathetic reader could apply to Kassner's examples. It is a candidate explanation, not a test of his examples.

Claims 2 and 3 verdicts: unchanged. Claim 3's resonance-specific non-thermal effect stays Never properly tested. A5 adds the point that his own theory predicts maximum heating at resonance.

C. Bearing on Claim 11 (electrolysis and burning cycle) and Claim 10 (bond strengths)

C1. The follow-up questions (Q03, Q05) confirms the thermodynamics; nothing found changes the second-law result. Kassner and Ternion wrote nothing on electrolysis or bond energies [read: full-text search of _text/ by the Investigator; consistent with my own earlier finding that Claim 10 is not in any Ternion text]. The extension to water is, as the revised A03 says, a reasonable application of a principle Kassner claimed for all dipolar substances including water; whether the principle is true is Claim 1.

C2. Energy books (Q05 section A) [calc, rechecked]. dG = 237.1, dH = 285.8, T dS = 48.7 kJ/mol; dH/dG = 1.205. The "air conditioning" half of the Q05 hypothesis is correct and already in my Claim 11: a reversible cell at 1.229 V takes 48.7 kJ/mol of room heat, burning returns 285.8 kJ/mol heat for 237.1 of work, ratio up to 1.205. The "more energy than put in" half needs a second reservoir; with one surrounding temperature the most work obtainable from the gases is 237.1 kJ/mol. I confirm this independently. Corollary I add: if a field lowers the electrical work per mole by W_f at fixed T, P and products, the saving must be paid at least in full by real power delivered into the water by the resonator; a tank in which the saving appears with no matching load is the second-law violation (Q05 section D). At 1 mol H2/hour, W_f = 24 kJ/mol is 6.7 W [calc].

C3. Why 237.1 kJ/mol cannot be lowered by "weakening" bonds first (Q03 section A). Turning liquid water into vapour removes the entire hydrogen-bond network and lowers dG only from 237.1 to 228.6 kJ/mol (8.6 kJ/mol, 3.6 percent) while costing at least that to do [memory, high; calc]. One 23.4 GHz photon is 9.3 J/mol, 1/265 of RT (2.48 kJ/mol); an O-H bond (about 463 kJ/mol) is 50,000 times a photon; a hydrogen bond (about 20 kJ/mol [memory, medium]) is about 2,200 times. This is the same 3,850 to 720,000-fold gap I give in Claim 10, now applied to the water bonds specifically.

C4. Field strength to act on water directly (lead 40). Saitta, Saija, Giaquinta, Phys. Rev. Lett. 108:207801 (2012), ab initio molecular dynamics: "Fields beyond a threshold of about 0.35 V/Angstrom" dissociate molecules; weaker fields change hydrogen-bond lengths and orientation; 1.0 V/Angstrom dissociates roughly 15-20 percent of molecules [web, rechecked: https://arxiv.org/abs/1204.1120, abstract]. 0.35 V/Angstrom is 3.5 x 10^9 V/m. This agrees with my Claim 10 estimate (5 x 10^8 V/m for a 1 percent bond-energy change). It is a simulation, not an experiment, and in a standing field of laboratory strength (10^4 V/m [memory, medium]) the dissociating field is 5 orders of magnitude away.

C5. Established resonant bond weakening of water is infrared and at a surface (lead 34). Hundt et al., Science 344:504 (2014): exciting the D2O stretch with an IR laser strongly raises dissociation probability on Ni(111); in HOD, exciting O-H gives mainly H + OD and O-D mainly D + OH (J. Chem. Phys. 2018) [web: Investigator, snippets; I did not open them]. Photon 0.33-0.46 eV, about 3,400-4,700 times a 1.28 cm photon [calc: 1.28 cm photon = 9.69e-5 eV; 0.33/9.69e-5 = 3,407; 0.46/9.69e-5 = 4,749]. (Corrected 2026-10-08 in the QE remediation: this line first said "about 35,000 times", repeated from sites.md lead 34 without recomputing; see Corrections, item 5.) I accept this as a real, bond-selective, resonant effect, and it is the correct textbook counterpart of the idea. It differs on three counts: the photon is in the vibrational band, not the Debye band; it acts on isolated molecules in a vacuum beam at a hot catalyst, whereas in liquid water the energy leaks to neighbours in picoseconds [memory, medium]; and the photon energy is spent, not saved. This refines my Claim 10 note: resonant bond-selective weakening exists, but at infrared photon energies; at 1.28-1.5 cm it does not.

C6. RF and microwave-assisted electrolysis (leads 32, 38). Niether et al., Nature Energy 3:476 (2018): under high-frequency alternating magnetic fields the oxygen-evolution overpotential at 20 mA/cm^2 fell by 200 mV and hydrogen evolution by 100 mV; the oxygen gain "is equivalent to a rise of the cell temperature to ~200 C" while the cell rose "5 C only" [web, rechecked: https://ideas.repec.org/a/nat/natene/v3y2018i6d10.1038_s41560-018-0132-1.html]. The abstract gives no field frequency, strength, power or system-level efficiency, so whether the saving exceeds the field power is not shown there. Mechanism (my reading): local induction heating of the catalyst surface lowers kinetic overpotential. Serra et al., Nature Energy 5:910 (2020): microwaves reduce Gd-doped ceria which then splits water below 250 C; the authors and universities have applied for a patent (interested party) [web: Investigator]. The 2026 review in ACS Appl. Energy Mater. 9(16):10362 speaks of AC-coil losses that "aggravate this negative energy balance even further" and of one laboratory case with "a net energy saving of 215.6 kWh"; context unseen, paywalled [web: Google snippet via Investigator]. Weight: the arbiter's "RF-assisted electrolysis as a rate effect" is no longer only a hypothesis: it has been measured at a catalyst by induction heating, with the field power not yet shown to be repaid. This acts on the overpotential (the 0.4 to 0.8 V above 1.23 V in real cells [memory, medium]), never on the 237.1 kJ/mol floor. It is not a test of Kassner's standing field at water's band: different frequency, different carrier (magnetic nanoparticles), different mechanism.

C7. Roy, Rao and Kanzius (lead 33): 13.56 MHz, about 300 W, salt water; the gas above the liquid burned only while the field was on; Raman changes in the O-H stretch region; the claimants propose "resonant coupling" and do not claim net energy ("Kanzius admitted that this process could not be considered an energy source", AIChE ChEnected 2010 snippet) [web: i-sis.org.uk and snippets, Investigator]. No independent replication was found, and none found against either. A critical reading of the observation: a 13.56 MHz photon is about 10^-8 of the energy needed per molecule, so single-photon bond breaking is excluded [calc]; the reported flame needs only a few percent of the 300 W to be converted to H2 if the salt water is hot, and plasma or electrode effects in a strong field are ordinary candidates (a plasma explanation is [memory, low]; not tested). Single-source evidence is still evidence: it stands as two papers by the claimants and press accounts; the effect (burning only while the field is on) is reported by them and unexplained by them. Not a Kassner-band experiment (22 m, not 1.3 cm).

C8. Later resonance-electrolysis claims (lead 35): Puharich US 4,394,230 (1983), not read; Meyer's "water fuel cell", Ohio court 1996 "gross and egregious fraud" (the court's characterization after expert testimony that it was conventional electrolysis) [web: Wikipedia snippet]. Recorded as related work. They claim yields the thermodynamics above rules out in the over-unity form.

C9. Thermally regenerative and non-equilibrium sources (leads 39, Q05 section B). Splitting water hot and recombining cold is a known Carnot-bounded heat engine (NASA, Morehouse 1986, electrolyser at 1300 K; Chum's review) [web: snippets]. At 373 K / 298 K the bound is 0.20 [calc]. Ambient radio energy is roughly 0.1 to 10 microwatts per cm^2 [web: aggregator snippet; primary survey Pinuela 2013 not read]; collecting 6.6 W at 1 uW/cm^2 needs 660 m^2 of perfect collector [calc]. A resonator tuned at 19-23 GHz does not collect broadcast bands. So the only "environment pays" routes with a physical basis are the 48.7 kJ/mol T dS (already counted) and a hot/cold or sunlight-type second reservoir; an essay draft in the folder says the machine "cools the local environment", which is true in the reversible-cell sense only for the T dS term.

Claim 11: verdict (split, now three parts)

1. Closed cycle with total work below dG (net gain from one reservoir): Refuted by calculation, unchanged (dH/dG = 1.205 ceiling; a saving requires matching real power from the tank).
2. Heat-pump part (COP up to 1.205 by drawing T dS from the room): Established, not novel, unchanged.
3. New third part, changed in wording: field-assisted electrolysis as an overpotential (rate) effect. Before: not separately rated in my review. Now: Established as a measured effect at a catalyst for induction heating (Niether 2018: 100-200 mV, 5 C bulk rise), with field-power accounting not shown in the abstract (Plausible, untested as a net economic gain); for Kassner's standing field at water's band: Never properly tested, with the Q = 1 loading and heating results above making a lightly loaded, non-thermal benefit unlikely. I changed it because there is now a published, peer-reviewed measurement of the rate effect, which my review had no source for.

Claim 10: verdict

Strong form (a Kassner-band field or photon changes covalent or ionic bond energy): Refuted by calculation, unchanged, now with the 0.35 V/Angstrom simulation threshold (C4) and the 9.3 J/mol photon (C3). Added clarification: resonant bond-selective weakening of O-H is Established for infrared photons on surfaces (Hundt 2014), so the weak/different form "a resonant field of the right type can weaken a water bond" is true at the vibrational band and false at Kassner's.

Corrections to earlier documents

1. My own Claim 11 section (2026-10-03) lists only two forms of the claim. It omitted the rate-effect form that the arbiter rates and that the Investigator has now sourced; added above (C6, third part). No earlier verdict I gave was wrong; it was incomplete.
2. My A1 (2026-10-04) said I could not tell whether Kassner's two water values had an explanation. A7 gives two candidate explanations (vapour line, or liquid at a higher temperature); both unchecked.
3. The physics background above gives Q of order 1 for Debye water as a general statement; A6 now has the computed values (1.14 at 19.2 GHz, 0.97 at 23.4 GHz).
4. Web check status: the Investigator's items I rechecked are Niether 2018 (abstract) and Saitta 2012 (abstract); I did not open Hundt 2014, Serra 2020, Roy/Rao/Kanzius, the Wien-effect papers, the ACS review or the NASA reports, and rely on the Investigator's quotations for them. Still [memory, medium]: serum dispersion, bulk-field strengths of laboratory plate oscillators, the 0.4-0.8 V overpotential range, and tau(T) of water in A7.
5. (Added 2026-10-08, QE remediation by the arbiter.) C5 and verdict change 2 said the infrared stretch photon is "about 35,000 times" a 1.28 cm photon. That was wrong: it was taken from sites.md lead 34 without recomputing. 0.33-0.46 eV over 9.69e-5 eV is 3,407-4,749, about 3,400-4,700 [calc]. The figure near 50,000 is the O-H bond energy (4.80 eV) over the same photon [calc: 49,500]. Both places are corrected; the point (the established resonance is thousands of times above Kassner's band) is unchanged.

What would settle things

• Claim 11 (decisive): total work per mole of H2 at constant temperature, DC electrical energy plus real power delivered into the resonator (drive minus reflected, or the tank's loss with and without splitting), divided by moles collected, under isothermal calorimetry. Hypothesis: below 237.1 kJ/mol; thermodynamics: at least that. Separate allowed form: does the standing field at about 1.3 to 1.6 cm lower overpotential by more than the field costs, against a heat-matched off-band control?
• Claim 1: swept permittivity and loss of water, 0.1 M NaCl and serum from 0.2 to 25 GHz with controlled temperature, checking for any feature narrower than the Debye width and testing whether 1.28 cm corresponds to a temperature or to a vapour-line source.
• Read the full text of Niether 2018 (field frequency, power, energy balance) and Roy et al. 2008 / Rao et al. 2010 (gas analysis, controls).

Verdict changes (this update)

1. Claim 11: split extended to three parts. Closed cycle: Refuted by calculation (unchanged); heat-pump part: Established (unchanged); new part, field-assisted electrolysis as a rate effect: Established at a catalyst (induction heating, field power not shown to be repaid), Never properly tested for Kassner's standing field at water's band.
2. Claim 10: strong form still Refuted by calculation; clarification that infrared bond-selective weakening of water (Hundt 2014) is real, at 3,400-4,700 times the photon energy of Kassner's band (first written as 35,000 times; corrected 2026-10-08, see Corrections item 5).
3. Claims 1-9, 12, 13: unchanged; Claim 1 reinforced (Q about 1 for water at its band; Kassner's own "absorption greatest at resonance" supports the heating reading), Claim 3 note added (resonance is where heating is largest by his own text).