Physicist review: Pappas-PT (US 5,556,418, "Pulsed Magnetic Induction", the PAP-IMI device)

Reviewer role: PhD physicist (Step 2). Inputs read: facts.json, claims.md, deep/US5556418-dd083a.md, the extracted patent text (_text/US5556418-dd083a.txt, pp.4-6 read directly), sites.md "Leads for review", and the research notes (PubMed/PMC, Quackwatch/Wikipedia). Source tags: [read] from the project documents; [web] from the Investigator's notes (URL given there; I did not re-fetch them); [memory] my own recall with confidence; [calc] my own arithmetic, shown. Web verification: I made no new web calls in this pass, so every [memory] fact below is flagged with its confidence, and none of the verdicts rests on a medium/low-confidence [memory] fact alone; the load-bearing steps are arithmetic from numbers printed in the patent.

Reader note: no source gap. The patent text is complete (7 pages, OCR, with page images checked by the deep reader). The patent has no biological or clinical data; the only measurements in it are two oscilloscope traces (FIGS. 2-3).

---

0. Background the reader needs (plain language)

• Faraday induction. A changing magnetic field B through a loop induces an electric field E around it: the integral of E around the loop equals minus dPhi/dt, where Phi is the magnetic flux. For a uniform B changing at rate dB/dt, inside a circular region of radius r the induced field is E = (r/2)(dB/dt) (Griffiths ch. 7). This is how a coil "sends" an electric field through air into tissue without any spark across the air. [memory] high.
• Dielectric strength of air. Air breaks down (a spark forms) at about 3 x 10^6 V/m (3 kV/mm) between flat plates at atmospheric pressure. [memory] high. The patent says 10 kV/cm = 10^6 V/m, which is a factor of about 3 low.
• Negative differential resistance. Over part of its range an arc's voltage falls as its current rises (dV/dI < 0). This lets an arc plus an LC circuit oscillate (the "singing arc" of Duddell, 1900). [memory] high. Energy for the oscillation comes from the supply; the arc does not create it.
• Plasma frequency. Free electrons in a plasma oscillate collectively at f_p = (1/2pi) sqrt(n e^2 / (eps0 m_e)), about 8.98 sqrt(n) Hz with n in m^-3. It depends on electron density n, not on which element the gas is. [memory] high.
• Transmembrane potential. A cell membrane about 5 nm thick holds a potential of about -70 mV (inside negative), so the field inside the membrane is about 10^7 V/m. [memory] high.

1. What the Investigator found, kept in the record

These stay in the review whatever the verdict (SOURCE HONESTY). Tags are the Investigator's [web] sources, from sites.md and research/.

1. Patent family: priority Greek GR920100309 (10 Jul 1992), WO 94/01176, EP 0648140 B1 (granted 2001), RU 2126278 C1 (24 claims), and others. US patent maintenance fees paid three times; expired 2013. The RU abstract states the purpose as therapeutic curing of malignant tumours and says the plasma generates frequencies "resonant to treated object". [web]
2. Earlier Pappas patents on spark-gap/LC oscillators (EP 0402552 lightning conductor, 1989) and a 1989-1993 co-invented ion-transport patent (US 5,068,039; US 5,239,442; co-inventors Cure, Pappas, Eichler) that says it is "ideal in applying Nordenstrom's method". The spark-gap/LC design predates the medical use. [web]
3. Later NMR/EPR family (US 7,151,372; priority 2003): the "selective resonant absorption" idea is reframed as nuclear and electron magnetic resonance from a pulsed damped field with no static field. [web]
4. Inventor's later statements: PAP-IMI "triggers biological transmutations" and "PAP IMI also produces 20% excess heat on the probe alone" (New Energy News, June 1998, p.28); instantaneous power "> 1 Giga Watts", rise time "some hundreds of nanoseconds", "888 magnetic nanopulses at about 80 Joules each" (2006 "Physiology of the Cell" talk); cell as an "ATP controlled microscopic Nuclear Reactor" (Na + O -> K). These are the inventor's own claims. [web]
5. Physics publications by Pappas on the Ampere force and "energy creation in electrical sparks" (Nuovo Cimento B 1983; Phys. Lett. A 1985; IECEC 1991 vol. 4 pp. 416-423). The IECEC 1991 paper, which the patent cites as the source of the 10-100x gain, was not obtained. [web]
6. Manufacturer-sourced waveform numbers quoted in independent trial papers (Viti 2023, Keilani 2025): capacitor discharged through a spark gap into a spiral coil, up to 40 kV, peak currents of several 10 kA, about 60 J (Viti) or 96 J (Keilani) per pulse, 50-150 mT, base frequency about 240 kHz, pulse about 50 microseconds, 1-8 Hz; "in the maxima and minima of the damped oscillation, high-frequency oscillation peaks in the megahertz to gigahertz range arise". The authors state they have no conflict of interest and no external funding (Viti 2023). The figures are the manufacturer's, not the authors' own measurements. [web]
7. Trials: Leisner et al. 2002 (Hebrew University; rat ulnar fracture; untreated controls, no sham): treated group healed more slowly with more fibrous callus; "does not support the claim that PEMF generated by the PAP-IMI stimulate osteogenesis". Viti 2023 (n = 32, single-blind, sham-controlled, one session): heart-rate-variability indices rose after PAPIMI, not after sham. Viti 2025 (n = 30, single-arm, no control): pain score fell by a median of 2 points. Keilani 2025 (Medical University of Vienna, n = 20, single-blind placebo-controlled, 10 sessions, post-COVID fatigue): improved 6-minute walk and fatigue scales. None measures transmembrane potential, nitrogen resonance, gain, or cancer outcomes. [web]
8. Regulatory and historical record: FDA 510(k) submitted 6 Nov 1995, not cleared for lack of clinical data; FDA inspection report 2002 (Bio-Energy Services; units imported as "plant and seed germinator"; the inspector wrote "the science behind the device is not clear"; unauthorised study with 37 sites); FDA 2005 letter calling it a significant-risk device and reporting, from the regulator's file, two rat deaths, two cases of tachycardia and one fatal myocardial infarction during treatment; FDA Import Alert 89-08; California court order of 2005 against the distributor over marketing claims; Seattle Times investigation of 19 Nov 2007 (quotes Pappas: "can cure cancer and AIDS"). These are regulators' and a newspaper's characterisations; the distributor's own site and Pappas disagreed. The device record did not go dark: it continues to be sold from Greece with CE marking (Pappas: CE 0120 via SGS UK) and is used in published European trials. [web]
9. Pappas died on 11 May 2026 according to one Greek source (single source; the memoriam of 11-12 June 2026 does not conflict). His children, both physicists, are said to continue the work. [web]

---

2. Claim-by-claim review

Claim 1. An oscillating arc gives transient voltage and current 10-100x above the supply, because arc resistance goes infinity -> under 0.5 ohm -> about -100 ohm ("infinite" then negative gain).

a. Is the mathematics consistent? No. The "gain" is defined as a ratio of resistances: "a gain of infinity/0.5, which equals plus infinity, and ... a subsequent negative gain of 0.5/-100 < 0" [read: p.5, page image]. A ratio of two resistances is a dimensionless number but is not a voltage gain, current gain or power gain of anything; no circuit equation connects it to an output. Infinity divided by 0.5 ohm is also the statement that an open circuit is unlike a closed one, not that an amplifier has infinite gain. A negative "gain" of -0.005 is a nonsense quantity for the claim made (excess power). The patent's own text also says: "The energy increase of these electrons or ions is achieved by a corresponding decrease in the energy of the remaining electrons or ions" [read: p.4]. That sentence is energy conservation and contradicts the "excess power" reading. Inconsistent also: "excess ... effective power is produced" [read: p.5] versus the supply. No energy accounting is given.

b. Consistent with experiment? The observation that an arc circuit can show transient voltages and currents well above the supply is ordinary and well established: capacitor discharge gives peak power far above average power, and an inductor interrupted by a spark gap gives V = L dI/dt spikes (ignition coil, Tesla coil, Marx generator). [memory] high. Negative differential resistance of arcs (Ayrton, Duddell) is also established. [memory] high. None of this is net energy gain. Established results that contradict a real gain: conservation of energy; no verified over-unity arc oscillator exists. [memory] high.

Worked check [calc] using the manufacturer figures reported in the trial papers (40 kV, about 60 J, base frequency 240 kHz): the stored energy 60 J = (1/2) C V^2 gives C = 75 nF; resonance at 240 kHz gives L = 5.9 microhenry; characteristic impedance Z0 = sqrt(L/C) = 8.8 ohm; the peak current when 40 kV discharges into Z0 is about 4.5 kA, the same order as the "several 10 kA" quoted. A circuit with Z0 = 8.8 ohm and a total resistance of 0.5 ohm has Q = Z0/R = 18, which is in the "10-100x" range, so voltage across the coil or capacitor of a driven or ringing circuit can exceed the source several-fold (resonant magnification). So the "10-100x" observation is explained by resonance in an LC circuit with low resistance, which stores energy and gives it back, with no gain. A 1 GW instantaneous power (the inventor's later figure) lasting 100 ns is 100 J, the same order as the stored energy: peak power is stored energy divided by a short time. Average power from the mains remains within the input rating (the trial papers give 2500 VA for the device). Even the inventor's 20% "excess heat on the probe" (1998 letter) has no calorimetric protocol in anything I could read; it is unverified.

c. Tested? The patent cites the inventor's own conference papers (1989 Einsiedeln "free energy" conference; 1990 Tesla Symposium; 1991 IECEC Boston) for the gain; I could not read them. No independent replication of gain is on record in the Investigator's findings. The trial papers repeat only the waveform. The patent grant and 1991 IECEC publication are not tests of an energy balance, which would need input and output energy measured over a whole cycle.

d. If true: Conservation of energy (first law) would fail: a circuit would deliver more energy than it takes from the supply. The claim cannot be saved by saying only part of the particles gain energy, since the patent says the rest lose it.

Verdict: Refuted by calculation (for the "gain" reading); the weak form (transient voltages and currents well above supply because of LC resonance and stored energy) is Established but is not a gain. The patent's resistance-ratio argument is also Internally inconsistent as written.

Claim 2. The plasma emits sub-microsecond bursts (10-15 per train, duty cycle about 0.001%) at its characteristic radio eigenfrequencies (tens to hundreds of MHz), delivered through an induction coil.

a. Consistent mathematics? Partly. The figures show a slow damped sinusoid with fast ringing near the extrema; frequencies on the readout include 110.033 MHz average and 240.309 MHz maximum [read: p.3]. Internal inconsistencies: the burst duration is "less than 10^-6 s" in the description, "about 10^-4 to 10^-6 s" in claim 14 (100 to 1 microsecond), and "less than 10^-4 s" in claim 16 [read: deep read, page image p.7]. So "sub-microsecond" is not what the claims say; the Step 1 summary used the description's number only. The duty cycle 0.001% = 10^-5; with 1-8 Hz repetition (the trial papers) this means bursts of about 1-10 microseconds per second, consistent in order of magnitude. [calc]

The "eigenfrequency of the gas plasma" explanation fails a numeric check [calc]. The natural oscillation of a plasma is the plasma frequency f_p = 8.98 sqrt(n) Hz. For f_p = 100 MHz the electron density would be n = 1.2 x 10^14 m^-3; for 240 MHz, 7 x 10^14 m^-3. A real arc has n of order 10^22 to 10^24 m^-3 (the arc column is far denser), giving f_p of order 10^12 to 10^13 Hz. [memory] medium on arc density (arc plasmas are usually quoted 10^21 to 10^24 m^-3). So tens to hundreds of MHz are about seven orders of magnitude of density off from a plasma-frequency oscillation of the arc, and f_p depends on electron density, not on the element. A far simpler source for MHz ringing at the extrema is the stray inductance and capacitance of leads, electrodes and the probe, excited when the arc suddenly changes resistance (the arc igniting is a step in the circuit). The scope trace cannot distinguish them because the patent states no gas-dependence data. The patent's claim that frequency "depends on the gas elements" [read: p.5] has no measurement shown; none is described for nitrogen or argon versus air.

b. Established experiments: High-frequency ringing riding on the discharge of a spark-gap circuit is established (Hertz's spark experiments, Marx and Tesla circuits). [memory] high. The manufacturer's waveform quoted by the independent trial authors (Keilani 2025: "in the maxima and minima of the damped oscillation, high-frequency oscillation peaks in the megahertz to gigahertz range arise") is consistent with the patent FIG. 2-3. [web]

c. Tested? The waveform is observed by the inventor; the trial authors repeat the manufacturer's description and do not report an independent measurement of their own. I know of no independent oscilloscope or antenna measurement of the bursts in the Investigator's findings. A neutral measurement of burst content (probe calibration, bandwidth) is missing; note that a scope with a few hundred MHz of bandwidth and ordinary probes ringing on a 40 kV pulse can create or distort such features, so the displayed numbers (e.g. 1.31703 MHz minimum) need a calibration check.

d. If true: Nothing in mainstream physics changes for the existence of bursts; what would have to change is the explanation (a "plasma eigenfrequency" in the hundreds of MHz would need a density 10^8 times lower than an arc's).

Verdict: Plausible, untested for the bursts as a physical effect (high-frequency ringing at the extrema is ordinary); the specific explanation as plasma eigenfrequencies characteristic of the gas is Refuted by calculation (plasma-frequency arithmetic above), and the stated durations are Internally inconsistent across the description and claims 14 and 16.

Claim 3. Because the bursts are shorter than the time air needs to break down, their fields exceed the dielectric strength of air by 1-100x without ionising it, giving non-invasive cell-scale fields of order 10^6 V/m.

a. Consistent mathematics? No, on several points. 1. The patent's key numbers are wrong. It writes the membrane field as "70 mv / several Angstrom ... of the order at least of 10^6 v/m" and the air strength as "10 Kv/cm, or just 10^6 v/m", and concludes that air breaks down "before the membrane of a cell breaks down" [read: p.4, page image confirmed by the deep reader; not an OCR error]. [calc] 70 mV over 5 nm (a real membrane) is 1.4 x 10^7 V/m; over "several Angstrom" it is 7 x 10^7 to 2 x 10^8 V/m. So the membrane value is 10 to 100 times above the air figure, not equal to it, and the correct air figure is about 3 x 10^6 V/m. The argument "air breaks down first" is then about right in direction (the cell field exceeds that of air) but the stated numbers do not show it, and Step 1's claims.md repeats "order 10^6 V/m" from the patent; that is a correction needed (see "Corrections to Step 1"). 2. The claimed boost is "voltage" (1 to 100 times the dielectric value of air), but a voltage is compared to a field; the field depends on the gap. The patent never states the gap or a field in the coil region. 3. The timing argument says the rise "is comparable to the time intervals of molecular vibrations" [read: p.5]. Molecular vibrations take about 10^-13 to 10^-14 s; the bursts are 10^-6 s or longer, so a million times longer. Air breakdown in a gap needs an avalanche formation time of nanoseconds to microseconds, so sub-microsecond overvoltage is real, but not by a factor of 100 over large distances. [memory] medium.

The decisive point. The problem the patent sets itself does not exist for magnetic induction. A changing magnetic field induces an electric field directly inside the tissue (Faraday), through air, with no electric field needed in the air at all. [calc] with the figures the trial papers give (50-150 mT, 240 kHz): maximum dB/dt = 2 pi f B = 2.3 x 10^5 T/s for B = 0.15 T; at radius 9 cm (an 18 cm coil) E = (r/2) dB/dt = 1.0 x 10^4 V/m; at 3 cm it is 3.4 x 10^3 V/m. A cell of radius 10 micrometres in a field E acquires a transmembrane change of about 1.5 E a (Schwan's quasi-static result; [memory] high), here about 0.15 V, the same size as the resting potential. This is not at all "10^6 V/m" and needs no air breakdown, and it is in the range where electric effects on cells (and nerve stimulation: MRI scanners limit dB/dt to a few tens of T/s to avoid nerve stimulation [memory] high; this device is about 10^4 times higher) are expected. Transcranial magnetic stimulation establishes this in routine use. [memory] high. So the main conclusion is that the pulses can drive membrane-scale effects, but by ordinary Faraday induction at the 240 kHz base frequency and 50-150 mT, not through the patent's mechanism of beating air breakdown.

b. Established experiments: Magnetic stimulation of nerves and muscle (Barker 1985; TMS; peripheral magnetic stimulation) shows induced fields of 10 to 100 V/m trigger neurons, at far lower dB/dt. [memory] high. Overvoltage of short gaps with nanosecond pulses above the static breakdown value (Raether, Meek streamer theory) is also established but gives factors of a few. [memory] medium.

c. Tested? Field strength in tissue from the device was not measured in the patent. The only on-record energy figure from a third party is the FDA 2005 letter's "100 watts at 2 apps (2 seconds x 50 Joules)" per two pulses (a regulator, an interested party) and the trial papers' 60 or 96 J per pulse (manufacturer data). The physiological responses in the small sham/placebo-controlled trials (Viti 2023 HRV; Keilani 2025 fatigue) are consistent with a real biological response to strong pulsed magnetic fields but do not test the patent's mechanism.

d. If true (the 100x overvoltage reading): Nothing in mainstream physics changes; short-pulse breakdown delay is known. The actual stated claim, that fields of 10^6 V/m act on cells at a distance without breakdown, is replaced in mainstream physics by induction fields in tissue.

Verdict: split. (i) The premise "10^6 V/m cannot be applied non-invasively without breaking down the air" is Refuted by calculation (induction gives fields in tissue with no air field; and the patent's own numbers for membrane and air fields are Internally inconsistent: 10^6 vs correct 10^7-10^8 and 3 x 10^6). (ii) The weak form "short strong magnetic pulses can induce membrane-scale electric effects in tissue through air" is Established (Faraday, magnetic stimulation) and consistent with the device's published parameters, with the caveat that the numbers come from the manufacturer. (iii) The "1 to 100 times air breakdown by sub-microsecond timing" is Never properly tested.

Claim 4. Cell constituents resonate with the plasma eigenfrequencies and selectively absorb energy, starting selected endothermic reactions (e.g. nitrogen plasma selectively exciting nitrogen in tissue).

a. Consistent mathematics? No. The patent claims that radio frequencies tied to elements in the plasma are "absorbed through resonance by similar elements existing in biological matter", and that "nitrogen atoms existing in the form of chemical compounds ... will selectively absorb this energy due to resonance" [read: p.5, quoted]. The reaction scheme XAB + Y + plasma energy A = XA + YB (reaction 1) and XB + YA (reaction 2) is bookkeeping, not mechanism: no frequency, transition or energy is given [read: deep read §7]. Specific problems [calc]: - The resonances of an atom's electrons (the frequencies of atomic lines) lie in the optical or ultraviolet region (about 1 to 10 eV), not at 10^8 Hz. A 240 MHz photon carries h f = 1.0 x 10^-6 eV, against thermal energy kT = 0.027 eV at body temperature and chemical bond or activation energies of about 0.5 to 5 eV. One radio photon is a million times too small for one reaction, and the absorbed field energy is thermalised within picoseconds in liquid water and tissue. - A bound nitrogen atom in a protein does not have a free "nitrogen frequency"; its vibrational modes (bond stretch, bend) are in the infrared, tens of THz, and depend on the whole molecule. - The only radio-frequency resonance of nitrogen is nuclear magnetic resonance of 14N (Larmor frequency about 3.08 MHz per tesla; [memory] high), which is about 0.3 MHz at 0.1 T, a number that happens to be near the device's 240 kHz base frequency. This is what the inventor's later NMR/EPR patent family (US 7,151,372) builds on. But NMR needs a static field and perpendicular RF field, the nuclear spin population is polarised by only about one part in 10^7 to 10^8 at 0.1 T and body temperature [memory] medium, and the energy of a spin flip (about 10^-9 eV) is further below chemistry than the radio photon above. NMR resonance does not drive endothermic reactions; it is the basis of MRI because it is so weak and non-destructive. - The selectivity claim ("selectively ... at will") needs narrow resonance lines; the gas is a plasma of arc temperature and the pulse is a broadband transient (a burst under 1 microsecond is broader than 1 MHz in spectrum), so there is no narrow line to match. - The electron-resonance alternative (EPR) is at 28 GHz per tesla ([memory] high), about 2.8 GHz at 0.1 T, within the "megahertz to gigahertz" range of the bursts; free-radical EPR is real but again produces no endothermic chemistry.

b. Established experiments: Resonant energy absorption by tissue at radio frequencies is known as heating and, at MRI frequencies, NMR; no experiment shows element-selective chemistry from radio pulses in tissue. Liboff's ion cyclotron resonance hypothesis (US 5,160,591, cited in the patent) is the nearest mainstream-adjacent idea; it was proposed for ions in the static geomagnetic field with a weak AC field, and replication of it has been mixed. [memory] medium (this is the weakest recollection in this review; the verdict does not rest on it). The inventor's later claims of biological transmutation (C to Fe; K-Na-K; 1998 letter; Moawad 2013, which cites Pappas) conflict with nuclear physics: nuclear reactions at keV or lower energies in tissue face the Coulomb barrier of MeV scale, and no neutron, gamma or product signature accompanies the claims. [memory] high. Those claims belong to the Investigator's record and are not part of the 1996 patent.

c. Tested? No test is on record, by anyone, of selective excitation of tissue nitrogen or of a chosen reaction. The trial papers use the device with air and do not vary the gas. The patent itself reports nothing.

d. If true: Chemistry and spectroscopy would have to change: broadband radio pulses (photon energy 10^-6 eV) would have to drive 1 eV reactions selectively in liquid water at 310 K; coherence times in tissue are picoseconds, far shorter than the bursts. A new coupling mechanism for radio energy into reaction coordinates would need to exist.

Verdict: Refuted by calculation in the form stated (radio-frequency resonance of elements for selective endothermic chemistry), by the photon-energy, thermalisation and spectral arguments above. For the inventor's later transmutation extensions: Never properly tested (the 1998 letter's claims are unsupported by data in the record) and in conflict with nuclear physics.

Claim 5. Restoring transmembrane potential (cancer about 15 mV, unhealthy about 50 mV, healthy about 70 mV) is therapeutic, and the pulses can "pasteurize" degenerate cells or microorganisms by pressure increase.

a. Consistent mathematics? The numbers are not self-consistent with claim 3 (see above), but the potentials are in a plausible range. The pasteurisation statement ("bursts ... pressure inside the degenerate cell ... is increased. This pressure can burst or kill" [read: p.4]) gives no pressure, field or time. The patent's own explanation of how a voltage in an external "magnetic pulse" raises internal pressure is not derived.

b. Consistent with experiment? - Potentials: depolarised resting potential in tumour cells (much less negative than normal cells) is reported in the literature (Binggeli and Cameron; Cone, 1970s). [memory] medium; the exact figures 70/50/15 mV and "source of pain generation signals" are given in the patent without citations [read: p.4]. That low-potential cells cause pain is not established. [memory] medium. - Killing cells with pulsed electric fields is established: irreversible electroporation and pulsed-electric-field food pasteurisation need fields across the cell of about 10^5 to 4 x 10^6 V/m (1 to 40 kV/cm) for microseconds, delivered via electrodes. [memory] high. The induced field from this device [calc, claim 3] is about 10^3 to 10^4 V/m in tissue, which is 10 to 100 times too small for membrane breakdown in that way, so pasteurisation is not supported by the numbers in the trial papers. A different, possible route is repetitive cardiac and neural stimulation, which is a risk, not a therapy. - Nordenstrom: the patent says his electrode method is "quite successful" [read: p.4]; his work (electrochemical treatment, 1980s) was studied in China and elsewhere with mixed and low-quality evidence. [memory] medium. I did not check this. - The physiological effect of the device is not zero: Viti 2023 (sham-controlled, n = 32) found HRV indices rose in the treated group only; Viti 2025 (single-arm, n = 30) found a median pain score fall of 2 points with no control; Keilani 2025 (randomised placebo-controlled, n = 20) found improvement in the 6-minute walk and fatigue scales. Those are small pilot studies; pain and fatigue scores are subjective and prone to placebo effect, and one of the three has no control. Leisner et al. 2002 (rat fracture, untreated controls) found slower healing in the treated group. The two sets of results go in opposite directions on different outcomes; none tests transmembrane potential, cancer or microorganism killing. - Cancer claims: no controlled study is on record. Pappas is quoted in the Seattle Times (2007) as saying the machine "can cure cancer and AIDS"; the FDA inspected a study with an infant with rhabdomyosarcoma treated about 29 times (FDA inspection report, 2002); the FDA 2005 letter reports adverse events including a fatal myocardial infarction during treatment (regulator's characterisation, uninvestigated by me). The FDA's refusal in 1995 was for want of clinical data, not a finding of harm.

c. Tested? Cancer: not in any controlled trial on record. Transmembrane potential: not measured by anyone in the record. Microorganisms: not reported. Pain/fatigue/HRV: small pilot trials noted above (positive sham/placebo-controlled in two, negative animal test in one). Replication by independent teams of the same outcome is not yet on record (Viti 2023 and 2025 are one group).

d. If true: For restoring Vm: an external pulsed field would have to produce a lasting change in membrane potential, which the cell's ion pumps would remove within seconds to minutes (the cell actively restores its own potential using metabolic energy). Mainstream physiology would need a mechanism for persistent repolarisation of tumour cells by external pulses. Cell killing "by pressure" would need a new mechanism beyond electroporation.

Verdict: split. (i) "Pulsed magnetic field sessions from this device have measurable physiological or symptomatic effects": Plausible, untested in the sense of mechanism, with small pilot trials on record (two sham/placebo controlled, positive; one animal trial negative on bone healing); the mechanism is probably ordinary induced currents (Faraday), as in other PEMF devices. (ii) "Restoring transmembrane potential cures cancer / low-potential cells cause pain": Never properly tested (no controlled test on record; it was not dismissed after testing, but the regulator did refuse and later ban on safety and evidence grounds, and no sponsor has run a controlled cancer trial). (iii) "Pasteurisation by pressure at these fields": Refuted by calculation against the induced field of about 10^4 V/m versus the 10^5 to 10^6 V/m needed for membrane breakdown (electroporation), unless fields at the coil surface or in the sub-microsecond bursts are much higher than the 240 kHz base figures; nobody has measured these, so the strong form is Never properly tested.

---

3. Summary table

4. Fair statement of the mainstream view

A spark-gap capacitor discharge into a coil makes a strong, rapidly changing magnetic field with a damped oscillation near 240 kHz, plus fast ringing. That field induces eddy currents and electric fields in tissue (Faraday), which are strong enough to stimulate nerves and muscles; that is how magnetic stimulators and many pulsed electromagnetic field (PEMF) devices work, and it can produce real physiological effects (autonomic, pain, perhaps others). Mainstream physics rejects: (1) net energy gain, (2) element-selective resonant chemistry at radio frequencies, and (3) the inference that the field must overcome air breakdown. The patent's practical device may do something in tissue, but for reasons not in the patent. Absence of clinical data in the patent is not itself evidence against effects; the 1995 FDA refusal, the 2002-2007 regulatory record and the trial literature are the data that exist. Patent grant and CE marking are not evidence of efficacy.

5. Open points that would settle things

• Claim 1: measure input energy from the mains and output energy (calorimetry) over many pulses in the device, plus the discharge circuit with a calibrated high-voltage probe; compare with 1/2 C V^2. The Pappas 1991 IECEC paper should be obtained and checked.
• Claim 2 and 3: map the actual E and B at the coil with calibrated field probes (not a general-purpose scope probe), then compute the induced field in a tissue phantom; check whether the MHz-GHz ringing changes with gas species or is fixed by the lead and electrode geometry.
• Claim 4: a gas-substitution test (air versus nitrogen versus argon) in a chemical or cell assay; no such test exists.
• Claim 5: a larger sham-controlled multi-centre trial by a group independent of the manufacturer; a safety review of the cardiac events cited by the FDA.

---

Corrections to Step 1

1. claims.md claim 3 repeats the patent's "cell-scale fields of order 10^6 V/m". The deep read already showed that the patent's own numbers (10^6 for membrane and for air) are inconsistent; the membrane field is about 10^7 to 10^8 V/m and air breakdown about 3 x 10^6 V/m. The claim should say "the patent states 10^6 V/m" and not give it as the true value. [calc above, deep read §1]
2. claims.md claim 2 says "sub-microsecond" bursts. The description says less than 10^-6 s, but claims 14 and 16 say 10^-4 to 10^-6 s and less than 10^-4 s [read: deep read §5, page image p.7]. Not uniformly sub-microsecond.
3. facts.json "mainstream_conflict" says the energy "stored capacitor energy ... do not create net power" (fine) but also lists "resonant absorption ... at tens to hundreds of MHz is not an accepted mechanism" without the specific reason. The reasons are now computed above (photon energy 10^-6 eV, plasma-frequency density mismatch).
4. facts.json states "Greek priority application 920200309". The Investigator found Espacenet lists the Greek application as GR920100309 (different number); the US patent text prints 920200309 [read: p.4]. The US number may be a misprint in the patent; not verified by me.
5. claims.md says the evidence is only the patent. Correct for the project folder, but the Investigator's findings (section 1 above) add post-1996 data: trials, FDA records, the later NMR/EPR patents, and the inventor's later statements. These belong in the claims 1, 4 and 5 records.
6. Step 1 summary says "30 cm induction coil". I did not re-verify this number; the deep read reports claim 6 (coil diameter "at least a fraction of one centimeter") conflicts with a 30 cm preferred coil and claim 5's "diameter of 6 square millimeters" is a unit error of the patent. These are patent errors, not Step 1 errors; flagged so later steps do not repeat them as facts.
7. No errors found in Step 1's reading of the 17 claims, the resistance figures (-100 ohm, 0.5 ohm; confirmed on page image) or the bibliographic data.