Pappas and the PAP-IMI: sparks, coils and cell membranes. A primer

Project: Pappas-PT. Panagiotis T. Pappas (Greek physicist, died 2026 according to one Greek source), US Patent 5,556,418, "Method and Apparatus for Pulsed Magnetic Induction" (filed 1993, granted 17 Sep 1996). The device is sold as PAP-IMI or PAPIMI.

Written for a self-taught reader at undergraduate level. Verdicts come from arbiter.md; this primer adds no new verdicts. Source tags: [read] = from the patent (the only document in the project folder); [web] = from the Investigator's and reviewers' web findings, with URL; [memory] = textbook knowledge recalled by the model, with confidence; [calc] = arithmetic done in example.py.

Source note. The patent is complete and was read in full, including page images for the exponents and the oscillograms [read]. Two important documents were not obtained by anyone: the inventor's 1991 IECEC (Intersociety Energy Conversion Engineering Conference) paper ("Energy Creation in Electrical Sparks and Discharges: Theory and Direct Experimental Evidence", vol. 4, pp. 416-423), which the patent cites for its 10-100x voltage and current boost, and his ICCF-7 (7th International Conference on Cold Fusion, 1998) paper. They are where any energy-balance data would be. The patent itself gives no capacitance, inductance, charging voltage, current, gas pressure, gap, or vertical scale on its oscillograms [read]. All device numbers used below (40 kV, 60-96 J per pulse, 240 kHz, 50-150 mT) are the manufacturer's figures as repeated in two recent clinical papers, not independent measurements: Viti 2023 gives 50-150 mT and about 60 J [web: https://pmc.ncbi.nlm.nih.gov/articles/PMC10001461/]; Keilani 2025 gives 50-100 mT and about 96 J [web: https://pmc.ncbi.nlm.nih.gov/articles/PMC12534222/]. Both papers write the pulse energy as "W (joules)" / "Watt (Joule)", mixing units; it is read here as joules per pulse. The only electrical figure in the patent itself is a coil current of "at least 1000 amps" (claim 5) [read: p.6-7]; section 5 uses it as a cross-check.

---

1. The idea in one paragraph

Pappas charges a capacitor to high voltage and lets it discharge through a spark gap (an arc in air or another gas) into a large flat coil held over the patient. The discharge rings: a slow, dying oscillation, with short bursts of very fast oscillation riding on each peak and trough. Pappas says three things happen. First, the arc acts like an amplifier with "infinite" and then "negative" gain, so the voltage and current briefly reach 10 to 100 times what the supply provides. Second, the pulses are so short that their fields can exceed the strength at which air breaks down without making a spark, so the strong fields needed to influence a cell membrane can be applied without touching the body. Third, the fast bursts carry "eigenfrequencies" (natural frequencies) of the gas in the arc, and atoms of the same element in the body absorb them by resonance; with nitrogen in the arc, nitrogen in tissue would be selectively energised and could drive chosen chemical reactions. He presents the treatment as restoring the electrical voltage across the membranes of sick cells (he quotes about 70 mV for healthy cells, 50 mV for aged ones and 15 mV for cancer cells) and as able to burst microorganisms ("cold pasteurization") [read: pp.4-6].

2. What you need to know first

Faraday's law of induction. A changing magnetic field makes an electric field that circulates around it. In integral form, the line integral of E around a closed loop equals minus the rate of change of the magnetic flux Phi through the loop. For a uniform field B filling a circle of radius r, this gives a circulating field E = (r/2)(dB/dt), zero on the axis and growing outward. The key point for this project: the electric field is created inside the tissue by the changing magnetic field. There is no need for a strong electric field in the air between coil and body. Read Griffiths, Introduction to Electrodynamics, chapter 7 [memory, high].

RLC circuits and capacitor discharge. A capacitor C charged to voltage V0 stores energy W = (1/2) C V0². Connected to an inductor L it rings at f0 = 1/(2 pi sqrt(LC)); resistance R makes the ringing die away. The characteristic impedance Z0 = sqrt(L/C) sets the peak current, I0 = V0/Z0, and the quality factor Q = Z0/R (for a series circuit) is roughly how many cycles the ringing survives (precisely, Q/pi cycles for the amplitude to fall by a factor e). Any intro E&M text covers this; Purcell and Morin, Electricity and Magnetism, has a clear treatment of alternating-current circuits.

Negative differential resistance and the singing arc. For an ordinary resistor, more current means more voltage (V = IR). For an arc over part of its range, more current means less voltage: the slope dV/dI is negative. Put such an element in parallel with an LC circuit and it can keep the circuit oscillating, drawing its energy from a DC supply. This is real and historically important: William Duddell's "singing arc" (1900) and Valdemar Poulsen's arc radio transmitters (1900s-1920s) worked this way [memory, high]. The supply pays for every joule of the oscillation.

Plasma oscillations. A plasma is a gas in which many electrons have been knocked off their atoms. If the electrons are displaced as a group, their attraction back to the positive ions makes them oscillate at the plasma frequency f_p = (1/2 pi) sqrt(n e² / (eps0 m_e)) ≈ 8.98 sqrt(n) Hz, where n is the number of free electrons per cubic metre, e the electron charge, m_e its mass and eps0 the permittivity of free space. Irving Langmuir described these oscillations in the 1920s; Pappas cites Langmuir 1925 [read: p.4]. Notice what f_p depends on: only the electron density n. Not on which gas the electrons came from. Read F. F. Chen, Introduction to Plasma Physics and Controlled Fusion, the chapter on waves in plasmas.

Dielectric breakdown of air. At atmospheric pressure, air between flat plates sparks over at about 3 x 10^6 V/m (30 kV/cm) [memory, high]. The spark is an avalanche: free electrons gain energy from the field, ionise molecules, and the number of electrons multiplies. The avalanche takes time to build up, so a very short pulse can briefly exceed the DC breakdown field. Measured nanosecond breakdown in air is about 75 kV/cm against 30 kV/cm for DC, a factor of about 2.5 [web: search-result summary of https://www.osti.gov/servlets/purl/1787665 (OSTI, the US Department of Energy's Office of Scientific and Technical Information); the page itself was not opened]. Read Raizer, Gas Discharge Physics.

The cell membrane as a capacitor. A living cell keeps its inside about 70 mV negative relative to the outside, across a membrane about 5 nm thick. That is a field of about 1.4 x 10^7 V/m inside the membrane [calc]. The membrane behaves like a capacitor of about 1 microfarad per square centimetre [memory, high]. An outside field E pushes charge onto the two poles of a spherical cell of radius a and adds up to ΔV = 1.5 E a to the membrane voltage (Schwan's relation, valid for fields that change slowly compared with the membrane's charging time) [web: https://pmc.ncbi.nlm.nih.gov/articles/PMC12045266/]. The external field needed is tiny compared with the field inside the membrane, because the cell gathers the voltage drop across its whole diameter and puts it across a membrane a thousand times thinner. If ΔV reaches roughly 0.2 to 1 V, pores open (electroporation); this is used in labs to put DNA into cells and in food processing to kill microbes (pulsed-electric-field pasteurisation) [web, same URL]. Read Philip Nelson, Biological Physics, the chapters on membranes and nerve impulses.

Photon energy. Electromagnetic energy at frequency f comes in quanta of h f, where h = 6.63 x 10^-34 J s is Planck's constant. In electron-volts (1 eV = 1.6 x 10^-19 J), h f at 100 MHz is 4 x 10^-7 eV. Chemistry works in steps of about 1 eV; thermal energy at body temperature (kT, with k Boltzmann's constant and T = 310 K) is 0.027 eV. Any intro modern physics text covers this.

Nuclear magnetic resonance (NMR), in one paragraph. Some atomic nuclei, including nitrogen-14, act like tiny bar magnets. In a steady magnetic field they precess at a frequency proportional to the field (for 14N about 3.08 MHz per tesla) and absorb radio waves at exactly that frequency [memory, high]. This is a real element-specific radio resonance, and it is the basis of MRI (magnetic resonance imaging). It needs a steady field to set the frequency, and each absorbed quantum is extremely small.

3. The textbook view

What a spark-gap capacitor discharge into a coil does, step by step:

1. Storage. The capacitor charges slowly from the mains, storing W = (1/2) C V0². Nothing in the circuit can deliver more than W per pulse (plus whatever the supply adds during the pulse, which for a slow charger is negligible).

2. Switching. When the gap voltage exceeds its breakdown value, the gap ionises within nanoseconds and its resistance falls from effectively infinite (open switch) to a fraction of an ohm (closed switch). The arc may show negative differential resistance along the way.

3. Ringing. The capacitor, the coil and the leads form an RLC circuit. Current I(t) ≈ I0 e^(-t/τ) sin(2 pi f0 t), with I0 = V0/Z0 and decay time τ = 2L/R. Voltage across the coil, L dI/dt, can be large; when the arc goes out, or the current changes sharply, L dI/dt makes voltage spikes (the same reason a car ignition coil makes 20 kV from 12 V).

4. Parasitic ringing. Every lead and electrode has some stray inductance (tens to hundreds of nanohenries) and stray capacitance (picofarads). Each sudden change in the arc excites these at high frequency: 100 nH with 25 pF rings at f = 1/(2 pi sqrt(LC)) ≈ 100 MHz [calc, arbiter]. That is why spark-gap circuits, from Hertz's 1887 experiments to Tesla coils and spark radio transmitters, are notorious radio-frequency noise sources [memory, high].

5. Induction in tissue. The coil's changing field B(t) induces E = (r/2)(dB/dt) in the tissue under it, which drives eddy currents (current density J = σE, with σ the tissue conductivity of order 0.1 to 1 S/m [memory, medium]). This is the mechanism of transcranial and peripheral magnetic stimulation, which excite nerves with induced fields of 10-100 V/m [memory, high].

6. Effect on a cell. The induced E adds ΔV ≈ 1.5 E a to the membrane voltage while the field lasts, reduced by a factor 1/sqrt(1 + (ωτ_m)²) if the field oscillates faster than the membrane charging time τ_m (ω = 2 pi f). The cell's ion pumps and channels then restore the resting potential within milliseconds to seconds; a lasting change needs a lasting cause [memory, high].

Key equations, collected:

4. Where the inventor departs

Pappas starts from real things: arcs do have negative differential resistance; spark discharges do show bursts of fast oscillation on top of the slow ringing (his FIG. 2 and 3 show exactly this, and later manufacturer descriptions say the same [read: p.3; web: PMC12534222]); short pulses do beat the DC breakdown of air; cancer cells are reported to have less negative resting potentials than normal cells (Cone; Binggeli and Cameron, 1970s-1980s) [memory, medium]; strong pulsed fields can kill cells (electroporation); and nitrogen does have a radio-frequency resonance (NMR). His earlier career also includes real peer-reviewed physics on the Ampère force law (Nuovo Cimento B 1983; Physics Letters A 1985; an experiment at the National Technical University of Athens published in J. Appl. Phys. 1986) [web, per sites.md: Google Scholar], and earlier patents on spark-gap LC oscillators for lightning conductors (EP 0402552, 1989) [web: Espacenet, per sites.md]. The spark-gap design came before the medical use.

Where the idea came from: related work the patent itself relies on. These are kept in the record as related work, not as corroboration of this device. - Nordenstrom's electrochemical therapy. Björn Nordenström's method treats tumours by passing current through needle electrodes inserted into the body. The patent calls it "quite successful" and builds its problem statement on it: how to get the same effect without electrodes [read: p.4]. The patent also says its own effect matches "microcontact" electrodes and can be applied over ordinary clothing [read: p.5]. Nordenström's own claims are not assessed in this review. - The direct precursor. Pappas's 1989 patent with Cure and Eichler, US 5,068,039 (with US 5,239,442), on moving ions by remote action, is described as "ideal in applying Nordenstrom's method" [web: https://patents.google.com/patent/US5068039A/en]. It sits between the lightning-conductor oscillator and the 1996 medical patent. - Liboff's ion cyclotron resonance (Liboff et al., US 5,160,591, "regulating transmembrane ion movement utilizing selective harmonic frequencies"), cited as prior art in the patent [read: p.1]. This is the nearest mainstream-adjacent relative of Pappas's resonance claim. An ion of charge q and mass m in a steady magnetic field B circles at the cyclotron frequency f = qB/(2 pi m), so it is ion-specific, much as Pappas wants his resonance to be element-specific. Like NMR, it needs a steady field to set the frequency, which the pulsed PAP-IMI does not supply. Its experimental replication record is contested according to the physicist [memory, medium; not web-checked; context only, no verdict rests on it]. - The other prior art cited: Czerlinski US 4,369,345 and Gordon US 4,622,953 [read: p.1].

The departures, side by side:

The exact point of departure is the first row. Everything the patent explains, it explains through the idea that the arc creates voltage, current and power. Once you accept that peak values above the supply are ordinary consequences of storing energy and releasing it quickly, the boost needs no gain, the 10^6 V/m target turns out to be the wrong target, and the induced field the device does make (from published figures) is already in the range that affects membranes and nerves.

Pappas kept extending the idea after 1996. These later claims are his own and are kept in the record: "PAP IMI also produces 20% excess heat on the probe alone", biological transmutation ("C to Fe") and energies "in the range of 30 Kev" (letter in New Energy News, June 1998, p.28) [web, file in research/files]; instantaneous power "> 1 Giga Watts", "888 magnetic nanopulses at about 80 Joules each", and the cell as an "(ATP) controlled microscopic Nuclear Reactor" (ATP, adenosine triphosphate, is the cell's energy-carrying molecule) (Na + O -> K) ("Pappas' Physiology of the Cell and its Atomic Energy", Athens 2006) [web: research/files/Pappas-Physiology-of-the-Cell-ralf-kollinger.pdf]; a later patent family (US 7,151,372, priority 2003) that recasts the resonance idea as NMR and EPR (electron paramagnetic resonance) without a static field [web: https://patents.google.com/patent/US20060220648A1/en]; and a 2018 Greek utility model titled "Earth unbalance device - free energy generation from zero" [web, per sites.md].

5. Worked example: one pulse, from the capacitor to the cell membrane

We follow one pulse through the device using the manufacturer's numbers, and see what it does to a cell. Every number below is reproduced by example.py (run command in its header) [calc].

Inputs [web, manufacturer figures via Viti 2023 and Keilani 2025]: charging voltage V0 = 40 kV; stored energy W = 60 J per pulse (Viti; Keilani gives 96 J); base ring frequency f0 = 240 kHz; field 50-150 mT (Viti; Keilani gives 50-100 mT); 1-8 pulses per second. From the patent [read: p.6]: coil "preferably" 30 cm diameter, 2-10 turns.

Step 1. Capacitance. W = (1/2) C V0², so C = 2W / V0² = 2 x 60 / (4 x 10^4)² = 120 / 1.6 x 10^9 = 7.5 x 10^-8 F = 75 nF.

Step 2. Inductance. f0 = 1 / (2 pi sqrt(LC)), so L = 1 / ((2 pi f0)² C). 2 pi f0 = 1.508 x 10^6 s^-1; squared, 2.274 x 10^12; times C, 1.706 x 10^5; so L = 5.86 µH.

Step 3. Peak current. Z0 = sqrt(L/C) = sqrt(5.86 x 10^-6 / 7.5 x 10^-8) = sqrt(78.2) = 8.84 ohm. I0 = V0 / Z0 = 40 000 / 8.84 = 4.5 kA (ignoring resistance, so this is an upper value). Check: at peak current all the energy is in the inductor: (1/2) L I0² = 0.5 x 5.86 x 10^-6 x (4524)² = 60 J. ✓ The manufacturer also quotes "peak currents of several 10 kA". Through the same L, 10 kA needs (1/2) L I² = 293 J and 30 kA needs 2.6 kJ, against 60-96 J stored. With 96 J the peak is 7.2 kA. So the quoted currents and energies do not fit together unless the circuit is different from this simple model (smaller inductance, for example); this is about the manufacturer's figures, not the patent [arbiter correction 3]. The only energy figure from outside the manufacturer is the FDA's 2005 letter: "capable of delivering 100 watts at 2 apps (2 seconds x 50 Joules) of energy for every two pulses" [web: research/files/seattletimes-2007-11-19-FDA-letter-papimi.pdf, per the physicist's reading]. That is the regulator's characterisation (a party with a regulatory stake) and about 50 J per pulse, the same order as 60 J, so it does not resolve the "several 10 kA" mismatch [arbiter section 9].

Step 4. Peak versus average power: no gain needed. Average power at 1 pulse/s is 60 J x 1 s^-1 = 60 W (480 W at 8 pulses/s). The peak instantaneous power in the ring is V0 I0 / 2 = 40 000 x 4524 / 2 = 90 MW, a million times the average. The patent's duty cycle of "the order of 0.001 percent" [read: p.6] alone makes peak power 1/10^-5 = 10^5 times the average, with energy in equal to energy out. A "gain" of this size is a statement about when the energy arrives, not how much there is. (For comparison, the inventor's later "> 1 Giga Watts" is about ten times this simple-model peak; nobody in the record measured it.)

Step 5. Coil field. For a flat coil of N = 5 turns, radius R = 0.15 m, carrying 4.5 kA: B = mu0 N I0 / (2R) = 4 pi x 10^-7 x 5 x 4524 / 0.30 = 95 mT, inside both quoted ranges (50-150 mT, Viti; 50-100 mT, Keilani). The published figures hang together at this level. What the patent itself specifies. The patent's only current is "at least 1000 amps" [read: p.6-7]. With the same 5-turn, 30 cm coil, 1 kA gives B = 4 pi x 10^-7 x 5 x 1000 / 0.30 = 21 mT, about 4.5 times below the 95 mT modelled from manufacturer figures [calc; the mathematician's figure]. We carry this lower value through Steps 6 and 7 as a cross-check.

Step 6. Induced electric field in tissue. B oscillates as B0 sin(2 pi f0 t), so its peak rate of change is dB/dt = 2 pi f0 B0. With B0 = 0.10 T: dB/dt = 1.508 x 10^6 x 0.10 = 1.5 x 10^5 T/s. At r = 10 cm from the coil axis, E = (r/2) dB/dt = 0.05 x 1.508 x 10^5 = 7.5 x 10^3 V/m. Over the published range (50 mT at 10 cm to 150 mT at 15 cm) E runs from 3.8 x 10^3 to 1.7 x 10^4 V/m, and it falls to zero on the coil axis. (Keilani's upper value, 100 mT at 15 cm, gives 1.1 x 10^4 V/m.) With the patent's own 1 kA (21 mT) at r = 15 cm: E = 0.075 x 1.508 x 10^6 x 0.021 = 2.4 x 10^3 V/m, at the low edge of the range [calc]. For scale: MRI safety limits keep dB/dt to some tens of T/s to avoid nerve stimulation, and magnetic nerve stimulators excite neurons with induced fields of 10-100 V/m [memory, high]. This device's numbers are thousands of times larger. Whatever else is true, a pulse like this is not physiologically inert.

Step 7. Voltage added across a cell membrane. Take a large mammalian cell, radius a = 10 µm, in E = 7.5 x 10^3 V/m. Steady-field limit: ΔV = 1.5 E a = 1.5 x 7540 x 10^-5 = 0.113 V = 113 mV, larger than the 70 mV resting potential. Is 240 kHz "slow" for the membrane? The membrane charging time is τ_m = a C_m (ρ_i + ρ_e/2), with C_m = 0.01 F/m² (1 µF/cm²), and cytoplasm and outside resistivities ρ_i, ρ_e taken as 1 ohm m each (illustrative values) [memory, medium]: τ_m = 10^-5 x 0.01 x 1.5 = 150 ns. Then ωτ_m = 2 pi x 240 000 x 1.5 x 10^-7 = 0.23 and the reduction factor 1/sqrt(1 + 0.23²) = 0.975, so ΔV ≈ 110 mV. At 100 MHz (the fast bursts), ωτ_m = 94 and the factor is 0.011, so ΔV ≈ 1.2 mV: the bursts the patent treats as the active ingredient barely reach the membrane at all. For a bacterium-sized cell (a = 1 µm): ΔV ≈ 11 mV at 240 kHz. With the patent's own 1 kA figure (E = 2.4 x 10^3 V/m): ΔV = 1.5 x 2370 x 10^-5 ≈ 36 mV for a = 10 µm in the steady limit, still about half a resting potential. So the conclusion below does not depend on trusting the manufacturer's higher numbers [arbiter section 9].

Step 8. Compare with the patent's targets. - The patent says a cell needs at least 10^6 V/m from outside. Step 7 shows that 7.5 x 10^3 V/m already moves the membrane of a large cell by more than its resting voltage. The patent's figure confuses the field inside the membrane (70 mV / 5 nm = 1.4 x 10^7 V/m; or 70 mV / 3 Å = 2.3 x 10^8 V/m using the patent's own "several Angstrom") with the field needed outside the cell, and then sets that field equal to air's breakdown, which it also gives as 10^6 V/m (the textbook DC value is 3 x 10^6 V/m). - Pasteurisation. The external field needed for ΔV = 1 V (a common electroporation threshold) is E = ΔV / (1.5 a): 6.7 x 10^4 V/m for a = 10 µm, 6.7 x 10^5 V/m for a = 1 µm. For the reversible lower end (0.2 V): 1.3 x 10^4 and 1.3 x 10^5 V/m. The device's 4 x 10^3 to 2 x 10^4 V/m is 10-100 times short for bacteria, and touches the lower (reversible) end only for large cells in the coil's outer region [arbiter correction 8].

What the example shows. Nothing in the chain needs gain, beating air, or resonance. Ordinary storage and switching give the spikes; ordinary induction gives an electric field in tissue that is, if the manufacturer's numbers are right, large enough to move membrane voltages by tens to a hundred millivolts for as long as each 2 µs half-cycle lasts. That is a mundane candidate mechanism for the physiological effects seen in the small trials below. The example rests on manufacturer figures; no one has measured dB/dt or the induced field from this device in a phantom [arbiter, claim 3].

Two side checks (also in example.py, part 7): - Plasma frequency. For f_p = 100 MHz the electron density would be n = eps0 m_e (2 pi f)² / e² = 1.2 x 10^14 m^-3; for 240 MHz, 7 x 10^14 m^-3. An atmospheric arc has n ≈ 10^22-10^24 m^-3 [web: search-result summary of https://www.osti.gov/servlets/purl/1595718; page not opened], giving f_p ≈ 0.9-9 x 10^12 Hz. Reading "eigenfrequency" as plasma frequency misses by about 8 orders of magnitude in density. - Photon energy. h f = 4.1 x 10^-7 eV at 100 MHz and 9.9 x 10^-7 eV at 240 MHz, against kT = 0.027 eV and a 1 eV chemical step: about 2.4 million radio quanta per reaction step, all of which would have to arrive at one molecule before water turns them into heat. The 14N NMR frequency at 0.1 T is 0.31 MHz, interestingly close to the 240 kHz base frequency, but it needs a steady field and its quanta are smaller still.

6. Evidence table

Verdicts are the arbiter's (Step 3 scale). "Tested?" means a controlled test with a stated method.

Safety record, kept in full. The FDA's 10 June 2005 letter called the device significant-risk ("Class III non-heating diathermy") and listed, from the agency's files, "2 Wistar rat deaths and one case of tremor" in animal studies and "2 cases of tachycardia and one fatal myocardial infarction during treatment" in humans [web: FDA letter, file in research/files]. These are the regulator's characterisation (a party with a regulatory stake); no causal investigation of the events was found. The 2002 FDA inspection found units imported as a "plant and seed germinator", 37 study sites, and an infant with rhabdomyosarcoma treated about 29 times; the inspector wrote "the science behind the device is not clear" [web: FDA EIR (establishment inspection report) 2002, file in research/files]. The Seattle Times (19 Nov 2007) quoted Pappas saying the machine "is absolutely safe" and "can cure cancer and AIDS" and linked deaths to the device (the newspaper's characterisation; Pappas an interested party) [web, per sites.md]. A California court in 2005 barred a US distributor from misleading claims (reported by Quackwatch, a site with an anti-quackery stance) [web, per sites.md]. Other regulatory items: FDA Import Alert 89-08 lists the "PAP-IMI Ion Magnetic Conductor Device" (importer Bio-Energy Services; the Athens firm BioElectroDynamics added in 2018) [web: https://www.accessdata.fda.gov/cms_ia/importalert_244.html]; a 24 June 2003 FDA Warning Letter to LifeWorks Wellness Center, Clearwater FL, on consent and IRB failures (device name redacted; CIRCARE, a research-subject advocacy group, files it as PAP-IMI) [web: http://www.circare.org/fdawls3/jonesjaye_20030624.pdf]; and a Health Canada advisory of 28 Dec 2007 (known only via Wikipedia's references, not checked directly) [web, per sites.md]. The device continues under CE marking (the manufacturer's declaration of conformity required to sell a device in the European Union) in Europe and is used in university trials since 2023, so the record did not go dark; it moved jurisdiction [arbiter section 6C].

Supporters and the inventor's side, kept beside the safety record. These are what supporters report, each with its source and stake; none is an independent controlled trial, and the same is true of most of the adverse record above. - The Seattle Times (19 Nov 2007) reported that Pappas treated C. Wallach's chronic back pain and that "both claimed the treatment a success". Wallach was later the sponsor-investigator of the US study; both are interested parties [web, per sites.md]. - A paper by Dr Shri Kant Mishra appears as exhibit 23 of the 2002 FDA inspection report. It was not obtained, so its content is unknown [web: FDA EIR 2002, per sites.md]. - Earlier single-subject PAPIMI reports describe use in refractory seizures, skin wound healing and fatigue syndrome (cited as refs 40-42 in Viti 2023) [web, per sites.md]. They are case reports, the record before the 2023 pilots. - A memoriam by Karampatsou, a collaborator since 1996 (an interested party), reports conference presentations in Salzburg and use in Cuba, Vietnam and Swiss clinics; it cites no studies [web, per sites.md]. - The manufacturer claims a penetration into tissue of "about 20 cm", repeated in Keilani 2025 [web: https://pmc.ncbi.nlm.nih.gov/articles/PMC12534222/]. The patent says only that depth is "proportional to the intensity of the applied field" and "can reach any desired depth" [read: p.5]. (Faraday's law gives a field wherever B changes, so depth is real in principle; the 20 cm figure is the manufacturer's characterisation, not a measurement on record.) - The patent states that cells with a low transmembrane potential "are the source of pain generation signals" [read: p.4]. This is the patent's own link to the later pain trials (row 5a). The arbiter lists it with claim 5 without a separate verdict; it falls under the untested transmembrane-potential mechanism (row 5c).

7. What a learner should take away

What is solid. - A capacitor discharged through a spark gap into a coil makes a strong, fast-changing magnetic field, ringing at a base frequency with high-frequency bursts at its peaks. Pappas's oscillograms show this correctly, and the base design goes back to his lightning-conductor patents. - By Faraday's law, that field makes an electric field inside tissue with no field in the air. At the manufacturer's numbers it is thousands of volts per metre, enough to shift membrane voltages by tens of millivolts and far above nerve-stimulation thresholds. Real physiological effects are therefore expected, and small sham-controlled pilots report some (HRV, fatigue). - Arcs have negative differential resistance and can sustain oscillations; short pulses briefly beat the DC breakdown of air. Both are real physics that Pappas built on.

What is wrong, and why the mistakes are instructive. - A ratio of resistances is not a gain. Gain must compare like with like, through a circuit equation. Dividing by an open circuit's "infinite" resistance tells you only that a switch has closed. And peak power far above average is what storage plus fast release always gives; the duty cycle alone makes it 10^5. Lesson: always do the energy budget over a whole cycle, not at an instant. - Field inside a membrane is not field needed outside. The cell collects the outside voltage drop over its whole diameter and puts it across a 5 nm membrane. Mixing the two scales cost the patent a factor of about 100, and then it compared the wrong number with a wrong value for air. Lesson: write down which field, where, before comparing numbers. - The problem solved was not the problem that existed. Induction never needed a strong field in air. Lesson: check whether the obstacle is real for the mechanism you are actually using. - Same word, different physics. "Resonance" and "eigenfrequency" mean something specific. Plasma frequency depends on electron density, not the element; atomic lines are optical; NMR needs a steady field; and a radio quantum is a millionth of a chemical step. Lesson: put a number on a resonance (its frequency and its energy) before building on it. - Self-inconsistent numbers. Three burst durations, "diameter of 6 square millimeters" in claim 5, a 30 cm coil against claim 6's "fraction of one centimeter" [read: p.7].

What is genuinely open. - What the 1991 IECEC paper actually measured, and whether it contains an input-output energy account. - The real L, C, current and induced field of the device: the manufacturer's "several 10 kA" does not fit 60-96 J (Step 3 of the example). - Whether the MHz-GHz bursts change with the gas (Pappas) or with lead and electrode geometry (textbook). A gas-substitution spectrum would decide it. - Whether the small positive pilots hold up in a larger independent trial with checked blinding, and whether their effects come from ordinary induced currents, sensation that defeats the sham, or placebo. - What happened in the 2002-2005 adverse events. - A Google search for the title of the missing 1991 paper ("Energy Creation in Electrical Sparks" Pappas) returned Google's notice that a result had been removed after a legal request, logged with the Lumen Database [web: research/2026-10-04-google-search.md]. This touches the one document that would decide claim 1. Competing explanations: a routine copyright takedown of a copied conference paper (common for engineering proceedings); a request by the inventor, his family or a company, for unknown reasons; or a different page that merely matched the query. Nothing in the record prefers one. The Lumen notice itself and a library copy of the 1991 proceedings would decide it [arbiter section 9]. - A December 2011 PDF by Pappas on an OverUnity forum, whose search snippet mentions "non conventional energy machines comprising arc and negative resistance", was not opened (login required). It may be a later statement of claim 1 in his own words [web, per sites.md]. - Two long documents were saved by the Investigator but not read in full: the inventor's own "Physiology of the Cell" (2006, 67 pp.) and the 2002 FDA inspection report (17 pp.). Whether either holds an input-output energy measurement, a calibrated oscillogram, a gas comparison, or the trial protocol and outcomes is unknown. - Why no large independent trial exists after 30 years (competing explanations in arbiter section 6: low commercial incentive with an expired patent, ethics hurdles after the FDA record, a mainstream belief that the mechanism is absurd; none documented).

8. Check yourself

1. A 0.5 µF capacitor charged to 10 kV discharges through a coil of 2 µH. Find the stored energy, the ring frequency, and the peak current. Is a peak power of 100 MW possible, and does it imply gain?
2. Using E = (r/2)(dB/dt), what induced field does a coil with B0 = 0.05 T oscillating at 240 kHz give at r = 5 cm? What ΔV does that put on a 10 µm-radius cell in the steady limit?
3. The patent says a cell membrane needs 10^6 V/m and air breaks down at 10^6 V/m. Give the two numbers you would use instead, and say why neither is the field that must be applied from outside.
4. If the fast bursts were really oscillations at the electron plasma frequency, what electron density would 240 MHz imply, and why does changing the gas from air to nitrogen not by itself change f_p?
5. Why does a 100 MHz burst produce a much smaller membrane voltage than the 240 kHz base ring, even at the same E?

9. Further reading

Mainstream: - D. J. Griffiths, Introduction to Electrodynamics, chapter 7 (Faraday's law, inductance). - E. M. Purcell and D. J. Morin, Electricity and Magnetism (RLC and alternating-current circuits). - F. F. Chen, Introduction to Plasma Physics and Controlled Fusion (plasma oscillations). - Yu. P. Raizer, Gas Discharge Physics (breakdown, arcs, negative characteristics). - P. Nelson, Biological Physics (membranes, resting potential, nerve impulses). - J. C. Weaver and Yu. A. Chizmadzhev, "Theory of electroporation: a review", Bioelectrochemistry and Bioenergetics 41 (1996) [memory, high]. - Electroporation threshold and Schwan's relation, review: https://pmc.ncbi.nlm.nih.gov/articles/PMC12045266/ [web]. - On the singing arc and arc transmitters: any history of early radio covering Duddell and Poulsen.

The inventor and the device: - P. T. Pappas, US Patent 5,556,418 (1996), "Method and Apparatus for Pulsed Magnetic Induction" (the project document) [read]; family WO 94/01176, EP 0648140 B1 [web: https://patents.google.com/patent/US5556418A/en]. - P. T. Pappas, "Energy Creation in Electrical Sparks and Discharges: Theory and Direct Experimental Evidence", Proc. 26th IECEC, Boston 1991, vol. 4, pp. 416-423 (not yet obtained; the key missing source). - P. T. Pappas, "The original Ampère force and Biot-Savart and Lorentz forces", Il Nuovo Cimento B 76, 189 (1983) [web, per sites.md]. - P. T. Pappas, "Pappas' Physiology of the Cell and its Atomic Energy" (Athens, 2006) [web: research/files]. - US 7,151,372 B2 (NMR/EPR reframing) [web: https://patents.google.com/patent/US20060220648A1/en]. - Related work and predecessors: Cure, Pappas and Eichler, US 5,068,039 and US 5,239,442 [web: https://patents.google.com/patent/US5068039A/en]; Liboff et al., US 5,160,591 (ion cyclotron resonance) [read: p.1, cited prior art]; EP 0402552 (lightning conductor) [web, per sites.md]. - Imitators and later devices on the same design: Horkel DE 10 2007 030325 (a near copy that lists Pappas among the original assignees); Pulsartec EP 2062614; Healing Machines WO 98/57704 [web, per sites.md and arbiter section 8]. - Minor historical leads: Em-Probe, Inc. named as assignee on one WO listing; Pappas called a "professor" (in sceptical quotation marks) at TEI (Technological Educational Institute) Piraeus by a 2009 Greek blog (snippet only, not opened) [web, per sites.md]. - Trials: Viti et al. 2023 (PMC10001461), Viti et al. 2025 (PMC12385873), Keilani et al. 2025 (PMC12534222), Leisner et al. 2002 (PMID 11913824) [web].

---

Answers

1. W = (1/2)(0.5 x 10^-6)(10^4)² = 25 J. f0 = 1/(2 pi sqrt(2 x 10^-6 x 0.5 x 10^-6)) = 1/(2 pi x 10^-6) = 159 kHz. Z0 = sqrt(2 x 10^-6 / 0.5 x 10^-6) = 2 ohm, so I0 = 10^4 / 2 = 5 kA. Peak power V0 I0 / 2 = 25 MW; 100 MW is not reached in this ideal ring, but even 25 MW is a million times a 25 W average at one pulse per second. That is energy released quickly, not gain: the total delivered per pulse is still at most 25 J.
2. dB/dt = 2 pi x 240 000 x 0.05 = 7.5 x 10^4 T/s; E = 0.025 x 7.5 x 10^4 = 1.9 x 10^3 V/m. ΔV = 1.5 x 1885 x 10^-5 = 28 mV.
3. Membrane: 70 mV / 5 nm = 1.4 x 10^7 V/m (2.3 x 10^8 V/m with the patent's 3 Å). Air: about 3 x 10^6 V/m DC. Neither is needed outside because a cell of radius a turns an outside field E into a membrane voltage of 1.5 E a; for a = 10 µm, about 5 x 10^3 V/m outside gives 70 mV. And with induction the field is made inside the tissue, so air never sees it.
4. n = eps0 m_e (2 pi f)² / e² = 7 x 10^14 m^-3, about 8 orders of magnitude below an atmospheric arc. f_p depends only on electron density (and the electron's charge and mass, the same in every gas); the gas affects f_p only indirectly by changing how many electrons the arc frees.
5. The membrane is a capacitor charged through the resistive fluids inside and outside the cell, with charging time τ_m of order 100 ns. At 240 kHz the field changes slowly compared with τ_m, so the membrane fully charges (factor 0.98). At 100 MHz the field reverses before charge can build up (factor about 1/(ωτ_m) ≈ 0.01), so the membrane voltage stays about a hundred times smaller.