Primer: Larson-Tyson (Realm Industries), "water with unnatural bond angles"

Study document (Step 4). It adds no new verdicts; all verdicts come from arbiter.md. Source tags: [read] from the project documents (two US patent applications and a Ventura County Star article); [web] with URL (mostly found by the Investigator, see sites.md and research/); [memory] recalled textbook facts with confidence; [calc] arithmetic reproduced in example.py.

Who is speaking. Everything about what the "vapor" is and does comes from the applicants, William Stehl and Tyson Larson, and their assignee Realm Industries. They had a commercial stake, so these are their own characterization. The regulators (US Environmental Protection Agency, EPA; California Division of Occupational Safety and Health, Cal/OSHA), the prosecutors and the press are interested parties on the other side, and their descriptions are reported as theirs. Source gap: none. Both patents and the article were read in full, and the graphs were read from page images. This version includes the QE remediation (qe.md, verdict MINOR) and the arbiter's QE revision (arbiter.md section 10).

---

1. The idea in one paragraph

A water molecule is shaped like a shallow "V": an oxygen atom with two hydrogen atoms, the arms 104.5 degrees apart. Stehl and Larson proposed that if you pass electric current through a stack of closely spaced metal plates in water, the stack behaves like a resistor-capacitor circuit with its own "resonating frequency". Mixed with the power supply's own frequency, they said, this signal pushes the V wider without breaking the molecule. They held that the bent molecules leave the water as a gas that can be stored under pressure for months, and that when sparked the molecules snap back to normal and release energy, leaving only water. They reported a cutting torch and engines (two small Honda engines and a Toyota FJ-40 truck) running on this gas, with exhaust that looked almost like clean air. The textbook reading of the same apparatus is an electrolysis cell. It splits water into hydrogen and oxygen, and that gas mixture (called "oxyhydrogen", "Brown's gas" or "HHO") burns, runs engines and cuts metal. The inventors' own later international filing conceded that their process "could result in the production of hydrogen gas (H2) and oxygen gas (O2)". The project ended in tragedy: Tyson Larson was killed in an explosion at the workshop in June 2010.

---

2. What you need to know first

Electrolysis and Faraday's law. Two metal electrodes in water with a voltage across them split water: hydrogen bubbles form at the negative electrode (cathode) and oxygen at the positive one (anode). Each hydrogen molecule takes two electrons through the circuit, so the gas produced is set by the charge passed (Faraday's law). Below about 1.23 V nothing is split; real cells need 1.8-2.5 V. Read: Atkins and de Paula, Physical Chemistry, the electrochemistry chapters; or any general chemistry text (e.g. Zumdahl, Chemistry, the electrochemistry chapter).

Chemical energy: Gibbs energy, enthalpy and heating value. Splitting one mole of water needs at least 237 kJ of electrical work (the Gibbs energy) and 286 kJ in total (the enthalpy). Burning the hydrogen back gives at most 286 kJ, or 242 kJ if the product water leaves as steam (the "higher" and "lower heating values", HHV and LHV). Energy is conserved: no cycle that starts and ends with liquid water can give out more than went in. Read: Schroeder, An Introduction to Thermal Physics, the chapters on free energy and chemical thermodynamics.

Capacitors, resistors and RC circuits. A capacitor (capacitance C, farads) stores charge; a resistor (resistance R, ohms) dissipates energy. Together they give a time constant tau = RC (seconds): charge leaks away as exp(-t/RC). An RC circuit has a corner frequency but no resonance. Resonance needs two stores that trade energy back and forth, such as a capacitor and an inductor (L), with natural frequency 1/(2 pi sqrt(LC)). Read: Purcell and Morin, Electricity and Magnetism, the chapter on alternating current circuits; or Horowitz and Hill, The Art of Electronics, chapter 1.

Molecular vibrations and spectroscopy. A molecule's shape is the lowest point of an energy landscape (its potential energy surface). Bending the shape away from that minimum costs energy, and the molecule vibrates around it like a mass on a spring. These vibrations are quantised: water's bending vibration absorbs infrared light at 1595 cm^-1 (wavelength 6.3 micrometres, frequency 47.8 THz). An excited vibration gives its energy away in collisions very quickly. Read: Atkins and de Paula, the chapters on molecular spectroscopy (vibrational spectra); or Banwell and McCash, Fundamentals of Molecular Spectroscopy.

Vapour pressure. A liquid in a closed container evaporates until the gas above it reaches the saturation vapour pressure, which depends only on temperature. Push the gas above that pressure and it condenses. For water this is 3.2 kPa (0.46 psi) at 25 C and one atmosphere at 100 C. Pressures in the patents are in psi (pounds per square inch; 1 psi = 6.895 kPa). A tyre-gauge reading ("psi gauge", psig) is the pressure above the surrounding atmosphere; the absolute pressure is the gauge reading plus 14.7 psi. The patents do not say which they mean; section 5 takes 35 psi as gauge and notes the other case. Read: Schroeder, the chapter on phase transformations (Clausius-Clapeyron relation).

Combustion and exhaust analysis. When fuel burns in air, each oxygen molecule used disappears from the exhaust. Counting atoms (carbon, hydrogen, oxygen, nitrogen) before and after combustion lets you check whether exhaust readings are possible. Read: Heywood, Internal Combustion Engine Fundamentals, the chapters on thermochemistry and on pollutant formation.

---

3. The textbook view

3.1 The water molecule's angle is a minimum, not a setting

The H-O-H angle of 104.5 degrees and O-H length of 0.96 angstrom are the bottom of water's energy landscape [memory, high]. Even at room temperature the molecule is always bending back and forth by several degrees around that angle (zero-point motion). A molecule "with an angle greater than 104.5 degrees" is simply caught mid-vibration, or vibrationally excited.

Two numbers measure how much bending costs: - One quantum of bend: E_bend = h c nu-tilde, where h = 6.626e-34 J s is Planck's constant, c is the speed of light and nu-tilde = 1595 cm^-1 is the wavenumber. E_bend = 0.198 eV [calc]. (1 eV = 1.602e-19 J.) - Pushing the molecule all the way to a straight line (180 degrees) costs 11,114 cm^-1 = 1.378 eV = 133 kJ/mol [web: https://www.sciencedirect.com/science/article/abs/pii/S0009261405011899; calc].

At temperature T the fraction of molecules with extra energy E is about exp(-E/(k_B T)), where k_B = 1.381e-23 J/K is Boltzmann's constant. For the straight-line shape at 300 K that fraction is 7e-24 [calc]. There is no hump on the way back down from a bent shape, so nothing holds a bent molecule bent. In a gas at a few atmospheres a molecule collides about 10^10 times per second [calc, per contrarian], and an excited bend relaxes within nanoseconds to microseconds [memory, medium]. The energy turns into ordinary heat.

3.2 The plate stack as a circuit

Two facing plates of area A (m^2), separated by a gap d (m) filled with water, form both a capacitor and a resistor:

• C = eps_r eps_0 A / d, where eps_0 = 8.854e-12 F/m is the vacuum permittivity and eps_r = 80 is water's relative permittivity.
• R = d / (sigma A), where sigma (siemens per metre, S/m) is the water's electrical conductivity.

Multiply them: RC = eps_r eps_0 / sigma. The area and the gap cancel. The time constant of the bulk water depends only on what the water is, not on how the plates are built. Its inverse is a relaxation rate, not a resonance. (Real electrodes add a much larger "double-layer" capacitance at the metal surface, which pushes the time constant longer [memory, medium].)

3.3 Electrolysis energy

To make one mole of H2 you must push 2 moles of electrons, a charge of 2F, where F = 96,485 C/mol is the Faraday constant, through a cell at voltage V. The electrical energy is

W = 2 F V (joules per mole of H2).

At the reversible 1.229 V that is 237 kJ; at the thermoneutral 1.481 V, 286 kJ [calc]. Any higher voltage is lost as heat in the cell. Burning the H2 gives back at most 286 kJ (HHV) or 242 kJ (LHV).

3.4 What oxyhydrogen does

A 2:1 mixture of H2 and O2 carries its own oxidiser. It burns without air, it explodes if sparked in a closed container, and it gives a hot, small flame used for jewellery and thin-sheet work. The adiabatic flame temperature (the temperature the burnt gas would reach if no heat escaped to the surroundings; a real flame is cooler) is about 3,075 K (2,800 C) [web: https://etasr.com/index.php/ETASR/article/view/6132, via contrarian]. Its only combustion product is water, so exhaust contains no CO or CO2 from the fuel. Engines can run on hydrogen, and hydrogen added to petrol or diesel engines is a real, studied technique [memory, high].

---

4. Where the inventors depart

The departure happens at the first step. The inventors treat a relaxation time as a tunable resonance and assume it can reach into the molecule. "Summing" two frequencies has no defined meaning in a linear circuit: two signals added together are still two separate frequencies; sum and difference frequencies appear only if something nonlinear mixes them, and the patent gives no equation for it [arbiter 10.3]. The rest of the mechanism (bent, stable, energy-storing molecules) depends on that step.

What they observed is real enough. Bubbles on the plates (the patent itself describes gas bubbles that need flow or ultrasound to detach [read: [0035]-[0041]]), pressure building in a closed vessel, a gas that burns, a torch that cuts thin metal, engines that run. Every one of these is what oxyhydrogen does. Their observations do not separate their explanation from electrolysis, because no measurement of the gas itself was ever made.

---

5. Worked example: what does the patent's own generator do?

The patent's examples [read: US20090092540 [0058]-[0059]] give one apparatus: four chambers, 20 plates each, each plate 4 ft^2, gap 0.1765 in, 5 V, gas stored at 35 psi. They do not give the current, the water's conductivity or the frequency. Where a number is needed I say it is ASSUMED. Every number below is reproduced by example.py (run command in its header).

Step 1: units

• Plate area A = 4 ft^2 x 0.0929 m^2/ft^2 = 0.3716 m^2.
• Gap d = 0.1765 in x 0.0254 m/in = 4.483 mm.

Step 2: the capacitance of one gap

C = eps_r eps_0 A / d = 80 x 8.854e-12 x 0.3716 / 0.004483 = 5.87e-8 F = 58.7 nF.

With alternating plates, the 19 gaps in a chamber are in parallel: 19 x 58.7 nF = 1.12 microfarads.

Step 3: the resistance and the "frequency"

R = d / (sigma A). The conductivity sigma is ASSUMED (the patent gives none):

Two lessons. First, the plate size and gap dropped out (RC = eps_r eps_0 / sigma), so the patent's idea of tuning C with plate area, number and gap cannot change this frequency. Second, the inventors' 1-22 kHz window corresponds to nearly deionised water, which conducts so poorly that little current flows. Whether the patent meant rad/s or Hz changes the numbers by 2 pi but not the conclusion.

Step 4: can that frequency bend a molecule?

• One quantum at 22 kHz: E = h f = 6.626e-34 x 22,000 = 1.46e-29 J = 9.1e-11 eV.
• One quantum of bend: 0.198 eV. The ratio is 2.2 billion.
• Static push: the field is E = 5 V / 4.483 mm = 1,115 V/m. Water's dipole moment p = 6.2e-30 C m. The ratio p E / (k_B T) at 300 K is 1.7e-6. Thermal jostling is about 600,000 times stronger than the field's pull.

So the drive cannot reach the bend directly. (Honest limit: this rules out direct resonant excitation. It does not rule out every possible nonlinear effect at the electrode surface, but the patent proposes none and measured none [arbiter].)

Step 5: the same stack as an electrolyser

Which voltage does each gap see? The examples say "four parallel chambers" [read: [0058]], but [0049] separates the chambers' physical arrangement from their electrical connection, which may be "in series or parallel as may suit the application". The example does not say which. If parallel, every gap sees the full 5 V. If the four chambers were in series, each sees about 5/4 = 1.25 V. This worked example assumes parallel wiring, 5 V per gap. The waveform (continuous, pulsed or AC) is not stated either.

At 5 V, each mole of H2 costs W = 2 F V = 2 x 96,485 x 5 = 964.9 kJ. Burning it gives back at most 241.8 kJ (LHV), so the round trip is 25.1% (29.6% on HHV). A well-run electrolyser near 1.8-2 V does two to three times better.

ASSUMED operating point: the patent's preferred power density of 30 W/ft^2 [read: [0054]] over 4 x 20 x 4 = 320 ft^2 gives P = 9.6 kW, so I = P/V = 1,920 A. By Faraday's law, H2 is made at I/(2F) = 9.95 mmol/s = 13.4 L/min at standard conditions, plus 6.7 L/min of O2. Its chemical power is 2.41 kW, a quarter of the 9.6 kW put in.

That is a lot of gas: enough to fill a vessel to 35 psi steadily, and enough to run a small torch. It also shows why "uses less energy than electrolysis" fails if the gas is H2 + O2: the cell is an electrolyser, and an inefficient one at 5 V.

The series case gives the opposite picture. At 1.25 V per gap, 2 F V = 241.2 kJ per mole H2, 99.7% of the LHV [calc, example.py]. That is barely above the 1.229 V reversible voltage, and real cells need roughly 1.5-2 V to make gas at a useful rate [memory, medium]. So a series-wired example generator would make almost no electrolysis gas, and the 35 psi of stored gas and the engine runs would be harder to explain by electrolysis. The 25% round-trip figure holds only for parallel wiring. The patent does not choose, and neither reading is established; only the inventors' apparatus could answer it [arbiter 10.2]. Faraday's law (gas per coulomb of charge) applies either way, and was never measured.

Step 6: could the stored gas be water vapour?

The Antoine equation (an empirical fit), log10(P / mmHg) = 8.07131 - 1730.63 / (233.426 + T/C), gives water's saturation pressure at 25 C as 3.16 kPa = 0.46 psi [calc]. Storage at 35 psi gauge is 343 kPa absolute, 109 times that. Water reaches that pressure only near 138 C (extrapolated; about 126 C if the 35 psi meant absolute). At room temperature any H2O in the vessel condenses to liquid. Hydrogen and oxygen, whose critical temperatures are -240 C and -119 C, stay gaseous at any such pressure.

What the example shows. Taken at face value, the patent's own apparatus is a large electrolyser. It can make roughly the gas volumes and pressures described. It cannot store a pressurised water "vapor", and its drive frequency has nothing to do with the molecule's bend. This is a calculation, not a measurement of their gas. The measurement that would settle it (a gas chromatograph or mass spectrometer reading of the product) was never made.

---

6. Evidence table

Verdicts are the arbiter's (arbiter.md section 9).

Material kept in the record, whatever the verdicts: - The inventors' own later filing WO 2010/039283 (filed 3 Apr 2009; "WO" marks an international application filed under the Patent Cooperation Treaty, PCT, which reserves the right to seek patents in many countries at once; it is not itself a granted patent) renames the product "an energy carrier, which includes but is not limited to molecules with unnatural bond angles and H2 and O2 gas" and says "it may be possible that covalent bonds are broken". It still claims "less energy is consumed" than electrolysis [web: https://patents.google.com/patent/WO2010039283A2/en]. This is the inventors' hedge, possibly a patent lawyer's caution; it does not show they abandoned their belief. - The inventors' own applications list points two ways. The patent lists over a hundred uses. Some read as if the product contains free oxygen: "Aircraft Oxygen Systems", "Firefighter Breathing Apparatus", "Portable Oxygen Apparatus", "Tankless Underwater Breathing Apparatus", "Atmospheric Oxygen Enrichment" [read: 0092540 [0145]-[0147], [0155], [0191]]. Others read as free hydrogen or a reducing agent: "High Purity Gases", "Down Hole Hydrogenation", "Reductive Purification of Metals without Carbon" [read: [0115], [0118], [0216]]. Yet the same list calls the product "Rocket Fuel - Non-explosive" and "Vehicle Fuel - Non-combustive" [read: [0170]-[0171]], and [0021] says the O-H bonds stay intact. The list states no composition and is not reconciled. It is applicants' text, and grouping it is a reader's inference; the arbiter treats it as weak evidence beside the WO concession, not a substitute for a gas analysis [arbiter 10.4]. - A patent was granted. The continuation-in-part issued as US 7,793,621 B2 on 14 Sep 2010. Its 4th-year maintenance fee was paid on 12 Mar 2014, after the European withdrawal and the parent's abandonment, and it lapsed in 2018 [web: https://patents.google.com/patent/US7793621B2/en, via Investigator]. So someone was still paying in 2014; who is unknown. The Investigator reports that its claim 1 meters air and mixes vapor, air and a fuel (not independently confirmed by the arbiter); that would match claim 1 of the published CIP, which is an air-and-fuel engine method, not the air-excluded mode. A grant is not a test of the physics. The parent application was abandoned in 2013 after an unanswered office action; the European application was withdrawn in 2013. The examiners' objections have not been read. - Strongest honest case for the inventors (contrarian, kept by the arbiter): oxyhydrogen engines are real, so "a gas made from water ran an engine" is credible; the vapor-phase traces repeat across three runs; the inventors published readings that were not zero and a gasoline phase they could have trimmed; the authorities describe a working gas-production operation filling cylinders for sale. - Related work: R2D2 LLC (WO 2023/055615, 2022-2023) claims a "structurally altered gas molecule" from water with angle 94-104 degrees, stable above 300 psig, made with crossed electric and magnetic fields; the examiner cited Realm's application against it [web: https://patents.google.com/patent/WO2023055615A1/en]. This is a re-use of the idea, not confirmation; no measurement data were found. Stanley Meyer's "water fuel cell" (1980s-1990s, also pulsed DC across closely spaced electrodes in water) is the best-known earlier claim in the same family [memory, high]; "Brown's gas", named after Yull Brown, is ordinary oxyhydrogen under another name and is sold for torches [memory, high]. - Where the process came from. Stehl's companies (BGX Technologies and Rainbow Technologies, Saranac Lake, New York) raised money for the process from 2001. Timothy A. Larson, Tyson's father and Realm's founder, was one of the investors and is not accused of wrongdoing. When BGX failed in late 2005 he "believed in the science", in the prosecutor's words, paid to ship the equipment to California and set up Realm Industries with his sons [web: https://archive.vcstar.com/business/simi-business-in-fatal-explosion-originated-with-failed-ny-venture-ep-368507413-349112401.html, via Investigator; AP repeats it]. This is the prosecutor's characterization of a person's belief, not evidence about the physics. The earlier "Rainbow Technology patents" have not been identified. The DOJ release names another use claimed for Stehl's energy source: processing precious metals from a slag pile in Silver City, New Mexico [web: https://www.justice.gov/usao-ndny/pr/second-defendant-convicted-8-million-fraud-scheme-concerning-purported-alternative]. In 2015 hazardous and reportedly radioactive items were found in a garage on a Saranac Lake lot owned by Stehl (North Country Public Radio and nny360 snippets, not opened, via Investigator); they may be remnants of the original lab, a lead on what materials the process used, and verify nothing. - The one outside scientist on record. Dan Raftery, a Purdue University chemistry professor, told a reporter: "I don't believe this patent application really is viable"; it appeared to be a form of electrolysis, and large-scale electrolysis could create the conditions for a big explosion [web: VC Star 21 June 2010, URL above, via Investigator]. He worked from news reports and the application, not the apparatus. One informed opinion, documented in one source. - The full incident record (each item from one source, not reconciled with the others): 2008, a high-pressure tank "imploded due to mechanical valve failure" and Larson had burns (Simi Valley Acorn); the VC Star recorded early reports of a hydrogen cylinder top blowing off; EPA's POLREP (pollution report, the EPA's running situation report on a cleanup) mentions "reports of an additional explosion in Simi Valley in 2008" [web: https://response.epa.gov/sites/7236/files/SSI_polrep_2.htm]; 17 June 2010, fatal; 9 Aug 2011, Sylmar, two injured. Fire Engineering (2014), citing a Strategic Sciences/EPA report, counts "at least four explosions", two of them unreported in 2008 [web: https://www.fireengineering.com/firefighting/the-dangers-of-hydrogen-based-experimental-fuels-full/, via Investigator]. - The legal record: Stehl and a partner were prosecuted in New York for misleading investors (2001-2010, Rainbow Technologies and successors); Rossignol was convicted in 2014, and Stehl entered an Alford plea on 2 Sep 2014 (a guilty plea in which the defendant does not admit the acts but accepts that the prosecution's evidence would likely convict) [web: https://www.justice.gov/usao-ndny/pr/second-defendant-convicted-8-million-fraud-scheme-concerning-purported-alternative]. The charges concerned false statements about contracts and licensing, not whether the process worked. The prosecutor said the case was "not alleging that this is some kind of fraudulent scientific process" (VC Star, via Investigator). The convictions do not refute any physics claim; they are a reason to want independent data for unwitnessed tests. - Where the record goes dark. After the August 2011 Sylmar explosion and the destruction of the cylinders, there are no papers, test reports or products. Competing explanations (arbiter section 6): ordinary electrolysis gas that failed commercially and legally; the death, injuries, penalties and prosecution ended the work whatever its merit; work moved abroad (a South Korea shipment was attempted in 2011) or under another company name; an unanswered technical rejection by patent examiners; suppression (the only hint is an unexplained removal notice under the DMCA, the US Digital Millennium Copyright Act, which lets a party ask a search engine to drop results it says infringe copyright; Google noted it had removed 10 results on one search, filer and content unknown). The best fit is the first, with the second as the proximate cause of the end, but the decisive documents (patent file wrappers, Cal/OSHA files, court exhibits, the Korean consignee) are unread.

---

7. What a learner should take away

Solid. - Current through plates in water makes hydrogen and oxygen; the mixture burns, runs engines, cuts thin metal and detonates if stored together. All of Realm's observations sit comfortably inside this. - A plate stack in water has an RC time constant set by the water's conductivity, eps_r eps_0 / sigma.

Wrong, and why it is instructive. - Resonance versus relaxation. An RC circuit has no resonance. Calling 1/(RC) a "resonating frequency" smuggles in the idea that it can pump energy into something. Always ask: where are the two energy stores that trade energy? - Scale mismatch. A kHz signal carries quanta about 10^9 times smaller than a molecular vibration. The other frequencies the patent names do no better: the power source's "intrinsic frequency" (most plausibly a 10 Hz pulse rate or 60 Hz mains, about 10^12 too small) and the GHz radiation embodiment (about 2-4 x 10^4 too small). Before proposing that one oscillation drives another, compare their frequencies and energies. - Shape is not a stored state. A bent molecule with no barrier to return is just a hot molecule. Energy stored this way leaks out as heat in nanoseconds; it does not wait for a spark. - Phase diagrams are a quick test. "Water vapor at 35 psi at room temperature" fails a one-line vapour-pressure check. - Check instruments with atom balances. Gasoline exhaust above about 15.4% CO2 (dry) is impossible in air; the patent's own data cross that line, which says the analyser chain was unreliable at the high end.

Genuinely open. - What exactly the gas was. No neutral measurement exists; the one sample was destroyed. A gas chromatograph reading of the product from a rebuilt example generator would settle it in one run. - Whether "tungsten melted" was surface oxidation or something else (no rod size, time or photo). - Why the vapor-phase exhaust shows CO2 at 5-15 times the air level and O2 slightly below ambient: residual gasoline, sampling carry-over, or lean hydrogen burning in intake air. - Whether the example generator's chambers were wired in series (about 1.25 V per gap, almost no electrolysis gas) or in parallel (5 V, an inefficient electrolyser). The patent allows both. - Who told the authorities it was "a water-based explosion". - What the patent examiners objected to, and what happened to the 2011 Korean shipment.

---

8. Check yourself

1. A friend proposes to raise the "resonating frequency" of the Realm cell by halving the plate gap. Using R and C for parallel plates, what happens to RC, and why?
2. A cell is run at 2.0 V instead of 5 V. What is the electrical energy per mole of H2, and what fraction comes back on burning (LHV)?
3. Why can a stoichiometric H2 + O2 mixture run an engine whose air intake is blocked, while a gas made only of H2O molecules (bent or not) cannot?
4. An engine burns x moles of H2 per 100 moles of intake air (21% O2). Each H2 uses half an O2, and the water formed is removed before the analyser. Show that dry exhaust O2 = (21 - x/2)/(100 - x/2), and find x for a reading of 20.2%.
5. Stoichiometric gasoline (taken as CH1.85) burns in air. Dry exhaust contains CO2 and N2 only. What is the maximum dry CO2 percentage, and what does a reading of 18% tell you?

---

9. Further reading

Mainstream - P. Atkins and J. de Paula, Physical Chemistry: electrochemistry; vibrational spectroscopy. - D. V. Schroeder, An Introduction to Thermal Physics: free energy, phase transformations. - E. M. Purcell and D. J. Morin, Electricity and Magnetism: AC circuits, dielectrics. - J. B. Heywood, Internal Combustion Engine Fundamentals: combustion stoichiometry, exhaust analysis. - R. O'Hayre et al., Fuel Cell Fundamentals: the thermodynamics of water splitting and recombination, worked numerically. - Barrier to linearity of water: https://www.sciencedirect.com/science/article/abs/pii/S0009261405011899

The inventors' own key works - W. Stehl and T. Larson, US 2009/0092540 A1, "Method and Apparatus of Modifying Bond Angles of Molecules" (filed 5 Oct 2007) [read]. - W. Stehl and T. Larson, US 2009/0090312 A1, "Alternative Fuel Engine" (filed 2 Oct 2008), granted as US 7,793,621 B2 [read; web: https://patents.google.com/patent/US7793621B2/en]. - WO 2010/039283 A2, "energy carriers" [web: https://patents.google.com/patent/WO2010039283A2/en]. - Record: Ventura County Star, 17 June 2010 [read]; EPA Region 9 POLREP #2 [web: https://response.epa.gov/sites/7236/files/SSI_polrep_2.htm]; sites.md and research/ for the full trail.

---

Answers

1. R = d/(sigma A) halves and C = eps_r eps_0 A/d doubles, so RC = eps_r eps_0/sigma is unchanged. Only the water's conductivity (or permittivity) changes it. More plates or bigger plates do not help either.
2. W = 2 F V = 2 x 96,485 x 2.0 = 386 kJ per mole H2. Burning returns 241.8 kJ (LHV): 241.8/386 = 62.7%. At 5 V it was 25.1%.
3. Combustion needs an oxidiser. H2 + 1/2 O2 brings its own oxygen, so the engine needs no air. H2O is already fully burned hydrogen; it has no chemical energy to release with oxygen, and only the bending energy (at most 133 kJ/mol, which would leak away as heat) could be offered.
4. Start with 100 moles of air: 21 O2, 79 N2, plus x H2. Burning uses x/2 O2 and makes x H2O, which is removed. Dry exhaust: O2 = 21 - x/2, N2 = 79, total 100 - x/2. So O2 fraction = (21 - x/2)/(100 - x/2). Set it equal to 0.202: 21 - x/2 = 20.2 - 0.101 x, so 0.8 = 0.399 x, x = 2.0. About 2% H2 in the air, far too little to power a 4.2 L engine under load. It is also below hydrogen's lower flammability limit of about 4% (the leanest mixture with air in which a flame can spread) [web: https://www.aiche.org/sites/default/files/docs/pages/the_elemental_-_hydrogen_flammability.pdf, via arbiter], so a 2% hydrogen-air mixture would hardly burn on its own. That is why the arbiter says the data imply either an oxidiser-bearing gas, negligible power, or residual gasoline.
5. CH1.85 + 1.4625 O2 gives CO2 + 0.925 H2O, and the O2 brings 1.4625 x 3.76 = 5.499 N2. Dry products: 1 CO2 + 5.499 N2, so CO2 = 1/6.499 = 15.4%. A reading of 18% is above what gasoline can produce in air: either extra oxygen came from somewhere (the electrolysis gas), or the analyser was mis-reading (the arbiter's oxygen balance on the same readings also fails, pointing to the analyser).