Superhot Atomic Oxy-Hydrogen Flame
KeelyNet 03/18/00

Thanks to Dr. William A. Rhodes for sharing this.


Readers may wonder why I waited three decades before regaining interest to probe several unanswered questions of this system.

A friend on the Internet discovered the second patent number under my name and notified me that another party had patented a new version of this concept and was claiming discovery of a new gas.

Inspection of his patent showed his claim as discoverer was not valid, since my first patent predated his by eleven years. I was not about to allow him that recognition.

After all, I reasoned, should the second man on the moon logically claim the distinction of being the first? And so, research began resulting in this document.

The answers here are by no means conclusive, but lead to a better understanding of a very complicated reaction. If references are found proving priority over mine, then I will yield.

The name of the culprit was Yull Brown of Australia, now deceased who "invented" Brown's Gas.


Parameters & Variables
Dr. William A. Rhodes, Physicist


This concept was discovered in 1961 by request from a manufacturer for a new and novel means for producing torch flame temperatures beyond those of that era.

Such system was conceived and developed involving electrolytic production of mixed hydrogen and oxygen. Prior to that time, literature on the subject focussed exclusively on separation of such gases and conducting them out of the electrolyzer for tank storage.

Using hydrogen and oxygen immediately when generated through a common duct was not found in the literature and it appeared to be a new technology.

The first patent (Apparatus For The Electrolytic Production of Hydrogen And Oxygen For The Safe Consumption patent # 3,262,872 issued July 26, 1966.) dealt with intermixing the gases in an electrolyzer, issuing through a common duct for instant use in a torch. 9 claims in the patent read on;

"an outlet for said generator to conduct a mixture of hydrogen and oxygen gases therefrom,"

as opposed to other electrolyzers using separate ducts for each gas.

The patent contains the financing party as co-inventor. His contribution was limited to an additional small alcohol booster tank, entraining the vapor for a reducing flame.

The patent appears to establish my precedence in the art. Starting in 1962, the Henes Mfg., Company of Phoenix sold many thousands of their trademarked 'Water Welder' in several sizes, presently continuing under another name.

Immediately after launching the Henes venture, I began research on a large electrolyzer patented in March 1967 under the title, MULTICELL OXYHYDROGEN GENERATOR (3,310,483).

It contained 60 iron plates, nickel plated on the oxygen generating side and iron on the hydrogen side. This patent claims use of loosely fitting grooves for holding the plates in tank 8"x8"x16"x3/4" Plexiglas.

I previously discovered that current could not bypass such plates loosely fitting in grooves of proper design. The torch flame from that unit was 20 inches long, melting everything into blue-white puddles, including firebrick, ceramics and carbon (in argon atmosphere).


Of all elements, hydrogen and oxygen should hold no secrets.

Yet, in this example they do and have been troublesome. Many experts in such gases contributed important knowledge hoping such would answer our questions. Their offerings were accurate for tank gases, but these were not tank gases and three major obsta- cles remained.

  • 1. flame propagation rate (burning velocity) was unusually high.
  • 2. Flame temperature is far greater than tank gases.
  • 3. Allowing the gases to mix at the moment of genera- tion, and delivered in a common duct for immediate consumption should contain both molecular and atomic components. Until these were examined through experiment and observation, conjecture and theories were invalid.


SETUP: A phototransistor cell was attached to a Plexiglas base containing a groove to locate the start and finish marks on a known length of transparent plastic tubing.

An ignition chamber with sparkplug was attached to a 2500 v transformer controlled with a button switch. The electrolyzer was attached to the input end of the spark chamber, a 22 ft length of tubing was attached to the output side of the chamber.

The first marked tubing position was placed in the phototransistor groove, and the 20 ft mark was placed on top of the first tubing mark. Recording equipment included a dual pen strip-chart recorder with parallel connection to a memo-scope and audio tape recorder.

NIST WWV clock ticks were coupled to all. With this setup we hoped to capture precision measurements of flame front velocity plus rise and fall time.

TESTS: Electrolyzer gas purged the tubing, and since the flame is in the UV, the electrolyzer was allowed to run until a trace of KOH allowed visual spectra to produce a slight pink-white. Stripchart, memoscope and recorder running and standardized. Spark initiated.

SIX SEQUENCES: Recorded timing for 10 feet of tubing was consist- ently 1.225 milliseconds = 10,000 ft in 1.226 seconds, or 8160 ft/sec div 1088 ft/sec (speed of sound not compensated for our 1150 ft above sea level) was mach 7.5.

Rise and fall pulse duration via memo-scope was .5 millisecond with a total baseline to baseline duration of .6 millisecond. With exception of the small error between sea level and 1150 ft above sea level, re- sults of these tests appear reliable. This combination has MAXIMUM INSTABILITY. Any electrostatic discharge can trigger a very mild explosion compared with tank H2 & O2. The "ashes" from burning are of course pure water.

LUMPED FLAME RATE CONFIRMATION The previous rate was resolved from pip spacings. These tests were made with the plastic tubing wound into a small donut with phototransistor mounted on the focal plane of a camera lens. A flat-black background behind the donut and floodlight illumina- tion allowed the donut image to be adjusted to cover the active area of the phototransistor. The tubing beyond the measured marks were covered to prevent errors from their exposure.

Instrumentation and standardization was identical to the previous test. Recorded data of the previous were pips, marking the beginning and ending of the flash. This time, burn illumination produced a slightly rounded flat area with a baseline to baseline rise and fall of .6 milliseconds as before.

(Previous test shots allowed strip chart recorder gain adjustment for approximately 3 cm reading. The flat tracing showed gradual rise and fall of about 2 mm from beginning to end of a sequence.)

Time measurements of six sequences were identical to the previ- ous. The last test was made with the tubing exit clamped off, and gave readings identical with the others. No tubing rupture occurred and explosion sound was muffled. These should provide sufficient evidence of the flame propagation rate of such mixed gases.


Flame tests in an argon atmosphere directed on several layers of carbon fiber fabric with its micron size filaments (Used on the stealth fighter & bomber.) melted carbon filaments into brilliant globules. This means carbon's melting temperature 3550C/6422F is exceeded, but its boiling point 4827C/8720F is not attained. Past that point no reference exists.


First, be aware we are dealing with common-ducted gases, data being absent from NIST and the literature. There is also theory vs experimental evidence to contend with. From the CRC handbook: "Lifting power of 1 cu/ft hydrogen is about 0.075 lb at 760 mm pressure."

SETUP: Our test volume chosen was 1 liter single duct electro- lyzed gases. An igloo from a plastic pop bottle was cut to provide exactly 1000 ml volume between the flat igloo door top, and the upper dome. (1000 ml was from a standard 1000 ml flask, transferred to the pop bottle, marking the door top, and extend- ing the igloo another 2", where it was lathe cut and the doorway snipped out.

It was located inverted on the pan of our Mettler milligram balance. An L shaped tube on lab stand extended through the doorway and bent upward ending near the dome top, leaving the balance completely free of interference.

The gas generator was purged of air 15 minutes. The balance was tare arbitrarily adjusted for 30 grams +- 1 mg. The igloo was filled with pipe smoke; -6 mg deflection noted due to warmer air.

The gas tube was attached and maximum weight reduction of 0.510 grams was attained, rounded off to the nearest mg. Gas input was allowed to flow for 30 minutes for accuracy.

5 minutes after gas cutoff, the balance returned to the pre-gas reading caused by rapid diffusion of electrolyzed gases into atmosphere. Comparing H2 lifting power, 1 liter mixed gases multiplied to 1 cu/ft provided lifting power of 0.0311 lb. Or 41% that of H2.

Here we must consider single atoms of hydrogen 1 and oxygen 16 for lifting power against atmosphere (29+). Of course, if a stoichiometric mix of H2 & O2 were present, O2 alone would have a molecular weight of 32, and such gases introduced in the igloo would show a slight weight increase as the combination spilled *downward* through the doorway.


Over the years many suggested if such gases were collected and remained unused, several kinds of recombinations would spontane- ously occur regardless of temperature.

Determining volumetric changes of stored electrolyzed gases was done with a calibrated 100 ml domed bell of 1/4" thick Plexiglas open at the bottom and sliding inside a closely fitting Plexiglas container, with an L shaped gas entry tube extending upward under the bell.

The bell was held in place to prevent upward movement. 500 viscosity silicone oil was poured into the outer cylinder as air inside the bell was slowly exhausted, causing the oil to fill the bell completely, continuing to flow slowly into the plastic vacuum tubing, to eliminate all air. A cock on the metal L tube was turned off, and the plastic tube pulled from the L and cleared of oil.

Room temperature was adjusted for 80 F. When the temperature of the oil over the bell read 75 F, gas electrolysis began, allowing the plastic line to be purged of air, then connected to the cock which was turned on.

Gas filled the bell from top downward below the 100 ml mark. The cock was turned off, gas line pulled, and generator switched off. The cock was cracked to bleed gas down to the 100 ml level and turned off.

At the end of 6 months, room temperature again increased oil temperature to 75F. Volume change was not measureable. The gas was then allowed to fill the in- verted bell on the gram balance.

Calculations gave the same answer as previous, comparing lifting ability as being 41% that of H2. (Plus or minus 2% error.) To prevent any light activity, the system was covered with black polyethylene.


The only purpose of KOH is to create the lowest possible resist- ance eg, highest electrical conductivity. Being slowly depleted by mist generated during electrolysis, specific gravity must occasionally be corrected by addition of KOH.

It is noted that any sharp metallic whisker in the storage atmos- phere could cause an explosion, similar to the dangers of storing high percentage hydrogen peroxide, where the entire contents can burst into high pressure steam with disastrous results, just because somewhere in the interior someone forgot to round off a sharp edge.

On the other hand, these mixed gases were ignited repeatedly in a 4 liter container of 16 gage iron with flat ends and sparkplug. The only evidence of ignition was a sharp click, with no damage to the vessel.

A recent report revealed one experimenter was wounded with shrap- nel from such explosion. The only way this might happen is from accumulation in an unusually thin container, or one made from an easily fractured plastic.

However, a duplication of the original multicell unit was constructed of 3/4" Plexiglas with an interior volume of 8 liters. Half of this was filled with electrolyte leaving 4 liters for foam and gas accumulation, (Identical to the volume of the iron container.

The multicell had a 2.5" diameter rupturable diaphragm of food grade Saran wrap. The unit was set on a stand in the open and ignited. The resultant pop splintered the case into many pieces which were all deposited within a radius of 5 feet around the stand. The diaphragm remained in- tact.

Such indicated the sonic wave front was responsible in- stead of pressure which would have ruptured the diaphragm.

These tests allowed us to design electrolyzer tanks of materials and thicknesses that could contain flashbacks. Viewing the permanent Plexiglas multicell in operation, electrolyte foam rises upward, but at maximum elevation allows sufficient gas space above.

Therefore no purpose is served with designs containing more gas than necessary for conduction out of the reservoir. Extrapolation of chart curves indicate a possible diesel type explosion as pres- sure approaches 400 psi. However, this is not conclusive. Generation of such single ducted gases appears to be an event not found in nature, unless lightening produces them.


There are two types of arrester. For small units of one or two liters total tank capicity, two acquarium aerator stones are adequate. Over time they tend to clog with KOH vapor, but can be easily cleared by backflushing with 50% phosphoric acid.

For larger units a water filled U-tube is service free and best, since its inertia disallows flame movement through the water. An alternate to the U-tube is two tubes of different diameters. [INCOMPLETE]


Flame propagation rates refer to complete combustion mixtures to fill a measured length of tubing and after ignition, combustion speed is measured against standard time pulses from WWV transmis- sions from the National Institute of Standards and Technology.

From the literature, the Butane rate is 60 ft/sec. Acetylene 330 ft/sec. Tank Hydrogen (H2) 680 ft/sec. Since no literature could be found for mixed atomic gas, burning velocity was precision measured in our lab.


Be aware of this: If a current i flows for a time t and reacts with water whose electrochemical equivalent is e, mass of the gases released is: m=eit. This means present chemistry is forev- er restricted by this equation.

Direct current wave shapes, frequencies, half-waves, full-waves, nothing will allow gas delivery approaching unity. Some claim that under certain elec- trical manipulations, cells run cooler, or may produce more gas than before.

Yet, if precision measurment instruments are avail- able, they will always show results exactly following this equa- tion. Amperage readings made of rectified direct current by some hang-on ammeters produce enormous errors, leading the observer to believe cell efficiency has improved. This requires special attention to exhibit trustworthy data.

Note: Data on gas species percentages are incomplete. No reference source exists for atomic gases.

W.A.Rhodes. 3-13-2000
Email: Dr. William A. Rhodes

Original OxyHyd discovery from 12/97/00

If you found this file useful or interesting, please consider a donation or a purchase to help keep KeelyNet online and providing free information. Even a dollar will help. Others sell it, we prefer to share it, thanks!