Graham R.L. Cowan is a long-time enthusiast for nonpolluting personal vehicle power through nuclear production of motor fuel. When hydrogen cars proved they could go 300 km and be re-fueled by an existing hydrogen infrastructure, he was persuaded they were just around the corner. The range needed to double to match what drivers were used to, but this was obviously doable.

Twenty years on, it had not been done. Around the corner they remained. Wondering what the holdup was, and noticing that the lightest fuel needed very heavy tanks to reduce the likelihood of devastating fuel-air explosions (which are inevitable anyway) he wondered if there wasn't some better clean fuel. To ask this question, while equipped with background chemical knowledge that instantly offers candidate fuels that are not just safer, but safe, is to learn two things in one moment: hydrogen cars will not be arriving, and gasoline's days are numbered.

He wrote "Boron: A Better Energy Carrier than Hydrogen?" in 2001, published "How Fire Can Be Tamed" in the IJNHPA in 2008 and since then has continued to refine his vision of how cars will be fueled when several billion people are daily drivers.

How fire can be tamed

Graham R.L. Cowan

105, 1144 Division St., Cobourg, Ontario K9A 4J9, Canada

E-mail: This email address is being protected from spambots. You need JavaScript enabled to view it.

Abstract: Combustion dies at the interface between breathable air and

macroscopic pieces of certain involatile fuels. If fed only them, in a compressed

oxygen chamber, it makes an almost sun like flame that cannot run wild. If upon

dilution to manageable coolness the ash drops to the chamber bottom, and from

there can be removed without the diluent, true harness brokenness is possible.

Excess oxygen can be the diluent without thereby being wasted. It can rid itself

of the diluted flame’s heat, and spare many trees from becoming newsprint

bearing motor fuel mishap reports, by working in a thermodynamic cycle.

Some ashes, especially boria, both precipitate well from the diluted flame and

travel well. By visiting faraway solar or fission power stations, and returning to

the chamber as regenerated fuel, they can make combustion both docile, and

subsidiary to docile primary energies.

1 Introduction

Where fuel combustion seems most certain to continue and hugely increase, tame or not,

is in car motors. Taming it there can be considered to have six parts:

1 Liven up the oxygen (LUTO): equip cars with extractors of oxygen from air.

(These normally deny a little oxygen to air-breathing bystanders so as to provide it,

purified, to something that needs it so. On occasion they must cut that thing

entirely off in the bystanders’ favour. Accordingly, and to save space, they will be

called ‘oxygen deniers’ or ‘O-deniers’ from here on.)

Int. J. Nuclear Hydrogen Production and Applications, Vol. 1, No. 3, 2008 235

Copyright © 2008 Inderscience Enterprises Ltd.

2 Burners, learn to burn eclectically (BLeBE): choose a fuel whose ash can readily

be let out of a chamber full of hot compressed oxygen without letting much of the

oxygen come along. The fuel’s combustion must propagate in compressed oxygen

but die in air at breathable temperature and pressure.

3 Star-pyrolyse ash to fuel and oxygen (SPAFO): develop nuclear or solar power

stations to import the ash and convert it back to oxygen and fuel, the latter in small

uniform pellets shaped for feedability into pressure chambers.

4 Strongly confine oxygen in oxide (SCOO): make the inner surfaces of the oxygen

chamber of materials that hot, compressed, nearly pure oxygen cannot attack.

Include yielding surfaces so that when the fuel of part two burns within it, the

oxygen can expand and do work.

5 Let engines eat pellets (LEEP): equip cars with bins for the fuel pellets of part 3

and ways for the internal combustion engines of part 4 to take them from the bins.

6 Return ash to sender (RATS): Equip cars with bins in which ash can stay until

refuelling time, so that it then can be offloaded in exchange for fuel, and begin its

journey to the power stations (part 3).

A linear heap of one-carat (0.0002-kg) diamonds probably would satisfy the dying-in-air

condition of part 2. Strong heating at one end would cause the diamonds there to burn,

but the fire would not travel along the heap. Raising the pressure and, as part one implies,

making the oxygen less dilute could change this. But diamond ash is gaseous, and this

makes part two’s other requirement, that departing ash not take extra oxygen with it, hard

to meet.

Also, one variety of diamond ash is carbon monoxide, a gas that kills stealthily when

inhaled. Diamond fires, and carbon-oxidising fires in general, are not tame.

The earth’s atmosphere normally contains enough of the other diamond ash – carbon

dioxide – that the power stations of part three could import all the ash they would need

merely by letting it come to them in the wind. Part six would be unnecessary: all the cars,

all the power stations, and all of us would be in a single ash bin, the same one that now

contains just us and the cars. However, a car motor that burns diamonds in oxygen has

another difficulty: at present-day levels of efficiency, about half its power would go to the

oxygen denier.

A much smaller share would be taken if the heat engine could get the same heat using

much less oxygen. This turns out to be a real option, as potential fuels that are more

suitable than carbon in other ways typically also work their oxygen much harder.

2 Oxygen denier readiness

Although today’s oxygen deniers leave room for improvement, they are good enough to

go on with. A typical one is said to produce 34.5 kg/h of oxygen and mass 1590 kg.

Air sieve materials that let argon pass have been developed, but the portable oxygen

deniers now in commercial service use materials that sift out oxygen and argon alike.

Their normal product is a mixture with an oxygen mole fraction of 0.945, argon in the

same proportion to oxygen as in air, balance nitrogen (Santos et al., 2006).

236 G.R.L. Cowan

The theoretical work done by a single-stage compressor in feeding air through them has

been assessed in terms of megajoules per cubic metre of product at standard temperature

and pressure (STP). Plots of that work versus the production rate show it increasing

slowly for a while, then more quickly. For a commercial device using three different sieve

materials – Oxysiv 5, Oxysiv 7, MS S 624 – the knees in their plots appear, for the first

and last, respectively at 1.60 MJ/(STP m3) and 1.20 MJ/(STP m3). For Oxysiv 7, the greatest

work requirement is 1.02 MJ/(STP m3) (‘TE’ curves, Santos et al., 2006, Figure 4).

If Oxysiv 7 is used and a car motor turns a compressor that indeed has just one stage,

but is only 60% efficient, the car motor must provide 1.70 MJ/(STP m3) of shaft work.

That is 0.38 kilowatt-hours per kilogram of oxygen.

2.1 The fate of the argon . . .

At its simplest, a tame fire chamber would only let gas in, not out. It would admit no gas

but oxygen, and as much oxygen as came in would be incorporated into non-gaseous ash

and exit in that form.

Revised to use an oxygen denier like those available today, it would admit gases,

principally argon, that fire cannot condense. Argon would build up and the pressure

would increase. This would snuff the fire in one of a number of ways: increasing the

chamber pressure until no more oxygen could be injected, or until the chamber burst, for

instance. The fuel will not burn in air.

2.2 . . . And the resulting waste of oxygen

At the cost of increasing the oxygen denier’s energy demand by about 10%,1 this problem

can be solved. The extra energy would be needed to bring in that much more oxygen.

Along with argon, it will be allowed to flow back out of a small gas exit port. Argon’s

accumulation will thus be limited, and will not prevent continuous combustion.

The oxygen denier that takes 0.38 shaft kWh per produced kgO2 will therefore need

0.42 shaft kWh per kilogram of oxygen that the fire actually uses.

NIST data imply that a fire whose minus-delta-’G’ at 298.15 K is 1 kWh must, if

zirconium-fuelled, consume 0.111 kg of oxygen. Most other fuels take more: carbon

takes 0.292 kgO2/kWh, iron, if being oxidised to the (II, III) oxide magnetite, takes

0.226 kgO2/kWh (Linstrom and Mallard, 2005).

Multiplying these oxygen requirements by 0.42 kWh/kgO2 yields the fractions of the

fire’s energy the O-denier would take: for zirconium 0.0466, for carbon 0.123, and for

iron 0.0952. But since the O-denier runs on mechanical work, these fractions need to be

divided by a typical heat engine efficiency, say 0.3, and for zirconium, carbon, and iron

that gives 0.155, 0.409, and 0.317.

2.3 Capping the oxygen denier’s fractional take . . .

If there is a type of mechanical linkage that can divide the power from a turning shaft in

a fixed proportion between two loads, it could ensure that an O-denier that is expected to

need 0.317 of the power of the motor it feeds can never take more than 0.32.

The energy cost of extracting oxygen from air rises as the oxygen fraction diminishes,

so with such a power-fraction-limiting linkage, an oxygen denier might fall behind the

motor’s oxygen demand if its intake air began to contain less than 20% oxygen.

How fire can be tamed 237

2.4 . . . means a motor can starve for oxygen . . .

The motor would begin to lose oxygen pressure, and therefore power. The O-denier, able

to take no more than the fixed fraction of the now reduced power, would still not be able

to keep up. Soon it and the motor would stop.

2.5 . . . in the midst of plenty

This turns out to be just what they ought to do. People in a freezing house may someday

seek warmth in the garage by using there, as a furnace, an air-breathing car that oxidises

some carbon-free fuel. If the car has an oxygen denier that cannot, as it deoxygenates the

garage’s air, take more power to compensate, it will soon live up to its name by cutting

off the motor.

It will do this even though, with a less stingy allocation of the motor’s output, they

both might still have been running hours later, long after hypoxia, unheralded by any rise

in carbon dioxide levels, had stealthily and permanently shut down the people.

Portable oxygen deniers’ energy needs can in theory still decline by an order of

magnitude. If reductions continue incrementally, designers who pair them with heat

engines will be obliged incrementally to ratchet down, also, the fraction of the engines’

output they can take, keeping it barely over the minimum needed in fresh air.

Calling them oxygen deniers may make this critical denial function harder to overlook.

2.6 The inert hordes dodged a bullet

The necessity of thinking about hypoxia arises not just with tame fires but with the

encompassing class of carbon-free fuel-air reactions.

After hydrogen cars were impressively demonstrated in the mid to late 1970s, many

hoped or believed they were within 5–10 years of catching on. Had this been true, perhaps

by the late 1980s all cars would have been safe for prolonged running indoors, but that

safety is not intrinsic to hydrogen-air flames nor to fuel cell cathodes.

2.7 Internal iron combustion power

If a truck has a 1590-kg oxygen denier that provides 34.5 kg of oxygen per hour, and of

that, the fire in its heat engine combines 31.4 kg per hour with iron to form magnetite, it

will go:

238 G.R.L. Cowan

although perhaps not very quickly. Using 0.3 for heat engine efficiency, the above yields

for carbon 19 kW (26 hp), for iron 28 kW (38 hp), and for zirconium 72 kW (96 hp).

Dividing the fire’s oxygen consumption rate by the net power yields oxygen consumption

per net shaft kWh: carbon takes 1.65 kgO2/kWh, iron takes 1.12 kgO2/kWh, and zirconium

takes 0.44 kgO2/kWh.

Net power_(31.4 kg O2 per hour) *

(heat engine efficiency/mass of oxygen required per unit energy yielded by fire

_ (0.42 kWh per kg O2 ))

3 Burners, learn to burn eclectically (BLeBE)

The fuel chosen should burn readily enough in purified compressed oxygen, but be

non-ignitable in breathable air. If pulverisation would make it ignitable in air whereas

in palpably large pieces it is not, then in a fire-taming effort one would not pulverise it.

The side effects of making it inhalable and hard to clean up when spilled, would also be

avoided. Similarly, one would not have it generate another fuel, one that burns readily or

explosively in air, by taking oxygen from that other fuel’s ash.

Aluminium is the best-known fireproof fuel, but there are others. They include

beryllium – toxic and very expensive – and boron and silicon. They include heavier

elements, such as the above-mentioned zirconium, that clothe themselves in oxide films

that strongly impede ignition. If a net shaft kWh from a zirconium-burning vehicle motor

requires 0.44 kg oxygen, which over time averages roughly half in the ash reservoir, half

in the atmosphere or in transit back to it, then that kWh also requires 1.23 kg zirconium,

which is always on board.

With boron, aluminium, and silicon the always-on-board masses are, respectively,

0.276, 0.483 and 0.485 kg, and the combining oxygen masses are 0.613, 0.430 and 0.552

kg, per net shaft kWh. This again is on the basis of 0.3 heat engine efficiency and an

oxygen denier that requires 0.42 kWh per combining kilogram of oxygen. Unless for some

heavy element there is a compelling reason to accept the extra mass, tame combustion

should be fed one of these three.

The net-power calculation of the previous section predicts 51 kW for boron, 73 kW

for aluminium, and 57 kW for silicon. The truck mentioned there could keep up with

traffic if it were burning any of these.

3.1 Ability of ash to precipitate from diluteness in hot oxygen

An ash that fails to do this, and is produced in a flame much hotter than a solid pressure

envelope can bear, will not lend itself to that flame’s dilution down to bearable

temperature with excess oxygen, for letting the ash out will tend to entail also letting out,

and wasting, the oxygen.

If, however, cooling due to dilution causes the ash to condense and fall out, a quantity

of oxygen can repeatedly do three things – dilute ash, drop it, and rid itself of the heat the

ash gave it by acting as heat engine working fluid – before in its turn being incorporated

into ash and falling to the vessel’s bottom.

Among combustibles that will not burn in air, few present any difficulty for the

ash-fall part of this scheme; maybe only osmium, since its tetroxide’s normal boiling

point is 403 K. If there are osmium-burning-car enthusiasts, their dream will not be

advanced by the taming of fire.

Of elemental boron, aluminium, and silicon, none is screened out, for their ashes

– boria, corundum (aka alumina), and silica – all have normal boiling points well above

2000 K. By the time their flames in oxygen have diluted themselves below this temperature,

they have become smoke plumes, upwellings of hot oxygen in which involatile ash particles

fall slowly.

How fire can be tamed 239

3.2 Selective removability of precipitated corundum

If the purpose the smoke is to serve is expanding and pushing a piston in a cylinder, this

typically needs to be repeatable with the same piston and cylinder about 100 million

times: thousands of times per minute for thousands of hours. This will not happen with

corundum-bearing oxygen, as corundum is abrasive. In an environment of hot dusty

oxygen that is just beginning to do its work of expansion in an efficient automotive heat

engine, at the temperature of the surface upon which the oxygen presses, corundum is

harder than any other material that would be stable.

That surface could well be made of corundum. Strong confinement of oxygen in

oxide (SCOO) calls for a strong oxide. Corundum is a strong oxide. But if they are not

to scratch it, ash particles in the oxygen must be made of something softer.

Aluminium-burning rocket motors embody another solution to this problem – accept

a short lifetime. But there may be a way an aluminium burner can live a long life, despite

expelling corundum dust throughout it: let aluminium fill its combustion chamber, as

liquid, and burn oxygen in an inverse flame in mid-metal.

Liquid aluminium has a history of being fairly successfully managed, including

travelling hundreds of kilometres on public highways as 5-m3 crucibles-full. At departure

from a foundry such a load is at 1150–1200 K. Ideally it has not yet cooled below 1050 K,

more than 100 K above its freezing temperature, when it arrives at a casting site.

It seems reasonable that compressed oxygen injected from non-oxidisable nozzles

deeply submerged in pressurised liquid aluminium would be consumed near the nozzles,

none of it reaching the upper surface, and that the combustion products would initially

rise through the liquid as bubbles, but quickly lose heat to it and collapse. Farther up, the

flame would be one of hot liquid metal carrying corundum particles up through cooler

liquid metal.

Compared to a corundum-bearing updraft of oxygen, this other convecting fluid

offers the advantage that a material exists – aluminium diboride – that promises to resist

both it (through being saturated with it) and its corundum particles (through possibly

superior hardness). Although combustion would be internal to the liquid metal, it would

be external to aluminium diboride heat exchanger tubes placed in the updraft’s way, so

the system would be an external combustion engine. The working fluid in the tubes could

not be oxygen, but it would not have to be.

At the chamber bottom, where the liquid aluminium would be relatively cool, there

could be a drain with a screw impeller in it, both also made of aluminium diboride, in

order to let the precipitated corundum particles and the liquid metal entrained by them

exit as a sludge or paste. The drain would give onto a chamber where aliquots of the

sludge would fall onto a draining surface. As much liquid metal as might readily drain

out of them would do so, and be returned to the upper chamber. The aliquots would then

be cooled and put in an ash ingot bin.

This would amount to somewhat unselective ash removal because some liquid

aluminium between particles would get frozen into the ingots. Per net shaft kWh, more

than the previously noted 0.483 kg of aluminium would move from fuel bin to ash bin.

By helping to hold the ash ingots together, the unburned metal would aid in returning ash

to sender (RATS).

240 G.R.L. Cowan

3.3 Selective removability of precipitated boria and silica

Boron melts above 2000 K and silicon at 1687 K. Neither, when burned with pure oxygen,

can usefully be present in excess to dilute the flame, for neither shares aluminium’s

ability to remain fluid when cooled to temperatures well below the maximum a heat

engine’s solid parts can long endure. Because of this, they cannot coolly surround a flame

that starts out much hotter than that maximum and through turbulent mixing form a

diluted flame whose temperature is just right. Of the flame’s two reagents, only the

oxygen can do this. The result is a smoke of boria in oxygen or silica in oxygen.

Either smoke’s expansion would be less rough on a heat engine than that of

corundum-laden oxygen. Silica, whose crystals melt at 1983 K, is, like every

non-corundum ash that might be so suspended, less abrasive than corundum. If it

originally was vapour and has become a suspended condensate through being diluted and

cooled by the oxygen around it, boria exists as droplets, which supercool. When they land

on a surface that is hotter than about 0.8 of boria crystals’ 723-K melting point, they coat

it. Water droplet impact in engines powered by wet steam can be destructive, but boria

droplets are very different. A film formed by earlier-arriving boria droplets can so cover

a surface that latecomers cannot abrade it at all.

On oxygen-boron combustion chamber walls warm enough for accreting boria

droplets to merge into a film, the film will flow. Along with direct droplet precipitation,

this flow can create a chamber-bottom boria lake. A dynamic equilibrium can be

established where boria comes to the lake’s surface and, at the same rate, removes itself,

plus as much oxygen as was soluble in the surface layers, through a lake-bottom drain.

This removal is selective because gaseous oxygen cannot get down there.

If in the drain the boria has a temperature of 873 K, somewhat cooler than liquid

aluminium, it is runny enough to drain at a useful rate. Unlike a paste of corundum and

liquid aluminium, it is entirely liquid and unabrasive. It needs no screw impeller to help

it along.

Its dynamic viscosity and density, represented below by ‘_’ and ‘_’, are 480 pascal

seconds and 1608 kg/m3 (Smith and McBroom, 1999). A 0.25-m-deep lake of it that is

held at a combustor bottom by its standard terrestrial weight of 9.80665 N/kg will have

a top-to-bottom hydrostatic pressure difference, represented below by ‘_p’, of 3942 Pa.

A laminar flow calculation that neglects end effects:

How fire can be tamed 241

predicts that boria will exit through a circular hole in the bottom at a rate of 0.0189 kg/s

if the hole’s length L and diameter D are both 0.0388 m, and its walls are at the same

873-K temperature as the liquid.

In preparation for return of ash to sender – RATS – the boria extruding below the

bottom of the drain could periodically be cut off. The now detached gob could be cooled

by air flow, and then, with a little more air flow, blown into its bin.

Raising the temperature by 200 K could speed the flow up about 18-fold. A spun lake,

held to its bed by centrifugal force rather than planetary gravity, could drain much faster

still. But the 0.0189-kg/s oozing, with an average flow speed of 0.0100 m/s, turns out to

be adequate for draining the small volumes per unit time that a car-scale combustor

would produce. According to the previously noted per-net-shaft-kWh boron and oxygen

Mass throughput _ π _ D4* (_p)/(128 _L)

masses, which sum to 0.889 kg, the net driveshaft power that could be co-produced is

76 kW.

For silica that mass sum is 1.037 kg, highest of the three fuel-ash pairs under

consideration. Since corundum is more abrasive than silica, conceivably an engine

powered by expansion of hot compressed silica-bearing oxygen could be internally

surfaced with corundum, and not be eviscerated by the silica particles. But wherever they

settled, they would tend to stay. They would not be able to merge and flow. If they did

merge, it would amount to sintering. Perhaps they would form tough artificial quartzite

or vitreous silica scale on heat engine surfaces. Along with the extra mass, this prospect

makes internal silicon combustion relatively uninteresting.

4 Central station deoxidation

Billions are determined to become motorists. One way or another, motor fuel production

rates will rise by many terawatts over the next few decades. They can be increased cleanly

by constructing nuclear or solar power stations that take in alumina or boria ingots

and remove the oxygen. Such stations could be rated in terms of aluminium terawatts,

TW(Al), or boron terawatts, TW(B).

If each station were to produce 0.1 TW(Al), it would make about three times as much

as all of today’s smelters. Since this is roughly a two-orders-of-magnitude increase in

individual plant scale, by the inverse-square-root-of-scale rule (Marchetti, 2006), product

unit cost would decline by one order of magnitude, from around US$0.30/kWh(Al) down

to about US$0.03/kWh(Al). In aid of this cost reduction, inert anodes might be put

into service (Welch, 1999), so that where now a significant fraction of the energy stored

in aluminium comes from oxidising carbon anodes, scaled-up plants would need only


A 0.1-TW(B) power station might need only heat, provided heat of a certain intensity

was available. Boria follows the general rule for very stable oxides that mere direct

heating, however intense, will not cause macroscopic separation into fuel and oxygen.

Star pyrolysis of ash to fuel and oxygen (SPAFO) yields fuel and oxygen atoms. They

remain mixed, and will not refrain from reattaching to each other if one cools the vapour

down to a temperature where, if they would so refrain, separation might conveniently occur.

However, heat from a source significantly hotter than 2500 K can usefully act on

another oxide that is less stable, and – being an ore of iron – very much cheaper and more

abundant: magnetite (Ehrensberger et al., 1997; Mohai et al., 2007):

(l_x)/(1_4x) Fe3O4(1/2) O2 _ 3/(1_4x) Fe(1_x)O(liq)

For x_0, NIST data imply this process has enthalpy change _372.3 kJ/mol, plus another

39.2 kJ/mol that the oxygen would give back in being cooled from 2500 to 298.15 K.

Being liquid, the ferrous oxide tends to separate from the oxygen, so they can be cooled

without recombining.

Losing the 39.2 kJ would be reasonable for a solar power station that focussed a large

image of the sun down onto a high-altitude outdoor stream of magnetite, for then the

half-mole of oxygen could go directly into the upper air. If such a station annually turned

32.6 billion kg of magnetite into 1.9 billion kg of oxygen and 30.7 billion kg of ferrous

oxide, its annual average output could be expressed as 1 GW(FeO).

242 G.R.L. Cowan

Where summer is much sunnier than winter, ferrous oxide production rates in winter,

spring, summer, and fall might average respectively zero, 1, 2, and 1 GW(FeO). By

summer’s end, 7.7 billion kg of ferrous oxide, a gigawatt-season’s worth, could

accumulate, perhaps as an outdoor conical heap 300 m across the base. If a steady

year-round ferrous oxide gigawatt were taken, the iron by winter’s end would be in a slightly

larger magnetite pile. Other kinds of gigawatt-season energy reservoir – two billion

lead-acid car batteries, a cubic km of water raised 800 m – are larger or more costly

or both.

Boria, like corundum, can dissolve in molten salts. Converted to 300–400 MW of

electricity, the GW(FeO) could electrolyse it there year-round, but it may also be able to

react directly with ferrous oxide and nitrogen, in a way that corundum cannot:

How fire can be tamed 243

This condenses nitrogen without producing any other gas, but is exothermic and

spontaneous at room temperature and pressure.

Raising the nitrogen pressure to a few tens of MPa should make it go at 500 K. Some

compound of boria and ferrous oxide may be so stable as to sidetrack it, or it may still

be too slow. It should be tried. If boron nitride and magnetite are seen to be producible

quickly enough, and they can be separated, three more spontaneous reactions will lever

boron up to freedom:

The above would transform the problem of deoxidising boron into one of converting

magnetite and ferrous halides back into free halogens, ferrous oxide, and iron. This

highly exothermic reaction would free the halogens:

Iron could be produced by consuming more ferrous oxide.

Finally, the magnetite could become oxygen and ferrous oxide,

and the ferrous oxide could freeze, and if x_0, 29 moles of ferrous oxide would have

been oxidised to 29/3 moles of magnetite and then recovered. Magnetite’s above-noted

372.3-kJ/mol enthalpy of partial deoxidation, times 29/3, gives the delta H of this step:

3.98 MJ/mol if oxygen is let go while still hot, 3.60 MJ/mol if its heat is recovered.

The net result would be the dissociation of one mole of boria, delta H _1.25336 MJ.

Of 3.98 MJ, that is 31.5%, about the same as if electricity had been made, but boria can

be wirelessly gathered in, and boron wirelessly distributed.

The magnetite heating could occur in an annular curtain of falling particles bathed

in rising high-pressure helium. Each particle would begin its fall as a magnetite particle.

As it descended it would give up oxygen until it was a ferrous oxide particle. The oxygen

would be swept upwards by the helium.

The heat could be radiant heat from a central fountain of chunks of an actinide

element or a mixture of actinides. Fission in the chunks would melt and evaporate them

so that they would merge as they approached the apices of their free flights. Because their

vapour would always be much denser than the surrounding helium, each chunk’s expanded

remains, although flying upwards as the chunk had been, would still be decelerating and

soon would fall back. The structure would be a dense vapour fountain.

Continuing fission in the vapour slug a chunk had become, would raise its temperature.

The resulting expansion would reduce its opacity to neutrons, so the process would limit

itself. As it fell back, later-arriving chunks, still condensed, could pass through it.

Surrounding this fountain and reflecting neutrons to it, would be a region filled

with cool flowing helium. The high transparency of helium to thermal radiation would

allow it to fill the space between the fountain and the sheath of iron oxide particles while

remaining cooler than either. Its flow would continuously remove the dense vapour

fountain’s outer surface, the remains of the longest-serving chunks, and carry it downwards.

Compared to operation at normal atmospheric pressure, the high pressure would

serve two purposes. One is to enable the cool helium to promote chain fission by

converting fast neutrons that escape the fountain into cool thermal ones before they have

gone far, so that they have a good chance of diffusing back into it.

If nuclei of natural uranium are the only ones in the vapour, these returning

cool neutrons are more likely to cause fission than in the case where they are captured

while still fast, because they are more likely to avoid capture by nuclei of the 99.28%

non-fissionable majority, and instead be captured by the fissionable minority. If the

surrounding helium has to cool too many neutrons, and its temperature increases over

time, so does theirs, and their selectivity diminishes. Therefore, in raising its own

temperature, this reactor would reduce its moderator’s effectiveness, and in so doing,

would reduce its own power. This self-regulating tendency is one it would share with

existing nuclear power reactors.

The second purpose pressure would serve is to raise the temperature at which the

actinide would boil, and thereafter reduce the rate at which it would expand, so that it

could stay near the centre longer, reach a higher temperature, and shine more strongly on

the iron oxide.

The cascade of actinide vapour and the sheath of helium, neutron-heated nearest

the vapour but still cool farther out, would together descend through a ring of nozzles

that would douse them with fluoride particles. This would condense the actinide vapour,

making the heat earlier stored in it unusable for magnetite pyrolysis, so ideally, for

efficiency’s sake, each piece of vapour would shine away much more energy during its

time of high luminosity than it would then spend in this manner.

The fluoride particles might be 16 mole parts sodium fluoride, four parts potassium

fluoride, and five parts magnesium fluoride, a composition sometimes referred to as

244 G.R.L. Cowan

NaF-16KF-20MgF2. It melts at 1077 K (Misra and Whittenberger, 1987). The amount

thrown into the descending flame would be adjusted so as to end up above this

temperature, in the hope that the particles of actinide soot would end up inside its

droplets, and there give up some fission fragments to it (Lemort, 1997):

How fire can be tamed 245

Gravity and a small pressure difference, maybe 1 kPa, together would pull the droplets,

any unincorporated actinide soot grains, and the helium around them all down into a

lower chamber. There they would impinge straight down upon the centre of a broad pool

of liquid magnesium near its 923-K freezing point. Falling into this, the fluoride droplets

would freeze into beads. These beads and the actinide soot, some of it inside them and

isolated from the liquid, would fall through it and come to rest on a layer of older beads.

The beads and soot would contain fast-decaying fission fragments. An influx of

pieces of solid magnesium would be maintained, therefore, to keep the liquid magnesium

near its freezing point. They would sink through it alongside the beads and lie with them

on the bead layer, and be melted by decay heat there. As much magnesium as was being

added as solid would be removed from the surface as liquid and dropped through helium,

which would take away the heat, probably to a heat engine.

The oldest, least heat-producing beads and soot grains would be at the bottom of

the bead layer. Under them would be a deep foundation of solid NaF-16KF-20MgF2.

It would extend far enough down to provide a long-term backup heat sink. After

solid magnesium pieces have stopped being thrown in, and all present have melted, the

liquid magnesium warms to 1077 K and the beads immersed in it melt. They form a

liquid NaF-16KF-20MgF2 layer between the solid fluoride and the magnesium. The

actinide grains formerly trapped in them fall onto the solid fluoride. Their continuing but

diminishing heat production drives the liquid-solid interface downwards. Occasionally it

reaches pockets of magnesium and causes them to drain upwards. Eventually the inward

leakage of the outside world’s coolness overwhelms the fission fragments and the melt

front becomes a freeze front. It rises above the grains, trapping them.

Usually the beads would not be allowed to melt and a dredge could selectively dredge

the deepest-down, least heat-producing ones off the solid NaF-16KF-20MgF2 surface.

The liquid magnesium could be evaporated off them, and then they could be pulverised

under liquid xenon. NIST tables say natural air-derived xenon at 175 K is dense enough

under its own 173.25-kPa vapour pressure to float NaF-16KF-20MgF2 bead fragments

but not actinide soot grains (Lemmon et al., 2005).

An alternative fluoride composition, NaF-13MgF2-22CaF2, melts at 1027 K (Misra

and Whittenberger, 1987). That is 50 K lower than NaF-16KF-20MgF2, but still higher

than magnesium’s freezing point, so much the same actinide-soot-gathering and

bead-forming behaviour can be expected of it. However, its floatability in liquid natural

xenon, even at the latter’s triple point, is uncertain. A better bet for floating it would be

the large amounts of slightly denser xenon that can be extracted from existing commercial

reactors’ spent fuel.

The actinide soot would be taken out from under the liquid xenon and the fluoride

bead fragments skimmed off the top. The soot would be pressed into chunks. Flung again

up the fountain’s axis, they would again become bright aliquots of vapour. The fluoride

would be used to cool the fountain’s effluent again.

Helium in the gas lying over the liquid magnesium would be cooled, filtered, and

returned to its place surrounding the fountain and enabling it to shine. Some fission

fragments’ inability to pass through the helium filters, and others’ binding by the fluoride,

are what make requiring the fuel to boil itself reasonable: those fission fragments not

incorporated when it condenses are promptly trapped in a trap good enough to be their

final resting place.

5 Summary

Fuelling combustion only with pellets of aluminium or boron, small but not so small as

to be difficult to see and handle individually, can tame it. Compared to burners of oil in

air, internal boron combustion cars and external aluminium combustion ones would both

significantly advance safety, as spilled aluminium or boron pellets would not be able to

burn in the dilute, low-pressure oxygen we live on.

Non-combustion primary energies can supply aluminium and boron by extracting

them from their oxides. These would conveniently be available as tame combustors’

well-consolidated ash ingots, as the same use of purified oxygen that overcomes these

fuels’ ignition resistance also makes it easy to retain their ashes.

Cars would need machinery to make ingots and bins to hold them. They would need

oxygen deniers. These O-deniers would reduce the engines’ net efficiency by taking some

of their output, and they and the ash bins would make cars larger and heavier, other things

being equal, than oil burners.

5.1 Implications

The cars would advance environmental stewardship: ashes would not be emitted to the

atmosphere. Chemically unaltered air would carry waste heat away in a stream that would

probably be directed aft, and the oxygen denier would emit another relatively slow,

relatively narrow stream of slightly oxygen-depleted air, probably upwards.

The cars would advance ease of use. Letting the ashes of hydrocarbon motor fuels

simply blow away can be easier than dealing with ash ingots, but certain ways of dealing

with the fuels themselves, ways that would otherwise be convenient, would cause frequent

severe accidents. Like the severe but infrequent harm of post-crash fuel-fed fires, these

large risks can be reduced to zero, and the potential conveniences made actual, by a

switch to innocuous fuel that will not burn.

People who carry flasks of reserve fuel in their cars may appreciate the following

illustration. The low-fuel light has been on for some time. The car falters. Reaching into

the back seat without looking, one raises the reserve flask’s lid and scoops out a handful

of aluminium or boron pellets. The car falters again, and in haste one fumbles some of

the handful onto the cabin floor. The rest are successfully dropped into the in-cabin

refuelling port. They roll down its throat into the main fuel bin in the front bumper. No

further hesitation occurs; additional handfuls cause the low-fuel light to go out. The

dropped pellets can be let lie until one is at one’s destination, then gathered up and put

back in the reserve flask.

Compared to those of liquid hydrocarbon-burning cars, zero-local-emission cars’

stores of propulsion energy have been significantly less convenient to replenish and much

smaller. The bargain they have offered early adopters has not been cachet for a price, but

246 G.R.L. Cowan

cachet for two prices: pay more and tolerate more inconvenience. Early adopters of tame

combustion cars would also have to pay more, but convenience would become one of two

benefits rather than one of two costs.

Having two refuelling ports on a car, one inside the cabin and one outside, is one such

potential convenience. Another is the option of buying the car and its whole lifetime

supply of about 10 m3 of fuel at the same time. Neither geometry nor prudence would

forbid storing the fuel in one’s 200 m3 basement apartment, and refuelling at home.

The ash ingots would not be so compact. Home refuellers could accept the

inconvenience of packaging them and shipping them off to power stations, but fuel

retailers would soon see an opportunity in doing this for them. Since the aluminium or

boron pellets would be uniform, durable, and non-ignitable in air, a fuel bin wall

ventilated with holes too small for them to fit through would let a retailer blow them in

with an air hose. The ash ingots might need two hoses, one to vacuum and one to blow.

This and the fuel transfer could be done simultaneously. The whole three-hose process

could be quick.

5.2 Prediction

When developed, tame combustion cars will promptly take over the high end of the car

market and then push hydrocarbon-air combustion progressively down and out.


Ehrensberger, K., Palumbo, R., Larson, C. and Steinfeld, A. (1997) ‘Production of carbon

from CO2 with iron oxides and high-temperature solar energy’, Industrial & Engineering

Chemistry Research, Vol. 36, pp.645–648.

Lemmon, E.W., McLinden, M.O. and Friend, D.G. (2005) ‘Thermophysical properties of

fluid systems’ in P.J. Linstrom and W.G. Mallard (Eds), NIST Chemistry WebBook, NIST

Standard Reference Database Number 69, National Institute of Standards and Technology,

Gaithersburg MD, 20899, June. Available at:

Lemort, F. (1997) ‘Étude de la Séparation Actinides-Lanthanides des Déchets Nucléaires par

un Procédé Pyrochimique Nouveau’, Commissariat à l’Energie Atomique France Rapport

CEA-R-5760, pp.173–174. Available at:


Linstrom, P.J. and Mallard, W.G. (Eds) (2005) NIST Chemistry WebBook, NIST Standard

Reference Database Number 69, National Institute of Standards and Technology,

Gaithersburg MD, 20899, June. Available at:

Marchetti, C. (2006) ‘Long-term global vision of nuclear-produced hydrogen’, Int. J. Nuclear

Hydrogen Production and Applications, Vol. 1, pp.13–19. Available at: http://www.

Misra, A.K. and Whittenberger, J.D. (1987) ‘Fluoride salts and container materials for thermal

energy storage applications in the temperature range 973 to 1400 K’, NASA Technical

Memorandum 89913, AIAA-87-9226. Available at:

Mohai, I., Gál, L., Szépvölgyi, J., Gubicza, J. and Farkas, Z. (2007) ‘Synthesis of nanosized zinc

ferrites from liquid precursors in RF thermal plasma reactor’, Journal of the European

Ceramic Society, Vol. 27, pp.941–945.

Santos, J.C., Portugal, A.F., Magalhães, F.D. and Mendes, A. (2006) ‘Optimization of Medical

PSA Units for Oxygen Production’, Industrial & Engineering Chemistry Research, Vol. 45,


How fire can be tamed 247

Smith, R.A. and McBroom, R.B. (1999) ‘Physical Properties of Vitreous Boric Oxide’, article

‘Boron Oxides, Boric Acid and Borates’, Kirk-Othmer Encylopedia of Chemical Technology,

Table 4, 4th edition.

Welch, B.J. (1999) ‘Aluminum production paths in the new millennium’, JOM, Vol. 51, No. 5,

pp.24–28. Available at:


1 One of the things determining how many kilograms of oxygen an O-denier must supply per

kilogram a tame fire uses is the minimum allowed oxidiser fraction in the tame fire chamber,

represented below by t. Air’s 0.21 is by definition significantly less than t.

Supposing t_0.6, assuming nitrogen behaves as an inert gas, and setting the oxygen denier

product’s oxygen, argon, and nitrogen mole fractions xO2

, xAr, and xN2

respectively to 0.945,

0.042, and 0.013:

248 G.R.L. Cowan

yields a value near 1.096.

Nitrogen in the chamber does not behave exactly as an inert gas, however. Were it to show

complete non-inertness, every molecule combining with two oxygen molecules to form two

nitrogen dioxide molecules, it would make its volume and diluting effect zero by hiding itself

in the oxidiser. xN2

would disappear from the ratio expression:

For the same t and mole fractions, that is around 1.071.

A catalytic converter might be needed in the gas exit channel to deal with nitrogen dioxide.

It would have about 75 times less gas to scrub than its counterpart in an oil/air-burning car, so

a bed of alkali, although consumable, might serve well enough instead.

Keywords: alternative fuels; decarbonisation; energy carriers; energy storage;

fireproof fuels; fuel safety; hydrogen economy; hydrogen public acceptance;

hypoxia; ignition resistant fuels; nuclear production of motor fuels; solar

production of motor fuels; tame combustion; tame fire; zero emission vehicles.

Reference to this paper should be made as follows: Cowan, G.R.L. (2008) ‘How

fire can be tamed’, Int. J. Nuclear Hydrogen Production and Applications,

Vol. 1, No. 3, pp.235–248.

Biographical note: Graham Cowan obtained a programming diploma from the

Honeywell Institute in 1984. He is a researcher in clean vehicle propulsion. His

interests include solar and nuclear hydrogen production for use as motor fuel,

and direct nuclear propulsion of larger vehicles, including air-cooling of their

reactors without irradiation of the air, through the use of low-vapour-pressure

oxides as heat transfer media. He has published popular articles and a technical

conference paper on these same oxides’ potential to be manageable ashes in

combustion power systems.


scgi logoThe Science Council for Global Initiatives is a nonprofit 501(c)(3) charitable organization. All contributions are tax-deductible.
© 2023 The Science Council for Global Initiatives | We do not use cookies.

Sorry, this website uses features that your browser doesn’t support. Upgrade to a newer version of Firefox, Chrome, Safari, or Edge and you’ll be all set.