A drive that ran on mercury really flew

Start with the metal

If you have ever seen an old fever thermometer, a barometer in a hallway, or the silvery bead that rolls out of a broken tilt switch in a heater, you have met mercury. It is the one metal that stays liquid on an ordinary day. Caesium and gallium melt in a warm hand, at about 28 and 30 degrees, but they sit solid on a cool morning; mercury freezes at minus 38.83 degrees, colder than anywhere in Australia has ever recorded, and boils at 356.73 degrees, well past what a kitchen oven can reach.

It is also heavy for its size. A 375 millilitre can holds 5.1 kilograms of it, against 375 grams for the same can of water. A ten litre bucket of mercury weighs about 135 kilograms, which is why it was shipped in small steel flasks rather than drums. Drop a steel nut into it and the nut floats. Drop in a lump of lead and that floats too.

That combination, a metal you can pour, is why mercury turns up everywhere in the history of making things: pulling gold out of crushed rock, silvering mirrors, filling teeth, carrying the pressure reading in a hospital blood pressure cuff, glowing inside a fluorescent tube. It behaves like a liquid and conducts like a metal at the same time, and almost nothing else does.

How an ion drive actually works

Take an atom. Knock one electron off it and what is left carries an electric charge, so an electric field can grab it. Put two metal grids at the back of a small chamber, hold a couple of thousand volts across them, and every charged atom that drifts near gets flung out of the back at tens of kilometres a second. Throw enough of them and the chamber feels a push forward.

The push is tiny. The trick is that it never stops. A chemical rocket empties its tanks in minutes; an ion drive sips propellant for years, so the small push adds up into a large change in speed. Engineers call that total change delta-v, and it is the real currency of space travel: not how hard you shove, but how much speed you can buy with what you carry.

Two numbers describe any drive. Thrust is the push, measured in newtons. Specific impulse is how fast the exhaust leaves, quoted in seconds by convention; multiply it by 9.81 to get the exhaust speed in metres per second. A good chemical rocket runs about 320 seconds, which is 3.1 kilometres a second out the nozzle. The 1964 mercury engine ran about 4,900 seconds, which is 48 kilometres a second: Brisbane to Sydney in fifteen seconds.

The last piece is the bit that fails if you forget it. Fling positive ions out the back and the spacecraft is left negative, and it will simply pull its own exhaust back. So every ion drive carries a neutraliser: a hot filament or hollow cathode squirting electrons into the departing beam so the whole plume leaves electrically neutral. The 1964 flight was the first time anyone showed that worked outside a laboratory.

SERT I, 20 July 1964

A Scout X-4 lifted off from Wallops Flight Facility in Virginia at 10:53 UTC and threw a small NASA Lewis Research Center payload up to an apogee of 4,002 kilometres. It carried two engines. One was an eight centimetre caesium engine designed for 5.6 millinewtons at 8,050 seconds; a high voltage short circuit meant it never started. The other was a ten centimetre mercury engine, and about fourteen minutes into the flight it lit.

Inside it, a small boiler warmed liquid mercury and fed the vapour through a porous stainless steel plug into a chamber where a hot tantalum wire threw off electrons. Those electrons collided with mercury atoms and knocked electrons loose. Grids at 2,500 and 2,000 volts pulled the resulting ions out. It ran for 31 minutes on 1.4 kilowatts, produced 28 millinewtons, and rode out 53 high voltage recycle events, which is what the engineers called it when the grids arced and the supply had to blink off and back on.

Three independent systems measured the thrust, and all three agreed with what the same hardware had done on the ground. Nothing else on the spacecraft was disturbed by the beam. Those two results are the reason there is an ion drive on anything today.

SERT II, 3 February 1970: the endurance run

The follow-up went into a 1,000 kilometre polar orbit chosen so the spacecraft would sit in continuous sunlight and never lose power. It carried two identical fifteen centimetre mercury engines and what was then the largest solar array NASA had flown, five feet by forty. Each engine drew 0.85 kilowatts for 28 millinewtons at 4,200 seconds. Before launch, ground units had run 6,742 hours and 5,169 hours, plus a 2,400 hour test on a prototype spacecraft.

Thruster one started on 14 February 1970 and ran about five months, to 22 July, when a short across the ion optics stopped it. Thruster two started two days later and ran about three and a half months. Between them they logged 3,781 and 2,011 hours of thrust: 5,792 hours, or 241 days of continuous pushing, in orbit.

Then the spacecraft kept going. It was woken in 1973 to demonstrate cathode restarts, and again from 1979. NASA Glenn records about 18,000 hours of operation and more than 500 restarts across eleven years, until the propellant ran out in the spring of 1981. The whole eleven years used about 9 kilograms of mercury: 664 millilitres, less than two soft drink cans.

The mercury engines that were built and numbered

Mercury electron-bombardment engines were built from five centimetres across to a hundred and fifty, and ground endurance tests on them ran from about 4,000 hours to 15,040. A five centimetre unit called SIT-5 weighed 2.2 kilograms complete with its tank and feed system, drew 72 watts, and pushed with 2.1 millinewtons: about a fifth of the weight of a five-cent coin. It could hold 6.8 kilograms of mercury, enough for roughly 30,000 hours at full power.

An eight centimetre engine called IAPS passed every flight qualification test in the years to 1983 and was fitted to an Air Force technology satellite. The flight was cancelled for lack of funding and it never left the ground. A thirty centimetre design for a solar-powered transfer stage, SEPS, was taken to 2.6 kilowatts, 128 millinewtons and about 3,000 seconds; NASA spent roughly US$30 million on it over about ten years, an engineering model ran 10,000 hours, and seven advanced models accumulated 14,541 hours between them.

So the built mercury range runs roughly 2,500 to 4,900 seconds of specific impulse and 2 to 128 millinewtons of thrust. Every mercury drive that has flown or run on a stand is plumbing and voltage: a boiler, a discharge chamber, two charged grids and a neutraliser. The vortex is a different machine, and no run of one has been published.

Exhaust speeds are specific impulse multiplied by 9.81 m/s², rounded.

Watch a small push win

Give four drives the same spacecraft: 500 kilograms all up, with 80 kilograms of that as propellant. Fire them in a straight line with nothing pulling on them, and see who is going fastest. The chemical engine finishes in nine and a half minutes. The others are still going a year later.

Run it out and the order settles. The argon Hall thruster passes the chemical engine's total speed change after about 18 days, the xenon ion after about 34, the mercury ion after about 112. On distance covered the crossings come later, at about 37, 68 and 224 days, because the chemical craft spent its whole speed on day one and has been coasting ever since. By the five year mark the mercury ion has built 7,181 metres a second against the chemical engine's 547.

That last comparison is the whole argument for electric propulsion. Mercury's high exhaust speed is exactly why it was picked in the first place: the same 80 kilograms buys thirteen times the speed change.

What the family did next

Deep Space 1 launched on 24 October 1998 carrying a thirty centimetre xenon engine called NSTAR, throttleable from 19 to 92 millinewtons and 1,900 to 3,100 seconds. Ninety millinewtons is close to the weight of two sheets of A4 paper resting on your palm. It finished with 16,265 hours of running against a 14,000 hour design life, and gave a 486 kilogram spacecraft about 4.3 kilometres a second of speed change.

Dawn, launched 27 September 2007, carried three of the same engines and is the only spacecraft to have orbited two bodies beyond Earth, first Vesta and then Ceres. Its engines ran about 51,400 hours, used 425 kilograms of xenon, and built about 11.5 kilometres a second of speed change. To buy 11 kilometres a second from a chemical stage at 320 seconds you would have to start with about 33 times the mass you finish with. At 3,100 seconds you need about 1.44 times. That ratio is why Dawn could visit two destinations and a chemical Dawn could not.

Meanwhile the same idea went commercial and quiet. PAS-5, launched 27 August 1997, was the first working satellite to run on ion propulsion, using four small xenon thrusters to hold its slot in the sky. The larger XIPS-25 at 4.2 kilowatts gives 165 millinewtons at 3,800 seconds and keeps a satellite on station for about 5 kilograms of xenon a year, saving as much as 450 kilograms of chemical propellant off the launch mass. Starlink satellites fly Hall thrusters, a cousin design that traps electrons in a magnetic field instead of using grids: krypton from 2019, and from 27 February 2023 argon, at a published 170 millinewtons and 2,500 seconds on 4.2 kilowatts, from a thruster weighing 2.1 kilograms.

Why the field left mercury: a handling problem

Mercury's own numbers were good. It is heavier per atom than xenon, at 200.592 against 131.3, so each ion carries more push. It lets go of an electron more easily: 10.44 electron volts against xenon's 12.13, the electron volt being the small unit of energy that measures how tightly an atom holds its outermost electron, so less power is wasted making ions. And it stores as a liquid at ordinary pressure, so the tank is a can, not a pressure vessel.

The trouble is what happens after the beam leaves. Some fraction of the mercury never gets ionised, and some ions swap charge on the way out and drift off slowly. Both land on whatever is cold and nearby: solar arrays, radiators, camera optics. Xenon is a gas at any temperature a spacecraft reaches, so it lands on nothing. The same difference shows up on the ground, where a mercury vacuum chamber has to be decontaminated after every run and a xenon one does not.

In the 1980 timeframe it was decided to replace the mercury engines with xenon engines because xenon was less contaminating to spacecraft surfaces, and ground-test operations were greatly simplified.NASA technical memorandum TM-1999-209439, "A Historical Perspective of the NASA Lewis Electric Propulsion Program"

The propulsion numbers back that reading. On the same thirty centimetre hardware, mercury gave 128 millinewtons at about 3,000 seconds and xenon gave 92 millinewtons at 3,100. Mercury was not the weaker propellant. It was the harder one to live with.

Then costs moved. Xenon is rare and expensive to separate from air; krypton is cheaper, argon cheaper still at nearly one per cent of the atmosphere. The penalty is in the atom: argon weighs 39.95 units per atom against xenon's 131.3, and holds its electrons harder at 15.76 electron volts against 12.13, so more power goes into making each ion and less speed comes out per ion thrown. The ladder of flown propellants, mercury and caesium in the 1960s and 1970s, xenon through the 1990s and 2000s, krypton then argon in the 2020s, is a ladder of supply and handling, not of new physics.

Mercury nearly came back

In 2018 Bloomberg reported that a startup called Apollo Fusion was developing Hall thrusters running on liquid mercury, vaporised and ionised the same way as in 1964. The stated plan was up to 500 thrusters between 2019 and 2023, roughly ten tonnes of mercury over five years; critics put the possible release from mercury-fuelled satellite constellations at around twenty tonnes a year. The Minamata Convention on Mercury, which bans a long list of mercury uses, had nothing to say about spacecraft. Apollo Fusion withdrew the mercury plan in April 2021, and in March 2022 the parties to the convention adopted a resolution to phase out mercury as a satellite propellant by 2025.

The mercury passages in the old sky-craft texts

Two Sanskrit works carry the mercury material. The older is the Samarangana Sutradhara, attributed to King Bhoja of Dhar and usually placed in the eleventh century. Its chapter on made things describes a wooden craft with mercury pitchers set over iron ovens: the heavier variety carries four, the lighter one. The Sanskritist V. Raghavan discussed those passages directly. Heat mercury in vessels, the verses say, and the craft rises.

The younger is the Vaimanika Shastra, and its history is unusually well documented. Pandit Subbaraya Shastry of Anekal dictated about 3,000 verses to G. Venkatachala Sharma between 1918 and 1923, and said they had been delivered to him by the ancient sage Bharadvaja. One transcript found in Poona carries a signature and the date 9 August 1919. The manuscript was in the Rajakiya Sanskrit Library at Baroda by 1944. G.R. Josyer revealed its existence in a press release in 1952 and published the Sanskrit with an English translation in 1973. The technical drawings were made by T.K. Ellappa, a draughtsman at a Bangalore engineering college, working under Shastry's direction, and they appear only in the 1973 English edition.

The interesting part is what the text actually says about mercury. In the verses examined by a team at the Indian Institute of Science, mercury appears as an ingredient in metallurgy and in chemistry, not as a working fluid for propulsion. The floor board of the Shakuna craft is made of an alloy called raja loha, prepared from ammonium chloride, Bengal gram, benzoin, mercury borax, mica, silver and a mixture called panchamrita, all heated and poured. In the Sundara craft, one vessel used to make electricity is to be filled with three named plant and mineral ingredients soaked in elephant's urine mixed with mercury, and another with cow's urine. The propulsion described for that craft is not a mercury vortex at all: electricity from friction, heat, sunlight or falling water, used to vaporise oil, mixed with separately raised steam, and passed out through a pipe named for an elephant's trunk.

Where the mercury vortex engine actually comes from

The picture most people carry, a spinning bath of mercury driving a whirlwind, traces to two books published in the last decade of the twentieth century rather than to either Sanskrit source. W.D. Clendenon's Mercury: UFO Messenger of the Gods (1990) supplied the engine diagrams and a loose English rendering of the Samarangana Sutradhara. David Hatcher Childress's Vimana Aircraft of Ancient India and Atlantis (Adventures Unlimited Press, 1991) reprinted Josyer's translation and gave a chapter to mercury vortex propulsion and mercury gyroscopes. The Clendenon wording runs: inside the circular air frame, place the mercury engine with its electric or ultrasonic mercury boiler at the bottom centre, and the power latent in the mercury sets the driving whirlwind in motion.

Set that beside the Sanskrit and one thing stands out. The source says mercury is heated in vessels over fire. The phrases carrying the whole engineering claim, the ultrasonic boiler and the driving whirlwind, are twentieth century vocabulary. Which of them is translation and which is a modern engineering picture written into the gaps is a question a side-by-side reading against the critical edition of chapter 31 would answer, and nobody appears to have published one.

What the 1974 study measured

In 1974, five staff at the Indian Institute of Science in Bangalore worked through the Vaimanika Shastra as an engineering document. H.S. Mukunda, S.M. Deshpande, A. Prabhu and S.P. Govindaraju were in aeronautical engineering, H.R. Nagendra in mechanical. They interviewed G.V. Sharma, the man who wrote the verses down, and Venkatarama Shastry, Subbaraya Shastry's adopted son. They published in Scientific Opinion, pages 5 to 12.

What they did next was check the numbers. Four craft are described in enough detail to try. The results, in their arithmetic:

The finding they flagged as serious was not any single dimension. It was an absence: no weight is given anywhere, for any craft or any component, and the unit of mass does not appear in the text at all. For anything meant to leave the ground, weight is the number everything else answers to.

On dating they wrote that the work was brought into existence sometime between 1900 and 1922, and that the only evidence for Bharadvaja as its author is the text saying so. On the drawings they noted that Ellappa worked in an engineering college and the drawings point to a knowledge of modern machinery, and that the text and the drawings do not match each other even in theme. On the geometry they wrote that none of the craft has properties or capabilities of being flown, and that the principles of propulsion make them resist rather than assist flying.

The same paper settles one popular claim from the primary source. Shivkar Bapuji Talpade is often said to have flown a mercury-powered craft at Chowpatty in Mumbai in 1895. Shastry's own biography, as the 1974 team read it, says Talpade tried to make models under Shastry's guidance and did not succeed in making any of them fly. Shastry's autobiography also records that he was unsure of the practicality of the ideas in the work.

Forty-one years later, on 4 January 2015, a paper at the 102nd Indian Science Congress in Mumbai claimed Vedic-age aircraft able to reach other planets and to move backwards and sideways. Mukunda, by then long past the 1974 study, asked the plainest question available: if the claims hold, where are the working models? That question is still the one on the table, and it is answerable by anyone with a workshop.

For the craft themselves, and what a flying machine of that description would need in real units, see the vimana brief.

Red mercury, and the 1990s trade

The name surfaced in Soviet and Western media late in the 1980s. What it was said to be shifted with the telling. In 1992 New Scientist reported work at Lawrence Livermore National Laboratory describing it as a hoax that changed shape each time anyone looked. In 1993 the Russian newspaper Pravda described it as a superconductive material for high-precision explosives, stealth surfaces and self-guiding warheads. Two Channel 4 documentaries in 1993 and 1994 claimed evidence for a miniature device built around it.

The physical description stayed steady even as the story moved: a cherry to brick red semi-liquid, very dense, sometimes quoted above 20 grams per cubic centimetre. The formula usually attached to it is Hg₂Sb₂O₇, mercury pyroantimonate, a compound of mercury, antimony and oxygen.

The trade ran on those two claims. Asking prices sat between US$100,000 and US$300,000 a kilogram, which converts nominally to about A$139,000 to A$418,000 at 71.7 US cents to the Australian dollar on 28 August 2026, though those are 1990s dollars and are not inflation-adjusted. A British operation in September 2004 involved £900,000 for a single kilogram. A rumour that ran through Saudi Arabia in 2009 held that red mercury was inside the needles of Singer sewing machines, and machines changed hands at up to 200,000 riyals each. C.J. Chivers reported in the New York Times Magazine on 22 November 2015 that prices had climbed to hundreds of thousands of dollars a kilogram or more.

Bulk mercury traded around US$19 a kilogram in 2006. The asking prices were therefore between about five thousand and sixteen thousand times the price of the metal.

The claim as a spec

Two numbers carry the whole thing, and both can be checked without a laboratory.

Density above 20 grams per cubic centimetre. Take the ingredients one at a time. Mercury is 13.53 grams per cubic centimetre. Antimony is 6.685. Oxygen in a solid oxide is lighter again. If the volumes of the ingredients simply add, the compound lands between them, so a mercury pyroantimonate would sit well below 13.5. Run the formula Hg₂Sb₂O₇ through those three densities and it comes out between about 5 and 7 grams per cubic centimetre, depending on how tightly the oxygen packs in. Getting above 20 would mean the compound occupying far less room than its parts do apart, and no oxide has been measured doing that. It is a claim about packing, and X-ray diffraction settles it in an afternoon. The claimed figure is above uranium at 19.0 and plutonium at 19.8, which are elements, not mixtures of a heavy one with two lighter ones.

Energy release without a detonation. The claim carried a coined word, ballotechnic, meaning a material that releases very large energy when shocked rather than when detonated, enough to compress and ignite a fusion reaction directly. Real shock-triggered reactions in powder mixtures do exist, and they release a few kilojoules per gram, which is the same order as ordinary high explosives. To stand in for a fission trigger you would need something around 10⁸ kilojoules per gram. That gap is about twenty million times, and it is a gap in energy stored per atom, which is set by how tightly the atoms are bound. Chemistry works in the electron volts; that requirement lives in the millions of electron volts.

What the analyses returned

Around the trade, the record is thinner than the story. An International Atomic Energy Agency spokesman said in 2004 that red mercury does not exist. A trial opened at the Old Bailey in April 2006 and in July three men were acquitted of the related charges, with the prosecution conceding that whether the substance existed was beside the point, because the charge turned on intent.

Thirty-odd years of seizures have all been analysed and named, and every result so far has come back as mercury in one form or another. What no seizure has returned is a sample of the compound the claim names, Hg₂Sb₂O₇ itself. That compound can be made in a laboratory, and what it does under shock has never been published. The measurement that would settle it is the one nobody has taken.

Spinning liquid metal, for real

There is a machine that really does spin metal and produce current in the millions of amperes, and it has been around since 1831. Michael Faraday built the first one: a copper disc turning between the poles of a magnet, with current drawn from the centre and the rim. It is called a homopolar generator, and its personality is the opposite of what you might expect. The voltage is tiny, a few volts on a demonstration unit and a few hundred on a research machine. The current can pass a million amperes.

At those currents an ordinary carbon brush or slip ring wastes too much power, so designers use liquid metal instead: mercury, gallium, or a sodium-potassium mixture that stays liquid at room temperature. The liquid gives continuous contact with almost no resistance. The engineering problem is heat: the current warms the liquid metal, so it has to be pumped out, cooled and returned continuously, and a bubble in the contact ruins the run. A run of United States patents through the 1970s and 1980s covers exactly that: field-compensated collectors, barriers against circulating current, porous surfaces to hold the interface stable.

The Australian one

Sir Mark Oliphant started the design of what was then the largest homopolar generator in the world, at the Research School of Physical Sciences and Engineering at the Australian National University in Canberra. Four steel discs of 20 tonnes each were spun to 1,000 revolutions a minute. The machine stored up to 500 megajoules and delivered up to 2 megaamperes, with 1.8 million amperes measured. Initial tests were reported in Nature in 1962. It ran until it was pulled apart in 1986, feeding experiments the whole time. Its liquid metal contacts used the sodium-potassium mixture; after an accident with that, it was rebuilt with carbon brushes.

Five hundred megajoules is about the energy in two large electric-car batteries, or enough to boil about 880 full kettles. It comes out in a pulse. And 1.8 million amperes is roughly ninety thousand household twenty-amp circuits carrying their limit at once.

Mercury really was a working fluid

Between 1922 and 1968, power stations on the American grid ran on mercury vapour. William Le Roy Emmet at General Electric designed the cycle: boil mercury, run the vapour through a turbine, and condense it at a temperature still hot enough to raise steam for a second turbine underneath. Two engines stacked, sharing one lot of heat.

At Hartford in Connecticut the mercury entered the turbine at 471 degrees and condensed at 229, and that condensing heat raised 129,000 pounds of steam an hour. Overall efficiency reached 34.8 per cent, held across a continuous four month run, with about 43 per cent of the output from the mercury side and 57 per cent from the steam side. Plants ran at South Meadow in Hartford from 1928, at Kearny in New Jersey from 1933, at Schenectady in New York, and at Schiller Station in Portsmouth, New Hampshire, which was commissioned in 1950 at 40 megawatts and retired in 1968. Schiller reached a little over 37 per cent, the best of them.

The inventories were the problem. A plant held 20 to 50 tonnes of liquid mercury, and one unit described in 1931 held 90. Twenty tonnes is about 1,476 litres: seven bathtubs of it, moving through pipes and turbine seals. The cycle was abandoned because the capital cost was high, because leaks were dangerous, and because by the mid 1950s better materials let a plain steam plant match it. A modern combined-cycle gas plant reaches about 61 per cent, which is why nobody has gone back. When Schiller was decommissioned, the residue left in boilers, turbines, tanks and pipework needed a dedicated dismantling project of its own.

Mercury and the body

Mercury vapour is invisible and has no smell, and about 80 per cent of what you breathe in crosses into the blood. Skin takes up roughly one per cent of that, so breathing is the route that matters. Autopsy work points to a half-life for inorganic mercury in the human brain of about 27.4 years, meaning half of what arrives is still there decades later.

Here is the number that governs everything else about handling the metal. At 20 degrees, mercury's vapour pressure gives a saturated concentration in air of 13.2 milligrams per cubic metre. The level the United States National Institute for Occupational Safety and Health marks as immediately dangerous to life or health is 10. The workplace ceiling is 0.1, and the recommended limit averaged over a shift is 0.05.

So still air over an open pool of mercury, at ordinary room temperature, with no heating and nothing stirred, sits above the immediately-dangerous line and at about 132 times the workplace ceiling. A small bedroom of air, fully saturated, would hold about four tenths of a gram of mercury: less than one old fever thermometer, spread through the room.

The historical record is not abstract. At Minamata in Japan, the Chisso Corporation discharged methylmercury into the bay; more than 600 people died in Minamata alone, more than 21,000 filed claims and almost 3,000 were certified. New cases appeared at Niigata in 1965. The convention that now regulates mercury worldwide is named after that bay, and was adopted in 2013. In Iraq over 1971 and 1972, treated seed grain milled into bread produced at least 6,530 cases of poisoning and at least 459 deaths. At Saint Isaac's Cathedral in Saint Petersburg, 60 men died gilding the main dome, a process that drives mercury off a gold amalgam as vapour.

What would have to be true

Two claims, two sets of conditions, each with a number attached.

For a mercury drive that lifts

Electric drives are rated by how much push you get per kilowatt. The efficient ones cluster close together: SERT II gave 33 millinewtons per kilowatt, the NSTAR xenon engine 40, the XIPS-25 about 39, the Starlink argon Hall thruster 40.5, with the ground-tested SEPS reaching 49. Older and smaller units sit below that: SERT I managed 20, and the krypton Starlink thruster about 15. Call it 40, which is the generous end for anything asked to lift.

To hold one kilogram against Earth's gravity you need 9.81 newtons, which at 40 millinewtons per kilowatt takes 245 kilowatts. The best space solar arrays deliver roughly 150 watts per kilogram, so the array alone would weigh about 1,634 kilograms to hold up that one kilogram. To hover a one tonne craft you would need about 245 megawatts on board, which is the draw of about 350,000 Australian homes, and enough thrusters to make 9,810 newtons: about 350,000 of SERT II's size.

So a drive of the built kind would need a power source about three orders of magnitude lighter per watt than anything flown before it could lift its own weight. That is the condition, stated as a number: a power supply near 245 kilowatts per kilogram, against about 0.15 today. It is a power-source problem, not a thruster problem, and it is the same problem whatever the propellant. Mercury does not change it.

For red mercury

Three things would have to hold at once. A compound of mercury, antimony and oxygen would have to pack denser than mercury itself, which would require a packing arrangement nobody has demonstrated in any oxide. Its atoms would have to store energy about twenty million times more tightly than the best chemical bonds, which would put a chemical compound into the energy range of a nuclear one. And having done both, it would have to release that energy on a mechanical shock rather than on a chain reaction.

Any one of those would be a genuinely new result in materials science, and worth chasing on its own. The compound itself is not exotic and can be made; what has never been published is a measurement of what it does when you hit it hard.

What you would measure next

Everything below uses instruments that exist today, in laboratories that already run them for other work.

Of those six, the one that would change the most is the third. A mercury Hall thruster measured properly would tell you exactly how much of the 1980 decision was contamination and how much was the ground-testing inconvenience, and that answer matters for every propellant the field might pick up next.

Keep exploring

Sources

• Sovey, Rawlin and Patterson, "A Historical Perspective of the NASA Lewis Electric Propulsion Program", NASA/TM-1999-209439. Source of the SERT I and SERT II numbers, the mercury and caesium engine family, and the 1980 propellant decision · ntrs.nasa.gov/api/citations/19990116725
• "NASA's SERT II: A Genuine Space Success Story", NASA Glenn Research Center. Source of the eleven years, 18,000 hours and 500 restarts · nasa.gov/centers-and-facilities/glenn
• "Deep Space 1 validates the promise of ion thrusters", NASA Glenn, 2019, and mission summaries for the 16,265 hour total · nasa.gov/feature/glenn/2019
• The Planetary Society and DLR mission pages for Dawn's 51,400 hours, 425 kg of xenon and 11.5 km/s · planetary.org/articles/1476
• SpaceX published specifications for the argon Hall thruster on Starlink V2 mini, 27 February 2023, and contemporaneous reporting · spacenews.com/spacex-launches-first-upgraded-starlink-satellites
• Minamata Convention on Mercury, submission and 2022 resolution on satellite propulsion · minamataconvention.org, satellite propulsion compilation
• Mukunda, Deshpande, Nagendra, Prabhu and Govindaraju, "A Critical Study of the Work 'Vymanika Shastra'", Indian Institute of Science, in Scientific Opinion, 1974, pages 5 to 12 · hsmukunda.in, 1974 critical study
• Bhojadeva, Samarangana Sutradhara, chapter on mechanical contrivances and dharu vimanas, with V. Raghavan's discussion · academia.edu, mechanical contrivances and dharu vimanas
• W.D. Clendenon, Mercury: UFO Messenger of the Gods, 1990, and David Hatcher Childress, Vimana Aircraft of Ancient India and Atlantis, Adventures Unlimited Press, 1991
• Obhodas and colleagues, "Analysis of an object assumed to contain 'Red Mercury'", Nuclear Instruments and Methods in Physics Research B, volume 261, 2007, from page 922 · ui.adsabs.harvard.edu, 2007NIMPB.261..922O
• Jeffrey Lewis, "Curious case of red mercury", Arms Control Wonk, for the Promekologia decree and the customs analysis · armscontrolwonk.com/archive/207001
• C.J. Chivers, "The Doomsday Scam", New York Times Magazine, 22 November 2015 · longform.org/posts/the-doomsday-scam
• "The Large Homopolar Generator at Canberra: Initial Tests", Nature 195, page 113, 1962, and the ANU thesis on its brush design · nature.com/articles/195113a0
• Mercury vapour turbine plant records for Hartford, Kearny and Schiller Station, including the Schiller mercury boiler decommissioning project · manafort.com, Schiller mercury boiler decommissioning
• NIOSH chemical listing for mercury vapour exposure limits, and the NIOSH immediately dangerous to life or health documentation · cdc.gov/niosh/chemicals, 7439-97
• United States Environmental Protection Agency mercury compounds summary, December 2021, for reference doses · epa.gov, mercury compounds summary