Mercury Thrusters: A Worldwide Disaster Averted Just In Time

The field of space vehicle design is obsessed with efficiency by necessity. The cost to do anything in space is astronomical, and also heavily tied to launch weight. Thus, any technology or technique that can bring those figures down is prime for exploitation.

In recent years, mercury thrusters promised to be one such technology. The only catch was the potentially-ruinous environmental cost. Today, we’ll look at the benefits of mercury thrusters, and how they came to be outlawed in short order.

Electric Thrust

As we’ve explored in our previous in-depth explainer, ion thrusters have proven valuable in innumerable space missions. Rather than using chemical reactions to generate thrust, they use electric fields to accelerate ions instead. Compared to traditional rockets, they can’t generate anywhere near as much thrust. However, they are far more fuel-efficient. This means they can generate far more delta-v (change in velocity) with the same amount of fuel.

NASA experimented with mercury-based ion thrusters on the SERT-I (pictured) and SERT-II spacecraft. However, mercury was deemed too toxic to use in future missions. Credit: NASA, public domain

Although their thrust is so meagre that you could never use one to launch a vehicle into orbit, they find their primary application in stationkeeping for satellites, helping them maintain position over time against the forces of upper-atmospheric drag. They can also be used to propel long-range probes that don’t have gravity to fight against.

These days, most thrusters use inert gases like xenon or krypton as fuel. However, these gases are expensive and their molecules are relatively lightweight. Mercury, on the other hand, is much heavier, still very easy to ionize, and easy to store on a spacecraft in liquid form. It’s also very, very, cheap. By sheer virtue of its toxicity, many industries are often stuck paying to dispose of mercury as a byproduct. The old saying that “you can’t even give it away” really does apply here.

The Problem

Mercury has a multitude of uses, such as the thermometer seen here. However, the silvery liquid metal is now used less often due to knowledge of its negative health effects. Credit: CambridgeBayWeather, public domain

While mercury makes an excellent ion thruster fuel on paper, its toxicity is too potent to ignore. Causing deletrious effects to the nervous system and brain, its presence in the environment can have major negative effects on human populations. From lowering IQs to damaging memory, it’s all bad all the way down.  It’s a toxin that accumulates in the body over time, and often enters the human body through the food chain. Indeed, mercury concentrations in many sea creatures mean that pregnant women are specifically advised to avoid many types of seafood.

For this reason, NASA abandoned the use of mercury as a propellant after initial experiments in the 1970s. Outside of contaminating the atmosphere, mercury comes with other risks too. There are occupational hazards for the crews working on the thrusters. Furthermore, explosions on the launchpad or crashes would spread the toxic material into the surrounding environment.

For these reasons, mercury was quickly considered a “dead fuel” by NASA, simply too dangerous to use despite the benefits.

Concerning Developments

NASA moved on to xenon-fuelled Hall effect thrusters after mercury was deemed too dangerous to use. Credit: NASA JPL, public domain

As is so often the case, however, a Silicon Valley startup was reported to be “disrupting” an established industry by rehashing an old idea. Bloomberg ran a story in 2018, regarding the activities of startup Apollo Fusion. Industry insiders told the outlet that the startup was shopping around a new thruster technology using mercury as a propellant.

This quickly set alarm bells ringing for many around the world. With SpaceX planning to launch over 10,000 satellites over a period of a few years, and many other companies rushing to establish their own massive satellite fleets, prospects were terrifying. If Apollo Fusion got a contract to equip thousands of satellites with mercury thrusters, widespread pollution of the entire Earth was suddenly on the table.

A scientific paper showed that a constellation of 2,000 satellites with 100 kg of propellant on board would deposit 20,000 kg of mercury into the upper atmosphere each year for a decade. Due to the weight of mercury ions, the majority would end up falling back to Earth, and account for 1% of existing global mercury emissions. Modelling suggested 75% of this mercury would end up in the world’s oceans, with negative impacts on marine life and fishing operations.

60 Starlink satellites seen prior to deployment in 2019.
Concerns abounded that if mercury thrusters were used for upcoming constellations of thousands of satellites, it could spread significant pollution into the atmosphere and around the world. Credit: SpaceX, public domain

Great effort has been expended over the decades to reduce the amount of mercury in the environment. The Minimata Convention on Mercury, a treaty from the United Nations, provided a framework for controlling mercury use by signatory countries. 128 countries signed the treaty, involving restrictions on the use of mercury in everything from batteries to lamps, soaps, and cosmetics.

At the time of signing in 2013, the idea of a return to mercury propulsion simply wasn’t on the table. Apollo Fusion wasn’t established until 2016. Worse, US regulations meant that there was precious little stopping any company that wished to launch mercury into space. Communication satellites fall under the jurisdiction of the Federal Communications Commission, which allowed satellite operators to self-certify their craft as having no deleterious impacts on humans or the environment.

A Safe Resolution

Thankfully, the hard work of scientists lobbying against the technology bore fruit. In March this year, the UN held a meeting regarding the Minamata Convention on Mercury, and adopted a resolution to phase out any use of mercury as a satellite propellant by 2025.

With most spacefaring nations being signatories to the convention, it makes the business case for mercury thrusters virtually unviable. As for Apollo Fusion, the company has stuck to working in the world of ion propulsion, though may have given up mercury propellants at this time. The company, which was acquired by American space launch company Astra, has since flown a xenon thruster in space as part of SpaceX’s Transporter-2 mission last year.

In any case, it seems that the thousands of satellites to be put in orbit in coming years will go up to space without mercury-spewing thrusters onboard. That should come as a great relief to all of us down here on Earth, where there is already more than enough mercury pollution as it is.


60 thoughts on “Mercury Thrusters: A Worldwide Disaster Averted Just In Time

  1. And again the baby is out with the bathwater. There’s no reason to ban them – if there is something better then everyone will use that. If there isn’t, then forcing the case just hurts the industry for virtually no difference for the environment or human health.

    >account for 1% of existing global mercury emissions.

    There’s over twice as much coming from natural sources than what people put in the atmosphere, further diminishing that number. Weathering of rocks, volcanoes, forest fires… these are all diffuse sources that don’t pose much of a hazard since metallic mercury and mercury oxides are rather inert and not readily absorbed. It’s methyl mercury that is the proper toxin, made by bacteria under anaerobic conditions. The problem with most human sources of mercury is that they’re highly concentrated, whereas most natural sources and the satellites would be diffuse sources, and therefore present negligible hazard to anyone.

    1. Yep, agree…. It is all in how you present it to the public. Scare tactics by the paranoia seem to throttle usage of ‘anything’ they deem ‘bad’ for you. So exaggerate/hype the situation as much as possible. Same with good o’ useful lead for example…. The sky isn’t falling, but listening to them, we should all be afraid of going out the o’ front door.

        1. He’s not wrong though.

          Quick search turns up with the claim that a pre-RoHS mobile phone would contain .3% per weight of lead, so one 100 gram phone would contain 300 mg of lead. A car battery contains about 10 kg of lead, so one single car battery dumped in the environment is equal to 33,333 lead-containing cellphones. There was little lead in electronics, and again it was metallic lead instead of organic lead, so it posed less of a health hazard. The real problem was with the waste handling and recycling, because they kept sending the e-waste to Africa.

          There was never a good reason to ban lead and specifically leaded solder in consumer products, and it actually isn’t banned in automotive applications, medical industry, space industry… However, getting lead banned meant compromises that made consumer products less reliable, which was good for the business of selling them, like hiring juvenile gangs with rocks is good business for the window salesman.

          Just as with the light bulb ban, CAFE standards, etc. the big corporations won’t allow governments to legislate against them unless there’s money to be made. When there is, usually it’s the largest industry players pushing for regulations that either harm the consumers, or their smaller competitors. This is called crony capitalism.

          1. “Compromises that made electronics less reliable” deliberately ignores that it was a *temporary* problem while the lead-free technology improved. The restriction required industry to innovate, which it did.

            But hey, we should just let industry be lazy and do whatever they want because heaven forbid there’s a minor inconvenience to pushing them off a local maximum.

          2. Lead-free technology didn’t entirely improve. Still we have problems with brittle joints, higher soldering temperatures, tin whiskers, tin pest, higher cost… it’s just a question of which problems you can live with and which you cannot.

          3. I bought some lead free car battery terminal clamps once. You know the ones that hold the thick wire and clip on to the lead battery terminals. They were marked as ROHS and lead free.

    2. The space industry seems to be doing quite well so what’s the gain for humanity in making it a bit more profitable by spraying the earth with 30 metric tonnes of mercury?

      1. Metallic mercury isnt readily absorbed into an organism. -Besides, I would imagine mercury would be deployed to interplanetary or geostationary missions, not to these spamsats in LEO.

    3. I mean the industry already wasn’t using it, so it’s not much of a loss

      A large chunk of the overall mercury emissions, partially included in the natural figure, are in fact reemissions, where mercury is cycled repeatedly instead of being safely deposited somewhere. The process that converts mercury into methylmercury isn’t fully understood, but it is known to be happening in waterways at a large scale, a larger scale than your comment implies. Regardless of how diffuse this methylmercury ends up being in the oceans, it tends to accumulate in fish as they consume contaminated vegetation, and more importantly it ends up being retained within the fish’s bodies. It’s reconcentrated to levels known to be potentially dangerous to consume, which is why there are recommendations limiting the consumption of certain kinds of fish already in effect. Methylmercury in fish has already been on an increase, so adding more mercury into the mercury cycle doesn’t seem like a good idea if fishing’s still going to be an important source of protein for a good chunk of the world population

      1. >I mean the industry already wasn’t using it, so it’s not much of a loss

        That’s kinda the point. If there is something better, then everyone will use that. They do – problem solved.

        The bigger problem is that governments and multinational organizations appropriate the right to dictate such bans by appealing to fear and irrationality – not for the sake of actually protecting the public, but to protect their own status and power to command the public as being necessary, and by proxy to secure funding. In other words, such bans are more for PR than effect. Do anything to be seen doing it.

        1. Come off it, change always costs time, money and effort, even if the end result is superior. If someone can make plenty of money already then why bother? The answer is they *don’t*, unless forced to.

          1. You wouldn’t have sped up the development of the automobile by banning the horse whip.

            The force is that doing something better is cheaper and more profitable – because it’s better, not because it’s legally mandated. Those who stick with the old stuff are competed out.

          2. @Dude but what happens when the “better” thing isn’t more profitable? When the best interest of the CEO is in conflict with the best interest of society, environment and overall long term economics? Take leaded petrol for example, everyone knew the dangers of lead from the start but the fact that it was cheaper for the industry to adopt than other alternatives meant they picked their own profits over the health of the general population. In my opinion legal regulations are necessary even if at times they may overstep the mark a tad.

      1. Higher performing thrusters would allow faster and further deep space missions. I think its silly to ban them for that purpose.

        I agree about not using them in earth orbit though.

        1. using satellites, and low altitude ones at that, as the reasoning for not using a thing its kind of silly i think. for geosynchronous and deep space mercury is completely fine. if the exhaust velocity is more than the orbital velocity of the spacecraft plus the escape velocity of the planet, then the mercury would be cast away from the earth, and if the exhaust velocity is higher than solar escape velocity, you wont have to worry about it ever again.

          should point out that we use some pretty nasty fuels for deep space stuff. clean fuels like lh2+lox are just not tank stable for long term missions, they boil off over time. anything hypergolic is going to be nasty. and nobody bats an eyelash at those. they are the perfect fuels for super simple pressure fed engines. various ion drives are starting to replace these thrusters, but they are still used fairly heavily in some situations.

          i have my doubts that liquid metal thrusters would be significantly better than things like hall thrusters in terms of specific impulse. generally the lighter the fuel molcule/atom, the better your isp. their only major advantage is their fuel density, so tankage can be very compact. so if you want a thruster for a cube sat, mercury would fit in the allotted space a lot better than xenon would. of course i wouldnt want this sat in leo, but for small form factor satellites for deep space exploration would be very nice to haves. missions to the outer planets could shotgun these small satellites on approach and get more data than the mothership alone.

          simply require that launch vehicle be of a very reliable type in order to launch these kinds of fuels, and same goes for ships with radioactive components (nobody gets their panties in a bunch about plutonium rtgs either). use something man rated with a reliable launch record and clear geo before lighting the mercury thruster.

        2. The problem is that you can’t get them into deep space without having them in earth orbit at some point. Even if you don’t use them till you’re away from earth, the entire period of the launch/orbit is a time that you have potentially houndreds of kilos of the stuff ready to fall back down in an uncontrolled manor due to some failure. Also, deep space in this case means between here and other planets in sol. I and probably most other reasonable folks would like to avoid a murcury laden probe crashing into a place we may like colonize one day because someone decided to use imperial units instead of metric. This isn’t like a RTG that only needs 15k or so of plutonium. We can make viable crash resistant sheilding for that small a quantity. 100k of mercury would need a massive container if was also crash resistant.

      2. Considering that our own sun is an open fusion reactor, I am not too concerned about putting toxic waste in space. Its all elements right? Might be dangerous to us but I doubt the gas giants or our sun are the least bit concerned. Don’t project what is bad for us on to the subject. Most places outside of earth are very dangerous to us, like immediately. Liquid methane and ammonia atmospheres to name a few.

    4. The mercury is stored in metallic form, but by definition leaves the ion thruster as highly charged ions. All sorts of interesting chemistry is bound to happen to these ions once they interact with CO2 and H2O in the atmosphere. I wouldn’t be surprised if there’s pathways that don’t lead back to elementary mercury but some some fraction ends up as organometallic mercury. Maybe someone who actually knows a thing about organometallic chemistry could pipe in?

      Imho still not necessary to ban it outright though. Would be entirely sufficient to ban its use in Earth orbit, and allow use of mercury thrusters eg for interplanetary probes.

    5. The modeling of where the exhaust goes was probably reasonable, but intuition asks how so much will be getting in the atmosphere. The exhaust velocity of an ion engine must be considerably greater than escape velocity of the solar system. Under what circumstances will a station keeping thruster ever be pointed at the atmosphere? And why would an ion engine powered probe be sending mercury to the Earth? I would worry more about contamination of the vehicle by mercury.

    6. It obviously didn’t hurt the industry too much, satellites are doing just fine. Governments and megacorps can afford xenon.

      We don’t know how many failed launches would have spewed mercury on us all.

      If innovation makes life more dangerous
      globally… what good is it?

  2. “If Apollo Fusion got a contract to equip thousands of satellites with mercury thrusters, widespread pollution of the entire Earth was suddenly on the table.” Facepalm!

  3. I somewhat agree with “Dude” in that it is foolish to completely abandon the fuel source. I think policies similar to those used for RTG’s would be appropriate. Or run a dual fuel system. A possible example being use Xenon when in Earth and Mars gravity but use mercury when transiting between the two systems. You could add the requirement that all inbound traffic dumps the fuel prior to entering the Earth-Moon System; similar to what most commercial ships do with their sewage prior to entering costal waters.

  4. > *Basic Math* showed that a constellation of 2,000 satellites with 100 kg of propellant on board would deposit 20,000 kg of mercury into the upper atmosphere each year for a decade.


    (100 kg * 2,000) / 10 years = 20,000 kg/year

    I don’t know the specs of that specific engine but Ion engines have high exhaust velocity; ie. “exhaust velocities around 20–50 km/s”

    Wouldn’t that kick it into high-enough orbit that the solar wind might “blow it away” rather than almost all of it falling back to earth?

    Or even from a delta-v perspective, wouldn’t it end up escaping earth orbit to a solar orbit if it was high enough?

      1. > For a typical Hall thruster with a mean exhaust
        velocity of about 15 km/s

        With that velocity, then yes, it will fall back to earth.

        But according to Wikipedia, that’s the low end of the scale for ion thruster exhaust.

        I wonder how much their estimate would change if the exit velocity exceeded what they used for the calculation.

  5. I wonder if bismuth would be a suitable substitute. Its density is 71% of mercury and its melting point is only 544°K (about double mercury’s of 234°K) so easier to liquify than lead for example. It’s toxicity is low enough that it’s used as a lead substitute.

    1. i think its been used before. it does come with the small issue that you need to heat the propellant to get it in a liquid state. so you will probibly also need include a small radioisotope heater as well, or electric heaters if its in your power budget.

      1. You’ve got a kilowatt available for the thruster anyway, and darned good vacuum for insulation: use the kilowatt to heat the tank first, then sustain it on waste heat once you light up the thruster.

        Or just feed it as wire to the thruster. Which, aside from reliability, is a great idea: no tankage required.

    2. Bismuth would be a brilliant substitute. 1% the cost of mercury. Doesn’t attack aluminum. Higher atomic number (so better Isp and higher thrust). 30% lower ionization potential (so more efficient). Non toxic. Spill cleanup is trivial. Can double as rounds for the point defence cannons and railgun :-)

  6. “… propel long-range probes that don’t have gravity to fight against.”

    I know what you meant but even long range probes ARE fighting gravity. It just might not be as strong as on the Earth’s surface. Thrusters also benefit from the large delta-V already imparted by their chemical cousins, such that their smallish thrust can add up to significant delta-V over a long period of time.

    1. Long range probes would be mor subject to the sun and other planets gravity and thus it would greatly reduce any amount of mercury deposited on earth. The vast majority would be swept outwards by the solar wind.

  7. Sounds like a mercury thruster would be ideal for a deep space probe. Getting the load of mercury to space safely isn’t a problem. Use the same design for containing it as was used for that RTG which was recovered and rebuilt from a rocket that blew up and landed in the Pacific. If it won’t lose radioactive material due to that kind of abuse, it wouldn’t leak mercury from the same.

  8. A couple questions come to mind here. 1. How much mercury is naturally existing in sea water? After all I would assume it is constantly present due to volcanic and erosive activities. 2. Is the mercury present in our soil naturally less dangerous? That is where it comes from in the first place right? 3. Is mercury in space more or less dangerous than all the other stuff there like ionizing radiation?

    I understand the launch failure concerns but would have to figure out how likely that is and how dangerous that situation is given all the other hazmat on space vehicles (sometimes including nasties like plutonium powered rtg).

    Seems like naturally occuring substances like mercury are dangerous because we concentrate them. Would dispersing them in the upper fringes of the atmosphere be any different than returning from a volcanic cloud?

    1. There are places where naturally occurring pure mercury can be found. I’ve done gold panning in mountain streams and creeks in Idaho where I’d have a small drop of mercury in the pan. Where did it come from? Leached out of cinnabar ore that’s very high in mercury content. Near Weiser, ID used to be a mercury mine that was the largest producer of the metal in the USA. Look up the Almaden mine, which was also a gold mine.

      1. I found mercury in gold panning a couple of times, and it could be possible that it was naturally occurring, but it was much more likely to be the tailings of older gold mining efforts, where the heavy sediment fron hydraulic mining was doped with mercury to separate out the gold in an amalgam. Now, this was in Northern California streams and creeks known to have been mined during the Gold Rush there, and it’s possible you were working sites not mined in the 1860 Idaho Gold Rush; I never found any mercury while panning in Idaho myself, but i was down in southern Idaho.

  9. Musk is not the only kid in town, just because Starlink is already approved for over 40,000 satellites, doesn’t mean competitors like Amazon don’t want their 3000+ up there too… or that they won’t try to get approved for a few dozen thousands more.

    1. Rather than ban mercury thrusters for everyone and all uses, you could demand the users to justify why they need as much and issue permits according to maximum allowed emissions.

      However, the point is rather moot since you’d have to get countries like China and Russia on-board and trust that they’re not cheating anyways.

      1. yeah that’s a great idea, more permits issued by government wonks that know so much more than everyone else and couldn’t possibly have political motives for messing with your business. No thank you.

  10. Well, looks like Mercury is off the menu, back to good old clean Hypergolics! That way you can dump 160,000kg into the atmosphere with a single launch rather than a mete 20,000kg for an entire multi-launch satellite constellation (those are rookie numbers!).

    1. Pretty sure you have your numbers scrambled there. There aren’t any satellites launching with 160 tons of hypergolic maneuvering fuel. There aren’t even any launch platforms that could lift that much fuel to orbit in a single launch.

      Kind of sounds like you might be confusing launch and maneuvering fuels. So just to be clear: mercury could never have been used to launch anything anyway, regardless of the legality. These are ion thrusters we’re talking about, and simply don’t work in atmosphere (nor would they have enough thrust if they did).

      And even for maneuvering, I doubt the preferred alternative to mercury-ion would be hypergolic. For a new design, the more likely the choice would be between mercury-ion and xenon-ion or krypton-ion. Hypergolics simply don’t make sense for something like, e.g., Starlink.

      1. The point is, all the hulabaloo is about mercury returning to earth. Launching something just releases all the crap at once instead of having it slowly rain down from the sky. Totally irrelevant if the contaminant is spaceborne or launchborne. The proverbial sheit hits the fan in both cases. Ergo the mention of hypergolics.

      2. “Kind of sounds like you might be confusing launch and maneuvering fuels”

        When you’re concerned with “is it dumped into the atmosphere” launch propellants are even more important than satellite RCS, as you are GUARANTEED to have launch propellants return to the atmosphere whereas satellite RCS plumes may end up in stable orbits themselves, or be ejected from cislunar space.

        The 160 tons was for a single Proton launch, a vehicle jam-packed with fun-juice of the ‘bucket of Delta V’ variety.

        1. But then the original statement (“”Well, looks like Mercury is off the menu, back to good old clean Hypergolics!”) simply makes no sense.

          There are already good alternatives to hypergolics in launch stages (Proton is an oddity). So banning Mercury thrusters doesn’t imply “going back to hypergolics”. Like, what is the logic there? Are you under the impression that banning mercury-ion maneuvering thrusters means, e.g., SpaceX will have to shelve their methalox plans and switch to hydrazine or something?

          Of course not.

          And to the extent that hypergolics are bad*, it’s almost like maybe they could and should be banned [i]too[/i]!

          * They’re certainly not [i]as[/i] bad as mercury, especially not pound for pound. Mercury is environmentally persistent and, in this scenario, would have been raining down globally. Hypergolics like hydrazine/dinitrogen tetroxide are plenty nasty, in their unburned state, but the effect of spills or unburned exhaust is mostly localized (and the compounds eventually break down). The actual combustion products do have some NOx and stuff, but the bulk is ordinary atmospheric gases like water vapor and nitrogen.

    2. False dichotomies do not do anything to minimize the damage that using mercury as a propellant would do both to the atmosphere and the “environment” of orbital space. Mercury does not break down into simpler molecules the way that chemical propellants do. The ions would damage the aluminum components of the any satellite or spacecraft that they encounter, including the one emitting it.
      The velocity of the ions doesn’t mean that they would mostly end up in interstellar space. Much of it would remain in orbit.
      Also, what is the effect of the stream of metallic ions on RF signals? It could have a deleterious effect on radio astronomy.
      And for those who claim that there was no kerfuffle about RTDs, you don’t have a very good memory.

  11. Yeah, only a little contamination at a slow rate so what.
    Problem is all those little bits add up making the planet hazardous to those life forms that survive our human caused extinction.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.