I was looking through an old copy of Moon Miner's Manifesto a while ago and found the concept of a liquid airlock in an early issue.

schematic of air lock with a column of liquid

It is on page 32 of this issue from 1988. It posits this as an option in places with gravity, or simulated gravity on rotating space stations, and suggested 3 possible liquids that could be used: mercury, gallium, or a sodium-potassium alloy known as NaK.

At first glance it seems like it could be really useful in certain situations. I wonder what the likely problems and limitations would be.

In a pinch it even seems like it would be a faster way to get to safety - if your space suit could deal with immersion in liquid. I suppose it would be a good idea to put it inside an alcove or shed, and maybe have a lid on it for when it isn't being used. What else needs to be considered?

  • $\begingroup$ What's the pressure under all that water? Space suits usually handle pressures between zero and one atmosphere. $\endgroup$
    – Samuel
    Aug 25, 2015 at 18:05
  • $\begingroup$ @Samuel: Seems likely the pressure would be only a bit over 1 atm, depending on the internal pressure of the station. If you decide to use, say, 40% O2 in a 0.6 atm station, the pressure wouldn't even be that high. After all, the point is that at the lower surface, the pressure from the water column matches that of the atmosphere, and the distance from the lower surface to the middle of the tunnel out is a lot less than that between the surfaces. $\endgroup$ Aug 26, 2015 at 1:45
  • 1
    $\begingroup$ The Na K solution puzzles me, Na and K are both alcalic metals and I don't see them forming a salt together. Actually, it's not a solution of NaK but pure NaK which apparantly has a low melting point. $\endgroup$
    – mart
    Nov 9, 2015 at 14:25
  • 1
    $\begingroup$ @mart Hey you are absolutely correct and I submitted an edit suggestion to correct that. NaK is an alloy of sodium and potassium and it is liquid at room temperature (it is common for metal alloys to have a lower melting point than all of their constituents). Calling it a "solution of NaK" is confusing, ambiguous, and actually incorrect: it implicitly suggests presence of a third substance, a solvent, in which Na and K were supposedly dissolved. However, all there is are just two ingredients: Na and K metals, dissolved in one another. $\endgroup$ May 18, 2022 at 0:22

6 Answers 6


The diagram you show would work only for specific instances. A few things that it relies on:

  1. Gravity is required for this to work, or else all of the liquid will escape. The diagram you show works because of pressure differential, basically the pressure due to gravity counteracts the pressure from the air. Without gravity, the air would simply push the liquid completely out.
  2. The "Vacuum" must be an enclosed surface, or else the liquid will all boil away. The enclosed surface will then become composed of a thin (at best) atmosphere of the liquid. In a pure vacuum, such as space, the pressure won't be high enough to keep the liquid in its liquid state. See this question on Physics.SE. One could theoretically keep the a lid on the liquid, which would be sufficient to keep it from boiling, until one actually wishes to use the airlock. This would allow for an airlock of sorts.

Thinking about this, the only places I could really see this potentially working is on Mars or Titan, both of which have a thin atmosphere. Titan in particular might be of interest as its atmosphere could be explosive if combined with oxygen, and this would be an effective way to keep the oxygen away from Titan's atmosphere when performing an EVA.

There is an option that could work. Gallium is a relatively toxic free material. Its melting point is near room temperature, so it would be reasonable to survive going through it, although a human could very well have parts solidified to one's suit after passing through it, as they would freeze on it (Unless the suit was pre-warmed). It is a fairly expensive material, and the amount required to use for the space application would be quite large.

  • $\begingroup$ The original article mentions mercury, gallium, and NaK. I had this question sitting around for a while and didn't go back over that article - i really should have. To be fair to it, i should probably edit the question to clarify that, and that it posited this as a possibility for places with gravity. This is helpful though. $\endgroup$
    – kim holder
    Aug 25, 2015 at 2:39
  • 2
    $\begingroup$ You are correct that all liquids will eventually boil away, but depending on the vapour pressure this can take so long as to actually be usable in an airlock. See (space.stackexchange.com/a/10755/3372) for further explanation. $\endgroup$
    – Marius
    Aug 25, 2015 at 12:59
  • 4
    $\begingroup$ Maybe add a thought on Gallium: The melting point is 30C. While that is a bit of a problem for thermal management, it does open the possibility to purposely cool the vacuum surface, and heat it up only when the airlock is in use. That way, the liquid would not boil off. $\endgroup$ Aug 25, 2015 at 22:30
  • 5
    $\begingroup$ Gallium has the nasty property of dissolving aluminium. $\endgroup$
    – MSalters
    Nov 9, 2015 at 9:19
  • 1
    $\begingroup$ NaK is nasty stuff. It's a eutectic combination of sodium and potassium that is liquid at room temperature but doesn't play well with moisture and oxygen. $\endgroup$ Nov 14, 2015 at 3:21

The only reason to have an airlock is to move through it: if you're not going to move through it, you'd just replace the whole thing with a steel plate.

  • NaK is a total non-starter for this. It is extremely reactive so anything passing through it would need to be extremely well protected. It tends to catch fire when exposed to air, so the surfacing pool needs to be carefully designed to avoid this. If you're going to pass through another liquid first, you need to make sure there are no air bubbles.

  • Mercury is a non-starter because of its extremely high density: you'd have to transport a huge mass of it to the Moon and moving things through the mercury would require you to deal with huge buoyancy forces.

  • Gallium seems like a non-starter because of its rarity. Wikipedia quotes annual world production at a few hundred tons a year and you're going to need tens of cubic metres of the stuff, at a density of 5.9 tons/m3 and a price of a few hundred dollars a kilogram in today's market, but a whole lot more than that if NASA decides to buy up half the world's production. It would also need to be kept warm: its melting point is just under 30 °C.

That seems to rule out all the options that are currently on the table.

  • 2
    $\begingroup$ Why not water? It's possibly available on the moon, and likely to be found in quantity anywhere humans decide to colonise. $\endgroup$
    – NPSF3000
    Aug 26, 2015 at 1:28
  • 2
    $\begingroup$ @NPSF3000 Water will boil in a vacuum. $\endgroup$
    – user253751
    Aug 26, 2015 at 6:05
  • 2
    $\begingroup$ @user253751 All liquids will boil in vacuum. Liquids boil once the ambient pressure is equal to or drops below liquid's vapour pressure, and in vacuum the ambient pressure is zero; meanwhile, all liquids have a non-zero vapor pressure. $\endgroup$ May 17, 2022 at 22:20
  • 3
    $\begingroup$ Not all liquids boil in vacuum. Boiling involves bubble formation, and the contents of a bubble are not zero pressure. Some liquids with very low vapor pressures have enough surface tension to prevent bubbles from forming, and will only slowly evaporate from their surface, similar to solids (which also have non-zero vapor pressures). Silicone oils and polyphenyl ethers are routinely used in ultra-high vacuum applications, even in the pumps used to achieve such vacuums: en.wikipedia.org/wiki/Diffusion_pump $\endgroup$ May 18, 2022 at 0:07
  • 1
    $\begingroup$ @ChristopherJamesHuff All liquids evaporate away in a vacuum. The only question is how fast--you're listing some things that evaporate slow enough to work. $\endgroup$ May 18, 2022 at 1:28

In order for a liquid airlock to work, some points must be considered; they have been mentioned in some of the other answers, but I will try to combine them and suggest some new implementations.

  • Gravity:

    As stated, there must be some gravity (or constant acceleration) present for this to work in principle. The required column height (see sketch in the question) of the pressure head is $h = p/(\rho \cdot g)$.

    To compensate for a 1 atm habitat pressure on the Moon, a water column would need to be about 60 m high, and a mercury column 4.5 m. On Earth, 1 atm corresponds to only 760 mmHg (millimeters of mercury).

  • Vapour pressure of the liquid vs. pressure of the atmosphere (if any):

    The rate of evaporation depends on the external pressure as well as the intrinsic vapour pressure of the liquid.

    Liquid metals have low vapour pressures, which is why they were suggested in the original article. Water, on the other hand, has a much higher vapour pressure and will evaporate / freeze and sublimate very quickly in space or on a body without a significant atmosphere, e.g. the Moon.

    There are some other possibilities for liquids from this stand point; silicone oils, for example, are used in vacuum applications because they can have low vapour pressure. Quite recently (compared to the article's publication date) room temperature ionic liquids have been studied which have very low vapour pressures and have also been used in vacuum applications, e.g. in this article.

    In any case there will always be a small rate of evaporation on a body without any atmosphere, which means you should top up your air lock once in a while.

  • Practicalities of passing through the liquid airlock, buoyancy forces and reactivity of the liquid:

    Liquid metals like mercury and gallium tend to be very reactive and will form solutions with most other metals. NaK is very corrosive as well. This could presumably be solved by coating exposed portions of cargo with inert polymers like PTFE. Silicone oils, on the other hand, are very inert and should pose no problem at all.

    David Richerby mentioned that buoyancy can also be a problem, especially with the high density of metals. Buoyancy is proportional to gravitational acceleration (see xkcd's What If) but diving through roughly 5 m of mercury in your non-metallic space suit on the Moon would be next to impossible unless you had big tungsten or platinum ankle weights).

So, to sum up, it would be pretty difficult to make as well as use a liquid airlock. The volumes and masses of the liquid alone would be way too great to carry even to the Moon. Maybe you could extract some of the resources in-situ, however, to me it seems this type of an airlock is totally impractical, unless you already have a well developed industry on another planet. Although, it is a nice physics and chemistry problem to think about.

  • $\begingroup$ (Also, you don't need to dive: a machine could be used to move things through the airlock.) $\endgroup$ Aug 25, 2015 at 16:49
  • $\begingroup$ @David Richerby using a type of elevator inside the fluid shaft is a good idea. Unfortunately it makes the whole thing even more complicated. Concerning buoyancy: the total buoyant force on a body is its mass times the difference of fluid and body density. This is independent of local gravitational acceleration. It is this force you have to overcome by swimming or with an elevator type machine and it will present the same challenge as on Earth. $\endgroup$
    – Marius
    Aug 25, 2015 at 18:31
  • 3
    $\begingroup$ "the total buoyant force on a body is its mass times the difference of fluid and body density" No. Mass times density has dimension $[\mathrm{mass}]^2[\mathrm{length}]^{-3}$. That cannot possibly be a force because force has dimension $[\mathrm{mass}][\mathrm{length}][\mathrm{time}]^{-2}$. It's easy to see that the buoyancy force depends on the local gravitational field: if there was no gravity, the fluid wouldn't care whether you were inside it, partly inside it, floating next to it, whatever. The buoyancy force exists only because you've displaced fluid by lifting it against gravity. $\endgroup$ Aug 25, 2015 at 19:10
  • 1
    $\begingroup$ You are of course right, it is proportional to local g. I changed my answer accordingly. $\endgroup$
    – Marius
    Aug 26, 2015 at 7:08
  • 2
    $\begingroup$ I'm starting to think the idea of using a liquid metal airlock was done for two reasons: because the person wrongly felt metals were gastight, and because it seemed "cool". Water or sillicone probably weren't mentioned because it is known gasses will diffuse in and out of them. Of course, metals (even solid ones) are not gas tight. Paladium, for instance, will allow hydrogen gas to diffuse right through it. $\endgroup$
    – Dan S
    Aug 26, 2015 at 21:45

In terms of protecting some kind of habitat viable for humans you can eliminate NaK and Hg.

It is incorrect to say NaK is pyrophoric or reacts violently with air (I have worked with NaK). But it does react very violently with water. It also would consume the oxygen of the habitat over time (and you would form salts from these reactions which would not be fluid in nature). Basically, you'd have a zero humidity habitat on the other side of the airlock and over time the oxygen would be consumed as well. NaK is basically a substance that only behaves nicely in very controlled and isolated environments.

A human habitat sealed in by liquid mercury would also be a habitat where everyone lost their hair, fingernails, went insane, and died prematurely... Mercury has a low vapor pressure but that vapor pressure is vastly greater than what is required to slow poison most mammals.

I don't know what the health factors with Gallium are, but out of the three it would be the only one that you could maintain a viable habitat on the other side with.


There are several answers now that examine different aspects of this. I wanted to add the matter of waterglass, (sodium silicate Na2SiO3, or potassium silicate K2O3Si) as a potential liquid for use in such a design. A liquid solution is roughly half water. Though water is volatile, the vapor pressure (page 10 here) of waterglass is very low. With a cover over the exit of the airlock, there would be no vapor loss issue other than a scum developing on the surface as the water there evaporated. It is non-toxic, non-reactive, and the elements for it are all found on the Moon. (Sodium is much more widely available and plentiful than potassium, oxygen and silicon bound in minerals make up about 2/3 of the Moon's surface.)

It is a clear, somewhat viscous liquid with a density of about 1.3 g/cm3 in a half-water mix. Thus a person in a space suit would have little trouble moving through it. The column needed to balance a full atmosphere of pressure would be about 8 m.

  • $\begingroup$ This was mentioned to me by TildalWave, so props to him. $\endgroup$
    – kim holder
    Aug 27, 2015 at 15:41
  • $\begingroup$ Interesting. My first thought was ionic fluids as I've just been reading about how they are proposed for making liquid mirror telescopes on the Moon due to their almost zero vapor pressure. See Liquid Mirror Telescopes on the Moon and Ionic liquids. Main problem would be their high viscosity and there's research into low viscosity ionic fluids. So quite closely related to water glass. $\endgroup$ Apr 13, 2016 at 21:54
  • $\begingroup$ Kim, I've just checked the source, it says "The vapour pressures that have been measured for three solid sodium silicates are extremely low: 0.0103 hPa at 1175 °C (MR 1.0, metasilicate), 0.0031 hPa at 1165 °C (MR 2.0) and 0.0016 hPa at 1172°C (MR 3.0). This indicates that the respective pressures at ambient temperature will be unmeasurably small. The penta- and nonahydrates of sodium metasilicate contain significant amounts of hydration water (pentahydrate: 43 %; nonahydrate: 57 %). In commercial silicate solutions the water content is still higher and can reach up to 70 %. $\endgroup$ Mar 6, 2017 at 1:16
  • $\begingroup$ "Therefore, the vapour pressures of the solid hydrates and the solutions are expected to be significantly higher. However, this would be governed by the high water content and reflect rather the vapour pressure of water than that of the respective silicates." So if I read that right, the water would evaporate from the waterglass in a vacuum, just leaving the solid silicate behind. $\endgroup$ Mar 6, 2017 at 1:17
  • $\begingroup$ Another idea I had just now is that you could use a low density immiscible ionic fluid. It would float on the top of the water as a thin layer and so protect the water from evaporation. That way, it wouldn't matter if the ionic fluid has a high viscosity. I've just added this idea to my Case for Moon First $\endgroup$ Mar 6, 2017 at 1:28

A potential solution would be an heterogeneous solution. A denser solution (e.g. water) covered with a sufficiently deep layer of a less dense solution. Under a gravity field such as the Moon's this would separate out into two dissimilar layers (assuming there was no significant solubility of one solution in the other). The lighter layer would have to have a low vapour pressure. An oil such as (honestly) olive oil or kerosene would have much lower vapour pressures than water and as such would evaporate significantly more slowly.

In fact when there were believed to be canals on Mars several authors suggested that the canals would be protected from the smaller atmospheric pressure (assumed due to the lower mass of Mars) causing excessive water loss by floating oils on the surface.

Whilst this would be insufficient in itself to stop vapour loss, a thermal condenser could be placed above the solution which would cool vapour losses and condense them back to liquid which would dribble back into the original lock. This could be powered on the moon by simply keeping it in the shade — no Maxwell's Demon would be required!

The actual exit in the low pressure region would be effected by opening the pressure equilibrated cap, leaving, and closing it. Why a cap? Because pressure equilibration could occur through a small diameter tube which would allow a small thermal condenser.

  • 2
    $\begingroup$ @SimonN, if there is anything you don't like about an edit to your post, change it. $\endgroup$
    – Joe
    Aug 26, 2015 at 3:24
  • 2
    $\begingroup$ Agreed. If you do not like an edit, change it yourself. $\endgroup$ Aug 26, 2015 at 15:51

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.