I am currently building a world where the majority of the population is some kind of amphibian species. My question is : Is it possible to design a kind of membrane that is completely waterproof but can let pass organisms that are sufficiently large (see the gungans in SW : the phantom menace ? that's the idea) and how would one implement it given the laws of physics we know ?

The membrane can be tangible or more like a force field, I don't see any problem.

Thanks in advance.

edit: the membrane must be translucent, as thin as possible and roughly the same size as a regular human door. I think the fish should not be able to pass through. my species is living in buildings underwater that can be filled with water or emptied at will. The membrane would serve as a door to have a smooth transition between either sea and building or different rooms of one building.

  • 31
    $\begingroup$ How about using preasure and gravity like in a diving bell? $\endgroup$ – lijat Apr 3 '18 at 14:36
  • 2
    $\begingroup$ This is an interesting question, but could you maybe state the premise a bit better? Few people know that movie by heart (since it was more like another (less popular) part of the body). Does this membrane surround an entire city deep under water and human-sized people can swim through it but fish can't or was it just over some spaceship/spacesubmarine in shallow waters or what exactly? $\endgroup$ – Raditz_35 Apr 3 '18 at 14:38
  • $\begingroup$ @Peamcy, I see you've marked an answer as accepted. While mart's answer is a very good one, we typically recommend that you wait at least 24 hours to accept an answer to give enough time for people from several time zones to respond. Marking an answer as accepted will discourage others from contributing, and you could lose some great ideas! $\endgroup$ – bendl Apr 3 '18 at 19:20
  • 1
    $\begingroup$ @Len He is proposing using the same solution that is used in the real world. If you have the same pressure of air inside and water outside, gravity will stop water from rising up from a hole in the bottom of the structure as water is denser than air. So only top and sides need to be air and water proof, the bottom can be open and provide easy access. $\endgroup$ – Ville Niemi Apr 4 '18 at 14:33
  • 1
    $\begingroup$ Someone should make the diving bell solution an answer. $\endgroup$ – bendl Apr 4 '18 at 14:56

12 Answers 12


When I was a child, we had a toy which was apparently called a Water Snake (pro tip, don't google water filled sphincter toy, ever). Water Snake Toy

The tunnel through the middle should allow a creature to pass through, but, if kept pressurized, the mouth of the tunnel should always remain closed tightly.

Use one of these, with a rope through the middle to allow the creature to pull themselves through. It will have to be pressurized enough to keep water out, but low pressure enough to not crush the creature inside. This should allow 2-way access while being impermeable.

It could be quite difficult to pull your way through this tunnel, so you could either coat the tunnel with a slippery hydrophobic substance that will lower the friction, or run the contraption on bearings, so that the whole thing can turn inside out on itself, letting the creature enter or exit easily.

Finally, if the contraption were motorized, then it would make the process much easier for the creature. It would work sort of like a weird 3D conveyor belt to suck the creature in from outside. All you need to do is get yourself stuck in it.

What is a Water Snake?

A water snake is a flexible membrane shaped into a torroid and filled with a liquid. The upshot is that, because the membrane is flexible, it can turn itself inside out by rotating around the center. Imagine a two gears meshing, and then revolve that around the point where the gears mesh to get a 3d object. The entire outer surface moves one direction while the inner surface moves the other direction.

  • 21
    $\begingroup$ This answer simultaneously grosses me out and makes me giggle $\endgroup$ – Len Apr 3 '18 at 19:27
  • 1
    $\begingroup$ @SilverCookies is the edit clear? $\endgroup$ – bendl Apr 3 '18 at 20:23
  • 3
    $\begingroup$ A bit like a peristaltic pump $\endgroup$ – StackOverthrow Apr 3 '18 at 22:04
  • 6
    $\begingroup$ This sounds a lot like a more complicated version of an airlock. The membrane part doesn't get you anything is this case, because if you rotate the hole to allow an organism through, the hole is going to let water coming flowing through, too. (And if it's an undersea thing like the OP referenced, that's going to be very high pressure water, meaning a lot of it, moving fast enough to be devastating to whatever's on the other side... so, worse than an airlock.) $\endgroup$ – HopelessN00b Apr 4 '18 at 6:18
  • 5
    $\begingroup$ Unfortunately, while fascinating, it needs a significant width, and is not going to work. It is a 3-D version of a gear: imagine two water gears locked together. Do they stop the water flowing? No, they start rotating. The same would happen with this tube. $\endgroup$ – LSerni Apr 4 '18 at 14:11

What you are looking for is a moon pool: It's an opening in the floor of the underwater habitat, that connects the sea with the pressurized air inside the habitat. The air pressure keeps the water from filling the room. This is as simple and low-tech as you can get.

In principle, this works as far down as pressurized air diving works: The limiting factor is the toxicity of the pressurized gases to the body. Afaik, this is somewhere around a depth around 40m for a human with normal air mixture (I couldn't find any precise numbers quickly). If you are willing to go with some other air mixtures (as they are used when people dive to greater depths), you can push that to somewhere short of 100 meters (again, no precise number, just something I vaguely remember). If you don't need to take humans through your moon pools, just your adapted species, you can make pretty much any depth believable.

But, I guess, maximum depth is not your main concern anyway, as your amphibious species will want to live in conditions close to what they naturally live at, which is near the surface.

  • $\begingroup$ Couldn't you combine the moon pool concept with an airlock system? That way there is an external moon pool with pressurized atmosphere, but the remainder of the underwater complex can have a different pressure. This could allow humans to enter since they would not need to remain inside the airlock room for too long. $\endgroup$ – mbomb007 Apr 5 '18 at 18:19
  • $\begingroup$ @mbomb007 Yes, you could. Even better: Use a waterlock instead of an airlock. Water is near incompressible, air is very compressible. So, evacuating/pressurizing an airlock takes a lot of energy (you have to move a lot of air in/out), (de-)pressurizing a waterlock takes very little (almost no water has to be moved). However, as we are talking about low-interference barrier crossing, and an amphibious species that most likely prefers not so great depths, I felt that the pressurized interior is the easier way to go. It avoids any doors. $\endgroup$ – cmaster - reinstate monica Apr 5 '18 at 18:39
  • $\begingroup$ This is probably the best answer, and it's based on real-life physics. Unlike an air or waterlock system, which requires time & equalization between the two differentials, a moon pool represents a single surface very similar to the membrane concept described by the OP. Great answer! $\endgroup$ – Omegacron Apr 5 '18 at 21:46
  • $\begingroup$ @mbomb007, that's shown in the Wikipedia illustration, situation C. $\endgroup$ – Wildcard Apr 6 '18 at 4:07

This answer tries to design a system that could work with far future technology (that we don't know if it will ever exist!), but without inventing things contemporary physics rule out, like force fields.

The membrane is a sort of smart liquid or nanotech slime, consisting of many interlocking macromolecules or similar. They are all very fatty and hydrophobic. When a solid object touches the membrane, the individual parts intelligently make way and slide over the surface of the object as it passes.

A high trade off has to be made about surface tension: Too small, and parts could easily tear off or enter very small pores and orifices on the passing solids. Too large, and the membrane can't form a tight seal in sharp corners.

There's another issue: someone could press a donut against the membrane. When the membrane retreats around the donut, the inner whole is open and you have a leak. When the membrane separates around the donut, the inner plug might be pushed in by water pressure - unless the plug can hold fast to the donut, but the membrane hast to glide about surface so it has to be really smart to know when to glide, when to hold, when to separate and when to remain one blob.

The question about the donut is not purely academic, every piece of clothing has a similar topology. So I suggest the following: So upon contact, the mebrane material will flow around the object and try to form a cavern. This will involve tendrils fingering up your ear or trouserleg! Then parts of the membrane will perform as a pump to remove water, shrinking the cavern. Once the cavern is sufficiently dry, the cavern will open to the other side. The simpler the object passing through, the faster this can work.

The membrane has to be thick to withstand pressure without causing excessive tension in the frame, unless interor rooms are the same pressure. But then you can have houses like diving bells, without doors ...

This has ramifications: The membrane (that is so thick, it should be called blobgate or maybe bloor (portmanteau of door and blob)) can easily kill anyone passing through, so control of the software (that is probably hardcoded into the individual nanotech components) will be an issue. The technology will have many other applications in medicine, toys, manufacturing ...

  • 1
    $\begingroup$ Could a mass of miniature robots work? They could perhaps use a stirling-engine effect for its electricity and simply interlock against the waterpressure, as someone tries to pass they open up around that prerson. With a 1m thick wall you do have a membrane of sorts but avoid things like yhe donut problem. $\endgroup$ – Demigan Apr 3 '18 at 15:42
  • 3
    $\begingroup$ The membrane has to be thick to withstand pressure -- actually, not. This is a brilliant answer: just have several membranes one after the other. Each membrane only withstands 1/Nth of the pressure. This is how heavy duty magnetic seals work. $\endgroup$ – LSerni Apr 3 '18 at 17:55
  • $\begingroup$ @Demigan, stirling-engine effect? Please tell more... $\endgroup$ – Len Apr 3 '18 at 19:24
  • 2
    $\begingroup$ The membrane could be such, that there is never ever a hole in the membrane. It will start covering the "back side" of the object passing through, until at one point it will completely envelope the object passing through, before creating a hole on the other side and retracting from the "front side" of the object passing through. Shape of the object does not matter, as membrane will never split into several pieces. $\endgroup$ – hyde Apr 3 '18 at 19:38
  • 1
    $\begingroup$ @hyde that's a great idea. Just touch the membrane and it will automatically move to the opposite side of you, no movement required on your part. $\endgroup$ – bendl Apr 3 '18 at 20:06

Mart's suggestion is essentially an airlock,and very viable, but very high-tech.

A simpler airlock solution would be to have something like a womb or pipe with sphincters on both ends. Sphincters recognize creatures allowed to enter, and let them in, then close one end, push out water if necessary, and open the other end.

Also, can have an easy entryway into underwater habitat if you make it horizontal (opening in the floor), and make air pressure inside the habitar equal (or slightly above) water pressure outside. This is what @lijat means by "diving bell" above.

  • 1
    $\begingroup$ For redundancy, you would probably benefit from multiple chambers in your airlock with decreasing pressures, that way any water that leaks through can be dealt with before it gets anywhere near the innermost chamber, and a catastrophic failure of one portal has a limited effect. $\endgroup$ – Ruadhan Apr 4 '18 at 13:09

There is an energy management problem with the single membrane concept.

Imagine the object being passed through is a cube, one meter each side, in and out a volume at atmospheric pressure just ten meters under water.

To exit the membrane, against a pressure differential of one hundred thousand Newtons per square meter, and displace one cubic meter of water weighing one round ton, the cube requires an energy of one hundred kiloJoules. To do so in, say, ten seconds, requires a power of ten kilowatts.

A human being, producing around 100W of power, would need to exert himself for about twenty minutes to drive such a cube outside the membrane; being around ten times smaller in volume, he could exit himself in around two minutes.

The "membrane" must then be an active mechanism. Nano-machines creeping along the surface of the incoming body, using gecko hydrophobic pili for both traction and seal, should be good.

Inside the membrane, there needs to be a much stronger "skeleton" to oppose the pressure differential.

One could do this with a large quantity of gecko-coated flexible marbles, capable of selectively gripping one another. The marbles would need to know their own rest position and be able to communicate with one another. The distance between a geckoball's actual and rest position would be its "energy". Each geckoball would either negotiate a swap of logical positions between itself and another geckoball, if by so doing the sum of their respective energies decreased, or exert a pressure towards its rest position if, and only if, in that direction there was water or air, and nothing else. This movement would require energy, of course.

The net result would be that geckoballs would part with ease in contact with a living creature, allowing it to feel no pressure, and would flow back and reconstitute themselves behind it, requiring considerable energy, as soon as it passed through. This approach requires the membrane to be somewhat thicker than the thickest object going through, which must have no holes. A thinner membrane would require sophisticated communications and checks to stabilize pressure along the separation surface:

enter image description here

  • $\begingroup$ This is an interesting proposal. I was about to ask what about if the person going through needs to breathe and then I realised that they'd be going underwater anyway! $\endgroup$ – Orgmo Apr 4 '18 at 15:17
  • $\begingroup$ I must be misunderstanding your description of exiting the membrane. To lift a cubic meter of water vertically by one meter requires only 9.8 kJ, so I'm not sure how you arrive at ten times that for simple displacement. It seems you're assuming one must lift the ten meter column of water above the membrane in order to move through it, that clearly makes no sense. $\endgroup$ – Samuel Apr 4 '18 at 18:54
  • $\begingroup$ @Samuel in a certain sense yes, you are lifting the whole column: you must do work against the pressure of that column when you exit the enclosure. Of course if the atmospheric pressure inside is higher, the work required is less. $\endgroup$ – LSerni Apr 4 '18 at 19:41
  • $\begingroup$ @LSerni Think of it this way then, shrink the volume displaced by the membrane by exactly the volume exiting it. Does your energy problem disappear? If not, then why does moving through water at a depth of 10 m not require 100 kJ per meter of lateral movement? $\endgroup$ – Samuel Apr 4 '18 at 19:52
  • $\begingroup$ @Samuel moving through water does not require much energy. It's moving from an environment at 1 atm to an environment at 2 atm (STP plus 10m of water depth) that requires energy. And, yes, outsourcing that energy requirement to the membrane is what I was proposing. In practice, it's "as if" the person exited with next to zero energy expenditure but leaving a person-shaped shallow in the outer region of the membrane, then the membrane bulged outward, displacing the water and requiring a pressure-dependent energy. Not so unlike airlocks (there, you spend energy to pump water out). $\endgroup$ – LSerni Apr 4 '18 at 21:28

Why didn't anyone offer jelly, blubber and other gooey fats? :)

They wouldn't pass water through, a solid-enough organism can literally swim or walk through it - and the goo will close after it, sealing the wall back.

Alternately, perhaps one can upscale the mechanisms we all have as the walls of our cells, complete with keeping liquids inside and outside separated while allowing for selective osmosis of nutrients and waste, etc....

  • 3
    $\begingroup$ Because any pressure differential would make the blubber membrane implode. And having air on one side and water on the other guarantees a pressure differential - at the very least, one tenth of an atmosphere for every meter of the membrane height. $\endgroup$ – LSerni Apr 4 '18 at 16:49

My advice would be to avoid the hypothetical membrane and go for a simple solution that we know is doable: the airlock.

This would avoid suspension of disbelief problems that hypothetical technology can cause if it's too far out there, similar to how many people don't like teleporters and other such staples of science fiction. It will also make it easier to write about since both you and your reader will have some intuitive understanding of how it works, and what the limitations are. And that might also make it easier to turn it into a plot point without it coming across as an ass-pull.

For an example of such an airlock in fiction, see the adventure game Indiana Jones and the Fate of Atlantis:

Amazing! We're hundreds of feet below sea level, yet there's enough air pressure to keep the water out. (Video on YouTube)

  • $\begingroup$ I got in a good amount of trouble racking up charges by calling the hint line for that game. $\endgroup$ – Samuel Apr 4 '18 at 18:57

You will have to contend with multiple challenges:

  1. The membrane will need to have some means of identifying material that may pass, including items that the entering persons are carrying, but presumably excluding stuff adrift in the external fluid.
  2. The membrane will need to resist a pressure differential, which is no mean feat: Try pushing an empty container into the water to get a feel for the forces involved. Additionally try to keep a horizontally strung rope completely straight by pulling at it while some weight pulls at it perpendicularly to get a feel for the forces needed to keep the membrane from bulging to the point where it is not really a window but more of a crazy lens. The object sticking through the membrane will experience the same forces, so expect to 'pop in'.
  3. The topology of the entrant will be of vital importance: a torus-shaped entrant will break the membrane. Humans are toruses! (Mouth-anus-connection)
  4. The membrane will behave as some sort of fluid, and will also have high bonding strength to itself ('itself' being chemically defined, or in case of nano-bots both chemically and cryptographically), so contamination will be a huge problem. Be prepared to have them be 'alive' to self-regenerate/purge or have a thriving industry of membrane-cleaners at hand.

Imo your best bet yould be to have the creatures wear membranes themselves: Topology-concerns are handwaved away by making everything a (topological) sphere, plus any membrane-clad thing or person is cleared to enter, non-clad vagrants and debris don't get in. The pressure-differential must:

  • a) not exist (pressurized habitats, but be wary of multi-story habitats, they will have the same internal pressure, but the external pressure will vary wildly with depth, assuming dense external fluid and light internal atmosphere)
  • b) be compensated by some force thats commensural to the area the entrant has in the membrane at that moment (two caveats: 1) for someone entering the membrane, that area will increase to a max, and then dwindle again, as the person slips through, as must the force to keep the entrant from 'popping in' 2) the area will rise to the power of 2 with size, while the circumference will only rise by simple power, so the force to hold back / force out the entrant can not be linearly dependent on the circumference, which would be natural in any active-membrane (e.g. nano-bots) scenario.

Probably best to stay with the internally pressurized scenario, or just ride with the 'popping in' and incorporate for comedic effect :-)

The membrane passing could be achieved by membrane-clad entrants just as a vesicle passes a biological membrane - google 'vesicle membrane trafficking' for a plethora of good illustrations. Molecules swimming in that bio-membrane can also secure special admission-rights, just as they do in cells. With electrical potentials, the membrane can be made to open up (electroporation) so your visitors can shed the membrane they are wearing. If the suit-membrane is just a molecular bilayer, it might be thin enough to allow transmission of gases at sufficient speeds (breathing!) - the entrance-membrane can be a lot thicker and be interwoven with filaments (think muscle filaments, those can move against each other creating force, while also able to reversibly detach from one another) to make it more plausible in the toughness arena.

  • $\begingroup$ Welcome to WorldBuilding! If you have a moment please take the tour and visit the help center to learn more about the site. Have fun! $\endgroup$ – Secespitus Apr 5 '18 at 8:38

I'm not sure how to form this answer as I don't have the scientific terminology handy. But here goes for the concept.

Let "water" surface tension be the door. The diving bell comment above has the right idea. You need to have the internal pressure and external pressure the same. So in your world, can air and water be the same weight/pressure? They could behave like water and air do in zero gravity. Your complication would be the stickiness of water.


A good real-world example is the cell membrane surrounding all living cells. The membrane is composed of a phospholipid bi-layer which form a water tight membrane. The membrane is perforated with protein channels that allow large molecules to enter and exit the cell as needed.

Given that this is science fiction or fantasy, adapt with words like "di-electric' and 'plasma field' as needed.


This is similar to a few other answers but how about a non newtonian fluid that is kept solid by the internal and external pressure and the only way to pass through is to slowly and gently travel through it.

The upside to this is that it could technically be translucent and you have time to identify what is attempting to pass through it. It could also act as a defensive barrier against anything being fired at it. All that would be required to fix it is to add more fluid.

Slow moving fish or aquatic creatures would randomly enter and die. No way to prevent hostiles from entering if they know how it works. This also isn't how non newtonian fluids actually work.


You could have a membrane that self-heals so that your creature can go in, sort of create an airlock for itself and close it off from the outside environment then proceed to tear open the membrane (on the inside of the bubble/dome) which, then absorbs the torn mass back into itself. It's like the plastic balloon toy where you roll up a piece of it and put it on one end of a blow pipe and blow air through it. You close it up by pulling out the pipe and closing the hole in.

  • $\begingroup$ membranes tend to react pretty badly to anything disrupting their force equilibrium. That's why balloons pop when punctured. $\endgroup$ – L.Dutch - Reinstate Monica Apr 6 '18 at 7:09
  • $\begingroup$ Welcome to WorldBuilding Arkhaine! If you have a moment please take the tour and visit the help center to learn more about the site. Have fun! $\endgroup$ – Secespitus Apr 6 '18 at 7:36

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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