As we all know1, stillsuit as described in the books would cook the user. Water would carry heat away from the body, but then it would be caught again, and again with the heat.

Assuming energy is not a problem in a world with tiny Hunter-seekers and routinely levitating lamps, we could probably put that heat away - but where to? Modern day phase change cooling requires either frozen packets for heat to go to, or radiators and environment colder than the user. Electrical heat pumps can heat one end to cool the other, but the heated end still needs effective way to dump the heat. Something with large thermal capacity, or radiator with airflow.

Is there anywhere we can dump heat on the Arrakis-like environment without making users look like hedgehod with glowing red fins of a radiator? And preferably with little or no moving parts?

1 The Science of Dune: An Unauthorized Exploration into the Real Science

  • 6
    $\begingroup$ Refrigerators exist. Haven't read Dune. How hot is Arrakis exactly? If it's less than a hundred degrees C you can probably get by with souped-up refrigerators and they won't glow red hot. $\endgroup$ Commented Nov 2, 2021 at 10:03
  • $\begingroup$ @user253751 up to 75 degrees Celsius, but sources differ on that. And my refrigerator has about 1.5 square meter of a radiator, impractical on a suit. $\endgroup$
    – Mołot
    Commented Nov 2, 2021 at 10:12
  • $\begingroup$ Please don't answer in comments. Use comments only to ask clarification/suggest improvements to the post. Other comments will be deleted without notice $\endgroup$
    – L.Dutch
    Commented Nov 2, 2021 at 13:15
  • $\begingroup$ @ZeissIkon "a human body has coincidentally close to 1.5 square meters of surface area" flexible and mobile, while resisting pressure (fridges use a compressor) is hard to come with in today's level of tech. Be it only because those flexible tubes have a tendency to become rigid under pressure - works a treat for car tires. $\endgroup$ Commented Nov 2, 2021 at 15:19
  • $\begingroup$ You are in air. Just transfer excess heat to air, which is then dumped. This requires a power source, of course, but you could make the suit surface photovoltaic, no problem. $\endgroup$
    – PcMan
    Commented Nov 2, 2021 at 17:02

5 Answers 5


If you want a Dune "hard science" answer, the "heat exchange filaments" that AcePL gracefully cited are very likely engineered by applying the Holzman effect at molecular scale. After all, repealing "fast moving atoms/molecules" and letting only the slow one pass through sounds pretty much what a Maxwell daemon would do - at ambient temperature, the nitrogen and oxygen molecules average speed is 300-400 m/s. Dial the allowed speed to lower values and the molecules allowed to pass through will be grateful for a little bit of warming up.

If you want real world hard science, a heat pump and a radiator is all you need. The best coefficient of performance for real world systems pumping heat from 0C to 35C is 4.5 - that is, for every Joule that you spend on the heat pump, you move 4.5J of heat from low temperature to high temperature.

The maximum theoretical achievable for cooling between 37C (310K) as the "cold temperature" and, say, 80C as the hot temperature is

$$COP_{cooling} = \frac{T_c}{T_h - T_c} = 310/47 = 6.6$$

But even only half of that would be enough, because... human power is not that big

Normal human metabolism produces heat at a basal metabolic rate of around 80 watts.

During a bicycle race, an elite cyclist can produce close to 400 watts of mechanical power over an hour and in short bursts over double that—1000 to 1100 watts; modern racing bicycles have greater than 95% mechanical efficiency. An adult of good fitness is more likely to average between 50 and 150 watts for an hour of vigorous exercise. Over an 8-hour work shift, an average, healthy, well-fed and motivated manual laborer may sustain an output of around 75 watts of power.

Assuming the freemen are trained and motivated by survival, one can allow for 150W. At a $COP_{cooling} = 3.3$, the heat pump will require a 46W - letting the freeman $150W - 46W = 104W$ of power for survival. (to compensate, the spice in the freeman's ration will need to provide 4.3 MJ for each 8 hours of marching)

@AlexP (thanks) points out in comments that the human body is not very efficient (and provides citation for it), for every Joule of useful power it produces, another 2 or 3 Joules are wasted as heat (efficiency in the 25-33%). So the correction to the energy the spice ration should provide - between 12.9 MJ and 17.2MJ
More important, the heat pump will need to reject 450-600W - I'll come to it later.

The median energy intake for athletes during competition is 8674 kJ, but the extreme include a whooping 18,009 kJ (so the 8h energetic effort/day is plausible).

Now, the realization of the stillsuit:

  1. the lithium-ion battery has an energy density of 0.90–2.43 MJ/L. So 8L worth of battery (say, 12kg including well engineered electrodes and enclosure?) would let the freeman use her/his full power (and "just carry" the stillsuit, baterry pack included). After all, those Holzman shield do need some power, and they carry one even if they don't use it when navigating open sands;

  2. will definitely need an insulating material, very likely reflective - the outside appearance should be like those emergency thermal blankets.

  3. the trick is... if you maintain the body temperature at 36.5C, the body has little incentive to perspire. So... actually... the still suit will only condense the respiration water and extract the water in urine/fecals?

  4. the radiator (air cooled) will make probably the bulk of extra weight. Because rejecting heat must be done by compressing the refrigeration fluid, so one expects a good heat conductor and good behavior under pressure. Probably aluminium would be good enough a compromise, otherwise one would suggest the heavier copper

Well engineered and tested in Arrakis conditions, one on top of the other (suit, compressor, radiator, battery pack) would contribute - feeling of guts - at about 20-25kg?

To be noted that a 200W camping fridge, able to cool 50L+ of drinks to 4C, weights about 20kg when empty.
A 200W with an assumed $COP_{cooling} = 3.3$ would be able to reject 660W, so we are well within the reach of today's technologies

Now, about the weight of the suit, some perspective

For the last 3,000 years, dismounted soldiers carried 55 to 60 pounds on average. This has almost doubled in the last 200 years. Roman legionnaires carried almost 60 pounds. The British fighting in the American Revolutionary War carried 80 pounds. At the Battle of Waterloo (1815), the British carried 60 to 70 pounds while the French carried 55 pounds. The French in the Crimean War (1853-1856) carried 72 pounds. Around World War I, approximate march weights jumped to 85 pounds. U.S. soldiers trained with at least 60 pounds but carried additional rations and munitions in combat. During World War II, U.S. troops carried more than 80 pounds in the Normandy landings. U.S. soldier loads increased even more dramatically in the second half of the 20th century. March loads stayed at approximately 80 pounds during Vietnam but grew to 100 pounds afterward, with a maximum march weight over 160 pounds in Grenada in 1983. In Iraq and Afghanistan, march weights have approximated 100 pounds or more.

Answers to grumblings

  • but I'm fitter than a fit freeman and I used to push 220W for 1h

Well then, you maxed out your heat pump capacity, you probably need to do it at a higher temperature - which, BTW, is a self limiting factor - but you still won't boil inside.

An increase in $T_c$ (i.e. acceptable body temperature) will make the COP of the heat pump higher - as it only needs to fight a low temperature differential. The human body will work at 40C, a temperature when the theoretical COP goes to $COP_{cooling} = \frac{313}{353 - 313} = 7.11$ (353K is 80C, 313K is 40C). Half of that is 3.5, with a 200W input, you'll be able to reject 700W.

It's very much like what happens when you max out your A/C and then you decide to put something in the oven - you home won't catch fire, you'll just need to accept an increased kitchen temperature. You may accept it or else you will need to be wise and shut down to oven.

  • 1
    $\begingroup$ To produce 150 W of mechanical power, a human's metabolism will produce some 300 to 400 W of waste heat, in addition to the 100 W baseload. Humans are very inefficient when used as engines. $\endgroup$
    – AlexP
    Commented Nov 2, 2021 at 14:28
  • $\begingroup$ @AlexP Citation please. I'm not saying that's not true, but... hard-science. I can adjust the answer, but I think I'd be still safe, given that 200W camping fridge would manage to move, with a COP of 3.3, 660W - one would think that may be enough. $\endgroup$ Commented Nov 2, 2021 at 14:31
  • 1
    $\begingroup$ I'm not saying that it isn't enough. It was just a note that the numbers are on the very optimistic side, and in practice you would need more refrigeration. As for the efficiency of converting food calories into mechanical output, that's not hard to find. The generally accepted value range for human-as-engine efficiency is 20% to 30%. $\endgroup$
    – AlexP
    Commented Nov 2, 2021 at 14:44
  • $\begingroup$ @AlexP Thanks, mate. $\endgroup$ Commented Nov 2, 2021 at 14:45
  • 1
    $\begingroup$ @10ebbor10 "sandworms... hate shields". True... when manifested in external form. I assert that the "heat exchange filaments" work so that the field lines close inside the stillsuit, very much like the magnetic field lines never leave a ferrite core until fully saturated. $\endgroup$ Commented Nov 2, 2021 at 16:03

Let's find out the scale of the problem. How much heat do we actually have to dump?

Some random source suggests an active human produces around 1Kw at peak output. Our stillsuit needs a refrigeration system able dump that amount of thermal energy (1000 joules per second). Assuming a perfectly passive radiator (radiates purely through, uh, radiation) we can now grab the Stefan-Boltzman law

P = A * ε * σ * T^4


P = power in watts to dissapate
A = area of radiator
ε = Emissivity. Lets assume a perfect black body aka 1.0
σ = 5.670373×10-8 = Stefan-Boltzmann constant
T = Temperature raise above ambient

We can now find the relationship between the radiator area and temperature rise:

A * T^4 = P / (ε * σ)
A * T^4 = 1000 / (1.0 * 5.670373×10-8)
A * T^4 = 17635524153

Which means that if we have a 1m2 radiator....

T^4 = 17635524153
T = 364.4 degrees Kelvin

Converted to Celcius, that's a surface temperature of 91 degrees, which is hot but not crazy. That's a pretty decent cooling loop you've got there, a gradient of 60 degrees - about 3 times that of a typical freezer, but it sounds pretty plausable to me. Curiously the emissivity and area don't really matter so much because of that T^4. If we halve our area to 0.5m2 we only bump our radiators surface temperature to ~140 degrees.

These temperatures aren't too crazy, radiators that work with those sorts of power levels probably exist in various industrial processes.

This is assuming that our suit is a perfect black body emitter on it's radiator panels, and ignores several effects:

  • First up, the big kicker: convection and airflow would help to cool the suit. This is why cars have radiator fans. This could dramatically reduce radiator temperatures. (Unfortunately my knowledge of thermodynamics is pretty poor, so I'm keen to see other peoples answers with better designed cooling systems.)
  • Energy absorbed from the sun would add to the energy needed to be dissipated. Here on earth the solar energy at the surface can hit ~700W. Painting your suit white or silver could help with this (but would hurt emissivity)
  • This assumes a 100% efficient system. Any energy created by reactors/pumps etc. would need to be handled by the cooling system as well. There's a thing called the coefficient of performance (https://en.m.wikipedia.org/wiki/Coefficient_of_performance ) which implies a heat pump can move ~3x more heat than it consumes, so a heat pump that provided 1Kw cooling would require ~300 watts power.
  • That same source suggests that sitting down reduces a humans thermal output to only 100W.
  • 3
    $\begingroup$ "(Unfortunately my knowledge of thermodynamics is pretty poor, so I'm keen to see other peoples answers with better designed cooling systems" Suggested reading: coefficient of performance. Radiative cooling as an approach is... toast. $\endgroup$ Commented Nov 2, 2021 at 12:19
  • $\begingroup$ May I interest you in some MathJax? :) $\endgroup$
    – Mołot
    Commented Nov 2, 2021 at 12:27
  • $\begingroup$ T is measured in Kelvin, not in Celsius. So 365 Kelvin is around 93 Celsius. $\endgroup$
    – quarague
    Commented Nov 2, 2021 at 14:18
  • $\begingroup$ @quarague I thought that here the T is Temperature in Kelvin above ambient, but on second thoughts you are right. T-above-ambient would imply conductive or convective energy transfer. $\endgroup$
    – sdfgeoff
    Commented Nov 2, 2021 at 17:55

Using the sky-radiation scheme from @MikeSerfas answer, it's pretty simple (with abundant energy sources as posited) to make the radiating surface hot enough for black-body to get rid of your heat. You need Peltier junctions, aka thermocouples, or whatever they develop into over the next hundred centuries.

These work two ways. The more familiar way, for most people, is to heat one junction while the other is kept cool, to produce a small DC current -- but it's a reversible effect; if you connect the junction pair to a current source, one end will get hot while the other gets cold; in effect, it's a no-moving-parts heat pump. These are commercially sold for iceless camp coolers and similar, drawing 12 V from an automobile electrical outlet to keep roundly 1/8 cubic meter of insulated box 10-15 °C cooler than the ambient temperature for as long as the power source lasts.

These could work without any sort of dedicated radiator; the exterior of the stillsuit would merely have to be (enough) hotter than the ambient temperature of your desert planet to conduct/convect the heat away to the air -- whether that's 40 °C or 60 °C. Of course, the more temperature differential you need to maintain, the more junctions and the more power are needed. They might be more efficient, and surely would have a reduced heat/infrared signature, if they had some kind of directed radiator, however.

  • 4
    $\begingroup$ "You need Peltier junctions" Good Lord, noooooo... Besides having only a 5% efficiency, the 95% rest of it is heat. This is why cascading Peltier elements to obtain lower temperature is doomed to fail for more that 2 stages - the next stage needs to be way larger to dump not only the extracted energy, but also the heat create by the previous stage. $\endgroup$ Commented Nov 2, 2021 at 12:15
  • 2
    $\begingroup$ So I don't suppose there's any likelihood that the efficiency of such a device can be improved in the next ten thousand years? $\endgroup$
    – Zeiss Ikon
    Commented Nov 2, 2021 at 13:56
  • $\begingroup$ Give me enough spice and I just might be able to foresee that future :D Until then, hard-science "requires answers backed up by equations, empirical evidence, scientific papers, other citations, etc.". $\endgroup$ Commented Nov 2, 2021 at 14:20
  • 1
    $\begingroup$ @ZeissIkon: Peltier junctions, extremely unlikely. (The effect was discovered in the first half of the 19th century. We had plenty of time to explore it.) Another, as yet unknown, solid-state heat pump, maybe. $\endgroup$
    – AlexP
    Commented Nov 2, 2021 at 14:24
  • 3
    $\begingroup$ @ZeissIkon If we can't say whether it's possible with the hard science we know right now, then the answer is "I don't know," or "not with our current technology." Those are valid answers, and not a reason to stop someone from asking a question in the first place. $\endgroup$
    – automaton
    Commented Nov 2, 2021 at 21:17

People consume about 2000 kcal/day = 100 W on average. This has to be radiated from a surface of about 1.9 m$^2$, so we need 53 Wm$^{-2}$. The Stefan-Boltzman law requires a temperature difference $T_2-T_1$ such that $\sigma (T_2^4-T_1^4) = 53$, where $\sigma=5.67\times 10^{-8}$ Wm$^{-2}$K$^{-4}$. Hence $T_2^4=T_1^4+53/\sigma$. If the outside temperature is $T_1$ = 50 C = 323 K (a hot day in the Sahara desert), then $T_2$ = 330 K = 57 C.

This calculation ignores lots of other factors like convection, it being cooler at night, etc., but illustrates the point that you don't really require red hot radiator fins to be able to get rid of the small amount of heat a human body generates.

But you do very definitely need active heat pumping to keep the body at a survivable 30 C and the outside of the suit at 57+ C. The maximum efficiency theoretically possible is the Carnot heat cycle, where $1-T_C/T_H=W/Q$. Here, $T_C$ = 303 K and $T_H$ = 330 K so the work needed (per second) divided by the heat energy transferred (per second) is 1-303/330 = 0.08. To transfer 53 Watts we need $0.08\times 53=4.3$ Watts of effort. About 8% of your energy expenditure has to be redirected to keeping cool. Boot pumps driving a compressor seems like a plausible approach.

You need some sort of very high thermal resistance material to separate the inner and outer layers, and then you need heat pumps to pump the heat from the inside through channels in the resistor against the thermal gradient. An extremely good foam insulator (like an aerogel) could have a thermal conductivity of around 0.01 W/m K, so with a temperature difference of 27 K over an area of 1.9 m$^2$ and maximum practical thickness of 1 cm we get a leakage of $0.01\times 27\times 1.9/0.01=51$ W, which is far too large. We're pushing the state of the art, and still need something a factor of 5 better.

Here we have to get a bit speculative, and appeal to future technology and metamaterials. One way to enhance the thermal resistance of a vaccuum is to create many layers. If we have lots of flat opaque membranes separated by vaccuum (with spacers holding them apart strong enough to withstand the pressure and sparsely enough for their contribution to conductivity to be ignored) then heat can only transfer from each membrane to the next by radiation. If the power conducted through the material is $p$ then each membrane must radiate $p$ units of power each way more than the previous one. The membranes radiate $p_0+p$, $p_0+2p$, $p_0+3p$, $p_0+4p$, ... $p_0+np$ units from each side. Take the $p_0+3p$ membrane - it radiates a total of $2p_0+6p$ units of power from its two sides towards its neighbours, and receives $2p_0+2p+4p=2p_0+6p$ from its two neighbours in return, and is thus in equilibrium. In the gap between any pair of membranes, there is $p_0+kp$ units moving one way and $p_0+(k+1)p$ units in the other direction, giving a net transfer of $p$ units. This yields an insulator far more effective than a single vaccuum gap.

Of course, if you suppose the desert is generally cooler than that (as most deserts on Earth are) then the problem goes away. Descriptions of Arrakis emphasise the extreme dryness, not the heat. A desert can be quite cool, and still deadly for lack of water.


It's a matter of creating particles...

We need to carry the entropy away somehow. So we want to make some low-energy particles and send it away. Specifically, we want a surface with an emissivity greater than 1, or at least, which acts like it has an emissivity greater than 1.

For example...

  • We create "ommatidia", deep cavities with an emitting membrane that faces upward toward the desert sky. They receive little infrared radiation in that direction, so if we could increase the rate at which photons are radiated (or absorbed - I'm not going to require irreversibility here) then we could get rid of all the heat we wanted. Question is ... is there any way, with all that Dune technology, to get our photon count well above what a black body would produce?

  • We do neutrino pair production. Some neutrinos are very, very low in mass, and if we use some free energy to make them out of nothing, then they will carry away some heat with no additional charge. Can Dune tech make neutrinos?

Any other hypothetical particle you can mine from the books might similarly have such a function.

Alternatively... You can use your spice-induced prescience to pick the universe where all the random transfers of heat from one particle to another happen to move away from you. :)

  • 1
    $\begingroup$ Neutrinos et al are the only "magic" cooling option that can keep an object cooler than it's surroundings passively and indefinitely without violating thermodynamics. Too bad no known material gets anywhere near the needed absorption cross-section for sub-eV neutrino energies. $\endgroup$ Commented Nov 2, 2021 at 21:11
  • 2
    $\begingroup$ @KevinKostlan see this Nature paper. more work has been done since then, but this is the group I saw give a keynote talk ata conference. Radiative cooling can be done to space with materials that are very reflective at wavelengths the atmosphere radiates, but emit well at other wavelengths that go straight out to space. So it's very much a non-black body, but behaves like one at some wavelength ranges. The sun fills very little of the sky in angular terms so isn't major consideration. $\endgroup$
    – Chris H
    Commented Nov 3, 2021 at 10:36
  • $\begingroup$ ... these coatings can also be visible-transparent, so there's work on applying them to cool solar panels (which get less efficient and less robust when hot). It's not good enough yet, but not so very far from reality (needs to face the sky for best effect, so a very broad-brimmed hat!) $\endgroup$
    – Chris H
    Commented Nov 3, 2021 at 10:39
  • $\begingroup$ @ChrisH Okay, that is pretty much the answer to this question. I missed that paper, but "we experimentally demonstrate radiative cooling to nearly 5 degrees Celsius below the ambient air temperature under direct sunlight..." Just by applying a coating? $\endgroup$ Commented Nov 3, 2021 at 10:52
  • $\begingroup$ @MikeSerfas it's not yet efficient enough, and it's not like painting it on - in fact rather a difficult process. (I have to admit I was sceptical at first, even hearing it at a reputable scientific conference, but it's mainly a counterintuitive application of radiative and atmospheric physics, with some well-established optical production techniques. The Fremen would need a cleanroom with some pretty sophisticated kit in their caves to produce the material. $\endgroup$
    – Chris H
    Commented Nov 3, 2021 at 10:57

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .