Thermal energy is everywhere on earth.

If it could be transferred into electrical energy it could serve as a great energy source.

This idea is violating the laws of thermodynamics.

But it's not against conservation of energy.

I wonder if an electronic component which has the size of only some nano meters could absorb the brownian motion. Of course many of these components would be needed to generate a usable voltage.

Is this possible? Will it cool down the planet? When will it be possible?

  • $\begingroup$ I’m worried that all scenarios would result in loss of energy. The only way to focus heat enough to drive anything with viable results naturally rather than mechanically (requiring yet more energy expenditure) would be in environments where other solutions would be more appropriate (Graphene pressure diodes underground where it’s geologically active perhaps). A little more context would be useful. Remember, if you decide to create any plants that live purely of infra-red light then their leaves will be black, not green. $\endgroup$ Nov 27, 2019 at 14:24
  • $\begingroup$ See Maxwell's Demon $\endgroup$
    – Dugan
    Nov 27, 2019 at 17:01
  • 3
    $\begingroup$ Is is possible to transform thermal energy into electrical energy. This is how thermo-electric power plants work, be they coal-fired, gas-fired or nuclear. What you seek is a means to use unavailable (= non-free) energy. Your proposed device is called Maxwell's demon (because it was first imagined by James Clerk Maxwell in 1867), and the last 150 years have produced a considerable amount of literature about its properties and implications. $\endgroup$
    – AlexP
    Nov 27, 2019 at 17:59

5 Answers 5


This idea is violating the laws of thermodynamics.

Simply put, forget about it.

Thermodynamics laws are the best enforced laws in universe. There is no way around them.

Those laws clearly state that you cannot convert heat to usable energy, you can only convert a fraction L of energy from a temperature T1 to usable energy if you can discard some energy at temperature T2 below T1.

That fraction is given by $L=1-$$T_2 \over T_1$

  • $\begingroup$ I've been thinking about this one a lot thanks to all the conversations about spacecraft and waste heat lately. $\endgroup$ Nov 27, 2019 at 14:24
  • $\begingroup$ I think you did not understand what the laws of thermodynamics are. It is indeed legal to violate them, when it comes to single atoms. $\endgroup$
    – zomega
    Nov 27, 2019 at 18:28
  • $\begingroup$ @somega So, the closest thing I could find to 'violating thermodynamics at the atomic level' is the Maxwell's Demon thought experiment, which may be possible with quantum somethings-or-others. Which has no bearing on what you're asking, with the perfect transfer of energy from thermal to electrical. $\endgroup$
    – Halfthawed
    Nov 27, 2019 at 18:56
  • $\begingroup$ Three laws of thermodynamics are simply restated as: "You can't win, you can't break even, and you can't even get out of the game." All the pursuits of efficiency we put time, money, and thought into are about losing slower. $\endgroup$
    – Zeiss Ikon
    Nov 27, 2019 at 20:21

We can already do that, and no, it does not violate the laws of thermodynamics at all. This is called the Thermoelectric effect. As for doing this effectively at scale, it's generally more expensive than other means of heating, cooling, and electrical generation, so it's not done very much. There is research about it, such as this.

Currently it's practical usage is mostly as thermocouplers in ovens and other gas appliances.

Yes, it's possible to make a lot of these. But possible does not mean practical. There are much better ways to generate electricity.

And no, this would not "cool the earth down" so-to-speak. What's really happening in the Seebeck effect is using differences in temperatures to stimulate a small current across different metals. You're not "taking" nor "converting" the thermal energy, so the cooling effect is very minimal.

  • 4
    $\begingroup$ This doesn't really address his question though. The seebeck effect doesn't CONVERT heat into electricity, it just lets you generate electricity out of a temperature gradient. The heat is still there, so you can't (for example) use a thermoelectric generator to get rid of all the waste heat in your spacecraft and turn it into usable power. You still have to radiate it all away. Seebeck just lets you generate some electricity from it as it's leaving. $\endgroup$ Nov 27, 2019 at 14:23
  • 2
    $\begingroup$ @MorrisTheCat exactly; but his question was a bit vague on this: If it could be transferred into electrical energy ... it's not immediately obvious from a layman's perspective whether he means literally transferred as in "changed into" or transferred as in "used for". If the goal is to get energy while starting from heat, then to the layman that's what the Seebeck effect is "doing". From a scientific perspective, the heat is not changing, but OP did not specify this as a necessity in his answer $\endgroup$
    – cegfault
    Nov 27, 2019 at 14:26
  • $\begingroup$ The thermoelectric effect needs a temperature difference. I am speaking about getting electric energy from for example sea water with an uniform temperature. $\endgroup$
    – zomega
    Nov 27, 2019 at 18:31
  • $\begingroup$ You can buy thermoelectric generators off the shelve. They are used by outdoor geeks to run USB devices, and by interplanetary spacecrafts. $\endgroup$
    – Karl
    Nov 27, 2019 at 22:10

As it happens I've been putting a lot of thought and reading into this exact problem lately, and it's a real head scratcher. This problem comes up a lot when people talk about spacecraft design and waste heat issues. Intuitively it seems like obviously you should be able to capture heat and turn it into useful power.

The problem is Entropy. Here's a summary:

The total entropy of a system either increases or remains constant in any process; it never decreases. For example, heat transfer cannot occur spontaneously from cold to hot, because entropy would decrease. Entropy is very different from energy. Entropy is not conserved but increases in all real processes

Because heat is, basically, the MOST disordered form of energy, you can't just convert heat into a less entropic form like electricity If you want more reading, I found this document pretty enlightening.

Now, and important caveat is that you can HARNESS the flow of heat from a higher concentration to a lower concentration and generate electricity THAT way, this is how thermoelectric generators work. It's important to note thought that you're not getting RID of any heat this way, you're just generating some electricity as the heat moves.

  • $\begingroup$ Entropy is not disorder. This notion should be expunged. Entropy is the log of the degeneracy of a state. The degeneracy is the number of configurations with the same energy. "Number of equal energy configurations" does not have any simple relationship to order. $\endgroup$
    – puppetsock
    Nov 27, 2019 at 14:35
  • $\begingroup$ Harmessing heat flow from hotter to cooler is how all heat engines work, whether they're a Newcomen engine from the 17th century or a hyper-efficient supercritical closed-cycle turbine generator system (which still wastes about 60% of the heat into the environment). $\endgroup$
    – Zeiss Ikon
    Nov 27, 2019 at 20:24

We know of no way to do this.

As to your particular nano-scale contraption, that one is actually a rather interesting little story. We've actually investigated the idea of capturing energy out of Brownian motion with nano-scale devices. The structure that we relied on was a water-wheel of sorts with a ratchet. The idea was that random collisions in one direction would cause the wheel to advance, engaging the ratchet mechanism. Random collisions in the other direction would oppose the ratchet mechanism and fail to turn the wheel. It looked like that would turn the random Brownian motion into rotational motion that we could use to turn generators. From what I understand, we actually made several of these for study.

Of course, given that we don't hear much of it, it's pretty clear it didn't work. It turned out that we neglected one tiny part of the ratchet: the spring. If it's too stiff of a spring, then the Brownian motion wont be able to push the wheel. If it's too loose, the spring can sort of flop around on its own (under its own Brownian motion), and let the ratchet move backwards. Both cases prevent energy from being generated.

But what about if you get it just right? Well, it turns out that if you get the spring just right for the water bath that you're generating power from, you can indeed generate some power. However, that power comes at the cost of warming the spring. The laws of thermodynamics actually transmits heat from the water bath into the spring until the spring is too loose to generate power. Run the numbers, and basically we transferred heat from a hot water bath to a cold spring, and from this we derived from this a small amount of power. If we were to cool the spring down again, we would find we accidentally created a thermodynamic cycle, and derived power from this transfer of heat.

To the best of our understanding, this is the end of the story. We know of no way to get past this issue.

The real issue is that heat is a form of energy that's described stochastically. There is a probability of the system being in any given state (and, at equilibrium, every state with the same classical energy will be equally likely to occur). Given that we can't predict the exact state of the system, we can't outwit the laws of probability which state that entropy always increases.

Of course, there's no obligation that future scientists wont "crack the code." We already have forms of energy that are in a usable form, like kinetic energy or chemical energy. If one were to find a non-probabalistic structure to thermal energy, then we could leverage that to pull power from it. At the moment, we do not believe this is possible, thanks to chaos theory, but who knows what future cleverness may hold. This is, after all, the species that figured out how to harness the power of the wind and the power of the water to grind their grains to make bread.

If you are interested in studying this question, I highly recommend The Last Question by Isaac Asimov. It's a story about people asking the question of whether such entropy reversal is possible. It's short, well written, and its ending does a very good job of capturing just how extraordinary such an answer would be.


Sure, we can do it today. As long as you are taking heat from a hot place to a cooler place, there's no problem with the second law of thermodynamics.

You can use a solar updraught tower. Works best where there's a lot of solar thermal energy. Take a bit of desert, or a nice big black tarmac area, build a tower to collect the rising hot air and funnel it up into a turbine. Heat in, electricity out. No commercial scale plants have been built, but a 200kW plant is in operation in Mongolia and there were plans to scale it up to 27MW.

You can use a ground source heat pump which pumps a heat transfer medium underground where it is heated by natural hot rocks. In this setup the heat is often used directly for heating a building, but you could design a phase change setup with a medium that boils at the working temperature and the gas is used to spin a generating turbine and condensed back into the working circuit. In this setup you do require some energy to run the pump, but you can harvest a great deal of heat from underground. Iceland uses a lot of pumped geothermal energy.


You must log in to answer this question.

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