Its a summer day and the air in my house has been heated up. I could switch on my air conditioning, but then I'd be using energy from the grid in order to reduce the amount of energy in my house.

What I'd much rather do is capture the heat energy from the air, so cooling it down, and turn that heat energy into electricity. (So I could store it in a battery or sell it to the grid or use some other means of disposing of it.)

Is such an apperatus possible within our present understand of nature? Why or why not?
(I presume not, as people would be selling kits.)

Just to pre-empt a possible answer: I've seen similar conversations in the past and "inefficient" has been given as the answer as to why not. I don't understand this as an objection, because the heat energy is effectively free. The sun rudely went and gave me an excessive amount of heat energy without being asked to. Is converting heat to electricity really so ineffiecient that I'd be better off running air conditioning?

  • $\begingroup$ As Claudius says, you need someplace colder for the heat to go to, like an ocean or maybe underground water. $\endgroup$ – Mike Dunlavey Nov 14 '12 at 22:48
  • 1
    $\begingroup$ I have an idea ;). You can "collect" heat on a hot summer day then release it at night. It would require the right chemical reaction, which may exist. The chemical reagents are such that chemical reaction happen in the forward direction at higher temperature absorbing heat. And backward reaction happen at night releasing heat. One may use the chemical reaction that happens at night to generate electricity. While this is possible, it might not technically feasible. Important thing is you need a colder place. $\endgroup$ – Prathyush Nov 15 '12 at 1:07
  • 4
    $\begingroup$ @Prathyush it is technically feasible and has been done . One pine resin changes from solid to liquid at around 22degC. So in the day it melts absorbing heat and when the temp falls below this at night it freezes releasing heat. You just need a log cabin in a desert $\endgroup$ – Martin Beckett Nov 15 '12 at 2:40

In order to build any thermal engine as envisioned by you, you need both a cold and a hot reservoir, such that heat can flow from the hot part to the cold part and the entropy doesn’t decrease while you’re making energy.

The efficiency of such a machine has an upper limit of $(T_{\textrm{hot}} - T_{\textrm{cold}})/T_{\textrm{cold}}$ (as given by the perfect Carnot engine).

Given that you are usually well off when you get a cold reservoir of $T_{\textrm{cold}} = 290\textrm{ K}$ on a hot summer day ($T_{\textrm{hot}} = 320\textrm{ K}$, the efficiency of your machine has an upper bound of 10%, which does not necessarily include loss due to friction, electric resistance, escaping air etc. If you include these, you get (probably) well below 1%. But, for the sake of argument, let us continue with an assumed efficiency of 10%.

What you want to know next is the maximum heat you can transfer out of your hot reservoir into your cold reservoir. For simplicity, we will assume that the cold reservoir is very large and stays at a constant temperature, heat then flows from the hot to the cold reservoir as long as $T_{\textrm{hot}} > T_{\textrm{cold}}$. The thermal energy hence available to you is $W = c_V \times \delta T \times N \times 10\%$, where $\delta T = 30 \textrm{ K}$ is the temperature difference, $N$ is the mass/number of air (particles) and $c_V$ is the heat capacity.

For air at sea-level, Wikipedia gives me $c_V \approx 29 \textrm{ J}\textrm{K}^{-1}\textrm{mol}^{-1}$ (I am using the constant-pressure one, as we cannot compress air without putting more work into it). I shall then assume that you have a really large house of $10 \times 10 \times 10 \textrm{m}^3 = 10^6 \textrm{ L } \hat{=} 4.27 \times10^4 \textrm{ mol} = N$ where the last but one equality stems from the fact that there are about $6.02 \times 10^{23}$ particles in about $23 \textrm{ L}$. Great, we can now calculate W!

$$ W = 1.03 \textrm{ kWh } \hat{=} 21 \textrm{ cent}$$

where the last equality is a top-of-my-head number I have floating around for electricity prices in Germany during the summer of 2011.

To conclude: Even assuming that you somehow manage to build a perfect Carnot engine using the heat in the air in your house and find a magical reservoir of constant temperature (some part of the earth, possibly), you would get about 20 € per year out of it (four months of high temperature and one "charge" per day).

Really, just put some solar cells on your roof and appropriate insulation on/in your walls. ☺

  • 3
    $\begingroup$ While I can't follow all of the physics and math part of your answer, what I read in the conclusion is "you would get about 20 € per year out of it". What if my goal isn't really to MAKE money but rather to not spend it? I am also bothered by the basic idea that we use energy to remove energy from our homes... $\endgroup$ – nute May 30 '13 at 15:56
  • $\begingroup$ In your Formula for the Carnot efficiency, you meant to divide the $T_{hot}$, don't you? $\endgroup$ – Calmarius Dec 2 '18 at 23:50

I like Claudius' answer. To concisely reiterate, if you could take the heat inside your home and turn it directly into electricity, it would violate the second law of thermodynamics and be equivalent to creating perpetual motion.

Although it is nearly certain that such a device is impossible, the details of why it is impossible can be subtle. One famous attempt to do something like what you're interested in is Maxwell's Demon.

Since you're looking for an intuitive explanation of why you can't simply turn heat into electricity (without using a cold sink as people have mentioned), the best intuitive explanation I know is the "ratchet and pawl" example from Feynman's lectures. Unfortunately, I can't find Feynman's original online, but the Wikipedia article has a decent summary. (You might want to click through to read the context.)

Although at first sight the Brownian ratchet seems to extract useful work from Brownian motion, Feynman demonstrated that if the entire device is at the same temperature, the ratchet will not rotate continuously in one direction but will move randomly back and forth, and therefore will not produce any useful work. A simple way to visualize how the machine might fail is to remember that the pawl itself will undergo Brownian motion. The pawl therefore will intermittently fail by allowing the ratchet to slip backward or not allowing it to slip forward. Feynman demonstrated that if the temperature of the ratchet and pawl is the same as the temperature of the paddle, then the failure rate must equal the rate at which the ratchet ratchets forward, so that no net motion results over long enough periods or in an ensemble averaged sense.2 A simple but rigorous proof that no net motion occurs no matter what shape the teeth are was given by Magnasco.

  • 1
    $\begingroup$ Fine, lets just replace the ratchet with a tiny magnet, place a tiny coil next to it and use a rectifier to convert those pulses to DC electricity instead? Or just connect the coil to some restive load thus converting the energy of the Brownian motion into heat somewhere else (which can be hotter than the source reservoir). And this would be a perpetual motion engine of the second kind. $\endgroup$ – Zar Shardan Apr 27 '17 at 23:05
  • 1
    $\begingroup$ physics.stackexchange.com/questions/238090/… $\endgroup$ – Mark Eichenlaub Apr 28 '17 at 3:02

One can operate a solar air conditioning with the electricity from the solar panels. . There are several methods in the article , some passive.

Also there are geothermal heatpumps which could be combined with solar panels to have a completely independent from the grid air conditioning/heating system.

enter image description here


As of last week, scientists have created a thermoelectric material that increases its efficiency to 15-20%. This is still not good enough to heat or cool your house, but it gives me hope that someday in the future, we have materials so good and cheap that we can build our houses and cars with it. So we may not be able to use the heat inside the house to cool the house but we can use the temperature difference between the inside and outside to generate electricity to keep the house warm or cool. Like you, I hope that someday we will use very little energy from the grid.

As for today technology, we have windows that heat can and cool our houses if they face the sun as explained here: http://www.chelseagreen.com/content/celebrate-the-winter-solstice-by-using-windows-to-heat-your-home/


Heat is something that comes into existence ONLY when there is a temperature difference. Heat is a flowing energy. You feel heat because the difference in your body temp and the atmospheric temp. You cannot use heat between the SAME atmospheric temperatures.

A method would be to enhance the temperature using solar panels, and store the energy in batteries or in other form like use this this energy in a continuous chemical reaction and then store the "useful" gas in a pressurised cylinder.


protected by Qmechanic Mar 8 '14 at 17:49

Thank you for your interest in this question. Because it has attracted low-quality or spam answers that had to be removed, posting an answer now requires 10 reputation on this site (the association bonus does not count).

Would you like to answer one of these unanswered questions instead?

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