Source Nature Article: https://www.nature.com/articles/s41467-024-52047-5
Nature Communications... a good journal but not where I would expect them to publish what appears to be (if I'm not misunderstanding- and I probably am/must be) the most surprising experimental result in the history of physics
It seems to me that the authors are pretty cautious.
They don't claim "we found a scalable way to violate the second law of thermodynamics", but instead make very specific claims about what they measured, and that they produced 94 nano-Watts per square centimeter. They don't even mention entropy.
This feels more like they know exactly how unbelievable their results sound, and that they are very careful to not overstate anything.
Can't wait to strap some of these to the side of my desktop or the bottom of my laptop and make use of that waste heat. I'm sure I can do something or other with the approximately 2 mW I'll generate.
2 mW would work better than Sony's generator in some cases.
2mW would require about 1.5m² of area. I could maybe see this being used to pull a little bit of energy out of a room to charge batteries or capacitors to run sensors or other low power devices in short bursts. But you could harvest more power just by turning a hand crank generator for a few minutes. An average solar panel is like 10k times more efficient. I could still see some sort of use for it, maybe for being embedded into floors and ceilings of buildings to passively run sensors. There would be a small cooling effect on a room with enough of it.
but theoretically: because this does not require heat gradient (I assume that's impossible, they just base on micro gradients) you could fold this hundreds times to get much more from area
Oh, if only it were that powerful! Their best result produced 94 nanowatts per square centimeter, or 0.94 milliwatts per square meter.
Title misses the big story: "power generation ... without a temperature gradient".
Turning heat directly into electricity is one of those Trek-level technologies. Many would debate whether it is even theoretically possible, while others claim practical successes.
Indeed, I am really skeptical this is possible- it would seem to violate the laws of thermodynamics…
If the paper's claim is correct, it would literally be the creation of Maxwell's Daemon.
I am banking on the author not fully understanding their own experiment, rather than a complete overturning of thermodynamics. I would be very excited to be wrong, though.
I wouldn't hold your breath on being wrong. A device that converts heat into electricity without a gradient is a perpetual motion machine. Consider what would happen if you popped such a device on one side of a thermoelectric plate, and used its output to power a heater on the other side of the plate. The arrangement would permanently maintain a gradient across the thermoelectric plate - free electricity forever!
Not if it produces less heat than it converts, the rest being lost as ... ah, wait a minute. Well anyway it would only produce as much heat as it converts, so this perpetual cycle couldn't do any useful work. Perpetual useless motion is allowed, isn't it?
Oh, I see it now, you're suggesting that it could just passively create a gradient that could produce power in a more conventional way, like via an engine, until I guess the device kicks in again and re-establishes the gradient (producing more power of its own in the process). Yes. That does sound like impossible free energy.
But so do solar panels. A PV system turns a moving particle (a photon) into an energy gradient, a voltage. So the concept of using another moving particle, this time a hot molecule, doesn't seem totally impossible. If one understands heat as being molecules moving at variety of speeds, harvesting energy only from the fastest of them wouldn't violate thermodynamics. It can be understood as taking advantage of the temperature gradients across the gaps between individual molecules.
It sounds like you have basically given a formulation of Maxwell’s Demon https://en.wikipedia.org/wiki/Maxwell%27s_demon
“All you have to do is harvest the hottest/fastest molecules” pretty much describes exactly what Maxwell’s Daemon is.
Now, you can see from the Wikipedia page that there is at least some debate about the topic. But, I think that generally, most physicists think that Maxwell’s Demon would violate the 2nd law, and that the 2nd law is a real limitation.
It certainly sounds like this device would violate the “metaphysics” definition of the 2nd law, “entropy always increases”. Because it sounds like it’s harvesting energy from heat for “free” (not actually energy from nothing, not violating conservation, but free in that you are not paying for the loss of entropy by increasing entropy somewhere else, which is typically “required”).
I am not a physicist, and in my layman’s understanding, I have always felt that the 2nd law seemed fishy. A lot of other people do, too. It’s called “a scientific law”, but it feels more like a philosophical “principle”. It feels different than, for example, “the law of gravity”.
But, this invention, if it is not a hoax or error, sounds world changing. 94 nW/cm2 means a milliwat for 2 square inches. That’s already usable amounts of power for tiny devices. Can you roll up a couple square meters of this stuff into something like a capacitor and get a usable “free” AA battery?
Is heat about to be free energy? No more problems with global warming? Are our refrigerators and air conditioners soon going to be a source of energy rather than a sink?
If this claim is true, then this is world changing technology available to anyone with pretty basic thin film technology…. And there are YouTubers who will be replicating this soon. Let alone real fabrication labs.
You can see why it’s reasonable to be skeptical…
The laws of thermodynamics are empirical: all of our physics experiments and observations thus far support and agree with them. We have no external confirmation, e.g. "god's source code for the universe" or some such that proves them to be absolute laws, but given that they hold up consistently across a huge number of observations, in a huge number of domains- we are nearly certain that they are absolute. Gravity itself (e.g. general relativity) is also on the same footing- like thermodynamics it has proven itself to reliably predict observations across a huge number of different new experiments.
This reasoning isn't convincing enough for some (non-physicist) people- there are whole forums of people working on perpetual motion machine designs, etc. that think thermodynamics is nonsense. However, thus far none of their machines have ever worked.
Although I am not a working physicist, I was trained as a physicist (a long time ago!), and assign a prior of essentially 100% to the laws of thermodynamics, from a huge amount of evidence. E.g. I am certain that either I am misunderstanding this paper, or it is simply wrong. I'm hoping someone whose physics isn't as rusty as mine will step in and explain what I am missing here.
I agree. In my opinion this is direct Nobel Price or big honest error [1]. I'm really surprised they didn't add a discussion about the obvious violation of the second law and that the referee didn't ask for one neither. Anyway, I'm almost sure (100-10^15)% it's just an experimental error.
[1] or fraud or bullshit, but let's be assume good intentions.
This is false. Solar panels do not violate the laws of thermodynamics. Generally the sun is quite a bit hotter than the environment in which the solar panels operate, which are the correct temperatures to compare in this case.
"harvesting energy only from the fastest of them" actually does violate thermodynamics and, if possible, would constitute the biggest revolution in physics in all of history.
They don't, but depending on one's understanding an perspective they can be described as violating. The point is that just because something can possibly be described as a violation does not mean that there are perhaps other perspectives that may understand them without such violation.
The paper claims to generate energy "without a temperature gradient." If there's a hidden temperature gradient somewhere, then the paper is still incorrect.
Given the extremely small power levels, I expect that if it's not a hidden temperature gradient (eg between local air temperature and local radiant temperature, similar to PV) then they're actually seeing some sort of static electricity or air movement effect.
Yes, it is probably better described as a shitty battery NOT a thermoelectric.
>But so do solar panels. A PV system turns a moving particle (a photon) into an energy gradient, a voltage.
Solar panels are not an example of a Carnot violation. As far as we know, no such examples exist.
https://en.wikipedia.org/wiki/Solar_cell_efficiency#Thermody...
> Solar panels only work because there's a temperature difference between the panel and the optical surface of the Sun.
I believe this isn't quite true (even though it's often repeated).
Consider that you could have an LED, at the same temperature as the PV cell, emitting light onto the PC cell, and so the PV cell would "generate" power without a temperature differential.
The Carnot cycle applies (in its initial form, at least), ONLY to generating power from temperature differentials, but PV uses the photoelectric effect, not temperature differentials.
This is not my area of expertise, but I think you can generalize the Carnot efficiency to talk about low-entropy energy sources instead of temperature differentials, but it's not quite as simple as associating a light wavelength with a temperature, because that doesn't work for radiation from something that's not a black body (like a laser or an LED).
I think here there is a misunderstanding on the meaning of 'work'. It's most likely clear to all authors in this chain that if you take an LED and shine it onto an appropriate PV you will get voltage out of it. That much isn't governed by thermodynamic limits.
Likely the word 'work' here is in the context of collecting or generating energy in some manner. Your example here is essentially energy transfer. In analogy what you're describing is the transfer of power from driveshaft to axle, not what is happening in the engine itself.
You still need the photons to hit the surface of the object and deposit their energy. Talking about the temperature of the surface of the sun is just a simpler way to refer to the energy output. What you are saying is a very high level view of the actual physics involved. The physics just works out that we can use temperature to talk about this.
This is the same argument the authors make to reviewer #3:
"At this moment, we think that the thermal energy can be supplied to the devices from the surrounding air which possesses thermal energy such as 25 meV at room temperature. The air surrounding the devices can be considered as an infinite heat bath, and the analogized conditions of steady light irradiation in a solar cell, i.e., “photon absorption,” are considered to be formed by “phonon absorption.”"
This sounds like nonsense to me- I don't think a solar cell would function without a gradient either, e.g. if the cell itself were the same temperature as the light source.
Does a photon carry information about the temperature of it's source?
What about photons from non-thermal sources? A PV panel is able to generate electrical current from photons generated by LEDs, bioluminescence, and indirect light reflected off a cold surface.
[Hard question!]
Each photon, no. But the distribution of the spectrum of a black body follows a pattern that depends on the temperature (and other stuff).
<bold>The important part is that the PV panel not only absorbs photons, it also emits photons. Usually the emission is much much much smaller than the light it absorbs.</bold>
If you assume that the PV panel is another black body and the sun is another black body, then the energy must flow from the hot one to the cold one.
Real objects are not perfect black bodies, so you have emission and absorcion coefficients that makes the calculation harder.
For LEDs, bioluminescence, lasers, and other sources, the calculation is harder because the distribution of light in the spectrum is different, and they usually don't match. You must consider work, entropy, and more technical stuff. But the oversimplified version is that the energy also flows from the "hot" to the "cold" one. You need a gradient of temperature or something equivalent to make the energy flow.
PS1: You can use a LED as a light detector, like a photodiode. It's somewhat the inverse effect of using a PV panel as a lamp.
PS2: I know that <bold></bold> doesn't work in HN, but that it's the important paragraph.
Yes, indirectly- this is how infrared thermometers can measure temperature at a distance. If there is no temperature differential then both surfaces will be radiating the same number of photons at one another, and therefore radiate at least as much energy as they absorb.
For a “non thermal” source, the thermodynamics aren’t changed, you are still radiating the energy you put into it, and a sufficiently “hot” absorber will be unable to use the energy. The thermodynamics are the same, you just have more control over the frequency and direction of the radiation.
No, that still violates thermodynamics.
Thermodynamics says "no", at least on macroscopic scales.
My thermodynamics lectures are some 20 years in the past, so my memory is a bit rusty, but iirc you need "free energy", that is, usable energy in the form of a low-entropy energy source, which is usually a temperature gradient (but could also be a radiation source or something else).
To get to these results, you have to do statistics of large numbers, so it's theoretically possible to violate this on the nano scale.
The optimist in me hopes that we've found a scalable way to exploit this, the pessimist/skeptic in me says it's another cold fusion moment.
Obligatory https://www.youtube.com/watch?v=tuxbMfKO9Pg
What’s more probable, violation of the second law, or some experimental error? 94 nano Watts is not a lot of power. I’m not holding my breath…
Sounds like a great idea for wildlife tracking.
They already have miniature kinetic energy generators, now they can pair with heat difference generators (air vs animal body temp)