Thermoelectric Solar Panel | Hacker News Reader

Thermoelectric Solar Panel

simplifier.neocities.org

50 pointsadrianN60 comments

Compare this to a Phonovoltaic cell, which is similar to a photovoltaic cell, but instead of converting the energy of a photon into electricity, it converts the energy of phonons - packets of heat vibrations in solids [1].

Unfortunately, no one has as yet found a material with the right combination of thermodynamic properties to realize such devices. Otherwise, you could stick these to any source of high temperature heat and get electricity out.

[1] https://en.wikipedia.org/wiki/Phonovoltaic

Efficiency of 0.25%. Like a dancing bear, the amazing thing is that it dances at all, not that it dances well.

That's still 2.5 watts per square meter in direct sun, 0.5 watts per square meter as a round-the-clock average. More than enough to keep your cellphone charged with a square meter. A section of a farm is 2.6 million square meters.

The bigger question is not what its efficiency is, which is to say, how much sunlight it needs per watt, but what its material consumption is—how much bismuth it needs per watt. Sunlight is abundant; bismuth isn't.

But presumably it can be improved.

Or just cover 1% of that area with PV.

Can you make PV at home?

Thermoelectric devices are interesting because they can be manufactured using comparatively simple metallurgy. Not because their efficiency is competitive with semiconductor photovoltaics.

> Can you make PV at home?

https://hackaday.com/2018/01/26/home-brew-solar-cells-for-th...

Not anywhere near as efficient as commercial PV cells, of course.

This is by the same person as the OP btw (and linked in the original article).

So why did you ask the question, if it was answered in the original article?

Rethorical question, sorry.

Perovskite solar cells could be and dye sensitized solar cells can be[0]. The better question is why should one make solar cells at home?

[0]https://www.instructables.com/How-to-Build-Use-A-Dye-Sensiti...

Solar panels enabling offgrid power is tagcloud-related to surviving human civilization collapse.

We're entering an era of decreased globalism, where megacorporation scale actually becomes a danger to society due to reduced warehousing/stockpiling and long extended supply lines, and of course offshored manufacturing that goes with that.

Before solar panels become impossible to purchase, they will be difficult to purchase.

Before solar panels become difficult to purchase, they will become more expensive to purchase.

Mass produced solar panels have been getting both cheaper and easier to get. What you describe could happen, but it's far enough off that we have not seen even the first warning signs yet.

"What you describe could happen, but it's far enough off that we have not seen even the first warning signs yet."

Well, some disagree on there being no warning signs.

https://thebulletin.org/doomsday-clock/2025-statement/

Yeah, if some shithead nukes China, solar panels would get expensive in a hurry.

Okay, sure, but that's no different than the same identical "warning signs" that people have been flipping out about since the 1960s.

Maybe some individual country will have some collapse, but all of human civilization will not.

I.maintain that if you're worried about any of that, you are still better off stockpiling panels now, than developing ways to make your own much worse ones in your backyard.

DIY photovoltaics can be a fun hobby but for someone worried about an actual collapse they can acquire a lifetime supply of solar panels and batteries for less money and time than setting up a custom fab that's able to operate in a collapse situation.

Is your argument because we have not had a nuclear war yet, we will never have one?

Can you elaborate? I would like to have your confidence ..

And in general setting up a solar panel fab is maybe not the best prepper action, but for the point of distributing critical techologies for a potential reconstruction, I do see the point.

There is individual survival and general progress of the species.

There's exactly one scenario that results in the actual technological collapse of our species, which is all out nuclear war between the US and Russia. Which with the current presidential administration, is possibly less likely than it's ever been; why would Russia nuke its newest ally?

The chances of a nuclear weapon being used somewhere right now tactically I think are quite high. Russia in Ukraine, or Israel in Iran (or someday soon Iran in Israel), or between India and Pakistan. But none of those are sufficient to bring us to a point where home manufacturing of solar panels becomes remotely worthwhile.

The problem is, except for the very first ones in Japan, there were never any nuclear weapons being used in a major war exactly for the unknown consequences of the other players.

It is all interconnected. If one nuke is used, then there will be many on the other side applying pressure to also use a nuke. And so on. I assume much more countries secretly have nukes and the frontlines are somewhat blurry. Meaning, at the moment I am also not too worried, but if a nuke is used it will be a very high gamble, that it will be just the only one.

Most countries definitely do not have nukes. There are a handful that could have them in secret, or that maintain the materials to make them immediately if needed. But all of those combined would amount to no more than a few dozen, small, nuclear weapons. There are not a thousand Tsar Bomba size nukes secreted away.

If the US and Russia stood down and the rest of the world let loose all of their nukes, it would be insufficient to cause sufficient damage to the technological integrity of our species such that backyard solar manufacture becomes viable.

"Most countries definitely do not have nukes."

I did not claim that.

I claim it will be hard to limit the use of nuclear weapons.

Just like in the weapons itself, one ignition can trigger a chain reaction in the end forcing russia or US to take part in it as well. If all the people involved are level headed and able to think rational - it likely will prevented also in the future. But if the person in charge is already stressed (and old) and gets lots of pressure - this person, might then feel forced to press a button.

You said "I assume much more countries secretly have nukes".

I claim that beyond the open secret that is Israel, the number of countries controlling a right-now detonatable nuclear weapon who are not on the Wikipedia list of "countries with nukes" is less than 5.

Claiming much more countries secretly have one, is not at all the same as claiming most countries secretly have them. And I agree that it won't be many many, but in this context one previously unknown bomb already might change lots of geopolitical equations and their outcome.

No one ever said most.

You said much more. I think it's fewer than 5.

"No one ever said most."

"Most countries definitely do not have nukes. "

https://news.ycombinator.com/item?id=43209283

Also by my understanding, even 4 more countries having secretly nukes counts as "much more countries having nukes than officially known".

But my point is not semantics, but that even one nuke can trigger a chain reaction. Unknown nukes just makes this more likely.

Are you illiterate? You think that's what "most" meant in context?

I believe it's optimistic to think that that is the only scenario. Consider the problems we had during the pandemic, which was luckily just a minor blip as far as possible global disruptions go.

Agreed. But one individual country makes 90% of the panels, and the others cost twice as much.

Okay, but again, it's a far cry from "the price doubles" to "make these in your backyard or you will not have them".

To make a good, efficient solar panel requires materials and equipment incompatible with backyard manufacture, especially in a collapse scenario.

It's a neat hobby, but if your goal is just "have solar panels", you are strictly better off buying them than making them.

Probably. But all of these questions are contingent on what has been invented so far.

hacker spirit and/or hedge against the end of the industrial age

Personally, I don't have much faith in sustaining a modern industrial lifestyle for billions of people very long term on PV panels or any renewables really (EROEI issues, rare earth minerals etc.). But I'm bullish on electricity in general.

By far the most common PV systems are single-junction monocrystalline silicon. These have had pretty decent ERoEI for many years (9–10 according to https://www.nrel.gov/docs/fy17osti/67901.pdf, but the energy required for PV modules has dropped precipitously since then) and don't use any rare-earth minerals.

Wind turbines have typically had even higher ERoEI (15–20 according to https://davidturver.substack.com/p/eroei-eroi-of-onshore-off...), and, while they do most commonly use rare earths, that's an engineering tradeoff rather than necessary; currently viable alternatives include switched-reluctance-machine generators (which just use conventional electrical steel, like transformers and relays) and other kinds of rare-earth-free generators: https://www.sciencedirect.com/science/article/pii/S030142071...

You can't reach 25% efficiency with monocrystalline silicon PV even with the entire industrial economy backing you up; those efficiencies require multijunction solar cells. Conveniently homebrewable PV panels, like the ZnO:Mg/Cu2O cells you link below, are closer to 1% efficient. The author reports 525mV and 0.9mA/cm², which works out to 4.7W/m², which would be 0.5% efficient if the illumination is one sun.

They probably are a better approach than the thermoelectric approach, because not only are they more efficient, they use more abundant materials and thinner films of them. But either approach could plausibly defeat the other.

Oh good grief. The point I was making didn't require that it be exactly 1%.

With Simplifier's homebrew PV cells, it's closer to 50%.

Obviously I wasn't referring to those, which are also a "dancing bear" technology.

Longi has demonstrated a champion monocrystalline silicon cell at 27.3% efficiency and it has silicon cells over 26% efficiency in mass production. It has demonstrated a champion silicon module at 25.4% efficiency.

https://www.longi.com/eu/news/25-4-module-efficiency-world-r...

Wow, I appreciate the correction! I'm apparently several months out of date.

You can make Schottky solar cells too with bismuth but as others have pointed out optimum operating temperature is different for thermoelectrics.

The best use of themovoltaics I've seen is

https://www.rockfordchimneysupply.com/products/thermo-self-p...

since you've got a concentrated heat source that you want to spread around. Themoelectric devices have been active research field for a long time and there is still some hope of improvement because it is a largely a matter of discovering materials with a high electrical conductivity and poor thermal conductivity.

https://en.wikipedia.org/wiki/Thermoelectric_effect

You'll get better efficiency running a Stirling engine from a lightweight solar collector. I actually wonder how it would compare if you equalize the weight.

That's an excellent idea, especially if you have aluminized mylar or stainless steel.

That's one of those areas that's been talked about in the literature ceaselessly since the 1970s with almost 0% commercialization. That's the normal scenario for energy technologies, cases like PVs and lithium batteries are the rare counterexample.

This is different. This person wants a solution that can be built with common materials with fairly standard tools and at small scale, not a scalable commercial power plant. You can build Stirling engines at home with common materials. We're not comparing 10MW power plants here.

So the question is, for a comparable weight and size as this solution, how difficult would it be to build a solar powered Stirling engine and what would the power output be? I expect the Stirling engine to easily come out on top.

rankine cycle lets you cross vapor dome to access latent heat. enables scaleup

Is there a reason? It's curious to me that you don't see many Stirling engines. Is there a big downside to them?

There is this

https://www.stirlingengine.com/why-not-popular/

I've heard that the manufacturing tolerances are tough.

One problem is that people systematically publish optimistic cost estimates for new technologies and that these don't get updated. I still see papers that cut-and-paste cost estimates for various technologies from 1970s papers and don't even bother to adjust for inflation.

You can certainly couple a Stirling engine to a parabolic dish

https://en.wikipedia.org/wiki/Solar-powered_Stirling_engine

but it just isn't cost effective with PV when you consider the cost of the engine, the cost of the dish, etc. And think of all the moving parts to maintain! A big solar thermal plant looks ready to shut down

https://www.greenprophet.com/2025/02/ivanpah-fails-value/

not just because the capital costs were high but because the operating costs are high. It's hard for anything to compete with PV because the operating costs are low (you clean them once in a while) and the capital cost keeps dropping because it is so fiercely competitive -- something that happens rarely (hydrofracking for hydrocarbons was another example)

Ivanpah failed because of the maintenance costs of the motors moving the reflectors, not the heat engine. Those motors are not really a factor in the kind of setup comparable to this article. Stirling engines actually require considerably less maintenance than combustion engines, and the maintenance is simpler because the engines are simpler.

Manufacturing tolerances are only relevant if you want to maximize efficiency for competition at commercial scale power production, which is not this case.

Conventional solar cells also suffer from decreased efficiency (-0.45%/°C) as temperature increases, a the inclusion of TEGs could help dissipate this waste heat while simultaneously increasing solar panel efficiency.

Minor correction/note: most modern modules/cells have a temperature coefficient in the range of approx. -0.27 to -0.35%/°C. A little better than -0.45%.

See fig 14 in [1] for data through 2021/22, and then the more recent transition to n-type cells is helping more [2]. This database [3] doesn't list module release date, but filtering for modules with STC rating over 500 W (decent proxy for "modern") gives an average of -0.34%°C, with a lot at or better than -0.30.

[1] https://www.ise.fraunhofer.de/content/dam/ise/de/documents/p... [2] https://www.nrel.gov/docs/fy22osti/82871.pdf [3] https://github.com/NREL/SAM/blob/patch/deploy/libraries/CEC%...

I've done some research on exactly this, and the efficiencies of TEGs makes this a tough prospect. It started to work out in our favor when using concentrated solar to reduce the amount of paneling needed, but that's no longer an economic driver, and you get a much better return on investment by skipping the TEGs and just plopping a few extra panels down in a lake. You can also flip the configuration and use TEGs to generate power from the solar heating, but the terrible efficiency bites us again and you're better off buying more panels instead.

In the domestic/commerical, the simplest solution to this is to add more solar to compensate for the drop in efficiency. The next cheapest approach would be to cool the panels by heating water for domestic/industrial use, although the economics aren't as straightforward (depends on the cost of hearting fuel being displaced).

> although the economics aren't as straightforward

And you're multiplying the points of failure and complexity of the system. That's the beauty of solar panel imho, not much an go wrong, adding water, potentially boiling water that is, pipes, pumps, &c. sounds like an headache

I agree that today it doesn't make cost sense, especially in a retrofit scenario and when the marginal cost of additional solar panels is so low.

However, heat removal methods could make sense for much more efficient, but much more temperature-sensitive solar PV materials like perovskites.

TEGs will concentrate heat. A TEG adds significant thermal resistance to your system. For a TEG to produce meaningful energy, you'd almost certainly need a heat sink of some sort. From a total system efficiency and cost perspective, you'd probably be better off just directly mounting a heat sink to the back of a panel.

Thermoelectric cells or Stirling engines could also be a good option for an off grid location that gets cold in the winter. Heat it with a wood stove, and charge your batteries off it too.

My house has a super-insulated, passive solar design. I heat it with wood pellet stove, which requires electric power to run the fans and pellet feed system. I've often wanted to add a Stirling engine and a small battery to make it truly grid independent. However, the house design means I only burned about $20 worth of pellets this winter, so the ROI isn't remotely justifiable.

What is the total mass of $20 worth of pellets?

There's already some wood pellet boilers on the market which can integrate a sterling engine, to supplement off-grid battery-backed photovoltaic solar panels for local electricity generation. The power output of the sterling engine isn't stellar, but that's to be expected, it's just to help when it's less sunny and you need space heating anyways.

For example: https://www.oekofen.com/en-gb/pellematic-condens/

Combine a Trombe Wall with a TEC.

https://en.wikipedia.org/wiki/Trombe_wall

https://en.wikipedia.org/wiki/Thermoelectric_cooling

I don't think it would make much electricity. But a Trombe wall can trap a huge amount of heat.

Reply on news.ycombinator.com