New thermoelectric cooling breakthrough nearly doubles efficiency | Hacker News Reader

New thermoelectric cooling breakthrough nearly doubles efficiency

sciencedaily.com

104 pointswesturner60 comments

Doubles efficiency? From what I understand, the efficiency was pretty lousy to start with: https://www.youtube.com/watch?v=CnMRePtHMZY

Aren't there theoretical limits to this sort of cooling too?

But, if this innovation causes Technology Connections to make yet another heat pumps video, I'm all for it.

Efficiency is pretty lousy. But sometimes you don't want to lug around a compressor, like inside a portable device (such as a food cooler, portable computer, or even a prosthetic limb). In that case inefficient is often better than nothing.

But also if this is an indication that more research effort is going into this because of breakthroughs, that’s a good thing. Double the efficiency a few more times and pretty soon it could be competitive in many uses.

Solar panels used to be horrible at efficiency. Now they’re pretty amazing and extremely competitive in the power generation market. It, similarly, took a few decades of these kinds of efficiency improvements to get there.

> Solar panels used to be horrible at efficiency. Now they’re pretty amazing and extremely competitive in the power generation market.

To be fair, in my understanding economies of scale kicked in _hard_ for solar, making it much cheaper to produce single panels, moreso than making each individual panel more efficient.

Ehh, we have managed to make pretty tiny compressors these days. Small enough that people are putting AC units into cosplay costumes. You can also buy compressor based ice coolers now that are lighter and last longer than one full of ice or one full of ice and peltier coolers.

A lot of them are small enough to power with a single Solar panel too. The ice chest ones usually use about 48 watts.

I'll throw in my own niche usecase for thermoelectric - deep sky imaging.

I do astrophotography as a hobby (taking pictures of galaxies/nebulae/etc). This involves doing long, multi-minute, exposures repeatedly all night long. The heat generated from such long exposures produces what are called "hot pixels" - noise generated by heat disrupting the electronics.

To combat this, dedicated deep sky imaging cameras run a thermoelectric cooler to bring the sensor to sub-zero temperatures. These sensors tend to be about 23.5×15.7mm in size.

Niche, sure. But I saw an opportunity to bring up an aspect of my own life where I sincerely do not think even the smallest compressors would be practical. Plus, the lack of moving parts in thermoelectric is very attractive.

HA! I know we're not supposed to do "me too" posts, but literally the second I saw the headline I was thinking of the Technology Connections videos I've seen comparing thermoelectric to <anything else> and how inefficient it is in comparison. My memory is more than 2x, but I figure it's worth checking the videos to be sure before making an ass of myself.

But before that I figure it's worth it to check the comments already to see what people are thinking. And of course, your comment is at the top of the list.

Someone should code an "HN TC poster-bot" that scans headlines for topic matches and just immediately posts the relevant TC video. All of which to say, TC is awesome and everyone should check it out.

Well, the thing is, heat pumps/compressors can be more than 100% efficient by using phase change, they can be "200-300% efficient". It's not clear that TEC will be able to do that, and they're currently around 5-10% efficiency.

They're good for active cooling phone holders, maybe a chilling cupholder or stuff like that. Novelty items where you have USB power and want to cool something maybe 10-20 degrees below ambient and can have a heatsink/fan on the hot side. You can stack them for a bigger delta T, but the efficiency drops even more, and you end up needing to use increasingly smaller modules on the cold side, so only good if you need to get something small and well-insulted down to very cold temps, not active cooling.

10% efficient, without working fluids or moving parts, can be epicly more practical than 300% with those things. In very small or hard to reach places, peltiers are often only real option.

Yeah, pretty much. I haven't seen a compressor-based cooler that's a few cm across. But it's not something that will replace AC, fridges, or freezers, which is where they get hyped in popular media. It's a niche technology for niche applications.

> active cooling phone holders

Things I'd never think would exist much less that I would think of actual occasion for utility: my phone was in the sun (windshield), charging and performing wayfinding, resulting in "phone too hot" errors.

On the whole however, can someone explain how efficiency can be over 100%, if 100% represents conversion of inputs into all work and no entropy?

A 100w space heater will produce exactly 100w of heat. A 100w heat pump, properly installed, may turn that 100w of power into 300w of heat output on the hot side. This >100% language is marketing for people who do not understand thermodynamics or who dont want to bother with specifics.

The same is said of those scam "personal" AC devices that blow air over ice. They too are "more than 100% efficient".

> This >100% language is marketing for people who do not understand thermodynamics or who dont want to bother with specifics.

That is a little unfair: this is the language of economics rather than physics, so if you are paying for 100w and getting 300w, that is indeed over 100% efficient. SCOP is a more "correct" term but is less clear when communicating with the public and (IMO) is mostly for professionals or people with a technical understanding of heat pumps. Nobody thinks they defy the laws of physics though.

It's not exactly just marketing, it is indeed possible to "move" heat energy by using less energy than the amount of heat energy moved, which all modern high efficiency compressor based air conditioners/refrigerators/mini splits/heat pumps can do. It's still a useful metric for comparing similar devices' efficiencies. It's called Coefficient of Performance and if it's only 1, your compressor based system has a major issue.

What about NHL-sized ice arena using TEC? What's the theoretical limits/bottlenecks here? (asking for a friend)

Heat loss due to resistive heating of the thermoelectric elements is a factor.

I do some work using TECs for cooling optical sensors. Most of the industry has settled on a very small range of materials, such as BiTe, meaning that we're probably at a plateau until a more efficient material comes along. Some of my applications are battery powered, and cooling consumes a fair portion of the total power budget, so a more efficient cooler would be welcome.

Also, TECs are mechanically delicate. There's a laundry list of needs such as being able to form the material into the desired shapes, and make a good mechanical and electrical connection to it, which is usually done with some kind of low melting solder.

It'll be ludicrously expensive to built, and then you'll also have a ridiculous amount of heat to get rid of. The main challenge will be keeping that from melting the ice again (this, in general, is the underlying issue with TECs: they don't have any means of stopping heat from flowing back through them, so it's a bit like trying to bail out a leaking boat, and it gets worse the larger the temperature difference you're trying to achieve. TECs max out their efficiency at a mediocre COP at very low temperature differences, in comparison to compressor based systems which can achieve better COP with more than 10x the temperature difference)

When its cold enough outside and heat pumps need to turn on their resistive elements to keep your box warm enough, having something better than resistive would be nice.

better opex but may not overcome the capex cost.

There are current applications for solid state cooling. Why wouldn't you welcome those applications being more efficient?

While Peltier devices are less efficient than a compressor/heat pump, they are a lot quieter due to being completely solid state. Comparing the two technologies is tempting but the applications can be very different.

Since these devices can also produce power given a heat differential, they are used in spaces where you need just a little power and heat is readily available.

By far their most useful application is precision thermal control. Need that LED or laser to be within 0.05C to maintain its wavelength? Literally nothing else can do it.

Also camera sensors. AFAIK most (all?) high speed sensors are actively cooled with TEC. All high sensitivity cameras are, most telescopes and satellite cameras also i think

interesting to know that there are actual critical applications for this technology though it does have limitations as I think the JWST is cooled with some sort of exotic compressor useing liquid helium, though perhaps a solid state device is part of the stack as they are after a very low, very stable temperature

JWST is special in that it needs to avoid radiated photons for its IR images.

The way these systems are controlled (in my terrestial experience) is to have the system be capable of cooling below the desired set point. Then you can PID the temperature using a heater or a thermal bridge/heat switch that will cause the system to warm up. It's complicated, and a lot of very interesting thermodynamics involved.

It also depends if you just need "as cold as possible" or if you need to know the temperature and can calibrate out fluctuation, or if you genuinely need stability.

https://bluefors.com/stories/differences-between-pulse-tube-...

Also sensors with long exposure times, which are cooled to reduce the dark current. Amateur telescope camera chips are often actively cooled for long exposures.

Yea, plenty of cameras need precise temperature control too. A laser or LED shifting spectra is just the easiest and most accessible example I could think of.

Highspeed needing it is a head scratcher though. Best I can come up with is the sensor needs relatively low temperatures, that or as the frame gets smaller on the sensor to up the framerate the power density goes nuts.

The highest sensitivity cameras tend to use cryocooling rather than TECS. But you're correct that a lot of mid to high end astro cameras are Peltier cooled, and many hobbyist imagers. Satellites are a mix, you have to get rid of the heat, so some missions also use cryocooling and are limited by the amount of cryogen on board. Some instruments also get away with passive cooling, if you can radiate it fast enough.

https://science.nasa.gov/mission/webb/cryocooler/

ScholarlyArticle: "Nano-engineered thin-film thermoelectric materials enable practical solid-state refrigeration" (2025) https://www.nature.com/articles/s41467-025-59698-y :

> Abstract: Refrigeration needs are increasing worldwide with a demand for alternates to bulky poorly scalable vapor compression systems. Here, we demonstrate the first proof of practical solid-state refrigeration, using nano-engineered controlled hierarchically engineered superlattice thin-film thermoelectric materials. [...] The improved efficiency and ultra-low thermoelectric materials usage herald a new beginning in solid-state refrigeration.

Previous discussion: https://news.ycombinator.com/item?id=44424087

I tried gpt to get an idea of what this meant relative to compressor tech and i had to put my tail between my legs and crawl back in my cave. it said it could be COP very efficient iirc COP 50 over tiny temp lifts. i don’t fully understand that yet. But the theme is matching the application with techs characteristics. solid state should be great but as someone mentioned, expansion and contraction at the solder or connection is troublesome. on the other hand, those that mention it wouldn’t work for residential cooling/heating should keep in mind the cost of installation for current heat pumps- if this cost could be reduced or homeowner-capable, then that’s a huge win. My point is is this constellation of characteristics the SOTA and how they interact with the moving world can be hard to track and if something isn’t viable, it may suddenly become viable. it’s a tricky thing to notice if you’re just looking at the one characteristic of efficiency.

To note, the lead investigator on this is Rama Venkatasubramanian, an Indian:

https://www.jhuapl.edu/about/people/rama-venkatasubramanian

This is the same sort of person that MAGA and tech workers want to discourage coming to the United States.

> This is the same sort of person that MAGA and tech workers want to discourage coming to the United States.

Tangential, but let's not forget that there are also Indian MAGAs like the current FBI director.

There are also influential and wealthy Indians who are more than willing to partner with MAGA to pay whatever tribute is necessary to curry favor, and throw working Indian immigrants under the bus to line their own pockets and accomplish their own personal and political objectives.

In the end, it has little to do with being Indian and more to do with scapegoating and dividing tech workers among themselves, to distract them from the squeeze that big corporations are putting on them.

The BJPs brand of ethnonationalism does have a kinship with MAGA as we've seen with the bromance Trump and Modi. But the recent tariffs have greatly changed many Indians' views on Trump, though I haven't seen anything about the views of Indian Americans specifically.

I can't help but be reminded of Steve Bannon's comments about their being too many Indian tech CEOs.

You might be confusing the specific with the substantive.

You might be confident the intended with the substantive. The intended effect is to raise American salaries. The substantive that is the MAGA’s H1-B and larger immigration policies will discourage STEM researchers to come the US. This will hurt US scientific research and is likely the most substantive long term effect on the nation. Any potential effects on tech workers salaries pales in comparison. This potential also includes lower developers salaries as American companies lose their competitive edge.

Is this similar to Peltier coolers?

Yes, "thermoelectric" and "Peltier" are the same thing.

You can double the speed of a slug and not have it do anything meaningful as a useful domestic animal.

Thermoelectric cooling is extremely inefficient, to the point that we have very little practical use for it right now. Heat pumps a hundred times more effective predominate.

TEC are used everywhere, actually.

Heat pumps can move a lot of heat efficiently. Thermoelectric coolers can move heat more precisely and in much smaller packages. They don't need to beat heat pumps at their game to be useful for their own game.

Scaling up (eh, out?) MEMS compressors in matrixes seems a more likely path to success to me. Some research is ongoing there. I’m not expert, but I find it interesting.

Holy ads Batman - https://imgur.com/a/6vwJBkT

One of the greatest need for energy in the next few years is going to be for air conditioning if this really works it is going to be put another nail in the coffin of Oil and gas industry.

Air conditioning is mostly needed while the sun is shining, and so electricity of A/C can come mostly from photovoltaics (plus batteries). So I think this technology is not quite as important. Sure, its nice to reduce electricity usage! But cheaper heat storage, for winter, seems more important.

If you're heating with non-fossil-fuels you're doing it with heat pumps, just like AC, just with the hot side inside and the cool side outside.

The press release doesn't give any concrete numbers, but if it doubles efficiency of Peltier coolers, it's still 3-5× less efficient than heat pumps.

Thermoelectric cooling is notable for not having any moving parts and ability to scale down to small sizes, so it might end up having many specialized applications, but for A/C heat pumps are already very effective.

And what about service life? I had a mini-fridge that used this technology, and it stopped working after about 2 years. Was that just bad luck or poor quality, or some inherent lifetime of the components?

In principle peltier elements should be very robust over time, as a solid state system where the only moving parts are fans (versus traditional refrigeration which includes a high pressure pump...).

In practice I strongly suspect most peltier based systems are built very cheaply... because their inefficiency means the majority of the market is bordering on a scam. Sophisticated consumers aren't going to be buying very many fridges built with them (of course you might have a niche use case where they actually make sense and you're willing to pay for a quality product, but do most purchasers?).

Thermal cycles is murder on rigid electronic connections; the mechanical connection between the heatsink on each side of the peltier cell being a prime example.

I mean USA is now subsidizing coal and wants to double toxic fracking

"better" "cheaper" "cleaner" is buried by political spite agendas

You can't beat gas phase change for moving heat around. But there are a lot of applications for small coolers where a compressor is too bulky.

Yes you can! Not with peltier elements, but with elastocaloric (phase changes in solid metal induced by stress) and maybe magnetocaloric (phase change induced by magnetic fields) heat pumps.

There's engineering challenges here, but I believe the science is pretty clear that in principle these beat gas phase change systems.

> I believe the science is pretty clear that in principle these beat gas phase change systems.

Do you know by how much?

Gas phase change systems in practice are up to about 60% efficiency (relative to an idealized carnot cycle), magnetocaloric systems in labs have been something like 70%, and elastocaloric 80%.

Edit: And just for reference peltier devices are usually 10-15% efficiency on this scale. So I suppose this new one is somewhere in the 20-30% range.

Probably very applicable, this video from Ben

https://www.youtube.com/watch?v=lfmrvxB154w

IIRC from reading up on these in the past, you need a pretty large input energy in order to make them work. The elastocaloric ones are also super loud IIRC.

This has nothing to do with that application.

Having looked at the paper, I don't much like their testing methodology: calorimetry is really hard (a general theme of cold-fusion 'research'), and they don't really seem to have calibrated their test setup, just calculated a theoretical value. This makes me highly skeptical of claims of COP, which can be highly sensitive to small differences in heat flows.

Reply on news.ycombinator.com