If you need to cool something, the gold standard is using a gas compressor arrangement. Of course, there are definite downsides to that, like weight, power consumption, and vibrations. There are solid-state heat pumps — the kind you see in portable coolers, for example. But, they are not terribly efficient and have limited performance.
However, researchers at Johns Hopkins, working with Samsung, have developed a new thin-film thermoelectric heat pump, which they claim is easy to fabricate, scalable, and significantly more efficient. You can see a video about the new research below.
Manufacturing requires similar processes to solar cells, and the technology can make tiny heat pumps or — in theory — coolers that could provide air conditioning for large buildings. You can read the full paper in Nature.
CHESS stands for Controlled Hierarchically Engineered Superlattice Structures. These are nano-engineered thin-film superlattices (around 25 μm thick). The design optimizes their performance in this application.
The new devices claim to be 100% more efficient at room temperature than traditional devices. In practical devices, thermoelectric devices and the systems using them have improved by around 70% to 75%. The material can also harvest power from heat differences, such as body heat. The potential small size of devices made with this technology would make them practical for wearables.
We’ve looked at the traditional modules many times. They sometimes show up in cloud chambers.
And below that:
Self-oscillating polymeric refrigerator with high energy efficiency
Article 08 May 2024
Can they stack?
Don’t see why not.
This article said two things that made me irrationally angry. Listed ‘power consumption’ as a downside to gas heat pumps and ‘heat pumps or — in theory — coolers’ as if all heat pumps aren’t coolers and heaters.
This, and statements like “100% more efficient”.
Most mechanical refrigeration applications are comparatively large, sometimes by several orders of magnitude. I am skeptical of this replacing a large centrifugal chiller anytime soon.
Somebody wake me up when they report a CoP of 3 or better.
https://en.wikipedia.org/wiki/Coefficient_of_performance
^This. “Double the performance” of traditional thermoelectrics is still nothing to write home about.
Peltier devices with a COP of 3 or better have been available for at least a decade. The key is not to drive them too hard. Keep delta T below 10 K and drive them at about 1/8 the rated current.
Of course, that reduces the potential applications and greatly increases the cost because you’d have to use many more devices to do the same job. TANSTAAFL.
The additional issue is that TECs are very good heat conductors when they’re powered off. Adding a bunch of TECs in parallel also compromises the insulation of the box you’re trying to cool down, so you have to keep them powered up to stop the heat leaking back in. There are ways to mitigate this issue, such as isolating the TEC from the box by a secondary coolant loop, but that adds complexity and cost.
And, keeping the dT below 10K is almost impossible in a practical application where the heatsink has to be small enough and integrated to the device to carry around.
So it’s going to turn those cheap peltier powered mini fridges from completely useless to just terrible.
Even if the peltier element in a typical cooler box was 70% more efficient, it’d still take the better part of a day to cool a six pack of beer to drinking temperature. The only thing those “mini fridges” can really do is keeping already cold stuff from warming up.
I’d like a cooler for my medicine whose battery would last longer doing the same job.
Try being positive for once.
It’s not about being positive, it’s about the hand waviness of the reports. I’ve seen many outlets report on this tech, and they all just parroted the same “It’s 70-100% more efficient than current tech!”, without ever mentioning even a single numerical comparison in efficiency with traditional, compressor-based systems, even if they do, in the narrative, mention the compressor-based system. It’s misleading reporting, and probably rooted in the fear that if they actually did, their tech would not compare favorably, as the numbers point to it being ~15-25% efficient vs compressors being 30-60%.
Yes.
If you can achieve a 25-30 degree F difference between the cold and hot sides then you’re doing good. Oh by the way, all of the power you’re putting into the module now is just to maintain that gradient, fighting against the heat conducting through the module from the hot side back to the cold side – so the cooling capacity with that differential is effectively zero.
On the other hand, mechanical refrigeration will be less impacted by evaporator and condenser temperatures (within design limits). The system is still pumping heat at its rated capacity even as your beer approaches optimal drinking temperature.
That said, I found that some simple (but not always practical) changes to those mini fridges can really help. As always, reducing Rth(J-A)* helps immensely. I found one unit actually worked a LOT better if I just put it on its back and filled the compartment with water. If it’s got a fan and heatsink inside, the three As are critical: Airflow, Airflow, Airflow. You’re still limited by the thermal gradient problem, so you’re never going to get down to refrigerator temps on a hot summer day.
There are some applications where TECs are still the best option available, and any performance improvement is beneficial. DWDM lasers for example.
*Thermal resistance, Junction to Alcohol
Can you expand on ‘junction to alcohol’? i cannot find that term on the Internet, while all the AIs are eager to use it as many heating/cooling applications use junctions and alcohol. Did you mean junction to ambient?
It’s just a slightly tongue-in-cheek way of saying to use something more conductive than the air inside the cooler to conduct the heat from the beverage can to the cold side of the TEC.
Fwiw it depends how well you engineer the system, not the fundamental tech, which is perfectly capable of producing decent Qc down to -20 and more if you extract the heat properly from the hot side. The problem is the mini fridges are designed to a price. It’s like expecting a mini to win at an F1 race.
And practicality. The main issue is heat rejection and insulation, which are limited in a small carry-on box or desktop size unit. Then there’s the power limits of how much you can safely draw from a car’s lighter socket – that’s what limits these things to 50 Watts (45 W peltier and 5 W fan).
What kills it is the shoddy build quality where they don’t even bother to machine the heatsink surface – just slap on a bunch of thermal grease and call it a day. As a result, the units literally do not work, and they’re all like that. All brands and models. If you manage to find a working one, it’s a first batch product that was made for the customer reviews and the subsequent units are just thrown together without care to save cost.
I’m sure we’ll see a video on the YT channel PUNdecided, I mean, Undecided w/ Matt Ferrell soon
Wonder how the cost benefit math works out for using them as solar array cooling systems. Set them up as thermoelectric generators on the back sides of the solar panels to pull heat and run very small DC fans.
TECs do not “pull heat” away from the solar panel. Any heat engine requires a temperature difference to operate, so for the TEC to operate well the panel would have to remain hotter rather than colder.
You can think of it like a hydroelectric power plant. A free running river represents the thermal gradient of heat flowing out of the back of the solar panel. When you put a dam in a river, the upstream gets higher (hotter) while the downstream goes lower (colder) and the difference in elevation (temperature) is the potential energy difference you’re using to generate power.