The Thermoelectric effect (TE) comes in two variants:
- The Seebeck effect makes it possible to turn a heat flow (based on a temperature difference) into electric power.
- The peltier effect does the reverse: it turns an electric current into a temperature difference across the two sides of the device.
The fundamental mechanism is a p-n-transition. So you have two different semiconducting materials, which means that the electrons are on different energy levels on both sides. When the electrons move from one side to the other, they have to absorb energy from the environment to get on the higher energy level themselves (p->n transition), or they give off energy (n->p transition), thus cooling or heating the environment.
With this technology, it is possible to build solid-state heat pumps that generate a temperature difference from an electric current with no moving parts! (i don’t know about efficiency or cost)
You must log in or # to comment.
Wanna listen to a Midwestern nerd talk at length about how awful these kinds of devices are with respect to efficiency? I got you…
Technology Connections - Thermoelectric cooling: it’s not great.
It’s extremely inefficient.
I looked into replacing the alternator in my car with a bunch of Peltier devices in the exhaust system. Yeah, no go.
but the important question is: could it be efficient? like, is it physically possible to build efficient and cheap peltier devices?
people said for decades that solar panels can not be cheaper than coal and gas power plants yet here we are today.
That would be nice. I kind of don’t think so. This isn’t new technology, but stranger things have happened.
To my understanding, no. Heat pumps will always be magnitudes more efficient.
is there any proof for this? i.e any physical law that derives an upper limit for the efficiency of peltier devices?
so i believe that semiconductor technology only really took off around 1960, and by then gas-compressing fridges were already widespread so they couldn’t be replaced because it’s always difficult to enter a market that’s already occupied. but i’m thinking, what if there’s a significant improvement in peltier devices? we know a lot more today about semiconductors than we did in the 20th century, it’s worth re-investigating whether they’re feasible.
i will read up on it because i’m fascinated. i always like solid-state technology. i might just think about it a lot simply due to being very interested in the technology, even if it’s not market-mature yet today.
some people have pointed out that they’re less efficient the greater the temperature difference gets. so i wonder, what if you just take a bunch of them and “stack” them, i.e. put them one-after-another. they would be thermally in series but electrically in parallel, so the electric resistance doesn’t go up too much but the thermal resistance does. this is basically what you want. a low electric resistance but a high thermal resistance.
for this to make any economic sense they would have to be pretty cheap per-piece. that is why they should use the simplest possible manufacturing materials, such as doped silicon. we know from solar panels (PV cells) that those that eventually market-matured were using very simple materials (doped silicon) instead of the more rare and difficult-to-get germanium, gallium, etc.
there has been lots of research into more “efficient” (energy efficient) solar panels from german research institute fraunhofer institut, however it turns out that energy efficiency wasn’t what eventually made solar panels take off. the chemistry used today is still the simplest chemistry from around 2000, just scaled up enormously to take advantage of economies of scale. i think that peltier devices would have to be similar: simple doped silicon, no germanium etc., just very simple chemistry, mass manufactured to produce a lot of devices with cheap per-piece price, then just take a bunch of them and stack them behind one another to achieve greater energy efficiency.
deleted by creator
I’ve experimented with these a lot for cooling, and the fundamental problem is that they are most efficient when the temperature drop across them is 0. This means the use case for them is quite limited - basically when you want something small with no moving parts, you can get rid of the heat well (big heatsink on the hot side), and you don’t want it that cold that fast.
ok but what if you stack multiple of them on top of another? then the temperature drop across each one is rather small but since they stack, it adds up. would that be a possibility?
i looked it up. no, stacking them does not magically improve the efficiency. for example, for a ΔT = 1 K you have like 20 mJ of electric energy input to move 1 J of heat across the device (so the coefficient of power is around 50, which is crazy damn high)
if you have ΔT = 2 K then you need 40 mJ of energy input to move 1 J of heat across.
but if you use two stacked devices with ΔT = 1 K each, then you need 2 * 20 mJ of energy input to move the 1 J of heat across both devices to achieve a total ΔT = 2K. so it does not actually save energy.
Peltier devices are what was usually used in car mounted “fridges” years ago. Not sure now. They’re not very efficient, so they’re only used when heat pumps aren’t practical.
I bought a beer fridge a few years ago that was a Peltier device and it came with a wall plug and a car adaptor, they’re still around. It worked but if you left it running for a week or so it would freeze the cans. It was lightweight and did the job and I knew the downsides when I bought it.
That effect is also largely used in temperature sensors
