Physics Van 3-site Navigational Menu

Physics Van Navigational Menu

Q & A: cooling with magnetism

Learn more physics!

Most recent answer: 06/05/2015
I see all these questions about how heating up a magnet breaks the magnet because the molecules aren't aligned anymore. Does the opposite work? Does placing an item (ferromagnetic, paramagnetic, or even diamagnetic) in a strong magnetic field have a cooling affect? Does the total thermal energy decrease as the molecules line up due to a strong external magnetic field? Or do they maintain their kinetic energy but just wiggle in place somehow?
- Peter Brehm (age 25)
Boulder, Colorado, US

First, it's not that the molecules align when things are magnetized. Many magnetic materials (e.g. iron) don't even have molecules. It's that the magnetic spins align.

Now for the temperature effects. What happens if you put a paramagnet (for simplicity) in a magnetic field? At first the spins are aligned randomly. So they have zero net interaction energy with the field, just as many pluses as minuses. As they align, the net interaction energy goes negative. So the extra energy gets dumped out into the material, heating it up. That's sort of like what you were thinking, but opposite.

Here's a more useful version of the same effect. Say that you have a paramagnet with the spins lined up in a magnetic field. The net interaction energy is negative because the spins tend to be lined up. Now you turn off the field gradually. The spins tend to come out of alignment, but still with some field present. So they tend to gain magnetic energy. That energy comes from the other modes of the material, cooling it down. The process is called adiabatic demagnetization, and it plays an important role in cooling things to very low temperatures. You can read more about these effects here:

Mike W.

(published on 06/02/2015)

Follow-Up #1: cooling techniques

Cool!! Correct me if im wrong but Adiabatic Demagnetization seems to be used only in extreme cryogenic temperatures with paramagnetic salts to get the last little bit even colder, right? And the annoying caveat is that it needs to get hot first, by applying a magnetic field... and only later removing the magnetic field makes it even colder. right?This topic led me to the slightly different effect, the Peltier effect. Does this thermo-electric cooling technique have a theoretical lower temperature limit? Can it be used to reach cryogenic temperatures? Many products advertise a min temp of -50C or so, but I dont understand why. If I submerge a Peltier junction in liquid nitrogen will it sill create a temperature differential centered at -195C? Or conversely is it possible to stack up many of these and create a huge temperature differential, with -190C on one side and 30C on the other (assuming proper heat sinking on the hot side)?Thanks! This stuff is super interesting.
- PETER BREHM (age 25)

The heating after the field is applied doesn't have to be a big deal since it would ordinarily be done when the sample is relatively warm, when it doesn't matter. The cooling is then used after the sample is already cold, to get it colder.

The Peltier effect, the heat transfer from one side of a device to the other when a current is applied, is accompanied by some heat generation, just by ordinary Joule heating. At low temperature, that heating beats the cooling effect so both sides warm up. So you can stack Peltier cells but they aren't useful for reaching very low temperatures.

Mike W.


(published on 06/05/2015)

Follow-up on this answer.