Low Temperature Physics: The What, the How, the Why

The How: Zeroing in on Zero

Two tools make such low temperatures possible: dilution refrigerators and adiabatic nuclear demagnetization refrigerators. The former gets experiments colder than you can imagine. The later gets them even colder.

Dilution Refrigerator

Dilution fridges (dil fridge for short) owe their cooling power to the incredible element helium and its two isotopes, helium-3 (³He) and helium-4 (⁴He). Though most people are familiar with it as the gas inside their party balloons (⁴He), helium can also condense into a liquid – but only at 4.2 K. This property makes helium a very valuable cryogen in science. Using a condensation/evaporation cycle not unlike that of a kitchen refrigerator, a dil fridge takes 4.2 K liquid helium way down to 1.5 K.

That's no small feat. But low-temperature physicists are aiming even lower. By utilizing the special properties of helium isotopes and exploiting phase transitions, osmosis, evaporation, condensation and (of course) dilution, dil fridges gradually reach the millikelvin range. At the MagLab"s Tallahassee headquarters, dil fridges are used in the Millikelvin Facility's superconducting magnets and in the world-record 45-tesla hybrid magnet.

At the High B/T Facility, a dil fridge is only the first of a two-stage cooling process. The second stage involves an adiabatic nuclear demagnetization refrigerator – NDF for short.

An NDF doesn't use a refrigerant, such as liquid helium. Rather, it cools things down with a magnet and a bundle of metal, taking advantage of the fundamental properties of magnetism. The magnetization (M) of a material is uniquely dependent on the ratio of applied magnetic field (B) to temperature (T). If a magnet can be perfectly isolated, M (or B/T) stays constant. So when you lower B, you necessarily lower T – once, that is, you have created a strong initial magnetization (M). And that is the trick. For example, if you have a screwdriver that has become magnetized and you want to make it non-magnetic again, you can heat the metal and the ferromagnetism will go away. Conversely, if you want to make it a better magnet, you can decrease the temperature while exposing the screwdriver to an external magnetic field to increase the alignment of the magnetic domains within the metal.

In an oversimplified nutshell, NDFs work like this. You start with a piece of copper. The nuclei of copper atoms have a property called spin: They randomly spin about an axis, roughly like the Earth around its axis. This spinning turns the nucleus of each atom, in effect, into a tiny magnet. So each copper nucleus has a north and south pole, just like a magnet. When you put the copper inside a high magnetic field, those nuclei will align with that field, their north and south poles lining up with the poles of the stronger external field.

When you then reduce the applied magnetic field, the nuclei want to fall out of alignment and move randomly around again (i.e., increase entropy). However, M (or B/T) of a perfectly isolated sample must remain constant, so as the applied field is lowered (controlled from the outside by the scientist), the temperature of the copper and of the experiment (connected to the copper by a solid silver cylinder) is also lowered. This is nuclear demagnetization. (The reason for the long, silver extension is that the experimenter may want to apply a high field to the sample itself while holding the nuclear refrigerator cold. The geometry of the High B/T Facility, along with a special, central superconducting coil to shield the experimental region, allow the scientists to do just that.) The perfect isolation needed to carry out nuclear demagnetization is achieved with the use of a superconducting heat switch placed between the dilution refrigerator stage and the NDF.

In this way, the NDF takes the experimental space down from the millikelvin range of the dil fridge to the microkelvin range, cooling the space by a factor of 1,000.

After the NDF has done its job, the system will slowly warm up a little due to vibration, weak radio signals from the neighborhood and residual heat leaks via the very wires needed to make measurements at the sample. Luckily, these systems are carefully designed to absorb this heat while remaining cold enough for the experiment. For the High B/T facility, this cooling capacity is about one milliwatt. To appreciate how little heat that is, think of your typical 100-watt light bulb – then divide by a million. That's a nanowatt. As High B/T Director Neil Sullivan likes to explain it, that's about the amount of energy produced by a Florida mosquito doing one push-up per second.

This high capacity compared to other facilities enables researchers to stay at the extremely low temperatures for several days, even weeks in some cases.

Seems like an awful lot of trouble to go to for cooling some small objects. Not to mention cost: At High B/T, each experiment uses about 1,000 liters of helium just to get started, then about 250 liters per week while experiments are conducted. And helium ain't cheap. But for low-temperature physicists, it's well worth it. Why?

Next Page The Why: Quantum Effects and Freaky Fermions

1 | 2 | 3 | Links | Full Article