People have spent a lot of time and effort in trying to control the flow and transfer of heat. Heat is energy transferred in a disorganized way among many little degrees of freedom, like the vibrations of atoms in a solid or the motion of molecules in a gas. One over-simplified way of stating how heat likes to flow: Energy tends to be distributed among as many degrees of freedom as possible. The reason heat flows from hot things to cold things is that tendency. Manipulating the flow of heat then really all comes down to manipulating ways for energy to be distributed.
There are other ways to cool things, though, than by running a cycle using a working fluid like freon. For example, I've written before about magnetic cooling. There, instead of using the motion of liquid and gas molecules as the means to do cooling, heat is made to flow in the desired directions by manipulating the spins of either electrons or nuclei. Basically, you can use a magnetic field to arrange those spins such that it is vastly more likely for thermal energy to come out of the jiggling motion of your material of interest, and instead end up going into rearranging those spins.
|Stretching a polymer tends to heat it, due to the barocaloric|
effect. Adapted from Chauhan et al., doi:10.1557/mre.2015.17
It turns out, you can do something rather similar using rubber. The key is something called the elasto-caloric or barocaloric effect - see here (pdf!) for a really nice review. The effect is shown in the figure, adapted from that paper. An elastomer in its relaxed state is sitting there at some temperature and with some entropy - the entropy has contributions due to the jiggling around of the atoms, as well as the structural arrangement of the polymer chains. There are lots of ways for the chains to be bunched up, so there is quite a bit of entropy associated with that arrangement. Roughly speaking, when the rubber is stretched out quickly (so that there is no time for heat to flow in or out of the rubber) those chains straighten, and the structural piece of the entropy goes down. To make up for that, the kinetic contribution to the entropy goes up, showing up as an elevated temperature. Quickly stretch rubber and it gets warmer. A similar thing happens when rubber is compressed instead of stretched. So, you could imagine running a refrigeration cycle based on this! Stretch a piece of rubber quickly; it gets warmer (\(T \rightarrow T + \Delta T\)). Allow that heat to leave while in the stretched state (\(T + \Delta T \rightarrow T\)). Now release the rubber quickly so no heat can flow. The rubber will get colder now than the initial \(T\); energy will tend to rearrange itself out the kinetic motion of the atoms and into crumpling up the polymer chains. The now-cold rubber can be used to cool something. Repeat the cycle as desired. It's a pretty neat idea. Very recently, this preprint showed up on the arxiv, showing that a common silicone rubber, PDMS, is great for this sort of thing. Imagine making a refrigerator out of the same stuff used for soft contact lenses! These effects tend to have rather limited useful temperature ranges in most elastomers, but it's still funky.