Magnetic refrigeration gets rid of gas

Tom Shelley reports on an alternative refrigeration technology that uses no gas and could retain its efficiency even when used on a very small scale

Solid state refrigeration – which has no need for gaseous refrigerants such as HCFCs – has come a step closer with the development of materials that change their temperature when magnetised.
These binary metal refrigerants – used in conjunction with novel nano-particle suspension heat transfer fluids – could work much more efficiently than Peltier effect devices.
The technique could be used on a large or small scale. It has the potential to replace conventional gas compression and cycle refrigeration plants – saving energy and avoiding problems associated with leaking refrigerant gases – but could also cool electronic chips and other small scale devices.
The technologies being developed by Camfridge, based next to the Cambridge Science Park, are based on ideas that have been around for some time.
“There has been a lot of work at Nasa Ames on gadolinium silicon germanium, which makes a large temperature change when it is magnetised – but also shows a large hysteresis effect,” explains Camfridge director Neil Wilson.
A magnetic field is applied to a material so that it heats up (or in some cases, cools down). After this the heat is passed to the working fluid, and the field is turned off – making the material cool as it takes in energy in order to return to its previous state. It can then absorb heat from whatever it is that needs to be cooled. As it does so, its temperature rises until it reaches that where the cycle may be repeated.
Two experimental machines have been built. The original, demonstrated to Eureka, uses a bed of gadolinium particles in a tube. Any practical machine is unlikely to be based on gadolinium – it is too expensive – but it is well understood and easy to work with, so other working parameters can be deduced from experiments and used to design more practicable machines.
Heat transfer fluid is pumped through the bed that is reciprocated by an actuator through a magnetic field produced by a permanent magnet. This machine cools fluid from 23 or 24 [degrees]C to 10 or 11 [degrees]C on a 4-second cycle time. It has a cooling power of 10W. The permanent magnet is made of neodymium iron boron, and is a Halbach magnet which enhances the 1.4 Tesla maximum remnant field associated with this material to 2 Tesla.

The mark II design machine has a disk of magnetic material which rotates its sectors between two pole piece pairs in a larger permanent magnet arrangement with a much shorter cycle time. The intention is for the next generation machine to be about the same size as the first, but running at 10 or more cycles per second – and with a cooling power of around 100W, using lower cost working materials.

“People naively think that what is important in a magnetocaloric material is the height of the transition [the temperature change when it is magnetised and demagnetised],” says Wilson. “But a material that has a high transition over a narrow temperature range is completely useless as a refrigerant. What we need is an expanded entropy [energy divided by temperature] range over which the material will work.”
In the original machine, the individual gadolinium particles cycle over a temperature range of about 2[degrees]C, but because the working fluid passes through a bed of particles with a hot and a cold end, the overall cooling is 15[degrees]C.
Many of the materials previously studied for this type of application undergo a change in crystalline structure when magnetised and demagnetised, but the Camfridge materials change only their electronic ground state (the way electron spins line up when in their lowest energy state). In the case of gadolinium, magnetising the material turns it from ferromagnetic (spins lined up) to paramagnetic (spins in all directions), though other transitions are possible. Materials that change their crystalline structures take too long to do so, consume too much energy to make the changes, and – if cycled a large number of times – are liable to fall apart.

The material finally chosen must be low cost, non-toxic and have high thermal conductivity. Other features include low heat capacity and preferably a high electrical resistance, to reduce eddy currents.
“Some of the materials we are looking at are still in the laboratory,” says Wilson. “Gadolinium alloys are much too expensive. It has to have a large change in entropy change between high and low field states. We may want to go to binary materials, where when you alter the properties of one material, you much more dramatically alter the properties of the other.”
The team is also working on cobalt manganese silicon plus various dopants, though Wilson says this causes problems because it can be hard to work with.
“We may also work with thick films, which would allow us to work with exotic structures,” he says. “In such circumstances, you can then look at electrocaloric materials that change their temperature in an electric field in capacitative structures. Films might be 0.1 to 100 microns thick. We are also looking at nano particles in fluids for thermal management.”
Although the ideas have a long history – and Wilson spent time researching in the Cavendish Laboratory – the present enterprise only started in February 2005 with a Carbon Trust grant of 200,000. It has also received funds from the University Challenge Fund, Nesta, the Oxford Trust and “several early stage investors”.
According to Wilson: “We spent the first six months developing computer models, then we built a system. Now are in the process of updating the models with a view to building our next machine. Our time to selling a refrigerator system to the commercial market is still a number of years away”. Email Neil Wilson


* Instead of gaseous refrigerant, the system uses alloys that are passed in and out of a magnetic field – and a new heat transfer fluid based on nanoparticles

* Efficiency is potentially higher than that of gas liquefaction and evaporation cycle refrigerators,

* It can also be applied on a very small scale, where it is likely to be much more efficient than Peltier coolers

* New alloys and alloy combinations have been developed that are very much less expensive than the original gadolinium-based materials

For more technical developments see

Back to list of stories