The Cryofuel Systems Group (CFS) was formed as a part of the Institute for Integrated Energy Systems (IESVic) to design and build a cost effective natural gas (NG) refuelling system. A key component of the refuelling system is the liquefier. CFS is developing an active magnetic regenerative refrigerator (AMRR) as a potentially economical means of efficiently liquefying the NG. A magnetic regenerator is the heart of the AMRR device. It provides the thermodynamic work and the passive regeneration to allow the AMRR to operate over large temperature spans . This thesis deals with the development of the magnetic refrigerants that will comprise the regenerator of a prototype that will span from 110 K to 240 K with a cooling power of 700 W. The work has been broken into four parts:​ selection of the magnetic refrigerants, preparation and characterization of selected materials, manufacture of high performance active magnetic regenerators from the magnetic refrigerants, and testing of fabricated regenerators.

An AMRR is a complicated thermodynamic system that is composed of two interlinked subsystems, the magnetic refrigerant and the circulating heat transfer fluid . To properly select magnetic refrigerants for the AMRR, two fundamental criteria must be satisfied:​ (1) the material must provide the required temperature difference to the heat transfer fluid at the hot and cold boundaries of the device to allow the specified cooling load (with the net work and any entropy created) to be pumped and deposited to the environment, and (2) the magnetic material must do enough net work to satisfy a first law thermodynamic energy balance.

When selecting magnetic materials for the AMRR, the fundamental screening characteristic is magnetic entropy change per unit volume of the refrigerant as a function of applied field and temperature. For the temperature range of interest in the prototype, heavy rare earth elements and intra-rare earth alloys are excellent choices because they not only have large magnetic entropy changes, but they are also have good formability characteristics. Several refrigerants must be layered in the regenerator to meet the above selection criteria and to allow the transient start-up of the device. The materials chosen for the prototype AMRR were elemental dysprosium and three alloys of gadolinium and dysprosium.

The selected AMRR materials must be manufactured into highly effective regenerator geometries. Wire screen and packed particle bed geometries offer the best compromise at this time between performance and manufacturing ease. To make an effective packed particle bed, 100±20µm particles were manufactured using a rotating plasma arc atomization method. This technique produced high quality dysprosium particles with an effective yield of 29%. Special tests were conducted that showed the bulk resistivity of the l00µm particles is at least 13 orders of magnitude greater than pure dysprosium. This makes entropy production due to predicted eddy currents negligible.

As is the case with particles, highly effective wire screen geometries require very fine media and hence, the availability of small diameter rare earth wire. Dysprosium was successfully drawn to a diameter of 66µm but began to break at this point because the original as-cast grain structure was not treated. Indications are that with an initial hot swage, the wire can be drawn to ~25µm. To our knowledge, this is the first time such wire has been made, and it gives the capability to weave up to 400 mesh screens. Because of the high set-up cost of fine screen weaving with new and unfamiliar material, packed particle beds were chosen for the prototype. Fine mesh wire screens will theoretically outperform packed particles when considering the AMRR prototype operating conditions. If the experimental performance is as excellent, the cost of future development will be justified. Randomly stacked matts of these wires may be a practical alternative.