Magnetic Cooling

A ‘green’ magnetic cooling device built using ‘upcycled’ NdFeB magnets

Dimitri Benke, Jonas Wortmann, Marc Pabst, Tino Gottschall, Iliya Radulov, Konstantin Skokov, Oliver Gutfleisch
Material Science, Functional Materials, TU Darmstadt, Germany

Davide Prosperi, Alex Bevan, Stephen Dove, Gojmir Furlan, Catalina Tudor, Peter Afiuny, Miha Zakotnik
Urban Mining Company, USA


Magnetocaloric devices hold the potential to satisfy the rising demand for cooling in the future. One remaining challenge is to reduce the high ecological footprint of the permanent magnets driving the magnetic cooling cycle. Existing devices use neodymium-iron-boron (NdFeB)-type permanent magnets, which account for more than 50% of the ecological footprint of the appliance. To overcome this hurdle, TU Darmstadt and Urban Mining Company have built the first working magnetocaloric device that uses recycled NdFeB as a magnetic field source. Coupling this with optimization of the magnets and their geometry, it is possible to further reduce the ecological footprint. Together, these two approaches help to position magnetic cooling as a realistic and sustainable cooling technology.


Cooling Demand
Figure 1: Predicted global energy demand for cooling and heating

Rising Demand for Cooling

Human activity is steadily warming the planet. Despite being in the midst of a ‘refrigeration revolution’ there are increasing demands for cooling [1]. Indeed cooling is expected to outpace global demand for heating by 2070 [2] (Figure 1). Cooling could significantly add to climate change unless it is carefully managed. To tackle this problem, better cooling technologies that are more energy efficient than the widely used gas-compression system, must be investigated.

Magnetocaloric and permanent magnet materials are key components for energy conversion technologies that hold a possible answer to more efficient cooling [3, 4]. The Functional Materials (FM) research group, led by Professor Oliver Gutfleisch in Darmstadt, has been investigating both alloys types for several years (

The principle behind magnetic cooling is the so called ‘magnetocaloric effect’ (MCE). MCE is the temperature rise that occurs in a certain group of materials when a magnetic field is applied to them. At this increased temperature, it is possible to expel the heat and, after removal of the magnetic field, the material is subsequently colder than at the start of the cycle (Figure 2). This makes it possible to build a cooling device which operates without a compressor. In addition, devices relying on MCE for cooling do not require coolants that often display a potential to trap solar radiation that is many tens or even hundreds of times greater than an equivalent volume of CO2.

However, magnetocaloric technology comes with its own challenges; the materials that show the greatest heating effect during MCE are based on the rare-earth element gadolinium; this magnetocaloric ‘benchmark’ material shows an isothermal entropy change of ΔSm = 3 J/kg K and an adiabatic temperature change of ΔTad = 3 K in magnetic field change of µ0H =1 Tesla. Gadolinium occurs in low abundance in natural ores, and is mixed together with many other elements. It is therefore predicted to become more expensive if global demand substantially increased.

Cooling Cycle
Figure 2: The 4-step magnetic cooling cycle [4].

In addition, the high-energy field needed for efficient magnet cooling is usually provided by neodymium iron boron (NdFeB) based magnets. NdFeBs contain rare-earth (RE) elements like neodymium and dysprosium. The environmental costs of mining rare-earth have been well documented [5-6]. These costs mean that a conventional approach to magnetic cooling would result in a larger ecological footprint than that of vapor-compression cooling. Indeed, NdFeB magnets account for more than 50% of the total ecological footprint of a magnetic refrigerator [7].To solve this problem, FM is developing new eco-friendly magnetocaloric materials and devices and have partnered with Fraunhofer IWKS and Urban Mining Company, who both specialize in the recycling of NdFeB [7-11].

Figure 3: The Magnet to Magnet recycling system [9]. Waste NdFeB is depicted in red and the proprietary alloy in green.

Upcycling magnetic material

Production of NdFeB magnets from ‘virgin‘ elements derived from mining creates is environmentally damaging; one ton of refined rare-earth creates 75 cubic meters (tons) of acidic waste water, one ton of low-level radioactive waste and releases 10 tons of CO2 [6]. At the same time OEMs are not willing to pay for sustainable mining of REs.

A response to this dilemma is to create technologies to keep reusing the REs already extracted from the ground. Lead by Dr. Miha Zakotnik, Urban Mining Company has obtained funding to build an NdFeB recycling plant and a research center in Austin, Texas (see

Recycling NdFeB magnets is more sustainable because it eliminates the environmentally damaging aspects of production; including ore mining, acid leaching, and solvent extraction. The problem with merely recycling is that the magnetic properties of the NdFeB material gradually decrease with each cycle of reuse; as oxidation and contamination take their toll. This limits the potential of a system that merely retains the NdFeB materials’ original magnetic properties.

UMC‘s Magnet-to-Magnet (M2MTM) upcycling process, discussed herein, showcases how scrap NdFeB-based magnets can be reused and their physical and magnetic properties enhanced. This makes the process powerful; because magnets can now realistically be recycled many times using M2MTM.

Figure 3: The Magnet to Magnet recycling system [9]. Waste NdFeB is depicted in red and the proprietary alloy in green.

Briefly, during M2MTM NdFeB is harvested from MRI machines, hard disk-drives, electric motors, or loudspeakers. The waste NdFeB and a proprietary alloy are then mixed and the two materials milled into very fine powders. The two mixed powders are then pressed into loose block shapes in a mould under enormous pressure in a magnetic field (Figure 3). During this stage, each tiny grain of magnetic crystal orientates itself in the magnetic field so that all the grains are aligned. These loose blocks are then given a series of heat treatments that allows the proprietary alloy to coat and glue the particles of the magnetic alloy together (see Fig 3). This creates fully dense blocks of NdFeB that have:

• The same or better magnetic properties as starting materials; properties can be tailored to customer requirements.

• Better resistance to high temperatures; all NdFeBs lose magnetic performance at higher temperatures. Those made by M2M are better at resisting this decline.

• Better resistance to corrosion; NdFeB type magnets are very susceptible to attack by water and air and ’rust‘ badly. Those made by M2MTM resist this particularly well.

The M2M process is not confined to a laboratory but has been tested in real production facilities at industrial scale.

In addition, independent work by Purdue University has demonstrated that the M2MTM process utilises only 50% of the energy required by traditional manufacture; this drops to around 10% if mining is included. CO2 emissions are therefore reduced by around 10 tons for every ton of recycled magnets produced. Hence recycled magnets have a greatly reduced environmental footprint compared to conventionally manufactured NdFeBs.

Figure 4: MCE demonstrator in a testing rig

Magnetic Cooling Array

To prove the viability of recycled magnets, FM and UMC have developed the first prototype magnetocaloric cooling engine (Figure 4) that uses

recycled permanent magnets (see Figure 5). In this device, the geometry of the magnetic assembly was optimized such that the amount of magnetic material required was reduced whilst also enhancing performance.

Figure 5: A magnetocaloric demonstrator consisting of nested Halbach arrays made from recycled permanent magnets

UMC used the M2MTM process to create customised magnets for a magnetic cooling device. The magnets were shaped into the arrays shown in Figure 4 and coated with an anticorrosion layer.

Thermal Span
Figure 6: Build-up of the temperature span between cold and hot sides in the magnetic cooling device.

Magnetic Refrigeration

The performance of the upcycled Halbach array was investigated in our demonstrator by implementing a powder bed regenerator made from Gadolinium spheres that is put into the active volume of the magnetic array. Using this array the resulting temperature span (Figure 6) can reach 33 °C, exceeding the first demonstrator by 16°C while significantly decreasing the amount of permanent magnets used in the device. Also the volume which is exposed to the magnetic field change is larger and so a longer regenerator can be used, resulting in a further increase in thermal span. These developments together mean that this demonstrator can achieve the amount of cooling that is necessary for an actual product. Recent investigations show that our early stage of development already provides a cooling power of up to 76 W/kg which is in the middle range of power compared to similar machines.

Figure 7: Adiabatic temperature change of Tad of Gd and La(Fe1-xMnxSiy)13Hz in magnetic field change of µ0H = 1 T. The Mn content allows precise tuning of the working temperature and decreases here from left to right.

We went one step further and replaced Gd with abundant and non-toxic La(Fe,Si)13-type compounds which can provide even larger relative cooling power (ΔSm = 10 J/kg K and ΔTad = 3.5 K in µ0H =1 T). Its inherent brittleness we overcame with our patented metal bonding process [12] and the composite magnetocaloric materials can be shaped in heat exchangers with planar channel and/or porous body geometry [13].

In summary, we demonstrated that by combining resource efficient material development for the magnetic refrigerant, upcycling of permanent magnets and materials and device building expertise, TU Darmstadt and Urban Mining built an optimized magnetic cooling device that has a better performance and a lower ecological footprint than former conventional cooling devices.


[1] R. Gauss, G. Homm, O. Gutfleisch, The resource basis of magnetic refrigeration, J. of Industrial Ecology, DOI: 10.1111/jiec.12488. (2016)

[2] Isaac, M. and D. P. van Vuuren. 2009. Modeling global residential sector energy demand for heating and air conditioning in the context of climate change. Energy Policy 37(2): 507–521.

[3] R. Gauss and O. Gutfleisch, Magnetische Materialien – Schlüsselkomponenten für neue Energietechnologien, in Rohstoffwirtschaft und gesellschaftliche Entwicklung, ed. P. Kausch und J. Matschullat, März 2016, Springer Spektrum Heidelberg, Springer-Verlag GmbH, ISBN 978-3-662-48854-6, pp. 99-118.

[4] O. Gutfleisch, J.P. Liu, M. Willard, E. Brück, C. Chen, S.G. Shankar, Magnetic Materials and Devices for the 21st Century: Stronger, Lighter, and More Energy Efficient (review), Advanced Materials 23 (2011) 821–842.

[5] Miha Zakotnik, Catalina Oana Tudor, Laura Talens Peiro, Peter Afiuny, Ralph Skomski, Gareth P. Hatch, Environmental Technology & Innovation Vol 5., pp. 117- 126, 2016.

[6] Keith Bradsher, Mitsubishi quietly cleans up its rare-earth refinery. New York Times, March 9, 2011, Page B4.

[7] R. Gauß, O. Diehl, E. Brouwer, A. Buckow, K. Güth, O. Gutfleisch, Verfahren zum Recycling von seltenerdhaltigen Permanentmagneten, Chem. Ing. Tech. 87 no. 11 (2015) 1477-1485, DOI: 10.1002/cite.201500061.

[8] O. Gutfleisch, K. Güth, T.G. Woodcock, L. Schultz, Recycling Used Nd-Fe-B Sintered Magnets via a Hydrogen-Based Route to Produce Anisotropic, Resin Bonded Magnets, Advanced Energy Materials 3 (2013) 151-155.

[9] M. Zakotnik & C.O. Tudor, Commercial-scale recycling of NdFeB-type magnets with grain boundary modification yields products with ‘designer properties’ that exceed those of starting materials, Waste Management, Volume 44, Pages 48-54, October 2015.

[10] US patent 14/307,267

[11] Hongyue Jin, Peter Afiuny, Timothy McIntyre, Yuehwern Yih, John W. Sutherland, 2016. LCA Analysis for magnet to magnet recycling. Procedia CIRP, 48, 45-50

[12] I. Radulov, K. P. Skokov, T. Gottschall, M. Kuz'min, O. Gutfleisch, Preparation process of a magnetocaloric powder and a composite with magnetocaloric powder. German patent application 10 2015 119 103.2, 6. November 2015.

[13] I. A. Radulov, D.Yu. Karpenkov, K.P. Skokov, A.Yu. Karpenkov, T. Braun, V. Brabänder, T. Gottschall, M. Pabst, B. Stoll, O. Gutfleisch, Acta Materialia 127 (2017) 389.