Technical field This invention relates to an improved method for processing and/or recycling rare earth metal magnets, specifically neodymium iron boron (NdFeB) magnets.

Background Rare earth magnets based on neodymium iron boron (NdFeB) alloys are employed in many clean energy and high tech applications, including hard disk drives (HDDs), motors in electric vehicles and electric generators in wind turbines. While demand for rare earth element-based materials continues to grow, such materials are at a critical supply risk, in particular neodymium and dysprosium. Recycling end-of-life products containing NdFeB magnets could provide a strategic supply of these elements.

However, NdFeB magnets vary in composition, which can cause issues for downstream recycling processes such as Hydrogenation-Disproportionation- Desorption-Recombination (HDDR), where the reaction kinetics and hydride stability are affected by the starting material composition. For a batch of NdFeB magnets with mixed compositions, this will lead to un-even processing and possible deterioration in magnetic properties. An alternative to these direct recycling routes would be to separate and recover the rare earth elements, which can be performed conventionally by using pyro- and/or hydrometallurgical processes.

Such processes are highly energy intensive, costly, and may suffer from loss of yield of the rare earth elements. There is therefore a need for an improved method of separating rare earth elements from NdFeB magnets. Summary of invention

According to the present invention, there is provided a method of preparing NdFeB powder for separation of the Nd therefrom comprising,

(i) providing a source of NdFeB containing powder,

(ii) subjecting the powder to hydrogen disproportionation comprising exposing the powder to an atmosphere of hydrogen at an elevated temperature,

(iii) maintaining the hydrogen atmosphere and elevated temperature for a hold time sufficient for spherical pools of Nd-hydride to form, and

(iv) cooling the powder under the hydrogen atmosphere with subsequent evacuation of the hydrogen.

The invention finds particular utility in the recycling of scrap sintered magnets from, for example, waste electronics or electrical equipment.

The powder provided in step (i) may be obtained by subjecting a NdFeB magnet to hydrogen decrepitation, as described for example in European Patent application EP 1 1813796.7. Uncoated, sintered NdFeB magnets comprise a matrix of Nd ₂ Fe ₁₄ B grains surrounding a Nd-rich grain boundary. During hydrogenation, the Nd-rich grain boundary phase forms NdH ₂ j, then the matrix phase forms Nd ₂ Fe ₁₄ BH _x . The differential expansion of these two phases due to hydride formation causes inter-granular cracking at the grain boundaries, creating a powder comprising Nd ₂ Fe ₁₄ BH _x matrix phase particles and NdH _{2 7} grain boundary phase particles.

The NdFeB powder is subjected to one of two different types of hydrogen disproportionation: either vacuum hydrogen disproportionation (v-HD), wherein the powder is heated to the elevated temperature under vacuum prior to the introduction of the hydrogen, or conventional hydrogen disproportionation (c-HD), wherein the powder is heated to the elevated temperature after introduction of the hydrogen.

The powder may be heated to the elevated temperature at a rate of at least 1°C/min, 5 °C/min, 10 °C/min , 15 °C/min, 20 °C/min or 30°C/min. The elevated temperature may be at least 900 °C, 950 °C, 975 °C, 1000 °C, 1025°C, 1050°C, 1075°C, 1100 ^o ^C or 1200°C. The hydrogen pressure may be at least 1200 mbar, 1500 mbar, 2000 mbar, 2500 mbar, 4000mBar, 6000mBar, lOOOOmBar, 20000mBar.

The hold time may be a length of time sufficient to increase the diameter of the pools of neodymium hydride to at least 1 μηι, 1.5 μηι, 5 μηι, 10 μηι or 40 μηι.

The method may further comprise an additional step of extracting the neodymium hydride from the cooled powder. The neodymium hydride may be extracted by physical means, such as milling and sieving, or by chemical means, such as roasting, dissolving or leaching.

Brief description of the drawings

Figure 1 is a flow diagram of a method for recycling NdFeB magnets according to an embodiment of the present invention.

Detailed description

Figure 1 shows a flow diagram of a method for recycling NdFeB magnets according to an embodiment of the present invention.

The NdFeB magnet starting material is obtained 100 by breaking apart waste electronics and electrical equipment. The starting material is subjected to hydrogen decrepitation 102 to form a hydrogenated powder comprising Nd _x Fe _y BH _z matrix phase grains and a Nd-rich grain boundary phase.

The decrepitated powder is separated 103 from other components of the electronics scrap and subjected to hydrogen disproportionation 104 (v-HD or c-HD) at an elevated temperature. The temperature and hydrogen atmosphere are maintained 106 for a hold time sufficient for spherical pools of Nd-hydride to form and coalesce. The disproportionated powder is then cooled 108 under hydrogen and then evacuated 1 10.

The Nd-hydride can be extracted 1 12 from the cooled powder, either by physical or chemical means. The starting material used was in the form of uncoated, sintered NdFeB-type magnets of composition Nd25.70Dy1.32Pr3.44FeBaiCo1.30B1.00 (wt%, measured by ICP, minor additions not included). The magnets used for each experiment had a mass of approximately 20 grams.

The magnets were broken into pieces before being loaded into a furnace tube, which was then flushed with argon and evacuated to 10 ^"2 mbar. Hydrogen was introduced at 200 mbar min ^"1 to the processing pressure, normally 1200 mbar, and the pressure was held until the hydrogen decrepitation reaction was finished, e.g. when the temperature was back to the starting temperature, in response to the exothermic reaction, and the hydrogen flow rate was zero. The disproportionation reaction was then carried out.

For vacuum disproportionation (v-HD) processing, the hydrogen was evacuated and the sample was then heated to the experiment temperature at 15°C min ^"1 . Hydrogen was then re-introduced at 16 mbar min ^"1 , and the conditions were held for varying hold times (see Table 1), before cooling down in hydrogen and finally evacuating.

For conventional disproportionation (c-HD) processing, the sample was heated in a hydrogen atmosphere at 15 °C min ^"1 up to the experiment temperature, which was varied in this study, and the hydrogen pressure was held constant throughout. The conditions were then held for varying hold times (see Table 1), before the sample was cooled in hydrogen and then evacuated.

The microstructure of the samples were analysed using a JEOL 6060 Electron Microscope, and JEOL 7000 Electron Microscope for high resolution images.

Results

The microstructure of the NdFeB magnet starting material was analysed using a scanning electron microscope (SEM). SEM analysis showed that the starting material consists of Nd ₂ Fei ₄ B matrix grains surrounded by a Nd-rich grain boundary and triple junctions.

Samples of the starting material were subjected to hydrogen disproportionation under varying conditions, as detailed in Table 1. Table 1 - List of experiments

Vacuum disproportionation (v-HD)

Samples 1-3 were subjected to vacuum disproportionation at 900 °C and 1200 mbar H ₂ for a hold time of 1 , 3, and 5 hours, respectively.

After 1 hour of hold time (sample 1), there was a mixture of microstructural features. Fine lamellae of NdH ₂ were formed in some areas, while oxidised triple junctions could still be observed from the starting material. In many places the fine lamellar structure was Over-processed' into spherical pools of NdH ₂ , which were measured to be on average 0.05 micron in diameter by SEM image analysis. Such over-processing appeared to have occurred at grain boundaries in the sintered magnet. After 3 hours (sample 2), the disproportionated structure was shown to coarsen further, with fewer areas of the fine NdH ₂ lamellae being observed. The grain boundary regions from the original sintered magnet were also becoming harder to distinguish.

After 5 hours (sample 3), almost all lamellae disappeared, and most of the NdH ₂ was now spherical in shape. With increasing time, the structure had become increasingly uniform in terms of the shape and size of the NdH ₂ . Conventional disproportionation (c-HD)

Samples 4-6 were subjected to conventional disproportionation. The only condition that was changed between the samples was the temperature, which was set to 900, 950, and 995 °C, respectively. The hydrogen pressure was held constant at 1200 mbar (except for sample 6 where the pressure was 2000 mbar to be sure the material would not recombine) and the hold time was 3 hours.

Sample 4 was subjected to c-HD at the same temperature, pressure, and hold time as the vacuum disproportionation of sample 2. SEM analysis of sample 4 showed that all matrix grains seemed to have reacted, and almost all NdH ₂ had formed into spherical pools instead of lamellae. The microstructure was much more uniform in terms of the size and the shape of the NdH ₂ pools than sample 2, with a relatively narrow size distribution. Samples 5 and 6 were subjected to c-HD at temperatures of 950 and 995 °C respectively. Surprisingly, the NdH ₂ regions started to coalesce and grow much larger at temperatures above 950 °C. The NdH ₂ in sample 6 was on average around 0.74 micron in size. The microstructure appeared homogeneous, although in sample 5 some oxidized Nd-rich triple junctions were still visible in the material.

Samples 7-9 were subjected to c-HD at a temperature of 977 °C and hydrogen pressure of 1200 mbar. The only condition changed between the samples was the hold time, at 6, 15, and 48 hours, respectively. SEM analysis of the disproportionated samples showed that the NdH ₂ pools continued to grow with increasing hold time, and in some cases seemed to have coalesced to a very large extent. The Nd-rich grain boundary phase from the starting sintered magnet was no longer visible, and the longer the hold time, the less difference observed between the NdH ₂ pools formed by the disproportionation reaction. The microstructure continued to grow all the way up to 48 hours, although the difference was larger between the sample processed for 6 hours (sample 7) and 15 hours (sample 8) compared to between 15 and 48 hours (sample 9). The average diameter of the NdH ₂ pools in sample 9 was measured with image analysis and found to vary between 0.5 and 7 microns, with an average of 1.61 microns. When the disproportionated powder from sample 9 was ball milled, then separate NdH ₂ and Nd ₂ Fe ₁₄ BH _x matrix particles could be clearly observed by SEM analysis. By passing this powder through an inert gas cyclone such as a jet mill it should therefore be possible to separate the NdH ₂ from the matrix particles based on the size and density of the two phases.

By extending the time, temp and pressure compared to the current disproportionation reactions then it would also be possible to increase the size of the NdH ₂ . If the NdH ₂ reaches around 40 microns then it should be possible to use low energy milling (eg- ball milling) and then conventional sieving to remove the NdH ₂ .

After the NdH ₂ is removed this can then be degassed by heating above 600°C in vacuum to remove the hydrogen. The Nd can then be cast into new NdFeB alloys by induction melting with Fe and B to form new alloys. Those alloys would then be put back into the primary supply chain for NdFeB magnets.

Alternatively the NdH ₂ could be used as a blending agent to mix with hydrogenated recycled NdFeB alloy powders separated from electronic scrap. This extra NdH ₂ could then provide a liquid phase sintering agent.