Recycling rare earths

The two types of rare earth magnets: neodymium−iron−boron and samarium−cobalt magnets, together use around a fifth of the REEs produced and account for over a third of the REE market, making them ideal targets for recycling. The more powerful neodymium magnets make up the bulk of the market and can be found in computer hard drives, electric and hybrid cars, electric bicycles, and wind turbines.

A study in 2014, by researchers from the Netherlands and the UK (Environmental Science & Technology, doi: 10.1021/es404596q), compared the environmental impact of producing 1kg of the neodymium magnets from recycled magnets taken from computer hard drives, versus mining virgin material. Recycling used 88% less energy than mining and had a ‘human toxicity’ score that was 98% lower. 

The recycling process was developed by one of the authors of the impact study, Allan Walton and his group in the school of metallurgy and materials at the University of Birmingham, UK. 

In the process, the magnets are placed in a container with hydrogen gas. The neodymium absorbs the hydrogen, causing the magnets to expand and disintegrate into a powder, which is sieved to separate the neodymium−iron−boron powder from the magnet coating, usually nickel, and then milled to break up any larger particles. 

The researchers claim that when the technique is used on magnets taken from computer hard drives, if the powder is re-sintered to produce new magnets, ‘it is possible to recover more than 90% of the magnetic properties of the starting material’. In addition to its low energy use, the new process generates little waste because the hydrogen can be recycled.

The main disadvantages are that the composition of the final magnets is determined by the waste magnets used, and the fact that the rare earths aren’t separated from the other alloys.Traditionally, hydrometallurgy has been used to separate the rare earths from their alloys. This is a multi-step process that uses strong mineral acids to dissolve the alloys; the REEs are then recovered using selective precipitation. However, since large amounts of acids are used, the process isn’t particularly environmentally-friendly. 

Recently, a team led by Koen Binnemans, at the University of Leuven in Belgium, has been working to make the process ‘greener’ by using reusable ionic liquids in place of the mineral acids to dissolve the magnets (Green Chemistry, doi: 10.1039/c5gc00155b 2015, 17, 2150). Binnemans’ group use an ionic liquid they have synthesised – betainium bis(trifluoromethylsulfonyl)imide, [Hbet][Tf2N] – to dissolve and then separate the metal oxides from the roasted (microwaved at 80°C) and crushed neodymium−iron−boron magnets. 

The technique takes advantage of the thermomorphic properties of the ionic liquid when mixed with water – ie at 80°C the ionic liquid and water mixture is homogeneous, but as it cools to room temperature it separates and the metals selectively distribute between the two phases. The iron oxide moves into the ionic liquid phase and the REE ions into the water phase, where they can be precipitated using oxalic acid. This technique produces rare earth oxalates that are 99.9wt% pure from a high-performance cobalt- and dysprosium-rich magnet and a standard hard disk drive magnet.

Binnemans’ group has also exploited the thermomorphic properties of [Hbet][Tf₂N] to recover red phosphor, which contains the rare earths yttrium and europium, from the phosphor in fluorescent lamps and low-energy light bulbs (Green chemistry, doi: 10.1039 both/c4gc02107j). Once separated from the rest of the waste, the yttrium and europium can be recovered from the red phosphor with oxalic acid or hydrochloric acid.

The yttrium/europium oxalate can then be calcined, ie heated to high temperatures in air or oxygen, to reform red phosphor (Y₂O₃:Eu³⁺) with a purity of more than 99.9wt%, ‘effectively closing the loop after only three process steps’, the researchers report.

Meanwhile, Marion Emmert, assistant professor of chemistry and biochemistry at Worcester Polytechnic Institute in the US, has developed a two-step hydrometallurgy process for separating rare earths from the drive units of electric vehicles (Green Chemistry, doi; 10.1039/C5GC01255D). ‘Our process is, to the best of my knowledge, the only process that can actually deal with real end-of-life materials,’ Emmert says. ‘We deal with material mixtures, which is very different to a lot of the prior processes that have just looked at processing waste magnets.’

The researchers heat the drive units to 450°C for one hour to demagnetise them and then shred them before soaking the resulting scrap in hydrochloric acid for 24 hours. The magnetic material dissolves, leaving behind the steel and copper from the rest of the drive unit, which can be put into a regular recycling stream. The rare earths are then precipitated out of the solution using oxalic acid. The researchers comment they can recover 82% of neodymium, dysprosium and praseodymium present in the original motor material, with a purity of 99.8wt%.

However, all these techniques cannot separate the rare earths from each other. ‘This requirement is a big liability for recycling,’ says chemist Eric Schelter at the University of Pennsylvania, US, ‘because essentially all of that capability for separations exists in China.’ 

Binnemans agrees that there is still a lot of work to be done on the separation of rare earths. ‘Because they are so similar chemically, it is very difficult to separate them. The liquid–liquid techniques that are currently used consume a lot of chemicals and have to be repeated over and over again, and then you get an accumulation of small effects and at the end, after a hundred steps, you have two rare earths with very high purity,’ he explains.

Schelter is developing a novel technique that uses a ligand to separate neodymium from dysprosium (Angewandte Chemie, doi: 10.1002/anie.201501659). The specially designed tripodal nitroxide ligand is added to a powdered mix of the two elements. Size differences between the neodymium and dysprosium ions mean that ligand–neodymium complexes bind to one another, while the ligand–dysprosium complexes do not. This increases the solubility of the ligand–neodymium complexes, compared with their dysprosium counterparts, making it easy to dissolve them in solvents and filter them. 

The process takes about five minutes and creates separate samples of the rare earths that are 95wt% pure. The ligands are stripped from the rare earths using an acid bath and can be used again.

‘Our initial findings were promising because our method is simpler, cheaper and faster than liquid–liquid separation,’ says Schelter. ‘We are currently working on a method to separate europium and yttrium from recycled phosphor,’ he adds

Binnemans, too, is on the case. His team recently demonstrated it is possible to separate europium and yttrium by UV light (Green Chemistry, doi: 10.1039/C4GC02140A). The method exploits the fact that while both elements have three positive charges per ion, europium is also stable with two positive charges. Shining the correct wavelength of UV light on a solution of the two elements reduces europium to its bivalent state, which can then be precipitated out of the solution using sulfate, which doesn’t react with the yttrium. Using this method, the team was able to recover 95% of the europium.

At the Ames Laboratory in the US, Ryan Ott has been working on a two-stage liquid metal extraction process that separates neodymium from magnets as well as the heavier rare earths, such as terbium and dysprosium. The work builds on a previous magnesium extraction technique developed at Ames. 

In the original process, demagnetised ground scraps from magnets are melted with magnesium. Neodymium is soluble in liquid magnesium, while iron is insoluble, so neodymium binds to the magnesium and leaves the iron scrap and other materials behind. The neodymium can then be recovered from the magnesium by vacuum distillation.

‘The magnesium process is very good at recovering the light rare earths, but it is less effective at recovering the heavy rare earths,’ Ott says. ‘You have to go through long processing times and high temperatures to get out a lot of the heavy rare earths.’ The researchers have found, however, that liquid bismuth will bind to the heavier rare earths and can be used to recover them from the material left behind after the magnesium extraction.

Although it is possible to recycle rare earths, the low prices set by China are hindering the development of commercial processes elsewhere. When the Worcester Polytechnic Institute’s business school analysed Emmert’s method of recovering rare earths from the drive units of electric vehicles, it found that the end products would be about twice the price of mined rare earths from China.

Binnemans says that his team is in a similar situation with the lamp phosphorus processes they have developed. ‘If the prices of rare earths doubled or similar, then such processes could become economically viable,’ he says. ‘It is very difficult at this moment to compete with China. A few years ago, [when the price of rare earths was higher] there was a lot of money available from companies for investment, but the technologies were not ready. Now the technologies are available at the lab scale, but the companies no longer want to invest.’

Binnemans concludes: ‘You can be sure that the prices of the rare earths will increase again. The only question is how long will it take? It has nothing to do with science; it is geopolitics and economics, and it is difficult for a scientist to make a forecast.’