In the past years, the environmental damage caused by the production of rare earth elements (REEs) has received substantial media coverage. The use of REEs in sustainable energy technologies such as wind turbines and electric vehicles has given rise to newspaper articles with titles like “Clean Energy’s Dirty Little Secret.” These articles describe appalling conditions under which rare earths are produced. Indeed, a quick search on the Internet will yield dozens of pictures of huge tracts of lands devastated by toxic wastewater, primitive metallurgical workshops, and Chinese mine workers covered in radioactive mud.

… Numerous studies have pointed to REEs as being critically scarce materials, especially in the context of a transition toward a global low-carbon energy system. Recent publications have also focused on how global trade flows of REEs influence scarcity and the possibility of recovering REEs from the bottom ash of municipal solid waste incinerators.

… Although recycling could help to alleviate scarcity of REEs, it is not immediately apparent that it would also carry a significantly lower environmental burden. REEs are notoriously difficult to process, and depending on the choice of recycling technology, many of the most energy intensive processing steps would have to be performed on recycled material as well. Nevertheless, the environmental damage caused by primary production of REEs has not been a subject of more than cursory scientific investigation. To our knowledge, the environmental impact of REE recycling is not discussed in the scientific literature.

—Sprecher et al.

Magnets are the single largest application of rare earths, taking up 21% of the total rare earth production by volume and generating 37% of the total value of the rare earth market. Although there are two types of rare earth permanent magnets (samarium−cobalt as well as NdFeB), neodymium magnets are more powerful, and hence play the major role in the market.

The team, with members from Leiden University, Tata Steel, the University of Birmingham and Van Gansewinkel Groep, sought to quantify the environmental impact of producing 1 kg of neodymium magnets using virgin material, compared with producing 1 kg of neodymium magnets from recycled material.

For the purpose of the study, they assumed the magnets would be used invoice coil motors, as found in computer hard drives (HDDs). The upstream processes covered the entire production chain of NdFeB magnets, from mining to the production of the magnets, but not the incorporation of these magnets into the final products.

The exact composition of NdFeB magnets varies by application. Elements such as dysprosium and holmium are added when the magnet is required to operate in a high temperature environment. Usually a mixture of neodymium (Nd) and praseodymium (Pr) is used as an alloying agent, instead of pure neodymium. Because Nd and Pr differ by only one atomic number, an extra solvent extraction step is needed to separate them. Therefore, in all but the most high-end applications, neodymium and praseodymium are not separated. However, because this has little influence on the production processes described here, as praseodymium will for all intents and purposes have the same properties as neodymium, we will refer to NdPr alloy as Nd. NdFeB magnets used in HDDs generally do not contain dysprosium, because HDDs are not designed to operate in high-temperature environments. Dysprosium use is not considered in our study.

—Sprecher et al.

The researchers constructed three production scenarios: a baseline scenario that represents the current state of the industry; a high-tech scenario that assumes best available technology; and a low-tech scenario. The main differences between the scenarios are efficiencies of various processes along the production chain and differing emission controls.

• Compared to the baseline, the high-tech scenario requires 22% less energy and 35% less ore. This is reflected in most of the indicators, which are reduced in roughly the same amount. For the freshwater aquatic ecotoxicity indicator, the difference is only 7%. This indicator is dominated by nickel use in the coating of the magnets. The high-tech and baseline scenarios both use the same coating process, explaining the small difference. Human toxicity is reduced by 72%, due to the modeling of more robust emission controls.

• The low-tech scenario requires 32% more energy and 77% more ore per kilogram of NdFeB compared to the baseline. Most indicators also increase in this range. The exception is human toxicity, which increases by 68%.

• In the baseline scenario, 81% of human toxicity is caused by emissions of hydrogen fluoride (HF), with the balance consisting of various smaller emissions of heavy metals. A total of 93% of HF is emitted during acid roasting. The low-tech scenario shows the same structure, albeit with higher absolute numbers.

• In the high-tech scenario, only 52% of human toxicity is due to HF emission, with the balance relating mostly to heavy metal emissions. Of the shredded recycling process, 43% is related to HF emissions during solvent extraction. A total of 36% is related to the emissions of heavy metals related to nickel electroplating, and the remainder to various smaller emissions.

• In the baseline scenario, 48% of total global warming potential GWP is due to electricity use of the foreground processes. A total of 17% is attributed to the burning of diesel in electric generators in the mining process. The remainder is due to energy consumption elsewhere in the system. Similarly, eutrophication is mostly due to energy use, although this indicator is dominated (52%) by the emissions of nitrogen oxides of the diesel electric generating sets used during mining.

General results of their comparative analysis included:

• Compared to the primary production process, recycling via hand picking scores significantly better with respect to most impact categories. This is due mainly to lower energy use. Additionally, human toxicity is significantly lower, because this recycling process does not include the most polluting processing steps associated with virgin production. The same is true for the recycling of magnets via shredding.

• Normalized results indicate that for the primary production process the human toxicity component is by far the most relevant environmental impact. Both recycling processes also count human toxicity and freshwater aquatic ecotoxicity as their main impacts, but much less overwhelmingly so.

• The largest Nd losses is the primary production chain occur during beneficiation, where 50% of the rare earth containing mineral is lost to tailings. Further losses amount to a total of 64% of the total input of neodymium in the production chain for neodymium magnets lost.

We conclude that the value of recycling of neodymium is highly dependent on the method of recycling. Although from an environmental point of view recycling will always be an improvement over primary production, the large losses of material incurred while shredding the material puts serious doubts on the usefulness of this type of recycling as a solution for scarcity. Furthermore, our LCA also shows that technological progress can make a significant difference in the environmental impact of producing neodymium magnets from primary sources.

—Sprecher et al.

Resources

• Benjamin Sprecher, Yanping Xiao, Allan Walton, John Speight, Rex Harris, Rene Kleijn, Geert Visser, and Gert Jan Kramer (2014) “Life Cycle Inventory of the Production of Rare Earths and the Subsequent Production of NdFeB Rare Earth Permanent Magnets,” Environmental Science & Technology doi: 10.1021/es404596q