A global overview of the use of REEs was provided by Binnemans et al. in 2013 and by Royen and Fortkamp in 2016 [3
]. The amounts of different REEs produced globally are uneven. Figure 4
shows the distribution of different REEs, and it is evident that the most expensive REEs such as Sc and Tb are produced in very fewer proportions compared to the production of light REEs (LREEs), such as La and Ce [13
]. The unbalance is not only attributable to the geological distribution of the proportionally higher amount of LREEs versus heavy REEs (HREEs), but it is also governed by the fact that the minor quantities of HREEs are more difficult to separate than the LREEs using conventional separation technologies.
It can be observed that La, Ce, Pr, Y, Dy, and Nd are rare earth elements that are most used in the specific energy-relevant applications described. It indicates that they are useful in many applications and that they will probably be more difficult to be substituted with other materials. It is also important to notice that La and Ce are the most commonly produced and used REEs. To extract one equivalent of Nd, two equivalents of Ce, and one equivalent of La are also inevitably removed from the mine as well [8
]. This is due to the inherent abundance of the La and Ce as compared to Nd in the ores. In addition, they are lighter lanthanides and are difficult to separate from each other. Therefore, there is a surplus supply of La and Ce as compared to Nd that keeps their pricing lower. In addition, more processing and ore removal have an adverse impact on the environment in terms of the use of chemicals and the production of secondary wastes and tailings. It is negatively affecting the overall market, supply dynamics, and the environment. The applications relevant to the energy sector is discussed in more detail in the following sections.
3.1. Nickel Metal Hydride (NiMH) Batteries
NiMH batteries are rechargeable batteries of importance mainly in electric and hybrid cars, electrical aircraft systems, and satellite pinpointing systems. These batteries have also a high energy density and have a wide operating temperature range (−30 to 70 °C) [14
REE materials in the batteries provide adequate hydrogen storage and rapid desorption/re-absorption of this hydrogen for a quick recharge and high power. The batteries have a long life with a safe operation. Another characteristic of these batteries is their low maintenance requirement. However, in comparison to conventional Li batteries they can only store 66% energy capacity [14
Hybrid and electric cars account for 57% of NiMH batteries in the world. As an example, every Toyota Prius car carries about 2.5 kg of REEs in the form of Misch-metal (Mm). Mm is a complex alloy of light REEs (La, Ce, Pr, Sm, and Nd). Therefore, electric and hybrid cars are a highly interesting market to study for the development of the recycling of rare earth metals even though it is still in an establishing phase. In the past, used NiMH batteries were exploited as cheap nickel sources without valorization of the 8 to 10% REEs (La, Ce, Pr, Sm, and Nd) [3
Nevertheless, according to Binnemans et al., several active research groups in the field have now developed several chemical separation processes on the NiMH batteries that achieve a recovery rate of the REEs of up to 97.8% [3
]. A first possibility (developed by Zhang et al. in 1998) is to use hydrometallurgical methods to recover nickel, cobalt, and REEs [15
]. Another method (by Li et al. (2009)) consists of hydrometallurgical separation procedures including leaching by solvent extraction, evaporation of the strip liquor and the raffinate, and crystallization [16
]. In pyrometallurgy, the metals are heated in a high-temperature furnace, and redox conditions are used to adjust the vapor pressure of different metals to separate them from each other. The pyrometallurgical operations can be automated easily; however, lack versatility in feed processing and require high investments in comparison to hydrometallurgy [1
Guyonnet et al. [8
] gave an overview of REE “flows into use” (e.g., Tb in lamps and Dy in magnets) and “in-use stocks” (average lifespans of products considered for the estimation of in-use stock) in Europe for the year 2010, showing a flow into the use of 120 (metric) tons for neodymium and 50 (metric) tons for praseodymium (Table 2
). According to Guyonnet et al., the biggest potential for recycling from NiMH batteries is the recycling of Nd and Pr. This is probably because La and Ce are produced in excess and their prices are too low to be included in feasible recycling processes. Guyonnet et al. [8
] also cited typical REE content data as a percentage concerning the entire NiMH battery weight. However, the weight percentage values for Pr (0.4%) and Nd (0.8%) did not agree well with the flow-in-use values (50 ton Pr and 120 ton Nd in Table 2
). With Pr, the rare earth content (wt%) and flow into use-values reported by Guyonnet et al., either, the wt% value should be lower for the given flow into use value (viz., 0.33 instead of 0.4 wt%), or, the flow into use-value should be higher (viz., 60 metal ton instead of 50) for the given wt%-value. Also possible is that the reported values for Pr are inaccurate or were brought down to too little digits. Therefore, Table 2
shows rare earth usage (%) values with NiMH battery alloys from Curtis [17
]. For many different applications, these and other values were originally presented by Curtis, and they are also shown and referred to by Binnemans et al. [3
] and Royen et al. [13
]. The flow in use amounts for La, Ce, and Sm were calculated here using the values for Nd (120-ton metal and 10% content) as a starting (reference) point.
3.2. Permanent Magnets
Permanent magnets are used in automotive electric and hybrid motors, direct-drive wind turbine generators, and speakers. Magnets in automotive motors and wind power generator applications are directly linked to the energy sector and are thus of importance and interest for the present review. These permanent magnets usually consist of a neodymium-rich NdFeB alloy also containing Pr, Gd, Dy, and Tb in smaller amounts. The Nd fraction is the most important one, and with the literature consulted here takes about 65% to nearly 70% of the total REE share in permanent magnets.
shows “flow in use” and “rare earths usage” (amount of REE used in defined application) values reported and estimated in this study for Nd, Pr, Dy, Gd, and Tb with magnets for the year 2010. The reported flow in use and the REE usage values are interrelated over a common REE content of the application under consideration. In Table 3
, the flow into the use of the ton metal values for Europe (the year 2010) was taken from Guyonnet et al. [8
], and they are written in bold script for clarity. The reported REE content values are from Curtis [17
] and are also written in bold script. All other numbers are in italic bold script and estimated here as follows: Nd was most prominently present with magnets, and by considering 70% content associated with the 1230 ton metal into use flow (Table 3
) as reference datasets, we were able to estimate the content for Pr and Dy from their flow into use values in proportion to the flow into use and the content value of Nd. Similarly, the flow into use-values estimated here for Gd and Tb followed from their reported content values relative to the Nd dataset. It must be mentioned that Curtis [17
] also gave usage values for Pr and Dy of 23.4% and 5%, respectively. However, the latter values did not correspond well with the estimated 18% and 13% content values shown in Table 3
For the year 2020, Guyonnet et al. [8
] estimated a recycling potential of 170 to 230 tons for neodymium—about 16% of the flow into use for Nd with permanent magnets in Europe in the year 2010 (Table 3
). As a comparison, the study from Rademaker et al. (2013) presents potential recycling supply ratios (PRSR) for neodymium and dysprosium, as given in Table 4
]. The PRSR stands for the ratio between the REE in collected end-of-life flows and the total REE demand. The low PRSR value for Nd for the year 2020 in Table 4
suggests that the 16% recycling of Nd in 2020 forecasted by Guyonnet et al. [8
] could probably be too optimistic. In addition, with the results in Table 4
, it is also surprising to observe that the Nd recycling ratios are expected to decrease.
The recycling situation of magnets today is very weak. This is due to the following factors [14
Difficulty to decompose the chemical compounds (technological difficulties);
Low concentrations of REEs in the goods and location deep within consumer products;
Lack of financial incentives;
Inefficient collection of end-of-use objects.
This is true for all the different technologies used in the energy sector containing rare earth elements. The REEs are lost in shredded and melted fractions of e-waste not being focused by the existing separation processes [1
]. On the other hand, Mueller et al. conducted an interesting analysis of the geological distribution data, the existing collection rates, the economic feasibility of recycling, and its environmental impact on Switzerland [19
]. For permanent magnets, the situation seems to be quite promising since the collection rate is claimed to be close to 100%. This does not include the motors sent to developing countries where the materials are also reused as second-hand machinery or components. It is assumed that in 2017, 1.6 tons of magnets were available in 1 million electric cars. This number was assumed by Mueller et al. [19
], and it did not represent a high confidence level. From these 1.6 tons, 2 wt% of Nd can be recovered, giving a total of 32 kg of neodymium recoverable in Switzerland.
The dismantling and processing of permanent magnets are expected to become feasible in the medium- or long-term future, and its impact on the environment has been observed to be rather small because the global warming potential of such recycling of Nd has been observed to be 14 kg CO2
equivalent per Nd2
according to kilogram mass allocation and without radioactive side products. This will make the motors manufactured using recycled REEs more sustainable in comparison to those dependent on mining resources [19
]. According to this information, a fairly certain assumption can be made that there is a real potential for permanent magnet collection and the recycling of the contained REEs.
Different separation and recycling methods are proposed for REE magnets [3
], ranging from (1) reuse in the original form up to (2) hydro- or (3) pyrometallurgical processing and processing using corrosive gases, such as (4) corrosive chlorine or (5) reducing hydrogen.
The first option (reuse) is of course the most economical way of recycling and does not generate any waste, but this only works for large and easily accessible magnets that can be found in large wind turbine generators and large electric or hybrid vehicle motors. This is not available in substantial quantities; therefore, other methods have been developed. The first option requires less energy than the hydro- or pyrometallurgical methods, and it does not produce any secondary waste either. This is particularly suited for hard disk drives, but it does not apply to mixed scrap feed or oxidized magnets. The second and third methods are equivalent to the recycling methods for NiMH batteries that were addressed in the last preceding section. Hydrometallurgical methods require many steps before obtaining a new recycled magnet and consuming many different chemicals. Pyrometallurgical methods require large energy inputs to bring enough heat to the material [1
]. The fourth method (gas-phase extraction) consumes large amounts of chlorine gas, but it does not generate any wastewater and applies to every type of magnet composition; thus, it is easier to implement at a larger scale. However, the impact of gases has to be studied carefully, and the technology should not contradict the net zero-emission goals [3
]. New approaches are being developed for direct reduction of the magnets in a hydrogen atmosphere (fifth method) under controlled conditions and sintering of the powdered metals in new magnet shapes again. This process needs a controlled and carefully designed safety concept due to the use of highly reductive hydrogen gas [21
3.3. Lamp Phosphors
Phosphors are essential components of different fluorescent lamps (also known as low-energy lamps) and smartphone screens among other examples. Those phosphors include REEs, mainly Eu, Tb, Y, La, Ce, and Gd in smaller quantities. Television and LED phosphors are very attractive for REE recycling [1
]. Lamps are composed of phosphors that can contain REEs (La, Ce, Eu, Gd, Tb, Y) up to 30% of their weight. At the moment, lamp phosphors are generally landfilled due to their toxic mercury (Hg) content. Figure 5
shows the different REE fractions present in phosphors.
European feed for the important REE target in phosphors is estimated in the range of 140–2300 tons for Y, Tb, and Eu metals. In Table 5
, in-use stock amounts for Y, Eu, and Tb for the year were reported by Guyonnet et al. [8
]. Those of La, Ce, and Gd shown in Table 5
was calculated here, using the REE percentages in Figure 5
, and linking these to the percentage and in-use stock numbers of yttrium (arbitrarily chosen base case).
Mueller et al. (2017) conducted the same analysis for phosphors and batteries in Switzerland for the year 2014 [19
]. The available quantities of lamps were about 1169 tons with an REE mass fraction of 0.012 wt% per lamp. This gives a total mass of europium to be recovered in Switzerland in 2014 as 140 kg. A very high reverse supply collection and extractive yield of >80% was observed. The recycling of phosphors is nevertheless still uncommon. On the other hand, Guyonnet et al. estimated that 10 tons of Terbium will be recycled every year from phosphors, starting in 2020 [8
]. Except for some piloting efforts in Switzerland, this has still become reality [1
]. Notwithstanding these, Tb and Eu have a rather high potential for recycling from end-of-life lamp phosphors. Lamp collection costs could vary between 15 cents to a couple of Euros per unit kilogram. An important challenge linked to the recycling of phosphors from lamps is linked to mercury contamination. The treatment of Hg induces additional costs to ensure minimum human health impacts [1
]. Lamp recycling is estimated to generate a greenhouse gas (GHG) impact of 23.5 kg per unit kilogram of metal oxides. Energy-saving lamps are part of electrical appliances. Moreover, worldwide, the stock of end-of-life lamp phosphors is large.
In Switzerland, retailers have to take back end-of-life electrical and electronic equipment free of charge by law. This type of waste cannot go into the municipal solid waste stream. Today, the dismantling of this equipment is done outside of Switzerland in countries belonging to the European Union or members of the Organization for Economic Co-operation and Development (OECD). Sending waste to other countries is now forbidden according to new international regulations [6
As with magnets, different recycling methods also exist for phosphors:
Direct re-use of lamps (out phased due to technological change to LEDs);
Separation of phosphor components or mixtures and reuse in lamp (or other) industry;
Recovery of REE content on an individually high purity level.
In general, it is a challenge to recycle phosphors because of the contamination with mercury and the initiated replacement of lamps by other light-producing devices (LEDs). The first method can be to use still working electronics in spent and still active lamps in newer energy-to-light conversion devices. The second option is only applicable to one single type of lamp because different lamps use different types of phosphor mixtures. The advantage is that no chemical process is required, which makes this technique quite simple. The recycling technology may be a simple process, but it is difficult to reach pure phosphor fractions. The particle size of the phosphor may be changed to affect its light and energy conversion capabilities. The last method applies to all types of phosphors and the process is hydrometallurgical separation for the extraction of REEs. It produces high purity products that can be marketed irrespective of the application type. The disadvantage here is the production of high amounts of wastewater and the consumption of many chemicals [1
Recycling lamp phosphors is particularly challenging because lamps are made of glass, metals, plastic, and REEs containing phosphor powder. This powder represents only 3% of the mass of the lamp. Since all the other materials can be more easily recycled, they have gained more interest for now. With the Solvay group, Rhodia has focused on recycling phosphors under wet conditions at the dedicated facility at La Rochelle [13
]. According to Solvay, the process allowed treating and revalorizing more than 90% (1350 t/y) of fluorescent powders, thus recovering rare earths (Y, Eu, Tb, Gd, La, Ce) as oxides and nitrates, glass (to be valorized in the glass industry), and phosphates (valorized in the phosphates industry) [22
]. Fluorescent lighting is also changing nowadays. There has been an inflection point in demand for REEs for lighting, i.e. fewer REEs are used in LEDs than in fluorescent lamps. LEDs are going to replace all fluorescent lamps in the future because they have higher energy efficiency, longer lifetimes, and require less packaging [23
New phosphors are being developed at several U.S. national laboratories with the capacity to achieve the same efficiency and quality with much less amount of critical REE in use [24
]. This jeopardized the investments in the recycling of rare earths coming from phosphors since there will be fewer REEs in this sector in future years. Nevertheless, being able to recover REEs present in current devices may be a real strength for a society to be less dependent on primary extraction ores. Therefore, the recycling of light phosphors is still advised, also from an environmental point of view [1
]. As mentioned earlier, environmental impacts linked to the recycling of REEs are very low in comparison with the impact linked to mining.