Promoting future sustainable utilization of rare earth elements for efficient lighting technologies

Energy efficient lights, such as fluorescent lamps (FLs) and light-emitting diode (LED) lamps, can greatly help energy saving, which is critical for achieving carbon neutrality in the building sector. Yttrium, europium, and terbium are three essential rare earth elements (REEs) for energy efficient lighting. However, due to the ongoing lighting technology transition from FLs to LED lamps, the demands for yttrium, europium, and terbium have decreased significantly. It resulted in oversupplies of these three REEs in the lighting sector, indicating an economically and environmentally unsustainable supply chain. This study aims to estimate the supply and demand dynamics of yttrium, europium, and terbium in China from 2021 to 2060 under China’s carbon neutrality target by applying a dynamic-material-flow-analysis framework. Key flows and stocks along their life cycles are examined. Results show that the annual demands for yttrium, europium, and terbium in China’s lighting sector will decrease by 87%–100% from 2021 to 2060 under two demand scenarios. Driven by the demands for other co-produced critical REEs, the overall growing REEs supply will result in high surplus risks of yttrium and europium. Meanwhile, terbium deficit risk deserves more attentions due to its demand growth in other fields. Such surpluses of these three REEs in 2060 under six combined demand and supply scenarios are estimated to reach between 71 727 tonnes and 274 869 tonnes for yttrium, 530 tonnes and 1712 tonnes for europium, and −1360 tonnes (i.e. deficit) and 540 tonnes for terbium. Recycling activities of major co-produced REEs, such as neodymium, and the export expansion of surplus products can effectively mitigate such surplus risks. Finally, policy recommendations are proposed to improve the overall REEs efficiency by addressing the supply–demand imbalance and mitigating corresponding environmental impacts.


Introduction
Energy efficient lighting can greatly contribute to energy saving in the building sector, which is key to achieve global carbon neutrality (Bergesen et al 2016, Zhou et al 2018. Traditional lights, such as incandescent lamps, have been gradually replaced by more efficient fluorescent lamps (FLs) in the European Union, the United States, China, and other countries since early 2010s (Tan andLi 2014, Montoya et al 2017). More recently, light emitting diode (LED) lamps have been widely promoted due to their higher luminaire efficacy, longer lifespans, and less environmental impacts (Liu et al 2017, Pode 2020. The market share of LED lighting had increased from 15% in 2012 to 72% in 2020 in the Europe Union (European Commission 2021), and from 4% in 2012% to 78% in 2020 in China (Huaon Industry Research Institute 2021).
Yttrium (Y), europium (Eu), and terbium (Tb) are three key rare earth elements (REEs) that are widely used in phosphors employed in FLs and LED lamps due to their outstanding luminescent properties. Such applications account for 54%, 100%, and 89% of the total global consumption of yttrium, europium, and terbium in 2008, respectively (Nassar et al 2015). China has been the largest global supplier of all REEs, accounting for 58% of the global production with its 38% of the global reserve in 2020 (U.S. Geological Survey 2011-2021, Chinese Society of Rare Earths 2012-2019). Due to their importance to lighting technologies and other applications, yttrium, europium, and terbium have been individually identified as critical minerals by the United States, rather than classifying them as REEs as a whole (U.S. Geological Survey 2022).
While the supply risks of REEs have been highlighted, the lighting technology transition from FLs to LED lamps has significantly decreased the demands for yttrium, europium, and terbium due to their much lower contents in LED lamps, raising wide concerns on surplus issues (Zhou et al 2017, Ciacci et al 2018, Xiao et al 2022. Moreover, the co-production nature of REEs may exacerbate such surplus issues, especially under the low carbon energy transition contexts (Binnemans et al 2013, Klossek et al 2016. For example, the expansion of wind turbines and electric vehicles (EVs) has driven considerable demands for neodymium (Nd) and dysprosium (Dy) for manufacturing neodymium-ironboron (NdFeB) magnets, leading to oversupplies of several other REEs, including yttrium, europium, cerium (Ce), and lanthanum (La) (Elshkaki andGraedel 2014, Elshkaki 2021a). While an overall shortage of terbium has been uncovered (Binnemans et al 2013), potential terbium surplus in the lighting sector has also been highlighted (Zhou et al 2017). The surpluses of yttrium, europium, and terbium in the lighting sector can lead to severe resource waste and environmental consequences. However, there is a lack of a holistic assessment on the future demand and supply of these three REEs, especially for the largest producer of REEs, China.
Dynamic material flow analysis (DMFA) is a systematic method for investigating the dynamics of material stocks and flows within a specific system defined by time and space (Brunner and Rechberger 2004). Prospective DMFA based on scenarios analysis has been widely applied to assess future material demand (Müller et al 2014). Although many studies have assessed the demands for REEs and other critical minerals, several knowledge gaps still exist. Firstly, demand estimating studies focus on REEs required for specific technologies, such as neodymium and dysprosium for wind turbines and electric vehicle (EVs) (Hoenderdaal et al 2013, Watari et al 2019, Li et al 2020. Although both retrospective and prospective studies have been conducted to assess the demands for yttrium, europium, and terbium in the lighting sector at global, regional, and national levels, future demand dynamics driven by on-going lighting technology transition has not been estimated for China (Rollat et al 2016, Zhou et al 2017, Elshkaki 2021b, Gao et al 2022a, Xiao et al 2022. Also, comprehensive studies covering such group of REEs are still lacking. Secondly, future supply dynamics of REEs are rarely studied, especially the inelastic supply resulting from the co-production of REEs (Månberger andStenqvist 2018, Habib et al 2020). Thirdly, the metabolism patterns of REEs along their life cycles, i.e. the patterns of their flows between society and environment, have barely been analyzed (Daniels 2002). It leads to a considerable underestimation of demands for primary ores due to the neglect of REEs losses during mineral extraction and processing stages .
Under the above circumstances, this study aims to estimate future supplies and demands for three REEs (yttrium, europium, and terbium) along their life cycles in China from 2021 to 2060 under China's carbon neutrality target. Both DMFA and scenarios analysis methods are employed. We expect that this study can fill the above research gaps at the national scale so that valuable policy recommendations can be proposed to support sustainable REEs resource management. The remainder of this study is organized as below. Section 2 details research methods, including the DMFA-based estimation framework and scenarios setting. Section 3 presents research results, including the metabolism patterns of yttrium, europium, and terbium along their life cycles and a detailed analysis of their demands and supplies. Section 4 proposes policy recommendations. Finally, section 5 draws research conclusions.

Modeling framework
An integrated model is developed for estimating the future supplies and demands of yttrium, europium, and terbium along their life cycles (figure 1). The spatial boundary of this study is China's mainland, while the temporal boundary is from 2021 to 2060. Six life cycle stages are included: mining and beneficiation, refining and separation, fabrication, manufacturing, use, and waste management . Relevant methods and data sources are further detailed in the following sections.

Future metabolism of yttrium, europium, and terbium
A hybrid top-down and bottom-up accounting method is adopted to measure the future metabolism of yttrium, europium, and terbium along their life cycles by using equation (1). Taking the estimated primary supplies, i.e. supplies of ores, as inputs, the top-down method is used to calculate the production volumes in both mining and beneficiation and refining and separation stages. Taking the estimated final demands, i.e. demands for lighting applications, as inputs, the bottom-up method is used to calculate the required production volumes in manufacturing, fabrication, refining and separation, and mining and beneficiation stages, among which the last one refers to primary demands. As such, the surpluses are calculated as the differences between primary supplies and primary demands. In particular, assuming that the surplus yttrium, europium, and terbium are all stockpiled as refined products, the corresponding flows into tailings, stockpiles, and production losses are estimated by using equation (1) where I j,n,t , O j,n,t , RE j,n represent the inflow, outflow, and resource efficiency in the life cycle stage n of REE j in year t, respectively; RE j,n is defined as the ratio of outflow to inflow in each life cycle stage; j denotes yttrium, europium, or terbium; n denotes mining and beneficiation, refining and separation, fabrication, or manufacturing.

Estimation of demands for yttrium, europium, and terbium
A stock-driven DMFA method is used to account the inflows, outflows, and stocks of yttrium, europium, and terbium in China's lighting sector (Müller 2006). The newly introduced inflows from those lighting applications are referred to as the final demands. According to the mass balance principle, the newly introduced yttrium, europium, and terbium amounts (inflows) must satisfy both the increases in in-use stocks and the replacement for the end-of-life (EoL) amounts (outflows): where I i,j,t and O i,j,t represent the inflow and outflow of REE j in lighting application i in year t, respectively; S i,j,t and S i,j,t−1 represent the in-use stock of REE j in lighting application i in year t and t−1, respectively; Note: REO refers to rare earth oxide, which is converted to metallic equivalent content by multiplying the atomic mass ratio in this study.
i denotes FLs or LED lamps; S i,j,t and O i,j,t are estimated by using equation (S.1) and equation (S.2) in the supporting information (SI), respectively. In addition, since the consumptions of yttrium and terbium in other fields experienced linear growth trends during 2011-2020 (Gao et al 2022b, Xiao et al 2022, a linear model is applied to extrapolate their future final demands as shown in figure S1 of SI. Subsequently, primary demands for yttrium, europium, and terbium are determined by their final demands and the resource efficiency in specific stages along their life cycles, as illustrated in equation (1).

Estimation of supplies of yttrium, europium, and terbium
As shown in figure 1, two supply estimation methods are adopted under different assumptions. Assuming that primary supply is determined only by the passage of time, a time series prediction method, namely, the autoregressive integrated moving average (ARIMA) model, is adopted. Based on the historical rare earth concentrates production data from 1988 to 2020, the ARIMA model in this study is determined to be ARIMA (0,1,0) by using IBM SPSS statistics (see equation (S.3) in the SI). The residuals of this model are white noise, and the goodness-of-fit (R 2 ) is 0.908, indicating a well performed model. Primary supplies of yttrium, europium, and terbium ores are then determined based on the estimated future supplies of REEs concentrates by using the ARIMA model, the mining and beneficiation efficiency, and the weight proportions of REEs.
In another case, assuming that the supplies of yttrium, europium, and terbium are driven by the demand for neodymium in NdFeB magnets for wind turbines and EVs, DMFA-based prediction method is adopted. Such rapidly growing demands under China's carbon neutrality target will lead to a rapid increase in supplies of all REEs due to the coproduction nature. The final demand for neodymium is first estimated by using the DMFA method (see equation (S.4) in the SI). According to the flow interconnections in a material cycle shown in equation (1), primary supplies of yttrium, europium, and terbium are then determined by using equation (S.5) in the SI.

Scenarios development
Two demand scenarios are developed based on different evolutionary trends of the in-use penetration rates of lighting applications, namely the businessas-usual demand scenario (DBAU) and the carbon neutrality demand scenario (DCNS). Three supply scenarios, namely the business-as-usual supply scenario (SBAU), the carbon neutrality supply scenario (SCNS), and the recycling scenario (SREC), are also developed. Under the DBAU and SBAU scenarios, the evolution trend of demands and supplies of yttrium, europium, and terbium are assumed to follow the historical pattern. In contrast, China's carbon neutrality target is assumed to be the dominant driver of those demands and supplies in China under the DCNS, SCNS, and SREC scenarios.
Relevant parameters for the DBAU, DCNS, SCNS, and SREC scenarios are listed in tables 1-3, while those supplies under the SBAU scenario is directly estimated by using the ARIMA model. More detailed assumptions for demand and supply scenarios are listed in the SI. Besides, sensitivity analysis is conducted to investigate the impacts of key parameters, in which the combined demand and supply scenario named DBAU-SBAU scenario is selected as the proxy. The results of sensitivity analysis are listed in Note: DDPMSG and GDPMSG refer to direct-driven and geared-driven train permanent magnet synchronous generators for wind turbines, respectively. Note: LDV and HDV refer to the light-duty vehicle and high-duty vehicle fleets, respectively. It is assumed that BEVs will dominate the LDV market while FCVs will dominate the HDV market in 2060.
the SI, indicating that our main findings are robust (see figure S3 in the SI).

Data sources
Diverse data are collected in this study, including three categories: rare earth mining and production, rare earth consumption, and technological development. Historical production of REEs concentrates and weight proportions of REEs in ores are collected from the Chinese rare earth yearbooks (Chinese Society of Rare Earths 2012-2019). Resource efficiency data of rare earth cycles and historical consumption data of yttrium, europium, and terbium products are collected from existing DMFA studies and CBC (China Bulk Commodities) database (Geng et al 2020, Gao et al 2022a, Xiao et al 2022. Data on lighting applications, wind turbines, and EVs are derived from relevant industrial reports, research reports, and academic papers. Tables S1-S7 in the SI list these relevant data. Figure 2 illustrates the flows of yttrium, europium, and terbium along their life cycles in China for the years 2021, 2030, and 2060 under the DBAU-SBAU scenario, which is the baseline scenario. Since China has become the world's largest producer and consumer of REEs, the extraction of rare earth ores is expected to grow. The amount of extracted yttrium ores will be more than 100 times larger than those of europium and terbium because of the much higher yttrium content in REE ores in China. Due to inefficient mining and beneficiation processes, a large amount of the extracted ores will be dumped into the local environment as tailings (with cumulative amounts of 5946 522 tonnes of yttrium, 34 459 tonnes of europium, and 55 634 tonnes of terbium), accounting for 69% of the total ores. In the refining and separation stage, due to the imbalance of supply and demand, the cumulative stockpiled surpluses will reach 2228 654 tonnes of refined yttrium, 13 746 tonnes of refined europium, and 4902 tonnes of refined terbium by 2060. Such results indicate an inefficient supply chain of these three REEs, leading to high economic and environmental costs (Lee and Wen 2018). In the manufacturing stage, a total of 2526 tonnes of yttrium, 256 tonnes of europium, and 170 tonnes of terbium will be fabricated into rare-earth doped phosphors and added to energy efficient lighting products, namely FLs and LED lamps, from 2021 to 2060. Since the lighting technology will transfer completely from FLs to LED lamps, the consumptions of yttrium, europium, and terbium in the lighting sector will decrease significantly from 332 tonnes, 30 tonnes, and 29 tonnes in 2021 to 18 tonnes, 2 tonnes, and 0.01 tonnes in 2060, respectively. This is due to the much lower yttrium, europium, and terbium contents and higher lighting efficiency in LED lamps. Moreover, with the continuous growth of yttrium and terbium consumptions in other fields, the proportion of lighting applications in their total consumption is estimated to fall below 5% after 2027. In addition, the cumulative discarded amounts of yttrium, europium, and terbium in the lighting sector will reach 3230 tonnes, 317 tonnes, and 237 tonnes in 2060. Due to the high recycling cost, recycling REEs from EoL lighting applications will be unprofitable in the future (Qiu and Suh 2019). Therefore, it is likely that all of these EoL REEs will be discarded directly into the landfills.

Demands for yttrium, europium, and terbium in the lighting sector
The cumulative demands for yttrium, europium, and terbium in China's lighting sector under both DBAU and DCNS scenarios are illustrated in figure 3. The cumulative demand for yttrium will be the largest, among which 1926 tonnes and 600 tonnes will be used in the fields of FLs and LED lamps by 2060 under the DBAU scenario. Such demand for europium will be 172 tonnes and 84 tonnes by 2060. Besides, terbium will be only used in FLs, with a cumulative amount of 170 tonnes by 2060. However, under the DCNS scenario, with an accelerated phaseout of FLs and a completed replacement by LED lamps in 2030, and lighting efficiency improvement of LED lamps, the cumulative demands for these three REEs in the fields of FLs and LED lamps will be 78% and 24% lower than those under the DBAU scenario by 2060.
Figures 4(a)-(f) further illustrate annual inflows, outflows, and stocks of yttrium, europium, and terbium in China's lighting sector under the DBAU and DCNS scenarios. Under both scenarios, the annual inflows, i.e. annual demands, for these three REEs will decrease by 87%-100% from 2021 to 2060, resulting in that such stocks will first decrease and then remain at an extremely low level. Under the DBAU scenario, yttrium, europium, and terbium stocks will be stable after 2045, with figures of 113 tonnes, 14 tonnes, and 3 tonnes in 2045, respectively. Under the DCNS scenario, yttrium and europium stocks will be stable after 2029, with figures of 65 tonnes and 9 tonnes in 2029. Besides, 62 tonnes of terbium stocks will be all scrapped by 2030. Such results reflect that the lighting technology transition will be accelerated under the carbon neutrality target, leading to many FLs being replaced by LED lamps before reaching their life spans. Considerable amounts of yttrium, europium, and terbium will be scrapped in the short term, peaking at 389 tonnes in 2024 under the DBAU scenario and at 360 tonnes in 2022 under the DCNS scenario. This means that future demands for yttrium, europium, and terbium in the lighting sector will be mainly used to replace FLs.

Supply-demand imbalances
The surpluses of yttrium, europium, and terbium under six combined demand and supply scenarios, namely S1-S6, are illustrated in figure 5. Since the increasing primary supplies will far exceed the shrinking demands, there will be annual surpluses of yttrium, europium, and terbium under all the scenarios, except obvious terbium shortages under S3 and S6. The cumulative surpluses of yttrium, europium, and terbium will be the largest under the DCNS-SCNS scenario (S5). Specifically, annual surpluses of yttrium and europium will increase continuously from 58 194 tonnes and 381 tonnes in 2021 to 274 869 tonnes and 1712 tonnes in 2060, respectively. The relationship between supply and demand for terbium will change from a deficit of 240 tonnes in 2021 to a surplus of 540 tonnes in 2060, with a cumulative surplus of 21 650 tonnes.
Compared to the primary supply with high economic costs and serious environmental consequences, recycling (known as secondary supply of REEs) can significantly improve the overall efficiency of the rare earth industry and mitigate the supply risks of those critical REEs. Given the co-production nature of REEs, such a secondary supply option can also avoid the excess production of certain REEs (Schulze et al 2018). Under the DCNS-SREC scenario  (S6), assuming that 75% of neodymium from EoL NdFeB magnets can be used to replace primary ones, the surpluses of yttrium, europium, and terbium will peak at 172 218 tonnes, 1084 tonnes, and 522 tonnes by 2033. As a result, the cumulative surpluses of yttrium and europium in the S6 will be 46% and 49% less than those in the S5 by 2060. However, as the primary supply decreases, it will be an increasing deficit of terbium due to its increasing demand in other fields, growing significantly from 2 tonnes in 2038 to 1360 tonnes in 2060.
In addition, the surplus risks of yttrium, europium, and terbium are assessed by using two indicators, namely annual surplus risk (ASR) and cumulative surplus risk (CSR), as shown in figures 5(d)-(e). The ASR was calculated as the maximum ratio of annual surplus to annual supply. The CSR was calculated as the ratio of cumulative surplus to cumulative supply. The former is used to reflect the dynamics of the supply-demand relationships to provide valuable information for short-and medium-term REEs management decision making, while the latter is used to reflect the total surplus risks, which is helpful to uncover long-term impacts of REEs management decisions. The surplus risk of europium is highest due to its single application area and low content in lighting applications, indicating 98%-100% of ASR and 96%-99% of CSR. Such indicators of yttrium are 91%-93% and 85%-92%, respectively. Under the S1, S2, S4, and S5 scenarios, low surplus risks and medium deficit risks of terbium are both presented, with −52%-33% of ASR and 22%-26% of CSR. However, the deficit risk of terbium deserves more attentions under the S3 and S6, in which ASR are −156% and −156% and CSR are −37% and −36%, respectively.

International trade perspectives
This study only estimates domestic consumptions of yttrium, europium, and terbium in China's lighting sector. However, about 65% and 37% of total FLs and LED lamps manufactured in China during 2011-2020 were exported, respectively (China Customs 2012-2017, UN Comtrade 2023. Since China's lighting application manufacturing has shifted to an export-driven pattern, the demands for yttrium, europium, and terbium will be higher under the globalization context. Based on the DBAU-SBAU scenario, three additional scenarios regarding export expansion are set up. The net export growth rates of China's lighting applications, i.e. FLs and LED lamps, are assumed to be 5%, 10%, and 15% under the low, medium, and high scenarios, respectively. The medium level of 10% is based on the average growth rate of China's exports of all goods in the last 20 years (State Council Information Office 2022).
Under the low scenario, the estimated net exports of yttrium, europium, and terbium embodied in lighting applications will only offset 1%, 18%, and 57% of their surpluses in 2060, respectively (see figure  S4 in the SI). At a medium growth rate of net exports, the surplus risks of europium and terbium can be significantly mitigated after 2050 and 2039, indicating that net exports will offset more than 50% of their surpluses. Such turning points for europium and terbium will happen in 2040 and 2033 under the high scenario. In addition, net yttrium exports will account for 8% and 45% of its surpluses in 2060 under the medium and high scenarios, respectively. Although the deficit risk of europium and terbium deserves more attentions under the medium and high scenarios, scenarios analysis results indicate that it is necessary to support export promotion policies to avoid the surplus risks of yttrium, europium, and terbium in China's lighting sector.

Policy recommendations
The above results provide valuable insights to prepare appropriate resource management policies for yttrium, europium, and terbium. The following policy recommendations are proposed by considering the Chinese realities.
Firstly, REEs supply chains should be optimized to solve the supply-demand imbalance of these three REEs. Our results present that the domestic lighting technology transition will lead to the growing surpluses of yttrium, europium, and terbium. The Chinese government should release targeted management strategies for more efficient and effective supply chains of such REEs from both demand and supply perspectives. In terms of expanding market, with China's dominance in the global REEs market, international trade can help alleviate such an oversupply. It is therefore crucial for the Chinese government to encourage the exports of relevant surplus products. In particular, most developing countries cannot afford the high costs of LED lamps and may still like to purchase FLs for their lighting service (Adjei-Mantey and Adusah-Poku 2021). This is particularly important for europium, which is mainly used for lighting. In terms of diversifying supply sources, it is critical to promote comprehensive tailings utilization and critical REEs recovery from the production wastes and those EoL products, such as the case of neodymium recovery in this study. These measures can reduce the unnecessary primary supplies of co-produced REEs surplus, including yttrium, europium, terbium, cerium, and lanthanum (Zheng et al 2022). Besides, the deficit risk of terbium caused by its consumption growth in other sectors should be investigated. It is necessary to build up a well-designed strategic reserve system to ensure its stable supply. In addition, an information platform for the entire REEs supply chain should be created to enhance communication and cooperation among different stakeholders.
Secondly, it is urgent to seek new applications with higher added values to consume the surpluses of yttrium, europium, and terbium. It is estimated that the quick replacement of FLs by LED lamps will significantly reduce the demands for yttrium, europium, and terbium in the lighting sector. Therefore, research and development activities should be encouraged to boost potential demands in other fields. In addition to the well-known application field of NdFeB magnets for terbium, promising applications include superconductors for yttrium, nanocomposites for europium, and biomedicine for europium (Wang et al 2014, Kong et al 2021, Lv et al 2022. But these applications require more investment so that the largescale production of such innovative products can be achieved. A university-industrial partnership is encouraged so that potential technological problems can be solved. Besides, appropriate pricing on REEs can provide sufficient economic drivers for these promising applications. However, the true prices of REEs usually cannot be accurately reflected in the market, since high production costs are shared by all REEs extracted from the ores. China's domestic prices of yttrium, europium, and terbium oxides are more than 8, 28, and 1800 times higher than those of other surplus REEs (such as lanthanum and cerium), reaching 62, 198, and 13 216 yuan per kilogram on October 2022 (Chinese Society of Rare Earths 2022). Such higher prices have impeded their wide applications. Therefore, the Chinese government must intervene in pricing such products so that reasonable prices can be set up. Specific measures include releasing price adjustment funds, charging REEs mining rents, and issuing and regularly updating the lists of guiding prices for rare earth-containing products (The National People's Congress of the People's Republic of China 2014, Li 2017).
Thirdly, environmental impacts along the REEs supply chains should be mitigated. Our results indicate that a large amount of yttrium, europium, and terbium will be lost as tailings during the mining and beneficiation processes. Therefore, it is crucial to seek more efficient mining and beneficiation technologies so that the overall tailings and corresponding environmental impacts can be minimized, especially the radioactive thorium emissions and the acidification issue (Lee and Wen 2017). Another useful economic instrument is to impose resource tax on yttrium, europium, and terbium-containing products. The current resource tax is imposed on rare earth concentrates (Ministry of Industry and Information Technology 2019). It is critical to set up appropriate resource tax rates for rare earth-containing products, such as rare earth oxides and functional materials, so that the true environmental externalities can be internalized (Arshi et al 2018). Especially, a new standard should be set up so that appropriate resource taxes can be determined on those surplus REEs caused by the co-production of other critical REEs, such as neodymium and dysprosium. Such tax can be used to support effective tailings treatment and other environmental remediation efforts.

Conclusions
REEs are vital resources for high-tech industries and achieving carbon neutrality targets. However, driven by the co-production of REEs, oversupplies of certain REEs have been observed. This study estimates the supplies and demands for yttrium, europium, and terbium, as well as their metabolism patterns from a life cycle perspective in China from 2021 to 2060. The main results show that the rapid lighting technology transition will significantly decrease the demands for yttrium, europium, and terbium in China's lighting sector. With the growing supplies, the surplus of yttrium will be the largest, followed by europium and terbium. Based on these results, several policy recommendations are proposed, including optimizing REEs supply chains, promoting new REEs applications, mitigating corresponding environmental impacts.
However, several research limitations exist. Firstly, our estimation on future REEs supply does not include external factors, such as the improvement of production technologies and environmental constraints. Also, a linear extrapolation method was adopted to estimate the demands for yttrium and terbium in other fields without considering potential emerging sectors, technological substitutions, etc. In the future, these external factors should be considered and incorporated into the estimation model so that more accurate results can be obtained. Lastly, only China's mainland is covered in this study. It is necessary to scale up to the global level so that a holistic picture can be obtained, such as the increasing demands for such REEs in developing economies and the decreasing demand for such REEs in developed economics.

Data availability statement
All data that support the findings of this study are included within the article (and any supplementary files).