Geological Survey of Denmark and Greenland Bulletin 41, 2018, 87-90 87 Rare-earth elements (REE) are considered Critical Raw Materials (CRM; EC 2018; US Department of the Interior 2018) and essential in the technological transformation of the energy sector into carbon-free technologies such as wind turbines, electrified transport and LED-lights. The new technologies have led to swiftly expanding markets for REE products, in which China has achieved a monopolistic role in all segments of the REE value chains. Political strategies aimed to establish REE supplies outside China are currently being implemented within the EU and in other Western countries in order to ensure an adequate future REE supply. However, new REE value chains outside China have not yet materialised. The aim of this paper is to assess whether the global REE supply from present and potential mines can keep pace with the REE demand for the expanding offshore wind energy sector (Fig. 1). A successful development of this sector out- side China relies on an adequate supply of particularly neo- dymium (Nd) and to some extent praseodymium (Pr), ter- bium (Tb) and dysprosium (Dy), used in permanent magnets for windmill generators. In 2015, about 82% of the global Nd-oxide production was used in the permanent magnets Examining the rare-earth elements (REE) supply– demand balance for future global wind power scenarios Per Kalvig and Erika Machacek 37 38 Mine Deposit type PEA started/ completed Mine development FS started/ completed Abandoned/ on hold Resource estimate 36 39 8 12 11 9 7 10 22 21 2943 42 40+41 35 34 4527 28 49 26 46 33 32 30 31 25 13 48 17 16 15 14 20 18+19 24 44 4723 34 51 2 6 Carbonatite Iron-oxide-apatite Hydrothermal Alkaline igneous Placer Bauxite Ion-adsorption clays 1 Browns Range 2 Charley Creek 3 Dubbo Zirconia 4 Mount Weld 5 Nolans 6 Yangibana 7 Ashram Main 8 Eco Ridge 9 Foxtrot 10 Kipawa/Zeus 11 Montviel 12 Strange Lake 13 Bayan Obo 14 China Minmetals 15 China Minmetals 16 China Minmetals 17 Chinalco 18 Ganzhow Mining Group 19 Guangdong Rising NF 20 Xiamen Tungsten 21 Kringlerne 22 Kvanefjeld 23 Chavara 24 Manavalakurichi 25 Kutessay II 26 Ngualla 27 Kangankunde 28 Songwe Hill 29 Lofdal 30 Khibiny 31 Lovozero 32 Norra Kärr 33 Aksu Diamas 34 Ngualla, Tanz. 35 Wigu Hill 36 Bear Lodge 37 Bokan 38 Mountain Pass 39 Round Top 40 Glenover (RSA) 41 Steenkampskraal 42 Zandkopsdrift 43 Buena Norte 44 Lahat 45 Tantalus 46 Gakara 47 Dong Pao 48 Tomtorskoye 49 Nkombwa Hill GEUS/MiMa April 2018 Fig. 1. Global rare-earth element mines and advanced exploration projects. FS: Feasibility study. PEA: Pre-economic assessment. © 2018 GEUS. Geological Survey of Denmark and Greenland Bulletin 41, 87–90. Open access: www.geus.dk/bulletin http://www.geus.dk/bulletin 8888 production (Adamas 2016). Here we evaluate the future sup- ply and demand situations for Nd, Pr, Tb and Dy in the glob- al wind energy sector in the form of three scenarios, one for 2020 and two for 2030 based on high and low demand. The balance is discussed. Our assessment ref lects the challenge caused by limited insight into the REE supply chains inside China, and the figures presented in this paper are therefore only indicative. Scenarios for future global REE demand of the wind energy sector In 2016, the Global Wind Energy Council reported a total global wind power capacity of 487 GW (GWEC 2016), of which the offshore capacity amounted to 3% according to the Global Status Report 2017 for Renewables. Due to the otherwise very high maintenance costs of the offshore wind energy, it depends in part on direct-drive and hybrid wind turbine technologies that use REE-based permanent magnet (PM) or high-temperature superconducting (HTS) genera- tors (Barteková 2016). This is why the offshore wind energy sector is the focus of this study. Our scenarios solely deal with technologies using perma- nent magnets in which Nd is vital. Barteková (2016) specifies the consumption of total rare-earth element oxides (TREO) for the individual types of magnets used in wind turbines and indicates the individual REE used. The REO content per magnet varies by generator design between c. 23–35%, highest in the permanent magnet of a direct-drive synchro- nous generator (PMSG-DD), and lowest in hybrid single and multistage synchronous generators (PMSG-SG and PMSG- MG). The relative magnet weight proportions of the four REOs are about 95% Nd, 4% Pr, 0.99% Tb and 0.01% Dy (Barteková 2016). In order to estimate the future REO-demand of the sector, the following assumptions about the wind energy technolo- gy are made based on Barteková (2016, p. 158), who reported that in 2014 the REE-based permanent magnet technology accounted for 4% of the offshore wind technology, equally divided between direct-drive and hybrid generator designs. For our 2016 baseline scenario, we (i) increased this share to 5%, (ii) maintained the equal split between direct-drive and hybrid generator technologies, and (iii) subdivided the hybrid generator designs equally into single and multistage gearbox designs. Further, we assume that in 2016, China held 8% and the rest of the world (ROW) 92% of the global offshore wind capacity. From the 2016 baseline scenario, we developed three sce- narios: (1) the REO use in the total global offshore wind en- ergy capacity in 2020, and (2, 3) low and high forecasts of the same in 2030. We set the regional offshore shares to 85% for ROW and 15% for China. See details in Table 1. Pr Nd Tb Dy Pr Nd Tb Dy Pr Nd Tb Dy Pr Nd Tb Dy 2016 2020 2030-low 2030-high     ROW 104 2.5 26 0.26 267 6.3 66 1 589 13.9 146 1 986 23.4 244 2 9 ton ton tonkton ton ton tonkton ton ton tonkton ton ton tonkton 215 2 0.02 23 551 6 0 104 2.5 26 0 174 4.1 43 0China 0 5000 0 10 000 15 000 20 000 25 000 ton A     China 0 200 400 600 800 1000 1200 ton Pr 4 0 90 8 1 0 0 0 10 1 230 20 2 0 0 0 0 0 45 8 0 0 11 2 1060 187 27 5 631 111 7 1 ROW Nd Tb Dy Pr Nd Tb Dy Pr Nd Tb Dy Pr Nd Tb Dy 2016 2020 2030-low 2030-high B Fig. 2. Rare-earth metal demand by wind energ y deployment in China and the ROW in 2016, and forecasts for 2020 and 2030. A: assuming that all offshore wind technolog y is centred on R EE-based permanent magnets. B: assuming that only a small share of offshore wind technolog y uses R EE- based permanent magnets, and differentiating between varying R EE uses per PM generator technolog y. Note: Both figures show cumulated (fore- casted) individual R EE use by examined wind energ y technolog y in the respective year. ROW: the rest of the world. Table 1. Forecast scenarios for installed global offshore wind energy capacity in 2020 and 2030 Forecast 2020 739 GW 5% = 37 GW 5% = 1.85 GW Forecast 2030 low 1260 GW 7% = 88 GW 10% = 8.80 GW Forecast 2030 high 2110 GW 7% = 148 GW 10% = 14.80 GW REE permanent mag- nets installed capacity offshore capacity %assumptions, this survey From GWEC (2016a) Global totalWind energy capacity scenarios Global offshore 89 Our scenarios are based on the estimates by GWEC (2016) for 2020, 2030-low, and 2030-high global wind en- ergy capacity. It is important to note the large variability in the underlying assumptions. The most significant parameter is the share of REE-based permanent magnets deployed in offshore installations. For instance, if all offshore technol- ogy capacity would employ REE-based permanent magnet technology and 75% of this was installed with the highest REE-using direct-drive design, this would result in the con- sumption 16 400–27 500 tons Nd-oxide in 2030 ( Fig. 2A). In contrast, the wind-turbine sector will demand only about 740–1250 tons Nd-oxide (Fig. 2B) for the alternative tech- nology split-up outlined in Table 1. Scenarios for the future global REE supply In 2016, the primary global production of total rare-earth oxides (TREO) amounted to 129 000 tons, of which China produced 83%, Australia 11%, Russia 2%, Brazil 1%, India 1% and Malaysia, Thailand, and Vietnam still less (USGS 2017). Although the TREO supply figures for 2016 reported by different sources are rather similar (USGS 2018: 23 680 tons; Adamas 2016: 24  377 tons), there are major discrep- ancies at national level. This partly stems from the assumed contributions of the non-reported market which may ac- count for 25–30% (Roskill 2016) and from uncertainties pertaining to production and smelting quota (Adamas, per- sonal communications 2018). This study applies the figures for for 2016 from USGS (2018), and contribution from non- reported production is not considered. The 2016 supplies of Pr, Nd, Dy and Tb from China and seven ROW countries are shown in Figs 3A, B. The official Chinese production quota for 2016 was set to 105 000 tons TREO (Machacek & Kalvig 2017); our estimate of the regional REO production is based on Kingsnorth (2016) and shown in Fig. 3A. Given that no scheduled production quotas for 2020 and 2030 are available, our China 2020 and 2030 supply scenarios are ar- bitrarily set to an increase of 5% p.a., ref lecting the antici- pated growth in demand (Dutta et al. 2016), the potential for higher capacity on existing plants, as well as continued efforts to transform the informal sector into to a formal one. The ROW supply in the 2020 and 2030 scenarios is de- veloped as follows: of the recorded 320 REE exploration projects outside China, 99 are reported to be active (S&P- database 2017). Our search revealed that 31 of these projects, located in 12 countries, have reached an advanced stage (Fig. 1). These 31 projects are divided into four classes of develop- ment, which are in turn translated into expected production start-ups in 2020, 2025, 2030 and 2035. The estimates of relative REO grades and targeted pro- duction of Pr, Nd, Tb and Dy are based on TMR (2015) and company data. Where relative grade data are not available, our estimate is based on the actual REE mineralogy. Accord- ing to these data, new REO productions in Australia, USA, and Vietnam are expected in 2020. In 2030, REO will also - 5000 10 000 15 000 20 000 25 000 30 000 35 000 Pr A Nd Tb Dy Pr Nd Tb Dy Pr Nd Tb Dy 2016 2020 (5% growth p.a.) 2030 (5% growth p.a.) Yunnan Hunan Fujian Guang- dong Guangxi Jiangxi Shandong Sichuan Inner Monglia To n 46 69 17 4 23 28 4 88 6 56 75 21 1 78 34 5 56 2 17 54 10 77 9 24 4 34 4 96 1000 0 2000 3000 4000 Brazil Vietnam Thailand Malaysia India Russia Australia B Pr Nd Tb Dy Pr Nd Tb Dy Pr Nd Tb Dy 2016 2020 2030 To n 10 49 32 74 62 33 1 10 49 37 24 62 33 1 10 49 37 24 62 33 1 5000 0 10 000 15 000 20 000 25 000 Vietnam USA Turkey Tanzania Sweden South Africa Namibia Malawi Greenland Canada Burundi Australia C 2016 2020 2030 Pr Nd Tb Dy Pr Nd Tb Dy Pr Nd Tb Dy 0 0 0 0 23 2 76 0 32 20 5 72 04 24 9 49 37 7 21 59 To n Fig. 3. Current and future R EE production scenarios. A: Current and forecasted Chinese production of praseodymium, neodymium, terbium and dysprosium in seven regions. Forecast assumes a general 5% annual in- crease. B: Current and forecasted production of praseodymium, neodym- ium, terbium and dysprosium in ROW. Forecast assumes a static produc- tion. C: Forecasted production of praseodymium, neodymium, terbium and dysprosium from 31 advanced R EE projects outside China. 9090 be supplied by Canada, Greenland, Malawi, Namibia, South Africa, Sweden, Tanzania and Turkey (Figs 1, 3C). The sce- narios indicate ROW-TREO productions of 6920 tons in 2020 and 154 075 tons in 2030 (Fig. 3C). Discussion and summary We demonstrate here that the level of detail applied to esti- mates of the future use of different types of generators and their relative shares allow for great variance in the current and forecast REE demand by the wind energy sector. Against this background, our ROW scenario for 2020 points to a Nd demand by the wind energy sector within a wide range of 230–6300 tons (Fig. 2), while our ROW supply forecast for Nd-oxide from both current and new mines is around 4500 tons. A top-down approach indicates that the REE-based permanent magnets for the wind sector absorb 10%, equiva- lent to about 4000 tons TREO (Lucas et al. 2015). Adamas (2016) estimates that Nd-oxide accounts for c. 73% of the TREO in the permanent magnets for the wind sector. In ef- fect this means that in 2020, the total Nd-oxide supply for the wind energy sector would be roughly 3000 tons Nd-ox- ide, i.e. in the middle range of the forecasted global demand for this purpose. For the 2030 low scenario and with REE technology applied in some but not all offshore technology, the global Nd-oxide demand by the wind sector is forecasted to c. 740 tons and for the 2030 high scenario close to 1250 tons. If it is assumed that all offshore technology will draw on REE use, these figures increase to 16 400 and 27 500 tons, respectively. Our global supply forecast of Nd-oxide in 2030 from current operations is about 35 000 tons from China and 3700 tons from ROW, to which advanced ROW REE projects could contribute an additional c. 25 000 tons if all projects go into production. If the 10% share of the REE-based permanent magnet sector as well as a stable demand for REE-based per- manent magnets from the wind sector are maintained, about 450 tons Nd-oxide could be made available by ROW suppli- ers in 2020, and 2900 tons in 2030. However, in the scenar- ios based on generator technologies that consume a higher percentage of REE-based magnets, the estimated ROW sup- ply is inadequate and an additional Nd-oxide supply will be required, e.g. from Chinese REE operators. This study shows that there are currently significant un- certainties in trying to determine both the current REE de- mand and supply in the wind sector and in building scenarios for 2020 and 2030. Given that Nd-Dy permanent magnets represent a fast-growing sector, there is a need to establish a comprehensive, research-based and harmonised framework for precise estimates of the future supply and demand sce- narios for REE-based permanent magnets. References Adamas 2016: Rare Earth Market Outlook: supply, demand, and pricing from 2016 through 2025 Report, 822 pp. Adamas Intelligence. Barteková, E. 2016: The role of rare earth supply risk in low-carbon tech- nolog y innovation. In: Borges De Lima, I. & Leal Filho, W. (eds): Rare earths industry: technological, economic, and environmental implica- tions. Amsterdam: Elsevier, 437 pp. Dutta, T., Kim, K-H, Uchimiya, M., Kwon, E.E., Jeon, B-H, Deep, A. & Yun, S-T. 2016: Global demand for rare earth resources and strategies for green mining. Environmental Research 150, 182–190. EC 2018: Report on Critical Raw Materials and the Circular Economy, 69 pp. Commission Staff Working Document. SWD(2018) 36 final. GWEC 2016: Global Wind Energ y Outlook 2016. Fried, L. et al. (eds), 44 pp. GWEC and Institute for Sustainable Futures, University of Tech- nolog y Sydney. Kingsnorth, D.J. 2016: Curtin University publication on the rare earth industry in 2016. Lucas, J., Lucas, P., Le Mercier, T., Rollat, A. & Davenport, W. 2015: Rare- earth-based permanent magnets preparation and uses, 231–249. In: Rare Earths. Amsterdam: Elsevier. Machacek, E. & Kalvig, P. (eds) 2017: Road map for R EE material supply autonomy in Europe. EUR AR E, European R EE market survey (com- ponent D1.2). 141 pp + appendix. Roskill Informations Services Ltd. 2016: Rare Earths: Global industry, markets and outlook to 2026, 396 pp. Roskill Reports on Metals and Minerals, 16th revised edition. S&P-database 2017: S&P Global Market Intelligence Metals & Mining Database 2017. Technolog y Metals Research (TMR) 2015: Advanced Rare-Earth Pro- jects: http://www.techmetalsresearch.com/metrics-indices/tmr-advanced- rare-earth-projects-index/ (accessed 9 February, 2018). US Department of the Interior 2018: Draft List of Critical Minerals. Feb 16, Notices. Federal Register 83(33), 7065–7068. USGS 2018: Mineral Commodity Summaries. US Geological Survey. Authors’ address Geological Survey of Denmark and Greenland, Øster Voldgade 10, DK-1350 Copenhagen K, Denmark. E-mail: pka@geus.dk. http://www.techmetalsresearch.com/metrics-indices/tmr-advanced-rare-earth-projects-index/ http://www.techmetalsresearch.com/metrics-indices/tmr-advanced-rare-earth-projects-index/ mailto:pka@geus.dk