Sustainable Urban Mining of Critical Elements from Magnet and Electronic Wastes

Ames Laboratory Accepted Manuscripts Ames Laboratory

11-12-2019

Denis Prodius Ames Laboratory, [email protected]

Kinjal Gandha Ames Laboratory, [email protected]

Anja-Verena Mudring Ames Laboratory

Ikenna C. Nlebedim Ames Laboratory, [email protected]

Follow this and additional works at: https://lib.dr.iastate.edu/ameslab_manuscripts

Part of the Other Materials Science and Engineering Commons, and the Sustainability Commons

Recommended Citation Prodius, Denis; Gandha, Kinjal; Mudring, Anja-Verena; and Nlebedim, Ikenna C., "Sustainable Urban Mining of Critical Elements from Magnet and Electronic Wastes" (2019). Ames Laboratory Accepted Manuscripts. 545. https://lib.dr.iastate.edu/ameslab_manuscripts/545

This Article is brought to you for free and open access by the Ames Laboratory at Iowa State University Digital Repository. It has been accepted for inclusion in Ames Laboratory Accepted Manuscripts by an authorized administrator of Iowa State University Digital Repository. For more information, please contact [email protected]. Sustainable Urban Mining of Critical Elements from Magnet and Electronic Wastes

Abstract A straightforward and environment-friendly process for acid-free leaching of rare-earth elements and cobalt, which are critical materials, from waste magnet materials has been developed. The process also allows for selective leaching of rare-earth elements from magnet-containing electronic wastes, such as end-of-life hard disk drives and electric motors. The use of copper salts eliminates the use of volatile toxic acids in the dissolution and separation processes, which allows for a more eco-friendly approach to recovering critical elements and a safer work environment. Recovered critical materials were shown to be suitable for reinsertion into the materials supply chain.

Keywords Acid-free leaching, Rare-earth permanent magnets, Rare-earth recovery, Sustainability

Disciplines Other Materials Science and Engineering | Sustainability

This article is available at Iowa State University Digital Repository: https://lib.dr.iastate.edu/ameslab_manuscripts/ 545 Sustainable Urban Mining of Critical Elements from Magnet and Electronic Wastes

Denis Prodius, Kinjal Gandha, Anja-Verena Mudring,†* and Ikenna C. Nlebedim* Critical Materials Institute, Ames Laboratory, U.S. Department of Energy, 2332 Pammel Drive, Ames, Iowa 50011, USA [email protected]

ABSTRACT: A straightforward and environment-friendly pro- higher recovery rates of REEs, especially as oxides or in other non- cess for acid-free leaching of rare earth elements and cobalt, which metallic forms. Nevertheless, most hydrometallurgical approaches are critical materials, from waste magnet materials has been devel- are energy-intensive and require substantial amounts of hazardous oped. The process also allows for selective leaching of rare earth chemicals, especially strong mineral acids such as hydrochloric, elements from magnet-containing electronic wastes, such as end- sulfuric and nitric acids. High volumes of wastes including residual of-life hard disk drives (HDDs) and electric motors. The use of cop- strong mineral acids present obvious environmental problems. In- per salts eliminates the use of volatile toxic acids in the dissolution vestments to contain both the acids and their contaminated wastes and separation processes, which allows for a more eco-friendly ap- add to the cost of the process. Thus, there is an obvious need for a proach to recovering critical elements and a safer work environ- cost-effective, environment-friendly and energy-efficient method ment. Recovered critical materials were shown to be suitable for for recycling REEs-containing materials. reinsertion into the materials supply chain. Herein, we report a novel and safer process for leaching REEs and KEYWORDS: Acid-free leaching, Rare earth permanent mag- cobalt from permanent magnet materials. The chemical dissolution nets, Rare earth recovery, Sustainability method involves contacting the REEs-containing material and an aqueous solution of a copper(II) salt to dissolve magnet materials. INTRODUCTION The REEs are then precipitated from the aqueous solution, which then can be calcined to produce REE-oxides or other REE- The global need for rare earth elements (REEs) is increasing be- compounds. This process also enables selective leaching of the cause they are indispensable for a wide range of applications in- REEs contained in waste magnets from e-waste products such as cluding permanent magnets, lighting, lenses, catalytic converters shredded hard disk drives (HDDs) and crushed electric motors. 1-7 and many consumer electronics. New disruptive technologies Using copper(II) salts for the oxidative dissolution of magnets al- can increase the need for REEs and result in significant market un- lows for a selective and atom economy transfer of the relevant met- certainties. In magnet applications alone, the use of REEs is ex- als into solution. Moreover, it helps to avoid the use of strong min- pected to increase and has been forecasted to exceed $50 billion eral acids, hence eliminating the associated harsh reaction condi- 8 market value by 2022. Neodymium (Nd), dysprosium (Dy), ter- tions. The recycling process allows for the copper content of the bium (Tb), samarium (Sm) and praseodymium (Pr) are crucial for copper(II) salts to be recovered and reinserted into the value chain. high performance magnets which, in turn, are critical for technolo- Interestingly, a significant amount of global copper supply is ob- gies such as wind power generators, electric powertrains, mo- tained with the aid of microorganisms via bioleaching.19-21 tors/generators, and in everyday consumer products like computers, The recycling process was developed considering the application mobile phones, amongst many others. Nd, Tb and Dy are among of green chemistry principles (acid-free dissolution of magnets the elements which the U.S. Department of Energy9 and the Euro- through redox-dissolution with copper(II) salts). It was also devel- pean Commission10 classified as being at risk of supply disruption. oped in view of potential commercial adoption such that the recov- Cobalt (Co) is also critical for energy-related applications, includ- ered REEs and other recycling co-products are suitable for reinser- ing permanent magnets and lithium-ion batteries. Recovering these tion into the supply chain. Moreover, to maximize profit in the ap- critical elements from waste REEs-containing materials can be eco- plication, the process development included recovery and reinser- nomically and technically viable approach towards helping address tion of some of the chemicals back into the recycling process. the supply risk challenges.11 When properly done, recycling can help conserve natural resources, save energy and result in less toxic wastes being dumped in landfills. One approach to accomplishing EXPERIMENTAL SECTION 12 this is to follow the principles of Green Chemistry. Material. The recycling feedstock materials used for this work in- Recycling of REE-containing materials can be done either by di- clude: (a) Nd-Fe-B grinding swarfs from U.S. magnet plants (Sup- rectly reusing the materials or via chemical recovery of the rare plementary information, Fig. S1), (b) hydrogen decrepitated Nd- earth elements. The chemical recovery methods for REEs generally Fe-B magnets manually harvested from end-of-life HDDs, (c) Sm- involve either pyrometallurgical or hydrometallurgical ap- Co grinding swarfs also from U.S. magnet plants, (d) Nd-Fe-B con- proaches.13-18 In the pyrometallurgical approach, the REEs recov- tained in two types of e-waste materials: shredded HDDs (supplied ery rates are reduced by slag formation, due to the high affinity of by Dr. Tim McIntyre, Oak Ridge National Laboratory) and crushed the REEs with oxygen. Pyrometallurgical approaches also generate electric motors (from www.amazon.com, “Great Planes ElectriFly large amounts of solid wastes and can be quite energy-consuming, RimFire .10 35-30-1250” Outrunner Brushless Motor”). Swarfs are although they can have the advantage of recovering the REEs in the metal powders derived from post-manufacturing processing includ- form of metals, instead of oxides. Hydrometallurgical routes enable 1 ing grinding, cutting, etc. which are typically oxidized and contam- of 0.88 kg of copper(II) nitrate hemi(pentahydrate) in 3.5 L of water inated by grinding media and lubricants. Decrepitation was per- and stirred at 300 rpm for about 5 hours. The reaction proceeded formed in a stainless steel vessel at 4 bar of hydrogen pressure and exothermically with no heat applied. Afterward, residue found to room temperature. For making magnets using recovered REE ox- be a mixture of Cu2O/CuO/Fe2O3/Fe3O4 (see the PXRD pattern ide, the oxides were first reduced to Nd/Pr metal ingot at the Ames in Fig. S3 and EDX mapping in Fig. S10), was filtered off. The Laboratory Materials Preparation Center (MPC) using the estab- copper-containing precipitate can be used as a source for copper. 22, 23 lished Ames process. Part of the ingot was alloyed with 99.9% For recovering REEs from the filtrate, a two-step approach was ap- pure Fe and Cu (from MPC) and 99.5% pure boron (from AlfaAe- plied aiming to exclude the presence of iron in the final REE oxides sar) via arc melting (in Ar atmosphere) to produce a magnet feed- (Figs. S4, S13). In the first step, the filtrate was stirred together with stock with composition (Nd-Pr)2.3Fe14B + 0.5 wt.% Cu. The arc- an aqueous ammonia solution (28%) at 60 °C for 2 hrs which led melted material was used to prepare melt-spun ribbons by inductive to the precipitation of Fe(III) and RE(III) hydroxides. Then, a small melting in fused silica crucibles and ejection of the melt onto a sin- excess of solid crystalline H2C2O4·2H2O was added, followed by gle copper wheel at 30 m/s surface velocity through a 0.8 mm ori- heating to 80 °C and stirring until insoluble REE oxalate precipi- fice. Melt spinning was performed in 1/3 atmosphere of high purity tated and highly soluble double salt of iron-ammonium oxalate He gas. The amorphous phase content of the melt-spun ribbons was formed. The REE oxalates were separated by filtration and washing crystallized by heat treatments at 700 °C for 15 mins in ultra-high in hot water, and subsequently calcined in air at 800 °C to obtain purity argon. Enabled by the low melting eutectic between REEs RE2O3 (m = 146.7 g; >98% yield). For the starting RE-Fe-B mag- and the added Cu, a hot-pressed magnet was made using the magnet net grinding swarfs, the following XRF elemental contents were ribbon. obtained (wt.%): Fe = 73.71; RE = 25.51; Ni = 0.57; Cu = 0.21. All chemicals were purchased from commercial sources (Sigma- The elemental contents of the obtained RE2O3 determined by XRF Aldrich, ACS reagent grade) and were used without further purifi- were (wt.%): Fe = 0.15; RE = 99.57; Cu = 0.28. Fig. 1 illustrates cations. For all the recycling feedstock materials, leaching of criti- the new process in comparison with the current state-of-art. cal elements (REEs and cobalt) were executed at room temperature while the solution was being stirred at 300 rpm for ≤ 6 hrs, followed by standard operations such as simple filtrations, selective precipitation and calcination in air. All inputs and outputs for individual steps (Nd-Fe-B and Sm-Co magnets) are presented in the material balance table (see SI, Tab. S1). Phase analysis was studied by powder X-ray diffraction (PXRD). PXRD patterns were collected at room temperature using an STOE Stadi P powder diffractometer equipped with an image plate and a Cu Kα1 radiation source (λ = 1.5406 Å) (Figs. 5, S2-S8). Energy dispersive X-ray spectrometer (EDS) detector contained in a scan- ning electron microscope (FEI Teneo LoVac) was used to determine the elemental distribution in the starting material for the decrepitated Nd-Fe-B magnet (Fig. S9) and also for the iron-copper residue obtained after leaching (Fig. S10). ESI-MS (mass spectrometry) was recorded on Agilent QTOF 6540 mass spectrometer (Figs. S11-S12). X-ray fluorescence (XRF) spectrometry (Bruker M4 TORNADO Micro-XRF spectrometer operated at 50 kV and 300 μA with Rh target) was used in this work to determine the composition of FIGURE 1: Flow charts for the REE recovery from Nd-Fe-B swarfs swarfs and metal concentrations in products of recycling (Fig. S13- or e-waste via the proposed acid-free leaching process (shaded in S15). Inductively coupled plasma-mass spectrometry (ICP-MS, teal), compared to general acid-based leaching process (current Thermo Fisher Scientific X Series 2) was also used to analyze con- state-of-the-art, shaded in red). centrations of some selected metals in reference to XRF data. For Dissolution of decrepitated magnets from HDDs the ICP-MS, powdered samples were dissolved in 5 mL of aqua regia to dissolve all metal content followed by dilution with 2% The steps in the dissolution of the hydrogen decrepitated magnets nitric acid. and recovery of RE2O3 were similar to those described above for 57 Nd-Fe-B- grinding swarfs. 1.0 g of decrepitated Nd-Fe-B (Fig. S5) Fe Mössbauer spectroscopy measurement was performed using a was reacted at RT with 3.75 g of copper(II) nitrate (dissolved in 40 MS4 spectrometer operating in the constant acceleration mode in ml of H2O) in about 30 min. A mixture of copper(I) oxide and me- transmission geometry. The measurement was performed at room 57 tallic copper powder was recovered as insoluble precipitate and temperature. A 50 mCi Co in Rh held at room temperature was separated by filtration (>98.5% yield and >99.9 wt.% purity, see used as source. All centroid shifts, δ, are given with respect to me- Fig. S6). Similarly, precipitated REE oxalates were filtered off and α− tallic iron at room temperature. The spectrum was least-square calcined to obtain RE2O3. fitted to extract hyperfine parameters including the centroid shifts Dissolution of magnets from e-wastes (δ), quadrupole splitting/shift (∆EQ/ε), magnetic hyperfine field (Bhf), Lorentzian linewidth (Γ), and intensity (I). For Nd-Fe-B contained in e-wastes (i.e. shredded HDDs and electric motors), the magnet dissolution and REE recovery steps were Magnetic properties were obtained at room temperature using a similar to that of Nd-Fe-B- swarfs described above except that cop- Quantum Design vibrating sample magnetometer (VSM) with per(II) chloride dehydrate was used: 120 g in the case of as-shred- µ magnetic field applied up to 0H = 3 T. ded HDD and 42 g in case of electric motor: corresponding to twice Dissolution of Nd-Fe-B magnet swarfs the amount of copper(II) salt needed. About 216.4 g chucks of About 0.5 kg of (Nd,Pr)-Fe-B swarfs, allowed to undergo oxidation shredded HDDs in as-shredded states were used (estimated magnet in air, i.e. without heating (Fig. S1-S2, SI), were added to a solution content is ~23 g or 8.8 wt.%). For the electric motors, 70.1 g was 2 used (estimated magnet content is ~8.19 g or 11.7 wt.%). The elec- ° 2RE2(C2O4)3·10H2O + 3O2 2RE2O3 + 20H2O + tric motor was crushed to accelerate the process of contacting the 800 𝐶𝐶 12CO 2 (5) contained magnets with the solution. The dissolution of both e- �⎯⎯⎯� wastes types was performed at room temperature for 24 hrs. The Fig. 2 shows the XRF spectra (inset text) of the recovered RE2O3 elemental constituents of the magnet contained in the starting ma- which confirms >99.9% purity of the REE contents. Multiple spec- terial for the as-shredded HDD was determined by XRF (in wt.%) tra (shown in different colors in Fig. 2) were obtained to validate as: Fe = 69.06; RE = 22.55; Ni = 6.15; Cu = 2.24. The REEs recov- that the obtained purity was uniform. ery efficiency was ~73 wt.% (4.43 g of RE2O3). The purity of the final RE2O3 product, also determined by XRF in wt.%, was >99.9 % and includes: Fe = Not detected (N/D); RE = 99.97; Ni = 0.02; Cu = 0.01 (Fig. S14). For the electric motor, the elemental content of the starting magnet determined by XRF (wt.%) include Fe = 69.06; RE = 22.55; Ni = 6.15; Cu = 2.24. The REEs recovery efficiency was ~65 wt.% (0.90 g of RE2O3). The purity of the final RE2O3 product was >99.9% with the following elemental contents (wt.%): Fe = N/D; RE = 99.97; Ni = 0.02; Cu = 0.01. Dissolution of Sm-Co Sm-Co magnet alloy swarfs (100 g) were stirred in an aqueous solution of copper(II) sulfate (346 g in 600 ml water) at 260 rpm without any external heating for about 3 hours. This resulted in the oxidative dissolution of the swarfs and precipitation of a mixture of FIGURE 2: X-ray fluorescence spectra (4-8 keV range) for the final Cu2O and metallic Cu powder (95.1 g, ≥98% yield) as a brown-red RE2O3 product. Inset: selected XRF data for copper, nickel, iron precipitate. The resulting mixture of copper(I) oxide and metallic and rare earth metals (Nd, Pr, Dy). N/D means “not detected”. copper powder (>99.7 wt.% based on copper; see Fig. S7) was filtered off and can a source of copper as will be later described. To We found that allowing the magnet swarfs to oxidize under ambient separate samarium, anhydrous sodium sulfate (100 g) was added to conditions resulted in the use of less copper(II) salt than stoichio- the filtrate at 100 °C, leading to the precipitation of metrically expected. This is likely because any oxidized component NaSm(SO4)2·H2O, which was subsequently converted to Sm2O3 of the swarfs will not participate in dissolution via redox process (Fig. S8). The elemental composition of the starting Sm-Co swarfs, with the Cu ions. Although the use of less Cu(II) salt indicates po- as determined by XRF (wt.%) are: Sm = 41.67; Co = 56.59; Fe = tential to improve the economic efficiency of the process, it is im- 1.74. The recovery efficiency of the Sm2O3 was about 97% (46.9 portant to understand how the yield for the process reached the ob- g) and the purity determined via both XRF and ICP-MS was >99.9 served ~98%. This is because the yield would be less if significant wt.% (includes: Sm = 99.56 wt.%; Co = N/D; Fe = 0.43 wt.%, Cu amount of RE2O3 phases remained undissolved and were filtered = 0.01 wt.%). Cobalt was afterwards recovered as cobalt(II) phos- off with the residues. Studies have reported that at ambient condi- phate anhydrous at >99% yield (116.2 g) and >98 wt.% purity (in- tions, oxidation of Nd-Fe-B appears to be mostly at the grain cludes Co = 98.65 wt.%; Sm = 0.34 wt.%; Cu = 1.01 wt.%). Upon boundaries.25,26 The product of such oxidation include REE and cooling to the ambient temperature, the filtrate was mixed with 400 iron oxides.25,26 Although the grain boundary phase is typically mL of ethanol resulting in the crystallization of ~252 g of sodium REE-rich, the volume ratio of the REE in the grain boundary is sig- sulfate decahydrate, Na2SO4·10H2O (equivalent to ≥ 111.2 g of nificantly smaller, compared to the matrix Nd2Fe14B phase. More- anhydrous salt). Na2SO4·10H2O can be recovered and reused. over, oxidation of the grain boundary phase gradually progresses 26 until it is saturated with oxygen. Hence, different swarfs will be at different states prior to recycling. It is also possible that some of RESULTS AND DISCUSSION the matrix phase may be oxidized and form phases like Fe3B, iron Application to Nd-Fe-B Magnets. Comprehensive oxidative dis- oxides and RE2O3 phases. Studies are hard to find on the oxidation solution was accomplished for (neodymium, praseodymium)-iron- process of magnet swarfs. We propose that at ambient conditions, boron (RE-Fe-B) magnets according to Eqs. (1, 2) (see SI, Figs. oxidation at room temperature may proceed as in equation 6. This S11-S12 for additional details).24 is because powder X-ray diffraction (PXRD, Fig. S2) analysis of

2+ the swarfs reveals that, in addition to (Nd,Pr)2Fe14B, our starting 2RE2Fe 14B + 34Cu + O2 material also contained Fe3B, (Nd,Pr)2O3, Fe2O3 and Fe3O4 (last 21 𝑅𝑅𝑅𝑅 3+ 2+ 0 two in ~1:1 ratio). Others also reported formation of these and other 3 3 2 2 → 4RE + 28Fe + Cu2 (BO��) ↓ + 15Cu O↓ + Cu ↓ (1) 27 phases upon oxidizing Nd2Fe14B at 400 °C. 2+ 3+ 12Fe + 6H2O + 3O2 → 4Fe(OH)3↓ + 8Fe (2) RE2Fe14B + O2 → RE2O3 + Fe3B + Fe3O4 + Fe2O3 (6) Eqs. 3 – 4 describe the chemical reactions for the recovery of 46 11 11 As the solution5 is stirred, the redox reaction5 will proceed5 with any REE from the solution which includes, precipitation of rare exposed Nd-Fe-B. The iron oxides are chemically inactive during earth hydroxide (eq. 3), rare earth oxalate (eq. 4) and calcina- the redox dissolution process with copper(II) cations. However, ac- tion to obtain rare earth oxide (eq. 5) . cording to the calculated potential-pH diagram of the Fe-H2O and 28 Nd-H2O systems, oxidized part of rare earth metals could be dis- ° 2+ 3+ 3+ 4Fe + Fe + RE + 14NH4OH + 2H2O + O2 solved due to possible hydrolysis of either or both of iron(III) and + 60 𝐶𝐶 Cu(II) salts and the formation of acidic conditions (pH≥4) (Eqs. 7, →5Fe(OH) 3↓ + RE(OH)3↓ + 14NH4 (3) �⎯⎯� 8) where X- = an anion, see Fig. S16). ° 10Fe(OH) 3↓ + 2RE(OH)3↓ + 30NH4OH + 33H 2C2O4 RE2O3 + 3CuX2 + 3H2O → 2REX3 + 3Cu(OH)2↓ (7) 80 𝐶𝐶 →10(NH4) 3[Fe(C2O4) 3] + RE2(C2O4)3↓·10H2O + 56H2O RE2O3 + 2FeX 3 + 3H2O → 2REX3 + 2Fe(OH)3↓ (8) (4) �⎯⎯� From eqn. 7 and 8, one would expect that either Cu(OH)2 or Fe(OH)3 would be part of the filtered residue. However, as stated 3 earlier, the residue was found to be a mixture of of the sample is magnetite. One can therefore conclude that in ad- Cu2O/CuO/Fe3O4/Fe2O3. This would suggest that dissolution pro- dition to oxides of Cu, the residue contained magnetite, Fe3O4 and gressed mostly via redox process. If hydrolysis occurred, it would maghemite γ-Fe2O3. The ratio of magnetite and maghemite is have been at an insignificant level. To further verify that none of nearly 1:1 (48% magnetite and 52% magnetite) in line with Eq. 6 the products of hydrolysis were present in the residues, they were and XRD spectrum (Fig. S3). 57 29 studied via room temperature Fe Mössbauer spectroscopy. The If small amounts of oxides were present in the swarfs before disso- Mössbauer spectrum, shown in Fig. 3, consists of two six-line pat- lution, the amount of Cu(II) salt used would be close to the values terns and a two-line pattern. The patterns can be fitted with two expected based on Eq. 1. To validate this, REEs contained in hy- sextets and two doublets. The extracted Mössbauer parameters for drogen decrepitated Nd-Fe-B were recycled. Exposure of Nd-Fe-B the sample are summarized in Table 1. to hydrogen pressure leads to penetration of hydrogen into the grain boundaries, reaction of hydrogen with the grain boudary phase, ex- pansion of grain boundaries and embrittlement of the material. De- crepitation in hydrogen atmosphere limits oxidation of Nd-Fe-B 31-33 and can form phases like Nd2Fe14BHx and Nd1+EFe4B4. The surface morphology and EDS mapping of the decrepitated (Nd,Pr)- Fe-B magnet before leaching are presented in the SEM micrograph in Figs. 4 and S9. The particle sizes of hydrogen decrepitated magnet powders range from 2 to 10 µm. It can be seen that the powders consist of RE (Nd, Pr)- and Fe-rich compositions.

FIGURE 3: Measured room-temperature Mössbauer spectrum (dots) for the iron(III) oxides-containing sample obtained after leaching of Nd-Fe-B magnet swarfs with copper(II) nitrate. Indi- vidual components obtained through spectral deconvolution in the fitting process indicated by solid lines. Table 1: Summary of refined Mössbauer parameters for the iron(III) containing residue obtained after leaching with copper(II) nitrate.

Parameters* Q1 Q2 Q3 Q4 δ (mm/s) 0.368 0.360 0.634 0.291 FIGURE 4: SEM micrograph of the surface of the decrepitated Nd- Bhf (T) ------45.4 48.7 Fe-B magnet (obtained from the disassembling of HDD) and the ∆EQ/ε (mm/s) 0.576 0.854 0.003 -0.010 corresponding SEM-EDS map (inset) of Nd (purple), Pr (blue), Fe Γ (mm/s) 0.33 0.47 0.54 0.30 (red) and O (green). I (%) 21 22 32 25 Figs. 5 and S5 show that the compositions include: Nd2Fe14BH2.7, Fe, Fe2B and ~1% of Nd2O3. Unlike in swarfs allowed to oxidized * centroid shift, δ, magnetic hyperfine field, Bhf, quadrupole split- in air, leaching of REEs from decrepitated Nd-Fe-B magnet using ting/shift, ∆EQ/ε, Lorentzian linewidth, Γ, and intensity, I. Esti- copper(II) nitrate is in agreement with the dissolution reaction in mated errors are in I ±3%, in δ and ε ±0.005 mm/s, in Γ ±0.02 Equation 1 with the ratio of Nd2Fe14B and copper(II) salt used be- mm/s, and in Bhf ±0.5 T. ing ≈ 1:17. The δ values for the doublets, components Q1 and Q2, are 0.368 mm/s and 0.360 mm/s, respectively. The ∆EQ values for these doublets are 0.576 mm/s and 0.854 mm/s, respectively. These are characteristic values for superparamagnetic Fe3+ in high-spin state.30 The intensity contributed by the doublets are 21% and 22%. The magnetic component Q3 shows magnetic hyperfine splitting with δ value of 0.634 mm/s and Bhf of 45.4 T. These are characteristic values for Fe+2.5 in octahedral environment (B-sites) in magnetite, Fe3O4. This component has an intensity of 32%. The magnetic component, Q4, has δ value of 0.291 mm/s and Bhf of 48.7 T, which are characteristic values for high-spin Fe3+. Most probably, this component consists of signal from the A-sites of the magnetite, as well as signal from maghemite, γ-Fe2O3. The intensity for this component is 25%. If we assume the same recoil-free fractions for all sites, 16% of the last magnetic component, i.e., half of the intensities from the B-sites, are the signals from the A-sites of the magnetite part and the remaining 9% are the intensities from the maghemite part. Adding the intensities for the B-sites and the A- sites in the magnetite part, i.e., 32% and 16%, will infer that 48% 4 FIGURE 5: Comparison of measured X-ray powder diffraction pat- FIGURE 6: Image of as-shredded HDDs before recycling (top). In- terns relative to expected peak positions in database for the decrep- set: Recovered 16.6 g of metallic copper from the corresponding itated Nd-Fe-B magnet. HDDs recycling. Bottom: Small motor containing REE magnets before (left) and after (right) partial crushing with a hydraulic press. To valorize copper from the mixed Cu2O/CuO/Fe3O4/Fe2O3 residue, it can be recovered as metallic copper via extractive pyromet- The lower REE recovery efficiency for e-wastes recycling (73 allurgy34,35or via hydrometallurgical process.36 To demonstrate a wt.% for the as-shredded HDDs and ~65 wt.% for the electric mo- hydrometallurgical approach, Cu2O was oxidized into CuO as in tor) is likely a consequence of limited access of the solution to the Eq. 9. Afterward, the CuO/Fe3O4/Fe2O3 mixture was treated with magnet surfaces. Recovery efficiency can vary depending on how KHSO4 (Eq. 10) and metallic copper was obtained by the addition much the copper salt solution makes contact with the magnet. In of iron in the metallic form (Eq. 11). Recovery efficiency was the case of HDDs, finer shredding will promote access of the leach- ~85%. ing agents to the magnets. However, it is also likely to result in the dissolution of other Fe-containing parts in the shred mix, hence re- 2Cu2O + O2 → 4CuO (9) quiring more leaching reagents. Considering a balance between {Fe2O3↓ + Fe3O4↓ + CuO} + 2KHSO4 → these competing outcomes, an optimum shredding is likely to result → {Fe2O3↓ + Fe3O4↓} + CuSO4 + K2SO4 + H2O in recovery efficiency between 80 – 85 wt.%. (10) Fig. S17 shows that the process resulted in no observable reactivity 0 0 CuSO4 + Fe → Cu ↓ + FeSO4 (11) with other components of the HDDs and motors. This result is im- Application to Magnets in e-Wastes. Electric motors and con- portant because it shows that the process minimizes the amount of sumer electronics are forecasted to continue to dominate the use of impurities leached from the heterogeneous mixture of electronic REE permanent magnets by 2020 (>50% for Nd-Fe-B and samar- waste materials. It also results in high purity and high yield of REEs ium-cobalt magnet applications).8 Having established the process precipitates for calcination step (Fig. S14). Most of the copper by- for dissolving RE-Fe-B magnets, we adapted the process to allow products were obtained in metallic form (inset of Fig. 6) and can be for selective leaching of REEs from waste electrical and electronic easily separated and cleaned with water for use in producing elec- equipment. To maximize the economic potential for “urban min- trolytic copper. It is also possible that an additional processing step ing” of REEs, this novel process was applied to HDDs in as-shred- may be required. ded conditions and also crushed electric motors (Fig. 6), without Reusing Recovered REE for Nd-Fe-B Magnets Feedstock. To the steps of pre-separation, pre-oxidation or thermal demagnetiza- 4,37,38 demonstrate the suitability for reinsertion into the magnet supply tion of the magnet contents which are typically undertaken. chain, the recycled REEs were used to make magnets as previously Thus, the number of steps in the recycling process was reduced; described. Fig. 7 shows the room temperature M–H curves for two which can enhance the economic potential of industrially deploying differently processed forms of the RE2.3Fe14B+0.5% Cu alloy: rap- the technology. idly solidified ribbons (sample #1) and hot-pressed powder of the ribbons (sample #2) (inset of Fig. 7). Both the ribbon and hot- pressed samples are isotropic magnets. Respectively, the magnetic properties at 300 K are coercivity of 15.0 and 14.2 kOe, remanence of 8.0 and 7.5 kG and energy product of 13.6 and 12.0 MGOe. These values are typical of isotropic ribbons and hot-pressed magnets,39,40 and are comparable to those of commercially available magnets.41 Obtaining comparable magnetic properties as commercial products is significant because it demonstrates that the recovered REEs have no impurities capable of hindering the economic values of recycling.

FIGURE 7: Magnetization versus magnetic field at 300 K of RE2.3Fe14B+0.5%Cu for melt-spun ribbons (sample #1) and hot- pressed magnet (sample #2). Inset: an image of the prepared hot

pressed magnet using recycled rare earth material. 5 Application to Sm-Co Magnets. The battery industry currently Eq. 16 – 18 describe the chemical reactions for the recovery of uses more than 40% of the global production of cobalt; an indis- Sm2O3 from the solution which includes, precipitation of 31 pensable metal both for lithium-ion cells and permanent magnets. NaSm(SO4)2·H2O (eq. 16), Sm2(C2O4)3·10H2O (eq. 17) and cal- Cobalt, like the REEs, has been identified as a critical material. cination to obtain Sm2O3 (eq. 18). Eq. 19 describes the recovery of Successful application of the recycling process to Sm-Co magnets Co in the form of Co3(PO4)2·6H2O. indicates the potential for waste Sm-Co alloys to be sources of sig- + 3+ 2- nificant amounts of cobalt for secondary Co supplies. Sm-Co mag- Na + Sm + 2SO4 + H2O NaSm(SO4) 2·H 2O↓ nets excel the high-performance Nd-Fe-B magnets when high tem- (16) 100 ℃ perature (>150 °C) applications are required. In addition, Sm-Co �⎯⎯⎯� magnets have superior oxidation resistance.42 Increasing amounts 2NaSm(SO4) 2·H 2O↓ + 6NH 4OH + 3H2C2O4 + 2H 2O → of Sm-Co have been projected to be used in applications by → Sm2(C2O4)3·10H2O↓ + Na2SO 4 + 3(NH4)2SO 4 (17) 202043,44 and most of those would be available for future recycling. Most of the known recycling processes for Sm-Co magnets use an 2Sm2(C2O4)3·10H2O + 3O 2 800 ℃ excess of hazardous sulfuric acid for leaching (a) and shifting of 2Sm2O3 + 20H2O↑ + 12CO2↑ (18) �⎯⎯⎯� the solubility equilibrium (b) to the target insoluble (double salt) 3CoSO4→ + 2Na3PO4·10H2O + 16H 2O → product.28 Co 3(PO 4)2·6H 2O↓ + 3Na2SO4·10H 2O (19) We adapted the acid-free leaching process to Sm-Co swarfs obtained from a U.S. magnet processing company. Fig. 8 shows the As stated in the experimental section, Na2SO4·10H2O is reusable process for recovering Sm and Co in which an aqueous solution of after crystallization which is significant because it minimizes ma- copper(II) sulfate was used for the dissolution of Sm-Co (Eq. 12). terials costs. Table 2 shows the comparison of the acid-free leaching process (current work) with acid-based leaching processes 2+ 2SmCo5 + 13Cu + O 2 found in the literature. 𝑅𝑅𝑅𝑅 3+ 2+ 0 2Sm + 10Co + 2Cu2O↓ + 9Cu ↓ (12) Table 2. Comparison between the acid-free leaching approach in 𝑅𝑅𝑅𝑅 �� the current work and some acid-based leaching methods found in �� the literature.46-48 Leaching procedures # Items Current Method Method Work A46,47 B48 pre-demag- ball-mill- 1 Pre-treatment None netization, ing, roast- corrosion ing

2 Leaching acid None HCl H2SO4 2 hrs to 3 Leaching time 4 hrs >24 hrs 1 week 4 Reagents used 5 5 1 Selective in e- 5 Yes No No waste recycling Recovering of 6 Yes No Yes* by-products FIGURE 8: Proposed recycling process for recovery of samarium Toxic and cobalt from Sm-Co swarfs.The blue dotted line represents a 7 No Yes Yes closed-loop process for the recovery of sodium sulfate. waste/acid Boron Copper is a comparatively noble metal. When Cu(II) salt solutions 8 >99% 30% No are brought into contact with metals or alloys of less noble metals, Separation typically these get dissolved, hence, oxidized and the copper pre- 9 REE Recovery >98% 97% 98% cipitates as a metal as it is reduced. A classic example which often is used in introductory chemistry laboratories is the plating of iron 10 REE Purity >99.9% Not reported >98% 45 nails or pennies by Cu(II) solutions. Half-reaction equations (Eqs. * = mixture of SO2/SO3 gases 13, 14) for this dissolution process and equation of copper oxida- 48 tion (Eq. 15) are the following: Although leaching with H2SO4 required the least amount of reagents, the reported process did not account for the boron content of 3+ 2+ - 2SmCo 5 → 2Sm + 10Co + 26e (13) Nd-Fe-B magnets. The acid-free leaching process also resulted in higher purity of recovered REEs with comparable recovery effi- 13Cu2+ + 26e- → 13Cu0 (14) ciency as the H2SO4. As can be seen in the table, the acid-free 0 4Cu + O2 → 2Cu 2O (15) leaching process is the only process that is applicable without pre- treatment of the waste, hence the only one suitable for selective Algebraic addition of the half-reactions in Eqs. 13, 14 and 15 will leaching of REEs from e-wastes. yield Eq. 12. The resulting brown-red precipitate mixture of copper metallic and copper(I) oxide powder (>99.7% based on copper; see Wastewater Treatment. Wastewater treatment has been per- Fig. S7) was filtered off and can be used in an application or valor- formed as part of the recycling process. For the process of valoriz- ized as previously explained. ing copper, wastewater treatment was performed with freshly prepared calcium hydroxide (Eq. 20) to precipitate iron hydroxides 6 and CaSO4, leaving behind aqueous solution of potassium sulfate The manuscript was written through the contributions of all au- (from Eq. 9) as potential fertilizer source.49 thors. All authors have given approval to the final version of the manuscript. FeSO4 + Ca(OH)2 → Fe(OH)2↓ + CaSO4↓ (20)

The iron(III) sulfate, which will be present in wastewater (see the Notes oxidation process in Eq. 2), has been removed in a similar manner The authors declare no competing financial interests. (Eq. 21):

Fe2(SO4)3 + 3Ca(OH)2 → 2Fe(OH)3↓ + 3CaSO4↓ ACKNOWLEDGMENT (21) This work was supported by the Critical Materials Institute, an En- Wastewater treatment for Nd-Fe-B recycling feedstock was per- ergy Innovation Hub funded by the U.S. Department of Energy, formed with a mixture of calcium chloride and freshly prepared Office of Energy Efficiency and Renewable Energy, Advanced calcium hydroxide (from calcium oxide) according to Eq. 22. After Manufacturing Office. The authors gratefully acknowledge Dr. decantation, the resulting solution of ammonium chloride and am- Oleksandr Dolotko, Dr. Ihor Hlova and Dr. Rakesh Chaudhary monium nitrate can find applications as sources of nitrogen for ag- (Ames Laboratory, Ames, IA, USA) for their help at the different ricultural purposes.49 stages of research (powder X-ray diffraction and X-ray fluorescence analysis). We thank Prof. G. J. Miller (Iowa State Univer- 2(NH4)3[Fe(C2O 4)3] + 3CaCl 2 + 3CaO + 3H2O → sity/Ames Laboratory, Ames, IA, USA) for acces to XRD facili- → 2Fe(OH)3↓ + 6CaC2O4↓ + 6NH4Cl (22) ties, Dr. S. Kamali, Prof. J. Johnson, Prof. C. Johnson (University Performing this wastewater treatment will help to ensure that of Tennessee Space Institute, Tullahoma, TN, USA) for providing no toxic waste is generated as part of the process. Mössbauer spectroscopy services and Dr. Timothy McIntyre (Oak Ridge National Laboratory/CMI, Oak Ridge, TN, USA) for provid- CONCLUSION ing shredded hard disk drives materials. Authors are thankful to Electron Energy Corporation (Lancaster, PA, USA) for providing A process for selective leaching of rare earth elements with Cu(II) the grinding swarfs used for this work. salts has been developed and successfully applied for recycling Nd- Fe-B and Sm-Co magnets obtained as waste swarfs from magnet REFERENCES processing/manufacturing companies. The process eliminates the addition of any mineral acids and was also applied for the recycling (1) Tan, Q.; Li, J.; Zeng, X. Rare earth elements recovery from waste of end-of-life electronic wastes. It excludes some energy-consum- fluorescent lamps: a review. Crit. Rev. Environ. Sci. Technol. 2015, ing steps, e.g. pre-separation, demagnetization and pre-oxidation 45 (7), 749−776. (2) Dutta, T.; Kim, K.-H.; Uchimiya, M.; Kwon, E. E.; Jeon, B.-H.; and did not result in the production of toxic wastes. Table 2 com- Deep, A.; Yun, S.-T. Global demand for rare earth resources and pares the new acid-free leaching process for Nd-Fe-B with some strategies for green mining. Environ. Res. 2016, 150, 182−190. other leaching approaches described in the literature. The proposed (3) Depuydt, D.; Van den Bossche, A.; Dehaen, W.; Binnemans, K. closed-looped recycling for Sm-Co swarfs via the acid-free leach- Metal extraction with a short-chain imidazolium nitrate ionic liq- ing process is efficient for both metals (i.e. Sm and Co) recovery uid. Chem. Commun. 2017, 53, 5271-5274. and generation of environment-friendly Glauber’s salt as a part of (4) Yang, Y.; Walton, A.; Sheridan, R.; Gü th, K.; Gauß, R.; Gutfleisch, the waste stream. O.; Buchert, M.; Steenari, B.-M.; Van Gerven, T.; Jones, P. T.; Binnemans, K. REE Recovery from End-of-Life NdFeB Perma- nent Magnet Scrap: A Critical Review. J. Sustain. Metallurgy ASSOCIATED CONTENT 2017, 3, 122-149. Supporting Information (5) Prodius, D.; Mudring, A.-V. Rare earth metal-containing ionic liq- uids. Coord. Chem. Rev. 2018, 363, 1–16. The Supporting Information is available free of charge on the ACS (6) Venkatesan, P.; Hoogerstraete, T. V.; Binnemans, K.; Sun, Z.; Si- Publications website. XRF spectra of materials (starting, recov- etsma, J.; Yang, Y. Selective Extraction of Rare-Earth Elements ered), ESI-MS spectrum of boron-containing sample, calculated from NdFeB Magnets by a Room-Temperature Electrolysis Pre- and measured X-ray powder diffraction patterns for the starting ma- treatment Step. ACS Sustainable Chem. Eng. 2018, 6, 9375–9382. (7) Venkatesan, P.; Vander Hoogerstraete, T.; Hennebel, T.; Binne- terials and recovered products, additional explanations and other mans, K.; Sietsma, J.; Yang, Y. Selective electrochemical extrac- supporting data (PDF) tion of REEs from NdFeB magnet waste at room temperature. Green Chem. 2018, 20 (5), 1065−1073. AUTHOR INFORMATION (8) Kumar, A. Magnets and Magnet Materials: Global Markets, Ban- galore, 2017. Corresponding Authors (9) Bauer, D. D, D.; Li, J.; McKittrick, M.; Sandalow, D.; Talleen, P. † Critical Materials Strategy; US Department of Energy: Washing- [email protected]; [email protected] ton, DC, 2011. Present Address (10) European Commission. Communication from the Commission to European Parliament, the Council, the European Economic and †Department of Materials and Environmental Chemistry, Stock- Social Committee and the Committee of the Regions; European holm University, Svante Arrhenius väg 16C, 10961 Stockholm, Commission: Brussels, Belgium, 2017. Sweden. (11) Binnemans, K.; Jones, P. T.; Blanpain, B.; Van Gerven, T.; Yang, Y.; Walton, A.; Buchert, M. Recycling of rare earths: a critical re- ORCID view. J. Cleaner Prod. 2013, 51, 1−22. (12) Anastas, P; Warner, J. Green Chemistry: Theory and Practice, Ox- Denis Prodius: 0000-0002-9967-6021 ford University Press, New York, 1998, pp. 1-135. Kinjal Gandha: 0000-0001-6363-6872 (13) Jha, M. K.; Kumari, A.; Panda, R.; Rajesh Kumar, J.; Yoo, K.; Lee, Anja-Verena Mudring: 0000-0002-2800-1684 J. Y. Review on hydrometallurgical recovery of rare earth metals. Ikenna C. Nlebedim: 0000-0002-0045-4585 Hydrometallurgy 2016, 165 (Part 1), 2−26. (14) Yoon, H.-S.; Kim, C.-J.; Chung, K.-W.; Kim, S.-D.; Lee, J.-Y.; Author Contributions Kumar, J. R. Solvent Extraction, Separation and Recovery of Dys- prosium (Dy) and Neodymium (Nd) from Aqueous Solutions: 7 Waste Recycling Strategies for Permanent Magnet Processing. Hy- (33) Skulj, I.; Douvalis, A. P.; Harris, I. R. Characterisation of oxidation drometallurgy 2016, 165, 27−43. products of modified Nd–Fe–B type magnets. J. Alloys Compd. (15) Orefice, M.; Binnemans, K.; Hoogerstraete, T. V. Metal coordina- 2006, 407, 304–313. tion in the high-temperature leaching of roasted NdFeB magnets (34) Guo, Z.; Zhu, D.; Pan, J.; Wu, T.; Zhang, F. Improving beneficia- with the ionic liquid betainium bis(trifluoromethylsulfonyl)imide. tion of copper and iron from copper slag by modifying the molten RSC Adv. 2018, 8, 9299-9310. copper slag. Metals. 2016, 6, 86-103. (16) Nockemann, P.; Ritesh, R. Separation of rare earth metals. (35) Donaldson, J. W.; Sharma, S. N.; Themelis, N. J. Pyrometallurgical WO2018109483A1, June 21, 2018. recovery of copper from slag material. US4,032,327, June 28, (17) Tunsu, C. Hydrometallurgy in the recycling of spent NdFeB per- 1977. manent magnets. Waste Electrical and Electronic Equipment Re- (36) Stanley, R. W.; Monette, S.; Kerfoot, D. G. E. Hydrometallurgical cycling: Aqueous Recovery Methods. 2018, 8, pp. 175-211. treatment of copper-bearing hematite residue. US4,338,168 July 6, (18) Padhan, E.; Nayak, A. K.; Sarangi, K. Recovery of Neodymium 1982. and Dysprosium from NdFeB Magnet Swarf. Hydrometallurgy (37) Bandara, H. M. D.; Field, K. D.; Emmert, M. H. Rare earth recov- 2017, 174, 210−215. ery from end-of-life motors employing green chemistry design (19) Schippers, A.; Hedrich, S.; Vasters, J.; Drobe, M.; Sand, W.; Will- principles. Green Chem. 2016, 18 (3), 753−759. scher, S. Biomining: Metal recovery from ores with microorgan- (38) Dutta, T.; Kim, K.-H.; Deep, A.; Szulejko, J. E.; Vellingiri, K.; Ku- isms. In Geobiotechnology I; Schippers, A.; Glombitza, F.; Sand, mar, S.; Kwon, E. E.; Yun, S.-T. Recovery of nanomaterials from W., Eds.; Advances in Biochemical Engineering/Biotechnology; battery and electronic wastes: A new paradigm of environmental Springer-Verlag: Berlin, 2014; pp. 1–47. waste management. Renew. Sustain. Energy Rev. 2018, 82, 3694– (20) Sethurajan, M.; van Hullebusch, E. D.; Nancharaiah, Y. V. Bio- 3704. technology in the management and resource recovery from metal (39) Ahmad, I.; Davies, H. A.; Kanwal, M. Microstructure and Mag- bearing solid wastes: Recent advances. J. Environ. Manage. 2018, netic Properties of RE2.28Fe13.58B1.14 Alloys. J. Magn. Magn. 211, 138-153. Mater. 2012, 324, 3971–3974. (21) Chen, B.-W.; Wen, J.-K. Feasibility study on heap bioleaching of (40) Lee, D.; Hilton, J. S.; Liu, S.; Zhang, Y.; Hadjipanayis, G. C.; chalcopyrite. Rare Met. 2013, 32(5), 524–531. Chen, C. H. Hot-pressed and hot-deformed nanocomposite (22) Spedding, F. H.; McGinnis, W. J. Preparation of rare earth metals, (Nd,Pr,Dy)/sub2/Fe/sub14/B//splalpha/-Fe-based magnets. IEEE Ames, 1951. Trans. Magn. 2003, 39, 2947–2949. (23) https://www.ameslab.gov/mpc/the-ames-process-rare-earth-met- (41) https://mqitechnology.com/wp-content/uploads/2017/09/mqp-b- als; Ames Laboratory, The Ames Process for Rare Earth Metals. 20173-070.pdf. Magnequench, MQPTM-B-20173-070 Isotropic Last accessed on September 20, 2019. Powder. Last accessed on May 13, 2019. (24) Mudring, A.-V.; Prodius, D.; Nlebedim, I. C. Dissolution and sep- (42) Lv, W.; Wang, Z.; Cao, H.; Sun, Y.; Zhang, Y.; Sun, Z. A critical aration of rare earth metals. US20180312941A1, November 01, review and analysis on the recycling of spent lithium-ion batteries. 2018. ACS Sustainable Chem. Eng. 2018, 6 (2), 1504−1521. (25) Jacobson, J.; Kim, A. Oxidation behaviour of NdFeB magnets, J. (43) Nlebedim, I. C.; King, A. H. Addressing Criticality in Rare Earth Appl. Phys. 1987, 61, 3767-3765. Elements via Permanent Magnets Recycling. JOM. 2018, 70, 115– (26) Li, Y.; Evans, H. E.; Harris, I. R.; Jones, I. P. The oxidation of 123. NdFeB magnets, Oxid. Met. 2003, 59, 167-182. (44) Kumar, A. Permanent Magnets: Technologies and Global Mar- (27) Edgley, D. S.; Le Breton, J. M.; Steyaert, S.; Ahmed, F. M.; Harris, kets, Bangalore, 2016. I. R.; Teillet, J. Characterisation of High Temperature Oxidation of (45) Herrmann, M. S. An easy-to-do plating experiment. J. Chem. Educ. Nd-Fe-B Magnets. J. Magn. Magn. Mater. 1997, 173, 29-42. 1992, 69, 60-61. (28) Tanaka, M.; Oki, T.; Koyama, K.; Narita, H.; Oishi, T. Recycling (46) Kitagawa, J.; Uemura, R. Rare earth extraction from NdFeB mag- of Rare Earths from Scrap. Handb. Phys. Chem. Rare Earths. 2013, net using a closed-loop acid process, Sci. Rep. 2017, 7 (1), 8039 (1- 43 (Chapter 255), pp. 159–211. 6). (29) Virender, K. S.; Klingelhöfer, G.; Nishida, T. Part V: Iron Oxides. (47) Bandara, H. M. D.; Field, K. D.; Emmert, M. H., Rare earth recov- Mössbauer Spectroscopy. 2013, (Chapters 18-25), pp. 349-532. ery from end-of-life motors employing green chemistry design (30) Cook, D. C. Application of Mössbauer Spectroscopy to the Study principles, Green Chem. 2016, 18, 753−759. of Corrosion. Hyperfine Interactions 2004, 153, 61-82. (48) Ö nal, M. A. R.; Borra, C. R.; Guo, M.; Blanpain, B.; Van Gerven, (31) McGuiness, P. J.; Harris, I. R., Rozendaal, E.; Ormerod, J.; Ward, T. Recycling of NdFeB magnets using sulfation, selective roasting, M. The production of a Nd-Fe-B permanent magnet by a hydrogen and water leaching. J. Sustain. Metall. 2015, 1 (3), 199−215. decrepitation/attritor milling route. J. Mater. Sci. 1986, 21(11), (49) Hignett, T. P. Fertilizer manual. Publishers: Dordrecht, The Neth- 4107-4110. erlands, 1985, 15, pp. 1-363. (32) Williams, A. J.; Gutfleisch, O.; Harris, I. R. S-HDDR induced cav- itation in NdFeB. J. Alloys Compd. 1996, 232, 22-26.

8 For Table of Contents Use Only

An efficient acid-free leaching process for rare earth metals and cobalt from Nd-Fe-B and Sm-Co magnets has been developed.