Oxidation behavior of spent NdFeB magnet under selective oxidation conditions

doi:10.21203/rs.3.rs-2072051/v1

Download PDF

Research Article

Oxidation behavior of spent NdFeB magnet under selective oxidation conditions

https://doi.org/10.21203/rs.3.rs-2072051/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 22 Dec, 2023

Read the published version in High Temperature Corrosion of Materials →

Version 1

posted

You are reading this latest preprint version

Oxidation behavior of the spent NdFeB magnet was investigated when exposed to the selective oxidation conditions to achieve effective separation of Nd element by forming an oxide form from metallic Fe one. XRD results confirmed that the selective oxidation heat treatment successfully leaded to the formation of Nd₂O₃ and α-Fe phases, which was ascribed to the dissociation of the Nd₂Fe₁₄B phase as a consequence of only Nd’s oxidation. The oxidation of the Nd element kept proceeding with time and temperature of the heat treatment, requiring 60 min when heat treated at 950 °C for the 100% oxidation degree based on the weight gain calculation. From the SEM analysis, it can be inferred that the oxygen diffusion for the oxidation mainly occurred through grain boundary at the initial stage and then, α-Fe lattice diffusion for inside grains. TEM analysis confirmed that the Nd₂O₃ formed at the early oxidation had hcp structure and only coarsening with oxidation, maintaining its structure. As the no oxidation of the Fe element did not induced diffusion barriers such as the outmost Fe₂O₃ layer and Fe₂O₃ matrix, facile oxygen diffusion through the grain boundary and α-Fe lattice was possible, leading to the activation energy for oxygen diffusion as low as 28 kJ/mol. The specific oxidation condition maintained the good microstructure for oxygen diffusion, α-Fe matrix containing submicron Nd₂O₃ particles, which made it possible that the good diffusion paths such as grain boundary and α-Fe lattice kept working during the oxidation.

NdFeB magnets

Oxidation

Selective oxidation

Recycling

Phase separation

Since the advent of NdFeB magnets in 1984, their outstanding magnetic properties have contributed to not only performance improvement of industrial products but miniaturization of them, allowing them to prevail in lots of applications such as hard disk drives (HDDs), motors, electric vehicles and wind-power generations [1–4]. However, the NdFeB magnets’ poor corrosion resistance restricted their usages so that corrosion and oxidation kinetics of the NdFeB magnets have extensively studied to solve their degradation issue during service [5–7]. Generally, there are three phases in the NdFeB magnets; Nd₂Fe₁₄B (ϕ phase), Nd_1+εFe₄B₄ (η phase) and Nd rich phase [8, 9]. As a magnetic phase, Nd₂Fe₁₄B which occupied almost 85% of the magnet oxidizes easily over 150°C, resulting in the dissociation of the phase into an α-iron and Nd oxide. The first work on the oxidation of the NdFeB magnets was carried out between 400 and 600°C and revealed the formation of a gray layer on the magnet’ surface and the parabolic growth of the gray layer with oxidation time [10]. Next work on the microstructure of the oxidized zone identified the oxidation product as the α-Fe matrix containing Nd oxide particles, which was later revealed to be the internal oxidation region [11]. Edgley et al. conducted the oxidation in pure oxygen and air and showed the successive oxidation reactions with temperature as well as transgranular growth of the gray layer into the magnet [12]. Firdaus et al. conducted the high temperature oxidation ranging 423–1,227°C, where the microstructural evolutions and oxidation kinetics were investigated [13, 14].

Not only the degradation of the NdFeB magnets, recycling of rare earth elements from spent NdFeB magnets was also critical to oxidation in terms of roasting [15]. As the whole leaching would require large amount of sulfuric acid to dissolve the spent magnet completely as well as hydroxide to neutralize an effluent [16–18]. For example, 272 kg of H₂SO₄ and 1,600 kg of NaOH are needed for the dissolution and neutralization of 1,000 kg of the NdFeB magnet. Selective leaching was addressed to overcome these issues based on a potential-pH diagram for Fe-H₂O and Nd-H₂O systems, where the co-existed region of Nd³⁺ and Fe₂O₃ was presented [16]. That means that if Nd and Fe elements in spent NdFeB magnets would become oxide forms through roasting, selective dissolution of Nd oxide is possible, leaving Fe₂O₃ as a residue. However, it is not easy to obtain only Nd₂O₃ and Fe₂O₃ through a simple roasting. If roasting temperature were not enough, Fe₃O₄ would be formed instead of Fe₂O₃. Whereas, high roasting temperature causes a ternary Nd-Fe oxide phase (NdFeO₃) [19, 20]. The poor stability of Fe₃O₄ compared to Fe₂O₃ tends to be dissolved with Nd₂O₃ during leaching process, while the high stability of NdFeO₃ demands more concentration of mineral acids [21, 22]. As both cases deteriorate leaching selectivity between Nd and Fe elements, lots of works have focused on controlling phases to improve the leaching selectivity [23–26]. Yoon et al. investigated grinding and roasting conditions for selective leaching and suggested the grinding with NaOH and subsequent roasting at 400°C as a method for Fe₂O₃ and Nd₂O₃ formation, where the leaching with the good selectivity of 94.2% Nd, 93.1% Dy and 1% Fe was achieved [27]. Maroufi et al. used oxidation at 1,000°C and subsequent reduction above 1,400°C processes to prevent the formation of NdFeO₃ and then, produced the mixed phases consisting of a rare-earth oxide and iron metal [28]. Orefice et al. developed the selective roasting to avoid NdFeO₃ formation by adopting argon atmosphere and carbon as an iron reducing agent for the roasting [29].

In this study, a selective oxidation method based on thermodynamic consideration was employed to oxidize only Nd element in the spent NdFeB magnet. Separating Nd element selectively from other ones by forming an oxide would be beneficial in controlling the final phases such as Nd₂O₃ and Fe₂O₃ instead of NdFeO₃. Unlike the previous methods using the high temperature about 1,400°C, the selective oxidation used in this study was conducted below 1,000°C and in a specific partial pressure ratio of hydrogen gas and water vapor. Understanding on oxidation characteristics including phase evolutions with the selective oxidation would be a first step to obtain the desired phases for good leaching selectivity.

2.1 Specimens preparation

The NdFeB magnet scrap used for this experiment was pulverized by combination of jaw crusher and rod milling to obtain desired powder size. The final powders were prepared by the sieving process, where the sieve with ASTM No. 450 was finally applied. The morphology and size distribution of the sieved powder were shown in Figs. 1 and 34.3 µm of D₉₀ was obtained for the final powder size. The composition of the NdFeB magnet powder was analyzed by X-ray fluorescence (XRF, Bruker, M4 tornado) and the contents of Nd and Dy were revealed to be 24.19 wt% and 1.90 wt%.

2.2 Selective oxidation heat treatment

Conditions for the selective oxidation heat treatment were derived from thermodynamic calculation, where partial pressure ratio of hydrogen gas and water vaper (\({p}_{{H}_{2}\left(g\right)}/{p}_{{H}_{2}O\left(g\right)})\) was considered as an oxidant agent. The desired \({p}_{{H}_{2}\left(g\right)}/{p}_{{H}_{2}O\left(g\right)}\) could be achieved by flowing hydrogen gas into a deionized water bath, where water vapor pressure could be adjusted by controlling bath’s temperature. As shown in Fig. 2, if the \({p}_{{H}_{2}\left(g\right)}/{p}_{{H}_{2}O\left(g\right)}\) would be placed between the oxidation driving forces of rare earth elements (Nd, Dy) and Fe one at a given temperature, only rare earth elements which have the higher oxidative driving force than Fe one could be oxidized, leaving Fe element under reducing condition. As Nd and Dy have similar oxidation behavior, Nd element in this study represents both elements. The selective oxidation heat treatment was possible if the \({p}_{{H}_{2}\left(g\right)}/{p}_{{H}_{2}O\left(g\right)}\) value would be over about 10 so that bath temperature was set at 10°C which leaded to 80.4 of the \({p}_{{H}_{2}\left(g\right)}/{p}_{{H}_{2}O\left(g\right)}\). The selective oxidation heat treatment was conducted at 750, 850 and 950°C with time to see variations of oxidation degree.

2.3 Characterizations

Laser particle size analyzer (Malvern, Mastersizer 3000) was used to find out size distribution of the NdFeB powder. Phase evolutions with the selective oxidation heat treatment were characterized by X-ray diffraction (XRD, Panalytical B.V, Emptrean) with Cu Ka radiation. Microstructural changes with time and temperature of the heat treatment were examined by scanning electron microscopy (SEM, FEI company, Quanta 250) and high-resolution transmission electron microscope (HRTEM, JEOL, JEM-2200FS).

3.1 Microstructural evolutions with selective oxidation heat treatment

As the selective oxidation heat treatment oxidized only rare earth elements like Nd and Dy which have higher oxidation driving force, leaving Fe one under reduction condition, the selective oxidized samples displayed α-Fe as well as Nd₂O₃ phases as a result of the dissociation of Nd₂Fe₁₄B phase and selective oxidation of the rare earth elements, as shown in Fig. 3. As mentioned in the experimental procedure, due to not only the Dy’s small amount but the almost same oxidation characteristic, Nd was used to represent both rare earth elements. Regardless of the selective oxidation temperature, all samples showed the same phase configuration but the intensity of the peak corresponding to (222) plane (27.996 °) of Nd₂O₃ increased with temperature, indicating that the higher temperature accelerated the oxidation degree of the NdFeB magnet. As shown in Fig. 3(b), the peak’s intensity for the (222) plane grew with not only temperature but also time of the heat treatment. That means that the oxidation kept proceeding with the temperature and time. The highest peak intensity was achieved when heat treated at 950°C for 60 min and additional increase in the peak intensity with time was not obvious, indicating that oxidation of the NdFeB powder almost saturated for the heat treatment at 950°C for 60 min. This oxidation degree was further investigated in terms of weight gain.

Under the selective oxidation condition, only Nd becomes oxidized as follow.

2Nd(s) + 3/2O₂(g) → Nd₂O₃ (1)

That means that if Nd element in a NdFeB magnet would become oxidized completely, the weight gain as a result of the Nd oxidation becomes 1.17 times higher than the Nd amount. Based on this theoretical consideration on the oxidation, variations of the weight gain with time and temperature of the heat treatment were calculated and the results were shown in Fig. 4. The oxidation became faster with temperature and the 100% oxidation could be obtained only for the 950°C condition. For the selective oxidation heat treatment at 750°C and 850°C, even if the heat treatment time extended to 90 min, the full oxidation of the NdFeB powder did not occur. Only 68% and 95% of the weight gain was obtained, respectively. On the other hand, the selective oxidation heat treatment at 950°C required only 60 min to achieve 100% oxidation and further heat treatment up to 90 min did not cause additional weight gain as the Nd element in the NdFeB powder was fully consumed for the oxidation.

As this oxidation would give an impact on microstructure, microstructural evolutions with the heat treatment conditions were investigated. To compare oxidation characteristics with temperature, similar oxidation degrees for each temperature were considered, as shown in Fig. 5. When the oxidation degree was about 20%, the oxidation mainly occurred around grain boundaries regardless of the heat treatment temperature, forming Nd₂O₃ particles around grain boundary. Unlike the oxidation in air which oxygen diffusion occurred through grain lattice, grain boundary seemed to be a main path for oxygen diffusion, making that the center of the powders was oxidized even in case of the 20% oxidation degree. With proceeding the oxidation, the oxidized region spread into grains from grain boundaries. The samples with about 60% oxidation degree showed that the Nd₂O₃ particles as a mark of oxidation dispersed throughout grains including grain boundaries. The sample heat treated at 750°C seemed to be oxidized only around grain boundaries but actually, the small Nd₂O₃ particles existed in the grains, as shown in Fig. 5(d). That means the heat treatment temperature affected not only the oxidation rate but growth of the Nd₂O₃ particles.

To figure out 100% oxidation state, the microstructural evolutions of the sample heat treated at 950°C were presented, as shown in Fig. 6. As mentioned before, the Nd₂O₃ particles, which was formed as a result of the oxidation, spread out from grain boundaries into grains and their size kept growing as the selective oxidation progressed. Once the oxidation degree reached 100%, with big Nd₂O₃ particles at grain boundaries, turned from pre-existed Nd rich phases, the Nd₂O₃ particles with over 100 nm were formed throughout in grains. As shown in Fig. 6(d), even though the oxidation degree had already reached 100% at the 60 min heat treatment, further heat treatment up to 90 min still leaded to the Nd₂O₃ particles’ growth. TEM analysis was carried out to clarify the resultant phases after the selective oxidation. Figure 7 presented HRTEM images showing Nd₂O₃, α-Fe and Nd₂Fe₁₄B phases, where selective area diffraction patterns confirmed each phase and revealed that all phases were crystalline. At the beginning of the oxidation, the Nd₂Fe₁₄B matrix was decomposed and thus, numerous crystalline Nd₂O₃ particles with about 10 nm in diameter was formed in the α-Fe phase. As reported in the literature, the nanocrystalline Nd₂O₃ had hcp structure [30]. Even though amorphous Nd oxide and/or cubic Nd₂O₃ would present when heated at moderate temperature like 400°C, the high temperature such as 950°C resulted in the hexagonal Nd₂O₃ [30, 31]. As shown in Fig. 6(b), un-oxidized region, the center of powders still had the Nd₂Fe₁₄B phase, indicating that the oxidation turned the Nd₂Fe₁₄B phase into the α-Fe containing nanocrystalline Nd₂O₃ structure. Due to the high temperature oxidation, particle coarsening occurred, leading to about 200 nm Nd₂O₃ particles when the 100% oxidation degree reached, but their crystal structure still held hcp one; no phase change happened.

3.2 Oxidation mechanism during selective oxidation heat treatment

It is known that when NdFeB magnets and/or powders were oxidized in air, an external oxidation zone (EOZ) consisting of mainly iron oxide (Fe₂O₃) and internally oxidized zone (IOZ) which was the region where the Nd₂Fe₁₄B phase was dissociated were formed [12]. Once oxidized, the Nd₂Fe₁₄B phase was dissociated into the α-Fe and Nd₂O₃ with the oxidation of the Nd rich one, where the α-Fe diffused toward the surface to form the iron oxide [32, 33]. Oxygen diffusion into a magnet was known to occur through grain boundaries and/or high angle α-Fe in IOZ (columnar α-Fe grain) [34]. While, as discussed before, the oxygen diffusion under the selective oxidation condition seemed to occur through grain boundaries and then, grain lattice (α-Fe) into inner grains, as schematically described in Fig. 8. These characteristics would be ascribed to the specific microstructure aspect related to the no Fe oxidation under the selective oxidation heat treatment. Usually, roasting in air oxidized not only the Nd element but the Fe element so that iron oxides like Fe₂O₃ and Fe₃O₄ were formed as well. These iron oxides would impede oxygen diffusion compared to the grain boundary and/or high angle α-Fe (columnar α-Fe grain). While, the selective oxidation made condition, the α-Fe was retained under the selective oxidation condition, holding the α-Fe matrix containing a dispersion of Nd₂O₃ nanoparticles without the formation of the outermost iron oxide layer and/or iron oxide matrix even if the oxidation kept proceeding. That means that the no Fe oxidation in the spent NdFeB magnet made it possible that the preferential paths for oxygen diffusion were always available during the oxidation process. Thus, only 60 min was enough to oxidize the entire grains in the magnet powders due to the facile oxygen diffusion.

This kind of oxygen diffusion would alter activation energy for oxidation. Based on a parabolic oxidation rate law, the activation energy for oxidation could be drawn from the weight gain as a result of oxidation as follow [14, 35].

ΔW = [kt]^1/2 (2)

k = A∙exp(- E_a/RT) (3)

where, ΔW is weight gain per unit area, k is oxidation rate constant, A is constant, E_a is activation energy for oxidation, R is gas constant and T is absolute temperature. Typically, thermal activated processes like diffusion could be described by the Arrhenius equation, as shown in Fig. 9 and the slope was used to determine the activation energy for oxidation whose value was calculated to be about 28 kJ/mol. Compared to the activation energies reported in the literatures ranging from 80 ~ 120 kJ/mol [11, 12, 14, 19, 33], this value was very small, confirming that the oxygen diffusion was much facilitated under the selective oxidation condition. As discussed before, the selective oxidation did not make the Fe element oxidized, allowing oxygen diffusion to get easy access to facile diffusion paths such as grain boundary and columnar α-Fe grain boundary. Based on the calculated activation energy, the required time for the 100% oxidation at a given temperature could be drawn; 190 min for 750°C, 108 min for 850°C and only 35 min when heat treated at 1050°C.

The selective oxidation heat treatment was successfully applied to oxidize only Nd element in the spent NdFeB magnet. XRD analysis confirmed that the Nd₂Fe₁₄B phase was dissociated into the α-Fe and Nd₂O₃ phases after the selective oxidation heat treatment. With proceeding the heat treatment, the weight gain related to the oxidation degree occurred and 100% oxidation could be obtained when heat treated at 950°C for 60 min. From the SEM results showing microstructural evolutions, it can be inferred that the oxygen diffusion for the oxidation mainly occurred through the grain boundary and then, α-Fe lattice for inner grain oxidation. TEM results showed that the nanocrystalline Nd₂O₃ formed at the early oxidation stage had hcp structure and only coarsening of the Nd₂O₃ particles occurred without any phase changes as the oxidation proceeded. Due to the no microstructural change except for the Nd oxidation during the selective oxidation, the preferential paths for oxygen diffusion like the grain boundary and α-Fe lattice could be maintained, allowing facile oxygen diffusion to continue. The calculated activation energy for the oxidation was about 28 kJ/mol, whose value was very low compared to the ones reported in the literatures. The facile oxygen diffusion during the selective oxidation was ascribed to the maintained microstructure, the α-Fe matrix containing the submicron Nd₂O₃ particles.

Data availability

The raw/processed data required to reproduce these findings cannot be shared at this time due to legal or ethical reasons.

Acknowledgement

This work was supported by the Korea Institute of Energy Technology Evaluation and Planning(KETEP) grant funded by the Korea government(MOTIE) (No. 20217510100070).

Conflict of interest

On behalf of all authors, the corresponding author states that there is no conflict of interest.

AuthorContributions

Conceptualization, B.S, K. P; methodology, B. S and K. P; software, T. N; validation, K. P and H.-K. P; formal analysis, B.S and K. P; resources, B. S and T. N; writing—original draft preparation, B. S; writing—review and editing, K. P; supervision, K. P; funding ac-quisition, K. P; All authors have read and agreed to the published version of the manuscript.

J.J. Croat, J.F. Herbst, R.W. Lee, F.E. Pinkerton, J. Appl. Phys. 55, 2078 (1984)
K. Habib, J. Clean. Prod. 230, 90 (2019)
L. Ciacci, I. Vassura, Z. Cao, G. Liu, F. Passarini, Resour. Conserv. Recycl. 142, 143(2019)
B. Swain, L. Kang, C. Mishra, J. Ahn, H.S. Hong, Waste Manag. 45, 351 (2015)
G. Yan, P. McGuiness, J. Farr, I. Harris, J. Alloys Compd. 478, 188 (2009)
L. Yu, Y. Wen, M. Yan, J. Magn. Magn. Mater. 283, 353 (2004)
D. Scott, B. Ma, Y. Liang, C. Bounds, J. Appl. Phys. 79, 5501 (1996)
G. Yan, A. Williams, J. Farr, I. Harris, J. Alloys Compd. 292, 266 (1999)
A. El-Aziz, Mater. Corros. 54, 88 (2003)
J.M. Le Breton, S. Steyaert, J. Condens. Matter Phys. 11, 4941 (1999)
Y. Li, H. Evans, I. Harris, I. Jones, Oxid. Met. Oxidation of Metals. 167 (2003)
D. Edgley, J. Le Breton, S. Steyaert, F. Ahmed, I. Harris, J. Teillet, J. Magn. Magn. Mater. 173, 29 (1997)
M. Firdaus, M.A. Rhamdhani, W.J. Rankin, M. Pownceby, N.A. Webster, A.M. D’Angelo, K. McGregor, Corros. Sci. 133, 374 (2018)
M. Firdaus, M.A. Rhamdhani, Y. Durandet, W.J. Rankin, K. McGregor, Corros. Sci. 133, 318 (2018)
Y. Yang, A. Walton, R. Sheridan, K. Güth, R. Gauß, O. Gutfleisch, M. Buchert, B.-M. Steenari, T. Van Gerven, P.T. Jones, J. Sustain. Metall. 3, 122 (2017)
M. Tanaka, T. Oki, K. Koyama, H. Narita, T. Oishi, Recycling of rare earths from scrap, in: Handbook on the physics and chemistry of rare earths, (Elsevier, 2013), pp. 159–211.
P. Venkatesan, Z. Sun, J. Sietsma, Y. Yang, Sep. Purif. Technol. 191, 384 (2018)
P. Venkatesan, T. Vander Hoogerstraete, T. Hennebel, K. Binnemans, J. Sietsma, Y. Yang, Green Chem. 20, 1065 (2018)
M. Stewart, B. Roebuck, M. Gee, J. Mater. Sci. 26, 1401 (1991)
M.A.R. Önal, C. Jönsson, W. Zhou, T. Van Gerven, M. Guo, A. Walton, B. Blanpain, J. Alloys Compd. 728, 727 (2017)
M.A.R. Önal, C.R. Borra, M. Guo, B. Blanpain, T. Van Gerven, J. Sustain. Metall. 1, 199 (2015) \
M. Orefice, A. Van den Bulck, B. Blanpain, K. Binnemans, J. Sustain. Metall. 6, 91 (2020)
A. Kumari, M.K. Sinha, S. Pramanik, S.K. Sahu, Waste Manag. 75, 486 (2018)
A. Kumari, R. Raj, N. Randhawa, S.K. Sahu, Hydrometallurgy. 201, 105581 (2021)
T. Vander Hoogerstraete, B. Blanpain, T. Van Gerven, K. Binnemans, RSC Adv. 4, 64099 (2014)
J.P. Rabatho, W. Tongamp, Y. Takasaki, K. Haga, A. Shibayama, J. Mater. Cycles Waste Manag. 15, 171 (2013)
H.-S. Yoon, C.-J. Kim, K.W. Chung, S. Jeon, I. Park, K. Yoo, M.K. Jha, Metals. 5, 1306 (2015)
S. Maroufi, R. Khayyam Nekouei, V. Sahajwalla, ACS Sustain. Chem. Eng. 5, 6201 (2017)
M. Orefice, K. Binnemans, T. Vander Hoogerstraete, RSC adv. 8, 9299 (2018)
S. Wang, Y. Li, J. Magn, Magn. Mater. 285, 177 (2005)
R. Lok, E. Budak, E. Yilmaz, J. Mater. Sci. Mater. 31, 3111 (2020)
D.C. Nababan, R. Mukhlis, Y. Durandet, M.I. Pownceby, L. Prentice, M.A. Rhamdhani, Corros. Sci. 189, 109560 (2021)
D.C. Nababan, R. Mukhlis, Y. Durandet, M.I. Pownceby, L. Prentice, M.A. Rhamdhani, Corros. Sci. 182, 109290 (2021)
I. Skulj, H. Evans, I. Harris, J. Mater. Sci. 43, 1324 (2008)
W. Liu, M. Yue, G. Wang, D. Zhang, J. Zhang, Corros. 66, 055004 (2010)

No competing interests reported.

Download PDF

Journal Publication

published 22 Dec, 2023

Read the published version in High Temperature Corrosion of Materials →

Version 1

posted

You are reading this latest preprint version

Oxidation behavior of spent NdFeB magnet under selective oxidation conditions

Status:

Journal Publication

Version 1

Abstract

Figures

1. Introduction

2. Experimental Procedure

2.1 Specimens preparation

2.2 Selective oxidation heat treatment

2.3 Characterizations

3. Results And Discussion

3.1 Microstructural evolutions with selective oxidation heat treatment

3.2 Oxidation mechanism during selective oxidation heat treatment

4. Conclusions

Declarations

References

Additional Declarations

Status:

Journal Publication

Version 1