Hydrogen Absorption and Release Properties of MgH 2, Mg 2 Ni, and Ni-added Mg via Reactive Mechanical Grinding

Transcription

1 [Research Paper] 대한금속 재료학회지 (Korean J. Met. Mater.), Vol. 56, No. 2 (2018) pp DOI: /KJMM Hydrogen Absorption and Release Properties of MgH 2, Mg 2 Ni, and Ni-added Mg via Reactive Mechanical Grinding Seong-Hyeon Hong 1 and Myoung Youp Song 2, * 1 Small and Medium Sized Corporate Supporting Department, Korea Institute of Materials and Science (KIMS), Changwon 51508, Republic of Korea 2 Division of Advanced Materials Engineering, Research Center of Advanced Materials Development, Engineering Research Institute, Chonbuk National University, Jeonju 54896, Republic of Korea Aabstract : In this work, MgH 2, Mg 2 Ni, and Ni were added to Mg in order to improve the hydrogen-storage properties of Mg. A 84 wt% Mg+10 wt% MgH 2 +5 wt% Mg 2 Ni+1wt% Ni sample (designated the Mg-based sample) was prepared by milling under a hydrogen atmosphere in a planetary ball mill for 5 h. The hydrogen absorption and release properties of the prepared sample were investigated and were compared with those of a 94 wt% MgH 2 +5 wt% Mg 2 Ni+1wt% Ni sample (the MgH 2 -based sample). The Mg-based sample had larger quantities of hydrogen absorbed and released for 60 min than the MgH 2 -based sample. The Mg-based sample had an effective hydrogen-storage capacity (the quantity of hydrogen absorbed for 60 min) of near 5.7 wt%. At n=2, the Mg-based sample absorbed 5.18 wt% for 2 min, 5.57 wt% for 10 min, and 5.67 wt% for 60 min at 648 K. At n=1, the Mg-based sample released wt% for 2 min, 5.13 wt% for 10 min, and 5.49 wt% for 60 min at 648 K. The addition of MgH 2, Mg 2 Ni, and Ni to Mg had stronger effects than the addition of Mg 2 Ni and Ni to MgH 2. It is believed that the effects of the reactive mechanical grinding (reducing the Mg particle size, making clean surfaces, and creating defects) are stronger in the Mg-based sample than in the MgH 2 -based sample. (Received Accepted August 16, 2017; Accepted September 29, 2017) Keywords : hydrogen storage, magnesium, mechanical milling, X-ray diffraction, MgH 2 Mg 2 Ni or Ni addition. 1. INTRODUCTION Magnesium is a prospective hydrogen-absorbing solidstate material with a high volumetric hydrogen-storage capacity (7.6 wt%). However, the hydriding rate of magnesium drops substantially when magnesium hydride forms on the surfaces of magnesium particles [1]. An intermetallic compound, Mg 2 Ni, reacts easily with hydrogen at about 573 K under a moderate hydrogen pressure to form its hydride Mg 2 NiH 4. Mg 2 Ni is known to have higher hydriding and dehydriding rates than Mg [2-7]. The hydride of Mg 2 Ni, Mg 2 NiH 4, is less stable than that of Mg, MgH 2. Many works have investigated the Mg 2 Ni-H 2 system [8-13]. The addition of Mg 2 Ni is believed to increase the hydriding and dehydriding rates of Mg. For example, the *Corresponding Author: Myoung Youp Song [Tel: , songmy@jbnu.ac.kr] Copyright c The Korean Institute of Metals and Materials hydrogen-storage properties of Mg were enhanced by preparing Mg 2 Ni-containing Mg alloys [2,3,14,15]. Cermak et al. [14] investigated the separate catalytic effects of Ni, Mg 2 Ni, and Mg 2 NiH 4 on the hydrogen desorption characteristics of MgH 2. They chose the chemical composition samples so that the mass ratio of Mg/Ni was always close to 3.25 (eutectic composition of the Mg-Mg 2 Ni binary alloy) and observed that the catalytic efficiency of Mg 2 NiH 4 was considerably higher than that of pure Ni and non-hydrated intermetallic compound Mg 2 Ni. Zaluska et al. [15] reported that a sample of ball-milled mixture of 65 wt% of MgH 2 and 35 wt% of Mg 2 NiH 4 desorbed hydrogen quickly at temperatures around K with a hydrogen capacity exceeding 5 wt%, concluding that mechanicallytreated hydride mixtures offer a new opportunity for magnesium-based materials, which take advantage of the high capacity of magnesium hydride and operate at much lower temperatures than conventional magnesium. Mg 2 Ni was chosen as a compound additive.

2 142 대한금속 재료학회지제 56 권제 2 호 (2018 년 2 월 ) The hydriding and dehydriding rates of Mg were also increased by adding Ni [4-7,16] and by adding by adding nano-sized Ni particles [17] in a high-energy ball mill. Niaz et al. [18] synthesized nanostructured Mg Ni alloy by the thermal decomposition of bipyridyl complexes of Mg and Ni metals at 773 K for 24 h under dry argon gas atmosphere. They reported that the alloy exhibited superior hydrogen absorption and desorption behavior with 3.2 wt% absorption within 1 min at 573 K and about 3 wt% desorption within 5 10 min at 573 K. In the present study, we selected Ni as a transition element additive. The addition of MgH 2, which is brittle, is believed to increase the hydriding and dehydriding rates of Mg by being pulverized during reactive mechanical grinding and helping make the Mg particles smaller. Only a small amount of Ni (1wt% Ni) was added to avoid significantly reducing the hydrogen storage capacity of the material. Slightly more Mg 2 Ni (5 wt%) was added since the added Mg 2 Ni can increase hydriding and dehydriding rates and the Mg 2 Ni itself can absorb and release hydrogen under the same conditions as those for Mg. Ten wt% MgH 2 was added to help the pulverization of Mg during the reactive mechanical grinding. The purpose of this work was to improve the hydriding and dehydriding kinetics of Mg by using comparatively cheap Mg instead of MgH 2, which is expensive, and adding MgH 2 and catalysts Mg 2 Ni and Ni. Zaluska et al. [15] prepared a mixture of 65 wt% of MgH 2 and 35 wt% of Mg 2 NiH 4 by ball milling. In the previously reported works [2-7,14,16-18], MgH 2, Mg 2 Ni, or Ni was added respectively (not together) to Mg. In contrast, in the present work, MgH 2, Mg 2 Ni, and Ni were added simultaneously. In addition, the percentages of additives were different from those of the reported previous works and kept small to avoid significantly reducing the hydrogen storage capacity of the material. A sample with a composition of 84 wt% Mg+10 wt% MgH 2 +5 wt% Mg 2 Ni+1wt% Ni (named the Mg-based sample) was prepared by milling under a hydrogen atmosphere in a planetary ball mill for 5 h. The hydrogen absorption and release properties of the prepared sample were investigated and compared with those of a 94 wt% MgH 2 +5 wt% Mg 2 Ni+1wt% Ni sample (named the MgH 2 -based sample). 2. EXPERIMENTAL DETAILS Pure Mg powder (purity 98.5%, Cu<0.02%, Zn<0.05%, Al<0.05%, Si<0.05%, Mn<0.1%, Fe<0.05%, particle size - 50 mesh +325 mesh, Daejung, Korea), pure MgH 2 powder (hydrogen storage grade, Aldrich), Mg 2 Ni (High Purity Chemicals, Japan, shape powder, purity 99.9%), and Ni (Nano Technology, particle size 100 nm, purity 99.9%) were used to prepare the samples. Mixtures with a composition of 84 wt% Mg+10 wt% MgH 2 +5 wt% Mg 2 Ni+1wt% Ni (total weight =15 g) were mixed in a planetary ball mill with SUS balls (diameter = about 6 mm, total weight = about 416 g) for 5 h by repeating a cycle of 15 min milling and 5 min pause. A mill container with a volume of 307 ml was filled with high-purity hydrogen gas of 15 bar before milling. The revolution speed was 250 rpm [19,20]. The quantity of hydrogen absorbed or released was measured as a function of time by a volumetric method with an automatic hydriding and dehydriding apparatus. The initial hydrogen pressures of the reactor were 17 bar H 2 for the hydriding measurement and 0 bar H 2 for the dehydriding measurement [19,20]. After measurement of the hydriding and dehydriding properties, the sample was furnace cooled under hydrogen atmosphere. X-ray diffraction (XRD) analysis was carried out with Cu Kα radiation for the as-milled powders and for the samples after desorption using an X-ray diffractometer Rigaku D/max The scanning speed was 5 o /min, and the range of the scanning diffraction angle (2θ) was 20 o -80 o. The microstructures of the prepared samples were observed by scanning electron microscope (SEM, JSM-5800, JEOL) [19,20]. 3. RESULTS AND DISCUSSIONS Figure 1 shows the XRD pattern of the as-milled Mg-based sample. The as-milled sample contains Mg, MgH 2, Mg 2 NiH 4 and Ni, showing that Mg 2 Ni is and Mg may be hydrided during reactive mechanical grinding. During milling under hydrogen of the Mg-based sample, a part of the Ni forms Mg 2 NiH 4 by reaction with Mg and hydrogen, and a part of

3 Seong-Hyeon Hong, and Myoung Youp Song Fig. 1. XRD pattern of the as-milled Mg-based sample. 143 Fig. 3. SEM micrographs at different magnifications of the asmilled Mg-based sample. the sample was dehydrided. The background of the XRD pattern for the sample after cycling is lower than that of the XRD pattern of the as-milled sample, showing that the crystallinity of the sample after hydriding-dehydriding cycling is good and the as-milled sample is slightly noncrystalline. During dehydriding of the Mg-based sample, a part of the starting material, MgH2, decomposes and MgH2 formed during the milling decomposes. The reactions during the dehydriding-hydriding cycling of the Mg-based sample can be expressed as follows: Fig. 2. XRD pattern of the Mg-based sample after hydridingdehydriding cycling (dehydrided at the 4th hydriding-dehydriding cycle). Mg + 2MgH 2 + 2Mg 2 NiH 4 + Ni 3Mg + MgH 2 + (2) 2Mg2Ni + Ni + 6H2. SEM micrographs at different magnifications of the as-milled Mg-based sample are shown in Fig. 3. Fine particles of the as- the Mg forms MgH2 with hydrogen. The reaction during milled sample form agglomerates. The agglomerates of the as- milling under hydrogen of the Mg-based sample can be milled sample are quite spherical in shape and the particle size expressed as follows: is not homogeneous. SEM micrographs at different magnifications of the Mg- 4Mg + MgH2 + Mg2 Ni + 2Ni + 5H2 Mg + 2MgH2 + (1) 2Mg2NiH4 + Ni. based sample after hydriding-dehydriding cycling (dehydrided at the fourth hydriding-dehydriding cycle) are shown in Fig. 4. Fine particles of the sample after hydriding-dehydriding Figure 2 shows the XRD pattern of the Mg-based sample cycling (dehydrided at the fourth hydriding-dehydriding after hydriding-dehydriding cycling (dehydrided at the fourth cycle) also form agglomerates. The agglomerates of the sample hydriding-dehydriding cycle). The sample after hydriding- after cycling are irregular in shape and the agglomerate size is dehydriding cycling contains Mg, MgH2, Mg2Ni, and Ni. A not homogeneous. The fine particles forming the agglomerates small amount of MgH2 remains undecomposed, even after in the sample after cycling are smaller than those of the as-

4 144 대한금속 재료학회지 제56권 제2호 (2018년 2월) Fig. 4. SEM micrographs at different magnifications of the Mgbased sample after hydriding-dehydriding cycling (dehydrided at the 4th hydriding-dehydriding cycle). Fig. 6. Variation in the hydrogen content vs. desorption time curve with the number of cycles for the Mg-based sample. then becomes very low at K. At n=1, the sample absorbs 0.32 wt% for 2 min, 1.86 wt% for 5 min, 2.78 wt% for 10 min, and 3.03 wt% for 54 min at 648 K. At n=2, the sample absorbs 5.18 wt% for 2 min, 5.40 wt% for 5 min, 5.57 wt% for 10 min, and 5.67 wt% for 60 min at 648 K. The Mg-based sample has an effective hydrogen-storage capacity of near 5.7 wt%. At n=4, the sample absorbs 2.95 wt% for 2 min, 4.13 wt% for 5 min, 4.45 wt% for 10 min, and 4.92 wt% for 60 min at 573 K. Figure 6 shows the variation in the dehydriding curve (hydrogen content vs. desorption time curve) with the number of cycles, n, at K for the Mg-based sample. At n = 1and n=2, in the beginning the dehydriding rate is Fig. 5. Variation in the hydrogen content vs. absorption time curve with the number of cycles for the Mg-based sample. extremely high until 11 and 9 min, respectively, and then becomes very low at 648 K. At n=3, at the beginning the hydriding rate is very high until 14 min and then becomes milled Mg-based sample. The particles became smaller, very low. At n=4, the sample is dehydrided at 573 K and then possibly since the expansion and contraction of the particles at 623 K. The sample releases 2.79 wt% H at 573 K and 2.12 due to hydriding and dehydriding reactions made cracks and wt% H at 623 K, the total quantity of hydrogen released at caused the particles to become smaller. The sizes of the the 5th cycle being 4.91 wt%. At n=1, the sample releases agglomerates in the sample after cycling are similar to those wt% for 2 min, 1.63 wt% for 5 min, 5.13 wt% for 10 for the as-milled Mg-based sample sample. min, and 5.49 wt% for 60 min at 648 K. At n=3, the sample The variation in the hydriding curve (hydrogen content vs. absorption time curve) with the number of cycles, n, at 573- releases 0.15 wt% for 2 min, 0.83 wt% for 5 min, 3.48 wt% for 10 min, and 4.93 wt% for 60 min at 623 K. 648 K for the Mg-based sample is shown in Fig. 5. At n = Figure 7 shows the hydriding curves at 648 K for the 1, in the beginning the hydriding rate is quite high until 6 min MgH2-based sample and the Mg-based sample at n=2. From and then becomes very low at 648 K. At n = 2-4, at the the beginning, the hydriding rates are extremely high until 1 beginning the hydriding rate is very high until 1-3 min and min for the MgH2-based sample and 2 min for the Mg-based

5 Seong-Hyeon Hong, and Myoung Youp Song 145 Fig. 7. Hydriding curves at 648 K for the MgH 2 -based sample and the Mg-based sample at n=2. Fig. 8. Dehydriding curves at 648 K for the MgH 2 -based sample and the Mg-based sample at n=2. Fig. 9. Dehydriding curves at 573 K and then 623 K for the MgH 2 - based sample at n=5 and the Mg-based sample at n=4. sample, then become low until 6 min for the MgH 2 -based sample and 13 min for the Mg-based sample, and finally they are very low. The Mg-based sample has a very slightly lower initial hydriding rate but a larger quantity of hydrogen absorbed for 60 min than the MgH 2 -based sample. The Mgbased sample and the MgH 2 -based sample have initial hydriding rates of 2.59 and 4.63 wt% H/min, respectively. Their quantities of hydrogen absorbed for 60 min are 5.67 and 4.98 wt% H, respectively. The dehydriding curves at 648 K for the MgH 2 -based sample and the Mg-based sample at n=2 are presented in Fig. 8. The dehydriding rates are high at the beginning, very high from about 3 min until about 11 min for the MgH 2 -based sample and from about 1 min until about 9 min for the Mgbased sample, then low until 16 min for the MgH 2 -based sample and 11 min for the Mg-based sample, and finally they are very low. The Mg-based sample has a slightly higher maximum dehydriding rate and a larger quantity of hydrogen desorbed for 60 min than the MgH 2 -based sample. The Mgbased sample and the MgH 2 -based sample have maximum dehydriding rates of 0.62 and 0.57 wt% H/min, respectively. Their quantities of hydrogen desorbed for 60 min are 5.45 and 4.96 wt% H, respectively. Figure 9 shows the dehydriding curves at 573 K and then 623 K for the MgH 2 -based sample at n=5 and the Mg-based sample at n=4. In the beginning at 573 K, the Mg-based sample has a slightly lower dehydriding rate, but from about 18 min it has a higher dehydriding rate than the MgH 2 -based sample. The quantities of hydrogen desorbed for 60 min are 2.80 and 2.36 wt%, respectively, for the Mg-based sample and the MgH 2 -based sample. At 623 K, the two samples exhibit very high dehydriding rates of 0.32 and 0.26 wt% H/ min, respectively. The total quantities of hydrogen desorbed from the Mg-based sample and the MgH 2 -based sample are 4.91 and 4.94 wt%, respectively. The results show that, on the whole, the Mg-based sample has a lower dehydriding temperature than the MgH 2 -based sample. Figure 10 shows a SEM micrograph of the as-milled MgH 2 -based sample. Some large particles consist of smaller particles and other large particles have polyhedrons surrounded by flat surfaces. Figure 3 shows that the as-milled Mg-based sample is composed of very fine particles.

6 146 대한금속 재료학회지 제56권 제2호 (2018년 2월) in the reactivity with hydrogen), and create defects, which act as active sites for nucleation, on the surface and in the inside of the Mg particles (leading to facilitation of nucleation). Fig. 3 and Fig. 10 show that, compared with the large particles of as-milled MgH2-based sample, those of the Mg-based sample are composed of finer particles. The Mg-based sample is believed to have more clean surfaces and more defects than the MgH2-based sample. The Mg-based sample has a larger quantity of hydrogen absorbed for 60 min than the MgH2based sample, as shown in Fig. 7. The Mg-based sample has a larger quantity of hydrogen desorbed for 60 min than does Fig. 10. A SEM micrograph of the as-milled MgH2-based sample. the MgH2-based sample, as shown in Fig. 8. These results show that the effects of reactive mechanical grinding are stronger in the Mg-based sample than in the MgH2-based sample; the addition of MgH2, Mg2Ni and Ni to Mg has stronger effects on the increase in the hydriding and dehydriding properties of Mg than the addition of Mg2Ni and Ni to MgH2. The rate-controlling steps, for the hydriding reaction of an activated, mechanically alloyed mixture of Mg 10 wt% Ni at K and bar H2 were the bulk and Knudsen flows through the pores, cracks, and interparticle channels in the ranges of weight percentage of absorbed hydrogen (Ha) smaller than 4.0 < Ha 4.25 [21]. The expansion due to the relatively rapid hydrogen absorption of Mg2Ni in 94MgH2+5Mg2Ni+1Ni is believed to provide the paths for Fig. 11. A SEM micrograph of the as-purchased Mg. the hydrogen released from neighboring MgH2, facilitating the hydrogen absorption of Mg. In the dehydriding reaction Compared with the large particles of the as-milled MgH2- of a mechanically alloyed alloy of 90 wt% Mg + 10 wt% Ni based sample, those of the Mg-based sample are composed [22] after activation, both the bulk and Knudsen flows of finer particles. The Mg-based sample is believed to have control the dehydriding rate in the ranges of weight more clean surfaces and more defects than the MgH2-based percentage of desorbed hydrogen (Hd) higher than 0.5 < Hd sample. 0.1, the Knudsen flow mainly controlling the dehydriding Figure 11 shows a SEM micrograph of as-purchased Mg. rate as the ranges of weight percentage of desorbed hydrogen The particles are rounded in shape. The surfaces of the become higher [22]. The contraction of the Mg2Ni lattice particles are quite flat and have few cracks. owing to the relatively rapid desorption of Mg2NiH4 in the Compared with the particles of as-purchased Mg, shown in Mg-based sample and the MgH2-based sample is believed to Fig. 11, the particles of the as-milled Mg-based sample, provide passages for the hydrogen released from neighboring shown in Fig. 3, are much smaller and have much more MgH2, increasing the hydrogen desorption rate of MgH2. defects. The reactive mechanical grinding of Mg (with Ni is known to act as active chemisorption sites for MgH2, Mg2Ni, and Ni) is believed to reduce the particle size hydrogen [23]. The Ni in the Mg-based sample and the of Mg (leading to decrease in the diffusion distances of MgH2-based sample is believed to contribute to the increase hydrogen atoms), make clean surfaces (leading to an increase in the hydriding and dehydriding rates of Mg by acting as

7 Seong-Hyeon Hong, and Myoung Youp Song 147 active chemisorption sites for hydrogen. If we exclude the 5 wt% Mg 2 Ni and 1 wt% Ni from consideration, the addition of 10 wt% MgH 2 to 84 wt% Mg is thought to bring about stronger effects of reactive mechanical grinding than 94 wt% MgH 2 alone without adding another material. Takacs [24] insisted that mechanical grinding brings about size reduction, mixing, and defect formation. Suryanarayana et al. [25] reported that mechanical grinding in a high-energy ball mill involves cold welding, fracturing, and rewelding of powder particles. Aoyaki et al. [26] reported that ball milling of alloy powders like FeTi, Mg 2 Ni, and LaNi 5 reduces the particle size and creates new clean surfaces. Orimo and Fujii [27] pointed out that mechanical grinding produces nanocrystalline materials, having interesting hydriding properties such as increased hydrogen solubility and increased hydrogen diffusion rate depending on hydrogen content in grain boundaries. Zaluska et al. [15] reported that ball milling of MgH 2 can introduce formation of defects and local imperfections, reduction of grain size, and induction of micro-stress in the structure. Song et al. [28] reported that the main effects of mechanical alloying in a planetary mill on the material are an increase in surface area and the creation of many defects on the surface and in the interior and the expansion and contraction of the lattice during hydriding-dehydriding cycling also stimulates a decrease in the effective particle size and create defects. 4. CONCLUSIONS An Mg-based sample was prepared by milling under a hydrogen atmosphere in a planetary ball mill for 5 h. The hydrogen absorption and release properties of the prepared sample were investigated and compared with those of a MgH 2 -based sample. The Mg-based sample had larger quantities of hydrogen absorbed and released for 60 min than the MgH 2 -based sample. The Mg-based sample had an effective hydrogen-storage capacity of near 5.7 wt%. At n=2, the Mg-based sample absorbed 5.18 wt% for 2 min and 5.67 wt% for 60 min at 648 K. At n=1, the Mg-based sample released wt% for 2 min, 5.13 wt% for 10 min, and 5.49 wt% for 60 min at 648 K. The reactive mechanical grinding of Mg (with MgH 2, Mg 2 Ni, and Ni) is believed to decrease the diffusion distances of hydrogen atoms, increase reactivity with hydrogen, and facilitate nucleation. The addition of MgH 2, Mg 2 Ni, and Ni to Mg was found to have stronger effects on the increase in the hydriding and dehydriding properties of Mg than the addition of Mg 2 Ni and Ni to MgH 2. The contraction of the Mg 2 Ni lattice owing to the relatively rapid desorption of Mg 2 NiH 4 in both samples is believed to provide passages for the hydrogen released from neighboring MgH 2, increasing the hydrogen desorption rate of MgH 2. ACKNOWLEDGEMENTS This research was performed for the Hydrogen Energy R & D Center, one of the 21 st Century Frontier R & D Programs, funded by the Ministry of Science and Technology of Republic of Korea. REFERENCES 1. M.Y. Song, Y. J. Kwak, S. H. Lee, and H. R. Park, Korean J. Met. Mater. 54, 210 (2016). 2. J. J. Reilly and R. H. Wiswall Jr, Inorg. Chem. 7, 2254 (1968). 3. E. Akiba, K. Nomura, S. Ono, and S. Suda, Int. J. Hydrogen Energy 7, 787 (1982). 4. M. Y. Song, J. Mater. Sci. 30, 1343 (1995). 5. M. Y. Song. E. I. Ivanov, B. Darriet, M. Pezat, and P. Hagenmuller, Int. J. Hydrogen Energy 10, 169 (1985). 6. M. Y. Song. E. I. Ivanov, B. Darriet, M. Pezat, and P. Hagenmuller, J. Less-Common Met. 131, 71 (1987). 7. M. Y. Song, Int. J. Hydrogen Energy 20, 221 (1995). 8. J. Schefer, P. Fischer,W. Haelg, F. Stucki, L. Schlapbach, J. J. Didisheim, K. Yvon, and A. F. Andresen, J. Less- Common Met. 74, 74 (1980). 9. S. Ono, H. Hayakawa, A. Suzuki, K. Nomura, N. Nishimiya, and T. Tabata, J. Less-Common Met. 88, 63 (1982). 10. B. Darriet, J.-L. Soubeyroux, M. Pezat, and D. Fruchart, J. Less-Common Met. 103, 153 (1984). 11. M. H. Mintz, Z. Gavra, G. Kimmel, and Z. Hadari, J. Less- Common Met. 74, 263 (1980). 12. E. Akiba, K. Nomura, S. Ono, and Y. Mizuna, J. Less- Common Met. 83, L43 (1989). 13. D. Noreus and L. Kihlborg, J. Less-Common Met. 123, 233

8 148 대한금속 재료학회지제 56 권제 2 호 (2018 년 2 월 ) (1986). 14. J. Cermak and B. David, Int. J. Hydrogen Energy 36, (2011). 15. A. Zaluska, L. Zaluski, and J. O. Ström-Olsen, J. Alloy. Compd. 289, 197 (1999). 16. J.-L. Bobet, E. Akiba, Y. Nakamura, and B. Darriet, Int. J. Hydrogen Energy 25, 987 (2000). 17. Zhenglong, L. Zuyan, and B. Yanbin, J. Alloy. Compd. 470, 470 (2009). 18. N. A. Niaz, I. Ahmad, W.S. Khan, and S. T. Hussain, J. Mater. Sci. Tech. 28, 401 (2012). 19. S.-H. Hong and M.Y. Song, Met. Mater. Int. 22, 1121 (2016). 20. S.-H. Hong and M.Y. Song, Met. Mater. Int. 22, 544 (2016). 21. M. Y. Song, Int. J. Hydrogen Energy 28, 403 (2003). 22. M. Y. Song, J.-P. Manaud, and B. Darriet, J. Alloy. Compd. 282, 243 (1999). 23. M. Y. Song, M. Pezat, B. Darriet, and P. Hagenmuller, J. Mater. Sci. 20, 2958 (1985). 24. L. Takacs, Progress in Materials Science 47, 355 (2002). 25. C. Suryanarayana, E. Ivanov, and V. V. Boldyrev, Materials Science and Engineering: A ,151 (2001). 26. H. Aoyaki, K. Aoki, and T. Masumoto, J. Alloy. Compd. 231, 804 (1995). 27. S.-I. Orimo and H. Fujii, Intermetallics 6, (1998). 28. M. Y. Song, E. Ivanov, B. Darriet, M. Pezat, and P. Hagenmuller, J. Less-Common Met. 131, 71 (1987).