- Email: [email protected]
Preparation of hydrogen-storage alloy Mg–10 wt% Fe2O3 under various milling conditions
International Journal of Hydrogen Energy 31 (2006) 43 – 47 www.elsevier.com/locate/ijhydene
Preparation of hydrogen-storage alloy Mg–10 wt% Fe2 O3 under various milling conditions MyoungYoup Songa,∗ , IkHyun Kwona , SungNam Kwona , ChanGi Parka , HyeRyoung Parkb , Jong-Soo Baec a Division of Advanced Materials Engineering, Research Center of Advanced Materials Development, Engineering Research Institute,
Chonbuk National University, 664-14 1ga Deogjindong Deogjingu Jeonju, Chonbuk, 561-756, Republic of Korea b Faculty of Applied Chemical Engineering, College of Engineering, Chonnam National University, 300 Yongbongdong Bukgu Gwangju,
500-757, Republic of Korea c Nanopowder Materials Group, Materials Technology Department, Korea Institute of Machinery & Materials, 66 Sangnamdong,
Changwon, Kyungnam, 641-010, Republic of Korea Available online 18 April 2005
Abstract We tried to improve the H2 -sorption properties of Mg by mechanical grinding under H2 (reactive mechanical grinding) at various weight ratios of sample to ball (1/9, 1/27 and 1/45) with 10 wt% Fe2 O3 . The revolution speed was 250 rpm and the milling time was 2 h. The sample Mg–10 wt% Fe2 O3 , prepared by milling at the weight ratio of sample to ball 1/45, has the best hydrogen-storage properties. It absorbs 5.56 wt% hydrogen at the first cycle at 593 K under 12 bar H2 for 60 min. Its activation is accomplished after two hydriding–dehydriding cycles. The activated sample absorbs 4.26 wt% hydrogen at 593 K, 12 bar H2 for 10 min. The reactive grinding of Mg with Fe2 O3 increases the H2 -sorption rates by facilitating nucleation (by creating defects on the surface of the Mg particles and by the additive), by making cracks on the surface of Mg particles and reducing the particle size of Mg and thus by shortening the diffusion distances of hydrogen atoms. Hydriding–dehydriding cycling also increases the H2 -sorption rates by creating defects on the surface of the Mg particles, and by making cracks on the surface of Mg particles and reducing the particle size of Mg. 䉷 2005 International Association for Hydrogen Energy. Published by Elsevier Ltd. All rights reserved. Keywords: H2 -sorption properties of Mg; Fe2 O3 addition; Reactive mechanical grinding; Sample-to-ball ratio; Hydriding rates and dehydriding rates; BET surface area
1. Introduction Magnesium has many advantages as a hydrogen storage material: large hydrogen storage capacity (7.6 wt%), low cost and abundance in the earth’s crust. But its hydriding and dehydriding kinetics are very slow [1]. A lot of work to ameliorate the reaction kinetics of magnesium with
∗ Corresponding author. Tel.: +82 63 270 2379; fax: +82 63 270 2386. E-mail address: [email protected] (M.Y. Song).
hydrogen has been done by alloying certain metals with magnesium [2–10], by mixing metal additives with magnesium [11], and by plating nickel on the surface of magnesium [12]. Song [13] reviewed the kinetic studies of the hydriding and the dehydriding reactions of Mg. Many works do not agree with one another on the rate-controlling step(s) for hydriding or dehydriding of magnesium. However, there is no contradiction in the points that the hydriding and dehydriding reactions of Mg are nucleation-controlled under certain conditions and progress by a mechanism of nucleation and growth, and that the hydriding rates of Mg are
0360-3199/$30.00 䉷 2005 International Association for Hydrogen Energy. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.ijhydene.2005.03.008
M.Y. Song et al. / International Journal of Hydrogen Energy 31 (2006) 43 – 47
controlled by the diffusion of hydrogen through a growing Mg hydride layer. The hydriding and dehydriding kinetics of Mg can be improved, therefore, by a treatment such as mechanical alloying [14–16], which can facilitate nucleation by creating many defects on the surface and/or in the interior of Mg, or by the additive acting as active sites for the nucleation, and shortening diffusion distances by reducing the effective particle sizes of Mg. Bobet et al. [17] reported that mechanical alloying in H2 (reactive mechanical grinding) for a short time (2 h) is an effective way to improve strongly the hydrogen-storage properties of both magnesium and Mg+10 wt% Co, Ni or Fe mixtures. Klassen et al. [18], Huot et al. [19] and Imamura et al. [20] improved hydrogen-sorption kinetics of magnesium by adding metal, oxide or carbon through mechanical grinding. The oxides are brittle, and thus they may be pulverized during mechanical grinding. The added oxides and/or their pulverization during mechanical grinding may help the particles of magnesium become finer. In this work, we have chosen Fe2 O3 as the oxide to add to Mg. Mg–10 wt% Fe2 O3 alloys were prepared by mechanical grinding under H2 (reactive mechanical grinding) in a planetary ball mill with various weight ratios of sample to ball at a revolution speed 250 rpm and for a milling time 2 h, and their hydrogen-storage properties were investigated.
2. Experimental details Pure Mg powder (particle size 50 m) was mixed with 10 wt% Fe2 O3 (total weight = 8 g) in a stainless steel container (with hardened steel balls, total weight = 72, 216 and 360 g) closed in a hermetic way. The average particle size of Fe2 O3 was smaller than 5 m. All the handlings were performed in a glove box under Ar in order to prevent oxidation. The mill container was then filled with high purity hydrogen gas (≈ 10 bar). The disc revolution speed was 250 rpm and the milling time was 2 h. The hydriding apparatus has been described previously [21]. The absorbed or desorbed hydrogen quantity was measured as a function of time by a volumetric method. X-ray diffraction (XRD) analysis was carried out for the as-milled powders. The microstructures were observed by scanning electron microscope (SEM). Brunauer, Emmett and Teller (BET) surface areas of the as-milled powders were measured at 77 K using a Micromeritics ASAP 2020 automatic analyzer. The samples were outgassed at 573 K for 6 h before measuring the surface areas.
3. Results and discussion Fig. 1 shows the X-ray (CuK ) powder diffraction patterns of Mg–10 wt% Fe2 O3 as prepared by milling
Mg
unmarked
Fe2O3
intensity (arbitrary unit)
44
MgH2
1/45
1/27
1/9
10
20
30
40
50
60
70
80
2θ (degree)
Fig. 1. X-ray (CuK ) powder diffraction patterns of Mg–10 wt% Fe2 O3 as prepared at 250 rpm for 2 h with various weight ratios of sample to ball (1/9, 1/27 and 1/45).
at the revolution speed of 250 rpm for 2 h with various weight ratios of sample to ball (1/9, 1/27 and 1/45). The XRD patterns are similar and the peaks become stronger as the weight ratio of sample to ball decreases. It is considered that the increase in peak intensity results from more effective decrease in particle size as the weight ratio of sample to ball decreases. The samples as prepared contain Mg, Fe2 O3 and a small amount of MgH2 . Fig. 2 gives the microstructures observed by SEM for Mg–10 wt% Fe2 O3 as prepared by milling at the revolution speed 250 rpm for 2 h with various weight ratios of sample to ball (a) 1/9, (b) 1/27 and (c) 1/45. The samples have small particles and large particles. As the milling time increases, the number of small particles increases. The sample prepared with the weight ratio of sample to ball 1/45 has very small particles. Fig. 3 shows the variations of weight percentage of hydrogen Ha absorbed at 593 K under 12 bar H2 for the first cycle of Mg–10 wt% Fe2 O3 prepared at a revolution speed of 250 rpm for 2 h with various weight ratios of sample to ball. The percentages of absorbed hydrogen Ha are expressed with respect to the sample weight. The samples have the higher hydriding rates as the weight ratio of sample to ball increases. The sample milled with the weight ratio of sample to ball 1/45 has the highest hydriding rates. It absorbs 5.56 wt% hydrogen at 593 K under 12 bar H2 for 60 min. We studied the hydrogen-storage properties of the Mg–10 wt% Fe2 O3 milled with the weight ratio of sample to ball 1/45 in detail. Fig. 4 shows Ha versus t curves at 593 K, 12 bar H2 for the activated Mg–10 wt% Fe2 O3 prepared at a revolution speed of 250 rpm for 2 h with various weight ratios of sample to ball. The sample milled with the weight ratio of sample to ball 1/45 has the highest hydriding rates. It absorbs 4.26 wt% hydrogen for 10 min, and 5.16 wt% hydrogen for 60 min at 593 K, 12 bar H2 .
M.Y. Song et al. / International Journal of Hydrogen Energy 31 (2006) 43 – 47
45
Fig. 2. Microstructure observed by SEM for Mg–10 wt% Fe2 O3 as prepared at 250 rpm for 2 h with various weight ratios of sample to ball [(a) 1/9, (b) 1/27 and (c) 1/45].
1/45 1/27 1/9
70
4.5 4.0 3.5 3.0 2.5 2.0 1.5
60
4
50 40 2 30
Ha (60 min) (wt.%)
5.0
BET surface area (m2/g)
5.5
Ha (wt.%)
6
80
6.0
20 0
1.0
10
0.5 0.0
1/45
-0.5 0
10
20
30
40
50
1/9
60
t (min)
Fig. 3. Ha versus t curves at 593 K under 12 bar H2 for the first cycle of Mg–10 wt% Fe2 O3 prepared at a revolution speed 250 rpm for 2 h with various weight ratios of sample to ball.
6.0 5.5
1/45 1/27 1/9
5.0 4.5 4.0
Ha (wt.%)
1/27
weight ratio of sample to ball
3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 -0.5 0
10
20
30
40
50
60
t (min)
Fig. 4. Ha versus t curves at 593 K under 12 bar H2 for the activated Mg–10 wt% Fe2 O3 prepared at a revolution speed 250 rpm for 2 h with various weight ratios of sample to ball.
Fig. 5. Variations of the quantity of hydrogen Ha (60 min) absorbed at 593 K under 12 bar H2 after activation and BET surface area with the weight ratio of sample to ball.
Fig. 5 shows the variations of the quantity of hydrogen Ha (60 min) absorbed at 593 K under 12 bar H2 after activation and BET surface area with the weight ratio of sample to ball. Ha (60 min) decreases and BET surface area decreases roughly as the weight ratio of sample to ball becomes smaller, indicating that Ha (60 min) increases as BET surface area increases. Fig. 6 shows the variations of Ha absorbed by Mg–10 wt% Fe2 O3 milled with the weight ratio of sample to ball 1/45 as a function of time t (min) at 593 K, 12 bar H2 according to the number of hydriding–dehydriding cycles n. As the number of hydriding–dehydriding cycles increases, the hydriding rates increase in the beginning of hydriding reaction. It is considered that the activation is accomplished after the second hydriding cycle. The Ha value after 10 min is 4.26 wt% at n = 2. The reactive grinding of Mg with Fe2 O3 increases the hydriding rates by facilitating nucleation (by creating defects on the surface of the Mg particles and by the additive) and by reducing the particle size of Mg and thus by shortening the diffusion distances of hydrogen atoms.
46
M.Y. Song et al. / International Journal of Hydrogen Energy 31 (2006) 43 – 47 6 6 5
5
4
Ha (wt.%)
Ha (wt.%)
4 3 2 1
0
10
20
30
40
50
2 583K 12 bar 11 bar 10 bar
1
n=1 n=2 n=3
0
3
0
60
0
10
20
t (min)
Fig. 6. Variations of Ha versus t curves of Mg–10 wt% Fe2 O3 milled with the weight ratio of sample to ball 1/45 at 593 K, 12 bar H2 according to the number of cycles n.
40
50
60
Fig. 8. Ha versus t curves under various hydrogen pressures at 583 K for the activated Mg–10 wt% Fe2 O3 milled with the weight ratio of sample to ball 1/45.
6
5
5
4
4
3
3 593K 12 bar 11 bar 10 bar 9 bar 8 bar
2
1
0 0
10
20
30
40
50
60
Ha (wt. %)
Ha (wt.%)
30
t (time)
2
573K 12 bar 11 bar 10 bar
1
0 0
Fig. 7. Ha versus t curves under various hydrogen pressures at 593 K for the activated Mg–10 wt% Fe2 O3 milled with the weight ratio of sample to ball 1/45.
Fig. 7 shows the Ha versus t curves at 593 K under various hydrogen pressures for the activated Mg–10 wt% Fe2 O3 milled with the weight ratio of sample to ball 1/45. As the hydrogen pressure increases, the hydriding rate becomes greater. Figs. 8 and 9 show the Ha versus t curves under various hydrogen pressures at 583 K and 573 K, respectively, for the activated Mg–10 wt% Fe2 O3 milled with the weight ratio of sample to ball 1/45. Fig. 10 gives the variations in the weight percentage of the desorbed hydrogen with respect to the sample weight, Hd (wt%), at 593 K under 1.0–1.8 bar H2 by the activated Mg–10 wt% Fe2 O3 milled with the weight ratio of sample to ball 1/45. Before obtaining these curves, the sample was hydrided under 12 bar H2 for 2 h at 593 K. As the hydrogen
10
20
30
40
50
60
t (min)
t (time)
Fig. 9. Ha versus t curves under various hydrogen pressures at 573 K for the activated Mg–10 wt% Fe2 O3 milled with the weight ratio of sample to ball 1/45.
pressure decreases, the dehydriding rate becomes greater. The hydrided sample desorbs 0.88 wt% hydrogen at 593 K, 1.0 bar H2 for 60 min. From the results of hydriding rates and microstructure, it is considered that the hydriding–dehydriding cycling also increases the H2 -sorption rates by facilitating nucleation (by creating defects on the surface of the Mg particles), by making cracks on the surface of Mg particles and reducing the particle size of Mg and thus by shortening the diffusion distances of hydrogen atoms.
4. Conclusions The samples Mg–10 wt% Fe2 O3 were prepared by mechanical grinding under H2 (reactive mechanical grind) with
M.Y. Song et al. / International Journal of Hydrogen Energy 31 (2006) 43 – 47
References
1.0 1.0 bar 1.2 bar 1.4 bar 1.6 bar 1.8 bar
0.8
0.6
Hd (wt. %)
47
[1] [2] [3] [4] [5]
0.4
0.2
0.0 0
10
20
30
40
50
60
t (min)
Fig. 10. Hd versus t curves under various hydrogen pressures at 593 K for the activated Mg–10 wt% Fe2 O3 milled with the weight ratio of sample to ball 1/45.
various weight ratios of sample to ball (1/9, 1/27 and 1/45) at a revolution speed 250 rpm for 2 h. The sample Mg–10 wt% Fe2 O3 prepared by milling with the weight ratio of sample to ball 1/45 has the best hydrogen-storage properties. It absorbs 5.56 wt% hydrogen at the first cycle at 593 K under 12 bar H2 for 60 min. Its activation is accomplished after two hydriding–dehydriding cycles. The activated sample absorbs 4.26 wt% hydrogen at 593 K, 12 bar H2 for 10 min. The reactive grinding of Mg with Fe2 O3 increases the H2 -sorption rates by facilitating nucleation (by creating defects on the surface of the Mg particles and by the additive), by making cracks on the surface of Mg particles and reducing the particle size of Mg and thus by shortening the diffusion distances of hydrogen atoms. Hydriding–dehydriding cycling also increases the H2 -sorption rates by creating defects on the surface of the Mg particles, and by making cracks on the surface of Mg particles and reducing the particle size of Mg.
Acknowledgements This research was performed for the Hydrogen Energy R&D Center, one of the 21st Century Frontier R&D Program, funded by the Ministry of Science and Technology.
Vose A. Metal hydrides. US Patent 2 1961;944:587. Reilly JJ, Wiswall RH. Inorg Chem 1967;6(12):2220. Reilly JJ, Wiswall Jr RH. Inorg Chem 1968;7(11):2254. Douglass DL. Metall Trans 1975;A 6:2179. Mintz MH, Gavra Z, Hadari Z. J Inorg Nucl Chem 1978;40:765. [6] Pezat M, Hbika A, Darriet B, Hagenmuller P. French Anvar Patent 78 203 82 1978. Mater Res Bull 1979;14:377. [7] Pezat M, Darriet B, Hagenmuller P. J Less-Common Met 1980;74:427. [8] Wang Q, Wu J, Au M, Zhang L. In: Veziro˘glu TN, Taylor JB, editors. Hydrogen energy progress V, Proceedings of the fifth world hydrogen energy conference, Toronto, Canada, July 1984, vol. 3. New York: Pergamon; p. 1279–90. [9] Akiba E, Nomura K, Ono S, Suda S. Int J Hydrogen Energy 1982;7(10):787. [10] Boulet JM, Gerard N. J Less-Common Met 1983;89:151. [11] Tanguy B, Soubeyroux JL, Pezat M, Portier J, Hagenmuller P. Mater Res Bull 1976;11:1441. [12] Eisenberg FG, Zagnoli DA, Sheridan III JJ. J Less-Common Met 1980;74:323. [13] Song MY. J Mater Sci 1995;30:1343. [14] Song MY, Ivanov EI, Darriet B, Pezat M, Hagenmuller P. Int J Hydrogen Energy 1985;10(3):169. [15] Song MY, Ivanov EI, Darriet B, Pezat M, Hagenmuller P. J Less-Common Met 1987;131:71. [16] Song MY. Int J Hydrogen Energy 1995;20(3):221. [17] Bobet J-L, Akiba E, Nakamura Y, Darriet B. Int J Hydrogen Energy 2000;25:987. [18] Yavari AR, LeMoulec A, de Castro FR, Deledda S, Friedrichs O, Botta WJ, Vaughan G, Klassen T, Fernandez A, Kvick A. Scripta Materialia 2005;52(8):719. [19] Huot J, Tremblay M-L, Schulz R. J Alloys and Compounds 2003;356–357:603. [20] Imamura H, Kusuhara M, Minami S, Matsumoto M, Masanari K, Sakata Y, Itoh K, Fukunaga T. Acta Materialia 2003;51(20):6407. [21] Song MY, Ahn DS, Kwon IH, Ahn HJ. Metals Mater Int 1999;5(5):485.
Preparation of hydrogen-storage alloy Mg–10 wt% Fe2 O3 under various milling conditions MyoungYoup Songa,∗ , IkHyun Kwona , SungNam Kwona , ChanGi Parka , HyeRyoung Parkb , Jong-Soo Baec a Division of Advanced Materials Engineering, Research Center of Advanced Materials Development, Engineering Research Institute,
Chonbuk National University, 664-14 1ga Deogjindong Deogjingu Jeonju, Chonbuk, 561-756, Republic of Korea b Faculty of Applied Chemical Engineering, College of Engineering, Chonnam National University, 300 Yongbongdong Bukgu Gwangju,
500-757, Republic of Korea c Nanopowder Materials Group, Materials Technology Department, Korea Institute of Machinery & Materials, 66 Sangnamdong,
Changwon, Kyungnam, 641-010, Republic of Korea Available online 18 April 2005
Abstract We tried to improve the H2 -sorption properties of Mg by mechanical grinding under H2 (reactive mechanical grinding) at various weight ratios of sample to ball (1/9, 1/27 and 1/45) with 10 wt% Fe2 O3 . The revolution speed was 250 rpm and the milling time was 2 h. The sample Mg–10 wt% Fe2 O3 , prepared by milling at the weight ratio of sample to ball 1/45, has the best hydrogen-storage properties. It absorbs 5.56 wt% hydrogen at the first cycle at 593 K under 12 bar H2 for 60 min. Its activation is accomplished after two hydriding–dehydriding cycles. The activated sample absorbs 4.26 wt% hydrogen at 593 K, 12 bar H2 for 10 min. The reactive grinding of Mg with Fe2 O3 increases the H2 -sorption rates by facilitating nucleation (by creating defects on the surface of the Mg particles and by the additive), by making cracks on the surface of Mg particles and reducing the particle size of Mg and thus by shortening the diffusion distances of hydrogen atoms. Hydriding–dehydriding cycling also increases the H2 -sorption rates by creating defects on the surface of the Mg particles, and by making cracks on the surface of Mg particles and reducing the particle size of Mg. 䉷 2005 International Association for Hydrogen Energy. Published by Elsevier Ltd. All rights reserved. Keywords: H2 -sorption properties of Mg; Fe2 O3 addition; Reactive mechanical grinding; Sample-to-ball ratio; Hydriding rates and dehydriding rates; BET surface area
1. Introduction Magnesium has many advantages as a hydrogen storage material: large hydrogen storage capacity (7.6 wt%), low cost and abundance in the earth’s crust. But its hydriding and dehydriding kinetics are very slow [1]. A lot of work to ameliorate the reaction kinetics of magnesium with
∗ Corresponding author. Tel.: +82 63 270 2379; fax: +82 63 270 2386. E-mail address: [email protected] (M.Y. Song).
hydrogen has been done by alloying certain metals with magnesium [2–10], by mixing metal additives with magnesium [11], and by plating nickel on the surface of magnesium [12]. Song [13] reviewed the kinetic studies of the hydriding and the dehydriding reactions of Mg. Many works do not agree with one another on the rate-controlling step(s) for hydriding or dehydriding of magnesium. However, there is no contradiction in the points that the hydriding and dehydriding reactions of Mg are nucleation-controlled under certain conditions and progress by a mechanism of nucleation and growth, and that the hydriding rates of Mg are
0360-3199/$30.00 䉷 2005 International Association for Hydrogen Energy. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.ijhydene.2005.03.008
M.Y. Song et al. / International Journal of Hydrogen Energy 31 (2006) 43 – 47
controlled by the diffusion of hydrogen through a growing Mg hydride layer. The hydriding and dehydriding kinetics of Mg can be improved, therefore, by a treatment such as mechanical alloying [14–16], which can facilitate nucleation by creating many defects on the surface and/or in the interior of Mg, or by the additive acting as active sites for the nucleation, and shortening diffusion distances by reducing the effective particle sizes of Mg. Bobet et al. [17] reported that mechanical alloying in H2 (reactive mechanical grinding) for a short time (2 h) is an effective way to improve strongly the hydrogen-storage properties of both magnesium and Mg+10 wt% Co, Ni or Fe mixtures. Klassen et al. [18], Huot et al. [19] and Imamura et al. [20] improved hydrogen-sorption kinetics of magnesium by adding metal, oxide or carbon through mechanical grinding. The oxides are brittle, and thus they may be pulverized during mechanical grinding. The added oxides and/or their pulverization during mechanical grinding may help the particles of magnesium become finer. In this work, we have chosen Fe2 O3 as the oxide to add to Mg. Mg–10 wt% Fe2 O3 alloys were prepared by mechanical grinding under H2 (reactive mechanical grinding) in a planetary ball mill with various weight ratios of sample to ball at a revolution speed 250 rpm and for a milling time 2 h, and their hydrogen-storage properties were investigated.
2. Experimental details Pure Mg powder (particle size 50 m) was mixed with 10 wt% Fe2 O3 (total weight = 8 g) in a stainless steel container (with hardened steel balls, total weight = 72, 216 and 360 g) closed in a hermetic way. The average particle size of Fe2 O3 was smaller than 5 m. All the handlings were performed in a glove box under Ar in order to prevent oxidation. The mill container was then filled with high purity hydrogen gas (≈ 10 bar). The disc revolution speed was 250 rpm and the milling time was 2 h. The hydriding apparatus has been described previously [21]. The absorbed or desorbed hydrogen quantity was measured as a function of time by a volumetric method. X-ray diffraction (XRD) analysis was carried out for the as-milled powders. The microstructures were observed by scanning electron microscope (SEM). Brunauer, Emmett and Teller (BET) surface areas of the as-milled powders were measured at 77 K using a Micromeritics ASAP 2020 automatic analyzer. The samples were outgassed at 573 K for 6 h before measuring the surface areas.
3. Results and discussion Fig. 1 shows the X-ray (CuK ) powder diffraction patterns of Mg–10 wt% Fe2 O3 as prepared by milling
Mg
unmarked
Fe2O3
intensity (arbitrary unit)
44
MgH2
1/45
1/27
1/9
10
20
30
40
50
60
70
80
2θ (degree)
Fig. 1. X-ray (CuK ) powder diffraction patterns of Mg–10 wt% Fe2 O3 as prepared at 250 rpm for 2 h with various weight ratios of sample to ball (1/9, 1/27 and 1/45).
at the revolution speed of 250 rpm for 2 h with various weight ratios of sample to ball (1/9, 1/27 and 1/45). The XRD patterns are similar and the peaks become stronger as the weight ratio of sample to ball decreases. It is considered that the increase in peak intensity results from more effective decrease in particle size as the weight ratio of sample to ball decreases. The samples as prepared contain Mg, Fe2 O3 and a small amount of MgH2 . Fig. 2 gives the microstructures observed by SEM for Mg–10 wt% Fe2 O3 as prepared by milling at the revolution speed 250 rpm for 2 h with various weight ratios of sample to ball (a) 1/9, (b) 1/27 and (c) 1/45. The samples have small particles and large particles. As the milling time increases, the number of small particles increases. The sample prepared with the weight ratio of sample to ball 1/45 has very small particles. Fig. 3 shows the variations of weight percentage of hydrogen Ha absorbed at 593 K under 12 bar H2 for the first cycle of Mg–10 wt% Fe2 O3 prepared at a revolution speed of 250 rpm for 2 h with various weight ratios of sample to ball. The percentages of absorbed hydrogen Ha are expressed with respect to the sample weight. The samples have the higher hydriding rates as the weight ratio of sample to ball increases. The sample milled with the weight ratio of sample to ball 1/45 has the highest hydriding rates. It absorbs 5.56 wt% hydrogen at 593 K under 12 bar H2 for 60 min. We studied the hydrogen-storage properties of the Mg–10 wt% Fe2 O3 milled with the weight ratio of sample to ball 1/45 in detail. Fig. 4 shows Ha versus t curves at 593 K, 12 bar H2 for the activated Mg–10 wt% Fe2 O3 prepared at a revolution speed of 250 rpm for 2 h with various weight ratios of sample to ball. The sample milled with the weight ratio of sample to ball 1/45 has the highest hydriding rates. It absorbs 4.26 wt% hydrogen for 10 min, and 5.16 wt% hydrogen for 60 min at 593 K, 12 bar H2 .
M.Y. Song et al. / International Journal of Hydrogen Energy 31 (2006) 43 – 47
45
Fig. 2. Microstructure observed by SEM for Mg–10 wt% Fe2 O3 as prepared at 250 rpm for 2 h with various weight ratios of sample to ball [(a) 1/9, (b) 1/27 and (c) 1/45].
1/45 1/27 1/9
70
4.5 4.0 3.5 3.0 2.5 2.0 1.5
60
4
50 40 2 30
Ha (60 min) (wt.%)
5.0
BET surface area (m2/g)
5.5
Ha (wt.%)
6
80
6.0
20 0
1.0
10
0.5 0.0
1/45
-0.5 0
10
20
30
40
50
1/9
60
t (min)
Fig. 3. Ha versus t curves at 593 K under 12 bar H2 for the first cycle of Mg–10 wt% Fe2 O3 prepared at a revolution speed 250 rpm for 2 h with various weight ratios of sample to ball.
6.0 5.5
1/45 1/27 1/9
5.0 4.5 4.0
Ha (wt.%)
1/27
weight ratio of sample to ball
3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 -0.5 0
10
20
30
40
50
60
t (min)
Fig. 4. Ha versus t curves at 593 K under 12 bar H2 for the activated Mg–10 wt% Fe2 O3 prepared at a revolution speed 250 rpm for 2 h with various weight ratios of sample to ball.
Fig. 5. Variations of the quantity of hydrogen Ha (60 min) absorbed at 593 K under 12 bar H2 after activation and BET surface area with the weight ratio of sample to ball.
Fig. 5 shows the variations of the quantity of hydrogen Ha (60 min) absorbed at 593 K under 12 bar H2 after activation and BET surface area with the weight ratio of sample to ball. Ha (60 min) decreases and BET surface area decreases roughly as the weight ratio of sample to ball becomes smaller, indicating that Ha (60 min) increases as BET surface area increases. Fig. 6 shows the variations of Ha absorbed by Mg–10 wt% Fe2 O3 milled with the weight ratio of sample to ball 1/45 as a function of time t (min) at 593 K, 12 bar H2 according to the number of hydriding–dehydriding cycles n. As the number of hydriding–dehydriding cycles increases, the hydriding rates increase in the beginning of hydriding reaction. It is considered that the activation is accomplished after the second hydriding cycle. The Ha value after 10 min is 4.26 wt% at n = 2. The reactive grinding of Mg with Fe2 O3 increases the hydriding rates by facilitating nucleation (by creating defects on the surface of the Mg particles and by the additive) and by reducing the particle size of Mg and thus by shortening the diffusion distances of hydrogen atoms.
46
M.Y. Song et al. / International Journal of Hydrogen Energy 31 (2006) 43 – 47 6 6 5
5
4
Ha (wt.%)
Ha (wt.%)
4 3 2 1
0
10
20
30
40
50
2 583K 12 bar 11 bar 10 bar
1
n=1 n=2 n=3
0
3
0
60
0
10
20
t (min)
Fig. 6. Variations of Ha versus t curves of Mg–10 wt% Fe2 O3 milled with the weight ratio of sample to ball 1/45 at 593 K, 12 bar H2 according to the number of cycles n.
40
50
60
Fig. 8. Ha versus t curves under various hydrogen pressures at 583 K for the activated Mg–10 wt% Fe2 O3 milled with the weight ratio of sample to ball 1/45.
6
5
5
4
4
3
3 593K 12 bar 11 bar 10 bar 9 bar 8 bar
2
1
0 0
10
20
30
40
50
60
Ha (wt. %)
Ha (wt.%)
30
t (time)
2
573K 12 bar 11 bar 10 bar
1
0 0
Fig. 7. Ha versus t curves under various hydrogen pressures at 593 K for the activated Mg–10 wt% Fe2 O3 milled with the weight ratio of sample to ball 1/45.
Fig. 7 shows the Ha versus t curves at 593 K under various hydrogen pressures for the activated Mg–10 wt% Fe2 O3 milled with the weight ratio of sample to ball 1/45. As the hydrogen pressure increases, the hydriding rate becomes greater. Figs. 8 and 9 show the Ha versus t curves under various hydrogen pressures at 583 K and 573 K, respectively, for the activated Mg–10 wt% Fe2 O3 milled with the weight ratio of sample to ball 1/45. Fig. 10 gives the variations in the weight percentage of the desorbed hydrogen with respect to the sample weight, Hd (wt%), at 593 K under 1.0–1.8 bar H2 by the activated Mg–10 wt% Fe2 O3 milled with the weight ratio of sample to ball 1/45. Before obtaining these curves, the sample was hydrided under 12 bar H2 for 2 h at 593 K. As the hydrogen
10
20
30
40
50
60
t (min)
t (time)
Fig. 9. Ha versus t curves under various hydrogen pressures at 573 K for the activated Mg–10 wt% Fe2 O3 milled with the weight ratio of sample to ball 1/45.
pressure decreases, the dehydriding rate becomes greater. The hydrided sample desorbs 0.88 wt% hydrogen at 593 K, 1.0 bar H2 for 60 min. From the results of hydriding rates and microstructure, it is considered that the hydriding–dehydriding cycling also increases the H2 -sorption rates by facilitating nucleation (by creating defects on the surface of the Mg particles), by making cracks on the surface of Mg particles and reducing the particle size of Mg and thus by shortening the diffusion distances of hydrogen atoms.
4. Conclusions The samples Mg–10 wt% Fe2 O3 were prepared by mechanical grinding under H2 (reactive mechanical grind) with
M.Y. Song et al. / International Journal of Hydrogen Energy 31 (2006) 43 – 47
References
1.0 1.0 bar 1.2 bar 1.4 bar 1.6 bar 1.8 bar
0.8
0.6
Hd (wt. %)
47
[1] [2] [3] [4] [5]
0.4
0.2
0.0 0
10
20
30
40
50
60
t (min)
Fig. 10. Hd versus t curves under various hydrogen pressures at 593 K for the activated Mg–10 wt% Fe2 O3 milled with the weight ratio of sample to ball 1/45.
various weight ratios of sample to ball (1/9, 1/27 and 1/45) at a revolution speed 250 rpm for 2 h. The sample Mg–10 wt% Fe2 O3 prepared by milling with the weight ratio of sample to ball 1/45 has the best hydrogen-storage properties. It absorbs 5.56 wt% hydrogen at the first cycle at 593 K under 12 bar H2 for 60 min. Its activation is accomplished after two hydriding–dehydriding cycles. The activated sample absorbs 4.26 wt% hydrogen at 593 K, 12 bar H2 for 10 min. The reactive grinding of Mg with Fe2 O3 increases the H2 -sorption rates by facilitating nucleation (by creating defects on the surface of the Mg particles and by the additive), by making cracks on the surface of Mg particles and reducing the particle size of Mg and thus by shortening the diffusion distances of hydrogen atoms. Hydriding–dehydriding cycling also increases the H2 -sorption rates by creating defects on the surface of the Mg particles, and by making cracks on the surface of Mg particles and reducing the particle size of Mg.
Acknowledgements This research was performed for the Hydrogen Energy R&D Center, one of the 21st Century Frontier R&D Program, funded by the Ministry of Science and Technology.
Vose A. Metal hydrides. US Patent 2 1961;944:587. Reilly JJ, Wiswall RH. Inorg Chem 1967;6(12):2220. Reilly JJ, Wiswall Jr RH. Inorg Chem 1968;7(11):2254. Douglass DL. Metall Trans 1975;A 6:2179. Mintz MH, Gavra Z, Hadari Z. J Inorg Nucl Chem 1978;40:765. [6] Pezat M, Hbika A, Darriet B, Hagenmuller P. French Anvar Patent 78 203 82 1978. Mater Res Bull 1979;14:377. [7] Pezat M, Darriet B, Hagenmuller P. J Less-Common Met 1980;74:427. [8] Wang Q, Wu J, Au M, Zhang L. In: Veziro˘glu TN, Taylor JB, editors. Hydrogen energy progress V, Proceedings of the fifth world hydrogen energy conference, Toronto, Canada, July 1984, vol. 3. New York: Pergamon; p. 1279–90. [9] Akiba E, Nomura K, Ono S, Suda S. Int J Hydrogen Energy 1982;7(10):787. [10] Boulet JM, Gerard N. J Less-Common Met 1983;89:151. [11] Tanguy B, Soubeyroux JL, Pezat M, Portier J, Hagenmuller P. Mater Res Bull 1976;11:1441. [12] Eisenberg FG, Zagnoli DA, Sheridan III JJ. J Less-Common Met 1980;74:323. [13] Song MY. J Mater Sci 1995;30:1343. [14] Song MY, Ivanov EI, Darriet B, Pezat M, Hagenmuller P. Int J Hydrogen Energy 1985;10(3):169. [15] Song MY, Ivanov EI, Darriet B, Pezat M, Hagenmuller P. J Less-Common Met 1987;131:71. [16] Song MY. Int J Hydrogen Energy 1995;20(3):221. [17] Bobet J-L, Akiba E, Nakamura Y, Darriet B. Int J Hydrogen Energy 2000;25:987. [18] Yavari AR, LeMoulec A, de Castro FR, Deledda S, Friedrichs O, Botta WJ, Vaughan G, Klassen T, Fernandez A, Kvick A. Scripta Materialia 2005;52(8):719. [19] Huot J, Tremblay M-L, Schulz R. J Alloys and Compounds 2003;356–357:603. [20] Imamura H, Kusuhara M, Minami S, Matsumoto M, Masanari K, Sakata Y, Itoh K, Fukunaga T. Acta Materialia 2003;51(20):6407. [21] Song MY, Ahn DS, Kwon IH, Ahn HJ. Metals Mater Int 1999;5(5):485.