Hydriding–dehydriding characteristics of NdNi5 and effects of Sn-substitution

Hydriding–dehydriding characteristics of NdNi5 and effects of Sn-substitution

Journal of Alloys and Compounds 297 (2000) 73–80 L www.elsevier.com / locate / jallcom Hydriding–dehydriding characteristics of NdNi 5 and effects ...

941KB Sizes 2 Downloads 28 Views

Journal of Alloys and Compounds 297 (2000) 73–80

L

www.elsevier.com / locate / jallcom

Hydriding–dehydriding characteristics of NdNi 5 and effects of Sn-substitution Yasuyuki Takaguchi, Kazuhide Tanaka* Department of Materials Science and Engineering, Nagoya Institute of Technology, Showa-ku, Nagoya 466 -8555, Japan Received 26 May 1999; received in revised form 7 August 1999

Abstract The pressure–composition–temperature relations and the reaction kinetics for the NdNi 52s Sns (s50, 0.1, 0.2, 0.4)–H 2 system are measured at temperatures between 253 and 293 K. The PCT for NdNi 5 , which presents two well-separated plateau regions at rather high pressures, is strongly altered by the Sn-substitution. Namely, both plateau pressures are continually reduced with increasing substitution and the relative separation is reduced to cause a nearly single sloped plateau without reducing the maximum hydrogen capacity appreciably (H / M|1.0). Furthermore, the hydriding and dehydriding kinetics are substantially improved by the Sn-substitution. The promotion of the reaction kinetics is related with increased tendency toward pulverization during hydriding–dehydriding cycles. X-ray diffraction study suggests that NdNi 5 is structurally stabilized by the Sn-substitution against decomposition during the H–D cycles. In conclusion, NdNi 4.8 Sn 0.2 appears to be favorable for use as a hydrogen storage alloy in the above temperature range.  2000 Elsevier Science S.A. All rights reserved. Keywords: NdNi 52s Sn s alloys; PCT relation; Hydriding–dehydriding kinetics; X-ray diffraction; SEM observation

1. Introduction The intermetallic compound NdNi 5 has a CaCu 5 -type hexagonal crystal structure belonging to the P6 / mmm space group. It is reported that this compound has a hydrogen-storage capacity as large as the isostructural compound LaNi 5 , and a dissociation pressure of the corresponding hydride at least one order of magnitude higher than that of the latter at room temperature [1]. Although it appears to be fundamentally interesting and technologically important to reveal a further detail of the reactivity of NdNi 5 with hydrogen, little has been known on its pressure–composition isotherms (PCT), hydriding– dehydriding (H–D) kinetics and effects of substitution of ternary elements on these properties, in contrast to those for LaNi 5 and Mm(mischmetal)Ni 5 , for which the related data have been amply accumulated to date [2]. The objective of this study is to investigate the PCT and hydriding–dehydriding kinetics of NdNi 5 and examine the effects of partially substituting Sn for Ni on these properties. It has been shown that a partial substitution of Sn *Corresponding author. Fax: 181-52-735-5316. E-mail address: [email protected] (K. Tanaka)

for Ni in LaNi 5 substantially improves the hydrogenstorage characteristics by suppressing its decomposition during H–D cycles [3–5]. The present work aims to reveal whether the Sn-substitution is also effective in improving the hydrogen storage characteristics of NdNi 5 .

2. Experimental procedure Alloys of compositions NdNi 52s Sn s (s50, 0.1, 0.2, 0.4) were prepared from raw materials of 99.9% Nd, 99.99% Ni and 99.9% Sn by melting together in an Ar-arc furnace. The melting was repeated three times for homogenization, each time inverting the button ingots in situ in the furnace. They were annealed for 6 h at 1223 K in vacuum to enhance the homogeneity. Since the weight loss of each ingot after the melting was less than 0.1%, its actual composition is expected to be close to the nominal one. The alloy compositions and their homogeneity were examined and confirmed with electron-probe microanalyses (EPMA). X-ray diffraction analyses of these alloys indicated that they were composed of a single a-phase with the CaCu 5 structure except for NdNi 4.6 Sn 0.4 which contained certain impurity phase(s) in addition to the major a-phase

0925-8388 / 00 / $ – see front matter  2000 Elsevier Science S.A. All rights reserved. PII: S0925-8388( 99 )00551-4

Y. Takaguchi, K. Tanaka / Journal of Alloys and Compounds 297 (2000) 73 – 80

74

3. Results

3.1. Pressure–composition isotherms

Fig. 1. PCT relations for the NdNi 5 –H 2 system at 253–293 K.

(see Fig. 8d). For the PCT and reaction kinetics measurements, the ingots were mechanically crushed and ground to 70-mesh grain sizes (d,|210 mm). The initial activation of these samples was readily achieved by exposing them to 3.5 MPa H 2 at 253 K for several hours. Following the activation process, they were subjected to five to ten cycles of H–D treatments for stabilization in the same condition as above. The crystal structures and microstructures of the samples before and after the H–D cycles were examined with X-ray diffraction (XRD) and scanning electron microscopy (SEM) techniques.

Fig. 1 shows PCT relations for NdNi 5 measured at 253 to 293 K. The PCT curve at 253 K clearly manifests two plateau regions, 0.1&x&0.6 and 0.7&x&0.9, where x is the hydrogen concentration expressed by the hydrogen-tometal atom ratio (x5H / M). The second-plateau pressure is |6 times higher than the first one. From this feature, the presence of at least three metallographic phases can be inferred for the NdNi 5 –H system, i.e., a (x&0.1), b (0.6&x&0.7) and g (0.9&x&1.0), referring to the terminology adopted for the R(rare earth)Co 5 –H systems [6]. Thus, the first plateau corresponds to the presence of (a1b) and the second to (b1g) mixed phases. At 273 and 293 K the second plateau is shifted up out of the measured pressure range. The hysteresis factor defined as the logarithm of the ratio of the plateau pressures for the absorption and desorption isotherms, ln ( pabs /pdes ), and the maximum hydrogen capacity, x m , of the isotherms are given in Table 1, together with the standard thermodynamic parameters of the plateaus to be determined later. Fig. 2 shows PCTs for NdNi 4.9 Sn 0.1 measured at the same temperatures as above. The Sn-substitution of s50.1 reduces the first-plateau pressure by a factor of |1.5 and the second one by |4 with respect to those for NdNi 5 without changing the maximum hydrogen capacity appreciably. Therefore, the large gap between the first and second plateaus in NdNi 5 is markedly diminished. Furthermore, the substitution induces sloping plateaus and decreases their hysteresis. The second plateaus can be seen more evidently for the desorption isotherms than for the absorption ones. Fig. 3 shows a similar result for NdNi 4.8 Sn 0.2 , in which the effects of Sn-substitution on the PCT stated above are more strengthened. Namely, both the first- and secondplateau pressures are further lowered with the associated hysteresis being more suppressed and the slopes more declined. The gaps between the two plateaus of the

Table 1 The thermodynamic parameters, DH 0f and DS 0f , for the formation of b- and g-phases, the hysteresis factors, ln( pabs /pdes ), at 253 K for the first plateaus, together with the maximum hydrogen capacities, x m , under 3.5MPa H 2 at 253 and 293 K, associated with the PCT characteristics of NdNi 52s Sn s alloys shown in Figs. 1–4 Alloys

DH 0f (kJ / mol H 2 )

DS 0f J / mol H 2 ?K)

ln( pabs /pdes )

x m (253 K)

x m (293 K)

NdNi 5 NdNi 4.9 Sn 0.1

227.9 227.9 223.6 225.7 222.7 226.4 231.8

2117.8 (b) 2109.0 (b) 2106.3 (g) 2102.6 (b) 298.9 (g) 299.3 (b) 2108 (b)

0.54 0.33

0.91 1.00

0.67 0.86

0.16

0.97

0.85

0.06

0.81 b

0.80

NdNi 4.8 Sn 0.2 NdNi 4.6 Sn 0.4 LaNi 5 a a b

Ref. [7]. At 283 K.

(b) (b) (g) (b) (g) (b) (b)

Y. Takaguchi, K. Tanaka / Journal of Alloys and Compounds 297 (2000) 73 – 80

75

desorption isotherms are further diminished and they nearly merge into a single plateau. The maximum hydrogen capacity is still kept to x m |1.0 at 253 K. A further increase in Sn-substitution beyond s50.2 strongly alters the PCT features, as shown for NdNi 4.6 Sn 0.4 in Fig. 4. All the PCTs are heavily curved and the maximum hydrogen capacity is reduced to x m |0.8. The first and second plateaus are merged to give apparently a single heavily declined one. Fig. 5 displays van’t Hoff plots for the first and second plateaus of the desorption isotherms of these alloys, where pdes at x50.35 (first plateau) and 0.75 (second plateau) are plotted in a logarithmic scale against the reciprocal temperature T 21 . For the second plateau, reliable plots are capable only for NdNi 4.9 Sn 0.1 and NdNi 4.8 Sn 0.2 , and only partly for NdNi 5 . By using a conventional relation DH 0f DS 0f ]] ln pdes 5 2 ]], RT R

Fig. 2. PCT relations for the NdNi 4.9 Sn 0.1 –H 2 system at 253–293 K.

Fig. 3. PCT relations for the NdNi 4.8 Sn 0.2 –H 2 system at 253–293 K.

the standard formation enthalpy, DH 0f , and entropy, DS 0f , of the b- and g-phases can be calculated from the plots for the first and second plateaus, respectively. They are given in Table 1, together with those for the b-phase of LaNi 5 [7] for comparison. For the b-phase, DH f0 changes little with the Sn-substitution and takes less negative values by

Fig. 4. PCT relations for the NdNi 4.6 Sn 0.4 –H 2 system at 283–313 K.

76

Y. Takaguchi, K. Tanaka / Journal of Alloys and Compounds 297 (2000) 73 – 80

Fig. 5. van’t Hoff plots for the first and second plateaus of the desorption isotherms shown in Figs. 1–4.

|5 kJ / mol H 2 compared with that for LaNi 5 , while, DS 0f for NdNi 5 is more negative by |10 J / mol H 2 ?K than that for LaNi 5 , and it steadily increases with the Sn-substitution. The steady decrease in the first-plateau pressure with the Sn-substitution is, therefore, attributable to increase in the formation entropy. On the other hand, for the g-phase, DH 0f is less negative by |4 kJ / mol H 2 , and DS 0f is also less negative by |3 J / mol H 2 ?K than those for the corresponding b-phase in NdNi 4.9 Sn 0.1 and NdNi 4.8 Sn 0.2 . This means that the higher pressures of the second plateau compared with the first in these alloys are due to increase in the formation enthalpy of the g-phase with respect to the b-phase. In this study, no numerical results have been obtained for DH 0f and DS 0f of the g-phase in NdNi 5 . However, if we assume that the formation enthalpy of the g-phase in NdNi 5 is as large as those for the Sn-substituted alloys, i.e., DH f0 | 224 kJ / mol H 2 , we obtain DS f0 | 2120 J / mol H 2 ?K, which is much lower than the corresponding values of NdNi 4.9 Sn 0.1 (DS 0f 5 2106.3 J / mol H 2 ?K) and NdNi 4.8 Sn 0.2 (298.9 J / mol H 2 ?K). This means that the remarkable drop of the second-plateau pressure upon Snsubstitution is explainable in terms of a large increase in the formation entropy of the g-phase. This thermodynamic effect appears to be consistent with that found for the lowering of the first-plateau pressure for the b-phase as mentioned above.

3.2. Hydriding-dehydriding kinetics The hydriding and dehydriding rates of the alloys were measured at 293 and 253 K for fully activated and

Fig. 6. Isothermal absorption curves measured at 293 and 253 K under p53.5 MPa. The horizontal lines denote 80% absorption.

stabilized samples. Fig. 6 shows isothermal absorption curves of these samples measured under p53.5 MPa. The hydrogen concentration is normalized to the equilibrium value x m reached at a long enough holding time for each sample. As can be seen from the figure, the absorption rate is significantly enhanced by the Sn-substitution at both temperatures. The characteristic hydriding times for reaching 80% of maximum absorption, t 80 (abs), are given in Table 2. Fig. 7 shows isothermal desorption curves measured at the same temperatures as above under p&0.05 MPa, from which we find that the desorption rate is also enhanced by the substitution. The characteristic dehydriding times for

Table 2 The characteristic reaction times for reaching 80% of the maximum absorption, t 80 (abs), under p53.5 MPa, and 80% of the complete desorption, t 80 (des), under p50.05 MPa at 293 and 253 K Alloys

NdNi 5 NdNi 4.9 Sn 0.1 NdNi 4.8 Sn 0.2 NdNi 4.6 Sn 0.4

t 80 (abs) (ks)

t 80 (des) (ks)

293 K

253 K

293 K

253 K

1.37 0.28 0.19 0.03

6.42 1.01 0.86 0.47

0.39 0.32 0.28 0.24

4.11 2.19 2.11 0.73

Y. Takaguchi, K. Tanaka / Journal of Alloys and Compounds 297 (2000) 73 – 80

77

a little larger than that before hydriding, probably because a certain amount of hydrogen is still retained in the a-phase after dehydriding.

3.4. SEM observation Fig. 9 shows scanning electron micrographs of (a) a virgin alloy before hydriding, (b) NdNi 5 after 13 H–D cycles, (c) NdNi 4.9 Sn 0.1 after nine cycles, and (d) NdNi 4.8 Sn 0.2 after nine cycles. All the virgin alloys investigated consist of grains with d5150–200 mm in size before hydriding. However, they are pulverized during the H–D cycles into grains with d550–100 mm in NdNi 5 , 10–20 mm in NdNi 4.9 Sn 0.1 , and 10–15 mm in NdNi 4.8 Sn 0.2 ; namely, the pulverization is enhanced by the Sn-substitution. Since H–D kinetics is naturally influenced by the particle size of a powdered sample [8], the enhancement of kinetics by Sn-substitution is associated with the concomitant effect of pulverization.

4. Discussion

Fig. 7. Isothermal desorption curves measured at 293 and 253 K under p&0.05 MPa. The horizontal lines denote 80% desorption.

reaching 80% of maximum desorption, t 80 (des), are also given in Table 2.

3.3. X-ray diffraction analyses Fig. 8a–d shows XRD spectra (Cu Ka) of the alloys measured before hydriding and after H–D cycles at 253 K. Before hydriding, all the alloys exhibit a diffraction pattern representing the hexagonal CaCu 5 -type crystal structure. After H–D cycles, this pattern is kept essentially unchanged in all the Sn-substituted alloys. However, in (a) NdNi 5 , the diffraction lines are appreciably broadened, suggesting that inhomogeneous microstrains (lattice defects) are accumulated in it. Although no additional structural investigations have been carried out in this study, e.g., by subjecting these alloys to further H–D cycles and to treatments at elevated temperatures, the present result suggests that the Sn-substitution effectively increases the structural stability of the alloys against the H–D treatment. The lattice parameters, a and c, and the cell volume, V, calculated from these XRD data are tabulated in Table 3. The cell volume monotonically increases with Sn-substitution. In each alloy, the cell volume after the H–D cycle is

The PCT characteristics for the NdNi 52s Sn s –H 2 system show that a partial substitution of Sn for Ni acts to reduce the first- and second-plateau pressures of NdNi 5 as well as the relative pressure gap in between without reducing the maximum hydrogen capacity (H / M|1.0). In an alloy with s50.2 and in a temperature range of 253–293 K, the first plateau falls in a range 0.1–0.6 MPa and the second 0.2–1.2 MPa, and the two plateaus merge to give nearly a single sloping plateau. These characteristics appear to be favorable for a practical use of this alloy for hydrogen storage in this temperature range. The reduction in the plateau pressures is mainly attributable to increases in the formation entropies of both b- and g-hydride phases, rather than to decreases in their formation enthalpies. A disorder caused by random distribution of Sn atoms in the Ni sublattice of NdNi 5 may lead to increases in DS f0 of the hydrides in the compound. The H–D kinetics of NdNi 5 are also improved by the Sn-substitution. The reaction rates of the parent compound NdNi 5 with hydrogen, which is rather sluggish even after the activation treatment, is significantly enhanced by the substitution. This is another advantage of Sn-substituted alloys over the parent compound for practical applications of the alloys. The SEM study indicates that the Snsubstituted alloys undergo pulverization into finer grains than the parent compound after H–D cycles. The production of these fine grains necessarily accelerates the H–D kinetics and improves the reactivity of the alloys. However, the possibility of a concomitant catalytic effect of Sn on the reaction kinetics still remains to be clarified. The reason for the promotion of pulverization during H–D cycles by Sn-substitution is uncertain at present. However, it is conceivable that either the substituted Sn atoms locally expand the interatomic distances of the

78

Y. Takaguchi, K. Tanaka / Journal of Alloys and Compounds 297 (2000) 73 – 80

Fig. 8. XRD patterns (Cu Ka) of the alloys investigated, measured before hydriding and after H–D cycles at 253 K.

Y. Takaguchi, K. Tanaka / Journal of Alloys and Compounds 297 (2000) 73 – 80 Table 3 The lattice parameters, a and c, and cell volumes, V, of the alloys investigated, measured before and after H–D cycles Alloys

H–D cycles

a (nm)

c (nm)

c /a

V (nm 3 )

NdNi 5

Before After Before After Before After Before After

0.4950 0.4952 0.4969 0.4968 0.4972 0.4985 0.4974 0.5009

0.3980 0.3988 0.3993 0.3997 0.4005 0.4012 0.4015 0.4026

0.804 0.805 0.804 0.804 0.806 0.804 0.806 0.806

0.08445 0.08469 0.08538 0.08543 0.08574 0.08634 0.08602 0.08747

NdNi 4.9 Sn 0.1 NdNi 4.8 Sn 0.2 NdNi 4.6 Sn 0.4

79

are readily accumulated in the crystal, which may cause a structural instability of the compound. In fact, this difference in microstructure is inferred from XRD patterns measured before and after H–D cycles (Fig. 8a). Although the degradation behavior of these alloys during H–D treatments at elevated temperatures has not been made clear in this study, NdNi 5 is expected to be structurally stabilized against decomposition by a partial substitution of Sn for Ni, as has been found for LaNi 5 [3–5].

5. Conclusions NdNi 5 compound and thereby weaken the interatomic bonds, or the grain boundaries of the alloys are segregated by Sn atoms and hence embrittled. In any event, the Sn-substituted alloys are embrittled and easily cracked by the internal stress caused by hydrides in comparison with the parent compound. Therefore, the internal stress is readily released and lattice defects are not accumulated in the crystals. On the other hand, in the parent compound, the internal stress is not necessarily released and defects

A pseudo-binary compound NdNi 4.8 Sn 0.2 provides a PCT curve characterized by a nearly single sloping plateau with a small hysteresis, whose desorption pressure ranges between 0.1 and 0.6 MPa at temperatures between 253 and 293 K. The maximum hydrogen capacity reaches H / M| 1.0 (|1.3 wt.%). This alloy exhibits higher hydriding– dehydriding rates than NdNi 5 in this temperature range. This enhancement of kinetics by Sn-substitution is associ-

Fig. 9. SEM photographs of (a) a virgin alloy before hydriding, (b) NdNi 5 after 13 H–D cycles, (c) NdNi 4.9 Sn 0.1 after nine H–D cycles, and (d) NdNi 4.8 Sn 0.2 after nine H–D cycles at 253 K. Note that the photograph (d) is magnified twice as large as the others.

80

Y. Takaguchi, K. Tanaka / Journal of Alloys and Compounds 297 (2000) 73 – 80

ated with increased tendency toward pulverization during H–D cycles. These features appear to be favorable as a practical hydrogen storage alloy for use at ambient or lower temperatures.

Acknowledgements This work is partly supported by the Grant-in-Aid for Scientific Research on Priority Area A of ‘New Protium Function’ from The Ministry of Education, Science, Sports and Culture of Japan.

References [1] H. Uchida, M. Tada, Y.C. Huang, J. Less-Common Met. 88 (1982) 81.

[2] L. Schlapbach (Ed.), Hydrogen in Intermetallic Compounds, Vols. I and II, Springer, Berlin, 1992. [3] S. Luo, W. Luo, J.D. Clewley, T.B. Flanagan, L.A. Wade, J. Alloys Comp. 231 (1995) 467. [4] S. Luo, W. Luo, J.D. Clewley, T.B. Flanagan, R.C. Bowman, J. Alloys Comp. 231 (1995) 473. [5] S. Luo, J.D. Clewley, T.B. Flanagan, R.C. Bowman, J.S. Cantrell, J. Alloys Comp. 253–254 (1997) 226. [6] F.A. Kuijpers, Philips Res. Rep. Suppl. 28 (1973) 1. [7] H.H. van Mal, K.H.J. Buschow, A.R. Miedema, J. Less-Common Met. 35 (1974) 65. [8] N. Gerard, S. Ono, in: L. Schlapbach (Ed.), Hydrogen in Intermetallic Compounds, Vol. II, Springer, Berlin, 1992, p. 165.