Hydriding characteristics of zirconium-substituted FeTi

Hydriding characteristics of zirconium-substituted FeTi

Journal of Alloys and Compounds 313 (2000) 53–58 L www.elsevier.com / locate / jallcom Hydriding characteristics of zirconium-substituted FeTi Nobu...

264KB Sizes 2 Downloads 23 Views

Journal of Alloys and Compounds 313 (2000) 53–58

L

www.elsevier.com / locate / jallcom

Hydriding characteristics of zirconium-substituted FeTi Nobuyuki Nishimiya*, Tomohiro Wada, Akihiko Matsumoto, Kazuo Tsutsumi Toyohashi University of Technology, Tempaku-cho, Toyohashi 441 -8580, Japan Received 18 July 2000; received in revised form 26 August 2000; accepted 26 August 2000

Abstract Hydriding of FeTi 12a Zr a alloys (a50.1, 0.2 and 0.5) proceeded without any activation treatments. Zirconium substitution lowered the equilibrium pressure in the b phase region and narrowed the width of the plateau. The hydrogen capacity increased with the zirconium content, but the reversible amount of hydrogen, conveniently defined as the difference in the hydrogen contents under 4 MPa and 0.05 MPa of hydrogen, decreased with the a value owing to the increased hydrogen content under 0.05 MPa. When the value a was equal to 0.5, the isotherm showed no plateau and major parts of observed data obeyed the general solid solution model, wherein the number of interstitial sites was equal to the number of metal atoms. A new ternary phase Fe 2 TiZr would substantially be present and hydrogen atoms occupy the interstitial sites to form Fe 2 TiZrH 4 . At lower temperatures, the hydrogen capacity x (Fe 2 TiZrH x ) exceeded 4 under high pressures of hydrogen.  2000 Elsevier Science B.V. All rights reserved. Keywords: Hydrogen storage materials; Intermetallics; Gas–solid reactions; Phase transitions; Thermodynamic properties

1. Introduction The quaternary alloys, (Fe b Mn 12b ) c Ti 12a Zr a , where a and b are 0 to 1 and c is 1 to 2, are promising hydrogen storage materials [1]. When c is equal to 1.5 or 2, the alloys are monophasic provided that both a and b are between 0.2 and 0.8. Zirconium would substitute for titanium as typically suggested by a monotonous increase in the lattice parameters of hexagonal (Fe 0.5 Mn 0.5 ) 1.5 Ti 12a Zr a phases with the a value. The hydrogen equilibrium pressures for (Fe b Mn 12b ) c Ti 12a Zr a decrease with the increasing lattice parameters and can be adjusted to intended values at mild temperatures. For example, (Fe 0.2 Mn 0.8 ) 1.5 Ti 0.5 Zr 0.5 shows an equilibrium pressure of 0.1 MPa at ca. 400 K and is practically promising owing to easy activation, little hysteresis and high hydrogen content (2.0 wt%) [2]. Since the alloys with the lower c values have the higher hydrogen contents and are more readily activated [1,2], the c value will preferably be fixed to the lower limit, 1, in order to search for novel compositions for hydrogen storage. The b value will be fixed to 1 for simplicity, and zirconium-substituted FeTi alloys, FeTi 12a Zr a , are thus to be studied in the present work. *Corresponding author. E-mail address: [email protected] (N. Nishimiya).

We are aware that zirconium does not necessarily replace all of the titanium sites. As a matter of fact, a part of manganese atoms in Zr(Mn 0.72 Fe 0.28 ) 2.4 and Zr(Mn 0.72 Fe 0.28 ) 2.8 do not substitute for the iron sites, but replace the zirconium sites [3]. We will use the notation FeTi 12a Zr a to show the overall composition of the solid phase. There are three papers that treat the FeTi 12a Zra alloys [4–6]. The studied a values are 0.01, 0.1 and 0.2 in the literature by Jang et al. [4], 0.02, 0.1, 0.2 and 0.3 by Nagai et al. [5] and 0.05 and 0.1 by Lee et al. [6]. There is a contradiction concerning the hydrogen capacity; Jang et al. and Nagai et al. claimed that the capacity decreased with the zirconium content, whereas Lee et al. reported that the decreasing order was FeTi 0.9 Zr 0.1 .FeTi.FeTi 0.95 Zr 0.05 . In the present work, the thermodynamic properties of the FeTi 12a Zr a –H 2 systems will be reconsidered with stress laid on the proper determination of the hydrogen contents. Secondly, the a value will be extended to 0.5 and a new composition of hydrogen storage alloy will be sought.

2. Experimental The alloys were prepared by arc-melting calculated amounts of reagent grade metals under an atmosphere of

0925-8388 / 00 / $ – see front matter  2000 Elsevier Science B.V. All rights reserved. PII: S0925-8388( 00 )01181-6

54

N. Nishimiya et al. / Journal of Alloys and Compounds 313 (2000) 53 – 58

high purity argon. Each alloy disk was mechanically polished, coarsely crushed, immersed in 1:1 hydrochloric acid for a few minutes, washed in dry acetone several times and crushed again under nitrogen to gather particles ranging from 5 to 10 mesh. Hydriding was carried out using a conventional pressure-proof vessel. Prior to the measurements of the pressure–composition isotherms, the alloys were subjected to several hydriding–dehydriding cycles (hydriding at the studied temperatures, 298–613 K, and dehydriding at 623 K), and reproducible kinetics were confirmed on hydriding. In almost all cases, the desorption isotherms were measured after the last hydriding came to equilibrium at the studied temperatures under several MPa of hydrogen. When an absorption isotherm was to be measured, the repeatedly hydrided sample was evacuated at 623 K for 12 h under 5 Pa, cooled to the studied temperature in vacuo and hydrided with the occluded amount volumetrically measured. The equilibrium pressure data were acquired every 24 h. When the last equilibration was attained for the desorption isotherm, the pressure-proof vessel was closed, cooled to room temperature and attached to a two-step thermal decomposition system fully described in our earlier work [7]. The sum of the amounts of hydrogen in the solid phase and in the gas phase inside the reaction vessel at room temperature, n SG( RT ) , was measured with a Toepler pump in the system until no further hydrogen was released from the sample under 10 Pa at 673 K. The residual hydrogen, n S(ST) , at the last equilibration was determined by the subtraction, n S(ST) 5 n SG(RT) 2 n G(ST)

(1)

where n G( ST ) was mols of hydrogen in the gas phase at the studied temperature and was determined from an experimental relationship between the pressure and mols of hydrogen. A portion of the sample was transferred into a quartz tube and heated to 1273 K in vacuo and a trace amount of hydrogen was collected by the Toepler pump to correct the n SG( RT ) value.

Fig. 1. Desorption isotherms for FeTi– and FeTi 0.9 Zr 0.1 –H 2 systems at 293 K.

conium for titanium lowered the dissociation pressure in the b phase region, narrowed the width of the plateau and increased the residual hydrogen amount, n S( ST) . The hydrogen capacity did not significantly change, but the average hydrogen atoms to metal ratio (H / M) remained at ca. 1 at saturation. When the a value in the FeTi 12a Zr a alloy increased, the n S(ST) value grew larger as can be seen from Fig. 2. The capacity fairly increased with the a value, whereas the H / M value did not largely differ from 1 at saturation. On going from a50.1 to a50.2, the width of the plateau was

3. Results The zirconium-substituted alloys readily occluded hydrogen at room temperature without any activation treatments. Evacuation under 5 Pa at 623 K for 4 h followed by admission of 4 MPa of hydrogen was sufficient to hydride the alloys at studied temperatures. In particular, the FeTi 12a Zr a alloy with a50.5 attained its maximal hydrogen content in the first hydriding run at room temperature under 4 MPa. The improved activation properties due to the substitution of zirconium for titanium were essentially the same as those reported earlier [4–6]. As can be seen from Fig. 1, the substitution of zir-

Fig. 2. Desorption isotherms for FeTi 12a Zr a – H 2 systems with different a values at 313 K.

N. Nishimiya et al. / Journal of Alloys and Compounds 313 (2000) 53 – 58

reduced to a half, but the lowering in the equilibrium pressure was slight. When the a value was equal to 0.5, the isotherm consisted of one smooth curve with no plateau. The observed lowering in the plateau pressure on the substitution of zirconium for titanium is qualitatively similar to the reported observations [4–6]. Using the reported equation [8], the plateau pressure for FeTi is calculated to be 0.71 MPa at 313 K and 0.35 MPa at 293 K, the latter being close to the observed data in Fig. 1. Comparing these values with the isotherms in Figs. 1 and 2, we are able to say that the lowering in the plateau pressure is significant between FeTi and FeTi 0.9 Zr 0.1 and that it is small between FeTi 0.9 Zr 0.1 and FeTi 0.8 Zr 0.2 . The narrowed plateau width on the substitution is accompanied by the suppressed formation of the g phase hydride. This has also been found by Jang et al. and described as a sloping phenomenon [4]. According to Nagai et al., no plateau has been detected for FeTi 0.7 Zr 0.3 in the desorption isotherm constructed with 7 data points [5]. The alloy FeTi 12a Zr a with a50.3 is not included in the present work. Instead the higher substituted alloy with a50.5 has no plateau as can be precisely shown in Fig. 2. Figs. 1 and 2 show that the hydrogen capacity of FeTi 0.9 Zr 0.1 is slightly smaller than that of FeTi and that it increases through higher substitution. This finding differs from the reported results after Jang et al. [4] and Nagai et al. [5]. As cited beforehand, the capacity of FeTi 0.95 Zr 0.05 was lower than that of FeTi, but FeTi 0.9 Zr 0.1 occluded more hydrogen than FeTi, according to Lee et al. [6]. The reported hydride composition was, however, FeTiH 1.60 at 323 K under 4 MPa, contradicting to the expected composition, FeTiH .1.8 , after Reilly et al. [8]. In the present work, the composition FeTiH 1.91 at 293 K in Fig. 1 was consistent with the interpolated one, FeTiH 1.9 . Thus the decreasing order should be FeTi 0.5 Zr 0.5 .FeTi 0.8 Zr 0.2 . FeTi.FeTi 0.9 Zr 0.1 . The residual amount of hydrogen, n S( ST) increases on the zirconium substitution (Fig. 1) and grows larger with the zirconium content (Fig. 2). This tendency is consistent with that reported by Lee et al. [6]. Although Jang et al. [4] were aware of this phenomenon to point out that the range of the a phase hydride was extended on the substitution, the observed values for n S( ST ) were much smaller than those in the present work and those after Lee et al. [6]. In the study by Nagai et al. [5], the n S( ST ) values were not determined, but a treatment for 1 h in vacuo at 303 K was applied prior to the measurement of both the absorption and desorption isotherms. Thus their hydrogen capacity in the first hydriding run was much larger than the saturation amount estimated from the isotherm. In the present study, the n S( ST) values were determined by the two-step heating procedure, and it was found that most of the residual hydrogen released from the solid at 500–600 K under reduced pressures and that no further hydrogen came out of the solid at elevated temperatures between 673 and 1273 K.

55

The shape of the isotherm for FeTi 0.5 Zr 0.5 in Fig. 2 suggests that the hydrogen occlusion would proceed by a solid solution mechanism. Since the following results and discussion might support an existence of a single phase alloy which absorbs hydrogen to form a solid solution and reversibly desorbs hydrogen to restore the original phase, we would represent this new ternary phase as Fe 2 TiZr. As shown in Fig. 3, no plateau feature of the isotherm was indeed noticed at varied temperatures. As shown in Fig. 4, the X-ray diffraction patterns before and after the cycling were the same and no phase separation was detected. Hydriding and dehydriding proceeded in reversible manners, which was further confirmed by the reproducible saturation amount of hydrogen. The observed patterns were different from those of known metals and intermetallics. For example, the composition Fe 2 TiZr might be constructed by FeTi11 / 3 (ZrFe 2 1Zr 2 Fe), but the diffraction peaks were not attributable to these phases. A plausible phase separation to Ti plus ZrFe 2 , which was proposed for FeTi 0.9 Zr 0.1 by Lee et al. [6], did not take place either. The phase behavior of Fe 2 TiZr was quite different from the lower substituted alloys, FeTi 0.9 Zr 0.1 and FeTi 0.8 Zr 0.2 , where FeTi phases with slightly increased lattice parameters were found. Indexing of the diffraction peaks from Fe 2 TiZr was not yet successful. From a thermodynamical consideration as fully described in the literature [9], the following equation similar to the one applied to the solid solution in palladium [10] is derived for Fe 2 TiZrHx , ln[(P/P0 )(r 2 x) /x ] 5 2DH¯ H /RT 2 2DS¯ H /R 2

2

XS

(2)

wherein r is the number of interstitial sites per Fe 2 TiZr, P

Fig. 3. Desorption isotherms for Fe 2 TiZr–H 2 system at varied temperatures. m, 613 K; j, 543 K; d, 410 K; ♦, 313 K.

56

N. Nishimiya et al. / Journal of Alloys and Compounds 313 (2000) 53 – 58

Fig. 4. Cu Ka X-ray diffraction patterns for Fe 2 TiZr before the cycling (as-prepared) and after ten hydriding–dehydriding cycles (after cycled). Four JCPDS data are inserted.

the equilibrium pressure, P0 the standard pressure (10 5 Pa), DH¯ H the relative partial molar enthalpy and DS¯ XS the H relative partial molar excess entropy. Theoretically the variable x is not restricted to a low concentration limit. Plots of log P versus log [x 2 /(r 2 x)2 ] should, therefore, give straight lines with slopes 1, if Fe 2 TiZrH x forms through a solid solution mechanism. Fig. 5 shows the actual plots for Fe 2 TiZrH x on the assumption that the number of interstitial sites is equal to the number of metal atoms, that is, r54. At the lower temperatures, the x value actually exceeded 4 under high pressures of hydrogen, as can be seen in Fig. 3, and those data points with x.4 were omitted here for simplicity. The straight lines in Fig. 5 were best-fitted ones with slopes 1. Most data seemed to obey the above equation theoretically derived based upon general solid solution model. The x range where the fittings were good was from 2 to 3.5. The data points with x,1.6 deviated from the straight line at 613 K and those with x.3.6 at 313 K. As can be seen in Fig. 6, a plot of log [(P/P0 )(r 2 x)2 / 2 x ] versus 1 /T gave a straight line, when only the data on the straight lines in Fig. 5 were used. Again the above equation held in this case. The partial molar quantities relative to hydrogen in its standard state at a pressure of 10 5 Pa were thus evaluated; DH¯ H 5 217.9 kJ mol H 21 and XS 21 21 DS¯ H 5 236.1 J K mol H . The enthalpy value was close to that reported for FeTi, 214.1 kJ mol H 21 , in the composition range of the b phase formation [8]. The absolute value was much less than those known for the solid solution and the b phase formation of Ti, 245.2 and 252.3 kJ mol H 21 , respectively [11]. The enthalpy value

Fig. 5. Plots of log P against log [x 2 /(r 2 x)2 ] with r54 for the Fe 2 TiZr2H 2 system at varied temperatures. m, 613 K; j, 543 K; d, 410 K; ♦, 313 K.

for Fe 2 TiZrH x was similarly much less negative than those for ZrH x . In order to know the applicability of Eq. (2) to the absorption process, a 1000 Torr sensor (1 Torr5133 Pa) was used together with the high pressure sensors and an isotherm in Fig. 7 was obtained. By plotting log P against log [x 2 /(r 2 x)2 ], we obtained a straight line (a) in Fig. 8. Here r54 was again assumed. Two data points in the low x range deviated downwards and those in the x range between 1.7 and 3.6 were fitted well. Thus the above equation held in the absorption branch in the same manner

Fig. 6. A plot of log [(P/P0 )(r 2 x)2 /x 2 ] against 1 /T with P0 510 5 Pa and r54 for Fe 2 TiZr2H 2 system.

N. Nishimiya et al. / Journal of Alloys and Compounds 313 (2000) 53 – 58

Fig. 7. Absorption isotherm for Fe 2 TiZr–H 2 system at 298 K.

as in the desorption branch. The line (b) in Fig. 8 shows the calculated line for desorption at 298 K. The spacing between (a) and (b) denotes a large hysteresis in the Fe 2 TiZr–H 2 system.

4. Discussion The observed lowering in the plateau pressure on the zirconium substitution for titanium was qualitatively consistent with the reported observations [4–6]. Jang et al. [4]

Fig. 8. Plot of log P against log [x 2 /(r 2 x)2 ] with r54 for hydrogen absorption (a) and best-fitted line for desorption calculated on the basis of Eq. (2) (b) for Fe 2 TiZr–H 2 system at 298 K.

57

studied three compositions, FeTi 0.99 Zr 0.01 , FeTi 0.9 Zr 0.1 and FeTi 0.8 Zr 0.2 , and recognized the lowering in the equilibrium pressure in this order. However, the reported order of the pressure in the g phase was opposite. The residual amounts of hydrogen, n S(ST) , which were graphically shown in the literature [4], were much less than those in the present work, and this would have caused the irregular order of the equilibrium pressure. Since the lattice parameters of the FeTi phase in FeTi 0.9 Zr 0.1 and FeTi 0.8 Zr 0.2 increased with the zirconium content, zirconium would replace the titanium sites in the FeTi phase. As mentioned earlier, the pressure difference in the b phase region between FeTi and FeTi 0.9 Zr 0.1 was larger than that between FeTi 0.9 Zr 0.1 and FeTi 0.8 Zr 0.2 . A substantial portion of zirconium in FeTi 0.8 Zr 0.2 might not, therefore, substitute for titanium, but possibly constitute another phase. According to Jang et al. [4], FeTi 0.8 Zr 0.2 was not homogeneous. Lee et al. [6] described that both ZrFe 2 and Ti phases were detected in FeTi 0.9 Zr 0.1 . In the present work, only the FeTi phase was found and a plausible phase separation was not recognized. Assuming that FeTi 0.9 Zr 0.1 was composed of FeTi and Fe 2 TiZr and that the latter was not detected by X-ray diffraction, we could roughly estimate the isotherm for the FeTi 0.9 Zr 0.1 alloy as shown in Fig. 1 from those for FeTi and Fe 2 TiZr. Linearly combining the 293 K isotherm for FeTi in Fig. 1 and that for Fe 2 TiZr constructed on the basis of the fitting in Fig. 6, we got the following x values in FeTi 0.9 Zr 0.1 H x under the parenthesized equilibrium pressures (MPa); 0.41 (0.02), 0.43 (0.05), 0.44 (0.1), 1.43 (0.5) and 1.97 (2). These values were fairly consistent with the observed values; 0.43 (0.02), 0.49 (0.05), 0.54 (0.1), 1.33 (0.5) and 1.74 (2). The proposed existence of the ternary phase, Fe 2 TiZr, seemed to be supported. If the composition Fe 2 TiZr decomposes to form FeTi1 1 / 3 (ZrFe 2 1Zr 2 Fe), then the x value in Fe 2 TiZrH x at saturation should be 2.1 at 313 K, which is less than a half of the observation in Fig. 3. Here the saturating compositions FeTiH 1.86 [8], ZrFe 2 H 0 and Zr 2 FeH 0.69 were used. The latter two values were determined in the present work. If Fe 2 TiZr decomposes to Ti and ZrFe 2 as suggested by Lee et al. for FeTi 0.9 Zr 0.1 [6], the extraction of hydrogen under 10 Pa at temperatures higher than 673 K (up to 1273 K) should give a nonzero value. However, such was not the case. The presence of the b hydride of titanium, which is stable under 10 Pa at 673 K [11], is thus to be denied. According to Nagai et al. [5], the FeTi 12a Zr a alloys with a50.1, 0.2 and 0.3 were composed of FeTi, Fe 2 Ti and FeTi 2 with Zr. The phase FeTi 2 is not known and an FeTi 1.3 alloy composed of FeTi and Ti has been reported [12]. After all, the three hypotheses on the phase separation discussed here should be discarded. For most applications, the reversible amount of hydrogen is of primary importance. It is convenient to define this amount as the difference in the x values of FeTi 12a Zr a H x under 4 MPa and 0.05 MPa of hydrogen. The reversible

58

N. Nishimiya et al. / Journal of Alloys and Compounds 313 (2000) 53 – 58

amount decreased with the zirconium content; 1.16 for FeTi 0.9 Zr 0.1 , 0.85 for FeTi 0.8 Zr 0.2 and 0.51 for FeTi 0.5 Zr 0.5 at 313 K. These values were much less than that for FeTi, 1.76 at 313 K [8]. The decrease was due to the increased hydrogen content under 0.05 MPa, presumably caused by the formation of the proposed ternary phase, Fe 2 TiZr. In order to increase the reversible amount, one should study the FeTi 12a Zr a alloys with low a values. So far the reported amounts were rather scattered; 1.20 for FeTi 0.99 Zr 0.01 at 303 K [4], 1.65 for FeTi 0.98 Zr 0.02 at 303 K [5] and 1.00 for FeTi 0.95 Zr 0.05 at 323 K [6]. It would be necessary to know the phase behavior and minimize the plausible phase separation. One promising approach would be research on multi-component alloys, like Fe b Mn 12b Ti 12a Zr a and Fe 0.945 Nb 0.04 Ti 0.96 Zr 0.04 . The reversible amount for the latter was as large as 1.72 at 303 K [13]. The proposal of the existence of the ternary phase, Fe 2 TiZr, is primarily based upon the successful fittings in Figs. 5, 6 and 8. One of the important questions to be answered is why r is equal to 4. The reasons for the deviations in the higher and the lower x ranges should also be clarified. At present, any crystal models for Fe 2 TiZr are not constructed and one interstitial site is simply attributed to one metal. The downward deviations of the data points in the lower x range of Figs. 5 and 8 would suggest that there exist some preferential interstitial sites. It is unclear whether they are distributed in the Fe 2 TiZr phase or in other coexisting phases. The preferential sites and the hydrogen atoms on them did not go out of the equilibrating system, but both of them came to be redistributed in the medium x range. If there exist some other phases than Fe 2 TiZr, they should be nano-structured to intimately interact with Fe 2 TiZr.

5. Conclusion Zirconium substitution for titanium in FeTi made the activation treatments needless, lowered the equilibrium

pressure in the b phase region and narrowed the width of the plateau. The hydrogen capacity increased, but the reversible amount of hydrogen of practical importance decreased with the zirconium content owing to the increased residual amount of hydrogen. The isotherm for Fe 2 TiZr showed no plateau and most data points obeyed a general solid solution model. A new ternary phase Fe 2 TiZr would substantially be present and hydrogen atoms occupy the interstitial sites to form a solid solution, Fe 2 TiZrH 4 .

References [1] A. Suzuki, N. Nishimiya, S. Ono, S. Higano, K. Kamino, Chem. Lett. (1982) 75. [2] Y. Komazaki, M. Uchida, S. Suda, A. Suzuki, S. Ono, N. Nishimiya, J. Less-Common Metals 89 (1983) 269. [3] G. Triantafillidis, L. Pontonnier, D. Fruchart, P. Wolfers, J.L. Soubeyroux, J. Less-Common Metals 172–174 (1991) 183. [4] T.H. Jang, J.I. Han, J.Y. Lee, J. Less-Common Metals 119 (1986) 237. [5] H. Nagai, K. Kitagaki, K. Shoji, Trans. Japan Inst. Metals 29 (1988) 494. [6] S.M. Lee, T.P. Perng, Hydrogen Energy Progress X (1994) 1139. [7] A. Suzuki, N. Nishimiya, Mat. Res. Bull. 19 (1984) 1559. [8] J.J. Reilly, R.H. Wiswall Jr., Inorg. Chem. 13 (1974) 218. [9] Y. Fukai, The Metal–Hydrogen System, Springer-Verlag, Berlin, Heidelberg, 1993, Chapter 1. [10] J.D. Clewley, T. Curran, T.B. Flanagan, W.A. Oates, J. Chem. Soc. Faraday Trans. I 69 (1973) 449. [11] W.M. Mueller, in: W.M. Mueller, J.P. Blackledge, G.G. Libowitz (Eds.), Metal Hydrides, Academic Press, London, 1968, Chapter 8. [12] S.M. Lee, T.P. Perng, J. Alloys Comp. 177 (1991) 107. [13] T. Sasai, K. Oku, H. Konno, K. Onouwe, S. Kashu, J. LessCommon Metals 89 (1983) 281.