J. inorg, nucl. Chem. Vol. 43, pp. 1763-1768, 1981 Printed in Grea~ Britain.
0022-1902]81/081763~6502.0010 Pergamon Press Ltd.
EFFECTS OF NICKEL AND INDIUM TERNARY ADDITIONS ON THE HYDROGENATION OF Mg-A1 INTERMETALLIC COMPOUNDS Z. GAVRA Nuclear Research Centre-Negev, P.O. Box 9001, Beer-Sheva, Israel and Z. HADARI and M. H. MINTZ* Nuclear Research Center-Negev,and The Ben-GurionUniversityof the Negev, Departmentof Nuclear Engineering, P.O. Box 653, Beer-Sheva, Israel
(Received 28 July 1980; accepted for publication 22 September 1980) Abstract--Alloying small amounts (2-3 wt%) of Ni and In with Mg-AI intermetallicscauses a significantincrease in the hydrogenation rates of these compounds. The catalitic effects of these ternary additions are accounted for by two mechanisms: (a) The formation of surface locations enriched with the ternary constituent which provide preferred dissociation sites for hydrogen and (b) More extensive cracking of the reacting particles during repeated hygrogenation-dehydrogenationcycles. The hydrogenation reaction of both, binary and ternary Mg-AI /3-phase alloys, has been found to follow a law of a three dimensional diffusion through a growing product layer. The thermodynamic parameters associated with the hydrogenation reaction of the Mg-A1 compounds, are evaluated. EXPERIMENTAL
Hydrogenation characteristics of magnesium alloys and magnesium intermetallic compounds are of great interest in developing new hydrogen storage systems with relatively high hydrogen weight capacities. Intermetallic compounds such as Mg2Ni and Mg2Cu exhibit improved hydrogen storage quality with respect to that of pure magnesium. It has been noticed that alloying small amounts of aluminum with magnesium increases the hydrogenation rate of the metal[3,4]. However, no influence of these low concentration aluminum additives on the thermodynamic stability of the corresponding hydrides has been detected. Mg-AI intermetallics, on the other hand, may potentially form hydrides with lower stabilities. Considering the light atomic weight of these intermetallics and their low cost, the investigation of the hydrogenation characteristics of such compounds may provide useful information needed for the research of new hydrogen storage media. Preliminary results have been recently reported on the hydrogenation reaction of Mg-A1 intermetallic compounds [5-7]. It has been found that hydrogenation of the /3-phase (referred in literature as Mg2Ala or MgsAIs) and of the y-phase (which has a wide range of existence, from 45 to almost 60wt% Mg, and is referred in literature as Mg3AIz or Mg17AI12) causes disproportionation into MgH2 and AI. This reaction takes place reversibly (i.e. the product of the dehydrogenation is the Original intermetallic compound) resulting in decomposition equilibrium pressures which are higher than for pure MgH2. The reaction rates however, were found [5-7] to be very sluggish impeding the utilization of such intermetallics for practical storage systems. It is the purpose of the present work to evaluate detailed kinetic and thermodynamic data on these magnesium-aluminum-hydrogen systems and to determine the effects of nickel and indium additions, which for pure magnesium, were found to catalize hydrogenation rates [4, 9]. *Author to whom correspondence should be addressed.
Alloys preparation. The binary Mg-AI and the ternary Mg-AINi, Mg-Al-ln alloys, were prepared by melting the appropriate amounts of the constituents in an r.f. induction furnace, under inert argon atmosphere. Each alloy was remelted at least twice, in order to achieve homogeneity. The alloys were characterized by X-ray diffraction, chemical analyses, metallography, and in the case of the ternary alloys, also by electron microprobe analysis. Table 1 summarizes the characteristics of each of the alloys utilized in the present study. Hydrogenation procedure. The reacting samples were crushed into fine powders and inserted into a conventional stainless-steel constant volume hydrogenation system. The samples were outgassed at about 400°C under high vacuum (-10 -5 torr) for at least 2 hr. Then, ultra high pure hydrogen at a pressure of about 70atm was introduced. The reaction course was followed by recording continuously the pressure drop vs time. Reaction rates were very slow during the first hydrogenation but accelerated after several cycles of absorption-desorption. After equilibrium has been established at 400°C, the hydrided sample was cooled to the desired temperature (in the range 400-320°C) at which the corresponding pressure-composition (P-C) desorption isotherm was obtained. The hydrogenated samples were characterized by X-ray diffraction and by thermal analysis performed in a thermobalance incorporated facilities for simultaneous thermogravimetricdifferential thermal analyses (TGA-DTA). Some of the samples were also analysed by differential scanning calorimetry (DSC) performed under helium flow.
Table 1. Characterization of the various binary Mg-A1 and ternary Mg-Al-In, Mg-AI-Ni alloys utilized in the present study. (The symbols s, m, w, are related to X-ray diffraction lines intensities) Alloy Mg2AI3 Mg2Al3-2wt%In Mg2A13-3wt%Ni
/3(s) +/3'(w)" bMgo.375Alo.62oIno.oos(s)+ if(w) bMgo.3soAlo.610Nio.olo(S)+ NiAI3(w)
+ /3'(vw) aThe/3' (or E) phase exists between the/3 and y fields in the Mg-Al phase diagram. bCombinedelectron microprobe and EDAX analyses.
Z. GAVRAet al.
1764 RESULTS AND DISCUSSION
(a) Hydrogenation kinetics Figure 1 presents some curves of a (the reacted fraction)t vs time, for Mg2AI3 (//-phase) powders after various numbers of hydrogenation-dehydrogenation cycles, at 400°C. Hydrogenation rates of the binary MgA1 intermetallics are very slow. For Mg2A13,about 60 hr were needed to attain equilibrium, during the first hydrogenation, at 400°C. Even activated samples which went through 15 cycles of absorption-desorption reacted relatively slowly and few hours were needed to attain equilibrium. Alloying small amounts (2-3 wt%) of Ni and In to Mg2AI3accelerates the activation process of the samples. Figure 2 presents a vs time curves for the second
hydrogenation cycle of the Ni and In containing Mg,AI3 alloys as compared with that of the pure binary compound. The catalitic effect of the ternary additives is definitely seen. The most pronodnced effect however, of the Ni and In additives, is the rapid increase in hydrogenation reaction rates caused by repeated cycling. Figure 3 presents the kinetic behavior of the Ni and In containing samples which went through 15 hydrogenation-dehydrogenation cycles (at 400°C) as compared with the binary alloys. While for the pure binary Mg2Al3fl-phase the duration of the reaction is of the order of some hours, the time scale of the In and especially the Ni containing alloys is of the order of minutes. These results do not agree with the data of Eisenberg et aL who reported that the addition of Ni does not affect the hydrogenation kinetics of Mg2AI3 fl-phase. Those authors however, investigated hydrogenation rates at much lower temperatures than 400°C. At such low temperatures (200-275°C) the hydrogenation kinetics are so sluggish that the effects of ternary additives may be overlooked. The present kinetic results were fitted to various calculated models available in literature . Approximating the shape of the reacting particles to spheres with a uniform radius ro, the best fit was obtained to a diffusion controlled process through a growing product layer. Such mechanism was found to control the hydrogenation rate of magnesium and magnesium-group IIIA dilute alloys[4,9]. Up to a - 0 . 5 , hydrogenation kinetics of both, binary and ternary Mg2Al3//-phases followed the relation:
D 4 ( a ) = (1 - 2 a / 3 ) - (1 - a ) 2/3 =
K(ro, T)" t
Fig. I. Effect of hydrogenation-dehydrogenationcyclingon the kinetics of binary Mg2AI3(//-phase) hydrogenationreaction, at 400°C. &, one cycle; ID,fifteen cycles.
K(ro, T) = k(T)lro 2
where k(T) is the temperature dependent rate constant. Table 2 summarizes the values of the rate constants K(ro, T), obtained at 400°C for various numbers of hydrogenation-dehydrogenation cycles. It is seen that while for the binary alloys 15 cycles increase the rate constant by a factor of about 3 the corresponding increase in the Ni containing alloys is by a factor of 50.
t , hr
Fig. 2. Effect of nickel and indium additions on the activation process of the Mg-AI B-phases (kinetics of second hydrogenation cycle, at 400°C). l , binary Mg2AIB;O, Ni containing #-phase; &, In containing#-phase.
l , rnin
tAs the hydrogenationreaction proceeds by disproportionation (i.e. formation of MgH2 and AI), a was taken as unity when reaching H/Mg= 2.
Fig. 3. Hydrogenation kinetics at 400°C of cycled (15 hydrogenation-dehydrogenationcycles) #-phase Mg-AIalloys, l , pure binary Mg2A13;O, Ni containing #-phase; &, In containing #-phase.
Hydrogenation of Mg-AI intermetallic compounds Table 2. Rate constants, K(ro, T), (min-~) [see eqn (1)] at T = 400°C, measured for the various powdered B-phase alloys cycled (hydrided-.dehydrided) one and fifteen times respectively Alloy
/3 + Ni
One cycle Fifteen cycles
2.2 × 10-5 5.8 x 10-5
8.3 x 10-5 4.2 x 10-3
7.9 x 10-5 6.1 x 10-4
changes occur by repeated hydrogenation--dehydrogenation (only some extent of fragmentation). The increase in particle size may be caused by free sintering which is probably enhanced by the presence of the ternary additives. Sieve size however, may be related to the kinetics of the hydrogenation reaction, only if the reacting particles are not cracked or porous, otherwise the extent of cracking or porosity is the dominant factor leading to the effective ro given in eqn (1). Figures 4 and 5 present scanning-electron-microscope (SEM) photographs of the original and cycled (15 cycles) powders. It is seen that both, binary and ternary/3-phases become porous during cycling. However, while the porous particles of the binary alloys consist of primary unfractured particles, having the size of about 10/zm, these primary particles of the ternary alloys are fractured and subdivided by microcracks, which reduce their effective size to about 1-2 gin. These observations may probably be correlated with the hydrogenation fragmentation resistance of Mg-10% A1 and Mg-25% Ni alloys reported by Douglass. For these alloys it has been found that whle the A1 containing samples did not particulate or crack during hydrogenation, the Ni containing alloys had much less resistance to fragmentation. The SEM photographs of the original unhydrided powders (Fig. 4) exhibited small bright bulges on the surface of the particles. For the Ni containing alloys it has been found by EDAX analysis that these bulges are enriched with nickel relative to the bulk. It is possible that these Ni enriched surface locations provide preferred dissociation sites for hydrogen molecules, which may account for the faster activation process displayed by the ternary/3-phases. (b) Thermodynamic data Some of the thermodynamic data on the binary /3phase-hydrogen system and some features of the yphase-hydrogen system have already been published. We shall thus summarize only the main features of these reactions: 1. Hydrogenation of the intermetallic Mg-A1 compounds results in disproportionation. For the B-phase, the reaction may be written as: Mg2A13+ 2H2~2MgH2 + 3AI
while for the y-phase, two parallel reactions occur simultaneously:
Fig. 4. SEM photographs of Mg-AI/3-phasepowders (200-mesh) before hydrogenation. (a) Pure binary Mg2AI3,(b) nickel containing/3-phase.
The pronounced cycling effect introduced by the addition of indium, and especially by nickel, may logically be attributed to a more effective fragmentation taking place during hydrogenation. Contrary to the expected result, .the sieve size distribution of the In and Ni containing alloys indicate an increase in particle size during cycling, while for the pure binary alloys no significant
3Mg3A1+ 5H2~5MgH2 + 2Mg2A13
Mg3A12+ 3H2~3MgH2 + 2A1.
2. Reactions (2) and (3) occur reversibly at the temperature range investigated in the present study. The corresponding equilibrium decomposition pressures are thus higher than the decomposition pressures of pure MgH2, since the enthalpy change associated with each reversible reaction differs from that of MgH2 decomposition, the difference being related to the enthalpy of formation of the corresponding Mg-AI intermetallic compound . The effects of the ternary substitutions of Ni and In on the thermodynamic parameters associated with the hydrogenation reaction of the Mg2AI3 /3-phase are
Z. GAVRAet aL
Fig. 5. SEM photographs of Mg-AI/3-phase. Powders after 15 cycles of hydrogenation-dehydrogenation. (a) Pure binary Mg:AI3,(b) nickel containing//-phase. Up--small magnificationwhere the porous structures of powders particle is seen. Down--large magnification where the structure of the primary particles is illustrated. A cracked subdivided structure is observed for the nickel containing alloys as compared with the unfractured structure of the binary Mg2AI3.
Table 3. Thermodynamic parameters, associated with reaction (2), for the various #-phase alloys Alloy /3-Mg2AI3 /3-Ni //-In 3'-Mg3AI2 Pure Mg
AH (kcal/mole H2)
AS (cal/deg mole H2)
15.1 17.9 17.6 18.6 17.8
29.6 33.8 33.9 33,8 32.3
summarized in Table 3. The corresponding P-C decomposition isotherms from which these thermodynamic parameters were evaluated are presented in Figs. 6-8. The addition of ternary constituents increases the absolute values of both AH and AS relative to the binary Mg2AI3 compound. The increase of AH is thus compensated by the increase of AS, so that no stabilization (i.e. lowering of the equilibrium pressures) occurs in the
P atm P atm (300°C) (400°C) 5.0 3.6 5.0 1.9 1.8
36.0 37.2 49.4 21.8 18.3
temperature range 320-400°C. At the high temperature range (-400°C) where entropy effects dominate, the ternary alloys display even higher decomposition pressures than the binary Mg2AI3. Below about 350°C two plateaux regions appear in the P-C isotherms of the binary Mg2AI3 and the indium containing samples. This phenomenon occurred reproducibly for various alloys preparations applied in
Hydrogenation of Mg-AI intermetallic compounds
H/Mg Fig. 6. Pressure composition desorption isotherms of binary Mg2AI3/3-phase hydrogen system. II, 335°C;e , 350°C; &, 375°C;O, 410°C.
H/Mg Fig. 7. Pressure composition desorption isotherms of indium containing/3-phase hydrogen system. II, 324°C: &, 345°C;e, 358°C; O, 384°C,
Z. GAVRA et aL
-330°C to various H/Mg compositions ratios exhibited in their X-ray diffraction patterns only the metallic products appearing in eqn (2). So far, we have no conclusive explanation to that peculiar behavior. The appearance of two plateaus did not occur in the nickel containing samples. Also, the isotherms illustrated by Ref.  did not display such phenomenon. It is possible that the lower (more stable) decomposition plateaus do not represent equilibrium conditions. These lower plateaus did not fit the linear relation of l n P vs 1/T, extrapolated from the high temperature region, while the upper plateaus fitted well that relation.
D.. I( Acknowledgements--Thanks are due to Mr. D. Brami and Mr. H. Maman for their assistance in the experiments performed in the present study. The authors are also grateful to Mr. S. Nathan for his X-ray diffraction measurements.
H/Mg Fig. 8. Pressure composition desorption isotherms of nickel containing B-phase hydrogen system. O, 330°C; II, 350°C; &, 375°C; O, 395°C.
the present study. Attempts to relate the two plateaus to different decomposition reactions (i.e. in addition to reaction (2), another reaction which yields other metallic products) were not supported by the results of X-ray diffraction. Samples which were partially dehydrided at
1. J. J. Reilly and R. H. Wiswall, Inorg. Chem. 7, 2254 (1968). 2. J. J. Reilly and R. H. Wiswal, Inorg. Chem. 6, 2220 (1967). 3. D. L. Douglass, MetalL Trans. 6A, 2179 (1975). 4. M. H. Mintz, S. Malkiely, Z. Gavra and Z. Hadari, J. lnorg. Nucl. Chem. 40, 1949 (1978). 5. J. J. Reilly and R. H. Wiswall, Rep. BNL-21322, May 1976. 6. M. H. Mintz, Z. Gavra, G. Kimmel and Z. Hadari, J. LessCommon Met. 74, 263 (1980). 7. F. G. Eisenberg, D. A. Zagnoli and J. J. Sheridan, J. Less Common Metals 74, 323 (1980). 8. For a comprehensive discussion on the Mg-AI phase diagram, see: L. F. Mondolfo, Aluminium Alloys: Structure and Properties, pp. 311-317. Butterworths, London (1976). 9. M. H. Mintz, Z. Gavra and Z. Hadari, J. lnorg. NucL Chem. 40, 765 (1978). 10. See for example: J. H. Sharp, G. W. Brindley and B. N. N. Achar, J. Am. Chem. Soc. 49, 379 (1966). 11. J. F. Stamper Jr., C. E. Holley Jr. and J. F. Suttle, J. Am. Chem. Soc. 82, 3504 (1960).