Stoichiometry in Inter-Metallic Compounds for Hydrogen Storage Applications

This chapter is devoted to the discussion of the influence of stoichiometry deviation on the structure, gaseous phase hydrogen storage, and electrochemical properties of some important inter-metallic compounds (IMC). First, the hydrogen storage characteristics and other important features of the IMC will be reviewed and then there will be a discussion of the AB5, AB2, A2B7, and AB metal hydride (MH) alloy systems with variations in stoichiometry. The experiments that provide the basis for this chapter’s discussion are described in detail (experimental set-up and chosen parameters) in earlier publications (Young, et al., 2008, 2009a, 2010a).


Introduction
This chapter is devoted to the discussion of the influence of stoichiometry deviation on the structure, gaseous phase hydrogen storage, and electrochemical properties of some important inter-metallic compounds (IMC). First, the hydrogen storage characteristics and other important features of the IMC will be reviewed and then there will be a discussion of the AB 5 , AB 2 , A 2 B 7 , and AB metal hydride (MH) alloy systems with variations in stoichiometry. The experiments that provide the basis for this chapter's discussion are described in detail (experimental set-up and chosen parameters) in earlier publications (Young, et al., 2008(Young, et al., , 2009a(Young, et al., , 2010a. The formation of an IMC between two metallic elements with a fixed stoichiometry depends on the heat of formation of the alloy (ΔH a ). With a very positive value of ΔH a , the two metallic elements in the alloy are immiscible in the liquid and will stay separate during cooling. As the ΔH a value decreases but is still positive, a eutectic reaction occurs during cooling, and mixtures of each element with limited solubility of the other is formed. As ΔH a further decreases to a negative value, a solid solution with a broad range of composition between the two elements is expected. Finally, when ΔH a becomes very negative, an IMC with a fixed stoichiometry is obtained. Based on the experimental data, ΔH a of an IMC can be predicted from the electron density at the boundary of the Wigner-Seitz atomic cell and the chemical potential of the conducting electrons (De Boer, et al.,1988). Therefore, the stability of an IMC with a stoichiometry of A m B n can be calculated theoretically. Among the choices of crystal structures available to certain stoichiometries of an IMC, the one that minimizes the total electron energy dominates. Parameters such as average electron density, atomic radius ratio, stoichiometry, and difference in electronegativity all contribute to the determination of crystal structure.
In contrast to solid solution, the constituent elements for binary IMC usually have large differences in both size and electronegativity. When the electron transfer takes place, the size of the atoms will change toward the ideal ratio. In the case of IMC, the electron being transferred from the element with the lower electronegativity (usually the larger one) to the other element causes the atomic volume to shrink (Pauling, 1987). This charge transfer changes the local electron density and is very important for the storage of hydrogen.
IMCs can have either a limited or a broad range in stoichiometry (defined as the ratio between constituent elements). For example, all the IMCs in the La-Ni binary phase diagram ( Fig. 1) exist at vertical lines, or in other words, there is a unique value of the stoichiometry (ratio between Ni and La) for each IMC; in contrast, the IMCs ZrNi 5 , ZrNi 3 , and Zr 7 Ni 10 in the Zr-Ni binary phase diagram (Fig. 2) exist at ranges of composition. The line width (also called solubility for a particular IMC) can be traced to the formation of defects and lattice-adjusted contraction (Thoma & Perepezko, 1995). When there are more A atoms present in the IMC, part of the excessive A atoms either move to B-site (anti-site defect) or create vacancy in B-site (vacancy defect) while still maintaining the same crystal structure. Although the radii ratio between A atom and B atom (R A /R B ) might not be optimum for the structure, the A-B distance will be adjusted by the defect formation and result in a close-to-ideal lattice. For example, the Laves phase IMC TiCr 2 with a less than the ideal ratio of 1.225 (ideal is 1.5 ) has a larger solubility on the A-rich side of the AB 2 stoichiometry (Chen, et al., 1994).  (Massalski, 1990)). Fig. 2. Ni-Zr binary phase diagram (data from (Massalski, 1990)). www.intechopen.com

Metal hydride alloys
Three different types of chemical bonds can be formed between hydrogen and other elements: the ionic bond, the metallic bond, and the covalent bond for elements with small (Li, Na, for example), medium (Ti, V), and large (N, O) electronegativities, respectively. While the strengths of both ionic and covalent bonds are too large for them to be appropriate for reversible hydrogen storage, the strength of metallic bond can be adjusted for the purpose of hydrogen storage depending on the operation temperature of the application. One way to quantify the metal-hydrogen (M-H) bond strength is by evaluating the heat of hydride formation (ΔH). An application for room temperature, the rechargeable battery for example, would require a ΔH value of between −35 to −30 kJ mol −1 H 2 . Only expensive metals, such as Pd and V, have the suitable M-H bond strength for room temperature applications. Therefore, the search for hydrogen storage metals was expanded firstly to binary IMCs, such as Mg 2 Ni, LaNi 5 , and TiNi, and then to more complicated and multi-element IMCs.
When hydrogen enters into the bulk of the metal, it contributes its only electron to the conduction band. The remaining proton then goes into the interstitial site among four or six host elements, causing an expansion in the crystal lattice. The metal hydride formed this way, shown in Fig. 3, has the same or very similar crystal structure as the host metal. Driven by a concentration gradient, the proton jumps to the neighboring storage sites through quantum mechanical tunneling. The moving path of the proton in AB 2 C14 MH alloy is along the vertical (c-) axis (Fig. 3a), which is different from its movement on the horizontal (a-b) plane in AB 5 MH alloy (Fig. 3b). Therefore, a flattened and an elongated unit cell along the c-axis are preferable for the ease of proton movement in AB 2 and AB 5 metal hydride (MH) alloys, respectively (Osumi, et al., 1983, Young, et al., 2009b. The process of hydride formation can be described by the pressure-concentrationtemperature (PCT) curves shown in Fig. 4, where the logarithm of the equilibrium hydrogen pressure is plotted against the hydrogen concentration in metal. At a fixed temperature, when the concentration of hydrogen in the bulk is low, the protons are randomly distributed throughout the host metal (α-phase). As the applied hydrogen pressure increases, more hydrogen atoms enter into the metal and start to form MH clusters, similarly to the beginning of ice formation in the case of freezing water. The co-existence of metal (α-phase) and MH (β-phase) eliminates the freedom in pressure and thus contributes to the flat region (plateau) of the PCT curve. With further increase in hydrogen pressure, the transformation into β-phase is complete, marks the end of the pressure plateau, and isotherm becomes more vertical. The The PCT curves can be measured with stepwise increases (absorption) or decreases (desorption) in applied hydrogen pressure. These two curves do not align with each other, and the difference in the plateau pressures of the two curves is called the hysteresis of PCT curve. The hysteresis represents the resistance of the metal against the pushing in or pulling out of additional hydrogen atoms. The correlation between the hysteresis and the MH pulverization rate through hydride/dehydride cycles is well established: MH alloys with small hysteresis are less prone to break into small pieces during cycling. Each set of the PCT curves (absorption and desorption) is obtained at one particular temperature. As the temperature increases, T 3 > T 2 > T 1 in Fig. 4, the following occur: the regions of αand β-phases extend into the plateau region due to the higher entropy in the system, the plateau region shortens, the plateau pressure increases, and hysteresis, reversible and maximum hydrogen storage capacities reduce. By connecting all the boundary points between α -phase, α + β mixed phase, and β-phase, a dome-shaped curve is obtained. As the temperature rises above a certain critical value (T 3 in Fig. 4), the plateau region of the PCT curve completely disappears. This critical temperature is very materialdependent. In the case of LaNi 5 based MH alloys, the systems are very ordered, and the critical temperatures are very high. Therefore, most of the PCT curves observed around room temperature are very flat. On the other hand, the multi-phase (Ti, Zr)Ni 2 based MH alloys have lower critical temperatures due to their high degrees of disorder inherent with the alloys and produce PCT curves without recognizable plateau region. Slope factor, defined as the percentage of storage capacity within a pressure range (usually from 0.01 to 0.5 MPa), is used to quantify the slope of the PCT isotherm.
where R is the ideal gas constant and T is the absolute temperature. When the logarithm of the pressure is plotted against the reciprocal of the absolute temperature as shown in Fig. 4, www.intechopen.com ΔH and ΔS can be calculated from the slope and the y-axis intercept of the line, respectively.
The value of ΔH can be used to quantify the M-H bond strength. MH with a lower ΔH (more negative) will release more heat during the formation of a more stable hydride with a lower equilibrium pressure, indicating both a stronger M-H bond and a larger interstitial site for hydrogen. For an ordered MH, ΔS is around the standard entropy of hydrogen gas (−130 J K −1 mol −1 H 2 ). ΔS will be higher for a less ordered MH.
Besides the application of solid hydrogen storage, MH alloys also play a very important role in nickel/metal hydride (Ni/MH) rechargeable battery as the negative electrode. The chemical reaction during charge/discharge is: During charge, the water molecule near the surface of the MH electrode splits into a proton (H + ) and a hydroxide ion (OH − ) by the applied voltage. While OH − moves into the electrolyte, H + enters into the bulk of the MH and is neutralized by the electron coming from the current collector. Charging/discharging in the electrochemical environment is very similar to hydriding/dehydriding in the gaseous phase; these two reactions can be linked by the Nernst Equation: Open circuit voltage vs. Hg/HgO = −0.934 − 0.0291 log P H2 where Hg/HgO is the reference electrode commonly used in alkaline solution. This correlation explains the similarity between the PCT curves ( Fig. 4) and the electrochemical charge/discharge curves illustrated in Fig. 5. The origin for the voltage difference between charge and discharge curves is quite complicated and contains at least the ohmic resistance in the connector and substrate, resistance from surface reaction, limited hydrogen diffusion, and electrolyte concentration difference. At a higher rate (gray curves in Fig. 5), the charge voltage is higher while the discharge voltage and the obtained capacity are lower, mostly due to the increase in current (i) as V = OCV + iR (during charge) and (4) where OCV is the open circuit voltage and R is the internal resistance.
Currently, there are four main MH alloy families used or proposed to be used as the negative electrode for Ni/MH rechargeable battery: the AB 5 , AB 2 , A 2 B 7 and AB alloy families. Rare earth element-based AB 5 alloy family is the most commonly used, with the misch-metal (Mm) (mixture of light rare earth elements, such as La, Ce, Nd, and Pr) as the A-elements and Ni, Mn, Al, and Co as the B-elements. Ni is an indispensably key element in alkaline electrochemistry due to its corrosion resistance and catalytic nature for water splitting/recombination. Other modifiers all have different contributions. The main advantages of AB 5 MH alloy are its high power capability as a result of the high Ni-content and superior cycle life. The transition metal-based AB 2 alloy (with Zr and Ti as the A-elements and B-elements similar to those of the AB 5 alloy) offers a reduction in raw material cost due to www.intechopen.com the elimination of rare earth elements and also has the advantage in storage capacity. In A 2 B 7 MH alloy, a small amount of Mg is added to the mixture of A-elements (which is composed of rare earth elements) to stabilize the crystal structure. The special super-lattice structure of A 2 B 7 MH alloy provides a compromise in performance between AB 2 and AB 5 . The last alloy family for Ni/MH application is the Mg-Ni based AB alloy. The much weaker M-H bond of Mg when compared to other A-elements, such as La, Ce, Zr, and Ti, reduces the desirable stoichiometry down to 1:1. The abundance of the raw material and high storage capacity of AB alloy make it a potential candidate to further reduce cost and improve capacity. The impact of stoichiometry on the structures, gaseous phase hydrogen storage properties, and electrochemical properties for these four alloy families will be presented in the remaining sections.

Stoichiometry in LaNi 5 -based metal hydride alloys
According to the phase diagram ( Fig. 1), eight IMCs exist in the La-Ni binary system. As the Ni-content increases, the ΔH of the hydride increases and forms a straight line in Fig. 6. Since the operational conditions for Ni/MH battery are room temperature and 1 atm, the suitable ΔH value for the MH alloy is around −40 kJ mol −1 H 2 (by assigning ΔS to the average value of −130 J K −1 mol −1 H 2 in formula (1)). In the IMCs of La and Ni, only ΔH of LaNi 5 falls into the range suitable for room temperature Ni/MH application. As a result, LaNi 5 -based MH alloy was chosen as the negative electrode in Ni/MH battery from early on and still dominates today's market.
The following is a description of the sample preparation for the experimental work backing this publication. Five Mm-based AB 5 alloys were prepared by arc melting. The composition of ingot prepared by arc melting is not as uniform as the commercial product produced by induction melting. The designed compositions of these alloys are listed in Table 1

Influences on microstructure
In Fig. 1, LaNi 5 is represented by a straight line in the phase diagram and shows no solubility at room temperature. If the liquid is quenched very quickly, some hyperstoichiometry is possible, as can be seen from the dash line near 1270°C; however, annealing and slow cooling will eliminate this hyper-stoichiometry. After annealing, a La 2 Ni 7 and a Ni phase are expected in the cases of hypo-and hyper-stoichiometries, respectively. In a commercial AB 5 alloy suitable for general performance requirements, A is mainly Mm (combination of La, Ce, Pr, and Nd) while B contains Ni, Co, Mn, and Al. The multi-element nature of the alloy system will increase the solubility of particular phases since anti-site or vacancy defects become easier to form in multi-element systems.
The phases of the five AB 5 alloys were first studied by x-ray diffractometer (XRD); the results are presented in Fig. 7. Almost all peaks can be attributed to a CaCu 5 structure. Ni peak starts to appear when the stoichiometry becomes ≥ 5.0. The lattice parameters and ratio, unit cell volume, and phase abundances obtained from XRD for each alloy are listed in  Table 2. Lattice constants and ratios, unit cell volumes, and phase abundances of CaCu 5 and Ni of AB 5 alloys from XRD analysis.
The microstructures were further studied by scanning electron microscope (SEM); the resulting backscattering electron images (BEI) are shown in Fig. 8. These micrographs were chosen due to the presence of multi-phase features and do not represent the general phase distribution of the alloys. Four distinctive phases, A, B, C, and D, can be identified from these micrographs. Phase A, with the brightest contrast due to its high average atomic weight, is the misch-metal with the composition very close to the raw material as indicated by x-ray energy dispersive spectroscopy (EDS) analysis. Phase B has a composition close to AB, which is the phase on the left-hand side of the eutectic point in Fig. 1. The AB phase www.intechopen.com abundance is too small to be detected by XRD analysis. Phase C is the main AB 5 phase, where the B/A stoichiometric ratio increases from 5.17, 5.29, 5.49, 5.48, finally to 5.47 with the increase in designed B/A values. In other words, the solubility of the AB 5 phase extends to about 5.5 in this case as the alloy's B/A stoichiometric ratio increases. Further increase of the B-content in the alloy promotes the formation of Ni-phase as seen from Fig.1 and Table  2. Phase D with the darkest contrast is a mixture of transition metals and has an average composition of Al 9 Mn 21 Co 20 Ni 50 , which is high in Al and Mn and low in Ni when compared to the raw material. www.intechopen.com

Influences on gaseous hydrogen storage
The gaseous phase hydrogen storage characteristics were studied by PCT measurements performed at 30 and 60°C. The 30°C PCT isotherms from the five AB 5 alloys are plotted in Fig. 9 and summarized in Table 3. As the designed B/A stoichiometric ratio increases, both the maximum and reversible hydrogen storage capacities measured at 30°C first increase and then decrease. This trend is the same as that in unit cell volume. The larger unit cell contributes to a more stable hydride with lower equilibrium pressure and higher storage capacity. The isotherm from AB4.9 is flatter than those from other alloys, indicating a higher uniformity in the main phase. The PCT hysteresis from AB4.9 is the highest and corresponds to its lowest c/a ratio among all alloys, suggesting a more difficult tunneling path for protons to hop within the structure. ΔH and ΔS were calculated from the plateau pressures measured at 30 and 60°C by formula (1) and are listed in

Influences on electrochemical properties
The discharge capacities for the five AB 5 alloys were measured at two discharge current densities (100 and 8 mA g −1 ), and the results are listed in Table 4. The full capacity, defined as the capacity measured at the slowest discharge rate (8 mA g −1 ), increases first and then decreases with the increase in B/A stoichiometric ratio. The stoichiometric AB5.0 alloy shows the highest capacity. The high-rate dischargeability (HRD) values (ratios between the capacities measured at 100 and 8 mA/g discharge currents) measured at the third cycle are listed in Table 4. The maximum HRD value was obtained from a slightly hyperstoichiometric AB5.1 alloy.

Alloy
Full capacity (mAh g -1 ) High-rate capacity (mAh g -1 ) Ratio (High-rate dischargeability) OCV at 50% DOD OCV measured at 50% depth-of-discharge (DOD) of the MH electrode against the Hg/HgO reference electrode for each alloy is listed in Table 4. In the AB 5 alloy series, the changes in OCV are small due to their similar plateau pressures. In general, OCV decreases with the increase in B/A stoichiometric ratio, which correlates well with the increase in plateau pressure by formula (3). Lower OCV value is a desirable trait because it will contribute to higher cell voltage and better power in the sealed battery configuration.

Stoichiometry in Laves phase-based AB 2 metal hydride alloys
Compared to La, Ti has a lower M-H bond strength and thus requires a smaller amount of Ni to adjust the heat of hydride formation to fall within the desired range. According to the graph that plots ΔH vs. atomic percentage of Ni (Fig. 6), a stoichiometry of approximately AB 2 would be suitable for room temperature Ni/MH application. However, the IMC of TiNi 2 is not available according to the Ti-Ni binary phase diagram. As seen from the Zr-Ni phase diagram in Fig. 2, Zr, a common element used to partially replace Ti in order to adjust the hydrogen storage, does not have an IMC at ZrNi 2 either. Since TiCr 2 , ZrCr 2 , TiMn 2 , and ZrMn 2 are all available based on the phase diagrams (Fig. 10, for example), the design of AB 2 alloy starts from the formulation of (Ti, Zr)(Ni, Cr, Mn) 2 . There are three AB 2 Laves phases, namely, C14, C15, and C36. From the Zr-Cr phase diagram in Fig. 10, the order of phase formation during cooling is C14 first, followed by C36, and finally the C15 phase. Due to the existence of defects, ZrCr 2 has a broad range of solubility, allowing the designed composition to be off-stoichiometric.
Five transition element-based AB 2 alloys were prepared by arc melting, and the designed compositions are listed in Table 5. The designed B/A stoichiometric ratio varies from 1.8 (hypo) to 2.2 (hyper). The average electron density (e/a) of each alloy, based on the numbers of outer-shell electrons of the constituent elements, is listed in Table 5 and can be used to predict the dominant phase (C14 or C15) (Shaltiel, et al., 1977). The base alloy, AB2.0, is rich in C14 phase and a candidate for replacing AB 5 alloy in Ni/MH batteries. Its microstructure was studied extensively with a combination of SEM and transmission electron microscope previously . In order to preserve the secondary phases, which serve as an electrochemical catalyst, these AB 2 alloys were not annealed (Young, et al., 2010b). Fig. 10. Zr-Cr binary phase diagram (data from (Massalski, 1990) Fig. 11 compares the XRD patterns from the five AB 2 alloys. As the e/a value increases, the main phase shifts from C14 to C15. Besides the main Laves phases, TiNi phase, which is associated with the solid-state transformation of non-Laves secondary phases during cooling, is also present. The C14 lattice constants a and c, c/a aspect ratio, unit cell volume, and phase abundances of each alloy are listed in Table 6. In this series of AB 2 alloys, a and c www.intechopen.com decrease monotonically with the increase in stoichiometry. In the case of hypostoichiometry, the excessive A-site atoms move to the B-site and slightly enlarge the unit cell. On the contrary, the larger B-site elements (however, still smaller than the A-site elements) will occupy the A-site and shrink the unit cell in the case of hyper-stoichiometry (Kodama, et al., 1995). The stoichiometric AB2.0 alloy has the highest or close to the highest c/a aspect ratio, so theoretically it should be the least stable alloy during hydride/dehydride cycling. The C14 unit cell volume decreases with the increase in B/A stoichiometric ratio. The smaller unit cell will contribute to an unstable hydride with higher plateau pressure and lower storage capacity. As the B/A stoichiometric ratio increases, more of the main C14 phase was replaced by the C15 phase due to the increase of the e/a value. The phase abundance of TiNi also reduces with the increase in B/A stoichiometric ratio. The TiNi phase has a higher A-atom content than the Laves phase does and therefore is not favorable in the case of B-rich hyper-stoichiometry. Fig. 11. XRD patterns of AB 2 alloys using Cu-K α as the radiation source. Four regions with different contrasts can be observed in most of the SEM BEI micrographs shown in Fig. 12. Region A with the brightest contrast is the Zr 7 Ni 10 phase, which was not observed in XRD due to its small abundance. The light gray region B has a composition close to AB and can be related to the TiNi phase found in XRD analysis. Both regions A and B, with nominal compositions of A 7 B 10 and AB, become smaller and even diminish as the B/A stoichiometric ratio increases. Region C is the main AB 2 phase. As the designed B/A stoichiometric ratio increases, the B/A stoichiometric ratio in the main AB 2 phase evolves from 1.85 to 1.97, 2.05, 2.01 and 1.99. This result demonstrates that the AB 2 phase has a wider solubility on the hypo-side than on the hyper-side in the C14-predominaing case. The last feature of the micrographs is the dark spot with very high Zr-content, which is identified as ZrO 2 inclusion (region D). This oxide is formed through oxidation of Zr segregated out from the melt; it is not the residual oxide in the raw material as reported before in a study of ZrO 2 inclusions in the Al-containing AB 2 alloys [Young, et al., 2010c]. As the B/A stoichiometric ratio varies, the size and amount of ZrO 2 inclusions do not change.

Influences on gaseous hydrogen storage
The PCT absorption and desorption isotherms measured at 30°C for the five AB 2 alloys are plotted in Fig. 13 and summarized in Table 7. As the B/A stoichiometric ratio increases, the maximum hydrogen storage capacity first increases and then decreases. The highest maximum capacity was obtained from AB1.9 alloy. For the alloys with B/A stoichiometric ratio ≥ 2.0, the maximum capacities decrease as a result of the shrinking lattice unit cell. In alloy AB1.8, the phase abundance of C14, which has less storage capacity compared to C15 (Young, et al. 2010b), is the highest and thus decreases the maximum storage capacity. The relationship between B/A stoichiometric ratio and reversible gas phase storage is similar to what is observed in maximum storage capacity: as the B/A stoichiometric ratio first increases, the reversible storage increases due to the higher plateau pressure and the higher C15 phase abundance with higher gaseous reversibility (Young, et al. 2010b). With further increase in B/A stoichiometric ratio, the reversible storage capacity decreases for the same reason as the maximum storage capacity. The maximum reversible capacity happens at AB2.0. Due to the high degree of disorder in AB 2 MH alloys, the critical temperature (top of the dome expressed by a dashed line in Fig. 4) is much lower than it is in AB 5 MH alloys. Therefore, all the PCT isotherms lack a well-defined plateau pressure, or in other words, a metal (α)-hydride (β) co-existing region. In order to compare the hydrogen equilibrium pressures among the alloys, the mid-point pressure, defined as the pressure where half of the desorption is complete, is calculated and listed in For AB2.0-stoichiomtery alloys, the hysteresis is related to the average electron density (Young, et al. 2009b). The PCT hysteresis first decreases and then increases with the increase in B/A stoichiometric ratio. The hysteresis minimum happens at a B/A stoichiometric ratio of 2.0. Therefore, from a cycle stability point of view, the stoichiometric AB2.0 alloy is preferable and contradicts from the c/a ratio prediction in the XRD analysis. ΔH and ΔS values, calculated from the mid-point pressures of 30 and 60°C desorption isotherms, are listed in Table 7. The value of -ΔH decreases as the B/A www.intechopen.com stoichiometric ratio decreases, which implies the weakening of the M-H bond strength and correlates well with the shrinking lattice. The measured ΔS values from these AB 2 alloys are similar and higher than those in AB 5 alloys (around −105 J K −1 mol −1 H 2 ), which suggest higher degrees of disorder in AB 2 alloys and are consistent with the steeper PCT isotherms. Fig. 13. PCT isotherms measured at 30°C for AB 2 alloys. Open and solid symbols represent data taken during absorption and desorption, respectively.

Influences on electrochemical properties
The discharge capacities for the five AB 2 alloys were measured, and the results are listed in Table 8. The full capacity increases first and then decreases with the increase in B/A stoichiometric ratio. The stoichiometric AB2.0 alloy shows the highest capacity. The HRD values measured at the tenth cycle are listed in Table 8. The maximum value occurs at AB2.1. Similar to the case of AB 5 alloys, slightly hyper-stoichiometry increases the HRD.
OCV measured at 50% DOD of the MH electrode against Hg/HgO reference electrode for each alloy is listed in Table 8. In the AB 2 alloy series, OCV decreases with the increase in B/A stoichiometric ratio, which correlates well with the increase in mid-point pressure by formula (3). This observation is consistent with that from the AB 5 alloy series. Similar influences of stoichiometry on various properties of other two series of AB2 alloys predominated by C15 and C14/C15 phases were also reported previously (Young, et al. 2011b(Young, et al. & 2011c

Stoichiometry in NdMgNi-based A 2 B 7 metal hydride alloys
La 2 Ni 7 is an IMC with slightly lower Ni-content than LaNi 5 as shown in the La-Ni phase diagram (Fig. 1). ΔH of La 2 Ni 7 is below the suitable range for room temperature operation as shown in Fig. 6. In order to increase ΔH, a small amount of Mg is added to replace La and brings ΔH into the suitable range for room temperature operation. So far, three MH alloys of the A 2 B 7 family have been used as the negative electrode in Ni/MH rechargeable battery. Kohno et al. first introduced La 5 Mg 2 Ni 23 with a potential capacity of over 400 mAh g −1 (Kohno, et al., 2000) (Mm, Mg)(Ni, Co, Al) 3.3 alloys were used in high-capacity 2700 mAh AA-size battery by Sanyo (Yasuoka, et al., 2006) where Mm is the misch metal. Finally, (Nd, Mg) 2 (Ni, Al) 7 alloys were used for the low self-discharge AA battery also introduced by Sanyo.  Table 9. These alloys were annealed in 1 atm of argon at 900°C for 5 hours.  Table 9. Designed compositions in at. % and B/A stoichiometry ratios of A 2 B 7 alloys.

Influences on microstructure
The Nd-Ni phase diagram is very similar to that of La-Ni shown in Fig. 1. Under the equilibrium condition, the Nd 2 Ni 7 phase has a very limited solubility. The influence of offstoichiometry to microstructure was studied by XRD analysis, and the results are plotted in Fig. 14. Besides the reflection peaks from the Nd 2 Ni 7 main phase, secondary phases of Nd, NdNi, NdNi 2 , and NdNi 5 can be identified from the patterns. The lattice constants of the Nd 2 Ni 7 phase were calculated and are listed in Table 10 together with the phase abundances of all observed phases. In general, as the B/A stoichiometric ratio increases, both c/a ratio and the unit cell volume decrease as seen from the shifting in the peak into higher angle illustrated by the line shown in Fig. 14 Region C is the main Nd 2 Ni 7 phase and the B/A stoichiometric ratio changes from 2.91 to 3.27, 3.33, 3.33, and 3.32 as the designed B/A stoichiometric ratio increases from 3.3 to 3.4, 3.5, 3.6, and 3.7. With the incorporation of Mg in the Nd-site, the stable main phase stoichiometry is between 2.91 and 3.33. Since the extra B elements (Ni and Al) are likely to be pushed to NdNi 5 , which is rich in B, the B/A stoichiometric ratio in the main phase is lower compared to the designed ratio. Furthermore, from the XRD analysis, the NdNi 5 phase can be observed in most alloys, and the amount of NdNi 5 phase increases with the increase in B/A stoichiometric ratio. Only the alloy with a designed B/A stoichiometry of 3.3 has too low of NdNi 5 -content to be detected by XRD. The Mg-content in the main phase is about 3.5%, which is higher than the average Mg-content of the designed compositions. The fourth phase, region D, has a composition close to NdNi 2 with the highest Mg-content of about 12 to 15%. The SEM micrographs show appreciable amount of region D in all alloys. However, the XRD analysis only shows a small amount of NdNi 2 in the last two alloys, AB3.6 and AB3.7. The areas where the SEM micrographs were taken are usually those with the most numbers of phases and should not be used to quantify the phase abundances. Region E, with the darkest contrast, is from ZrO 2 . Fig. 15. SEM backscattering images for A 2 B 7 alloys. Letters A, B, C, D, and E indicate Nd metal, NdNi, Nd 2 Ni 7 , NdNi 2 , and ZrO 2 , respectively. www.intechopen.com

Influences on gaseous hydrogen storage
The PCT isotherms measured at 30°C for three five A 2 B 7 alloys are shown in Fig. 16 and summarized in Table 11. As the B/A stoichiometric ratio increases, the maximum hydrogen storage capacity increases in the beginning and stabilizes when the designed stoichiometry becomes ≥ 3.5; the reversible hydrogen storage capacity increases; the mid-point pressure increases by a small amount; the isotherm becomes flatter; and the hysteresis, ΔH, and ΔS remain about the same. The changes in maximum hydrogen storage capacity are similar to those in B/A stoichiometric ratio in the main Nd 2 Ni 7 phase. The composition in the main phase is optimized for maximum storage capacity at B/A = 3.33. Although both the reversible capacity and the plateau pressure increase with the increase in B/A stoichiometric ratio, the increase in plateau pressure is too small to account for the large increase in reversible storage capacity. The secondary phases such as NdNi 2 and NdNi 5 are most likely to contribute to the increase in reversible hydrogen storage as the case of the ZrNi and Zr 7 Ni 10 phases in AB 2 alloy (Young, et al., 2011a). Comparing the PCT results among the AB 5 , AB 2 , and A 2 B 7 families yields the following observations: AB 2 has the highest maximum storage capacity, AB 5 has the flattest plateau due to the smallest amount of secondary phases, and A 2 B 7 has the most ordered MH structure as seen from its lowest ΔS value. Fig. 16. PCT isotherms measured at 30°C for A 2 B 7 alloys. Open and solid symbols represent data taken during absorption and desorption, respectively.  Table 11. Summary of gaseous phase hydrogen storage properties of A 2 B 7 alloys.

Influences on electrochemical properties
Similar to the case of AB 5 alloys, these A 2 B 7 alloys were easy to activate. All alloys reached the highest capacities within the first three cycles without any pre-treatment. The full and high-rate discharge capacities at the third cycle for each alloy are listed in

Stoichiometry in MgNi-based AB metal hydride alloys
MgNi-based MH alloys have attracted much of researchers' attention due to the abundance of Mg and Ni resources. Mg metal forms a stable hydride of MgH 2 which only desorbs hydrogen at temperature above 400°C. According to the schematic of ΔH vs. Ni % in Fig. 6, a stoichiometry of Mg : Ni close to 1:1 is suitable for room temperature operation. However, according to the Mg-Ni binary phase diagram illustrated in Fig. 17, only two IMCs are allowed at room temperature under the equilibrium condition, namely, Mg 2 Ni and MgNi 2 . The former is a MH alloy but only desorbs hydrogen at 250°C; the latter will not form any stable hydride at 1 atm. The desirable MgNi single phase (indicated by the gray arrow in Fig. 17) is not allowed. Therefore, quick quench methods under the non-equilibrium conditions, such as RF sputtering, laser ablation, and combination of melt-spin and mechanical alloying, were used to prepare the MgNi alloys for hydrogen storage and Ni/MH rechargeable battery applications.
Five thin films were prepared with the RF-sputtering technique simultaneously with a compositional gradient target, 20 mtorr Ar gas background, 200 watt of rf power in a MRC 8" multi-target system. The substrate was a Ni tab taped down to a copper plate maintained at 0°C by circulating water through the imbedded pipeline. The composition varies from Mg : Ni = 48:52 to 52:48, 57:43, 61:39, and 65:35. The deposition time was 2 hours, and the average film thickness was about 1.5 microns. Fig. 17. Mg-Ni binary phase diagram (data from (Massalski, 1990)).

Influences on microstructure
The XRD patterns obtained from these five thin films on Ni substrate are shown in Fig. 18. The two strongest defection peaks are from the Ni-substrate. The microcrystalline phase in the film has two peaks at around 20 and 40° from the broadening of MgNi 2 and Mg 2 Ni peaks. The positions of microcrystalline peak will not change with the film composition. While the first broad peak of microcrystalline can be seen at 20°, the second broad peak of microcrystalline (around 40°C) is unable to be discerned since it overlaps with another broad peak from the amorphous phase in the film. The center of the amorphous peak can be related to the average distance between two neighboring atoms by the Ehrenfest formula (Guinier, 1963): 2 rsin θ = 1.2λ (6) where r is the average distance between two neighboring atoms and λ is the wavelength of the x-ray (in this case, 1.5418Å). The shift of the amorphous peak to the lower angle is due to the increase in the inter-atomic distance with the higher Mg (larger than Ni)content in the thin film. As the Mg content in the alloy increases above 52%, deflection peaks from the crystalline Mg 2 Ni start to appear. Film with a composition close to 1:1 (Mg 48 Ni 52 and Mg 52 Ni 48 ) shows mainly amorphous phase with smaller amounts of microcrystalline phase. With the increase in Mg-content, the Mg 2 Ni crystalline phase starts to form. Fig. 18. XRD patterns of MgNi thin films deposited on Ni-substrate using Cu-K α as the radiation source. The straight line illustrates the shift of amorphous peak into lower angle as the Mg-content in the film increases.

Influences on electrochemical properties
The film thickness and capacity measured at a discharged rate of 20 mA g −1 of each film are listed in Table 13. As the Mg-content increases, the film becomes thinner, and the discharge capacity first increases and then decreases. The maximum capacity happens at a composition of Mg 52 Ni 48 , which is mainly amorphous in nature. Both the microstructure and electrochemical capacity are strongly affected by the film composition, or in other words, stoichiometry.

Conclusion
The influences of stoichiometry on the structures, gaseous phase hydrogen storage properties, and electrochemical properties of four families of MH alloys were studied. In the cases of AB 5 , AB 2 , and A 2 B 7 , as the B/A stoichiometric ratio increases, the sizes of the unit cell of main phases decrease, resulting in less stable hydrides with higher hydrogen equilibrium pressures, the stoichiometry in the main phase increases and then becomes stabilized at the stoichiometric alloy. Further increases in the average stoichiometry increases the abundance of certain secondary phases but do not change the stoichiometry in the main phase. This suggests that the alloys under the current study have a wider composition range on the hypo-stoichiometric side (A-rich) than they do on the hyperstoichiometric side (B-rich). The stoichiometric alloys usually have the highest gaseous phase storage capacities (both maximum and reversible). Slightly over-stoichiometric alloys are more suitable for high-power applications due to their higher HRD values. In the case of AB alloys, both the microstructure and electrochemical properties are strongly related to the film composition (stoichiometry). The optimal composition is Mg 52 Ni 48 considering its high capacity and amorphous content.