Studies on Incorporation of Mg in Zr-Based AB 2 Metal Hydride Alloys

Mg, the A-site atom in C14 (MgZn2), C15 (MgCu2), and C36 (MgNi2) Laves phase alloys, was added to the Zr-based AB2 metal hydride (MH) alloy during induction melting. Due to the high melting temperature of the host alloy (>1500  ̋C) and high volatility of Mg in the melt, the Mg content of the final ingot is limited to 0.8 at%. A new Mg-rich cubic phase was found in the Mg-containing alloys with a small phase abundance, which contributes to a significant increase in hydrogen storage capacities, the degree of disorder (DOD) in the hydride, the high-rate dischargeability (HRD), and the charge-transfer resistances at both room temperature (RT) and  ́40  ̋C. This phase also facilitates the activation process in measurement of electrochemical discharge capacity. Moreover, through a correlation study, the Ni content was found to be detrimental to the storage capacities, while Ti content was found to be more influential in HRD and charge-transfer resistance in this group of AB2 metal hydride (MH) alloys.


Introduction
Laves phase-based AB 2 metal hydride (MH) alloy is one of the high-capacity negative electrode materials used in nickel/metal hydride (Ni/MH) batteries.Its reversible hydrogen storage capacity can be as high as 3 wt% [1], which is equivalent to an electrochemical capacity of 804 mAh¨g ´1.
The measured electrochemical discharge capacity can reach up to 436 mAh¨g ´1 [2], which is about 25% higher than the conventional AB 5 MH alloys based on rare earth metals (330 mAh¨g ´1) [3,4].Early in their development, AB 2 MH alloys suffered from a harder activation and a shorter cycle life when compared to AB 5 MH alloys [5][6][7][8].With composition and process refinement, the activation and cycle stability of AB 2 MH alloys as negative electrode active material improved substantially [9].However, the high-rate dischargeability (HRD) of the AB 2 MH alloys, especially at low temperature, is still significantly inferior to the AB 5 MH alloys because of the relatively low nickel content in the AB 2 alloy [10].Various additions, including transition metals-Al [11], Cr [12], Co [13], Cu [14,15], Fe [16], Mo [17,18], Zn [19], Pt [20], Pd [21,22], rare earth metals-Y [23], Ce [24], La [25,26], and Nd [27], and others such as Si [28] and B [29], have been used to reduce the surface charge transfer resistance and increase the HRD of AB 2 MH alloys.In this paper, we summarize our findings regarding the use of one of the alkaline earth elements, Mg, as an additive in AB 2 MH alloys.
Mg can form various Laves phase alloys with different transition metals, such as C14 (MgZn 2 ), C15 (MgCu 2 ), and C36 (MgNi 2 ) [30].Mg-containing Mg 2 Ni, with an hP18 hexagonal structure (derivative of AlB 2 type [31]), is an important MH alloy that typically works in a temperature range of 200-250 ˝C.When the crystalline size decreases to the nanoscale or an amorphous state, Mg 2 Ni can be used as the negative electrode material in Ni/MH batteries [32][33][34][35][36][37][38][39][40][41].A more suitable stoichiometry for the Mg-Ni system is MgNi (1:1) but, unfortunately, it is not possible to obtain this material through conventional melt-and-cast, according to the phase diagram [42].Amorphous MgNi prepared by a combination of melt-spin and mechanical alloying can achieve an electrochemical capacity of 720 mAh¨g ´1 for the first cycle.However, it has very poor cycle stability [43] and is, therefore, the subject of a DoE-funded project [44,45].Reports of Mg use as a modifier in adjusting the hydrogen storage properties of AB 2 MH alloys are very scarce [46], which is very different from Mg-containing superlattice-based MH alloys (reviewed in [47]).Although Mg alone can form Laves phases, Mg is only slightly solubility in Zr(Ti)-based AB 2 phases (0.3 at%) and segregates into a Mg 2 Ni phase [46].The Mg 2 Ni secondary phase reduces the surface reaction current, but increases the charge retention [46].In addition to adding Mg to AB 2 alloy, we also investigated the role of Mg-addition (9.5 at%) in Zr 8 Ni 21 and found that the Mg-added alloy segregates into Zr 7 Ni 10 matrix, consisting of Zr 2 Ni 7 grains with occasional Mg 2 Ni inclusions, and Mg has a solubility of about 1.5 at% in the Zr 2 Ni 7 phase [48].Additionally, Mg added in Zr 8 NNi 21 alloy hindered the formation process, but increased the surface reaction exchange current [49].

Experimental Setup
An induction melting process involving an MgAl 2 O 4 crucible, an alumina tundish, a 2-kg furnace under argon atmosphere, and a pancake-shaped steel mold was used to prepare the ingot samples.MgNi 2 alloy was used as the Mg source, which was added in the final melting step.A 50% excess of Mg was added to compensate for evaporation loss.The ingots were first hydrided/dehydrided to increase their brittleness and then crushed and ground into ´200 mesh powder.The chemical compositions of the ingots were analyzed using a Varian Liberty 100 inductively coupled plasma-optical emission spectrometer (ICP-OES, Agilent Technologies, Santa Clara, CA, USA).A Philips X'Pert Pro X-ray diffractometer (XRD, Amsterdam, The Netherlands) was used to study the phase component.A JEOL-JSM6320F scanning electron microscope (SEM, Tokyo, Japan) with energy dispersive spectroscopy (EDS) was applied in investigating the phase distribution and composition.The hydrogen storage was measured using a Suzuki-Shokan multi-channel pressure-concentration-temperature system (PCT, Tokyo, Japan).In PCT analysis, each sample was first activated by a 2-h thermal cycle between room temperature (RT) and 300 ˝C under 2.5 MPa H 2 pressure, and then measured at 30, 60, and 90 ˝C.Details of the electrode and cell preparations, as well as the measurement methods, were previously reported [50,51].AC impedance measurements were conducted using a Solartron 1250 Frequency Response Analyzer (Solartron Analytical, Leicester, UK) with a sine wave amplitude of 10 mV and frequency range of 0.5 mHz to 10 kHz.Prior to experiments, electrodes were subjected to one full charge/discharge cycle at a rate of 0.1C, using a Solartron 1470 Cell Test galvanostat, discharged to 80% state-of-charge, and then cooled to ´40 ˝C.Magnetic susceptibility was measured using a Digital Measurement Systems Model 880 vibrating sample magnetometer (MicroSense, Lowell, MA, USA).

Results and Discussion
Six alloys were prepared by the induction melting technique.The design composition, with the ICP results, is summarized in Table 1.The Mg-free base alloy, Mg0, has been used numerous times in the previous comparison works [15][16][17][18][19].In the design, Mg-content was varied from 0 to 5 at% in alloys Mg0-Mg5, respectively.However, due to the strong rejection from the major phase, the Mg-content in the final alloys is in the range of only 0.6-0.8at%.To compensate for the increase in Mg-content in the design, both the Ti-and Ni contents were reduced.While the reduction in Ti content is clearly observed in ICP results, the reduction in Ni content is not obvious in alloys Mg1-Mg3, and the average Ni content actually increases in alloys Mg4 and Mg5 because of Mg loss.The average electron density (e/a), a strong factor in determining the C14/C15 phase abundance ratio [30,52], decreases monotonically in the design, but stabilized in the beginning and then increases in the ICP results, mirroring the evolution of Ni content.The B/A ratio, defined by the ratio of atomic percentage of B-sites (elements other than Zr, Ti, and Mg) and A-site atoms (Ti, Zr, and Mg), decreases in the design (hypo-stoichiometry), but stabilizes and then increases (hyper-stoichiometry) in the ICP results due to the increase in Ni content.The impact of stoichiometry on the performance of AB 2 MH alloys has been previously studied [53,54].In general, hypo-stoichiometry promotes the C14 phase, lowers the PCT plateau pressure, and decreases HRD.

X-Ray Diffractometer Analysis
XRD analysis is an important tool to study the multi-phase nature of the Laves phase MH alloys [55][56][57][58].The XRD patterns for alloys Mg0-Mg5 are shown in Figure 1.Peaks from the C15 cubic phase overlap with some of those from the C14 hexagonal phase.The TiNi phase with a B2 cubic structure can be seen in most of the XRD patterns.In addition to the C14, C15, and TiNi phases, one more cubic phase was observed in the Mg-containing alloys and it is believed to relate to Mg addition in the alloy.As the alloy number increases (Mg0 Ñ Mg5), the intensities of C14-only peaks (for example, the one near 39.5 ˝) decrease and the main C14/C15 peak (around 42.8 ˝) first shifts to the left (larger unit cell) and then shifts to the right (smaller unit cell), as indicated by the blue vertical line in Figure 1.electron density (e/a), a strong factor in determining the C14/C15 phase abundance ratio [30,52], decreases monotonically in the design, but stabilized in the beginning and then increases in the ICP results, mirroring the evolution of Ni content.The B/A ratio, defined by the ratio of atomic percentage of B-sites (elements other than Zr, Ti, and Mg) and A-site atoms (Ti, Zr, and Mg), decreases in the design (hypo-stoichiometry), but stabilizes and then increases (hyper-stoichiometry) in the ICP results due to the increase in Ni content.The impact of stoichiometry on the performance of AB2 MH alloys has been previously studied [53,54].In general, hypo-stoichiometry promotes the C14 phase, lowers the PCT plateau pressure, and decreases HRD.

X-Ray Diffractometer Analysis
XRD analysis is an important tool to study the multi-phase nature of the Laves phase MH alloys [55][56][57][58].The XRD patterns for alloys Mg0-Mg5 are shown in Figure 1.Peaks from the C15 cubic phase overlap with some of those from the C14 hexagonal phase.The TiNi phase with a B2 cubic structure can be seen in most of the XRD patterns.In addition to the C14, C15, and TiNi phases, one more cubic phase was observed in the Mg-containing alloys and it is believed to relate to Mg addition in the alloy.As the alloy number increases (Mg0 → Mg5), the intensities of C14-only peaks (for example, the one near 39.5°) decrease and the main C14/C15 peak (around 42.8°) first shifts to the left (larger unit cell) and then shifts to the right (smaller unit cell), as indicated by the blue vertical line in Figure 1.The lattice constants of the four phases obtained from the XRD analysis are listed in Table 2 with the crystallite size of the main C14 phase.With the increase in alloy number, the lattice constants of the C14 phase first increase and then decrease.The changes are very isotropic, as seen from the nearly unchanged a/c ratio.The crystallite size of the C14 phase decreases, and the lattice constants of C15 and TiNi follow the same trend as observed in the C14 main phase.The phase abundances of four constituent phases, obtained from a Rietveld refinement of the XRD patterns, are listed in Table 3.In general, as the alloy number increases, the C14 phase was replaced by the C15 phase and the TiNi phase abundance first increases and then decreases while the phase abundance of the Mg-related cubic phase remains unchanged.The evolution of the C14/C15 phase agrees with the changes in e/a and B/A (Table 1), because the C14/C15 phase determination threshold of e/a is approximately 6.9 in this case [52].

Scanning Electron Microscope/Energy Dispersive Spectroscopy Analysis
SEM back-scattering electron images (BEI) for alloys Mg1-Mg5 are shown in Figure 2. EDS was used to study composition information of representative spots with different contrasts on the BEI micrographs and the results are summarized in Table 4.As we know, EDS is a semi-quantitative analysis and results are only for comparison purpose.The microstructure of the Mg-free alloy (Mg0) was published before (as alloy Mo0 in [17]) and is composed of C14 and TiNi phases.In the microstructures of Mg-containing alloys, a C15 phase (judging from its relatively higher e/a value) with a slightly brighter contrast and an Mg-predominated phase with a darker contrast start to appear.The TiNi phase is usually surrounded by the C15 phase, since the cooling sequence is C14-C15-TiNi [59,60].According to the EDS results shown in Table 4, the Mg-content in the C14 phase is very small (0.2-0.3 at%), while that of the C15 phase is slightly higher (0.3-0.7 at%).In addition, the B/A ratios in the C14 phase are higher than those in the C15 phases.The initial increase followed by a decrease in the C14 lattice parameters found through XRD analysis can be explained by the balance between the decrease in the content of relatively small Ti (larger C14 unit cell) and the increase in B/A ratio (smaller C14 unit cell [61]).The B/A ratio in the TiNi phase is calculated based on the assumption of V occupying the A-site [27] and the results are still higher than 1, which indicates the possibility of other Zr x Ni y secondary phases with higher B/A ratios.The nature of the Zr x Ni y phase was studied before by transmission electron microscopy [62].The Mg-rich phase (the fourth phase in each BEI micrograph) has an Mg-content from 45.4 at% to 82.1 at%.It is difficult to link this phase to the cubic phase found by XRD since all alloys in Mg-Ni phase diagram are hexagonal (Mg, Mg 2 Ni, and MgNi 2 ).An MgIn 2 intermetallic alloy with a cubic structure and a lattice constant of 4.60 Å [63] was the prototype used in our XRD analysis.The phase with the bright contrast (the fifth phase in each BEI micrograph) has a relatively higher Zr-content and a B/A ratio close to 1.0, and therefore is identified as the ZrNi phase.It cannot be the mixture of Zr metal and neighboring Laves phase because of a relatively low V-content (similar to the case of TiNi phase).The corresponding XRD peak of this ZrNi phase cannot be identified due to low abundance.The Mg-contents in the TiNi and ZrNi phases are slightly higher than those in the C14 and C15 phases.and MgNi2).An MgIn2 intermetallic alloy with a cubic structure and a lattice constant of 4.60 Å [63] was the prototype used in our XRD analysis.The phase with the bright contrast (the fifth phase in each BEI micrograph) has a relatively higher Zr-content and a B/A ratio close to 1.0, and therefore is identified as the ZrNi phase.It cannot be the mixture of Zr metal and neighboring Laves phase because of a relatively low V-content (similar to the case of TiNi phase).The corresponding XRD peak of this ZrNi phase cannot be identified due to low abundance.The Mg-contents in the TiNi and ZrNi phases are slightly higher than those in the C14 and C15 phases.

Pressure-Concentration-Temperature Analysis
PCT measurement has been used extensively in the study of Laves phase MH alloys reacting with hydrogen gas [64][65][66][67][68][69].PCT isotherms measured at 30 ˝C and 60 ˝C for alloys Mg0-Mg5 are compared in Figure 3.These isotherms lacking noticeable plateaus are commonly observed in highly-disordered AB 2 MH alloys.The multi-phase nature in this group of alloys lowers the critical temperature (T c ) when the pressure plateau starts to disappear [70][71][72].Some hydrogen storage properties detected from the PCT isotherms are listed in Table 5.Both the maximum and reversible hydrogen storage capacities first increase and then decrease as the alloy number increases.Due to the lack of obvious plateau pressure, the desorption pressure at 0.75 wt% of hydrogen storage capacity was used for comparison between equilibrium pressure and calculation of hysteresis, and heat of hydride formation (∆H h ) and change in entropy (∆S h ).In the Mg-containing alloys, the equilibrium pressure first decreases and then increases as the alloy number increases, which complies with the general rule that a higher metal-hydrogen bond strength yields a lower plateau pressure and a higher hydrogen storage capability [50].The slope factor (SF) indicates the degree of disorder (DOD) in an alloy.SF is defined as the ratio of the storage capacity between 0.01 MPa and 0.5 MPa to the total capacity in the desorption isotherm [2,19,51].An alloy with a large SF has a flatter plateau and less DOD (less components or less variations among the components).As the alloy number increases, the SF decreases, indicating an increase in alloy homogeneity with addition of Mg.The hysteresis of the PCT isotherm is defined as ln(P a /P d ), where P a and P d are the absorption and desorption equilibrium pressures, respectively, at 0.75 wt% H-storage.The irreversible energy loss during plastic deformation of the hydride phase in the alloy matrix is a common explanation for PCT hysteresis [73][74][75], and was linked to the a/c ratio and pulverization rate of the alloy [76].In this study, the addition of Mg does not significantly change PCT hysteresis and should have no impact on the pulverization rate of alloy during cycling.The desorption equilibrium pressures at the midpoint capacity measured at 30, 60, and 90 °C (P) were used to estimate the changes in enthalpy (ΔHh) and entropy (ΔSh) using the equation:  The desorption equilibrium pressures at the midpoint capacity measured at 30, 60, and 90 ˝C (P) were used to estimate the changes in enthalpy (∆H h ) and entropy (∆S h ) using the equation: where R is the ideal gas constant and T is the absolute temperature.Since the hydrogenation reaction is exothermic, the heat of hydride formation (∆H h ) is negative.Both ∆H h and ∆S h decrease (become more negative) and then increase with the increase in alloy number.The evolution in ´∆H h value correlates to hydrogen storage capacity and is agrees with the strength of hydrogen-metal bond assumption described earlier.∆S h is an indicator for showing the DOD in hydride from a completely ordered solid (e.g., solid hydrogen).The difference between ∆S h and ´130.7 J¨mol ´1¨K ´1 (the S for H 2 (g) at 300 K and 0.1 MPa [77]) can be interpreted as the DOD for hydrogen in the hydride form (β-phase).In this study, the trend of |∆S h | increases, and then decreases with the increase in the alloy number, which is similar to that of SF (the indicator for the DOD in the host metal alloy).The same correlation between the DOD of the hydride and the DOD of the occupied hydrogen was previously reported [19].

Electrochemical Analysis
Discharge capacities measured at a discharge current of 4 mA¨g ´1 for the first 13 cycles for each alloy in this study are plotted in Figure 4a to show the activation behavior at full capacity.The activation of the Mg-containing alloys appeared to be slightly easier than the Mg-free Mg0 alloy.The results of discharge capacities with other electrochemical tests are listed in Table 6.For Mg-containing alloys, both discharge capacities measured at 4 mA¨g ´1 and 50 mA¨g ´1 rates increase and then decrease with the corresponding increase in the alloy number.The maximum capacities obtained are from alloy Mg2, which demonstrated the lowest e/a value and B/A ratio.Mg2 also has the highest hydrogen storage capacity among the alloys in this study.The HRD values, defined as the ratio of the tenth cycle capacities measured at 50 mA¨g ´1 and 4 mA¨g ´1 rates, are listed in Table 6 and demonstrated an increasing trend correlating to alloy number.The HRD obtained from the first 13 cycles are plotted in Figure 4b, which exhibited easier activation in HRD with higher alloy numbers.From Table 6, the addition of Mg is shown to be effective in improving both activation and HRD.The improvement in HRD with Mg was further investigated by electrochemically measuring the bulk diffusion constant (D) and surface exchange current (I o ).Method details of these two parameters were previously reported [25] and the results are listed in Table 6.In general, as the alloy number increases, D first decreases and then increases while I o shows the opposite trend.The increase in HRD with alloy number is related to both the bulk and surface properties of the alloys.The addition of Mg decreases the D value (bulk), except for alloys with very high Ni content (Mg4 and Mg5), and increases the I o value (surface) of the alloys.The contribution of Mg to faster activation is similar to that with La in AB 2 MH alloys [78][79][80][81], where the new Mg-containing phase may absorb a larger amount of hydrogen, causing surface cracking and an increase in the surface area.
where R is the ideal gas constant and T is the absolute temperature.Since the hydrogenation reaction is exothermic, the heat of hydride formation (ΔHh) is negative.Both ΔHh and ΔSh decrease (become more negative) and then increase with the increase in alloy number.The evolution in −ΔHh value correlates to hydrogen storage capacity and is agrees with the strength of hydrogen-metal bond assumption described earlier.ΔSh is an indicator for showing the DOD in hydride from a completely ordered solid (e.g., solid hydrogen).The difference between ΔSh and −130.7 J•mol −1 •K −1 (the S for H2 (g) at 300 K and 0.1 MPa [77]) can be interpreted as the DOD for hydrogen in the hydride form (βphase).In this study, the trend of |ΔSh| increases, and then decreases with the increase in the alloy number, which is similar to that of SF (the indicator for the DOD in the host metal alloy).The same correlation between the DOD of the hydride and the DOD of the occupied hydrogen was previously reported [19].

Electrochemical Analysis
Discharge capacities measured at a discharge current of 4 mA•g −1 for the first 13 cycles for each alloy in this study are plotted in Figure 4a to show the activation behavior at full capacity.The activation of the Mg-containing alloys appeared to be slightly easier than the Mg-free Mg0 alloy.The results of discharge capacities with other electrochemical tests are listed in Table 6.For Mg-containing alloys, both discharge capacities measured at 4 mA•g −1 and 50 mA•g −1 rates increase and then decrease with the corresponding increase in the alloy number.The maximum capacities obtained are from alloy Mg2, which demonstrated the lowest e/a value and B/A ratio.Mg2 also has the highest hydrogen storage capacity among the alloys in this study.The HRD values, defined as the ratio of the tenth cycle capacities measured at 50 mA•g −1 and 4 mA•g −1 rates, are listed in Table 6 and demonstrated an increasing trend correlating to alloy number.The HRD obtained from the first 13 cycles are plotted in Figure 4b, which exhibited easier activation in HRD with higher alloy numbers.From Table 6, the addition of Mg is shown to be effective in improving both activation and HRD.The improvement in HRD with Mg was further investigated by electrochemically measuring the bulk diffusion constant (D) and surface exchange current (Io).Method details of these two parameters were previously reported [25] and the results are listed in Table 6.In general, as the alloy number increases, D first decreases and then increases while Io shows the opposite trend.The increase in HRD with alloy number is related to both the bulk and surface properties of the alloys.The addition of Mg decreases the D value (bulk), except for alloys with very high Ni content (Mg4 and Mg5), and increases the Io value (surface) of the alloys.The contribution of Mg to faster activation is similar to that with La in AB2 MH alloys [78][79][80][81], where the new Mg-containing phase may absorb a larger amount of hydrogen, causing surface cracking and an increase in the surface area.Low-temperature performance is a very important parameter in propulsion applications, especially in start-stop type micro-hybrid vehicles.The conventional AB 5 MH alloy performed poorly below ´25 ˝C, but the Co-doped A 2 B 7 superlattice MH alloy can result in significant improvements [82].Additions of La, Nd, and Si in AB 2 MH alloys can lower the surface charge-transfer resistance (R) at ´40 ˝C to a comparable level with AB 5 MH alloys [23,25,27].In order to study the effect of Mg addition to the low-temperature performance of AB 2 MH alloys, AC impedance at ´40 ˝C was measured, and R and the double-layer capacitance (C, closely related to the surface reaction area) were calculated from the obtained Cole-Cole plot.R and C values for alloys in this study are listed in Table 7.The addition of Mg into AB 2 MH alloys increases the R value and decreases C. Therefore, different from rare earth elements and Si, the addition of alkaline earth elements in AB 2 MH alloys should not be considered when improvement in the low-temperature performance is needed.

Magnetic Properties
Magnetic susceptibility was used to characterize the nature of the metallic nickel particles present in the surface layer of the alloy following an alkaline activation treatment [10].Details on the background and experimental methods were reported earlier [83].Metallic Ni is an active catalyst for the water splitting and recombination reactions that contributes to the I o in this electrochemical system.This technique allows us to obtain the saturated magnetic susceptibility (M s ), a quantification of the amount of surface metallic Ni (the product of preferential oxidation), and the magnetic field strength at one-half of the M s value (H 1/2 ), a measurement of the averaged reciprocal number of Ni atoms in a metallic cluster (Figure 5a).The magnetic susceptibility graphs for alloys in this study are shown in Figure 5b and the calculated M s and H 1/2 values are listed in Table 6.The M s decreases and then increases with the increase in the alloy number, which is the opposite trend to the data observed for I o .The increase in I o for Mg2 is not from the metallic nickel particles embedded in the surface oxide and may be due to the high content of TiNi phase, which was reported as a catalytic phase for electrochemical reaction [84,85].The H 1/2 values listed in Table 6 indicate that the size of metallic nickel decreases and then increases as the alloy number increases.In general, the Mg-containing alloys have smaller metallic nickel clusters in the surface than the Mg-free Mg0.

Correlations
Due to the limited solubility of Mg in AB2 MH alloys, the ICP results have some deviations from the original design of Zr21.5Ti21−0.6xV10Cr7.5Mn8.1Co8Ni32.2−0.4xMgxSn0.3Al0.4,where x = 0, 1, 2, 3, 4, and 5, especially in Mg and Ni contents.As the alloy number increases, Mg content remains at approximately the same level, Ni content remain unchanged and then increases for the last two alloys, Mg4 and Mg5, while the Ti content decreases monotonically.In order to study the influences of different compositions with regard to various alloy properties, the correlation factor (R 2 ) was calculated between composition (Ni content, Ti content, e/a, and B/A) and properties.The comparison results are listed in Table 8.The findings of the correlation can thus be summarized as demonstrating that Ni content has more influences on the B/A ratio, C14 unit cell volume, hydrogen storage properties, low-rate electrochemical capacity, bulk diffusion, and size of metallic Ni embedded in the surface oxide layer when compared to Ti content.The hydrogen capacities obtained from PCT (converted into electrochemical capacity by 1 wt% = 268 mAh•g −1 ) and half-cell tests are plotted against Ni content in Figure 6a, showing a decrease in capacity with an increase in the Ni content.The Ti content is more closely related to the C14 phase crystallite size, C14 phase abundance, HRD, and R and C measured at −40 °C.The last three characteristics are plotted against Ti content in Figure 6b, showing that as Ti content increases, the R at −40 °C decreases and C at −40 °C increases.However, the RT HRD decreases.The decrease in R is

Correlations
Due to the limited solubility of Mg in AB 2 MH alloys, the ICP results have some deviations from the original design of Zr 21.5 Ti 21´0.6xV 10 Cr 7.5 Mn 8.1 Co 8 Ni 32.2´0.4xMg x Sn 0.3 Al 0.4 , where x = 0, 1, 2, 3, 4, and 5, especially in Mg and Ni contents.As the alloy number increases, Mg content remains at approximately the same level, Ni content remain unchanged and then increases for the last two alloys, Mg4 and Mg5, while the Ti content decreases monotonically.In order to study the influences of different compositions with regard to various alloy properties, the correlation factor (R 2 ) was calculated between composition (Ni content, Ti content, e/a, and B/A) and properties.The comparison results are listed in Table 8.The findings of the correlation can thus be summarized as demonstrating that Ni content has more influences on the B/A ratio, C14 unit cell volume, hydrogen storage properties, low-rate electrochemical capacity, bulk diffusion, and size of metallic Ni embedded in the surface oxide layer when compared to Ti content.The hydrogen capacities obtained from PCT (converted into electrochemical capacity by 1 wt% = 268 mAh¨g ´1) and half-cell tests are plotted against Ni content in Figure 6a, showing a decrease in capacity with an increase in the Ni content.The Ti content is more closely related to the C14 phase crystallite size, C14 phase abundance, HRD, and R and C measured at ´40 ˝C.The last three characteristics are plotted against Ti content in Figure 6b, showing that as Ti content increases, the R at ´40 ˝C decreases and C at ´40 ˝C increases.However, the RT HRD decreases.The decrease in R is consistent with the increase in C (reaction surface area), but should decrease the RT HRD.The RT R and C follow the trend of ´40 ˝C R and C (Table 7) and, therefore, the discrepancy is not due to different temperatures and requires further investigation.Both C14 and TiNi phase abundances are correlated to various properties.The results show that the C14 main phase abundance is more influential in hydrogen storage capacities both in solid state and electrochemistry, while the TiNi minor phase abundance has high impacts on D, I o , H

Conclusions
The influence of Mg addition (approximately 0.7 at%) to the structural, solid state, and electrochemical properties of a series of Laves phase based AB2 MH alloys with different Ti and Ni contents were investigated.In general, addition of Mg does not lower the surface charge-transfer resistance, unlike other non-transitional metals, such as Si, Y, La, and Nd.Interestingly, one of the alloys, Mg2 (Zr22.5Ti10.9V9.9Cr7.5Mn8.2Co8.0Ni31.8Mg0.8Sn0.3Al0.4) with a phase distribution of 88.3% C14, 7.9% C15, 3.5% TiNi, and 0.3% of a Mg-rich cubic phase, shows improved hydrogen capacities in both solid state and electrochemistry and a higher surface exchange current, but a lower bulk diffusion coefficient.

Conclusions
The influence of Mg addition (approximately 0.7 at%) to the structural, solid state, and electrochemical properties of a series of Laves phase based AB 2 MH alloys with different Ti and Ni contents were investigated.In general, addition of Mg does not lower the surface charge-transfer resistance, unlike other non-transitional metals, such as Si, Y, La, and Nd.Interestingly, one of the alloys, Mg2 (Zr 22.5 Ti 10.9 V 9.9 Cr 7.5 Mn 8.2 Co 8.0 Ni 31.8Mg 0.8 Sn 0.3 Al 0.4 ) with a phase distribution of 88.3% C14, 7.9% C15, 3.5% TiNi, and 0.3% of a Mg-rich cubic phase, shows improved hydrogen capacities in both solid state and electrochemistry and a higher surface exchange current, but a lower bulk diffusion coefficient.

Figure 1 .
Figure 1.X-ray diffractometer (XRD) patterns using Cu-Kα as the radiation source for alloys: (a) Mg0; (b) Mg1; (c) Mg2; (d) Mg3; (e) Mg4; and (f) Mg5.In addition to two Laves phases, two cubic phases can be also identified.The vertical line indicates the main C14/C15 peak shifting to lower and then higher angles with an increasing alloy number.

Figure 1 .
Figure 1.X-ray diffractometer (XRD) patterns using Cu-Kα as the radiation source for alloys: (a) Mg0; (b) Mg1; (c) Mg2; (d) Mg3; (e) Mg4; and (f) Mg5.In addition to two Laves phases, two cubic phases can be also identified.The vertical line indicates the main C14/C15 peak shifting to lower and then higher angles with an increasing alloy number.
Batteries 2016, 2, 11 5 of 16 phase found by XRD since all alloys in Mg-Ni phase diagram are hexagonal (Mg, Mg2Ni,

Figure 4 .
Figure 4. Activation behavior observed from (a) half-cell capacity measured at 4 mA•g −1 and (b) halfcell high-rate dischargeability (HRD) for the first 13 electrochemical cycles.

Figure 4 .
Figure 4. Activation behavior observed from (a) half-cell capacity measured at 4 mA¨g ´1 and (b) half-cell high-rate dischargeability (HRD) for the first 13 electrochemical cycles.

Figure 5 .
Figure 5. (a) Representative magnetic susceptibility measurement results showing the saturated magnetic susceptibility (Ms) with paramagnetic component removed and H1/2 defined as the magnitude of applied field corresponding to half of maximum magnetic susceptibility (1/2 Ms); (b) the magnetic susceptibility of the alloys in this study.

Figure 5 .
Figure 5. (a) Representative magnetic susceptibility measurement results showing the saturated magnetic susceptibility (M s ) with paramagnetic component removed and H 1/2 defined as the magnitude of applied field corresponding to half of maximum magnetic susceptibility (1/2 M s ); (b) the magnetic susceptibility of the alloys in this study.
1/2 , and C14 cell volume.The correlation of TiNi phase abundance to D and I o are plotted in Figure 6c.High amount in TiNi phase increases the surface reactivity (I o ) but hinders the bulk diffusion of hydrogen (D).While the connections between TiNi phase abundance and H 1/2 may be true and require future validation, the connection with the C14 cell volume is most likely a coincidence.Both B/A ratio and e/a value are not as influential as other factors in this comparison.Batteries 2016, 2, 11 11 of 16 consistent with the increase in C (reaction surface area), but should decrease the RT HRD.The RT R and C follow the trend of −40 °C R and C (Table 7) and, therefore, the discrepancy is not due to different temperatures and requires further investigation.Both C14 and TiNi phase abundances are correlated to various properties.The results show that the C14 main phase abundance is more influential in hydrogen storage capacities both in solid state and electrochemistry, while the TiNi minor phase abundance has high impacts on D, Io, H1/2, and C14 cell volume.The correlation of TiNi phase abundance to D and Io are plotted in Figure 6c.High amount in TiNi phase increases the surface reactivity (Io) but hinders the bulk diffusion of hydrogen (D).While the connections between TiNi phase abundance and H1/2 may be true and require future validation, the connection with the C14 cell volume is most likely a coincidence.Both B/A ratio and e/a value are not as influential as other factors in this comparison.

Figure 6 .
Figure 6.Examples of correlations found in this study showing: (a) a high Ni content decreases capacities; (b) a higher Ti content decreases the HRD and R, while C increases; and (c) a higher TiNi phase abundance increases Io (surface), but decreases D (bulk).

Figure 6 .
Figure 6.Examples of correlations found in this study showing: (a) a high Ni content decreases capacities; (b) a higher Ti content decreases the HRD and R, while C increases; and (c) a higher TiNi phase abundance increases I o (surface), but decreases D (bulk).

Table 1 .
Design compositions and inductively coupled plasma (ICP) results in at%.e/a is the average electron density.B/A is the atomic ratio of B-atom (elements other than Ti, Zr, and Mg) to A-atom (Ti, Zr, and Mg).

Table 1 .
Design compositions and inductively coupled plasma (ICP) results in at%.e/a is the average electron density.B/A is the atomic ratio of B-atom (elements other than Ti, Zr, and Mg) to A-atom (Ti, Zr, and Mg).

Table 2 .
Lattice constants a and c, a/c ratio, unit cell volume, and crystallite size of the main C14 phase of alloys Mg0-Mg5 from XRD analysis.ND denotes non-detectable.

Table 4 .
Summary of EDS results.All compositions are in at%.Compositions of the main C14 and C15 phase are in bold and italic, respectively.

Table 5 .
Summary of the solid state (gaseous phase) hydrogen storage properties of the Mgcontaining AB2 alloys.SF: slope factor.

Table 5 .
Summary of the solid state (gaseous phase) hydrogen storage properties of the Mg-containing AB 2 alloys.SF: slope factor.

Table 6 .
Summary of the room temperature (RT) electrochemical and magnetic results (capacity, rate, D, I o , and M s , H 1/2 ,) of the Mg-containing AB 2 alloys.

Table 7 .
Summary of the electrochemical results from AC impedance measurement (R: charge transfer resistance, C: double-layer capacitance at ´40 ˝C and RT) of the Mg-containing AB 2 alloys.

Table 8 .
Correlation factors (R 2 ) between the composition and hydrogen storage properties.

Table 8 .
Correlation factors (R 2 ) between the composition and hydrogen storage properties.