Electrochemical Properties of Nb-Substituted Zr-Ti-Ni Hydrogen Storage Alloy Negative Electrodes for Nickel-Metal Hydride Batteries

Abstract

Crystal structure, pressure-composition isotherms and electrochemical properties of the Zr_0.6−xTi_0.4Nb_xNi (x = 0.01, 0.02, and 0.05) alloys were investigated. Their X-ray diffraction profiles demonstrated that all the Zr_0.6−xTi_0.4Nb_xNi alloys consisted of the primary phase with the B33-type orthorhombic structure and the secondary phase with the B2-type Ti_0.6Zr_0.4Ni cubic structure. Rietveld refinement demonstrated that the atomic fraction of the secondary phase increased with the Nb content. The Zr_0.6−xTi_0.4Nb_xNi alloys were lower in hydrogen storage capacity than the Nb-free Zr_0.6Ti_0.4Ni alloy due to an increase in the abundance of the secondary phase. In the charge-discharge tests with the Zr_0.6−xTi_0.4Nb_xNi alloy negative electrodes, all the initial discharge curves had two potential plateaus due to the electrochemical hydrogen desorption of trihydride to monohydride and monohydride to alloy of the primary phase. The total discharge capacities at 333 and 303 K for the Zr_0.58Ti_0.4Nb_0.02Ni alloy negative electrode were 384 and 335 mAh g⁻¹, respectively, which were higher than those of the other Zr_0.6−xTi_0.4Nb_xNi and Zr_0.6Ti_0.4Ni alloy negative electrodes.

1. Introduction

Secondary batteries with high energy density, high power density, and long lifetime are needed because of the system interconnection of sustainable energy. Nickel-metal hydride (Ni-MH) batteries are known to have long lifetime and high-rate capability [1,2], but they exhibit lower gravimetric energy density than lithium ion batteries (LIBs). Nevertheless, the LIBs have a serious problem regarding safety because flammable organic electrolytes are used in them [3,4]. Therefore, it is meaningful to increase the energy density of the Ni-MH batteries with increased safety. For this purpose, it is essential to develop new negative electrode active materials with high capacity. The AB-type ZrNi hydrogen storage alloy is a promising negative electrode active material, because it has the theoretical capacity of 536 mAh g⁻¹ [5], which was higher than that of the AB₅- or AB₃-type rare-earth-based commercial alloys [6,7]. We reported on the electrochemical properties of the ZrNi alloy negative electrode before and after the boiling alkaline surface treatment, and its discharge capacity after the treatment was 320 mAh g⁻¹ at 333 K [8]. However, the large difference between the experimental and theoretical specific discharge capacity for the ZrNi alloy negative electrode was attributed to the stability of monohydride (β-ZrNiH) [8,9].

The way to destabilize hydrides is to reduce the lattice constant or the unit cell volume by partially replacing the alloy constituents with another element whose atomic radius is smaller [10,11]. Recently, we have reported on the ternary AB-type Zr_1−xTi_xNi (0.05 ≤ x ≤ 0.5) alloys, in which the Zr constituent was partially replaced with Ti whose atomic radius is smaller than Zr (r_Zr = 1.60 Å and r_Ti = 1.47 Å), and have assessed their crystallographic, thermodynamic and electrochemical properties [12,13]. With an increase in the Ti content, the hydrides were destabilized due to the reduction of the unit cell volume of the primary phase. For the alloys with x ≥ 0.3, the discharge reaction of the monohydride to alloy partially proceeded, leading to the increase in the total discharge capacity. Moreover, the alloy with x = 0.4 exhibited the maximum discharge capacity of 349 mAh g⁻¹ at 333 K [13]. Thus, the lattice shrinkage for the Zr_1−xTi_xNi (0.05 ≤ x ≤ 0.5) alloys also caused the destabilization of their hydrides. Nevertheless, there are still large differences between the experimental and theoretical discharge capacities. To improve the utilization of the Zr-Ti-Ni alloys, additional replacement of the constituents with other foreign metals will be needed. Niobium (Nb) is a suitable element because its atomic radius (r_Nb = 1.46 Å) is smaller than that of Zr and Ti, which can cause lattice shrinkage or the destabilization of the hydride. In addition, Nb is higher in electronegativity (χ) than Zr and Ti (χ_Nb = 1.60, χ_Zr = 1.33 and χ_Ti = 1.54). Nakano et al. reported the effect of the Nb substitution of the AB₂-type Zr-based Zr_1−xM_x(V_0.10Ni_0.49–0.57Mn_0.30–0.36Co_0.05)_2.1 (M = Ti or Nb, 0 ≤ x ≤ 0.1) in Laves-phase alloys [14]. The difference between the average electronegativity (χ_ave) of the constituents and the χ of hydrogen (χ_H = 2.20) decreased with an increase in the Ti or Nb content, leading to the destabilization of the hydrides [14]. So, the replacement of Zr and Ti with Nb will increase χ_ave or decrease the difference between χ_ave and χ_H, which causes the destabilization of the hydrides. However, so far, there has been no investigation of the effect of the Nb substitution on the electrochemical properties of the AB-type ZrTiNi hydrogen storage alloys.

In this study, the Zr_0.6Ti_0.4Ni alloy was chosen as a fundamental alloy because it exhibited the best negative electrode performances in our previous study [13], and Nb was substituted for the Zr constituent. The effect of the Nb substitution on the crystal structure, hydrogen absorption, and electrochemical properties of the Zr_0.6Ti_0.4Ni alloy was investigated.

2. Materials and Methods

2.1. Sample Preparation

The Zr_0.6−xTi_0.4Nb_xNi (x = 0.01, 0.02, and 0.05) alloys were prepared according to the previous method [12,13]. Briefly, an ingot of each alloy was prepared by arc-melting a mixture of element constituents (Zr wire-cut rod 99.0%, Ti powder 99.0%, Nb powder 99.9%, and Ni powder 99.9% purities, Kojundo Chemical Laboratory Co, LTD., Sakado, Japan) in a water-cooled copper mold in an Ar atmosphere. Before the arc-melting, the residual oxygen in the apparatus was removed to melt a piece of Ti a few times. The ingot was turned over and remelted five times to ensure homogeneity, and then, it was pulverized and sieved between 20 and 40 μm in diameter.

2.2. Material Characterization

The chemical composition of the Zr_0.6−xTi_0.4Nb_xNi (x = 0.01, 0.02, and 0.05) alloy powders was analyzed by induced couple plasma spectroscopy (ICPS, ICPV-1017, Shimadzu, Kyoto, Japan). The ICPS measurement was carried out according to our previous report [12].

The crystallographic characterization of the Zr_0.6−xTi_0.4Nb_xNi (x = 0.01, 0.02, and 0.05) alloy powders was carried out by X-ray diffractometry (XRD, D8 ADVANCE, Bruker AXS, Madison, WI, USA) with CuKα radiation (λ = 0.1541 nm, 40 kV, 40 mA). The lattice parameters (a, b, c) and unit cell volume (V_M) were refined by Rietveld analysis using Z-Rietveld software (ver.1.0.2) [15,16]. In the Rietveld refinement, the initial composition for the primary and secondary phases was taken as Zr_0.59Ti_0.4Nb_0.01Ni and Ti_0.59Zr_0.4Nb_0.01Ni for x = 0.01, Zr_0.58Ti_0.4Nb_0.02Ni and Ti_0.58Zr_0.4Nb_0.02Ni for x = 0.02, and Zr_0.55Ti_0.4Nb_0.05Ni and Ti_0.55Zr_0.4Nb_0.05Ni for x = 0.05, respectively. The lattice parameters (a, b, and c) calculated from the experimental XRD pattern for each alloy were taken as the initial values [12]. The atomic coordinates reported by Matar [17] and Cuevas et al. [18] were used as those of the primary and secondary phases, respectively.

2.3. Hydrogen Storage Property

The hydrogen storage properties for the Zr_0.6−xTi_0.4Nb_xNi (x = 0.01, 0.02, and 0.05) alloy powders were examined with pressure-composition isotherms (PCT curves) at 423 K using a Sieverts-type apparatus (PCT-6-4SDMDWIN, Suzuki Shokan, Tokyo, Japan). Before the PCT measurement, the initial activation treatment was performed according to the previous procedure [12]. Each sample was placed into a stainless-steel reactor tube in an Ar-filled glove box, heated in vacuum at 473 K for 1 h, and gradually introduced at 0.9 MPa H₂ pressure. After the reactor tube was heated to 623 K, it was evacuated at the same temperature for 2 h. After the initial activation, the hydrogen pressure was set to 0, and the PCT curves at 423 K were measured in the H₂ pressure from 0 to 0.92 MPa.

2.4. Electrochemical Property

The Zr_0.6−xTi_0.4Nb_xNi (x = 0.01, 0.02, and 0.05) negative electrodes were prepared according to the previous method [8,9,13]. Briefly, the Zr_0.6−xTi_0.4Nb_xNi (x = 0.01, 0.02, and 0.05) powders (42.5 mg), Cu powders (5 mg), and 10 mass% polyvinyl alcohol aqueous solution (25 μL) were mixed to make a paste, and the paste was filled into a Ni mesh as a current collector. For the activation treatment, the resultant Zr_0.6−xTi_0.4Nb_xNi negative electrodes were soaked in a boiling 6 M KOH alkaline solution for 4 h [8,9]. The positive and reference electrodes were NiOOH/Ni(OH)₂ and Hg/HgO electrodes, respectively. The electrolyte solution (100 mL) was a 6 M KOH aqueous solution containing 1 M LiOH.

Charge-discharge cycle tests were performed following our previous report [13]. Each negative electrode was charged at 100 mA g⁻¹ for 5 h and discharged at 25 mA g⁻¹ to a cut off potential of −0.5 V versus Hg/HgO. After charging and discharging, the circuit was opened for 10 min. For evaluating high-rate dischargeability (HRD), each negative electrode was charged at 100 mA g⁻¹ for 5 h and then discharged at 25–200 mA g⁻¹ to −0.5 V versus Hg/HgO. Each measurement was carried out at 333 and 303 K. After each cycle test, the concentration of each element dissolving in the electrolyte was measured by inductively coupled plasma spectroscopy (ICPS).

Alternative current (AC) impedance measurement was conducted in a frequency range of 64 kHz to 0.1 Hz with an amplitude of 5 mV at 303 K. Prior to the measurement, each negative electrode was charged at 100 mA g⁻¹ for 5 h, and then, its potential was kept open for 1 h. All electrochemical properties were measured by VMP3 (Bio-Logic, Claix, France).

3. Results and Discussions

3.1. Structural Characteristics of the Zr_1−xTi_0.4Nb_xNi (x = 0.01, 0.02, and 0.05) Alloys

The chemical composition of the Zr_0.6−xTi_0.4Nb_xNi (x = 0.01, 0.02, and 0.05) alloy powders analyzed by ICPS is listed in

Table 1. Chemical composition for the Zr_0.6−xTi_0.4Nb_xNi (x = 0.01, 0.02, and 0.05) powders.

		Content/at%
		Zr	Ti	Nb	Ni
x = 0.01	Theoretical	29.50	20.00	0.50	50.00
	Experimental	29.12	19.13	0.46	50.68
x = 0.02	Theoretical	29.00	20.00	1.00	50.00
	Experimental	28.86	19.43	0.96	50.12
x = 0.05	Theoretical	27.50	20.00	2.50	50.00
	Experimental	27.39	19.44	2.49	50.37

. In each case, the experimental composition was similar to the theoretical one. The error of these experimental data was ±0.1 at%.

The red curves in Figure 1 show the experimental XRD patterns for the Zr_0.6−xTi_0.4Nb_xNi (x = 0.01, 0.02, and 0.05) alloy powders. The blue curves in Figure 1 show the profiles refined by the Rietveld method, and the gray curves show the difference between the experimental and calculated diffraction intensities. As can be seen from each gray curve, the difference was small, so the experimental profile is reasonably well fitted with the theoretical model in each case. R_wp and S are named the residue of the weighted pattern and goodness-of-fit indicator, respectively [19,20]. The Rietveld refinement demonstrated that each alloy consisted of two phases, the primary phase with the B33-type ZrNi orthorhombic structure [12,17] and the secondary phase with the B2-type Ti_0.6Zr_0.4Ni cubic structure [12,18], which are indicated by the green and yellow bars in Figure 1, respectively.

Composition, lattice parameters (a, b, c) and unit cell volume (V) estimated by the Rietveld refinement for each phase of the Zr_0.6−xTi_0.4Nb_xNi (x = 0.01, 0.02, and 0.05) alloy powders are summarized in

Table 2. Structural parameters after the Rietveld refinement for the Zr_0.6−xTi_0.4Nb_xNi (x = 0.01, 0.02, and 0.05) alloy powders.

	Composition of Primary and Secondary Phases	Lattice	Space Group (No.)	a/nm	b/nm	c/nm	V/nm3	Phase Abundance/at%
x = 0.01	Zr0.589Ti0.396Nb0.014Ni1.001	B33	Cmcm(63)	0.31279	0.96712	0.40670	0.12303	95.2
	Ti0.589Zr0.392Nb0.013Ni1.000	B2	Pm-3m(221)	0.31239	0.31239	0.31239	0.03048	4.8
x = 0.02	Zr0.571Ti0.401Nb0.026Ni1.002	B33	Cmcm(63)	0.31269	0.96684	0.40662	0.12292	93.9
	Ti0.588Zr0.396Nb0.013Ni0.998	B2	Pm-3m(221)	0.31265	0.31265	0.31265	0.03056	6.1
x = 0.05	Zr0.550Ti0.399Nb0.052Ni0.999	B33	Cmcm(63)	0.31246	0.96572	0.40576	0.12243	84.9
	Ti0.528Zr0.433Nb0.038Ni1.001	B2	Pm-3m(221)	0.31236	0.31236	0.31236	0.03047	15.1

. All the lattice parameters and unit cell volume of the primary phase in Table 2 slightly decreased with an increase in the x value because of the Nb substitution. The fractions of two phases for the Nb-substituted alloys were evaluated by the Rietveld refinement and summarized in Table 2. The fraction of the secondary phase increased with x value. Nb is known to stabilize the β-phase Ti and Ti-based alloys with the body-centered cubic (BCC) structure [21,22,23]. So, in the present study, Nb seems to promote the segregation of the secondary phase, causing the increase in the intensity of the diffraction peak around 41.0 degree in Figure 1.

3.2. Hydrogen Storage Properies of the Zr_1−xTi_0.4Nb_xNi (x = 0.01, 0.02, and 0.05) Alloys

Figure 2 shows the PCT curves in the hydrogen absorption process at 423 K for the Zr_0.6−xTi_0.4Nb_xNi (x = 0, 0.01, 0.02, and 0.05) alloy powders.

The plateau region for each alloy is mainly assigned to the transformation from monohydride to trihydride of the primary phase, because the abundance of the primary phase was much higher than that of the secondary phase, as shown in Table 2. The hydrogen storage capacities at 0.5 MPa and the plateau pressure at 423 K were summarized in

Table 3. Hydrogen storage capacity at 0.5 MPa, plateau pressure at 423 K, and χ_ave of the primary phases for the Zr_0.6−xTi_0.4Nb_xNi (x = 0.01, 0.02, and 0.05) alloy powders.

	Hydrogen Storage Capacity at 0.5 MPa/mass%	Plateau Pressure/MPa	χave
x = 0	1.78	0.0496	1.662
x = 0.01	1.71	0.0695	1.664
x = 0.02	1.74	0.0756	1.667
x = 0.05	1.52	0.128	1.675

. Hydrogen storage capacity can be influenced by the lattice volume and abundance of the primary phase [12]. The plateau pressure at 423 K increased with the Nb content mainly due to the destabilization of hydrogen absorbed in the primary phase. The unit cell volume, however, barely decreased with an increase in the Nb content, as shown in Table 2, so the change in the unit cell volume will minorly contribute to the destabilization of the hydride. The χ_ave value is defined by the following equation [14]:

χ_ave = ∑(n_i χ_i)/∑ n_i

(1)

where n_i and χ_i are the number and electronegativity of constituent i, respectively. Nakano et al. reported that when the χ_ave value for the Zr-based Laves-phase alloy increased by 0.01, the enthalpy change for the hydride formation increased by about 3 kJ mol(H₂)⁻¹ [14], suggesting that the stability of hydride is sensitive to χ_ave. In the present study, χ_ave of the primary phase increased with the x value or the Nb content, as shown in Table 3, suggesting that the difference between χ_ave of the primary phase and χ_H decreased or hydrogen in the primary phase was destabilized. Therefore, the destabilization of hydrogen absorbed in the primary phase with the Nb substitution can be largely attributed to the electronic effect.

3.3. Electrochemical Properties of the Zr_0.6−xTi_0.4Nb_xNi (x = 0.01, 0.02, and 0.05) Alloys

The initial discharge curves of the Zr_0.6−xTi_0.4Nb_xNi (x = 0, 0.01, 0.02, and 0.05) negative electrodes at 333 K are shown in Figure 3a. Each discharge curve was divided by two potential plateau regions. The total discharge capacity (C_dis,total) defined by the sum of two plateaus is summarized in

Table 4. Total discharge capacity (C_dis,total), first discharge capacity (C_dis,1), second discharge capacity (C_dis,2), first peak potential (E₁), and second peak potential (E₂) for the Zr_0.6−xTi_0.4Nb_xNi (x = 0.01, 0.02, and 0.05) alloy negative electrodes at 333 K.

	Cdis,total/mAh g−1	Cdis,1/mAh g−1	Cdis,2/mAh g−1	E1/V	E2/V
x = 0	349	296	53	−0.828	−0.722
x = 0.01	351	296	55	−0.831	−0.733
x = 0.02	384	323	61	−0.836	−0.739
x = 0.05	362	299	67	−0.851	−0.748

. The C_dis,total value exhibited the maximum at the Nb content of x = 0.02. The first and second discharge potential plateaus at the lower and higher potentials are attributed to electrochemical hydrogen desorption of trihydride to monohydride and monohydride to alloy of the primary phase, respectively [12]. To precisely estimate the plateau potential, the differential capacity (dQ dE⁻¹) curves in Figure 3a are shown in Figure 3b,c. Figure 3c is the enlarged area between −0.80 and −0.65 V in Figure 3b. The discharge potentials for the lower and higher potential plateaus (E₁ and E₂) were defined as the first and second plateau potentials, respectively [23] and are summarized in Table 4. From Table 4, both E₁ and E₂ shifted to the lower potentials with an increase in the Nb content. This suggests the trihydride and monohydride of the primary phase were destabilized by the Nb substitution. In this way, the Nb substitution was effective for improving the destabilization of both trihydride and monohydride of the primary phase, which seems to be largely due to the electronic effect as mentioned above.

The discharge capacities in the first and second plateau regions (C_dis,1 and C_dis,2) were defined as the charge passed from the beginning of the discharge process to the inflection point between two peaks in each differential curve [13] and are summarized in Figure 3d. The C_dis,1 was the highest for the alloy with x = 0.02, which showed a similar trend as the C_dis,total, because hydrides of the primary phase were destabilized by the increase in the Nb content. However, the C_dis,1 value for the alloy with x = 0.05 decreased, because the abundance of the secondary phase greatly increased as shown in Table 2. The C_dis,2, which was defined as (C_dis,total−C_dis,1), is also shown in Figure 3d. The C_dis,2 gradually increased with the Nb content due to the destabilization of monohydride of the primary phase.

Figure 4a shows the initial discharge curves of the Zr_0.6−xTi_0.4Nb_xNi (x = 0, 0.01, 0.02, and 0.05) negative electrodes at 303 K. In Figure 4a, the second plateau region was not distinctly observed in all discharge curves, suggesting that the monohydride of the primary phase scarcely desorbed hydrogen at 303 K. The C_dis,total for each alloy is summarized in

Table 5. Total discharge capacity (C_dis,total), peak potential (E₁), and second peak potential (E₂) for the Zr_0.6−xTi_0.4Nb_xNi (x = 0.01, 0.02, and 0.05) alloy negative electrodes at 303 K.

	Cdis,total/mAh g−1	E1/V
x = 0	281	−0.803
x = 0.01	309	−0.825
x = 0.02	335	−0.830
x = 0.05	313	−0.836

. This table showed that the Nb substitution was effective for improving discharge capacity even at 303 K. The differential capacity (dQ dE⁻¹) curve for each discharge curves in Figure 4a is shown in Figure 4b. The E₁ value for each negative electrode is summarized in Table 5. As can be seen from this table, the E₁ value negatively shifted with increasing the Nb content, suggesting that the trihydride of the primary phase is destabilized even at 303 K due to the electronic effect.

Figure 5a,b show the change in discharge capacity at 333 and 303 K with cycle number for the Zr_0.6−xTi_0.4Nb_xNi (x = 0, 0.01, 0.02, and 0.05) alloy negative electrodes, respectively. In Figure 5a, the cycle performance was improved by the increase in the Nb content, and the discharge capacity for the alloys with x ≥ 0.02 at both temperatures was more than 300 mAh g⁻¹ even at the 10th cycle. The normalized discharge capacity (NDC) at 333 and 303 K for the Zr_0.6−xTi_0.4Nb_xNi (x = 0, 0.01, 0.02, and 0.05) negative electrodes was plotted as a function of cycle number in Figure 5c,d, respectively. The NDC is defined by the following equation [24]:

NDC (%) = 100C_dis,n/C_dis,max

(2)

where C_dis,n is the discharge capacity at the n-th cycle, and C_dis,max is the maximum discharge capacity.

At 333 K, the NDC values at the 10th cycle for the alloy negative electrodes with x = 0, 0.01, 0.02, and 0.05 were 74%, 77%, 83%, and 87%, respectively. The higher the NDC, the higher the charge-discharge cycle performance or cycle stability. At 303 K, the NDC values for alloys with x = 0, 0.01, 0.02, and 0.05 were 90%, 94%, 98%, and 100% at the10th cycle, respectively, and were higher than those at 333 K. For the alloys with x = 0.01, 0.02, and 0.05, the amount of Zr and Ti constituents dissolving in the electrolyte for 10 charge-discharge cycles at 303 and 333 K were determined by ICPS and are summarized in

Table 6. Concentration in the electrolyte after 10 cycles at 303 and 333 K by induced couple plasma spectroscopy (ICPS).

	Temperature/K	Concentration/ppm
		Zr	Ti	Nb	Ni
x = 0.01	303	3.6	1.1	-	-
	333	5.9	2.5	-	-
x = 0.02	303	3.3	1.0	-	-
	333	5.7	2.3	-	-
x = 0.05	303	1.4	0.6	-	-
	333	3.9	2.0	-	-

. The Zr and Ti species are detected at both temperatures and probably exist as HZrO₃⁻ and HTiO₃⁻ as expected from the Pourbaix diagrams for Zr and Ti [25]. On the other hand, the Nb and Ni constituents are not detected by ICPS. Nb also can dissolve in alkaline solutions [26], but its content in the alloys is very low in this study. So, Nb may be undetectable even if Nb dissolves. In contrast, Ni can be oxidized to Ni(OH)₂, which does not dissolve [25,27]. For all alloys, the amounts of Zr and Ti dissolving in the electrolyte at 303 K were less than that at 333 K, which is consistent with previous reports [28,29]. These results clearly indicate that the significant improvement of cycle performance at 303 K is attributable to effective suppression of the dissolution of the Zr, Ti, and Nb constituents. Moreover, the amount of dissolved Zr and Ti for the alloy with x = 0.05, which has the highest fraction of the secondary phase, was smaller than that for alloys with x = 0.01 and 0.02.

The cross-sectional SEM images of the Zr_0.6−xTi_0.4Nb_xNi (x = 0, 0.01, 0.02, and 0.05) negative electrodes after 10 cycles at 333 K are shown in Figure 6. The size of the alloy particles for each negative electrode was between 20 and 40 μm before the cycle test, but after 10 cycles it became less than 5 μm due to the cracking of the alloy particles. The reduction of particle size, however, was suppressed more effectively for the higher Nb contents, as shown in Figure 6. As shown in Table 6, the dissolution of the Zr and Ti constituents accelerated for the alloys with lower Nb contents, which may induce the pulverization of alloy particles.

The discharge capacity at 333 and 303 K, as functions of specific discharge current for the Zr_0.6−xTi_0.4Nb_xNi (x = 0, 0.01, 0.02, and 0.05) negative electrodes, are shown in Figure 7a,b, respectively. At both temperatures, the discharge capacity at 200 mA g⁻¹ increased with an increase in the Nb content. High-rate dischargeability (HRD) at 333 and 303 K, as functions of specific current density for the Zr_0.6−xTi_0.4Nb_xNi (x = 0, 0.01, 0.02, and 0.05) negative electrodes, are shown in Figure 7c and Figure 7d, respectively. HRD is defined by the following equation [24]:

HRD (%) = 100C_dis,i/C_dis,25

(3)

where C_dis,i and C_dis,25 represent discharge capacity at i and 25 mA g⁻¹, respectively.

At 333 K, each Nb-substituted alloy negative electrode is a little superior in HRD to the Zr_0.6Ti_0.4Ni negative electrode. In contrast, at 303 K, HRD was greatly improved with increasing the Nb content. The hydrides of the Zr_0.6−xTi_0.4Nb_xNi alloy were destabilized by the electronic effect of the Nb substitution, facilitating hydrogen diffusion [14,30]. Consequently, the decay in discharge capacity with an increase in specific discharge current will be suppressed.

Figure 8a shows the Nyquist plots of Zr_0.6−xTi_0.4Nb_xNi (x = 0, 0.01, 0.02, and 0.05) alloy negative electrodes after charging at the 11th cycle at 303 K, and each measured impedance spectrum was simulated on the basis of an equivalent circuit given in the Figure 8b [31]. The model circuit is composed of three depressed semicircles and a straight line. Constant phase elements (Q₂ and Q₄) were employed for the capacitive components parallel to R₂ and R₄, which are related to the two semicircles in order to take into account the depressed shape of the semicircles. Q₂ and Q₄ are converted as the pseudocapacitance (C₂ and C₄) by the EC-Lab software (Bio-Logic, Claix, France). Resistances (R₁–R₄) and capacitances (C₂–C₄) for each alloy at the 11th cycle are shown in

Table 7. Resistances (R₁–R₄), capacitance (C₃), and pseudocapacitances (C₂, C₄) for the full-charged Zr_0.6−xTi_0.4Nb_xNi (x = 0, 0.01, 0.02, and 0.05) alloy negative electrodes at the 11th cycle at 303 K.

	R1/Ω	R2/Ω	R3/Ω	R4/Ω	C2/F	C3/F	C4/F
x = 0	0.150	0.238	0.354	3.67	3.48 × 10−3	0.248	9.90 × 10−3
x = 0.01	0.127	0.234	0.369	3.11	3.61 × 10−3	0.226	11.57 × 10−3
x = 0.02	0.187	0.227	0.388	2.38	3.87 × 10−3	0.212	12.07 × 10−3
x = 0.05	0.168	0.214	0.370	2.06	3.95 × 10−3	0.219	12.36 × 10−3

. R₁ is the electrolyte resistance. R₂ and C₂ are the resistance and capacitance between the alloy particles and the current collector. R₃ and C₃ are the resistance and capacitance between the alloy particles. R₄ and C₄ are the charge transfer resistance and double layer capacitance [9]. R₂ and R₄ decreased with an increase in Nb content, whereas C₂ and C₄ increased. In particular, R₄ for the alloy with x = 0.05 was much smaller than that with x = 0. The abundance of the TiNi phase is known to improve electrical conductivity [32]; so, in this study, the abundance of the Ti_0.6Zr_0.4Ni secondary phase also seems to improve the electrical conductivity. Consequently, the alloys with higher Nb contents will improve HRD at the lower temperature. R₃ and C₃ were scarcely changed regardless of the Nb content, suggesting that the pulverization of the alloy particles scarcely occurs during charge-discharge cycling at 303 K.

Charge-discharge performance for the Zr_0.58Ti_0.4Nb_0.02Ni alloy, which exhibited the best performance in this study, is still insufficient, compared to the commercial rare-earth-based alloys. However, its discharge capacity was higher than that of the commercial alloys, which motivates us to propel the development of the Zr-Ti-Nb-Ni alloys to increase the energy density of Ni-MH batteries; although, some breakthroughs will be required.

4. Conclusions

The crystal structure of the Zr_0.6−xTi_0.4Nb_xNi (x = 0, 0.01, 0.02, and 0.05) alloy powders with two phases was assigned to the B33-type orthorhombic structure for the primary phase and the B2-type Ti_0.6Zr_0.4Ni structure for the secondary phase. The Rietveld refinement demonstrated that the atomic fraction of the secondary phase increased with the Nb content. Moreover, the lattice parameters and unit cell volume of the primary phase decreased with an increase in the Nb content.

The hydrogen storage capacity at 0.5 MPa decreased with an increase in the Nb content because of the increase in the abundance of the secondary phase, whereas the plateau pressure in hydrogen absorption from monohydride to trihydride of the primary phase rose due to the destabilization of the trihydride of the primary phase, which could mainly be attributed to the electronic effect.

The discharge curves for the Zr_0.6−xTi_0.4Nb_xNi (x = 0, 0.01, 0.02, and 0.05) negative electrodes had two potential plateaus. The first and second plateaus were attributable to electrochemical hydrogen desorption of trihydride to monohydride and monohydride to alloy of the primary phase, respectively. The C_dis,total value at 333 and 303 K was highest for the alloy with x = 0.02. The first and second plateau potentials negatively shifted with an increase in the Nb content, suggesting the destabilization of the trihydride and monohydride of the primary phase. The cycle and rate performance were also improved by the Nb substitution. In this study, the Zr_0.58Ti_0.4Nb_0.02Ni negative electrode exhibited the best negative electrode performance.