Performance Comparison between AB₅ and Superlattice Metal Hydride Alloys in Sealed Cells

Abstract

High-power cylindrical nickel metal/hydride batteries using a misch metal-based Al-free superlattice alloy with a composition of La_11.3Pr_1.7Nd_5.1Mg_4.5Ni_63.6Co_13.6Zr_0.2 were fabricated and evaluated against those using a standard AB₅ metal hydride alloy. At room temperature, cells made with the superlattice alloy showed a 40% lower internal resistance and a 59% lower surface charge-transfer resistance compared to cells made with the AB₅ alloy. At a low temperature (−10 °C), cells made with the superlattice alloy demonstrated an 18% lower internal resistance and a 60% lower surface charge-transfer resistance compared to cells made with the AB₅ alloy. Cells made with the superlattice alloy exhibited a better charge retention at −10 °C. A cycle life comparison in a regular cell configuration indicated that the Al-free superlattice alloy contributes to a shorter cycle life as a result of the pulverization from the lattice expansion of the main phase.

1. Introduction

Nickel/metal hydride (Ni/MH) batteries have been serving consumer portable electronics, hybrid electric vehicles, and stationary applications for more than 30 years [1,2,3,4,5,6]. Until now, the misch metal (Mm)-based AB₅ metal hydride (MH) alloy was the mainstream negative electrode active material [7]. In the last decade, the Mm-based superlattice MH alloy began to take over the market share because of its higher capacities; better high-rate dischargeability; and superior low-temperature, high-temperature, and charge retention performances compared to the conventional AB₅ MH alloy [2,3,4,5,6,8]. The superlattice MH alloy is composed of more than one phase with alternating A₂B₄ and AB₅ building slabs along the c-direction of the unit cell [2]. There can be one (AB₃), two (A₂B₇), three (A₅B₁₉), or more AB₅ units between two A₂B₄ slabs. Depending on the stacking sequence, either the hexagonal or rhombohedral structures are possible. The A-site of the superlattice MH alloy contains both rare-earth (RE) and alkaline earth (usually Mg) elements. While almost all academic research has focused on the single RE element (La or Nd)-based superlattice MH alloys (for reviews, see [9,10,11]), commercial applications have adopted the Mm composition for a higher cycle stability [2,12]. In the past, a few papers about the substitution works performed in the Mm-based superlattice alloy family with Al [13], Mn [14,15], Fe [16,17], Co [18,19,20], and Ce [21] were published, but a systematic performance comparison between a Mm-based superlattice MH alloy and a standard AB₅ MH alloy is absent. Therefore, we conducted a series of battery performance evaluations in the sealed cells made with both materials and report the results here.

2. Experimental Setup

Both the AB₅ and superlattice alloys were prepared by Eutectix (Troy, Michigan, USA) with a conventional 250 kg induction melting furnace [22]. The ingot was placed in a retort and annealed at 960 °C in vacuum (AB₅) or 1 atm atmosphere of Ar (superlattice) for 8 h. The annealed ingots were crushed and ground into the size of −200 mesh. A Philips X’Pert Pro X-ray diffractometer (XRD; Amsterdam, the Netherlands) and a JEOL-JSM6320F scanning electron microscope (SEM; Tokyo, Japan) were used to study the alloys’ microstructures. A Suzuki Shokan multi-channel pressure–concentration–temperature system (PCT; Tokyo, Japan) was used to study the gaseous-phase hydrogen storage characteristics. Electrochemical properties were evaluated with negative electrodes made by dry compacting the annealed alloy powder onto an expanded nickel substrate. A CTE MCL2 Mini cell testing system (Chen Tech Electric MFG. Co., Ltd., New Taipei, Taiwan) was used to study the alloys’ half-cell characteristics.

For the sealed-cell performance evaluation, a C-size cylindrical high-power design was chosen. Negative (0.193 mm thick) and positive electrodes (two thicknesses: 0.300 and 0.361 mm) were made using the dry-compaction and wet-paste methods [23], respectively. Positive electrode paste was composed of 89% standard AP50 [24] with a composition of Ni_0.91Co_0.045Zn_0.045(OH)₂ (BASF—Ovonic, Rochester Hills, Michigan, USA), and 5 wt % Co and 6% Co(OH)₂ powders on a nickel foam substrate. A Freudenberg FS2225 fluorinated acrylic acid grafted polyethylene/polypropylene non-woven fabric (Freudenberg Group, Weinheim, Germany) was used as the separator. In this high-power cell design, a high negative-to-positive capacity ratio cell design (about 1.7) was used to ensure a large amount of overcharge reservoir [25]. A 30 wt % KOH solution with LiOH (1.5 wt %) additive was used as the electrolyte. A six-cycle activation process using a Maccor battery cycler (Maccor, Tulsa, Oklahoma, USA) was conducted for each cell [26]. The battery performance testing procedures can be found in an earlier publication [27].

3. Results and Discussion

3.1. Alloy Properties Comparison

Alloy A, the most popular AB₅ alloy used in the industry with a composition of La_10.5Ce_4.3Pr_0.5Nd_1.4Ni₆₀Co_12.7Mn_5.9Al_4.7, was used as the control in this comparison work. Alloy B with a composition of La_11.3Pr_1.7Nd_5.1Mg_4.5Ni_63.6Co_13.6Zr_0.2, which shows the lowest charge-transfer resistance in a comparative study [28], was the superlattice alloy under the current study. In this composition, Pr and Nd were added to reduce the corrosion nature of the alloy, Ce and Mn were not included in the consideration of cycle stability and self-discharge [2,21], Co was added for low-temperature performance enhancement [19], and a very small amount of Zr was added for scavenging residual oxygen in the chamber. The B/A stoichiometry of 3.42 was chosen through an optimization study judging the electrochemical performance. While annealed alloy A has only one CaCu₅ phase, as seen from its XRD pattern (Figure 10a in [28]), alloy B shows a multi-phase structure in both pristine and annealed conditions (Figure 1). Phase abundances calculated from the XRD data are listed in

Table 1. Phase abundances (in wt %) of alloy B before and after annealing determined by the X-ray diffractometer (XRD) analysis. HEX, CUB, and RHO are hexagonal, cubic, and rhombohedral, respectively.

Stoichiometry	AB2		AB3		A2B7		A5B19		AB5
Structure	HEX	CUB	HEX	RHO	HEX	RHO	HEX	RHO	HEX
Phase	MgZn2	LaMgNi4	CeNi3	NdNi3	Nd2Ni7	Pr2Ni7	Sm5Ni19	Nd5Co19	CaCu5
Pristine alloy B	7.9	5.7	2.0	19.3	0.0	13.1	1.6	17.7	32.7
Annealed alloy B	0.0	3.2	7.5	7.4	56.9	0.0	6.4	17.0	1.6

. After annealing, the abundance of the desirable Nd₂Ni₇ phase [29] increased from 0 to 56.7 wt %; the unwanted CaCu₅ phase [30] decreased from 32.7 to 1.6 wt %; and LaMgNi₄ and other superlattice phases, such as CeNi₃, NdNi₃, Sm₅Ni₁₉ and Ni₅Co₁₉, still existed. SEM analysis was used to confirm the XRD findings, and two representative backscattering electron micrographs for pristine and annealed alloy B are shown in Figure 2. X-ray energy-dispersive spectroscopy (EDS) was used to study the chemical compositions of a few spots in Figure 2, and the results are summarized in

Table 2. Energy-dispersive spectroscopy (EDS) results from selected spots in Figure 2. All numbers are percentages.

Location	La	Pr	Nd	Mg	Ni	Co	Zr	B/A	Phase
Figure 2a-1	13.6	0.7	4.5	6.6	62.6	11.9	0.1	2.93	Superlattice
Figure 2a-2	11.1	1.2	5.3	0.9	67.2	14.1	0.2	4.41	AB5
Figure 2a-3	9.5	0.9	5.6	0.4	69.1	14.4	0.1	5.10	AB5
Figure 2a-4	11.1	0.8	5.2	18.7	57.5	6.5	0.2	1.79	LaMgNi4
Figure 2a-5	5.5	0.4	3.5	0.0	17.8	3.2	69.6	9.64	ZrO2
Figure 2b-1	86.0	0.0	4.5	0.8	6.4	0.8	1.5	0.1	La metal
Figure 2b-2	16.3	0.7	4.1	4.0	61.5	13.2	0.2	2.98	Superlattice
Figure 2b-3	14.4	0.9	4.4	12.4	60.3	7.5	0.1	2.11	LaMgNi4
Figure 2b-4	13.2	0.7	4.7	14.2	60.4	6.7	0.1	2.04	LaMgNi4

. In the pristine sample, the AB₅ phase (spots 2 and 3) can be identified by its relatively bright contrast due to the higher content of low-atomic weight nickel. Later, the AB₅ phase was removed by annealing. The superlattice phases are difficult to separate by contrast in the micrographs because of their similar chemical composition and stoichiometry. The microstructural analyses conclude that while annealed alloy A has only one CaCu₅ structure, alloy B (before or after annealing) is a superlattice-based (>95 wt %) multi-phase alloy.

Both the gaseous phase and electrochemical hydrogen storage characteristics of alloys A (AB₅) and B (superlattice) were studied. In the gaseous phase, PCT isotherms measured at 30 °C for both alloys are plotted in Figure 3. Annealed alloy A has a higher plateau pressure and a higher reversible capacity than pristine and annealed alloy B. Annealing in alloy B flattens the isotherm, increases the storage capacity, and reduces the hysteresis, as reported previously [31]. Electrochemical testing results from the first 20 cycles of annealed alloys A and B are compared in Figure 4. Annealed alloy B exhibits a higher initial capacity, but it degrades quickly in the flooded KOH solution compared to annealed Alloy A. The higher oxidation rate in the Mg-containing superlattice alloys is well known, and many electrode fabrication methods have been proposed to overcome this shortcoming [32]. As a result, a commercial cell capable of 6000 cycles with a superlattice alloy has been demonstrated [8]. Gaseous phase and electrochemical hydrogen storage properties of annealed alloys A and B are summarized in

Table 3. Gaseous phase and electrochemical hydrogen storage properties of annealed alloys A and B.

Alloy	Full H-Storage	Reversible H-Storage	Plateau Pressure	PCT Hysteresis	Discharge Capacity
Annealed alloy A	1.41%	1.37%	0.058 MPa	0.10	310 mAh·g−1
Annealed alloy B	1.42%	1.25%	0.025 MPa	0.23	370 mAh·g−1

. Plateau pressure is defined as the equilibrium pressure corresponding to a 0.75 wt % storage capacity in the desorption isotherm, and the PCT hysteresis is defined as ln (absorption pressure/desorption pressure) at the same storage capacity. Although the gaseous phase capacities of the two alloys are similar, the superlattice alloy shows a higher electrochemical discharge capacity, which is close to the theoretical limit converted from the gaseous phase capacity (381 mAh·g⁻¹ using the conversion of 1 wt % = 268 mAh·g⁻¹) because of the synergetic effect among the constituent phases [33]. The higher PCT hysteresis in annealed alloy B predicts a higher pulverization rate during repetitive cycling [34].

3.2. Sealed-Cell Performance

Fifty C-size cylindrical cells in a high-power design were made with annealed alloys A (cell A) and B (cell B). After the formation process, the discharge capacities were 2.7 and 3.1 Ah from cells A and B, respectively, with a 0.6 A discharge current. Cell B shows a higher energy density (50.4 vs 41.0 Wh·kg⁻¹) than cell A because of its higher active material capacity, which allows for the matching with a thicker positive electrode (0.361 vs 0.300 mm).

3.2.1. High-Rate

Room temperature (RT) discharge voltage profiles with four different rates (C, 2C, 5C, and 10C) for cells A and B are shown in Figure 5. The cell voltage (V) decreases with the increase in the discharge current (i) following the formula:

V = V_oc − iR_int

(1)

where V_oc and R_int are the open-circuit voltage (when i = 0) and internal resistance, respectively. Voltage suppression due to the increase in the discharge current is less severe in cell B compared to cell A, which indicates a lower R_int in cell B. Normalized discharge capacities (to those obtained with a 0.2C discharge rate) of cells A and B (set of four each) are listed in Table S1 and indicate a slightly lower high-rate dischargeability of cell B (average value of 84.1% in cell B vs 87.6% in cell A at a 10C rate). However, the capacities of cell B are higher than those of cell A at all discharge rates.

3.2.2. Low Temperature

Low-temperature performances of cells A and B were evaluated by measuring the capacities at −10 °C with different discharge rates (C, 2C, 5C, and 10C). The resulting discharge voltage profiles are plotted in Figure 6. Voltage suppression due to the increase in the discharge current is more severe at a lower temperature. Only about 50% of the capacity is obtained at −10 °C with C and 2C discharge rates. The cells deliver almost no capacity with further increases in the discharge rate. Normalized −10 °C discharge capacities (to those obtained at RT with a 0.2C discharge rate) of cells A and B (set of four each) are listed and indicate a slightly better low-temperature performance of cell B (average value of 51.6% in cell B vs 49.0% in cell A at a 1C rate).

3.2.3. Charge Retention

Charge-retention behaviors of cells A and B were evaluated by both the RT and −10 °C standing voltage stabilities at an 80% state-of-charge, and the results are plotted in Figure 7 and Figure 8, respectively. On average, cell A demonstrates a marginally better charge-retention performance at RT but a worse performance at −10 °C compared to cell B.

3.2.4. Internal Resistance

Internal resistance (R_int) was measured by a pulse method using the formula:

R_int = ∆V/∆i

(2)

Both 1 and 10 s pulses were used to measure R_ints from cells A and B, and data obtained at both RT and −10 °C are listed in

Table 4. Internal resistances (R_int, in mΩ·m²) measured with 1 and 10 s pulsed discharges with different discharge rates (1, 2, 5, and 10C) at both room temperature (RT) and −10 °C.

Condition	1C	2C	5C	10C
Cell A RT 1 s	0.144	0.142	0.138	0.130
Cell A RT 10 s	0.194	0.192	0.182	0.168
Cell A −10 °C 1 s	0.405	0.369	0.289	0.244
Cell A −10 °C 10 s	0.516	0.492	0.499	—
Cell B RT 1 s	0.087	0.087	0.087	0.085
Cell B RT 10 s	0.127	0.127	0.124	0.118
Cell B −10 °C 1 s	0.334	0.312	0.262	0.219
Cell B −10 °C 10 s	0.468	0.444	0.453	—

. RT R_ints decreases slightly with the increase in the discharge rate. Cell B shows a lower R_ints in all measurements.

3.2.5. Surface Charge-Transfer Resistance

Out of many factors contributing to R_int, ohmic resistance (R₀) and surface charge-transfer resistance (R_ct) can be deduced from the Cole-Cole plot obtained by the alternating current (AC) impedance measurement [35]. Cole-Cole plots of cells A and B measured at RT and −10 °C are shown in Figure 9 and Figure 10, respectively. Calculated R₀ and R_ct values, and double-layer capacitances (C) of cells A and B are listed in Table S2. While R₀ values in both sets are similar, R_ct values in cell B are lower than those in cell A at both RT and −10 °C. Part of the reason for the lower R_ct values in cell B is due to the larger surface reactive area (A) of the superlattice alloy from the connection:

C = εA/d

(3)

where ε and d are the dielectric constant of electrolyte and the alloy surface dipole thickness. Another reason for the lower R_ct values in cell B is the higher surface catalytic ability of the superlattice alloy [28].

3.2.6. Cycle Life

Because the high-power design is usually associated with shallow charge/discharge cycling, a regular C-size configuration with a nominal capacity of 4.5 Ah was used to study the cycle life performance. Cells were built with annealed alloys A and B and tested under a C/2 charge to a −∆V of 3 mV and a C/2 discharge to a cutoff voltage of 0.9 V at RT, and the results are plotted in Figure 11. Without the protective binder commonly used in the commercial cells made with the superlattice alloys [32], the cell made with the superlattice MH alloy (alloy B) only shows half of the cycle life of a cell made with the conventional AB₅ MH alloy (alloy A). Earlier studies on the failure mode of a Mm-based Al-free superlattice MH alloy indicated that the pulverization of the main phase is the main cause of capacity degradation [30].

3.2.7. Comparison

A battery performance comparison between cells made with the AB₅ (cell A) and Al-free A₂B₇-based superlattice (cell B) MH alloys is summarized in Figure 12. Cell B has a higher capacity and a better low-temperature performance; however, it demonstrates slightly worse high-rate dischargeability and charge retention, and its cycle life is only half that of cell A. Despite the lower R_int and R_ct in cell B, it still shows a lower normalized capacity at a higher rate, which may be associated with the relatively low V_oc at RT caused by alloy B’s relatively low equilibrium plateau pressure (Figure 2). In another article, cells made with an Al-containing superlattice MH alloy showed a comparable cycle life and better peak power and charge-retention performance compared to those made with the AB₅ alloy [19]. Therefore, the inferior cycle life observed in the superlattice alloy is only limited to the Al-free composition used in this study. Combined with the use of a hydrophobic binder in the negative-electrode paste [32], the Al-containing superlattice alloy showed even better cycle stability [8].

4. Conclusions

Electrochemical performances of a misch metal-based Al-free superlattice metal hydride alloy were compared to those of a standard AB₅ metal hydride alloy in a high-power C-size cell configuration. In the sealed cell, the superlattice alloy showed higher energy densities, lower internal resistances, lower surface charge-transfer resistances at both RT and −10 °C compared to the AB₅ alloy. For the charge-retention performance, the superlattice alloy was slightly worse at RT but outperformed the AB₅ alloy at −10 °C. The cycle stability of the superlattice alloy tested in a regular cell configuration is inferior to that of the AB₅ alloy mainly because of alloy pulverization.