Introduction
Nickel hydroxide is used as the cathode active material in nickel-based rechargeable alkaline batteries, including nickel-metal hydride (Ni-MH), nickel-cadmium (Ni-Cd) and nickel-iron (Ni-Fe) batteries [
1–
3]. The applications of nickel-based alkaline battery systems range from power tools and portable electronics to electric vehicles. As the performance of nickel-based alkaline batteries are limited by the cathode, considerable efforts have been made to improve the performance of Ni(OH)
2 as the cathode material. The active phase is typically b-Ni(OH)
2 [
4]. In terms of recent progress regarding nickel based batteries, Cheng et al
. [
5] have studied a novel redox flow battery-single flow Zn/NiOOH battery and showed that high efficiencies can be achieved with an average coulombic efficiency of 96% and energy efficiency of 86% over 1000 cycles. In the research conducted by Jia et al
. [
6] by using LiFePO
4 and TiO
2 as the cathodic and anodic Li storage materials, the tank energy density of the redox flow lithium battery has been found to reach ~500 Wh/L, which is 10 times higher than that of a vanadium redox flow battery.
Kohler et al. have proposed that the redox transformation of Ni(OH)
2 to NiOOH involves the release of protons from the Ni(OH)
2 host lattice and the acceptance of protons by the electrolyte [
7]. These reactions would be greatly favored by the use of nano-size cathode materials, with their significantly higher surface-to-volume ratio. Nanomaterials have been gaining popularity for applications in electrochemical power sources due to their special performance and coordination effects when mixed with other materials [
8,
9]. In Ni-based systems, Kohler et al
. [
10] have reported that near theoretical specific charge/discharge capacity (289 mAh/g) has been obtained by using a sample exhibiting broad X-ray peak widths. Liu and Yu [
11] have reported that b-Ni(OH)
2 with particle size ranging from 6 to 9 nm has yielded higher proton diffusion coefficients and an almost 10% increase in charge/discharge capacity compared with micro-sizedb-Ni(OH)
2.
The improved electrochemical performance observed in nano-size b-Ni(OH)
2 makes it an excellent candidate for nano-suspension flow batteries. The concept of a suspension electrode was first demonstrated more than 3 decades ago [
12,
13]. In a nano-suspension, the electroactive nanoparticles serve as energy storage material while suspended in the liquid electrolyte. Recently, Duduta and coworkers [
14] have used Li-ion cathode materials in a suspension form as the electrolyte in a flow battery, performing several successful charge/discharge cycles. Boota and coworkers [
15] have modified the cathode material to reduce the viscosity and increase the conductivity of the nano-suspension. However, many challenges remain, including limited power ratings, limited charge/discharge rates, incomplete discharge, and a short cycle life. The use ofb-Ni(OH)
2 nanoparticles in the KOH electrolyte avoids concerns related to low solubility that usually plagues redox materials used in conventional flow batteries. Because of the potentially high mass loading ofb-Ni(OH)
2 in the nano-suspension, modeling suggests that nano-suspension flow batteries could yield an order of magnitude greater capacity than conventional flow batteries.
The behavior and structural transformations of Ni(OH)
2 cathode in solid electrodes is well understood. X-ray absorption fine structure (XAFS) spectroscopy is a powerful tool for
in-situ study of electrode materials since it makes it possible to capture not only crystalline, but also amorphous phases within the electrodes. XAFS is a technique which examines absorption by an atom near and above its core level binding energy. The XAFS spectrum reflects the physical and chemical states of the targeted atom. It is especially sensitive to the oxidation state, coordination chemistry of the absorbing atom, and to the distances between the targeted atom and its neighboring atoms. Using XAFS spectroscopy, researchers have conducted multiple studies of the local structure of Ni(OH)
2 during charge/discharge cycling. Pandya and O’Grady have showed that the Ni atoms inb-Ni(OH)
2 contains two coordination shells: the 1st containing 6 oxygen atoms at a distance of 2.07 Å from the central Ni atom, and the 2nd containing 6.2 Ni atoms at a distance of 3.13 Å from the central Ni atom [
16]. Ichiyanagi and coworkers have developed Ni(OH)
2 monolayer nanoclusters surrounded by a host network of amorphous SiO
2 with a smaller Ni-Ni distance (verified with XAFS) in the monolayer plane than that in bulk Ni(OH)
2 [
17]. Han and coworkers have conducted a series of XAFS studies which indicated that, at a fully-charged state, the Ni-Ni bond length (2.37 Å) for nano-size Ni(OH)
2 is longer than that for micro-scale Ni(OH)
2 (2.32 Å), demonstrating a higher oxidation state of Ni in nano-scale Ni(OH)
2 electrodes when compared with micro-size Ni(OH)
2. At a fully-discharged state, the longer Ni-O bond length for nano-size Ni(OH)
2 (1.77 Å) when compared with micro-size Ni(OH)
2 (1.72 Å) shows a lower oxidation state of Ni in nanosized Ni(OH)
2 electrode than that in micro-size Ni(OH)
2 [
9]. These results confirm the use of nano-size Ni(OH)
2 particles enabled deeper charge and discharge during cycling. Farley and coworkers have conducted
in-situ XAFS studies on Ni(OH)
2 electrodes during discharge, and they have observed a contraction of the Ni-O shell that is dependent on the state of charge [
18]. The contraction is maximum at charged-state and decreased during the discharging process, with the bond length returning to its original level upon full discharge. With the knowledge of Ni(OH)
2 crystalline structure obtained from previous XAFS works,
in-situ XAFS has been used in the current work to confirm that the change of Ni-Ni and Ni-O bond lengths in Ni(OH)
2 during charge/discharge conforms with the previously reported changes in crystalline structure, providing help for determining the phases present in the electrode during charging and discharging [
9,
18].
This work is part of an effort to create a prototype flow battery using a Ni(OH)2 nano-suspension as the catholyte. Thanks to the advantages of nano-size Ni(OH)2 over micro-size Ni(OH)2 as described earlier, a one-step synthesis method has been employed for the synthesis of b-Ni(OH)2. Nano-size b-Ni(OH)2 have not only formed stable suspensions but also provided faster charge/discharge rates than micro-size variants. The electrochemical charge/discharge performance ofb-Ni(OH)2 has been tested initially in a pouch cell with a solid porous electrode. The change in the oxidation state of Ni has been observed through X-ray absorption near edge spectroscopy (XANES), while the coordination number and the distances between the Ni and its shell atoms (Ni and O) have been determined by extended X-ray absorption fine structure (EXAFS) spectroscopy. Finally, a series of experiments in a flow cell have been performed to demonstrate the feasibility of usingb-Ni(OH)2 nano-suspensions as cathode materials in flow batteries.
Experimental
Synthesis and characterization of Ni(OH)2
b-Ni(OH)2 nanoparticles were prepared by direct precipitation from NiCl2·6H2O (Sigma-Aldrich, 99.9%) by adding NH4OH. 4 g of NiCl2·6H2O was dissolved in 450 mL of deionized water. To prevent the agglomeration of Ni(OH)2 and facilitate uniform precipitation, 0.1 g of sodium dodecyl sulfate (SDS) was used as a surfactant. The resultant solution was stirred and heated to 65°C. After the temperature was established, 5 mL of NH4OH solution (28.0%–30.0% NH3 basis) was added to the solution. The system was kept at 65oC and under continuous stirring for one hour. The heating was then switched off, followed by stirring for another hour. Finally, the light-green nanoparticles obtained were separated by centrifuging and washed with deionized water followed by drying at 60°C overnight.
Scanning electron microscopy (SEM) was operated on SEM-Hitachi S-4700-II (resolution 500 nm), and X-ray diffractometer (XRD) on XRD-Bruker D2 Phaser (Cu Ka source, l = 1.4518 Å, 40.0 kV, 30.0 mA).
Charge/discharge testing of as-prepared solid casted Ni(OH)2 electrode in a pouch cell
A slurry with an 80:10:10 weight ratio between Ni(OH)2, carbon black, and polyvinylidene fluoride (PVDF) binder was prepared as follows: carbon black (0.125 g), polyvinylidene fluoride (PVDF) (0.125 g), and N-Methyl-2-pyrrolidone (NMP) were blended in a vortex mixer for 6 hours. Ni(OH)2 (1 g) was added to the mixture, followed by another 6 hours of mixing in the vortex mixer. The slurry was cast on a nickel foam substrate and was used as the working electrode. Plain nickel foam was used as the counter electrode, and Hg/HgO was used as the reference electrode. 3 mol/L KOH was used as the electrolyte.
The charging and discharging of pouch cells with b-Ni(OH)2 cathodes were performed using a potentiostat (Gamry Reference 600) at C/2 rates. The cut-off charging potential was chosen to be 0.45 V vs. Hg/HgO reference electrode to minimize oxygen evolution. The cut-off discharging potential was 0 V vs. Hg/HgO. Ni(OH)2 was oxidized to NiOOH during battery charge.
In-situ XAFS measurements
In-situXAFS measurements were performed during the charge/discharge cycling of a pouch cell using the as-prepared Ni(OH)2 as the working electrode. The measurement was conducted at the Materials Research Collaborative Access Team (MRCAT) beamline, Sector 10-BM, of the Advanced Photon Source (APS) at Argonne National Laboratory (ANL).
The spectra were collected at Ni-K edge (8333 eV). For this study, a slurry containing 80% Ni(OH)2 (wt), 10% carbon black (wt), and 10% PVDF(wt) was painted on a carbon paper substrate instead of on the nickel foam. Carbon paper was chosen as the substrate to avoid the interference by the Ni foam. A piece of blank nickel mesh was used as the counter electrode, and Hg/HgO as the reference electrode. 3 mol/L KOH was used as the electrolyte. Multiple charge/discharge cycles were completed externally beforein-situ XAFS measurements to ensure that the cell functioned stably prior to placing the pouch cell in the beamline. The experiments have been performed in a pouch cell with the configuration shown in Fig. 1.
The data was analyzed and modeled using IFEFFIT package by fitting the data to the models proposed by Morishita et al
. [
19] that represent the crystalline lattices ofb-Ni(OH)
2, b-NiOOH, and g-NiOOH. The differences among the models are their Ni-Ni and Ni-O bond lengths. With regard to the model for b-Ni(OH)
2, the Ni-Ni bond length was 3.147 Å and the Ni-O bond length was 2.060 Å. In terms of the model forb-NiOOH, the Ni-Ni bond length was 2.927 Å and the Ni-O bond length was 1.866 Å. For the model forg-NiOOH, the Ni-Ni and Ni-O bond lengths were 2.770 Å and 1.877 Å. The model of b-Ni(OH)
2 was used to simulate pristine Ni(OH)
2, the model for b-NiOOH was used for describing the charged state, and the models of b-Ni(OH)
2 and g-NiOOH were employed for the samples in the discharged state.
Charge-discharge test of Ni(OH)2 nano-suspension
The suspension containing 5% of the as-synthesized Ni(OH)2 (wt) was prepared as follows: the as-synthesized Ni(OH)2 nanoparticles (0.5 g) were added into 3 mol/L KOH solution and the suspension was sonicated in a sonic bath for 15 hours to break large agglomerates into smaller agglomerates or single particles.
A Nafion® 211 membrane was used to separate the nano-suspension cell into two chambers of 10 mL. The experiments were conducted by employing a PEM (Nafion®) separator due to its demonstrated chemical stability. To the best of the authors’ knowledge, there are no commercially available anion exchange membranes with certified and demonstrated alkaline stability. The as-prepared nano-suspension was filled into one chamber with a small stirring bar at the bottom. The other chamber was filled with 3 mol/L KOH. Nickel foam (MTI Corp.,>99.99% purity) was used as the current collector in both chambers. Two nickel foams, with a surface area of 4 cm2 and 8 cm2, respectively, were used to compare the influence of current collector surface area on the cell performance. The charge/discharge tests were performed at different charge (4 cm2 electrodes: 7.2, 9.0 and 18.0 mA/cm2; 8 cm2 electrodes: 3.6, 4.5 and 9.0 mA/cm2) and discharge (4 cm2 electrodes: 0.72, 0.9 and 1.8 mA/cm2; 8 cm2 electrodes: 0.4, 0.5 and 0.9 mA/cm2) rates. The reference electrode (Hg/HgO) was inserted into the chamber containing the Ni(OH)2 nano-suspension. The nano-suspension was constantly stirred throughout the electrochemical cycling by a magnetic stirrer.
Results and discussion
Characterization of Ni(OH)2 nanoparticles
Figure 2 shows the histogram of the particle size distribution of the as-synthesized Ni(OH)2 particles. The average size of the nanoparticles was on the order of 80 nm. The crystallite size and purity of the particles were estimated by XRD.
The as-synthesized Ni(OH)2 particles exhibited a typical XRD pattern for b-Ni(OH)2, as illustrated in Fig. 3. All the diffraction peaks could be indexed to single phase b-Ni(OH)2 with a well crystallized hexagonal structure. No peaks attributable to a-Ni(OH)2 were observed in the XRD pattern, suggesting that b-Ni(OH)2 was predominantly obtained. The crystallite size of the as-synthesized Ni(OH)2 particles was estimated from the XRD pattern to be 13.45 nm, suggesting some agglomeration occurred during the synthesis and processing.
Charge/discharge tests of b-Ni(OH)2 cathode in pouch cells
Figure 4 demonstrates the typical charge and discharge curves obtained with the pouch cell. Two plateaus were observed in the charging curve, as reported by Morishita et al. [
16]. The first plateau (0.38 V vs. Hg/HgO) in the charging curve corresponded to the transition ofb-Ni(OH)
2 to b-NiOOH while the second (0.45 V vs. Hg/HgO) related mainly to the oxygen evolution reaction [
19]. The overcharge of b-NiOOH to g-NiOOH also took place in the second plateau.
Figure 5 depicts the change in specific discharge capacity of b-Ni(OH)2 over 9 charge/discharge cycles. The specific discharge capacity increased until a stable capacity of 210 mAh/g (73% of the theoretical capacity of 289 mAh/g) was obtained after 4–5 charge/discharge cycles. The specific discharge capacity was lower than the theoretical capacity due to the oxygen evolution reaction occurred in the second plateau (0.45 V vs. Hg/HgO). The first cycle discharge capacity was approximately 25% lower than the final capacity. Possible explanations for this anomalous capacity include activation processes occurred during the first cycle as well as incomplete charging during the first cycle. The reason for the capacity decay and large overpotential in the flow battery tests is that, at each time point, the amount of Ni(OH)2 contacting the current collector is small compared to the loading of Ni(OH)2 in the nanosuspension. This situation leads to large charging and discharging currents relative to the amount of Ni(OH)2 in direct contact with the current collector, resulting in the overcharge of Ni(OH)2 to g-NiOOH and consequently in a capacity decay.
In-situ XAFS measurements during charge/discharge cycling in a pouch cell
Figure 6 shows the XAFS spectra at the Ni K-edge for the as-synthesized Ni(OH)
2 powder and the Ni(OH)
2 electrodes in an operating pouch cell in the charged and discharged states. The spectra were acquired after stable charge and discharge capacities were established (7 cycles). The Ni K-edge position for the charged electrode shifted 3.2 eV towards positive energies when compared with the pristine Ni(OH)
2, confirming the higher oxidation state of the Ni atoms [
15]. During discharge, the Ni K-edge absorption line partially shifted 1.5 eV back to lower energies, but did not fully return to the initial position, indicating a partial reduction of the Ni atoms and the formation of a mixture of Ni(OH)
2 and NiOOH phases in the discharged electrode.
Figure 7 presents the XAFS spectra (R-space) for pristine Ni(OH)
2, and for the pouch cell electrodes in charged and discharged states. The peaks between 1 Å and 2 Å correspond to Ni-O paths while the peaks between 2 Å and 3 Å correspond to Ni-Ni paths. Both Ni-O and Ni-Ni peaks shifted toward smaller distances after the electrode was charged. An opposite effect was observed during cell discharge [
9,
19].
The results from data fitting using the three models described in the experimental section are shown in Table 1. The distances between the Ni-O and Ni-Ni paths (model forb-NiOOH) in the charged state were shorter than those between the Ni-O and Ni-Ni paths in the as-synthesized sample, indicating that a contraction between Ni and its surrounding Ni and O atoms took place during charging. The level of this contraction is in agreements with the work of Morishita et al
. [
19]. For the as-synthesized Ni(OH)
2, the Ni-O bond length was 2.07±0.02 Å , and the Ni-Ni bond length was 3.11±0.02 Å. In the charged state, those distances were 1.91±0.01 Å and 2.83±0.02 Å , respectively.
Due to the formation of g-NiOOH during charging, full discharge of the battery was never reached during cycling. Therefore, the two models describing theb-Ni(OH)2 and g-NiOOH phases were used to fit the data of discharged electrodes.
The EXAFS results were in agreements with the charge and discharge curves shown in Fig. 4. There were two plateaus in the charging curves, with the higher plateau mainly corresponding to oxygen evolution, and the overcharging ofb-Ni(OH)2 to g-NiOOH also took place.
Now there are two factors that have negative and positive effects on the specific discharge capacity: the formation ofg-NiOOH [
20], and the activation of b-Ni(OH)
2 during the initial charge/discharge cycles [
21].
The discharge of g-NiOOH happened at lower potentials than the discharge of b-NiOOH [
19,
20]. The formation of g-NiOOH contributed to the decay of specific discharge capacity [
20], because the coexistence of b-NiOOH and g-NiOOH resulted in a non-uniform discharge that led to the isolation of the lower discharge potential g-NiOOH. The already-formed g-NiOOH could also reduce the tendency of further formation of g-NiOOH during charging, leading to a slower generation of g-NiOOH during cyclings [
20]. This explains the partial shifting back of the Ni K-edge position for the electrode in the charged state. There is also a factor that contributes positively to the specific discharge capacity. During charge/discharge cycling,b-Ni(OH)
2 kept being activated [
21], leading to more and more materials that can be charged and discharged. But due to the limitation of the amount of material in electrode, the activation process got slower gradually. When those two factors came together, stable specific discharge capacity was obtained as shown previously.
Charge/discharge tests of Ni(OH)2 nano-suspension in a beaker cell
Figures 8 and 9 display the charge/discharge curves for nano-suspension cells loaded with b-Ni(OH)2 nano-suspension containing 5 wt.% of b-Ni(OH)2, with a charge rate of 7.2 mA/cm2 and a discharge rate of 0.72 mA/cm2. In Fig. 8, it can be observed that the charging plateau is higher (between 0.5 and 0.6 V vs. Hg/HgO) than that in the charge/discharge tests in pouch cells. During the cycling in the pouch cells, all the Ni(OH)2 nano-powder was attached to the current collector and could be efficiently charged and discharged, while during the cycling of the beaker cell, at each time, only a small portion of Ni(OH)2 nanoparticles were in contact with the current collector, leading to additional overvoltage losses. A maximum discharge capacity of 35 mAh/g was achieved, which was approximately 17% of the specific discharge capacity observed in the pouch cells and 12% of the theoretical capacity (289 mAh/g) for this cell. The two-plateau behavior was not observed in the nano-suspension tests. Only one plateau during charging appeared with a potential between 0.58 V and 0.6 V vs. Hg/HgO which is much higher than both of the charging plateaus (0.38 V and 0.45 V vs. Hg/HgO) shown in the pouch cell tests. The higher charging cell potential was attributed to the less robust contact between the nanoparticles and the current collector compared with the pouch cell. The charge rate was calculated based on the total amount ofb-NiOH2 in the suspension, and the amount of b-NiOH2 in contact with the current collector was only a very small fraction of the total NiOH2. Therefore, the charging current was very large, leading to high charging overpotentials. Therefore, it can be hypothesized that the main reaction taking place in nano-suspension cell charge/discharge test is oxygen evolution reaction. Someb-Ni(OH)2 was charged to become b-NiOOH and then overcharged to the g-NiOOH due to the high potential. The low discharge capacity was attributed to the low amount of b-Ni(OH)2 that can be effectively charged. During charging, most of the charging current led to the oxygen evolution reaction instead of charging the battery (larger charging potentials than in pouch cells). The XRD patterns of fresh Ni(OH)2 and discharged electrodes (Fig. 10) showed that a part of the NiOOH in the discharged electrode could not be reduced to Ni(OH)2 during cell discharge.
To optimize the operation within the nano-suspension cell, the charge rate and the surface area of the current collector were adjusted. As seen in Fig. 11, higher discharge capacities were obtained when the charge rate was decreased. At higher charge rates, the overpotential was higher, which led to more oxygen. Additionally, in a nano-suspension cell, only a limited amount of the active material was in contact with the current collector, and therefore the charge/discharge rate based on the total amount of active material could not properly represent the charge/discharge phenomena. Figure 12 shows the specific discharge capacity obtained by the two current collectors with an area of 4 cm2 and 8 cm2, respectively, at the same charge/discharge rates. A 2-fold increase in the surface area of the current collectors increased the specific discharge capacity by 300%. The increase in the specific discharge capacity indicated that the current collector surface area rised the amount ofb-Ni(OH)2 effectively charged, improving the specific discharge capacity.
Conclusions
Ni(OH)2 nanopowder was synthesized using the sol-gel method. XRD patterns of Ni(OH)2 nanopowders confirmed the presence of pure b-Ni(OH)2, with a particle size, determined by SEM, of 50–110 nm (average of 80 nm).
Electrochemical tests were performed in pouch cells with b-Ni(OH)2 in the working electrode. The testing included 8 charge/discharge cycles at a rate of C/2. The results showed a discharge capacity of 210 mAh/g (theoretical capacity: 289 mAh/g) that remained stable for at least eight cycles.In-situXAFS was performed at the Ni-K edge to obtain a deeper understanding of the structural changes in the Ni(OH)2 nanoparticle electrode during charge/discharge cycling. Data analysis showed a contraction of the distance between Ni and its surrounding atoms Ni (7.7% contraction) and O (9.0% contraction) after the electrode (b-Ni(OH)2) was charged. XANES results suggested that the electrode in the discharged state was a mixture because the edge position did not completely shift to the original position. EXAFS also proved that the mixture was composed of 79% ofb-Ni(OH)2 and 21% of g-NiOOH. This observation indicated that the discharge capacity was provided by b-NiOOH. The specific discharge capacity (210 mAh/g) was lower than the theoretical specific discharge capacity of 289 mAh/g mainly due to the oxygen evolution reaction taking place (second plateau of the charging curve).
Proof-of-concept nano-suspension tests were also performed, which showed that the Ni(OH)2 nano-suspension had the potential to be used as the active material in slurry-based redox flow batteries. Based on these tests, it had been found that high surface area current collectors and low charging rates yielded higher specific discharging capacities. The current collector with a surface area of 8 cm2 gave a 300% higher specific discharge capacity (10 mAh/g) than that (35 mAh/g) given by the current collector with a surface area of 4 cm2. When discharge rate stayed the same (C/50), a slower charge rate yielded a higher specific discharge capacity: charge rate C/2 gave 6 mAh/g, charge rate C/4 gave 8 mAh/g, and charge rate C/5 gave 13 mAh/g.
Higher Education Press and Springer-Verlag Berlin Heidelberg
PDF
(393KB)
