- Email: [email protected]
Significantly improved electrochemical hydrogen storage properties of magnesium nickel hydride modified with nano-nickel
Journal of Power Sources 280 (2015) 132e140
Contents lists available at ScienceDirect
Journal of Power Sources journal homepage: www.elsevier.com/locate/jpowsour
Significantly improved electrochemical hydrogen storage properties of magnesium nickel hydride modified with nano-nickel Wei Chen, Yunfeng Zhu*, Chen Yang, Jiguang Zhang, Menghuai Li, Liquan Li College of Materials Science and Engineering, Nanjing Tech University, 5 Xinmofan Road, Nanjing 210009, PR China
h i g h l i g h t s The HCS product of Mg2NiH4 is modified with nano-nickel via mechanical milling. The discharge capacity and cycle stability are dramatically improved. A novel NiH0.75 phase is observed and the formation mechanism is discussed. Some insights into “electrochemical catalytically” enhanced mechanism are proposed.
a r t i c l e i n f o
a b s t r a c t
Article history: Received 1 November 2014 Received in revised form 11 December 2014 Accepted 14 January 2015 Available online 14 January 2015
Magnesium nickel hydride (Mg2NiH4) used as negative electrode material in nickel-metal hydride (NiMH) secondary battery is modified by nano-nickel via mechanical milling. In this paper, we systematically investigate the microstructure and electrochemical properties of the modified system with different milling durations. X-ray diffraction (XRD) and high resolution transmission electron microscopy (HRTEM) analyses confirm the amorphous transformation of Mg2Ni-based hydride and a novel NiH0.75 nanocrystalline with a diameter of about 5 nm embedding or covering on the surface of the base particle has been observed. Its formation mechanism and positive effects on electrochemical properties of the Mg2NiH4 have also been elaborated. Electrochemical measurements show that the 5 h milled composite possesses markedly increased discharge capacity up to 896 mAh g1. With prolonging the milling duration from 5 h to 40 h, the discharge capacity at the 10th cycle increases from 99 mAh g1 to 359 mAh g1. Besides, the discharging procedure changes from stepwise processes to one single-step process with increasing the milling duration. Tafel polarization test shows that the nano-nickel modified system exhibits a much better anti-corrosion ability during charging/discharging cycles. Meanwhile, both the charge-transfer reaction on the alloy surface and hydrogen diffusion inside the alloy bulk are enhanced with nano-nickel modification. © 2015 Elsevier B.V. All rights reserved.
Keywords: Magnesium-based hydrogen storage alloy Microstructure Electrochemical characteristics Surface modification
1. Introduction Since the discovery of the reversible hydrogen storage alloy in the late 1960s, nickel-metal hydride (Ni/MH) batteries have been developed as one of the most important energy storage carriers thanks to their high energy density and good environmental compatibility with tremendous development in the last few decades. Future Ni/MH batteries are expected to store a large amount of energy without the increase in weight to satisfy the portable equipment and mobile applications [1e4].
* Corresponding author. E-mail address: [email protected] (Y. Zhu). http://dx.doi.org/10.1016/j.jpowsour.2015.01.089 0378-7753/© 2015 Elsevier B.V. All rights reserved.
Among the hydride-forming materials, Mg-based alloys, especially Mg2Ni, exhibit promising properties as negative electrode materials in Ni/MH batteries because of their higher hydrogen storage capacity, lower specific weight and less cost compared with the commercial LaNi5-type alloys. Considering the objectives proposed by the U.S. Department of Energy (DOE) for the future hydrogen storage materials, magnesium and its alloys are expected to meet the target of high hydrogen storage capacity [2,5]. What is undesirable with it is that the lower practical discharge capacity, as well as the inferior cyclic stability stemming from the severe oxidation of Mg in alkaline solution restrains its practical application [6e8]. Meanwhile, the charging/discharging process is a solideliquid reaction, and the reaction rate is affected very much on the solideliquid interface. Hence, it is very necessary to modify the
W. Chen et al. / Journal of Power Sources 280 (2015) 132e140
surface of hydride particles, aiming to protect the base materials and accelerate the electrochemical reaction. In the past several decades, surface modification has been proved to be an effective method to improve the comprehensive electrochemical performance of the Mg2Ni electrode alloys. Besides the traditional surface modification methods, namely acid, alkali and F-treatments which usually affect the surface composition of parent alloy, several groups have reported the improvement in capacity degradation of the alloy electrodes by microencapsulating the alloy particles with various coating layers such as Ni, Cu, Co, polymer, etc via electroless [9e12]. Luo [13] found that surface modification with NieP coatings by electroless plating effectively improves the cycle stability of Mg2Ni-based alloy. Hwang [14] reported that electroless Ni coating can effectively improve the discharge capacity of Mg2Ni alloys. Shen [15] reported that surface modification with polyaniline by electroless deposition improves the high rate discharge ability (HRD) of the AB5-type alloy electrode. However, the electroless coating process is complex due to its special self-catalyzed chemical reactions. Impurity elements like P, S, B, may inevitably be introduced, increasing the charge transfer resistance and hindering the diffusion of H as well. Coating layers introduced by mechanical milling (MM) have been confirmed to be effective in improving the performance of hydride electrodes by preventing corrosion of the disintegrated alloy, bringing about a superior electric catalytic activity. Among the commonly used dopants, including the transition metals, oxides and various carbon sources (graphite, CNTs, carbon black, etc) [8,16e20], Ni exhibits remarkable positive effects in terms of the utilizing of hydrogen absorbability and desorbability of the Mg2Ni-type alloy electrodes, which has been attributed to the outstanding catalytic activity and electrical conductivity of Ni [21,22]. In our previous work, we have reported a novel method of preparing Mg2NiH4 with high activity and discharge capacity, namely hydriding combustion synthesis with subsequent mechanical milling (HCS þ MM) [23]. However, the as-prepared HCS þ MM product still needs investigation due to the severe corrosion of Mg in alkaline solution. In order to improve the electrochemical performance, especially the cycle stability of the HCS þ MM product, a further modification is thus needed. Therefore, in this work, nano-nickel is introduced to modify the Mg2NiH4 via mechanical milling. The influence of milling duration on the structures and electrochemical properties of the hydride is investigated in detail. These results not only lead to an optimized Mg2NiH4/nano-nickel composite system, but also provide some insight into the “electrochemical catalytically” enhanced mechanism of nano-nickel.
133
phase of Mg2NiH4 and the minor phase of Mg2NiH0.3. The HCS product and commercial nano-nickel powder (99.7 wt.% in purity and 50 nm in particle size) were mixed in 1:2 M ratio. Then the powder mixtures were put into the stainless steel vials (volume 100 ml) for mechanical milling. The milling was carried out under argon atmosphere at room temperature using a planetary-type ball mill at a speed of 400 r min1 and ball to powder ratio of 40:1. In order to investigate the influence of milling duration, the samples were milled for 5, 10, 20 and 40 h, respectively.
2.2. Sample characterization All the testing electrodes were prepared as follows: 0.1 g sample powder was mixed with 0.4 g Ni powder, and then cold-pressed into a pellet of 10 mm diameter and about 1 mm thickness under a pressure of 12 MPa. Electrochemical measurements were performed in a three-electrode cell in 6 M KOH at 30 ± 1 C, using sintered Ni(OH)2/NiOOH as the counter electrode and Hg/HgO as the reference electrode. The discharge capacities of electrodes were evaluated by the amount of active substance of Mg2NiH4. The discharge capacity and the cycle life were determined by the galvanostatic method using CT2001A Land battery testing system. The electrodes were discharged at 30 mA g1 to a cut-off potential of 0.6 V (vs. Hg/HgO) and then charged at 300 mA g1 for 2 h after 5 min rest. To investigate the HRD of the alloy electrodes, discharge capacities at different current densities (100, 200, 400 mA g1) were measured at the first cycle. Linear polarization curves were measured at a scanning rate of 0.1 mV s1 from 5 to 5 mV (vs. open circuit potential) at 50% depth of discharge (DOD). Tafel polarization curves were measured at a scanning rate of 1 mV s1 from 300 to 300 mV (vs. open circuit potential) at 100% DOD. Electrochemical impedance spectroscopy (EIS) studies of the electrodes were performed in a frequency range of 100 kHz to 5 mHz with an AC amplitude of 5 mV at 50% DOD under open-circuit conditions, using ZPLOT electrochemical impedance software. For the potentiostatic discharge, the electrodes were discharged at þ600 mV (vs. open circuit potential) potential step and 100% depth of charge (DOC) for 3600 s. The above electrochemical tests were performed at room temperature in the 6 M KOH aqueous solution on a CHI660C electrochemical workstation.
2. Experimental 2.1. Sample preparation The HCS product of Mg2NiH4 was prepared from commercial magnesium powder (99.9 wt.% in purity and <150 mm in particle size) and nickel powder (99.7 wt.% in purity and 2e3 mm in particle size). They were mixed in 2:1 of Mg:Ni molar ratio by an ultrasonic homogenizer in acetone for 1 h. After being completely dried, the well-mixed powder was placed directly into the synthesis reactor without compressive treatment. The hydriding combustion synthesis reactor used in the present work is the same as that reported in our previous work [24]. During HCS process, the mixed powder was heated to 850 K at the rate of 7 K min1 and held for 1 h under 1.8 MPa hydrogen pressure. Subsequently, the sample was cooled down to room temperature under hydrogen atmosphere. The XRD pattern of the as-prepared HCS product of Mg2NiH4 is shown in Fig. S1. It can be seen the HCS product is composed of the main
Fig. 1. XRD patterns of the Mg2NiH4/nano-nickel composite milled for different durations: (a) 5 h, (b) 10 h, (c) 20 h, (d) 40 h.
134
W. Chen et al. / Journal of Power Sources 280 (2015) 132e140
XRD measurements were carried out using a Rigaku SmartLab TM 3 kW diffractometer with Cu Ka radiation. The morphologies and microstructures of the samples were analyzed by an ULTRA 55 field emission scanning electron microscope (FESEM) equipped with an energy dispersive spectrometer (EDS) and HRTEM (JEM2010 UHR). The particle size distribution of the powder samples were examined by laser diffraction particle size analyzer model Malvern Mastersizer 2000.
was further confirmed in the following TEM analysis. The appearance of NiH0.75 pahse is very interesting and has never been reported previously for the Mg2Ni-based composite. To be noted, as the MM was conducted under Ar atmosphere, the H in NiH0.75 phase can only come from the Mg2NiH4. This implies that a new chemical reaction may occur between Mg2NiH4 and nano-nickel during the intensive milling, which can be speculated as follows:
Mg2 NiH4 þ Nano Ni/Mg2 NiHx þ Nano NiH0:75 3. Results and discussion 3.1. Structural properties of the Mg2NiH4/nano-nickel composite Fig. 1 shows the XRD patterns of Mg2NiH4/nano-nickel composite milled for different durations. From Fig. 1a, it can be seen that the diffraction peaks of Mg-based hydride have almost disappeared after only 5 h milling, indicating that the HCSed Mg2NiH4 has transformed into nanocrystalline or amorphous-like structures. Producing amorphous-like structure of the Mg2Ni alloy by ballmilling has been reported by Kohno et al. [25]. Iwakura [26] and Li [27] also noted the Ni powder can enhance the ball-milling efficiency, accelerating the formation of amorphous structure of Mg2Ni. With prolonging the milling duration, the peak intensity of nano-nickel decreases apparently and almost disappears after 40 h milling, while some new peaks appear at around 42.1, 48.9 and 71.8 after milling for 20 h. These peaks become sharper with increasing the milling duration to 40 h. By checking the standard spectrum, we attributed these peaks to NiH0.75 (84-0450), which
(1)
The possible reaction mechanism is discussed as following. The MM is accepted as a powerful method involving recurrent grinding and cold welding, which can lead to massive atomic diffusion among the powders. According to our previous work, part of the H atoms may deintercalate from the Mg2NiH4 lattice during the MM process. Besides, the nano-nickel possesses much higher reaction activity because of its smaller particle size and larger surface area. Thus, it is possible for the nano-nickel to interact with the H to form the new NiH0.75 phase. Fig. 2 shows the SEM images of the Mg2NiH4 milled with nanonickel for different durations. Fig. 2a shows that the 5 h milled powders are irregular in shape and attached with small flake particles which may be due to the agglomerate nano-nickel. It can also be seen that the exposed surface of Mg2NiH4 particles is smooth, indicating the coating of nano-nickel is incomplete. The 10 h milled sample presents a similar morphology to the 5 h milled sample, also showing the inhomogeneous coating of nano-nickel on the surface of the Mg2NiH4 particles. However, with further increasing the milling duration to 20 h and 40 h as shown in Fig. 2ced, we can
Fig. 2. SEM micrographs of the Mg2NiH4/nano-nickel composite milled for different durations: (a) 5 h, (b) 10 h, (c) 20 h, (d) 40 h.
W. Chen et al. / Journal of Power Sources 280 (2015) 132e140
observe that the surface of Mg2NiH4 particles becomes very rough, and almost no smooth surface can be observed. Thus, we can conclude that the nano-nickel coating is well-proportioned for the samples with long milling duration, indicating that long milling duration is in favor of the formation of homogeneous nano-nickel protective layer. To further confirm the homogeneity of nano-nickel coating layer, EDS analysis has also been conducted and the results are shown in Fig. S2. The results show that all the samples contain the elements of Mg, Ni, O and C. The oxygen may come from the oxidation of samples when exposed to air, and the carbon comes from the conductive adhesive. For each sample, different points on the particle surface were analyzed by EDS. The Mg: Ni (atomic ratios) for all the samples are presented in the figures to evaluate the homogeneity of nano-nickel coating. Before coating, the EDS analysis shows that the Mg: Ni ratios are all close to 2 at different points, and one of the spectra is shown in Fig. S2a, which agrees well with the Mg: Ni ratio of Mg2NiH4. For the 5 h milled Mg2NiH4 þ 2 nano-nickel composite, the EDS analysis shows that the Mg: Ni ratio is always deviated from the designed value of 0.667, and one of the spetra is shown in Fig. S2b. The Mg: Ni ratio of 0.923 is larger than 0.667, indicating that the nano-nickel coating is incomplete. The similar phenomenon is also found for the 10 h milled composite, in which the Mg: Ni ratio is 0.821 as shown in Fig. S2c, also larger than 0.667. However, with increasing the milling duration to 20 h and 40 h, the EDS analyses show that the
135
Mg: Ni ratios at different points are all close to the designed value of 0.667, which can be seen in Fig. S2d and Fig. S2e. The results further confirm that the nano-nickel coating is well-proportioned on the Mg2NiH4 particle surface with longer milling durations. Fig. S3 shows the particle size distribution of the 5 h milled bare Mg2NiH4 and Mg2NiH4/nano-nickel composite milled for different durations. The 5 h milled bare Mg2NiH4 shows an average particle size of 4.37 mm. When nano-nickel was introduced, the average particle size decreases dramatically to 1.42 mm. With increasing the milling duration, however, the average particle size increases. The average particle sizes are 1.63 mm, 11.4 mm and 15.4 mm for the 10 h, 20 h and 40 milled composites, respectively. The increasing in average particle size of the samples with longer milling durations is attributed to the agglomeration of the particles, especially for the 20 h and 40 h milled samples. TEM analysis was used to investigate the microstructure of the sample milled for 40 h. It can be seen from Fig. 3a that nanocrystals with a diameter of about 5 nm are embedded or covered on the surface of the base particle. HRTEM micrograph (Fig. 3b) further confirms that Mg2NiH4 has transformed to amorphous state after 40 h milling. The lattice fringes of inlaid nanocrystals with a separation of 0.1836 nm agree well with the (200) interplanar spacing of NiH0.75. As discussed above, the formation of nanocrystalline NiH0.75 could be attributed to the high active nano-nickel, which catches the desorbed H from Mg2NiH4 and forms the NiH0.75. As the
Fig. 3. TEM images of 40 h milled Mg2NiH4/nano-nickel composite: (a, b) bright field image and HRTEM image for as milled sample; (c) HRTEM image for discharged sample; (d) HRTEM image for recharged sample. The insets give the corresponding selected area electron diffraction patterns.
136
W. Chen et al. / Journal of Power Sources 280 (2015) 132e140
desorbed H atoms mainly gather around the Mg2NiH4/nano-nickel interface, the formation of NiH0.75 is influenced by the concentration of H at the two-phase interface and governed by a slow process of H diffusion in the bulk crystal [28,29]. The NiH0.75 phase was not observed in the 5 h and 10 h milled samples, indicating long milling duration is necessary for both diffusion and aggregation of desorbed hydrogen. Furthermore, the inset of Fig. 3b shows the corresponding selected area electron diffraction (SAED) pattern, which displays a distinguishable halo and ring-like feature of nanocrystalline and amorphous structures. Fig. 3c shows the HRTEM micrograph and corresponding SAED pattern of 40 h milled sample after discharging, revealing that the discharging process wouldn't cause crystallization of the amorphous Mg2NiH4. From Fig. 3c, we can also observe that the nanocrystals distribute on the surface of the bulk materials and the lattice fringes with a separation of 0.2107 nm agree well with the (111) interplanar spacing of NiH0.75. The recharged sample is shown in Fig. 3d and the corresponding SAED pattern is inserted. No obvious structure change is observed, which confirms that the charging process wouldn't cause crystallization of the amorphous Mg2NiH4 either. Moreover, the NiH0.75 couldn't absorb and desorb hydrogen reversibly during the cycles, which is further supported by Fig. S4. No diffraction peaks of Mg2Ni-based alloy/hydride are detected after cycling, which proves again that charging/discharging doesn't cause crystallization of the amorphous Mg2NiH4. From Fig. S4, the NiH0.75 phase is observed in all the samples, illustrating that the NiH0.75 phase doesn't participate in the charging/discharging process.
Table 1 Electrochemical properties of samples milled for different durations. Samples Cmax (mAh g1)
C10 (mAh g1)
Ecorr (V)
5h 10 h 20 h 40 h
99 146 265 359
0.839 80.9 0.838 86.7 0.833 100.5 0.828 91.6
896 996 968 946
I0 Rct (mA g1) (mU) 197 191 183 187
D/a2 ( 105 s1) 1.91 2.17 2.48 2.87
The effect of milling duration on the discharge behavior of different samples is shown in Fig. 4. Note that the electrodes were discharged directly after milling without charging and the maximum discharge capacities were found in the first cycle. It can be seen that the initial discharge capacity increases firstly with prolonging the milling duration and reaches the maximum value of 996 mAh g1 after 10 h milling. However, if the milling duration is prolonged continuously, the capacity decreases slightly, which may be caused by the desorption of hydrogen during milling. The maximum discharge capacities of samples are listed in Table 1. The Mg2NiH4 milled only for 5 h with nano-nickel exhibits a capacity of
896 mAh g1, which is much higher than 480 mAh g1 for the 5 h milled bare Mg2NiH4 [30]. We confirm that the amorphous structure is a key factor to achieve high discharge capacity [31]. Besides, nano-nickel plays an important role in increasing the discharge capacity. First of all, the nano-nickel can accelerate amorphization of Mg2NiH4 dramatically. Through MM, nano-nickel will well inlay and disperse on the hydride particle surface, forming a coating layer. This protective layer can impede the formation of nonconductive Mg(OH)2 layer which is accepted as a main reason for the capacity loss. Meanwhile, the coated nano-nickel is helpful in decreasing the electrochemical reaction resistance on the alloy surface and acts as current collector because of its good electrocatalytic activity in alkaline electrolyte [32]. Considering the superior discharge property, we believe the new NiH0.75 phase may also act as an electro-catalyst during the charging/discharging process. Overall, nano-nickel coating can simultaneously improve the electrochemical kinetics and protect the active substances from corrosion. Consequently, synergistic effects of the nano-nickel layer and the MM process result in the high capacity of the HCSed Mg2NiH4. Furthermore, as seen in Fig. 4, both the discharge curves of samples milled for 5 h and 10 h consist of two distinct discharge plateaus. This indicates that two different hydrogen desorption phases exist in the as-milled composite, corresponding to two independent desorption steps of hydrogen. In other words, there are two different discharging steps during the discharging process. We infer that the first discharging voltage plateau located at about 0.95 V could be attributed to the hydrogen stored in the nanocrystalline Mg2NiH4 and the free H atoms existing in the crystal defects and grain boundaries introduced by MM. Because the nanocrystalline hydride is easy to be corroded by the KOH electrolyte and the free H atoms do not have a definite location within the lattice, we concluded the first discharging step was
Fig. 4. Discharge curves of samples milled for different durations (Discharge rate: 30 mA g1; the first cycle).
Fig. 5. Discharge capacities as a function of cycle number for samples milled for different durations.
3.2. Electrochemical properties of the Mg2NiH4/nano-nickel composite
W. Chen et al. / Journal of Power Sources 280 (2015) 132e140
irreversible. As the second sloping plateau contributes most of the capacity, it should correspond to the main phase of the amorphous hydride. As the milling duration increases to 20 h and 40 h, it is obvious that the first discharge plateau becomes narrower and the second discharge plateau becomes longer, indicating the discharging procedure changes from stepwise processes to one singlestep process. We believe the slight decrease in discharge capacity is caused by the formation of new NiH0.75 phase, which cannot desorb hydrogen electrochemically. However, the amorphization level of Mg2NiH4 should increase with prolonging the milling duration, which is favorable to the electrochemical properties. Fig. 5 shows the discharge capacity degradation as a function of the cycle number of samples milled for different durations. It is obvious that with increasing the milling duration the cycle stability of samples is significantly improved. The discharge capacities at the 10th cycle are listed in Table 1. As can be seen in Fig. 4, the initial discharge capacities of different samples show small differences and they all reach their maximum discharge capacities at the first cycle without precharging, indicating that all the milled samples have a good activation property. Unfortunately, a dramatic drop of the discharge capacity is observed at the second cycle. The severe corrosion of Mg in the alkaline solution is usually thought as the main reason for capacity loss [33]. We discovered that the irreversibility of the free hydrogen and the poor anticorrosion ability of nanocrystalline hydride mainly caused the huge loss of discharge capacity in the early stage of cycles. It can also be seen that the capacities for the samples milled for 5 h and 40 h at the second cycle are 215 mAh g1 and 538 mAh g1, respectively, revealing increasing the milling duration can slow down the rate of degradation significantly. According to the above results, the amorphization level of the sample increases with the increase of milling duration. Unlike crystalline alloys, the amorphous alloys usually have good resistance to corrosion and pulverization because of the absence of any conventional grains and grain boundaries and continuous volume change when H atoms get in or out of the alloy particles. Similar results have also been reported by Horikiri [34] and Huang [35]. Moreover, with increasing the milling duration, a more complete and homogeneous nano-nickel coating is produced on the alloy surface, which is effective in protecting the active substance from oxidation failure and maintaining the electrocatalytic activity of the alloy, improving the charge/discharge cycle performance. Fig. 6 is proposed to give a further understanding of the charging/discharging process, which shows a schematic representation of electrochemical hydrogen ab/desorption. During
137
Fig. 7. Tafel polarization curves of samples milled for different durations (scan rate: 1 mV s1).
discharging, as illustrated in Fig. 6a, hydrogen atoms are desorbed first from the amorphous Mg2NiH4 and then diffuse to the twophase (Mg2NiH4 and nano-nickel) interface. The surface of nanonickel is where the oxidation of H mainly occurs in our system because of their high electrocatalytic activity. With increasing the milling duration, more nano-nickel will inlay and coat on the Mg2NiH4 surface, which means more two-phase interfaces and active areas. Thus, more efficient hydrogen-transfer reactions (H/Hþþe), rather than the generation of H2 (H þ H/H2), will be observed. Meanwhile, the milling process makes the free hydrogen transform into the NiH0.75 phase. The distribution of the in-situ synthesized NiH0.75 with nanosize can act as the catalyst, which not only brings large contacts of catalytic active sites with Mg2NiH4, but also has a great advantage of the fast electron transfer and the diffusion of hydrogen atoms. Furthermore, the positive effect of NiH0.75 can also be understood by viewing NiH0.75 as a grain growth inhibitor [36] that prevents the crystallization of amorphous Mg2NiH4 during the charge/discharge cycles. However, further indepth research is needed to fully understand the effects of the nanocrystalline NiH0.75. The surface oxide layer on the hydride particle is believed to prohibit both the hydrogen electrode reaction (HER) on the interface of electrode/electrolyte and the hydrogen diffusion. Therefore, even after the alloy powders have been cracked due to the
Fig. 6. The charging/discharging model of nano-nickel coated Mg2Ni-based hydride: (a) The discharging mechanism after nano-nickel coating and nano-nickel coating diagram in top left corner; (b) The charging/discharging diagram after disintegration.
138
W. Chen et al. / Journal of Power Sources 280 (2015) 132e140
expansion and contraction of the particle during the charging/ discharging process [37], the new exposed surface exerts little influence on the electrode performance as the fresh surface will be corroded very quickly. However, the nano-nickel coating layer can provide both the active sites for the redox reaction of hydrogen and the pathway for the hydrogen diffusion [13], as shown in Fig. 6b. All of these factors result in a prominent improvement of the discharge capacity by about 300 mAh g1 in each cycle with increasing the milling duration from 5 h to 40 h. To investigate the anti-corrosion ability of the alloy electrodes, Tafel polarization test was performed. The polarization curves for different samples are shown in Fig. 7 and the corrosion potential Ecorr are listed in Table 1. The Ecorr value shifts toward the positive direction with increasing the milling duration, suggesting favorable effect of the milling treatment on the anti-corrosion ability of the alloy electrode owing to the formation of the nano-nickel coating layer and the more complete amorphization. To further understand the anti-corrosion mechanism, a pure Mg2NiH4 sample milled for 5 h was chosen for comparison. Tafel polarization curves of the 5 h milled bare Mg2NiH4 and 40 h milled Mg2NiH4/nano-nickel composite at different cycles are shown in Fig. S5. As we can see, the Ecorr of the 5 h milled bare Mg2NiH4 shifts toward the negative direction severely with increasing the cycle number, indicating the loose Mg(OH)2 layer has no protective effect. Instead, it leads to the exposure of more fresh surface to the electrolyte and thus severer corrosion occurs. At the 10th cycle, the outer layer of Mg2NiH4 particle was almost corroded completely, forming a thick Mg(OH)2 layer. Therefore, it became difficult for alkaline solution to permeate through, resulting in the positive shift of the Ecorr. Fortunately, the Mg2NiH4/nano-nickel composite after 40 h milling exhibits an acceptable Ecorr value and its variation is mild with the increasing cycle number, which obviously reveals that it has a better anti-corrosion ability than the 5 h milled Mg2NiH4 sample during cycles. The HRD of the samples milled for different durations is shown in Fig. 8. The value of HRD is calculated according to the following equation:
HRD ¼
Cd 100% Cd þ C60
(2)
where Cd is the discharge capacity with cut-off potential of 0.6 V (vs. Hg/HgO) at the discharge current density Id (100, 200, 400 mA g1), C60 is the residual discharge capacity with the same
Fig. 8. High rate discharge ability (HRD) of the samples milled for different durations.
cut-off potential at the discharge current density I (60 mA g1) after the alloy electrode has been fully discharged at the discharge current density Id. It can be found that the HRD gets improved with increasing the milling duration, and the 20 h milled sample exhibits the highest HRD, suggesting that 20 h milled sample has the best electrochemical kinetics. Generally speaking, the HRD of the hydrogen storage electrode alloy is determined by charge-transfer on the alloy surface and hydrogen diffusion in the alloy bulk [38]. It is known that the transition-metal Ni has the unpaired electrons and unoccupied atomic orbitals, which can interact with the base metal and then a synergistic effect could exist between the Mg2NiH4 and nano-nickel. Thus, the charge-transfer reaction will be accelerated as the hydrogen atom diffuses near the Ni atom. Furthermore, the nano-nickel coating layer can block the formation of Mg(OH)2 and the two-phase interface of the nano-nickel and the bulk alloy can provide a diffusion channel for the hydrogen. To prove this, linear polarization, EIS and potentiostatic discharge measurements were performed. The exchange current density (I0) is used to characterize the electrocatalytic activity for the charge-transfer reaction on the surface of the alloy electrodes. Fig. 9 presents the linear polarization curves for the samples milled for different durations measured at 50% DOD. The exchange current density can be estimated from the slope of the obtained micropolarization curve using the following equation:
I0 ¼
RTI Fh
(3)
in which R, T, I, F andhare the gas constant, the absolute temperature, the applied current density, the Faraday constant and the total overpotential, respectively. The results obtained are also listed in Table 1. A high value of I0 corresponds to a good kinetics of chargetransfer on the alloy surface. It can be found the I0 value increases first and then decreases with increasing the milling duration. The 20 h milled sample possesses the highest I0 value of 100.5 mA g1. Linear polarization curves of the 5 h milled bare Mg2NiH4 and 40 h milled Mg2NiH4/nano-nickel composite at different cycles are shown in Fig. S6 and the calculated I0 values are also tabulated in Table S1. It is obvious that with cycling the 40 h milled Mg2NiH4/ nano-nickel composite exhibits much bigger I0 values, indicating that MM and nano-nickel coating can improve the charge-transfer reaction dramatically.
Fig. 9. Linear polarization curves for samples milled for different durations (scan rate: 0.1 mV s1).
W. Chen et al. / Journal of Power Sources 280 (2015) 132e140
The EIS and corresponding equivalent circuit of samples milled for different durations are shown in Fig. 10. It can be seen that each spectrum is composed of two semicircles followed by a straight line at the low-frequency region. It has been reported by Kuriyama et al. [39] that the semicircles in the high-frequency region and lowfrequency region correspond to the contact resistance (Rc) and the charge-transfer resistance (Rct), respectively, and the sloped straight line corresponds to Warburg impendence (W). Based on the equivalent circuit and by means of non-linear least squares (NLLS) fitting of the plots, the charge-transfer resistances of the samples were calculated and listed in Table 1. As we can see, the Rct decreases first and then increases with increasing the milling duration. The 20 h milled sample possesses the smallest Rct value of 187 mU. This variation of the Rct is consistent with the result obtained from the linear polarization. Fig. S7 shows the EIS of the 5 h milled bare Mg2NiH4 and 40 h milled Mg2NiH4/nano-nickel composite at different cycles. The charge-transfer resistances of samples at different cycles were also calculated and listed in Table S1. To investigate the hydrogen diffusion behavior of the alloy, potentiostatic discharge measurements of the hydride electrodes were performed and the results are shown in Fig. 11. It can be seen that at the initial stage of discharging (<500 s), the current declines sharply, and after a time response of about 2000 s the current decreases slowly in a linear fashion. In the linear region, the hydrogen diffusion in the bulk of the alloy dominates the electrochemical reaction process and the hydrogen diffusion coefficient parameter D/a2 can be calculated according to the following equation [40]:
6FD p2 D log i ¼ log ± 2 ðC0 Cs Þ t 2:303 a2 da
(4)
where i (A g1) is the current density, F is Faraday constant, D (cm2 s1) is the hydrogen diffusion coefficient, d (g cm3) is the density of the hydrogen storage alloy, a (cm) is the alloy particle radius, C0 (mol cm3) is the initial hydrogen concentration in the bulk of the alloy, Cs (mol cm3) is the hydrogen concentration on the surface of the alloy particles, and t (s) is the discharge time. As the average particle size is not identical for the different samples, it is reasonable to use D/a2 to evaluate the discharge kinetics of the electrode contributed by hydrogen diffusion within the bulk of the particles. The values of D/a2 were determined from the slopes of linear current responses and listed in Table 1. It is found that the D/
Fig. 10. Electrochemical impedance spectra and corresponding equivalent circuit for samples milled for different durations measured at 50% DOD.
139
Fig. 11. Potentiostatic discharge curves for samples milled for different durations measured at 100% DOC.
a2 value increases continuously with increasing the milling duration. A longer milling duration can not only increase the amorphous level, but also introduce more defects which facilitate the hydrogen diffusion procedure. In the present work, the nanocrystalline NiH0.75 can be regarded as the activated channel for hydrogen transferring, which accelerates the hydrogen diffusivity in the bulk. To investigate the hydrogen diffusion behavior with cycling, potentiostatic discharge measurements of 5 h milled bare Mg2NiH4 and 40 h milled Mg2NiH4/nano-nickel composite at different cycles were performed and the results are shown in Fig. S8 and the calculated D/a2 values are listed in Table S2. The 40 h milled Mg2NiH4/nano-nickel composite exhibits much higher D/a2 value at each cycle. 4. Conclusions In order to improve the electrochemical properties of Mg2Nibased hydrides, nano-nickel is used to modify the HCSed Mg2NiH4 via mechanical milling. Effect of different milling durations on the structures and electrode performances is investigated in detail. XRD results reveal the amorphization of Mg2NiH4 and especially, a novel phase NiH0.75 appears after 20 h milling. HRTEM observation further confirms the NiH0.75 nanocrystals with a diameter of about 5 nm embedded in the amorphous bulk. We infer the formation of NiH0.75 could be controlled by a slow process of H diffusion between the hydride and the nano-nickel during the MM process. SEM and EDS analyses show that the nano-nickel coating becomes well-proportioned with increasing the milling duration to 20 h and 40 h. Electrochemical measurements indicate that both the discharge capacity and the cycle stability are improved dramatically as the milling duration increases. The maximum discharge capacity of the composite reaches 996 mAh g1 which is very close to its theoretical value. With prolonging the milling duration from 5 h to 40 h, the discharge capacity at the 10th cycle increases from 99 mAh g1 to 359 mAh g1. We also find the discharging procedure changes from stepwise processes to one single-step process with increasing the milling duration. Tafel polarization test shows that the nano-nickel modified system exhibits a much better anticorrosion ability during cycles. Meanwhile, both the chargetransfer reaction and hydrogen diffusion are enhanced with nano-nickel modification. Taken together, the 20 h milled system exhibits the best electrochemical kinetics, which is mainly controlled by the charge-transfer reaction on the alloy surface.
140
W. Chen et al. / Journal of Power Sources 280 (2015) 132e140
Acknowledgments This work was supported by the National Natural Science Foundation of China (51471087, 51171079), Natural Science Foundation of the Jiangsu Higher Education Institutions of China (13KJA430003), Innovation Foundation for Graduate Students of Jiangsu Province (KYLX_0741), Qing Lan Project and the Priority Academic Program Development (PAPD) of Jiangsu Higher Education Institutions. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.jpowsour.2015.01.089. References [1] S.R. Ovshinsky, M.A. Fetcenko, J. Ross, Science 260 (1993) 176e181. [2] L. Schlapbach, A. Züttel, Nature 414 (2001) 353e358. [3] K. Young, D.F. Wong, S. Yasuoka, J. Ishida, J. Nei, J. Koch, J. Power Sources 251 (2014) 170e177. [4] Y.F. Liu, Y.H. Cao, L. Huang, M.X. Gao, H.G. Pan, J. Alloys Compd. 509 (2011) 675e686. [5] L. Klebanoff, J. Keller, Int. J. Hydrogen Energy 38 (2013) 4533e4576. , J. Power Sources 185 (2008) 566e569. [6] S. Pedneault, J. Huot, L. Roue [7] C. Rongeat, M.H. Grosjean, M. Dehmas, S. Marcotte, et al., J. Power Sources 158 (2006) 747e753. [8] Y.F. Liu, H.G. Pan, M.X. Gao, Q.D. Wang, J. Mater. Chem. 21 (2011) 4743e4755. [9] X.D. Wei, P. Zhang, H. Dong, Y.N. Liu, J.W. Zhu, G. Yu, J. Alloys Compd. 458 (2008) 583e587. [10] C.Y. Wang, D.H. Bradhurst, S.X. Dou, J. Alloys Compd. 285 (1999) 267e271. [11] W.Z. Shen, S.M. Han, Y. Li, J. Yang, K. Xiao, Mater. Chem. Phys. 132 (2012) 852e857. [12] F. Feng, D.O. Northwood, Inter. J. Hydrogen Energy 29 (2004) 955e960. [13] J.L. Luo, N. Cui, J. Alloys Compd. 264 (1998) 299e305. [14] R. Ohara, C.H. Lan, C.S. Hwang, J. Alloys Compd. 580 (2013) 368e372.
[15] W.Z. Shen, S.M. Han, Y. Li, J.Z. Song, Q. Tong, Electrochim. Acta 56 (2010) 959e963. [16] T. Abe, S. Inoue, D. Mu, Y. Hatano, K. Watanabe, J. Alloys Compd. 349 (2003) 279e283. [17] Z.P. Guo, Z.G. Huang, K. Konstantinov, H.K. Liu, S.X. Dou, Inter. J. Hydrogen Energy 31 (2006) 2032e2039. [18] D.C. Wu, L. Li, G.Y. Liang, Y.L. Guo, H.B. Wu, J. Power Sources 189 (2009) 1251e1255. [19] L.Z. Ouyang, Z.J. Cao, H. Wang, J.W. Liu, D.L. Sun, Q.A. Zhang, M. Zhu, J. Alloys Compd. 586 (2014) 113e117. [20] W. Oelerich, T. Klassen, R. Bormann, J. Alloys Compd. 315 (2001) 237e242. [21] C. Iwakura, H. Inoue, S.G. Zhang, S. Nohara, J. Alloys Compd. 270 (1998) 142e144. [22] T. Kohno, M. Yamamoto, M. Kanda, J. Alloys Compd. 293e295 (1999) 643e647. [23] Y.F. Zhu, C. Yang, J.Y. Zhu, L.Q. Li, J. Alloys Compd. 509 (2011) 5309e5314. [24] D.M. Liu, Y.F. Zhu, L.Q. Li, Inter. J. Hydrogen Energy 32 (2007) 2417e2421. [25] T. Kohno, S. Tsuruta, M. Kanda, J. Electrochem. Soc. 143 (1996) L198eL199. [26] S. Nohara, N. Fujita, S.G. Zhang, H. Inoue, C. Iwakura, J. Alloys Compd. 267 (1998) 76e78. [27] Y. Wang, X. Wang, C.M. Li, Inter. J. Hydrogen Energy 35 (2010) 3550e3554. [28] P. Nelli, P. Salvi, M. Villa, G. Zangari, Meet. Abstr. Electrochem. Soc. (2009) P3034. [29] S. Orimo, H. Fujii, Appl. Phys. A 72 (2001) 167e186. [30] Y.F. Zhu, Y.C. Wang, L.Q. Li, Inter. J. Hydrogen Energy 33 (2008) 2965e2969. [31] J.H. Woo, K.S. Lee, J. Electrochem. Soc. 146 (1999) 819e823. [32] C. Iwakura, S. Nohara, S.G. Zhang, H. Inoue, J. Alloys Compd. 285 (1999) 246e249. [33] Y. Zhang, Y.Q. Lei, L.X. Chen, J. Yuan, Z.H. Zhang, Q.D. Wang, J. Alloys Compd. 337 (2002) 296e302. [34] H. Horikiri, A. Kato, A. Inoue, Mat. Sci. Eng. A 179 (1994) 702e706. [35] L.J. Huang, Y.X. Wang, D.C. Wu, J.G. Tang, J. Power Sources 249 (2014) 35e41. € nnebro, J. Am. Chem. Soc. 132 (2010) [36] J. Lu, Y.J. Choi, Z.Z. Fang, H.Y. Sohn, E. Ro 6616e6617. [37] M.X. Gao, S.C. Zhang, H. Miao, Y.F. Liu, H.G. Pan, J. Alloys Compd. 489 (2010) 552e557. [38] C.S. Wang, Y.Q. Lei, Q.D. Wang, Electrochim. Acta 43 (1998) 3209e3216. [39] N. Kuriyama, T. Sakai, H. Miyamura, I. Uehara, H. Ishikawa, T. Iwasaki, J. Electrochem. Soc. 139 (1992) L72eL73. [40] G. Zheng, B.N. Popov, R.E. White, J. Electrochem. Soc. 142 (1995) 2695e2698.
Contents lists available at ScienceDirect
Journal of Power Sources journal homepage: www.elsevier.com/locate/jpowsour
Significantly improved electrochemical hydrogen storage properties of magnesium nickel hydride modified with nano-nickel Wei Chen, Yunfeng Zhu*, Chen Yang, Jiguang Zhang, Menghuai Li, Liquan Li College of Materials Science and Engineering, Nanjing Tech University, 5 Xinmofan Road, Nanjing 210009, PR China
h i g h l i g h t s The HCS product of Mg2NiH4 is modified with nano-nickel via mechanical milling. The discharge capacity and cycle stability are dramatically improved. A novel NiH0.75 phase is observed and the formation mechanism is discussed. Some insights into “electrochemical catalytically” enhanced mechanism are proposed.
a r t i c l e i n f o
a b s t r a c t
Article history: Received 1 November 2014 Received in revised form 11 December 2014 Accepted 14 January 2015 Available online 14 January 2015
Magnesium nickel hydride (Mg2NiH4) used as negative electrode material in nickel-metal hydride (NiMH) secondary battery is modified by nano-nickel via mechanical milling. In this paper, we systematically investigate the microstructure and electrochemical properties of the modified system with different milling durations. X-ray diffraction (XRD) and high resolution transmission electron microscopy (HRTEM) analyses confirm the amorphous transformation of Mg2Ni-based hydride and a novel NiH0.75 nanocrystalline with a diameter of about 5 nm embedding or covering on the surface of the base particle has been observed. Its formation mechanism and positive effects on electrochemical properties of the Mg2NiH4 have also been elaborated. Electrochemical measurements show that the 5 h milled composite possesses markedly increased discharge capacity up to 896 mAh g1. With prolonging the milling duration from 5 h to 40 h, the discharge capacity at the 10th cycle increases from 99 mAh g1 to 359 mAh g1. Besides, the discharging procedure changes from stepwise processes to one single-step process with increasing the milling duration. Tafel polarization test shows that the nano-nickel modified system exhibits a much better anti-corrosion ability during charging/discharging cycles. Meanwhile, both the charge-transfer reaction on the alloy surface and hydrogen diffusion inside the alloy bulk are enhanced with nano-nickel modification. © 2015 Elsevier B.V. All rights reserved.
Keywords: Magnesium-based hydrogen storage alloy Microstructure Electrochemical characteristics Surface modification
1. Introduction Since the discovery of the reversible hydrogen storage alloy in the late 1960s, nickel-metal hydride (Ni/MH) batteries have been developed as one of the most important energy storage carriers thanks to their high energy density and good environmental compatibility with tremendous development in the last few decades. Future Ni/MH batteries are expected to store a large amount of energy without the increase in weight to satisfy the portable equipment and mobile applications [1e4].
* Corresponding author. E-mail address: [email protected] (Y. Zhu). http://dx.doi.org/10.1016/j.jpowsour.2015.01.089 0378-7753/© 2015 Elsevier B.V. All rights reserved.
Among the hydride-forming materials, Mg-based alloys, especially Mg2Ni, exhibit promising properties as negative electrode materials in Ni/MH batteries because of their higher hydrogen storage capacity, lower specific weight and less cost compared with the commercial LaNi5-type alloys. Considering the objectives proposed by the U.S. Department of Energy (DOE) for the future hydrogen storage materials, magnesium and its alloys are expected to meet the target of high hydrogen storage capacity [2,5]. What is undesirable with it is that the lower practical discharge capacity, as well as the inferior cyclic stability stemming from the severe oxidation of Mg in alkaline solution restrains its practical application [6e8]. Meanwhile, the charging/discharging process is a solideliquid reaction, and the reaction rate is affected very much on the solideliquid interface. Hence, it is very necessary to modify the
W. Chen et al. / Journal of Power Sources 280 (2015) 132e140
surface of hydride particles, aiming to protect the base materials and accelerate the electrochemical reaction. In the past several decades, surface modification has been proved to be an effective method to improve the comprehensive electrochemical performance of the Mg2Ni electrode alloys. Besides the traditional surface modification methods, namely acid, alkali and F-treatments which usually affect the surface composition of parent alloy, several groups have reported the improvement in capacity degradation of the alloy electrodes by microencapsulating the alloy particles with various coating layers such as Ni, Cu, Co, polymer, etc via electroless [9e12]. Luo [13] found that surface modification with NieP coatings by electroless plating effectively improves the cycle stability of Mg2Ni-based alloy. Hwang [14] reported that electroless Ni coating can effectively improve the discharge capacity of Mg2Ni alloys. Shen [15] reported that surface modification with polyaniline by electroless deposition improves the high rate discharge ability (HRD) of the AB5-type alloy electrode. However, the electroless coating process is complex due to its special self-catalyzed chemical reactions. Impurity elements like P, S, B, may inevitably be introduced, increasing the charge transfer resistance and hindering the diffusion of H as well. Coating layers introduced by mechanical milling (MM) have been confirmed to be effective in improving the performance of hydride electrodes by preventing corrosion of the disintegrated alloy, bringing about a superior electric catalytic activity. Among the commonly used dopants, including the transition metals, oxides and various carbon sources (graphite, CNTs, carbon black, etc) [8,16e20], Ni exhibits remarkable positive effects in terms of the utilizing of hydrogen absorbability and desorbability of the Mg2Ni-type alloy electrodes, which has been attributed to the outstanding catalytic activity and electrical conductivity of Ni [21,22]. In our previous work, we have reported a novel method of preparing Mg2NiH4 with high activity and discharge capacity, namely hydriding combustion synthesis with subsequent mechanical milling (HCS þ MM) [23]. However, the as-prepared HCS þ MM product still needs investigation due to the severe corrosion of Mg in alkaline solution. In order to improve the electrochemical performance, especially the cycle stability of the HCS þ MM product, a further modification is thus needed. Therefore, in this work, nano-nickel is introduced to modify the Mg2NiH4 via mechanical milling. The influence of milling duration on the structures and electrochemical properties of the hydride is investigated in detail. These results not only lead to an optimized Mg2NiH4/nano-nickel composite system, but also provide some insight into the “electrochemical catalytically” enhanced mechanism of nano-nickel.
133
phase of Mg2NiH4 and the minor phase of Mg2NiH0.3. The HCS product and commercial nano-nickel powder (99.7 wt.% in purity and 50 nm in particle size) were mixed in 1:2 M ratio. Then the powder mixtures were put into the stainless steel vials (volume 100 ml) for mechanical milling. The milling was carried out under argon atmosphere at room temperature using a planetary-type ball mill at a speed of 400 r min1 and ball to powder ratio of 40:1. In order to investigate the influence of milling duration, the samples were milled for 5, 10, 20 and 40 h, respectively.
2.2. Sample characterization All the testing electrodes were prepared as follows: 0.1 g sample powder was mixed with 0.4 g Ni powder, and then cold-pressed into a pellet of 10 mm diameter and about 1 mm thickness under a pressure of 12 MPa. Electrochemical measurements were performed in a three-electrode cell in 6 M KOH at 30 ± 1 C, using sintered Ni(OH)2/NiOOH as the counter electrode and Hg/HgO as the reference electrode. The discharge capacities of electrodes were evaluated by the amount of active substance of Mg2NiH4. The discharge capacity and the cycle life were determined by the galvanostatic method using CT2001A Land battery testing system. The electrodes were discharged at 30 mA g1 to a cut-off potential of 0.6 V (vs. Hg/HgO) and then charged at 300 mA g1 for 2 h after 5 min rest. To investigate the HRD of the alloy electrodes, discharge capacities at different current densities (100, 200, 400 mA g1) were measured at the first cycle. Linear polarization curves were measured at a scanning rate of 0.1 mV s1 from 5 to 5 mV (vs. open circuit potential) at 50% depth of discharge (DOD). Tafel polarization curves were measured at a scanning rate of 1 mV s1 from 300 to 300 mV (vs. open circuit potential) at 100% DOD. Electrochemical impedance spectroscopy (EIS) studies of the electrodes were performed in a frequency range of 100 kHz to 5 mHz with an AC amplitude of 5 mV at 50% DOD under open-circuit conditions, using ZPLOT electrochemical impedance software. For the potentiostatic discharge, the electrodes were discharged at þ600 mV (vs. open circuit potential) potential step and 100% depth of charge (DOC) for 3600 s. The above electrochemical tests were performed at room temperature in the 6 M KOH aqueous solution on a CHI660C electrochemical workstation.
2. Experimental 2.1. Sample preparation The HCS product of Mg2NiH4 was prepared from commercial magnesium powder (99.9 wt.% in purity and <150 mm in particle size) and nickel powder (99.7 wt.% in purity and 2e3 mm in particle size). They were mixed in 2:1 of Mg:Ni molar ratio by an ultrasonic homogenizer in acetone for 1 h. After being completely dried, the well-mixed powder was placed directly into the synthesis reactor without compressive treatment. The hydriding combustion synthesis reactor used in the present work is the same as that reported in our previous work [24]. During HCS process, the mixed powder was heated to 850 K at the rate of 7 K min1 and held for 1 h under 1.8 MPa hydrogen pressure. Subsequently, the sample was cooled down to room temperature under hydrogen atmosphere. The XRD pattern of the as-prepared HCS product of Mg2NiH4 is shown in Fig. S1. It can be seen the HCS product is composed of the main
Fig. 1. XRD patterns of the Mg2NiH4/nano-nickel composite milled for different durations: (a) 5 h, (b) 10 h, (c) 20 h, (d) 40 h.
134
W. Chen et al. / Journal of Power Sources 280 (2015) 132e140
XRD measurements were carried out using a Rigaku SmartLab TM 3 kW diffractometer with Cu Ka radiation. The morphologies and microstructures of the samples were analyzed by an ULTRA 55 field emission scanning electron microscope (FESEM) equipped with an energy dispersive spectrometer (EDS) and HRTEM (JEM2010 UHR). The particle size distribution of the powder samples were examined by laser diffraction particle size analyzer model Malvern Mastersizer 2000.
was further confirmed in the following TEM analysis. The appearance of NiH0.75 pahse is very interesting and has never been reported previously for the Mg2Ni-based composite. To be noted, as the MM was conducted under Ar atmosphere, the H in NiH0.75 phase can only come from the Mg2NiH4. This implies that a new chemical reaction may occur between Mg2NiH4 and nano-nickel during the intensive milling, which can be speculated as follows:
Mg2 NiH4 þ Nano Ni/Mg2 NiHx þ Nano NiH0:75 3. Results and discussion 3.1. Structural properties of the Mg2NiH4/nano-nickel composite Fig. 1 shows the XRD patterns of Mg2NiH4/nano-nickel composite milled for different durations. From Fig. 1a, it can be seen that the diffraction peaks of Mg-based hydride have almost disappeared after only 5 h milling, indicating that the HCSed Mg2NiH4 has transformed into nanocrystalline or amorphous-like structures. Producing amorphous-like structure of the Mg2Ni alloy by ballmilling has been reported by Kohno et al. [25]. Iwakura [26] and Li [27] also noted the Ni powder can enhance the ball-milling efficiency, accelerating the formation of amorphous structure of Mg2Ni. With prolonging the milling duration, the peak intensity of nano-nickel decreases apparently and almost disappears after 40 h milling, while some new peaks appear at around 42.1, 48.9 and 71.8 after milling for 20 h. These peaks become sharper with increasing the milling duration to 40 h. By checking the standard spectrum, we attributed these peaks to NiH0.75 (84-0450), which
(1)
The possible reaction mechanism is discussed as following. The MM is accepted as a powerful method involving recurrent grinding and cold welding, which can lead to massive atomic diffusion among the powders. According to our previous work, part of the H atoms may deintercalate from the Mg2NiH4 lattice during the MM process. Besides, the nano-nickel possesses much higher reaction activity because of its smaller particle size and larger surface area. Thus, it is possible for the nano-nickel to interact with the H to form the new NiH0.75 phase. Fig. 2 shows the SEM images of the Mg2NiH4 milled with nanonickel for different durations. Fig. 2a shows that the 5 h milled powders are irregular in shape and attached with small flake particles which may be due to the agglomerate nano-nickel. It can also be seen that the exposed surface of Mg2NiH4 particles is smooth, indicating the coating of nano-nickel is incomplete. The 10 h milled sample presents a similar morphology to the 5 h milled sample, also showing the inhomogeneous coating of nano-nickel on the surface of the Mg2NiH4 particles. However, with further increasing the milling duration to 20 h and 40 h as shown in Fig. 2ced, we can
Fig. 2. SEM micrographs of the Mg2NiH4/nano-nickel composite milled for different durations: (a) 5 h, (b) 10 h, (c) 20 h, (d) 40 h.
W. Chen et al. / Journal of Power Sources 280 (2015) 132e140
observe that the surface of Mg2NiH4 particles becomes very rough, and almost no smooth surface can be observed. Thus, we can conclude that the nano-nickel coating is well-proportioned for the samples with long milling duration, indicating that long milling duration is in favor of the formation of homogeneous nano-nickel protective layer. To further confirm the homogeneity of nano-nickel coating layer, EDS analysis has also been conducted and the results are shown in Fig. S2. The results show that all the samples contain the elements of Mg, Ni, O and C. The oxygen may come from the oxidation of samples when exposed to air, and the carbon comes from the conductive adhesive. For each sample, different points on the particle surface were analyzed by EDS. The Mg: Ni (atomic ratios) for all the samples are presented in the figures to evaluate the homogeneity of nano-nickel coating. Before coating, the EDS analysis shows that the Mg: Ni ratios are all close to 2 at different points, and one of the spectra is shown in Fig. S2a, which agrees well with the Mg: Ni ratio of Mg2NiH4. For the 5 h milled Mg2NiH4 þ 2 nano-nickel composite, the EDS analysis shows that the Mg: Ni ratio is always deviated from the designed value of 0.667, and one of the spetra is shown in Fig. S2b. The Mg: Ni ratio of 0.923 is larger than 0.667, indicating that the nano-nickel coating is incomplete. The similar phenomenon is also found for the 10 h milled composite, in which the Mg: Ni ratio is 0.821 as shown in Fig. S2c, also larger than 0.667. However, with increasing the milling duration to 20 h and 40 h, the EDS analyses show that the
135
Mg: Ni ratios at different points are all close to the designed value of 0.667, which can be seen in Fig. S2d and Fig. S2e. The results further confirm that the nano-nickel coating is well-proportioned on the Mg2NiH4 particle surface with longer milling durations. Fig. S3 shows the particle size distribution of the 5 h milled bare Mg2NiH4 and Mg2NiH4/nano-nickel composite milled for different durations. The 5 h milled bare Mg2NiH4 shows an average particle size of 4.37 mm. When nano-nickel was introduced, the average particle size decreases dramatically to 1.42 mm. With increasing the milling duration, however, the average particle size increases. The average particle sizes are 1.63 mm, 11.4 mm and 15.4 mm for the 10 h, 20 h and 40 milled composites, respectively. The increasing in average particle size of the samples with longer milling durations is attributed to the agglomeration of the particles, especially for the 20 h and 40 h milled samples. TEM analysis was used to investigate the microstructure of the sample milled for 40 h. It can be seen from Fig. 3a that nanocrystals with a diameter of about 5 nm are embedded or covered on the surface of the base particle. HRTEM micrograph (Fig. 3b) further confirms that Mg2NiH4 has transformed to amorphous state after 40 h milling. The lattice fringes of inlaid nanocrystals with a separation of 0.1836 nm agree well with the (200) interplanar spacing of NiH0.75. As discussed above, the formation of nanocrystalline NiH0.75 could be attributed to the high active nano-nickel, which catches the desorbed H from Mg2NiH4 and forms the NiH0.75. As the
Fig. 3. TEM images of 40 h milled Mg2NiH4/nano-nickel composite: (a, b) bright field image and HRTEM image for as milled sample; (c) HRTEM image for discharged sample; (d) HRTEM image for recharged sample. The insets give the corresponding selected area electron diffraction patterns.
136
W. Chen et al. / Journal of Power Sources 280 (2015) 132e140
desorbed H atoms mainly gather around the Mg2NiH4/nano-nickel interface, the formation of NiH0.75 is influenced by the concentration of H at the two-phase interface and governed by a slow process of H diffusion in the bulk crystal [28,29]. The NiH0.75 phase was not observed in the 5 h and 10 h milled samples, indicating long milling duration is necessary for both diffusion and aggregation of desorbed hydrogen. Furthermore, the inset of Fig. 3b shows the corresponding selected area electron diffraction (SAED) pattern, which displays a distinguishable halo and ring-like feature of nanocrystalline and amorphous structures. Fig. 3c shows the HRTEM micrograph and corresponding SAED pattern of 40 h milled sample after discharging, revealing that the discharging process wouldn't cause crystallization of the amorphous Mg2NiH4. From Fig. 3c, we can also observe that the nanocrystals distribute on the surface of the bulk materials and the lattice fringes with a separation of 0.2107 nm agree well with the (111) interplanar spacing of NiH0.75. The recharged sample is shown in Fig. 3d and the corresponding SAED pattern is inserted. No obvious structure change is observed, which confirms that the charging process wouldn't cause crystallization of the amorphous Mg2NiH4 either. Moreover, the NiH0.75 couldn't absorb and desorb hydrogen reversibly during the cycles, which is further supported by Fig. S4. No diffraction peaks of Mg2Ni-based alloy/hydride are detected after cycling, which proves again that charging/discharging doesn't cause crystallization of the amorphous Mg2NiH4. From Fig. S4, the NiH0.75 phase is observed in all the samples, illustrating that the NiH0.75 phase doesn't participate in the charging/discharging process.
Table 1 Electrochemical properties of samples milled for different durations. Samples Cmax (mAh g1)
C10 (mAh g1)
Ecorr (V)
5h 10 h 20 h 40 h
99 146 265 359
0.839 80.9 0.838 86.7 0.833 100.5 0.828 91.6
896 996 968 946
I0 Rct (mA g1) (mU) 197 191 183 187
D/a2 ( 105 s1) 1.91 2.17 2.48 2.87
The effect of milling duration on the discharge behavior of different samples is shown in Fig. 4. Note that the electrodes were discharged directly after milling without charging and the maximum discharge capacities were found in the first cycle. It can be seen that the initial discharge capacity increases firstly with prolonging the milling duration and reaches the maximum value of 996 mAh g1 after 10 h milling. However, if the milling duration is prolonged continuously, the capacity decreases slightly, which may be caused by the desorption of hydrogen during milling. The maximum discharge capacities of samples are listed in Table 1. The Mg2NiH4 milled only for 5 h with nano-nickel exhibits a capacity of
896 mAh g1, which is much higher than 480 mAh g1 for the 5 h milled bare Mg2NiH4 [30]. We confirm that the amorphous structure is a key factor to achieve high discharge capacity [31]. Besides, nano-nickel plays an important role in increasing the discharge capacity. First of all, the nano-nickel can accelerate amorphization of Mg2NiH4 dramatically. Through MM, nano-nickel will well inlay and disperse on the hydride particle surface, forming a coating layer. This protective layer can impede the formation of nonconductive Mg(OH)2 layer which is accepted as a main reason for the capacity loss. Meanwhile, the coated nano-nickel is helpful in decreasing the electrochemical reaction resistance on the alloy surface and acts as current collector because of its good electrocatalytic activity in alkaline electrolyte [32]. Considering the superior discharge property, we believe the new NiH0.75 phase may also act as an electro-catalyst during the charging/discharging process. Overall, nano-nickel coating can simultaneously improve the electrochemical kinetics and protect the active substances from corrosion. Consequently, synergistic effects of the nano-nickel layer and the MM process result in the high capacity of the HCSed Mg2NiH4. Furthermore, as seen in Fig. 4, both the discharge curves of samples milled for 5 h and 10 h consist of two distinct discharge plateaus. This indicates that two different hydrogen desorption phases exist in the as-milled composite, corresponding to two independent desorption steps of hydrogen. In other words, there are two different discharging steps during the discharging process. We infer that the first discharging voltage plateau located at about 0.95 V could be attributed to the hydrogen stored in the nanocrystalline Mg2NiH4 and the free H atoms existing in the crystal defects and grain boundaries introduced by MM. Because the nanocrystalline hydride is easy to be corroded by the KOH electrolyte and the free H atoms do not have a definite location within the lattice, we concluded the first discharging step was
Fig. 4. Discharge curves of samples milled for different durations (Discharge rate: 30 mA g1; the first cycle).
Fig. 5. Discharge capacities as a function of cycle number for samples milled for different durations.
3.2. Electrochemical properties of the Mg2NiH4/nano-nickel composite
W. Chen et al. / Journal of Power Sources 280 (2015) 132e140
irreversible. As the second sloping plateau contributes most of the capacity, it should correspond to the main phase of the amorphous hydride. As the milling duration increases to 20 h and 40 h, it is obvious that the first discharge plateau becomes narrower and the second discharge plateau becomes longer, indicating the discharging procedure changes from stepwise processes to one singlestep process. We believe the slight decrease in discharge capacity is caused by the formation of new NiH0.75 phase, which cannot desorb hydrogen electrochemically. However, the amorphization level of Mg2NiH4 should increase with prolonging the milling duration, which is favorable to the electrochemical properties. Fig. 5 shows the discharge capacity degradation as a function of the cycle number of samples milled for different durations. It is obvious that with increasing the milling duration the cycle stability of samples is significantly improved. The discharge capacities at the 10th cycle are listed in Table 1. As can be seen in Fig. 4, the initial discharge capacities of different samples show small differences and they all reach their maximum discharge capacities at the first cycle without precharging, indicating that all the milled samples have a good activation property. Unfortunately, a dramatic drop of the discharge capacity is observed at the second cycle. The severe corrosion of Mg in the alkaline solution is usually thought as the main reason for capacity loss [33]. We discovered that the irreversibility of the free hydrogen and the poor anticorrosion ability of nanocrystalline hydride mainly caused the huge loss of discharge capacity in the early stage of cycles. It can also be seen that the capacities for the samples milled for 5 h and 40 h at the second cycle are 215 mAh g1 and 538 mAh g1, respectively, revealing increasing the milling duration can slow down the rate of degradation significantly. According to the above results, the amorphization level of the sample increases with the increase of milling duration. Unlike crystalline alloys, the amorphous alloys usually have good resistance to corrosion and pulverization because of the absence of any conventional grains and grain boundaries and continuous volume change when H atoms get in or out of the alloy particles. Similar results have also been reported by Horikiri [34] and Huang [35]. Moreover, with increasing the milling duration, a more complete and homogeneous nano-nickel coating is produced on the alloy surface, which is effective in protecting the active substance from oxidation failure and maintaining the electrocatalytic activity of the alloy, improving the charge/discharge cycle performance. Fig. 6 is proposed to give a further understanding of the charging/discharging process, which shows a schematic representation of electrochemical hydrogen ab/desorption. During
137
Fig. 7. Tafel polarization curves of samples milled for different durations (scan rate: 1 mV s1).
discharging, as illustrated in Fig. 6a, hydrogen atoms are desorbed first from the amorphous Mg2NiH4 and then diffuse to the twophase (Mg2NiH4 and nano-nickel) interface. The surface of nanonickel is where the oxidation of H mainly occurs in our system because of their high electrocatalytic activity. With increasing the milling duration, more nano-nickel will inlay and coat on the Mg2NiH4 surface, which means more two-phase interfaces and active areas. Thus, more efficient hydrogen-transfer reactions (H/Hþþe), rather than the generation of H2 (H þ H/H2), will be observed. Meanwhile, the milling process makes the free hydrogen transform into the NiH0.75 phase. The distribution of the in-situ synthesized NiH0.75 with nanosize can act as the catalyst, which not only brings large contacts of catalytic active sites with Mg2NiH4, but also has a great advantage of the fast electron transfer and the diffusion of hydrogen atoms. Furthermore, the positive effect of NiH0.75 can also be understood by viewing NiH0.75 as a grain growth inhibitor [36] that prevents the crystallization of amorphous Mg2NiH4 during the charge/discharge cycles. However, further indepth research is needed to fully understand the effects of the nanocrystalline NiH0.75. The surface oxide layer on the hydride particle is believed to prohibit both the hydrogen electrode reaction (HER) on the interface of electrode/electrolyte and the hydrogen diffusion. Therefore, even after the alloy powders have been cracked due to the
Fig. 6. The charging/discharging model of nano-nickel coated Mg2Ni-based hydride: (a) The discharging mechanism after nano-nickel coating and nano-nickel coating diagram in top left corner; (b) The charging/discharging diagram after disintegration.
138
W. Chen et al. / Journal of Power Sources 280 (2015) 132e140
expansion and contraction of the particle during the charging/ discharging process [37], the new exposed surface exerts little influence on the electrode performance as the fresh surface will be corroded very quickly. However, the nano-nickel coating layer can provide both the active sites for the redox reaction of hydrogen and the pathway for the hydrogen diffusion [13], as shown in Fig. 6b. All of these factors result in a prominent improvement of the discharge capacity by about 300 mAh g1 in each cycle with increasing the milling duration from 5 h to 40 h. To investigate the anti-corrosion ability of the alloy electrodes, Tafel polarization test was performed. The polarization curves for different samples are shown in Fig. 7 and the corrosion potential Ecorr are listed in Table 1. The Ecorr value shifts toward the positive direction with increasing the milling duration, suggesting favorable effect of the milling treatment on the anti-corrosion ability of the alloy electrode owing to the formation of the nano-nickel coating layer and the more complete amorphization. To further understand the anti-corrosion mechanism, a pure Mg2NiH4 sample milled for 5 h was chosen for comparison. Tafel polarization curves of the 5 h milled bare Mg2NiH4 and 40 h milled Mg2NiH4/nano-nickel composite at different cycles are shown in Fig. S5. As we can see, the Ecorr of the 5 h milled bare Mg2NiH4 shifts toward the negative direction severely with increasing the cycle number, indicating the loose Mg(OH)2 layer has no protective effect. Instead, it leads to the exposure of more fresh surface to the electrolyte and thus severer corrosion occurs. At the 10th cycle, the outer layer of Mg2NiH4 particle was almost corroded completely, forming a thick Mg(OH)2 layer. Therefore, it became difficult for alkaline solution to permeate through, resulting in the positive shift of the Ecorr. Fortunately, the Mg2NiH4/nano-nickel composite after 40 h milling exhibits an acceptable Ecorr value and its variation is mild with the increasing cycle number, which obviously reveals that it has a better anti-corrosion ability than the 5 h milled Mg2NiH4 sample during cycles. The HRD of the samples milled for different durations is shown in Fig. 8. The value of HRD is calculated according to the following equation:
HRD ¼
Cd 100% Cd þ C60
(2)
where Cd is the discharge capacity with cut-off potential of 0.6 V (vs. Hg/HgO) at the discharge current density Id (100, 200, 400 mA g1), C60 is the residual discharge capacity with the same
Fig. 8. High rate discharge ability (HRD) of the samples milled for different durations.
cut-off potential at the discharge current density I (60 mA g1) after the alloy electrode has been fully discharged at the discharge current density Id. It can be found that the HRD gets improved with increasing the milling duration, and the 20 h milled sample exhibits the highest HRD, suggesting that 20 h milled sample has the best electrochemical kinetics. Generally speaking, the HRD of the hydrogen storage electrode alloy is determined by charge-transfer on the alloy surface and hydrogen diffusion in the alloy bulk [38]. It is known that the transition-metal Ni has the unpaired electrons and unoccupied atomic orbitals, which can interact with the base metal and then a synergistic effect could exist between the Mg2NiH4 and nano-nickel. Thus, the charge-transfer reaction will be accelerated as the hydrogen atom diffuses near the Ni atom. Furthermore, the nano-nickel coating layer can block the formation of Mg(OH)2 and the two-phase interface of the nano-nickel and the bulk alloy can provide a diffusion channel for the hydrogen. To prove this, linear polarization, EIS and potentiostatic discharge measurements were performed. The exchange current density (I0) is used to characterize the electrocatalytic activity for the charge-transfer reaction on the surface of the alloy electrodes. Fig. 9 presents the linear polarization curves for the samples milled for different durations measured at 50% DOD. The exchange current density can be estimated from the slope of the obtained micropolarization curve using the following equation:
I0 ¼
RTI Fh
(3)
in which R, T, I, F andhare the gas constant, the absolute temperature, the applied current density, the Faraday constant and the total overpotential, respectively. The results obtained are also listed in Table 1. A high value of I0 corresponds to a good kinetics of chargetransfer on the alloy surface. It can be found the I0 value increases first and then decreases with increasing the milling duration. The 20 h milled sample possesses the highest I0 value of 100.5 mA g1. Linear polarization curves of the 5 h milled bare Mg2NiH4 and 40 h milled Mg2NiH4/nano-nickel composite at different cycles are shown in Fig. S6 and the calculated I0 values are also tabulated in Table S1. It is obvious that with cycling the 40 h milled Mg2NiH4/ nano-nickel composite exhibits much bigger I0 values, indicating that MM and nano-nickel coating can improve the charge-transfer reaction dramatically.
Fig. 9. Linear polarization curves for samples milled for different durations (scan rate: 0.1 mV s1).
W. Chen et al. / Journal of Power Sources 280 (2015) 132e140
The EIS and corresponding equivalent circuit of samples milled for different durations are shown in Fig. 10. It can be seen that each spectrum is composed of two semicircles followed by a straight line at the low-frequency region. It has been reported by Kuriyama et al. [39] that the semicircles in the high-frequency region and lowfrequency region correspond to the contact resistance (Rc) and the charge-transfer resistance (Rct), respectively, and the sloped straight line corresponds to Warburg impendence (W). Based on the equivalent circuit and by means of non-linear least squares (NLLS) fitting of the plots, the charge-transfer resistances of the samples were calculated and listed in Table 1. As we can see, the Rct decreases first and then increases with increasing the milling duration. The 20 h milled sample possesses the smallest Rct value of 187 mU. This variation of the Rct is consistent with the result obtained from the linear polarization. Fig. S7 shows the EIS of the 5 h milled bare Mg2NiH4 and 40 h milled Mg2NiH4/nano-nickel composite at different cycles. The charge-transfer resistances of samples at different cycles were also calculated and listed in Table S1. To investigate the hydrogen diffusion behavior of the alloy, potentiostatic discharge measurements of the hydride electrodes were performed and the results are shown in Fig. 11. It can be seen that at the initial stage of discharging (<500 s), the current declines sharply, and after a time response of about 2000 s the current decreases slowly in a linear fashion. In the linear region, the hydrogen diffusion in the bulk of the alloy dominates the electrochemical reaction process and the hydrogen diffusion coefficient parameter D/a2 can be calculated according to the following equation [40]:
6FD p2 D log i ¼ log ± 2 ðC0 Cs Þ t 2:303 a2 da
(4)
where i (A g1) is the current density, F is Faraday constant, D (cm2 s1) is the hydrogen diffusion coefficient, d (g cm3) is the density of the hydrogen storage alloy, a (cm) is the alloy particle radius, C0 (mol cm3) is the initial hydrogen concentration in the bulk of the alloy, Cs (mol cm3) is the hydrogen concentration on the surface of the alloy particles, and t (s) is the discharge time. As the average particle size is not identical for the different samples, it is reasonable to use D/a2 to evaluate the discharge kinetics of the electrode contributed by hydrogen diffusion within the bulk of the particles. The values of D/a2 were determined from the slopes of linear current responses and listed in Table 1. It is found that the D/
Fig. 10. Electrochemical impedance spectra and corresponding equivalent circuit for samples milled for different durations measured at 50% DOD.
139
Fig. 11. Potentiostatic discharge curves for samples milled for different durations measured at 100% DOC.
a2 value increases continuously with increasing the milling duration. A longer milling duration can not only increase the amorphous level, but also introduce more defects which facilitate the hydrogen diffusion procedure. In the present work, the nanocrystalline NiH0.75 can be regarded as the activated channel for hydrogen transferring, which accelerates the hydrogen diffusivity in the bulk. To investigate the hydrogen diffusion behavior with cycling, potentiostatic discharge measurements of 5 h milled bare Mg2NiH4 and 40 h milled Mg2NiH4/nano-nickel composite at different cycles were performed and the results are shown in Fig. S8 and the calculated D/a2 values are listed in Table S2. The 40 h milled Mg2NiH4/nano-nickel composite exhibits much higher D/a2 value at each cycle. 4. Conclusions In order to improve the electrochemical properties of Mg2Nibased hydrides, nano-nickel is used to modify the HCSed Mg2NiH4 via mechanical milling. Effect of different milling durations on the structures and electrode performances is investigated in detail. XRD results reveal the amorphization of Mg2NiH4 and especially, a novel phase NiH0.75 appears after 20 h milling. HRTEM observation further confirms the NiH0.75 nanocrystals with a diameter of about 5 nm embedded in the amorphous bulk. We infer the formation of NiH0.75 could be controlled by a slow process of H diffusion between the hydride and the nano-nickel during the MM process. SEM and EDS analyses show that the nano-nickel coating becomes well-proportioned with increasing the milling duration to 20 h and 40 h. Electrochemical measurements indicate that both the discharge capacity and the cycle stability are improved dramatically as the milling duration increases. The maximum discharge capacity of the composite reaches 996 mAh g1 which is very close to its theoretical value. With prolonging the milling duration from 5 h to 40 h, the discharge capacity at the 10th cycle increases from 99 mAh g1 to 359 mAh g1. We also find the discharging procedure changes from stepwise processes to one single-step process with increasing the milling duration. Tafel polarization test shows that the nano-nickel modified system exhibits a much better anticorrosion ability during cycles. Meanwhile, both the chargetransfer reaction and hydrogen diffusion are enhanced with nano-nickel modification. Taken together, the 20 h milled system exhibits the best electrochemical kinetics, which is mainly controlled by the charge-transfer reaction on the alloy surface.
140
W. Chen et al. / Journal of Power Sources 280 (2015) 132e140
Acknowledgments This work was supported by the National Natural Science Foundation of China (51471087, 51171079), Natural Science Foundation of the Jiangsu Higher Education Institutions of China (13KJA430003), Innovation Foundation for Graduate Students of Jiangsu Province (KYLX_0741), Qing Lan Project and the Priority Academic Program Development (PAPD) of Jiangsu Higher Education Institutions. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.jpowsour.2015.01.089. References [1] S.R. Ovshinsky, M.A. Fetcenko, J. Ross, Science 260 (1993) 176e181. [2] L. Schlapbach, A. Züttel, Nature 414 (2001) 353e358. [3] K. Young, D.F. Wong, S. Yasuoka, J. Ishida, J. Nei, J. Koch, J. Power Sources 251 (2014) 170e177. [4] Y.F. Liu, Y.H. Cao, L. Huang, M.X. Gao, H.G. Pan, J. Alloys Compd. 509 (2011) 675e686. [5] L. Klebanoff, J. Keller, Int. J. Hydrogen Energy 38 (2013) 4533e4576. , J. Power Sources 185 (2008) 566e569. [6] S. Pedneault, J. Huot, L. Roue [7] C. Rongeat, M.H. Grosjean, M. Dehmas, S. Marcotte, et al., J. Power Sources 158 (2006) 747e753. [8] Y.F. Liu, H.G. Pan, M.X. Gao, Q.D. Wang, J. Mater. Chem. 21 (2011) 4743e4755. [9] X.D. Wei, P. Zhang, H. Dong, Y.N. Liu, J.W. Zhu, G. Yu, J. Alloys Compd. 458 (2008) 583e587. [10] C.Y. Wang, D.H. Bradhurst, S.X. Dou, J. Alloys Compd. 285 (1999) 267e271. [11] W.Z. Shen, S.M. Han, Y. Li, J. Yang, K. Xiao, Mater. Chem. Phys. 132 (2012) 852e857. [12] F. Feng, D.O. Northwood, Inter. J. Hydrogen Energy 29 (2004) 955e960. [13] J.L. Luo, N. Cui, J. Alloys Compd. 264 (1998) 299e305. [14] R. Ohara, C.H. Lan, C.S. Hwang, J. Alloys Compd. 580 (2013) 368e372.
[15] W.Z. Shen, S.M. Han, Y. Li, J.Z. Song, Q. Tong, Electrochim. Acta 56 (2010) 959e963. [16] T. Abe, S. Inoue, D. Mu, Y. Hatano, K. Watanabe, J. Alloys Compd. 349 (2003) 279e283. [17] Z.P. Guo, Z.G. Huang, K. Konstantinov, H.K. Liu, S.X. Dou, Inter. J. Hydrogen Energy 31 (2006) 2032e2039. [18] D.C. Wu, L. Li, G.Y. Liang, Y.L. Guo, H.B. Wu, J. Power Sources 189 (2009) 1251e1255. [19] L.Z. Ouyang, Z.J. Cao, H. Wang, J.W. Liu, D.L. Sun, Q.A. Zhang, M. Zhu, J. Alloys Compd. 586 (2014) 113e117. [20] W. Oelerich, T. Klassen, R. Bormann, J. Alloys Compd. 315 (2001) 237e242. [21] C. Iwakura, H. Inoue, S.G. Zhang, S. Nohara, J. Alloys Compd. 270 (1998) 142e144. [22] T. Kohno, M. Yamamoto, M. Kanda, J. Alloys Compd. 293e295 (1999) 643e647. [23] Y.F. Zhu, C. Yang, J.Y. Zhu, L.Q. Li, J. Alloys Compd. 509 (2011) 5309e5314. [24] D.M. Liu, Y.F. Zhu, L.Q. Li, Inter. J. Hydrogen Energy 32 (2007) 2417e2421. [25] T. Kohno, S. Tsuruta, M. Kanda, J. Electrochem. Soc. 143 (1996) L198eL199. [26] S. Nohara, N. Fujita, S.G. Zhang, H. Inoue, C. Iwakura, J. Alloys Compd. 267 (1998) 76e78. [27] Y. Wang, X. Wang, C.M. Li, Inter. J. Hydrogen Energy 35 (2010) 3550e3554. [28] P. Nelli, P. Salvi, M. Villa, G. Zangari, Meet. Abstr. Electrochem. Soc. (2009) P3034. [29] S. Orimo, H. Fujii, Appl. Phys. A 72 (2001) 167e186. [30] Y.F. Zhu, Y.C. Wang, L.Q. Li, Inter. J. Hydrogen Energy 33 (2008) 2965e2969. [31] J.H. Woo, K.S. Lee, J. Electrochem. Soc. 146 (1999) 819e823. [32] C. Iwakura, S. Nohara, S.G. Zhang, H. Inoue, J. Alloys Compd. 285 (1999) 246e249. [33] Y. Zhang, Y.Q. Lei, L.X. Chen, J. Yuan, Z.H. Zhang, Q.D. Wang, J. Alloys Compd. 337 (2002) 296e302. [34] H. Horikiri, A. Kato, A. Inoue, Mat. Sci. Eng. A 179 (1994) 702e706. [35] L.J. Huang, Y.X. Wang, D.C. Wu, J.G. Tang, J. Power Sources 249 (2014) 35e41. € nnebro, J. Am. Chem. Soc. 132 (2010) [36] J. Lu, Y.J. Choi, Z.Z. Fang, H.Y. Sohn, E. Ro 6616e6617. [37] M.X. Gao, S.C. Zhang, H. Miao, Y.F. Liu, H.G. Pan, J. Alloys Compd. 489 (2010) 552e557. [38] C.S. Wang, Y.Q. Lei, Q.D. Wang, Electrochim. Acta 43 (1998) 3209e3216. [39] N. Kuriyama, T. Sakai, H. Miyamura, I. Uehara, H. Ishikawa, T. Iwasaki, J. Electrochem. Soc. 139 (1992) L72eL73. [40] G. Zheng, B.N. Popov, R.E. White, J. Electrochem. Soc. 142 (1995) 2695e2698.