Molybdenum Oxide as Cathode for High Voltage Rechargeable Aluminum Ion Battery

A dense molybdenum oxide layer was fabricated on nickel foam (MoO 2 @Ni) and used as the cathode in the 1-ethly-3-methylimidazolium chloride/AlCl 3 ionic liquid electrolyte aluminum ion battery. Study on the electrochemical performance demonstrated that the cathode could not only exhibit a discharge potential of 1.9 V which is higher than most of the studied metal oxide cathodes of Al ion battery, but also deliver a speciﬁc discharge capacity of 90 mAh g − 1 at a constant current density of 100 mA g − 1 . Although the MoO 2 was dissolved and transferred to the separator after long cycling, resulting in a rapid capacity decay, we still believe it is an encouraging outcome in terms of these type batteries.

Aluminum (Al) is the most abundant metal element in the earth's crust and has the highest theoretical volumetric capacity of 8.04 mAh cm −3 among the metals, which makes it to be a hot candidate electrode material for modern society. However, it is difficult to fully use the high power densities of Al batteries due to some intrinsic restrictions such as Al passivation and harsh working condition. 1,2 Recently, the rechargeable Al-ion battery has regained intensive attention because some ionic liquids for Al electro-deposition/stripping are used as the effective room temperature electrolytes. For example, 1-ethyl-3-methylimidazolium chloride (EMImCl)/AlCl 3 eutectic salt is successfully used in Al-ion battery, Al-S and Al-O 2 systems. [3][4][5][6][7][8][9][10][11][12][13][14][15][16] There is no doubt that a better battery performance is strongly dependent on the properties of electrode materials in terms of charge efficiency, cycling stability, discharge voltage and specific capacity. Unfortunately, a low discharge voltage character (usually less than 1 V) is widely exhibited in most of the investigated Al battery systems in spite of the mezzo standard reduction potential of the Al (−1.68 V vs SHE). [4][5][6][7] Thus it will be a significant drag on such fields as electrical vehicles, where high power supply is required. Although some carboneaous materials exhibit high voltage and long cycle life, their limited specific discharge capacity remains a big challenge to meet the market needs. Therefore, it is worth exploring new cathode materials to satisfy the future application of rechargeable Al batteries.
Molybdenum oxide (MoO 2 ) has long been investigated as a host material for lithium storage due to its high electrochemical activity and natural abundancy. 8,9,12 So far, various approaches have been adopted to fabricate novel structured MoO 2 materials. Herein, we report a dense MoO 2 layer fabricated by simple magnetron sputtering and heat-treatment. The electrochemical performance of the MoO 2 layer in Al ion battery was also presented for the first time.

Experimental
Materials preparation.-A high purity molybdenum target (China Material Tech. Co., 99.99% purity, China) was used as molybdenum source for deposition. Firstly, the target was dipped into 10 wt% phosphoric acid for 30 min to remove the surface oxides. Then it was ultrasonically cleaned in acetone, ethanol and de-ionized water for 15 minutes, respectively, and dried under nitrogen flux. A nickel foam disc ( 10 mm) was subjected to the same pretreatment as the molybdenum target and used as the substrate for molybdenum deposition. The deposition was performed by a radio frequency (RF) magnetron sputtering system (KYKY Co., China) at 3 W cm −2 power for 1 h at room temperature. After deposition, the nickel foam with Mo layer on its surface was placed into a horizontal tube furnace (Lindberg Blue M, TF55030C, USA). First, a high purity argon (99.999%) was z E-mail: demin.chen.1@imr.ac.cn introduced into the tube and the sample was heated to 400 • C at a ramping rate of 10 • C min −1 . After the target temperature was achieved, a purity oxygen flow was introduced at a flow rate of 2 mL min −1 and maintained for 1∼2 min. Finally, the furnace was cooled to the room temperature naturally. The net mass increment of the foam was about 1.5 mg, and the loading of active material was about 1.91 mg cm −2 (calculated on the geometry area of nickel substrate). A high purity Mo foil disk (99.99%, 10 mm) was also treated by the same procedure and used as the reference sample.
Battery assembly.-A Swagelok type cell ( 16 mm) was used to employ the electrochemical measurements in this work. Polytetrafluoroethylene (PTFE, Yangzhong Gongzheng Fluid Controls Co., China) was used as the battery shell. Two high purity (99.9%) molybdenum rods (Hebei Qingyuan Metal Materials Co., China) were used as current collectors. A rapid filtering paper (Toyo Roshi Kaisha Ltd., Japan) and a high purity electro-polished aluminum foil (99.999%, IMR, CAS, China) were used as the separator and the anode, respectively. All the assembling processes were performed in an Ar gas filled glove box where the oxygen and moisture were under 1 ppm.
Characterization.-The surface morphology of the sample was observed by scanning electron microscope (SEM, FEI INSPECT F50, USA). The composition of the sample was analyzed by energydispersive X-ray spectroscopy (EDX, X-Max Oxford, equipped with the SEM, UK). The crystal structure of the samples was detected by Xray diffraction measurement (XRD, Rigaku D/Max 2500pc with Cu Kα radiation, λ = 1.5406Å, Japan). The galvanostatic charge/discharge (GCD) and cycling measurements were performed on an Arbin BT-2043 battery test system at a current density of 100 mA g −1 . Cyclic voltammetry (CV) measurement was performed on an electrochemical working station (Zahner Zennium, Germany). Al was used as the counter and reference electrode. The scanning rate was 1 mV s −1 and the scanning voltage ranged from 0.3 to 2.5 V.

Results and Discussion
Fig . 1 shows the SEM images of the surface morphology of the pristine nickel foam and the foam after sputtering and heat-treatment. It can be seen clearly in Fig. 1b that a dense layer was formed. Fig. 1c shows the EDX spectra of the heat treated sample, and it can be found that other than those of Mo, Ni and O, no else element was detected. The XRD patterns of the foam after heat-treatment are presented in Fig. 1d  composition of the sputtered layer can be determined to MoO 2 . The composite is designated as MoO 2 @Ni for brevity. Fig. 2a shows the charge/discharge curves of the MoO 2 @Ni at a constant current density of 100 mA g −1 for the first, fifth and tenth cycles, respectively. Two discharge potential plateaus at about 1.95 V and 1.0 V can be seen clearly. The first charge/discharge capacities of the MoO 2 @Ni were about 253 mAh g −1 and 90 mAh g −1 , respectively. This means a large irreversible capacity loss occurred in the first cycle. The specific charge capacity decreased with the increase of cycle while the discharge specific capacity did not changed dras-  tically in the initial 10 cycles. The 100-cycle stability performance is presented in Fig. 2b. As can be seen, the specific charge/discharge capacities were about 75 mAh g −1 for the 5th to 20th cycles. However, both of them decreased with the charge/discharge proceeding. There was only 25 mAh g −1 charge/discharge specific capacity left after 100 cycles, indicating a capacity decay. In addition, a low coulombic efficiency of 70% in average was kept through the entire cycles. Fig. 2c shows the CV curves of the MoO 2 @Ni electrodes. For the first ten CV cycles, a clear pair of redox peaks at around 2.15 V and 1.91 V can be found invariably, which is in good agreement with the charge/discharge results. The asynchronous change of the CV peak intensity and discharge capacity may be ascribed to the difference of the activation degree during a given cycle between the batteries used for CV and GDC test, resulting from the difference of the chargedischarge regimes.
It has been pointed out by Reed and Nakayama et al., that the ionic liquid is a highly corrosive electrolyte. 10,13 In order to confirm that the high discharge voltage is originated from the MoO 2 and exclude the corrosion effect from the electrolyte and side reactions between the electrolyte and the cell components, we carried out two group of parallel experiments, respectively. In one group, a series CV measurements were performed on four type of batteries where the cathodes were nickel substrate, nickel substrate annealed at 400 • C, nickel substrate sputtered with Mo and blank electrolyte, respectively. The results are shown in Fig. 2d, no redox peaks can be found in all the CV curves and the electrolyte is stable at 1.9 V-2.2 V region, where the oxidation/reduction of MoO 2 @Ni electrode occurred. Therefore, the contribution to the high discharge voltage from other materials but MoO 2 can be excluded. In the other group, a high purity Mo disk foil was annealed to fabricate a dense MoO 2 film under the same treatment as the MoO 2 @Ni sample, which has been described in the Experimental section. The Mo foil covered with MoO 2 film (referred as MoO 2 @Mo for brevity) was assembled into a battery to study its electrochemical performance. The material characterization and battery performance are shown in Fig. 3. It can be seen in Fig. 3a that the surface of the Mo foil is composed of a particle-like MoO 2 film, which can be evidenced by XRD results as shown in Fig. 3b. The charge/discharge curves and cycling performance are shown in Figs. 3c and 3d. The MoO 2 @Mo electrode exhibits a similar property of high voltage as the MoO 2 @Ni. The better battery performance of the MoO 2 @Ni electrode maybe results from the 3-D porosity network in nickel foam. [17][18][19] Based on the above evidences, it can be solidly confirmed that the high voltage was ascribed to the MoO 2 film.
To understand the reactions occurred on the MoO 2 @Ni electrode, the battery after 100 charge/discharge cycles was disassembled in a glove box. The electrode and the separator were cleaned by anhydrous ethanol and examined by SEM, EDX and XRD. Fig. 4a shows the SEM image of the cathode after discharge. The morphology became roughness than that of the pristine electrode. The corresponding EDX spectra indicates that the cycled cathode consists of Cl, Al and Ni, without Mo, as shown in Fig. 4c. The XRD patterns also confirm that the MoO 2 diffraction peaks disappeared after long cycles. However, a beaded morphology and intensive Mo peaks can be found in the SEM image of the separator and the corresponding EDX spectra after discharge, as shown in Figs. 4b and 4d, respectively. It indicates that the Mo transferred from the electrode to the separator during the charge/discharge process, since no Mo was contained in the pristine separator, as shown in Fig. 4f. This can partially explain the specific capacity decay during the cycling. A similar phenomenon of Mo transfer is also found in the MoO 2 /Mo electrode after charge/discharge for a few cycles. The results are presented in Figs. 5a to 5d. The original dense particle-like morphology was destroyed and the Mo can be detected in the separator.
Since there is a capacity decay occurred in the cycling test. It is important to evaluate the stability of MoO 2 in electrolyte in stationary condition. Therefore, a MoO 2 @Ni electrode was immersed into the ionic liquid for 24 h to observe the morphology changes as shown in Fig. 6a. The electrode was taken out and washed in anhydrous ethanol ultrasonically for a few minutes. The SEM image of the MoO 2 @Ni electrode after immersed is shown in Fig. 6b. It can be observed that the morphology of the MoO 2 @Ni electrode maintained the compact layer. The EDX spectrum also proved that the layer is mainly composed of Mo and O elements as seen in Fig. 6c. It is showed that the MoO 2 was not corroded severely in the electrolyte without electric field applied although we do not exclude the possibility that small amount of MoO2 could be dissolved into the electrode in the stationary condition, and we believe that the stationary corrosion is not the main cause for the capacity decay.
It is also necessary to point out that the mechanism of the ionic liquid based Al-ion battery is quite complex for researchers to understand at present as our previous study demonstated. 14 Even so, some previous studies still obtained a similar dissolution phenomenon with our results. For example, Suto et al. used VCl 3 and EMImCl/AlCl 3 to construct a battery. 5 They found that VCl 3 could be dissolved into the EMImCl/AlCl 3 electrolyte by X-ray absorption near-edge structure (XANES) examination, resulting in a poor cycling stability. Takuya et al. demonstrated a low voltage and poor cycling stability FeS 2 /Al battery. 11 The dissolution of sulfides was the main reason for the sluggish performance, which was also proved by XANES. We believe this may help to understand the phenomenon in this work. Fig. 7 shows the rate capability performance of the MoO 2 @Ni electrode Al-ion battery. Although the electrode showed a low ratecapability performance, the high discharge voltage of the MoO 2 @Ni   electrode under middle current density is still worth the further investigation, which should be conducted to explain the MoO 2 and EMImCl/AlCl 3 reaction mechanism in the future and prolong its cycle life.

Conclusions
A dense MoO 2 layer was fabricated on the surface of a Ni foam through magnetron sputtering and its electrochemical performance as the cathode of the rechargeable Al ion battery was investigated. The first discharge capacity was about 90 mAh g −1 at a current density of 100 mAh g −1 and the discharge potential plateau was about 1.9 V. Mo was transferred from the electrode to the separator during the charge/discharge process, resulting in a capacity decay of the battery. To our knowledge, it is the first time to obtain such a high discharge potential plateau for the metal oxide positive electrode in ionic liquid based Al-ion battery.