Synthesis and Electrochemical Properties of CuFeO2 as Negative Electrodes for Sodium-Ion Batteries
Yuhki Yui*, Yoko Ono, Masahiko Hayashi, Jiro Nakamura
Energy and Environment Systems Laboratories, NTT Corporation, Kanagawa 243-0198, Japan
To cite this article:
Yuhki Yui, Yoko Ono, Masahiko Hayashi, Jiro Nakamura. Synthesis and Electrochemical Properties of CuFeO2 as Negative Electrodes for Sodium-Ion Batteries. American Journal of Physical Chemistry. Vol. 4, No. 2, 2015, pp. 16-20. doi: 10.11648/j.ajpc.20150402.11
Abstract: This paper investigates the electrochemical properties of CuFeO2 prepared by using a malate precursor method as negative electrodes for sodium-ion batteries. In the voltage range of 0.01 to 3.0 V, the oxide calcined at 750°C shows a poor cycle property of 27 mAh/g at 10 cycles, although it exhibits the large first discharge capacity of 579 mAh/g. The cycle property is improved to up to 122 mAh/g at 10 cycles by employing a cycle condition with the voltage range of 0.01 to 1.5 V. The results suggest that the sodium-ion insertion/extraction mechanism in CuFeO2 would involve valence changes of Fe ions in the oxide from the Fe3+ state to Fe0 state in the discharge process and from Fe0 to Fe3+ in the charge process.
Keywords: Sodium-Ion Battery, Negative Electrode, CuFeO2
Lithium-ion batteries (LIBs) are widely used in mobile phones and notebook computers owing to their high voltages and high energy density. Furthermore, the market demand for large-sized LIBs as the power supply for electric vehicle has been increasing in recent years. However, lithium is expensive because it is not an abundant metal. On the other hand, sodium is abundant and cheap , and interest in sodium-ion batteries (SIBs) has been growing. Therefore, materials that have been studied for use as negative electrodes for SIBs include hard carbon [2 - 5], tin [6 - 8] and tin-based materials [9 - 10]. Hard carbon can be cycled more than 100 times, but its capacity is only about 250 mAh/g . On the other hand, the capacities of tin and tin-based materials are about 500 mAh/g , but their cycle performance is poor. In addition, some oxide materials, such as Na2Ti3O7 [11 - 12], Na4Ti5O12 , TiO2 [14 - 15], MoO3 , Fe3O4 and a-Fe2O3 , have been studied for use as SIBs’ negative electrodes. However, most of these materials show small capacity or poor cycle performance, which leads to insufficient battery performance. In addition, most of the oxide materials show high voltage, which is why their energy density are low.
CuFeO2 has a delafossite structure consisting of closely packed oxygen double layers, in which the octahedral sites are occupied by Fe3+ cations, and these double layers are linked by monovalent Cu+ ions with two-fold oxygen linear coordination . CuFeO2 has been studied as a negative electrode material for LIBs [19 - 20], and it exhibited a 475 mAh/g beyond 100 cycles. As reported in ref. , the reversible capacity of Li/CuFeO2 cell is expected to correspond to 4 mol of Li per mole of CuFeO2, and the theoretical capacity is 694 mAh/g.
In this study, we investigated the electrochemical properties of CuFeO2 as negative electrodes for SIBs. We performed X-ray diffraction (XRD) measurements and X-ray photoelectron spectroscopy (XPS) measurements of electrodes containing CuFeO2 during cycling to examine the reaction mechanism involved in the crystallographic-structure changes in the oxide.
CuFeO2 was prepared by using a malate precursor method . The precursor powder was obtained by evaporating a mixed aqueous solution of Cu(NO3)2•3H2O (Kanto Chemical Co.), Fe(NO3)3•9H2O (Kanto Chemical Co.), and malic acid (Kanto Chemical Co.). Then the precursor powder was calcined at 750 or 900°C for 10 h in Ar. The phases of the resulting powder were identified with a powder X-ray diffractometer (XRD, Rigaku, X-ray diffractometer RINT2000) using CuKa radiation. The morphology of the powder was observed by a scanning electron microscope (SEM, JEOL Ltd., JSM-890) with accelerating voltage of 7 kV.
Working electrodes for electrochemical measurements were prepared by mixing CuFeO2 powder, acetylene black as a conductive material, polyvinylidene difluoride (PVdF) binder in N-methylpyrrolidone, coating the mixture slurry on an Al sheet as a current collector, and finally drying it at 90°C in an oven. The electrochemical performance was evaluated with a 2032 coin-type cell using a working electrode (0.02 mm thick and 14 mm in diameter), a non-aqueous electrolyte (1 mol/l NaPF6 / EC:DEC 1:1 in volume, Tomiyama Pure Chemicals Industries Ltd.), a polypropylene separator (Celgard, 19 mm in diameter), and sodium metal sheets (0.6 mm thick and 15 mm in diameter, Kanto Chemical Co.) as a counter electrode. All cells were assembled in an Ar-filled glove box (dew point < -75°C). Cycling tests were performed with an automatic galvanostatic discharge-charge system unit (Hokuto Denko Corp. HJ1001SD8) at a current density of 25 mA/g between 0.01 and 3.0 V(or 1.5 V) at 25°C. Then, the cells were opened after the charge/discharge process in the Ar-filled glove box, and the working electrodes were soaked in dimethyl carbonate (Tomiyama Pure Chemicals Industries Ltd.) and dried in the glove box overnight. XRD measurements of the electrodes with discharged or charged states were performed. Inductively coupled plasma spectroscopy (ICP, Hitachi High-Technologies Co., SPS1700) was used to evaluate the composition of the Nax-CuFeO2 in the electrodes before and after charge/discharge processes. The sample solution for the ICP-AES analysis was prepared as follows. Working electrodes were removed from discharged/charged coin cells and washed with dimethyl carbonate (DMC) in the glove box. Then electrode materials were dissolved in a dilute acid mixture consisting of nitrohydrochloric acid, perchloric acid, sulfuric acid, and nitrohydrochloric acid. An X-ray photoelectron spectrometer (XPS) analysis was performed with a PHI XPS-5700 system using AlKa radiation (1486.6 eV) to examine the binding energy for Fe2p of as-prepared, discharged, or charged electrodes. The binding energies for Fe2p were identified using the NIST database 
3. Results and Discussion
Figure 1 shows the powder XRD patterns of the sample obtained at 750 or 900°C for 10 h in Ar. The XRD patterns well agree with that of hexagonal CuFeO2 (PDF No. 00-039-0246). Furthermore, they are consistent with the pattern reported in the literature . These results indicate that CuFeO2 was obtained as a single phase. Figure 2 shows SEM images of the CuFeO2 particles sintered at 750°C and 900°C. The oxide sintered at 750°C has the primary particle size of 0.3-0.5 mm. On the other hand, the oxide sintered at 900°C consist of primary particles of more than 1.0 mm, and the particles are aggregate by sintering at higher temperature.
Figures 3 and 4 respectively show the first and second discharge (Na-ion insertion)-charge (Na-ion extraction) curves and the cycle properties of a Na/CuFeO2 (750, 900°C) in the voltage range of 0.01 to 3.0 V. In the first discharge profile of CuFeO2 obtained at 750°C, the voltage drops rapidly to about 1.0 V from the open-circuit voltage of about 2.8 V and then decreases gradually to 0.01 V. In Li/CuFeO2, the discharge plateau voltage was observed at around 1.0 V, and the first discharge and charge capacities were about 935 and 700 mAh/g . This suggests that the reaction mechanism is different between Na/CuFeO2 and Li/CuFeO2. Na/CuFeO2 could have higher energy density than Li/CuFeO2, because it showed low voltage. The first discharge and charge capacities of CuFeO2 obtained at 750°C were 579 (equivalent to 3.7 Na per CuFeO2 mol) and 383 mAh/g (equivalent to 2.5 Na per CuFeO2 mol), respectively. ICP analysis results revealed that 3.2 and 1.9 mol of sodium ions were inserted/extracted after the first discharge/charge processes (Table 1). These results are in good agreement with the values calculated from the discharge/charge capacities as shown in Fig. 3(a). The large irreversible capacity of about 200 mAh/g in the first cycle might be due to significant changes in the crystal structure with Na-ion insertion during the first discharge. The first discharge and charge capacities of the sample obtained at 900°C were 538 and 238 mAh/g, respectively. Since CuFeO2 at 900°C has large particle size as shown in Fig. 2(b), Na ions might not easily insert into the CuFeO2 bulk. That would explain why the capacity of the sample obtained at 900°C was smaller than that of the sample obtained at 750°C and why the over potential between discharge and charge voltage was higher than that the one of the sample obtained at 750°C. Additionally, both oxides showed much higher second charge voltage than the first one. This led to a drastic decline to where the oxide at 750°C showed a very low discharge capacity retention of about 5% at the 10th cycle as shown in Fig. 4. This suggests that CuFeO2 underwent severe crystallographic changes during the first cycle.
Figure 5 shows the XRD patterns of CuFeO2 obtained at 750°C during the first cycle in the voltage range of 0.01 to 3.0 V. After the first discharge (Na-ion insertion), the peaks corresponding to CuFeO2 almost completely disappeared as shown in Fig. 5(b). This result suggests that the CuFeO2 crystalline phase might become amorphous phase. Moreover, new peaks assigned to NaCuO (PDF No. 01-071-2300) and Fe3O4 (PDF No. 00-003-0862) were observed. After the first charge, there were small peaks of CuFeO2 and NaCuO peaks as shown in Fig. 5(c).
|Process||x in Nax-CuFeO2|
|From capacity||From ICP analysis|
|After first discharge||3.3||3.2|
|After first charge||2.2||1.9|
XPS measurements provide information to examine the reaction mechanism, and the binding energies for Fe2p were identified using the NIST database. Figure 6 (a) shows the wide–scan spectrum of pristine CuFeO2 electrodes. The Fe2p, Cu2p, O1s, F1s and C1s peaks are assigned as shown in Fig. 6(a). The typical narrow-scan spectra of Fe2p after discharge to 0.01 V and after charge to 3.0 V are shown in Fig. 6(b) for pristine electrodes. The Fe2p peak shifted toward lower binding energy from 711.2 (pristine, the Fe3+ state) to 710.5 eV (after discharge and after charge) as shown in Fig. 6(b).
Additionally, a shoulder at 707.0 eV (the Fe0 state) in the Fe2p spectra appeared after discharge and disappeared after charge, indicating that the bonding state of the Fe atoms has changed. This change can be attributed to the change from the Fe3+ state to Fe0 state due to discharging (Na-ion insertion). At the next charge, the Fe0 state changes to the Fe3+ state by charging (Na-ion extraction). This indicates that the bonding state of the Cu atoms must have changed with the changing bonding state of the Fe atoms. In Na/CuFeO2, both of the bonding states and the crystal structure change during cycling. That is why Na/CuFeO2 has a poor cycle performance.
In an attempt to improve the cycle performance, we carried out cycling tests under the discharge/charge condition with a narrower voltage range. Figures 7 and 8 show the first and second discharge-charge curves and the cycle properties of a Na/CuFeO2 (750°C), respectively, in the voltage range of 0.01 to 1.5 V. The first and second discharge capacities were 435 and 223 mAh/g, respectively. These capacities in the voltage range of 0.01 to 1.5 V were smaller than those in the range of 0.01 to 3.0 V. The smaller capacity was due to the insufficient charge process as a result of the narrower voltage range. The larger discharge capacity of 122 mAh/g at the 10th cycle was obtained in the narrower voltage range, although the capacity at the 10th cycle was 27 mAh/g for the wide voltage condition. As described above, the cycle property was improved by employing a suitable discharge/charge voltage range.
CuFeO2 powders were synthesized by using a malate precursor method. Their electrochemical properties as negative electrodes for SIBs were examined. The particle size of the oxide sintered at 750°C (0.3-0.5 mm) is smaller than the one sintered at 900°C (0.5-1.0 mm). The discharge capacity of the 750°C oxide (579 mAh/g) was larger than that of 900°C one (538 mAh/g) in the 0.01 to 3.0 V range. Na ions might easily insert into the bulk of the CuFeO2 obtained at 750°C because of the small particles, which would explain its large capacity. The cycle properties were poor (750°C: 27 mAh/g up to 10 cycles). The reasons for the poor cycle performance are changes in the bonding states of the atoms and changes in the crystal structure during cycling. In the narrow voltage range of 0.01 to 1.5V, the cycle property was improved from 27 to 122 mAh/g at 10 cycle.