Introduction

In recent years, the electric vehicles which have a fast-growing market are faced with serious problems in their power systems which have led to increasing demand for high energy density storage systems [1]. Since the specific energy density of traditional lithium ion batteries are getting close to the theoretical limit, breakthroughs in a new battery system with ultra-high energy density are strongly expected [2]. Rechargeable Li-O2 batteries with significantly high theoretical energy density have gained wide attention [3, 4]. Compared with current Li-ion batteries, Li-O2 batteries have 10 times higher theoretical energy density owing to its different electrochemical reaction mechanism [5, 6]. Hence, much interest has been aroused among research workers since the first report by Abraham et al. [7] and the first experimental confirmation by Bruce et al. [8]. However, Li-O2 batteries still suffer many challenges for practical applications, such as high overpotentials, poor cycle stability, and low rate capacity [9–11]. The main reason may be due to the sluggish kinetics of the oxygen evolution reaction (OER) and oxygen reduction reaction (ORR). So it is necessary to develop highly efficient cathode catalysts which can promote the OER and ORR, thus leading to a reduced overpotential and an improved cycle performance [12–16].

To date, various electrocatalysts, including carbons, metal oxides, precious metals, have been investigated as the cathode catalysts in Li-O2 batteries to address the sluggish kinetics of ORR and OER in Li-O2 batteries and lower the overpotentials [17–21]. Among these, transition metal oxides are of great interest because of the low price, relative stability, good catalytic performance, etc. MnO2 and Co3O4 and their composites are the most common transition metal oxides that have been studied by many researchers [19, 22–24]. Up to now, Fe-based catalysts have not been widely studied in Li-O2 batteries, but they exhibit efficient electrocatalytic activities for the ORR in fuel cells [25]. Recently, some work on iron oxides as catalyst in Li-O2 batteries was reported. Wenyu Zhang et al. [26] synthesized Fe2O3 decorated graphene through an electrochemical deposition method to investigate its catalytic performance in Li-air batteries. Zhang et al. [27] synthesized Fe2O3 nanoflakes as cathode catalysts in Li-O2 batteries. Jun Lu et al. [28] fabricated Fe3O4/Fe/carbon composite for Li-O2 batteries. All these works showed an obviously improved electrochemical performance, including lower overpotentials, higher capacity, and better cycling performance, but the polarization performance still should be improved. These studies inspired us to continue to explore the electrocatalytic performance of Fe-based catalysts for Li-air batteries. Owning to the advantages of low price, environment friendliness, and good performance, Fe-based catalysts should have a better potential in Li-O2 batteries.

Here, in this work, we synthesized a kind of Fe/Fe2O3 core-shell nanowires through the reduction of Fe3+ ions by sodium borohydride in aqueous solution at ambient atmosphere. With a further heat treatment, the Fe nanowires were coated with a layer of Fe2O3. This kind of composite was used as a cathode in Li-O2 batteries. The charge overpotential was reduced and a good cycling performance was achieved. The Li-O2 cell with Fe@Fe2O3 as a cathode can cycle more than 40 times with a cutoff specific capacity of 1000 mAh g−1.

Experimental

Preparation of Fe@Fe2O3 composites

All chemical reagents were of analytical grade and were used without further purification. In a typical synthesis, 0.3 g of FeCl3·6H2O was added to 60 mL of deionized water and 0.6 g of NaBH4 was dissolved in 40 mL deionized water. NaBH4 solution was then poured into FeCl3 solution at the rate of 0.8 mL/s. During the addition, the solutions were shaken by hand. The solution was bubbled with plenty of gas with the addition of NaBH4 solution, accompanied by fluffy black precipitates that appeared on the surface of the solution. The fluffy black precipitates were washed with deionized water and ethanol and finally dried under vacuum at 60 °C overnight. Finally, the Fe@Fe2O3 was obtained by heat treatment in an oven at 200 °C for 2 h.

Material characterization

X-ray powder diffraction patterns were obtained on a Rigaku Ultima diffractometer with Cu-Kα radiation (λ = 1.542 Å). Morphologies of the synthesized samples were characterized with field emission scanning electron microscope (FESEM JSM-4800F) and a transmission electron microscope (TEM JEM-2100F). The surface area was determined by BET (Brunauer–Emmett–Teller) measurements using a Tristar 3000 surface area analyzer. To confirm the surface component, the X-ray photoelectron spectrum (XPS, Thermo Fisher Scientific ESCAlab250) was also recorded. Cyclic voltammetry measurements in the potential window from 2.0 to 4.3 V (vs. Li+/Li) with a scan rate of 0.1 mV s−1 was tested on an Autolab PGSTAT302 potentiostat-galvanostat electrochemical work station (Metrohm).

Preparation of Li-O2 cell and electrochemical investigation

The electrocatalytic activities of the as-prepared Fe@Fe2O3 for the discharge and charge steps were tested in Li-O2 cells. The electrodes were formed by mixing the Fe@Fe2O3, Ketjen black carbon (KB), and polytetrafluoroethylene (PTFE) with a weight ratio of 7:1:2 then rolling it into a film which was punched into small disks and pressed onto carbon paper as cathode. The electrochemical cells used were based on a Swagelok Cell design composed of a Li metal anode, an electrolyte of 1 M LiTFSI in TEGDME solvent, the Celgard 2400 separator, and the as-prepared porous cathode. To avoid the influence of H2O and CO2 or any other impurities, the cells were operated at 1.5 mbar of pure O2.

The galvanostatic charge and discharge tests were conducted on a LAND CT2001A battery test system at different current densities, at ambient temperature after a 3-h rest period. All the specific capacities were calculated by normalizing with the mass loading of catalyst in the air electrodes.

Results and discussion

Figure 1a shows the XRD pattern of the as-prepared Fe nanowires and Fe@Fe2O3 nanowires. The peaks located at 2 θ ≈ 44.7° and 65.0° correlated to the (110) and (200) peaks of body-centered cubic iron (JCPDS card No. 06–0696). The peaks at 2 θ ≈ 24.1°, 33.1°, 35.6°, 49.5°, 54.1°, and 62.4° correlated to the (012), (104), (110), (024), (116), and (214) peaks of α-Fe2O3 (JCPDS card 33–0664). No other peaks have been observed, illustrating that the samples were comparatively pure with less impurities. It is obvious that the peak intensity of Fe is much stronger than Fe2O3, indicating the formation of Fe2O3 on the surface of Fe nanowires.

Fig. 1
figure 1

a XRD pattern of Fe nanowires and Fe@Fe2O3 nanowires; b XPS spectra of Fe 2p3/2 peaks of the Fe@Fe2O3 nanowires

Full size image

The chemical composition of Fe@Fe2O3 core-shell nanowires were further investigated by X-ray photoelectron spectroscopy (XPS). The XPS results of Fe 2p3/2 spectra are shown in Fig. 1b. The Fe 2p3/2 core level peak appeared at the binding energies of 709.8 eV corresponding to Fe2O3 [29]. It indicated that the surface of the Fe nanowires was coated with Fe2O3 layer, which was further confirmed by the TEM image. The peak at a lower binding energy of 707.0 eV is attributed to Fe 2p3/2 in pure iron [30]. A higher binding energy peak at about 712.3 eV may be attributed to an interaction between the Fe core and the Fe2O3 shell [29, 31].

Figure 2a shows the SEM images of the as-prepared sample. It shows that the morphology of the Fe@Fe2O3 was nanowire structures and crisscross-like branches. The nanowires are several microns in length and 50–100 nm in diameter. The resulting nanostructures were further investigated by TEM. The structure of nanowires was further confirmed by the TEM images (Fig. 2b, c). The contrast between the edge and the center of the TEM image in Fig. 2c shows the core-shell structure of the Fe@Fe2O3 nanowires and also confirms that the Fe2O3 layer is about 15 nm in thickness.

Fig. 2
figure 2

SEM image (a) and TEM images (b, c) of Fe@Fe2O3 nanowires; (d) typical N2 gas adsorption-desorption isotherm of the Fe@Fe2O3 powder

Full size image

The BET-specific surface area detected by nitrogen adsorption-desorption measurements of Fe@Fe2O3 was about 53.2 m2 g−1. The pore size distributions of Fe@Fe2O3 are shown in the inset of Fig. 2d. The strong peak at ∼4 nm indicates the mesoporous nature of the prepared powder. The high specific surface area would be beneficial for the deposition of discharge products during ORR process which could increase the specific capacity of the Li-O2 batteries.

The electrocatalytic activity of the as-prepared Fe@Fe2O3 catalyst for ORR and OER reactions in Li-O2 cells was investigated under galvanostatic cycling conditions. The initial discharge-specific capacity was measured at different current densities to characterize the rate capability of the cell (Fig. 3a). The battery with the Fe@Fe2O3 cathode exhibits a high discharge capacity of 2270 mAh g−1 at the current density of 100 mA g−1. The specific capacity decreased with the progressively increasing current from 100 to 300 mA g−1. Remarkably, the discharge capacity can still reach 1140 mAh g−1 at a high current density of 300 mA g−1. Another important point is that the medium voltage during the first charge is only about 3.8 V at the current density of 100 mAh g−1. The high specific capacity may be attributed to the high specific surface area and mesoporous nature of the Fe@Fe2O3 nanowires, and the core of Fe in the nanowire could increase the electron conductivity, which may lead to the good rate capability.

Fig. 3
figure 3

First galvanostatic discharge/charge curves at different current density (a); discharging and charging profiles (b); cycle performance (c) of Li-O2 batteries with Fe@Fe2O3 cathode

Full size image

The cyclic performance of the Fe@Fe2O3 cathode in Li-O2 battery was shown in Fig. 3b. In order to avoid a large depth of discharge and achieve improved cyclic performance of the batteries, the discharge capacities of the electrodes were limited to 1000 mAh g−1. The battery exhibited excellent cycle stability over 45 cycles without pronounced variation in the very stable discharge/charge curves (Fig. 3c). After 30 cycles, the cutoff discharge voltage is still above 2.6 V and the cutoff charge voltage below 4.0 V, showing the good stability of Li-O2 battery with the Fe@Fe2O3 cathode.

One possible factor for this good catalytic performance is the core-shell Fe@Fe2O3 structure. Based on some published references [32, 33], we presented the possible catalytic process of the Fe@Fe2O3 core-shell nanowires in Fig. 4a. Molecular oxygen could be adsorbed on the surface of iron oxide and could be reduced to peroxide iron or superoxide iron via the two-electron or one-electron transfer from iron core to the iron oxide shell surface to produce Li2O2 or LiO2 during the ORR process. The electrocatalytic activity of the Fe@Fe2O3 electrode was further investigated by cyclic voltammetry (CV) (Fig. 4b). The Fe@Fe2O3 electrode exhibited a high ORR onset potential (2.9 V), which implied a low ORR kinetic overpotential. The OER process started at the potential of 3.0 V, and the two obvious OER peaks between 3.0 and 3.5 V should be attributed to the oxidation of LiO2 and Li2O2 [32, 34]. The peak above 4.15 V should be due to the deposition of the carbon-containing substance such as electrolyte and carbon [35]. Also the high surface area of the nanowire structure could be beneficial for the discharge products’ deposition, and it is also helpful for the transmission of oxygen. This result suggests that Fe@Fe2O3 has considerable ORR/OER catalytic activity in Li-O2 batteries. Since the catalytic reaction in Li-O2 batteries is quite complicated, the reaction process still needs further study.

Fig. 4
figure 4

a Possible reaction process in Li-O2 batteries with Fe@Fe2O3 nanowires as catalyst; b cyclic voltammetry curves between 2.0 and 4.3 V at 0.1 mV s−1

Full size image

Conclusions

Fe@Fe2O3 core-shell nanowires were synthesized via the reduction of Fe3+ ions by sodium borohydride in aqueous solution with subsequent heat treatment to form Fe2O3 shell and employed as a cathode catalyst for non-aqueous Li-air batteries. The resulted electrodes could promote the cycle stability of Li-O2 batteries. The good performance may be due to the high specific surface area which provides more space for discharge products to deposit and the core-sell structure’s good catalytic performance. The Fe core not only increased the conductivity of the cathode but also improved the catalytic activity. These data show that the Fe@Fe2O3 core-shell nanowires could be a promising catalyst in Li-O2 batteries.