α-Fe₂O₃ and Fe₃O₄ hollow nanospheres as high-capacity anode materials for rechargeable Li-ion batteries

Abstract

α-Fe₂O₃ and Fe₃O₄ hollow nanospheres of size 30 ± 2 nm were synthesized by using polymeric micelles as a soft template for the first time. The hollow nanospheres were thoroughly characterized by transmission electron microscope, superconducting quantum interference device, X-ray diffraction, Fourier transform infrared spectroscopy, thermogravimetry and differential thermal analysis, cyclic voltammogram, and nitrogen sorption analyses. The α-Fe₂O₃ and Fe₃O₄ hollow nanospheres were used as anode materials in lithium-ion rechargeable batteries to investigate their electrochemical properties. The hollow particle-based electrodes exhibit high capacity, stable cycling performance, and good rate capability at different current densities. The α-Fe₂O₃ and Fe₃O₄ hollow nanospheres with nanosized shell domain favors fast lithium insertion/extraction processes during the repeated charge/discharges.

Introduction

Developing Li-ion batteries (LIBs) with high specific capacities and high current densities as power sources for diverse applications is of great interest [1–4]. The development of next-generation LIBs is a key to the success of electric and hybrid electric vehicles, advanced electronic devices, and implantable medical devices [1, 5, 6]. Since the pioneering work of Poizot and co-workers in the year 2000 on the reversible full reduction of 3D metal oxides, the research on α-Fe₂O₃ electrode has gained much interest [7]. Iron oxide, one of the main components of the earth’s crust, is considered to be a potential candidate with a high charge/discharge current densities and high reversible capacity coupled with its low cost and environmental benignity. Stimulated by both the promising applications of iron oxides coupled with novel properties of nanosize, considerable efforts have been made in the synthesis of α-Fe₂O₃ nanoparticles, nanorods, nanowires, and nanotubes [8–14]. Investigation by Larcher et al. and Morales et al. on Fe₂O₃ nanoparticles and by Chen et al. and Zhang et al. on carbon-coated Fe₃O₄ nanospindles has clearly proved that the morphology of nanosized particles plays a significant role in the reactivity of the electrode materials toward lithium [15–17]. In addition, it has been proved that the electrochemical reaction of Li with Fe₂O₃ is thermodynamically feasible, and up to 6 mol of Li can be intercalated to give a maximum theoretical capacity of 1,007 mAh g⁻¹ [17]. Similarly, Fe₃O₄ can accommodate up to 8 mol of Li making them promising anode materials for high-performance lithium-ion batteries.

Recently, polymeric micelles with core–shell–corona structure have increasingly used for the fabrication of hollow nanospheres [18]. The micelle of poly(styrene-b-2-vinylpyridine-b-ethylene oxide) (PS−PVP−PEO), consisting of a PS core, a PVP shell, and a PEO corona was used as a soft template in the fabrication of hollow silica nanospheres in our laboratory. Herein, we report on the fabrication of hollow α-Fe₂O₃ and Fe₃O₄ nanospheres with micelles of poly(styrene-b-acrylic acid-b-ethylene oxide) (PS–PAA–PEO) using FeCl₃·6H₂O as iron precursor. The importance of this strategy is that the PS core functions as template for hollow void, PAA shell domain containing anionic −COO⁻ ions serves as effective reaction center for metal ions, and PEO corona stabilizes the particles from aggregation. In the present report, the precursor metal cations (Fe³⁺) are first precipitated on the anionic PAA block with COO⁻ groups in the form of hydroxides, and this phase is subsequently transformed to porous α-Fe₂O₃ hollow nanospheres by calcinations in air atmosphere. The α-Fe₂O₃ hollow spheres further converted to magnetite hollow nanospheres by controlled reduction with a mixture of H₂ and N₂ at 350 °C. To the best of our knowledge, this is the first report for the synthesis of magnetic hollow nanospheres of size about 30 nm using soft template strategy. Both α-Fe₂O₃ and Fe₃O₄ nanospheres were further investigated as anode materials for their electrochemical lithium intercalation/deintercalation behaviors in rechargeable LIBs.

Experimental

Material preparation and characterization

The α-Fe₂O₃ hollow nanospheres were fabricated as depicted in Scheme 1. The micelle solution was adjusted to pH 8 by using dilute NaOH solution. The required amount of micelle solution (10 mL) was stirred for a few minutes followed by addition of FeCl₃·6H₂O under vigorous stirring (the ratio of Fe³⁺/PAA was 10). Then dilute NH₄OH was added slowly until the pH is reached to about 10 and the obtained brown precipitate was aged for 2 days at room temperature under static conditions. The composite particles were repeatedly washed with distilled water and ethanol and dried at 60 °C. In order to remove the polymeric template as well as to crystallize the hollow particles, the composite particles were heated to 400 °C for 3 h in a muffle furnace under air. The α-Fe₂O₃ hollow nanospheres obtained by the above method were further reduced to Fe₃O₄ hollow nanospheres using a mixture of H₂ and N₂ at 350 °C for 2 h. Prior to transformation of Fe₂O₃ to Fe₃O₄ hollow nanospheres, the suitable reduction temperature (350 °C for 2 h) was identified since high temperature (above 450 °C) leads to total reduction of Fe³⁺ to a certain extent. Wide-angle X-ray diffraction (XRD) was recorded on a Shimadzu XRD-7000 (WXRD) diffractometer. N₂ adsorption/desorption analyses were carried out by using an Autosorp I apparatus. Transmission electron microscope (TEM) images were recorded on JEOL JEM-1210 (80 kV) and JEM-2100 (200 kV) electron microscopes. Fourier transform infrared (FTIR) spectra were recorded on a Jasco FT/IR 7300 spectrometer by using a KBr pellet technique. Thermogravimetry and differential thermal analysis (TG–DTA) were obtained with a MAC Science TG-DTA 2100 instrument under air. The magnetic measurement was made on Quantum Design SQUID MPMS-XL instrument, and temperature-dependant magnetization data were obtained at 5,000 Oe

Scheme 1

Fabrication of Fe₂O₃ and Fe₃O₄ hollow nanospheres using micelles

Electrode preparation

For lithium insertion studies in lithium-ion batteries, the Fe₂O₃/Fe₃O₄ nanosphere (5 mg) was mixed mechanically with teflonized acetylene black (2 mg), and then the mixture was pressed on a stainless steel mesh as the current collector under a pressure of 500 kg/cm² and dried at 160 °C for 4 h under vacuum. The electrochemical characterizations were carried out using CR-2032 coin-type cells with pure lithium (Cyprus Foote Mineral Co.) as an anode. The electrolyte used was 1 M LiPF₆-EC/DMC (1:2 by volume, Ube Chemicals Co. Ltd.). The coin cell assembling was performed in a glove box filled with argon (dew point, lower than −80 °C). Cyclic voltammograms (CVs) were recorded by Hokuto denko HSV-100 (Japan) in a beaker-type cell which contains Fe₂O₃ or Fe₃O₄ working electrode, and lithium foil acts as both reference and counter electrodes. The beaker-type cell was assembled in the argon filled glove box and sealed well using Teflon seal tape to prevent air oxidation.

Results and discussion

The fabrication of Fe₂O₃ and Fe₃O₄ hollow nanospheres using PS–PAA–PEO micelles is depicted in Scheme 1. On gradual addition of FeCl₃ solution to the above micelles, the anionic PAA shell block effectively binds with Fe³⁺ ions through electrostatic interaction as shown in the Scheme 1. The charge neutralization between –COO⁻ and Fe³⁺ ions also confirmed from change in the ξ-potential from −36 to 0 mV with gradual addition of FeCl₃ to the micelle solution (Fig. S1, supporting information). Thermal analyses (TG/DTA) of iron oxide/polymer composite particles (prior to calcinations) indicated that most of the templates were decomposed below 400 °C (Fig. S2, supporting information). The FTIR spectrum of composite particles (Fig. S3-A, supporting information) containing –COO–Fe³⁺ complex exhibits bands at 1,737 and 1,456 cm⁻¹ corresponding to asymmetric and symmetric stretching vibration of –COO⁻, respectively. The FTIR spectra of α-Fe₂O₃ and Fe₃O₄ materials exhibited bands at 580, 850–1,040, and 1,600 cm⁻¹ characteristic of Fe–O bond vibrations (Fig. S3B-C, supporting information) [19]. The absence of C–H, –C=C–, and –COOH bonds stretching vibrations of phenyl groups of the polymer backbone in the calcined sample suggests the complete removal of templates from the hollow nanospheres. Furthermore, thermal analysis of calcined samples revealed negligible amount of organics in the Fe₂O₃ hollow particles. Figure 1a, c shows the TEM images of α-Fe₂O₃ and Fe₃O₄ hollow nanospheres, respectively; the average particle size and void space diameter of hollow nanospheres were found to be 30 ± 2 and 15 ± 1 nm, respectively; the reduction of α-Fe₂O₃ to Fe₃O₄ has not affected the spherical morphology of particles. All the hollow particles are polycrystalline, and the shell domain is comprised of nanoparticles of diameter about 2–4 nm as evidenced from the high-resolution TEM images (Fig. 1b). The wall thickness estimated from TEM images was found to be 8 ± 0.5 nm. Figure 2 shows the powder XRD patterns of nanosized hollow iron oxides; the XRD profile prior to hydrogen treatment matches well with the standard α-Fe₂O₃ (JCPDS 33-0664) [20, 21]. After reduction, the crystalline structures could be indexed to Fe₃O₄ (JCPDS 75-1609). All detected sharp diffraction peaks could be indexed as Fe₃O₄, and the absence of peak at 2θ = 44 ° clearly confirms the absence of metallic iron (Fe⁰) [21]. It is important to note that controlled reduction at low temperature (at 350 °C for 2 h) avoids Fe metal formation unlike the high-temperature treatment using H₂. Furthermore, the α-Fe₂O₃ hollow nanospheres were completely converted to Fe₃O₄ because the peaks unique for α-Fe₂O₃ (including the peaks at 24.1, 33.08, and 40.86) disappeared completely. The formation of Fe₃O₄ is further confirmed by the color change from brown to totally black materials.

Fig. 1

TEM images of hollow nanospheres: a α-Fe₂O₃ (low magnification), b α-Fe₂O₃ (high magnification), and c Fe₃O₄

Fig. 2

Wide-angle XRD patterns of α-Fe₂O₃ and Fe₃O₄ hollow nanospheres

The Brunauer–Emmett–Teller surface areas evaluated by nitrogen adsorption/desorption analyses of α-Fe₂O₃ and Fe₃O₄ hollow nanospheres were found to be 106 and 93 m² g⁻¹, respectively. Magnetic characterization was performed using a superconducting quantum interference device (SQUID). The temperature-dependant magnetization at 5,000 Oe for Fe₃O₄ hollow particles is shown in the supporting information (Fig. S4). The hollow Fe₃O₄ nanospheres are superparamagnetic at room temperature. The hollow particles exhibited a magnetic moment of 22.3 emu/g which is comparable to values obtained by Sun et al. for Fe₃O₄ hollow nanospheres of size about 20–25 nm [21]. Furthermore, the decrease in magnetic susceptibility with decreasing temperature indicates the hollow nanospheres become antiferromagnetic at low temperature, and the antiferromagnetism increases with decreasing temperature up to 2 K. The Néel temperature (T _N), the critical temperature for onset of antiferromagnetism, is observed at about 200 K characteristics of magnetite particles [22].

Figure 3a, b shows discharge–charge curves in the voltage window of 0.005−3.0 V (vs. Li) at a rate of 0.25 C up to 50 cycles, and for clarity, only selected cycles are shown in the voltage versus capacity profiles. During the first discharge, both α-Fe₂O₃ and Fe₃O₄ exhibit voltage plateaus at about 1.1 and 0.8 V (vs. Li) resulting from the lithium reaction with iron oxide nanospheres [14, 17]. The voltage plateau at 1.1 continues until a capacity of about 220 ± 20 mAh g⁻¹ is reached for both hametite and magnetite which is ascribed to Li insertion into the shell domain of hollow nanospheres, whereas the voltage plateau at 0.8 V continuing up to 950 mAh g⁻¹ reflects the reduction of Fe³⁺ to Fe⁰ metal clusters [16]. The charge curve showed a sloping plateau at 1.95 V due to reverse oxidation reaction of Fe⁰ → Fe³⁺. The total first-discharge capacity is 1,430 ± 10 mAh g⁻¹ which corresponds to 8.6 mol of Li/mol of Fe₂O₃, whereas Fe₃O₄ exhibited slightly higher discharge capacity 1,495 mAh g⁻¹ (9.0 mol Li/mol of Fe₃O₄) at similar experimental conditions. The first-discharge profiles of both hametite and magnetites qualitatively resemble that observed by Larcher et al. and Morales et al. on Fe₂O₃ nanoparticulate and Chen et al. on Fe₂O₃ nanotubes [14–17]. It is noteworthy to mention that in the complete reduction of Fe³⁺ → Fe⁰, one would expect a maximum uptake of 6 Li/Fe₂O₃ or 8 Li/Fe₃O₄ (1,005 mAh g⁻¹), and the excess capacity must originate from electrolyte decomposition in the low-potential region and subsequent formation of solid electrolyte interphase on the hollow nanospheres [23, 24].

Fig. 3

Charge/discharge profiles of α-Fe₂O₃ (a) and Fe₃O₄ (b) hollow nanospheres at 0.25 C rate in the potential region of 0.005–3.0 V vs. Li/Li⁺

The discharge capacities of the α-Fe₂O₃ electrode (Fig. 3a) in the 1st, 2nd, 5th, 10th, 25th, and 50th cycles are 1,435, 1,109, 895, 681, 518, and 490 mAh g⁻¹, respectively. The corresponding charge capacity values are 998, 943, 811, 650, 503, and 485 mAh g⁻¹ for the 1st, 2nd, 5th, 10th, 25th, and 50th cycles, respectively. The coulombic efficiency of first cycle was found to be 69.5 %. The corresponding capacity versus cycle number plot is shown in Fig. 4a. As can be seen, both discharge and charge capacities decrease up to 15 cycles and thereafter the capacity stabilizes. After 50 cycles with 100 % depth of discharging and charging at a rate of 0.25 C, the electrode capacity decreased to 485 mAh g⁻¹. The capacity retention of Fe₂O₃ from 15th to 50th cycle is 84.2 %. Similar charge/discharge cycling performance was also observed on α-Fe₂O₃ nanotube electrodes and nanoflakes embedded on copper foil [17, 25]. Furthermore, it is relevant to mention that the capacity of commercial α-Fe₂O₃ faded within few cycles to 250 mAh g⁻¹ and gradually decreased to about 130 mAh g⁻¹ in less than 50 cycles [26]. However, the performance of α-Fe₂O₃ hollow nanospheres in the present study is more than three times higher than that of commercial Fe₂O₃ and porous α-Fe₂O₃ nanotubes in the absence of any additives [27].

Fig. 4

Charge/discharge cycling performance of α-Fe₂O₃ (a) and Fe₃O₄ (b) hollow nanospheres at 0.25 C rate in the voltage region of 0.005–3.0 V vs. Li/Li⁺

The magnetite (Fe₃O₄) hollow nanospheres, however, exhibit enhanced charge/discharge capacities (Fig. 3b) compared to α-Fe₂O₃. The discharge capacity values (Fe₃O₄) at 0.25 C rate in the voltage window of 0.005–0.3 V (vs. Li) for 1st, 2nd, 5th, 10th, 25th, and 50th cycles are 1,498, 1,139, 1,041, 890, 694, and 690 mAh g⁻¹, respectively. The corresponding charge capacity values of Fe₃O₄ are 1,060, 1,044, 990, 867, 695, and 688 mAh g⁻¹ for the 1st, 2nd, 5th, 10th, 25th, and 50th cycles, respectively. The coulombic efficiency of first cycle was found to be 71 %, marginally higher than the α-Fe₂O₃ hollow nanospheres. The cycling performance of Fe₃O₄ (Fig. 4b) also shows similar behavior as that of α-Fe₂O₃ and the discharge capacities decrease up to 15 cycles possibly due to the formation of stable electrolyte film, and complete coverage of all hollow nanoparticles may require several charge/discharge cycles and similar observation was also noticed by other groups [14–17]. The observed discharge capacity after 50 cycles (690 mAh g⁻¹) for Fe₃O₄ hollow nanospheres is considerably higher than the carbon/Fe₃O₄ nanospindles (600 mAh g⁻¹) composite electrodes, where the Fe₃O₄ nanoparticles were mixed with carbon matrix [25]. However, it is worth to note that carbon/Fe₃O₄ electrode delivers 600 mAh g⁻¹ after 80 cycles. On comparison with commercial Fe₃O₄ nanoparticles, which fades quickly and as a result provides low capacity (152 mAh g⁻¹, after 50 cycles) [26], the Fe₃O₄ hollow nanospheres exhibit 4.5 times higher discharge capacity than the commercial Fe₃O₄. The capacity fading may be partly due to volume change during repeated charge/discharges similar to Sn- and Si-based anodes [28]. The difference in electrochemical performance of α-Fe₂O₃ and Fe₃O₄ with nearly similar size 30 ± 2 nm is attributed to the intrinsic electronic conductivities of latter which contribute to higher discharge/charge capacities and superior performance than the α-Fe₂O₃ [29]. This fact is further confirmed from the enhanced performance of Fe₃O₄/single-walled carbon nanotubes (SWNT) due to high electronic conductivity of these composite electrodes containing SWNT [30]. Therefore, one can expect still improved performance of Fe₃O₄ hollow nanospheres when mixed with appropriate carbon additives. At this juncture, we are investigating the performance of synthesized Fe₂O₃ and Fe₃O₄ with different carbon additives. We believe this may enhance the initial coulombic efficiency and cyclability.

Figure 5a, b shows the rate performances of α-Fe₂O₃ and Fe₃O₄ hollow nanospheres, respectively. At a low rate of 0.15 C (Fe₃O₄, charge–discharge of all active materials in period of 400 min) or 0.1 C (Fe₂O₃), both α-Fe₂O₃ and Fe₃O₄ show about 1,100 to 1,150 mAh g⁻¹. However, at a high rate of 5 C (discharge/charge of all active materials within 12 min), the specific capacity of α-Fe₂O₃ decreases to 265 mAh g⁻¹, whereas that of Fe₃O₄ reduced to 410 mAh g⁻¹ at the same rate of 5 C. Thus, consistence with other nanostructured materials with different morphologies, the iron oxide hollow particles also show higher capacity at slower rate (0.1 C) than that at faster rate (5 C). However, the α-Fe₂O₃ and Fe₃O₄ electrodes almost regain their original high capacities when the rate was again lowered to 0.3 C after being exposed to high current loads (5 C), which indicates the high stability of hollow nanosphere-based electrodes. The improved electrochemical performance of both Fe₂O₃ and Fe₃O₄ is attributed partially to the unique hollow spherical morphology. More importantly, the void space not only effectively buffers against charge storage and local volume change but also provides better electrical contact and shorter diffusion path length providing better rate capability. In addition, enhanced capacity retention, rate performance, and cycling performance of Fe₃O₄ over Fe₂O₃ hollow nanospheres are solely attributed to inherent electronic conductivities of magnetite particles. Figure 6 shows the CVs of α-Fe₂O₃ and Fe₃O₄ in the potential window 0.005–3 V (vs. Li⁺/Li) at a scan rate of 3 mV/min. During the first cycle, a smooth sloping curve (cathodic) up to about 0.8 V with shoulders at 1.68 and 1.1 V is indicative of insertion of Li and reduction of Fe³⁺ in α-Fe₂O₃ in the shell domain of hollow nanospheres. The appearance of a strong oxidation peak at 1.6 V suggests the oxidation of Fe⁰ to Fe³⁺ ions. The second cathodic sweep differs from the first one; the shoulders at 1.68 and 1.1 V appeared in the first scan vanished in the second cycle. The cyclic voltammogram of Fe₃O₄ is slightly different from Fe₂O₃ (Fig. 6b). In the cathodic polarization process of the first cycle, a reduction peak appeared at 0.5 V corresponding to Li insertion and reduction of Fe³⁺ to metallic iron. However, the oxidation scanning exhibited two peaks at 1.55 and 2.14 V indicating the oxidation of Fe⁰ to Fe²⁺ and then to Fe³⁺. The second cathodic sweep exhibited an additional shoulder at 1.5 V, and essentially the peak positions are unaltered with subsequent cycles.

Fig. 5

Rate performance of α-Fe₂O₃ (a) and Fe₃O₄ (b) of hollow nanospheres at different rates in the voltage region of 0.005–3.0 V vs. Li/Li⁺

Fig. 6

Cyclic voltammogram of hollow nanospheres: a α-Fe₂O₃ and b Fe₃O₄ in 1.0 M LiPF₆ (EC/DMC = 1/2 (v/v)) at a 3-mV/min sweep rate

Conclusions

In summary, we have demonstrated a simple soft template method for fabrication of α-Fe₂O₃ and Fe₃O₄ hollow nanospheres of size 30 ± 2 nm with void space of about 14–15 nm. Fe₃O₄ hollow particles exhibited a magnetic moment of 22.5 emu/g as estimated by SQUID. The charge/discharge cycling performance and rate performance α-Fe₂O₃ hollow nanospheres were comparable to that of α-Fe₂O₃ nanotube and nanoflakes coated on conducting copper foil. Fe₃O₄ hollow nanospheres exhibited higher discharge capacity after 50 cycles (690 mAh g⁻¹) which is considerably higher than the carbon/Fe₃O₄ nanospindles (600 mAh g⁻¹) composite electrodes and commercial Fe₃O₄. The Fe₃O₄ nanospheres show 4.5 times higher discharge capacity than the commercial materials, and the improved rate performance is attributed partially to the unique hollow spherical morphology coupled with hollow void space. The void space not only acts as buffer medium against charge storage and local volume change but also provides better electrical contact and shorter diffusion path length and therefore provides better rate capability.