Nitrogen-doped carbon-coated Ti–Fe–O nanocomposites with enhanced reversible capacity and rate capability for high-performance lithium-ion batteries

Abstract

Rationally designed composite materials could allow each component to fully play a part in their superior performance. In this work, we designed and fabricated Ti–Fe–O nanocomposites coated with nitrogen-doped carbon to act as anode materials for lithium-ion batteries, in which Fe₃O₄ and FeTiO₃ components provide high capacity, a small fraction of TiO₂ improves cycling stability, and a nitrogen-doped carbon coating not only enhances electronic conductivity but also alleviates the agglomeration and accommodates volume change during cycling. The electrochemical performance is associated greatly with Ti/Fe molar ratio in the products, and the composite with a ratio of 1 : 2 exhibits the best cycling performance (reversible capacities of 426.4 mA h g⁻¹ at 100 mA g⁻¹ and 321.7 mA h g⁻¹ after cycling 500 times at 500 mA g⁻¹) as well as outstanding rate capabilities (374.1, 347.5, 315.9 and 260.3 mA h g⁻¹ at 200, 400, 800, and 1600 mA g⁻¹, respectively). The combined merits of its various components endow the nanocomposite with enhanced cycling and rate performance with respect to carbon-coated TiO₂ and Fe₃O₄/FeTiO₃ counterparts prepared under analogous conditions.

---

1. Introduction

Rechargeable lithium-ion batteries (LIBs), one of the great successes of modern material electrochemistry, are widely utilized in our daily lives such as in portable electronics, electric vehicles (EVs) and hybrid electric vehicles (HEVs) due to their unique features of high energy density, long cycle life, high operating voltage, and environmental friendliness in comparison with other kinds of battery systems.^1–3 In addition, LIBs also show promising applications in large scale energy storage systems for smart grids and the integration of renewable but intermittent energy resources, such as wind, solar, and wave energy, helping to relieve the ever increasing energy crisis and global warming.^4,5 However graphite, a commonly used anode material, cannot meet the requirements for next-generation high performance LIBs with high energy density and power density owing to its low theoretical capacity (372 mA h g⁻¹) and poor cycling performance at high current density. Therefore, numerous efforts have been devoted to exploring and designing alternative anode materials that could deliver energy more efficiently.

Among the promising anode materials, transition metal oxides (TMOs), e.g. iron and manganese oxides, have been intensively investigated due to their attractive advantages of low cost, natural abundance, non-toxicity, and high theoretical capacities (500–1200 mA h g⁻¹).^6,7 Nevertheless, the Li-ion storage of these oxides usually suffers from rapid capacity fading during charge/discharge cycling as a result of the large volume change and severe aggregation that results from the conversion reaction mechanism, as well as the intrinsically low electronic conductivity of the oxides (except Fe₃O₄). Up to date, some methods have been developed to address these issues, including (i) tailoring TMO morphology to form nanobelts,⁸ microspheres,⁹ nanorods,¹⁰ nanowires,^11,12 hollow spheres,^13–15 octahedra,¹⁶ and nanotubes^17,18 to enhance the contact area with electrolyte and shorten Li-ion diffusion path, and (ii) hybridizing nanoscale TMOs with carbonaceous supports (e.g. nanotubes, nanofibers and graphene)^19–24 or coating with a carbon layer^25–29 to buffer the large volume change, maintain structural integrity and effectively improve electronic conductivity. As has been proven, nanostructuring together with coating carbon could effectively enhance the cycling stability and rate capability of electrode materials for LIBs.^27–30

TiO₂ as a typical intercalation type anode material has also attracted more and more attention thanks to its low cost and environmental benignity, as well as the high safety and good cycling stability associated with the low volume change (3–4%) during Li-ion intercalation/deintercalation.^6,31 However, the relatively low theoretical capacity (168–335 mA h g⁻¹) and the poor rate capability that result from the low electronic and ionic conductivity restrict the practical application of TiO₂ in LIBs. Therefore, nanocomposites that combine the high electrochemical stability of TiO₂ with the high capacity of other TMOs incorporated in conductive carbon frameworks have attracted recent attention.^32–34 On one hand, electrochemically stable TiO₂ could buffer the volume change and act as a barrier to suppress the agglomeration of other TMOs, maintaining the mechanical integrity of electrode materials in the charge/discharge process. On the other hand, the conductive carbon framework could not only improve the electronic conductivity of TiO₂ and other TMOs but also alleviate the agglomeration and accommodate the volume change of active materials. As expected, nanocomposites exhibit enhanced electrochemical performance by virtue of the synergistic effect of different components. For example, Liu et al. synthesized nanostructured TiO₂-modified iron oxides on/among carbon nanotubes using a bottom-up assembly approach, and the composite displayed a capacity of 922 mA h g⁻¹ at 500 mA g⁻¹ and excellent rate capability.³³ Very recently, Pan et al. reported the co-assembly of TiO₂ nanorods and Fe₃O₄ nanoparticles on graphene nanosheets, and the hierarchical structure delivered reversible capacities of 703 mA h g⁻¹ at 500 mA g⁻¹ and 524 mA h g⁻¹ at 1000 mA g⁻¹ after 200 cycles.³⁴ However, based on the limited reports available, the preparation of TiO₂/Fe₃O₄/carbon composites is complicated by the use of some expensive raw materials, so it is necessary to explore some other routes to effectively fabricate the composites for practical LIB applications.

As has also been reported, composites of TMOs with N-doped carbon could improve the cycling performance and rate capability due to the enhancement in electronic conductivity, Li-ion diffusion, charge transfer at the interface and stability of solid electrolyte interphase (SEI) films.^27,35,36

In view of the above considerations, we prepared N-doped carbon-coated TiO₂/Fe₃O₄/FeTiO₃ multicomponent nanocomposites (Ti–Fe–O) with various Ti/Fe molar ratios using a simple hydrolysis precipitation method and a subsequent carbon coating process employing pyrrole as a carbon precursor. The combined merits of various components endow the nanocomposites with improved cycling and rate performance with respect to the carbon-coated TiO₂ and Fe₃O₄/FeTiO₃ counterparts prepared under the analogous conditions, revealing the promising potential of the nanocomposites as anode materials for high-performance LIBs.

2. Experimental section

2.1. Material fabrication

All the reagents employed were analytically pure and used without further purification. The preparation of Ti–Fe–O nanocomposites is similar and can be exemplified as follows with a Ti/Fe molar ratio of 2 : 1 (denoted as S2-1). Typically, 27.23 g of tetrabutyl titanate (TBT) dissolved in 50 mL of ethanol was titrated dropwise into a uniform mixture containing 7.20 g of FeC₂O₄·2H₂O and 20 mL of deionized water under magnetic stirring at ambient temperature, then the resultant mixture was dried thoroughly in an oven at 120 °C for 12 h. The dried product was calcined at 480 °C for 5 h in air. The as-sintered products with Ti/Fe molar ratios of 2 : 1, 1 : 1, 1 : 2, and 1 : 3 are designated as S2-1, S1-1, S1-2, and S1-3, respectively.

Carbon coating was conducted simply and quickly by putting 1.5 g of Ti–Fe–O composite into a stainless steel autoclave (30 mL in capacity) and uniformly dropping 0.55 mL of pyrrole into the composite using a syringe, followed by heating the tightly sealed autoclave at 550 °C for 5 h in a furnace. The black products are designated as S2-1C, S1-1C, S1-2C, and S1-3C. For comparison, carbon-coated TiO₂ (denoted as TiO₂@C) was also prepared using a similar procedure. The formation of carbon coating includes the polymerization of pyrrole uniformly around the nanoparticles and the thermal pyrolysis of polypyrrole to form an N-doped carbon layer.

2.2. Characterization

X-ray powder diffraction (XRD) patterns were collected at a scanning rate of 4° min⁻¹ (Rigaku Dmax-2500 diffractometer with Ni filtered Cu Kα radiation at V = 50 kV and I = 100 mA). The morphology and microstructure of the products were examined using a JEOL JEM-2100 high-resolution transmission electron microscope (HRTEM). Thermogravimetric analysis (TGA) was conducted from ambient temperature to 800 °C in air at a heating rate of 10 °C min⁻¹ to evaluate the carbon content (SDT thermal microbalance apparatus). Raman spectra of the products were acquired with excitation at 632.81 nm by an argon ion laser (Lab-RAM HR800 Raman spectrometer). X-ray photoelectron spectra (XPS) were recorded using Al Kα radiation (hν = 1486.6 eV) as the excitation source (KARTOS XSAM800 X-ray photoelectron spectrometer at V = 12 kV and I = 10 mA).

2.3. Electrochemical tests

The electrochemical performance of the carbon-coated Ti–Fe–O nanocomposites was tested using CR2025-type cells. The working electrodes are composed of the active materials acetylene black and polyvinylidene fluoride (PVDF) in a weight ratio of 8 : 1 : 1, which were mixed in N-methyl-2-pyrrolidinone (NMP). The homogeneous slurry was coated uniformly on a copper foil substrate, and dried in a vacuum oven at 120 °C for 12 h. The mass of active material loaded on each electrode (14 mm in diameter, punched from the copper foil) is ca. 3.0 mg. Metal Li foils were used as a counter electrode, Celgard 2300 used as a separator, and a mixture of 1 M LiPF₆ dissolved in ethylene carbonate and dimethyl carbonate with a volume ratio of 1 : 1 as an electrolyte. The assembly of half-cells was performed in an argon-filled glovebox at ambient temperature. The cell performance was evaluated galvanostatically at varied current densities in a potential range of 0.02–3.0 V (vs. Li/Li⁺) using a Land CT2001A battery test system at room temperature. An IviumStat electrochemistry workstation was adopted to acquire cyclic voltammograms (CV) at a scan rate of 0.3 mV s⁻¹ in 0.02–3.0 V vs. Li/Li⁺ and electrochemical impedance spectra (EIS) with an AC signal amplitude of 5 mV in a frequency range from 10⁵ to 10⁻² Hz.

3. Results and discussion

The XRD patterns of the as-sintered and carbon-coated products with various Ti/Fe molar ratios are depicted in Fig. 1. Before coating carbon (Fig. 1a), the diffraction peaks for S2-1 could be indexed to those from anatase TiO₂ (JCPDS 21-1272) and γ-Fe₂O₃ (JCPDS 39-1346). With decreasing the Ti/Fe molar ratio from 2 : 1 to 1 : 1 and 1 : 2, some weak peaks that result from α-Fe₂O₃ (JCPDS 33-0664) could also be identified in the patterns of S1-1 and S1-2 besides those from anatase TiO₂ and γ-Fe₂O₃ due to the increased Fe content in these two composites. After coating carbon (Fig. 1b), Fe₂O₃ was reduced to magnetite Fe₃O₄ (JCPDS 19-0629) accompanied by the formation of ilmenite FeTiO₃ (JCPDS 29-0733) owing to the reaction of TiO₂ with FeO reduced from Fe₂O₃ during carbonization at 550 °C. It is noted that the dominate phase is FeTiO₃ in S2-1C, and Fe₃O₄ in S1-1C and S1-2C owing to the decreased Ti/Fe molar ratio. Furthermore, no diffractions from carbon could be distinguished in the patterns because of its amorphous features. From the XRD patterns in Fig. 1a, it is difficult to determine the presence of γ-Fe₂O₃ since γ-Fe₂O₃ shows similar peaks to Fe₃O₄ as shown in Fig. 1b. Thus, Raman spectra of the as-sintered products were further recorded (Fig. S1 in the ESI†). The peaks at around 225, 245, 290, 412 and 612 cm⁻¹ result from α-Fe₂O₃ and the one at around 514 cm⁻¹ from TiO₂.^10,37 Meanwhile, the peak at about 660 cm⁻¹ might be related to Fe₃O₄. However, γ-Fe₂O₃ also exhibits a broad peak around 660 cm⁻¹ according to the literature.^10,38,39 There are some disagreements in the reported data, so it is still difficult to determine the presence of γ-Fe₂O₃ or Fe₃O₄ from the Raman spectra. Nevertheless, considering the fact that the samples before carbon coating were calcined at 480 °C for 5 h in air and that the as-sintered products are reddish brown in color (photographs in Fig. S2†), it could be determined that the component is γ-Fe₂O₃ rather than Fe₃O₄ in the as-sintered products. It is noted that the peaks of α-Fe₂O₃ are also present in S2-1, which could be ascribed to laser degradation of γ-Fe₂O₃ during Raman tests. As has been reported,^40–42 FeTiO₃ is also electrochemically active for Li-ion storage with a theoretical capacity of 530 mA h g⁻¹. The coexistence of TiO₂, Fe₃O₄ and FeTiO₃ in the carbon-coated composites might help to adjust the Li-ion storage performance.

Fig. 1 XRD patterns of (a) the as-sintered products and (b) the carbon-coated products.

The morphology and microstructure of the carbon-coated products were examined by TEM (Fig. 2). The product S2-1C exhibits severe agglomeration with the particles ranging from 25 to 130 nm in size (Fig. 2a), and around the particles is a thin amorphous carbon layer of about 1 nm in thickness (Fig. 2d). The lattice fringe spacings of 0.35, 0.30 and 0.25 nm in Fig. 2d correspond to the (101) plane of anatase TiO₂, (220) plane of Fe₃O₄ and (110) plane of FeTiO₃, respectively. The particle agglomeration in S1-1C is weaker and the particle size is in the range of 20–100 nm (Fig. 2b). The particles in S1-2C exhibit better dispersivity and smaller particle sizes of 15–70 nm than those in the other two products (Fig. 2c). In other words, the particle size decreases and the dispersion becomes more homogeneous upon reducing the Ti/Fe molar ratio from 2 : 1 to 1 : 1 and 1 : 2, which is favorable for improving the electrochemical performance of the Ti–Fe–O composites. From the high resolution images, the carbon layer is about 1.2 nm for S1-1C (Fig. 2e) and 3.4 nm for S1-2C (Fig. 2f). As well as the lattice fringe spacing of 0.35 nm corresponding to the (101) plane of anatase TiO₂, the spacings of 0.25 nm in Fig. 2e and 0.21 nm in Fig. 2f are consistent with the (311) and (400) planes of Fe₃O₄, respectively. Lattice fringes of FeTiO₃ are not present in the selected zones for S1-1C and S1-2C owing to its lower content than TiO₂ and Fe₃O₄ in the composites.

Fig. 2 TEM images of (a, d) S2-1C, (b, e) S1-1C and (c, f) S1-2C.

Raman spectra were recorded to evaluate the nature of carbon in the carbon-coated Ti–Fe–O nanocomposites (Fig. 3a). The two broad peaks at around 1380 and 1605 cm⁻¹ are assigned to the D and G bands of carbon materials, respectively, and the former is related to disordered carbon, while the latter to ordered graphitic carbon.^25–27 The relative intensity ratio of the two bands (I_D/I_G) is generally used to assess the graphitization degree of carbon materials. The I_D/I_G ratio of approximately 0.7 demonstrates the amorphous nature of the carbon coating, consistent with the XRD and TEM results. The carbon content in the carbon-coated products determined by TGA is about 8.2 wt% for S2-1C, 8.3 wt% for S1-1C and 11.1 wt% for S1-2C (Fig. S3†).

Fig. 3 (a) Raman spectra of the carbon-coated products, XPS spectra of S1-2C for the survey spectrum (b), and core level spectra for C 1s (c) and (d) N 1s.

More information concerned with surface elements and their bonding was obtained from XPS spectra (Fig. 3). Taking S1-2C as an example, the survey spectrum in Fig. 3b verifies the presence of the elements C, N, Ti, Fe and O in the composites. In Fig. 3c, the four peaks deconvoluted from the core level spectrum of C 1 s result from the sp²–sp²C (284.7 eV), N–sp²C (285.9 eV), N–sp³C (287.2 eV), and C–O (289.0 eV) bonds.^27,43,44 The two peaks centered at ca. 400.3 and 398.4 eV for the N 1s region (Fig. 3d) belong to pyrrolic and pyridinic N,^{27,35,43–45} respectively, and the calculated atomic content of nitrogen is 5.27 at% for S1-2C (4.89 at% for S2-1C and 4.91 at% for S1-1C from the corresponding XPS spectra, not presented here). The aforementioned XPS results confirm the formation of the N-doped carbon coating when using pyrrole as the carbon source, which is beneficial for improving the cycling stability and rate capability of anode materials.^27,35,36 The O 1s peak (Fig. S4†) could be deconvoluted into four peaks corresponding to Ti–O–Ti (529.6 eV), Fe–O–Fe (530.1 eV), Fe–O–Ti (531.2 eV), and C–O (532.7 eV) type bonds,^42,46,47 and the integral area of each peak could be calculated to determine the relative amount of each component (TiO₂, Fe₃O₄ and FeTiO₃) in the composite by conducting conversion from the atomic ratio of oxygen (O_Ti–O–Ti : O_Fe–O–Fe : O_Fe–O–Ti) to the molar ratio of TiO₂, Fe₃O₄ and FeTiO₃ (1/2O_Ti–O–Ti : 1/4O_Fe–O–Fe : 1/3O_Fe–O–Ti) and the mass ratio of TiO₂, Fe₃O₄ and FeTiO₃ (1/2O_Ti–O–Ti molecular weight of TiO₂ : 1/4O_Fe–O–Fe × molecular weight of Fe₃O₄ : 1/3O_Fe–O–Ti × molecular weight of FeTiO₃). Combined with the carbon content acquired from the TGA results, the mass fractions of TiO₂, Fe₃O₄, and FeTiO₃ in S1-2C could be approximately calculated. Similarly, the mass percentage of the components in other samples could also be estimated, as summarized in Table S1.† Obviously, the content of TiO₂ decreases while that of Fe₃O₄ increases with decreasing the Ti/Fe molar ratio of the composite, in good agreement with the XRD results in Fig. 1b.

The cycling performance of the carbon-coated products was tested in the potential range of 0.02–3.0 V (vs. Li/Li⁺), as shown in Fig. 4. At a current density of 100 mA g⁻¹ (Fig. 4a), the reversible capacities in the initial few cycles for S2-1C and S1-1C deteriorate rapidly compared to that of S1-2C due to the larger particle size and less particle dispersivity of S2-1C and S1-1C than S1-2C. In addition, the carbon layers in S2-1C and S1-1C are thinner than that in S1-2C, so it is not strong enough to buffer the severe volume change that results from Fe₃O₄ and FeTiO₃ during charge/discharge cycling. This may also be a reason for the severe capacity fading in the first few cycles. From the sixth cycle, the reversible capacities increase gradually with cycling, which is ascribed to the activation effect that commonly occurs in TMOs.^{33–35,42,45,48,49} After 60 cycles, the reversible capacities are 300.7 mA h g⁻¹ for S2-1C, 354.5 mA h g⁻¹ for S1-1C and 426.4 mA h g⁻¹ for S1-2C, much higher than that of TiO₂@C (197.7 mA h g⁻¹ after cycling 60 times at 100 mA g⁻¹, Fig. S5†). From Fig. 4b, apart from the low initial coulombic efficiencies of 61.2% for S2-1C, 50.8% for S1-1C and 60.9% for S1-2C due to the formation of SEI films and electrolyte decomposition,^34,35,49 the coulombic efficiency increases rapidly to higher than 98% from the sixth cycle, suggesting the improved reversibility of the electrochemical reactions that occur in the electrodes.

Fig. 4 (a) Cycling performance and (b) coulombic efficiency at 100 mA g⁻¹ for S2-1C, S1-1C and S1-2C, (c) cycling performance at 100 mA g⁻¹ and (d) XRD pattern for S1-3C.

From Fig. 4a, the reversible capacity increases upon decreasing the Ti/Fe molar ratio from 2 : 1 to 1 : 1 and 1 : 2, because the Fe₃O₄ content in the composites increases gradually (Table S1†), which delivers a higher theoretical capacity (926 mA h g⁻¹) than TiO₂ and FeTiO₃. However, when the Ti/Fe molar ratio is decreased to 1 : 3 (namely S1-3C), the reversible capacity fades rapidly to below 200 mA h g⁻¹ after the tenth cycle and retains a capacity of only 119.1 mA h g⁻¹ after 60 cycles (Fig. 4c), which is much lower even than that of TiO₂@C. From the XRD pattern of S1-3C (Fig. 4d), only Fe₃O₄ and FeTiO₃ can be detected, verifying the significance of TiO₂ in improving the cycling stability of the carbon-coated Ti–Fe–O nanocomposites.

Rate capability is another very important consideration for practical LIBs. Fig. 5a reveals the rate performance of S2-1C, S1-1C, and S1-2C at varied current densities (each for 10 cycles), and the detailed data are summarized in Table 1. For convenience of comparison, capacity retention (defined as the ratio of the reversible capacity at a certain current density to that at the initial 100 mA g⁻¹) is also given in Table 1 following the capacity. Also S1-2C exhibits the best rate capability, delivering a high reversible capacity of 260.3 mA h g⁻¹ with a capacity retention of 64.3% even when cycled at 1600 mA g⁻¹, higher than that of S2-1C (196.1 mA h g⁻¹/51.2%) and S1-1C (131.4 mA h g⁻¹/47.5%). When the current density was recovered to 100 mA g⁻¹ after the rate performance test, all the samples reached higher reversible capacities than those at the initial 100 mA g⁻¹ due to high rate lithiation-induced reactivation.^46,50 After the rate performance tests, long-term cycling performance was continuously measured at a high current density of 500 mA g⁻¹ (Fig. 5b). After cycling 500 times, a reversible capacity of 318.2 mA h g⁻¹ was retained for S1-2C, much higher than those of S2-1C (112.0 mA h g⁻¹) and S1-1C (213.9 mA h g⁻¹), further conforming the superior cycling stability of S1-2C even at a high current density. Based on the above results, the Ti/Fe molar ratio has significant impact on the electrochemical performance of the carbon-coated Ti–Fe–O nanocomposites, and the composite with a ratio of 1 : 2 exhibits the highest reversible capacity, outstanding rate performance, and superior high rate long-term cycle life. The remarkable comprehensive performance of S1-2C is mainly ascribed to the synergistic effect given by the stable structure of TiO₂, the high capacity of Fe₃O₄/FeTiO₃ and the excellent electronic conductivity of the N-doped carbon coating. Thus the composite could serve as a promising high-performance anode material for LIBs.

Fig. 5 (a) Rate performance at varied current densities and (b) long-term cycling performance at 500 mA g⁻¹ after rate performance tests for S2-1C, S1-1C and S1-2C.

Table 1 Reversible capacities (mAh g⁻¹) and capacity retention (%) of S2-1C, S1-1C, and S1-2C at various current densities (mA g⁻¹)

Samples	Capacities/capacity retention at varied current densities
	100	200	400	800	1600	100 (last)
S2-1C	382.8/100	354.2/92.5	315.7/82.5	265.4/69.3	196.1/51.2	414.8/108.4
S1-1C	276.7/100	238.4/86.2	211.9/76.6	175.6/63.5	131.4/47.5	324.6/117.3
S1-2C	404.7/100	374.1/92.4	347.5/85.9	315.9/78.1	260.3/64.3	468.0/115.6

---

The electrochemical reactions that occur in the electrodes were distinguished from the charge/discharge plots for the first cycle at 100 mA g⁻¹ (Fig. 6a). In the discharge process, the three plateaus at around 1.7, 1.2 and 0.8 V are attributed to Li⁺ insertion into TiO₂,^31,42 lithiation of FeTiO₃^40–42 and reduction of Fe₃O₄ to Fe⁰,^33,34 respectively. During charging, the potential slope between 1.5 and 2.2 V corresponds to Li-ion extraction from the active materials, which is distinct for S2-1C and S1-2C but inconspicuous for S1-1C due to the severe inhomogeneous reactions between Li⁺ and the electrode materials in S1-1C.⁴⁶ Therefore, S1-1C delivers a lower initial charge capacity of 313.2 mA h g⁻¹ than both S2-1C (397.4 mA h g⁻¹) and S1-2C (373.0 mA h g⁻¹) and also a lower initial coulombic efficiency (Fig. 4b).

Fig. 6 (a) Charge/discharge plots for the first cycle at 100 mA g⁻¹ and discharge plots of the first cycle at various current densities for (b) S2-1C, (c) S1-1C and (d) S1-2C.

To clearly reveal the rate capability of S1-2C, S2-1C and S1-1C, the discharge plots for the first cycle at various current densities are also depicted in Fig. 6b. Upon increasing the current density, the discharge capacity is gradually decreased. For S1-1C (Fig. 6c), the potential plateaus become more and more invisible with increasing the current density, demonstrating the sluggish electrochemical reaction kinetics at high rates. In contrast, S2-1C (Fig. 6b) and S1-2C (Fig. 6d) demonstrate superior rate capability with distinct plateaus even at high rates. The information from the discharge plots is in good consistency with the rate performance shown in Fig. 5a.

CV profiles are conducive to effectively understanding electrochemical reactions (Fig. 7). Particularly, the role of TiO₂ could be distinguished by comparing the CV profiles of S1-2C (Fig. 7a) with those of S1-3C (Fig. 7b). In the first cathodic process, both samples show a strong peak at ca. 0.4 V thanks to the reduction of Fe₃O₄ to Fe⁰ accompanied with the formation of SEI films.^51,52 As reported in the literature,^40–42 the reaction between Li⁺ and FeTiO₃ (FeTiO₃ + (2 + x)Li⁺ + (2 + x)e⁻ → Fe⁰ + Li_xTiO₂ + Li₂O) occurs at ca. 1.2 and 0.4 V in the first discharge process, thus the cathodic peak at ca. 0.4 V might also be related to the lithiation of FeTiO₃. For S1-3C, besides the reduction peak at ca. 0.4 V, the other one at around 1.2 V is related to the high FeTiO₃ content in the composite. In the first anodic scan for S1-3C, there are two well defined peaks at ca. 1.8 and 2.0 V, corresponding to the oxidation of Fe⁰ to Fe²⁺/Fe³⁺, and the delithiation of Li_xTiO₂, respectively.^{40–42,51,52} For S1-2C, the broad anodic peak (1.4–2.1 V) centered at ca. 1.8 V results from the superposed Li⁺ extraction reactions. It is noted that the characteristic cathodic/anodic peaks related to the lithiation/delithiation of anatase TiO₂ (at ca. 1.7/2.0 V) are almost negligible for S1-2C, suggesting that TiO₂ contributes less to reversible capacity than Fe₃O₄ and FeTiO₃ because of its relatively low content and theoretical capacity. From the second cycle, both the cathodic and anodic peaks for S1-2C and S1-3C shift to higher potentials owing to the structural modification of Fe₃O₄ and FeTiO₃ in the first cycle induced by the formation of SEI films and irreversible Li₂O.^{22,40–42,51,52} The cathodic peak for S1-2C shifts to 0.8 V while the anodic peak shifts to about 2.1 V. However, the cathodic/anodic peaks for S1-3C widen and become inconspicuous from the second cycle, appearing at 1.3–0.2 V and 1.7–3.0 V, respectively. The potential difference between the cathodic and anodic peaks for S1-2C is about 1.3 V, lower than that for S1-3C (about 1.6 V), demonstrating the weaker polarization in S1-2C than in S1-3C, i.e. the better reversibility and faster Li⁺ diffusion in S1-2C.⁵³ The CV results validate the vital role of TiO₂ in maintaining the structural stability of the composite and improving the electrochemical performance of the Ti–Fe–O nanocomposites. Accordingly, even a small fraction of structurally stable TiO₂ could enhance the Li-ion storage performance of S1-2C to exceed that of S1-3C.

Fig. 7 CV profiles at a scan rate of 0.3 mV s⁻¹ between 0.02 and 3.0 V vs. Li/Li⁺ for (a) S1-2C and (b) S1-3C.

To further elucidate the difference in electrochemical performance, EIS spectra of charged cells after cycling 60 times at 100 mA g⁻¹ were measured. From the typical Nyquist plots shown in Fig. 8, the depressed semicircle in the high-medium frequency region reflects the resistance of SEI films (R_sf) and charge transfer (R_ct), and the inclined line in the low frequency region denotes Li⁺ diffusion in the electrode materials, i.e. the Warburg impedance (Z_w).^46,51,52 The semicircle diameters of S2-1C, S1-1C, and S1-2C are much smaller than that of S1-3C, implying that the resistances of S2-1C, S1-1C, and S1-2C are much lower than that of S1-3C. Meanwhile, the S2-1C, S1-1C and S1-2C electrodes exhibit a similar line slope due to the presence of TiO₂, and the slopes are all larger than that of S1-3C without TiO₂. The impedance spectra were fitted based on a reasonable equivalent circuit (Fig. S6,† where R_e is the electrolyte resistance and CPE is the constant phase element to replace pure capacitance) to estimate the resistance values. The total R_(sf+ct) values for S2-1C (∼232 Ω), S1-1C (∼176 Ω) and S1-2C (∼83 Ω) in the presence of TiO₂ are much lower than that for S1-3C (∼587 Ω) in the absence of TiO₂. The R_ct value is about 98 Ω for S2-1C, 37 Ω for S1-1C, 13 Ω for S1-2C and 177 Ω for S1-3C, exhibiting a similar variation tendency as the R_(sf+ct) values. It is noted that the R_(sf+ct) and R_ct values decrease with increasing the Fe content in the composites, i.e. with decreasing the Ti/Fe molar ratios from 2 : 1 to 1 : 1 and 1 : 2, likely associated with the incremental Fe₃O₄ content with high electronic conductivity in the composites.^28,54–57 However, in spite of the high Fe₃O₄ content in S1-3C with a higher Ti/Fe molar ratio of 1 : 3, the R_(sf+ct) and R_ct increase suddenly and exceed those of the other samples. The main reason is that S1-3C contains no TiO₂ to maintain the structural stability of the electrode materials, and the Fe₃O₄ and FeTiO₃ with high specific capacities in S1-3C will result in large volume expansion and pulverization of particles during discharging/charging, leading to the increase in contact resistance. The EIS results further confirm the important role of TiO₂ in improving the structural stability and in promoting Li⁺ diffusion of the composites. Apparently, S1-2C with a Ti/Fe molar ratio of 1 : 2 possesses the lowest resistance and fast Li⁺ diffusion among the samples with various Ti/Fe molar ratios, and thus achieves the optimal performance including excellent cycling stability and superior rate capability.

Fig. 8 EIS of the charged cells after cycling 60 times at 100 mA g⁻¹ for the carbon-coated products. The inset is the enlarged image in the high frequency region.

Based on the above analysis, the excellent Li-ion storage performance of an N-doped carbon-coated Ti–Fe–O nanocomposite with a moderate Ti/Fe molar ratio of 1 : 2 (namely S1-2C) could be attributed to the synergistic combination of the performance of various components. Firstly, the outer N-doped carbon coating around the nanoparticles could not only act as an effective conductor for fast charge transfer and Li-ion diffusion but also as a buffering network to alleviate agglomeration and volume change of active materials during cycling. Secondly, Fe₃O₄/FeTiO₃ with high capacity could compensate for the low theoretical capacity of TiO₂, endowing the composite with a high reversible capacity. Thirdly, the electrochemically stable TiO₂ could buffer the volume change and behave as a barrier to inhibit the agglomeration of Fe₃O₄/FeTiO₃, maintaining the structural stability of the electrode materials during repeated charge/discharge processes. It should be noted that a relatively high TiO₂ content (say S2-1C and S1-1C) or the absence TiO₂ (say S1-3C) does not play a part in composite performance.

4. Conclusions

In summary, N-doped carbon-coated Ti–Fe–O multicomponent nanocomposites with various Ti/Fe molar ratios were fabricated simply. A small fraction of TiO₂ in the composite is critical for enhancing the cycling stability and rate performance. The outstanding rate capabilities and long-term cycling performance at 500 mA g⁻¹ along with the simple and effective fabrication process allow the Ti–Fe–O nanocomposite with a moderate Ti/Fe molar ratio of 1 : 2 to be applicable in high performance LIBs.

Acknowledgements

This work was supported by the Shandong Provincial Natural Science Foundation, PR China (ZR2015EM016) and the Science and Technology Development Project of Shandong Province, PR China (2014GGX102009 and 2015GGX102005).

References

1. A. S. Arico, P. Bruce, B. Scrosati, J. M. Tarascon and W. V. Schalkwijk, Nat. Mater., 2005, 4, 366.
2. J. M. Tarascon and M. Armand, Nature, 2001, 414, 359.
3. J. Lee, A. Urban, X. Li, D. Su, G. Hautier and G. Ceder, Science, 2014, 343, 519.
4. B. Dunn, H. Kamath and J. M. Tarascon, Science, 2011, 334, 928.
5. J. B. Goodenough and Y. Kim, Chem. Mater., 2010, 22, 587.
6. S. Goriparti, E. Miele, F. D. Angelis, E. D. Fabrizio, R. P. Zaccaria and C. Capiglia, J. Power Sources, 2014, 257, 421.
7. V. Aravindan, Y. S. Lee and S. Madhavi, Adv. Energy Mater., 2015, 5, 1402225.
8. Y. Shi, M. Shi, Y. Qiao, J. Tu and H. Chen, Nanotechnology, 2012, 23, 395601.
9. J. Xu and Y. Zhu, ACS Appl. Mater. Interfaces, 2012, 4, 4752.
10. C. T. Cherian, J. Sundaramurthy, M. Kalaivani, P. Ragupathy, P. Suresh Kumar, V. Thavasi, M. V. Reddy, C. H. Sow, S. G. Mhaisalkar, S. Ramakrishna and B. V. R. Chowdari, J. Mater. Chem., 2012, 22, 12198.
11. D. Su, H. J. Ahn and G. Wang, J. Power Sources, 2013, 244, 742.
12. H. Wu, M. Xu, Y. Wang and G. Zheng, Nano Res., 2013, 6, 167.
13. F. Ma, H. Hu, H. Wu, C. Xu, Z. Xu, L. Zhen and X. Lou, Adv. Mater., 2015, 27, 4097.
14. B. Wang, J. Chen, H. Wu, Z. Wang and X. Lou, J. Am. Chem. Soc., 2011, 133, 17146.
15. J. Zhu, Z. Yin, D. Yang, T. Sun, H. Yu, H. E. Hoster, H. H. Hng, H. Zhang and Q. Yan, Energy Environ. Sci., 2013, 6, 987.
16. H. Cao, X. Wang, H. Gu, J. Liu, L. Luan, W. Liu, Y. Wang and Z. Guo, RSC Adv., 2015, 5, 34566.
17. X. Lou, D. Deng, J. Y. Lee, J. Feng and L. A. Archer, Adv. Mater., 2008, 20, 258.
18. L. Li, C. Nan, J. Lu, Q. Peng and Y. Li, Chem. Commun., 2012, 48, 6945.
19. Y. Park, M. Oh, J. S. Park, S. H. Baek, M. Kim, S. Kim and J. H. Kim, Carbon, 2015, 94, 9.
20. G. Zhou, D. Wang, P. Hou, W. Li, N. Li, C. Liu, F. Li and H. Cheng, J. Mater. Chem., 2012, 22, 17942.
21. Y. Wu, P. Zhu, M. V. Reddy, B. V. R. Chowdari and S. Ramakrishna, ACS Appl. Mater. Interfaces, 2014, 6, 1951.
22. Y. Wan, Z. Yang, G. Xiong, R. Guo, Z. Liu and H. Luo, J. Power Sources, 2015, 294, 414.
23. W. Xiao, Z. Wang, H. Guo, X. Li, J. Wang, S. Huang and L. Gan, Appl. Surf. Sci., 2013, 266, 148.
24. J. Su, M. Cao, L. Ren and C. Hu, J. Phys. Chem. C, 2011, 115, 14469.
25. T. Muraliganth, A. V. Murugan and A. Manthiram, Chem. Commun., 2009, 7360.
26. F. Wu, X. Ma, J. Feng, Y. Qian and S. Xiong, J. Mater. Chem. A, 2014, 2, 11597.
27. J. Wang, C. Zhang and F. Kang, ACS Appl. Mater. Interfaces, 2015, 7, 9185.
28. W. Zhang, X. Wu, J. Hu, Y. Guo and L. Wan, Adv. Funct. Mater., 2008, 18, 3941.
29. Y. Luo, X. Zhou, Y. Zhong, M. Yang, J. Wei and Z. Zhou, Electrochim. Acta, 2015, 154, 136.
30. H. Li and H. Zhou, Chem. Commun., 2012, 48, 1201.
31. X. Bai, T. Li, Y. Qi, X. Gao, L. Yin, H. Li, H. Zhu, N. Lun and Y. Bai, Electrochim. Acta, 2015, 169, 241.
32. J. Liao, D. Higgins, G. Lui, V. Chabot, X. Xiao and Z. Chen, Nano Lett., 2013, 13, 5467.
33. J. Liu, D. Qian, H. Feng, J. Li, J. Jiang, S. Peng and Y. Liu, J. Mater. Chem. A, 2014, 2, 11372.
34. L. Pan, X. Zhu, X. Xie and Y. Liu, Adv. Funct. Mater., 2015, 25, 3341.
35. C. Lei, F. Han, D. Li, W. Li, Q. Sun, X. Zhang and A. Lu, Nanoscale, 2013, 5, 1168.
36. D. Cai, D. Li, S. Wang, X. Zhu, W. Yang, S. Zhang and H. Wang, J. Alloys Compd., 2013, 561, 54.
37. D. Pham-Cong, J. Kim, S. Jeong, J. Choi, J. Kim and C. Cho, Electrochem. Commun., 2015, 60, 204.
38. S. J. Oh, D. C. Cook and H. E. Townsend, Hyperfine Interact., 1998, 112, 59.
39. D. L. A. de Faria, S. Venancio Silva and M. T. de Oliveira, J. Raman Spectrosc., 1997, 28, 873.
40. X. Guan, J. Zheng, M. Zhao, L. Li and G. Li, RSC Adv., 2013, 3, 13635.
41. T. Tao, A. M. Glushenkov, M. M. Rahman and Y. Chen, Electrochim. Acta, 2013, 108, 127.
42. S. Guo, J. Liu, S. Qiu, W. Liu, Y. Wang, N. Wu, J. Guo and Z. Guo, Electrochim. Acta, 2016, 190, 556.
43. G. Lu, S. Qiu, H. Lv, Y. Fu, J. Liu, X. Li and Y. Bai, Electrochim. Acta, 2014, 146, 249.
44. S. Qiu, X. Wang, G. Lu, J. Liu and C. He, Mater. Lett., 2014, 136, 289.
45. T. Li, Y. Wang, R. Tang, Y. Qi, N. Lun, Y. Bai and R. Fan, ACS Appl. Mater. Interfaces, 2013, 5, 9470.
46. S. Guo, J. Liu, S. Qiu, W. Liu, Y. Wang, N. Wu, J. Guo and Z. Guo, J. Mater. Chem. A, 2015, 3, 23895.
47. Z. Geng, Y. Lin, X. Yu, Q. Shen, L. Ma, Z. Li, N. Pan and X. Wang, J. Mater. Chem., 2012, 22, 3527.
48. X. Chen, H. Lin, X. Zheng, X. Cai, P. Xia, Y. Zhu, X. Li and W. Li, J. Mater. Chem. A, 2015, 3, 18198.
49. Y. Jiang, Z. Jiang, L. Yang, S. Cheng and M. Liu, J. Mater. Chem. A, 2015, 3, 11847.
50. H. Sun, G. Xin, T. Hu, M. Yu, D. Shao, X. Sun and J. Lian, Nat. Commun., 2014, 5, 4526.
51. J. Zhao, S. Zhang, W. Liu, Z. Du and H. Fang, Electrochim. Acta, 2014, 121, 428.
52. X. Fan, J. Shao, X. Xiao, L. Chen, X. Wang, S. Li and H. Ge, J. Mater. Chem. A, 2014, 2, 14641.
53. M. Madian, L. Giebeler, M. Klose, T. Jaumann, M. Uhlemann, A. Gebert, S. Oswald, N. Ismail, A. Eychmuller and J. Eckert, ACS Sustainable Chem. Eng., 2015, 3, 909.
54. P. L. Taberna, S. Mitra, P. Poizot, P. Simon and J. M. Tarascon, Nat. Mater., 2006, 5, 567.
55. P. Lian, X. Zhu, H. Xiang, Z. Li, W. Yang and H. Wang, Electrochim. Acta, 2010, 56, 834.
56. P. Lv, H. Zhao, Z. Zeng, J. Wang, T. Zhang and X. Li, J. Power Sources, 2014, 259, 92.
57. H. Wang, G. Wang, S. Yuan, D. Ma, Y. Li and Y. Zhang, Nano Res., 2015, 8, 1659.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra10682j

---