Synthesis of high performance Li₄Ti₅O₁₂ microspheres and TiO₂ nanowires from natural ilmenite

Abstract

The low cost production of Ti-containing nanomaterials from natural resources is an attractive method for producing materials suitable for the growing energy storage industry. In this work, natural ilmenite is transformed into either Li₄Ti₅O₁₂ microspheres or TiO₂ nanowires through a series of chemical and thermal processes by crossing metallurgy and material science. The produced well-crystallized Li₄Ti₅O₁₂ microspheres are composed of nanosized particles that show high cycle stability, coulombic efficiencies, and rate capability for C-rates as high as 20 C. The initial discharge capacities are 173.1, 168.5, 167.2, 163.4, 160.0, 155.4 and 142.7 mA h g⁻¹ at the 0.1, 0.5, 1, 2, 5, 10 and 20 C-rates, respectively. After 100 cycles, the as-prepared Li₄Ti₅O₁₂ retains 99.1%, 98.9%, 96.8% and 91.8% of its initial discharge capacities at the 2, 5, 10 and 20 C-rates, respectively. The synthesized echinus-like anatase TiO₂ nanowire aggregations also demonstrate excellent cycling performance, even at high current densities.

---

1. Introduction

Nowadays, the development of human society is closely related to energy resources and the environment. Social development needs the consumption of energy and resources. Therefore, with the increasing environmental problems caused by conventional energy sources and the gradual depletion of global energy and metallic ore resources, the development of new energy and comprehensive utilization of mineral resources have become the focus of worldwide attention. Electrochemical energy systems, such as batteries and supercapacitors, have garnered significant attention due to their high energy or high power characteristics, often available in small formats. Rechargeable Li-ion batteries are of particular interest for a number of applications ranging from portable electronics to transportation due to their high energy density, high voltage, and ambient temperature operation.^1–5 However, improved performance of current Li-ion batteries is required to meet higher energy and power requirements.

Graphite is the current industry anode material for Li-ion batteries; however, it has limitations. First, the theoretical capacity is lower than many other materials, which can intercalate or alloy with Li. Second, the low potential for lithium intercalation can lead to the lithium plating during charging. Li commonly plates as dendrites, which have significant safety concerns. Another issue originates from the charge consumption required to form the solid electrolyte interphase (SEI) layer that is essential to electrode stability in terms of the capacity, power retention, and cycle life.^6–10

Spinel lithium titanate (Li₄Ti₅O₁₂) and anatase titanium dioxide (TiO₂) have been viewed as two promising alternative anode materials to graphite in Li-ion batteries.^11–15 Spinel Li₄Ti₅O₁₂ has a theoretical capacity of 175 mA h g⁻¹ with high deintercalation reversibility.^16–18 Due to the flat Li⁺ insertion potential of ∼1.55 V for Li₄Ti₅O₁₂, above the reduction potential of electrolyte solvents, a SEI film will not form on Li₄Ti₅O₁₂, which may allow for a reduced irreversible capacity. Further, as a “zero-strain” insertion material, it has an excellent cycling performance. In comparison, anatase TiO₂ is a fast Li insertion/extraction host with an insertion potential ∼2.0 V, low volume expansion (3–4%) during lithium insertion, and a high theoretical capacity of 336 mA h g⁻¹.^19,20 These features make them two promising anode materials for large-scale long-life energy storage batteries. However, both of them are still far away from being perfect because of their low electronic conductivity, which limits their rate stability.

Nanostructured TiO₂ and Li₄Ti₅O₁₂ are conventionally prepared from TiO₂ or Ti-based organic compounds. Generally, these highly pure raw compounds are prepared from Ti-containing ores or other Ti slag via complex, expensive, energy consuming processes that remove impurities. In this paper, new methods to synthesize nanostructured TiO₂ and Li₄Ti₅O₁₂ from natural ilmenite (FeTiO₃) through a series of easily replicated processes are proposed: TiO₂ nanowire synthesis by a high temperature hydrothermal method and Li₄Ti₅O₁₂ microspheres synthesis by spray drying. Natural ilmenite (FeTiO₃) is one of the primary global sources of titanium dioxide. In China, commercially, pigment grade titanium dioxide is mainly produced by complex sulfate processes with the generation of a large amount of waste acidic iron(II) solutions and FeSO₄·7H₂O waste slag, which causes not only severe environmental problems but also a waste of iron resources. Finally, the obtained pigment grade titanium dioxide is often used to prepare high purity nanostructured TiO₂ and organic Ti-contained compounds by using complex metallurgical extraction and materials preparation processes, which are extensively used in coatings, paper, porcelain, plastics, catalysts, cosmetics, gas sensors and energy.

In previous publications, we used hydrochloric acid as a leachant to dissolve mechanically activated ilmenite to obtain a lixivium (or filtrate), which was used to prepare LiFePO₄.^21,22 In this process, another product of the hydrolyzed titania residue was used to prepare Li₄Ti₅O₁₂ directly after two-step alkaline leaching.²³ The obtained Li₄Ti₅O₁₂ particles were not uniform and the electrochemical performance was very poor.²³ The obtained hydrolyzed titania residue was hard to use by the chlorination process. Owing to its high chemical activity and reaction with H₂O₂ and NH₃·H₂O, Ti could be easily and selectively extracted from the high Ti-containing residue by alkaline-hydrogen peroxide leaching.²⁴ The lixivium from the hydrolyzed titania residue was used to synthesize TiO₂ nanosheets from which Li₄Ti₅O₁₂ was prepared; however, the performance was poor and this method consumed significant LiOH.²⁵

In this work, we combine the metallurgical extraction and materials preparation together to develop a series of short processes to synthesize Li-ion battery anode materials TiO₂ and Li₄Ti₅O₁₂ from natural ilmenite. In this method, TiO₂ does not need to be prepared from the lixivium of the ammonia titanium peroxide solution, as in previous work.²⁵ Here, only LiOH was added into the lixivium to obtain a lithium (ammonia) titanium peroxide solution which was used to prepare Li₄Ti₅O₁₂ microspheres. Moreover, the addition of NaOH to the lixivium of the ammonia titanium peroxide solution has a profound significance of not only decreasing the content of SiO₂ in the TiO₂, but also forming uniform TiO₂ nanowires.

The major element of Ti in ilmenite is utilized to prepare high performance TiO₂ and Li₄Ti₅O₁₂, which can be used in Li-ion batteries. The produced Li₄Ti₅O₁₂ microspheres show a very good performance, even better than our results using analytically pure and organic titanium compounds.^26,27 The produced TiO₂ nanowires also show a good performance and high purity; thus, through these methods, stable Li-ion materials can be synthesized using processes that reduce redundancy, reduce the cost of production, and reduce the negative environmental impact from the hazardous waste slag and acid to environment.

2. Experimental

2.1 Materials

The as-received 100–200 μm ilmenite particles consisted of several metal oxides (Table 1) and were used as the raw material.

Table 1 Chemical composition of the ilmenite (wt%)

TiO2	FeO	Fe2O3	MgO	SiO2	Al2O3	CaO	MnO2
47.60	32.81	7.25	5.64	3.35	1.66	0.70	0.663

---

2.2 Experimental method and experimental procedure

A detailed illustration of the synthesis process of the spherical Li₄Ti₅O₁₂ and nanowires TiO₂ from natural ilmenite is shown in Fig. 1.

Fig. 1 A detailed illustration of the synthetic process.

2.2.1 Lixivium of ammonia titanium peroxide solution. Compounds of Ti and Fe were effectively separated by hydrochloric acid leaching of natural ilmenite following eqn (1) and (2). The obtained hydrolyzed titania residue and Fe-rich filtrate were prepared by following previously reported methods.^21,22,24 According to previous results, Ti and Si were directionally enriched in hydrolyzed titania residue; moreover, Fe, Mg, Al, Mn and Ca were dissolved in the iron-rich filtrate.

FeTiO₃ + 4HCl → Fe²⁺ + TiOCl₄²⁻ + 2H₂O (1)

The lixivium of the ammonia titanium peroxide solution was prepared by alkaline-hydrogen peroxide leaching of the hydrolyzed titania residue via previously reported methods.^24,25 The hydrolyzed titania residue powder (5 g) was placed in a 500 ml round-bottomed flask, which was attached to a reflux condenser. 10 wt% ammonia solution (60 g) was added in the flask and a white slurry was observed. The white slurry was maintained at 35 °C in a thermostatically controlled water bath, equipped with a digitally controlled thermometer (within ±2 °C). Hydrogen peroxide solutions containing 10 wt% H₂O₂ (180 g) was then added to the white slurry to react with Ti under vigorous stirring of 300 rpm. After 20 min, the solution was rapidly cooled and filtered, and then aqueous ammonia titanium peroxide solution was obtained according to eqn (3).

TiO₂·nH₂O + H₂O₂ + NH₃·H₂O → (NH₄)_x(TiO_y)(O₂)_z(OH)_j(dissoluble) + H₂O (3)

2.2.2 Synthesis of Li₄Ti₅O₁₂ microspheres. LiOH·H₂O and the lixivium of the ammonia titanium peroxide (0.816 Li : Ti molar ratio) solutions were mixed and stirred for 10 min to form a lithium (ammonia) titanium peroxide solution. The prepared lithium (ammonia) titanium peroxide solution was added into a spray dryer machine by a peristaltic pump at 300 ml h⁻¹. The homogenous solution was atomized at 200 °C using a two-fluid nozzle with an atomizing pressure of 0.2 MPa. The spray-dried precursor powders were further calcined in air at 675 °C for 16 h to form Li₄Ti₅O₁₂ particles.

2.2.3 Synthesis of TiO₂ nanowires. A NaOH solution was added into the obtained lixivium of the ammonia titanium peroxide solution to form a 8 : 1 Na : Ti molar ratio. The obtained aqueous sodium (ammonia) titanium peroxide solution was heated to boiling point with vigorous stirring in an oil bath. After 1 h, according to eqn (4), a precipitate of titanium dioxide hydrate with nanosized TiO₂ crystals was obtained and subsequently washed by 2 wt% HNO₃ several times. The TiO₂ precipitate was dried in an oven at 80 °C to obtain the titanium dioxide hydrate power, which was calcined at 300 °C for 3 h to form TiO₂ particles.

2.3 Analysis

Micrographs of the prepared powders were taken with a field-emission LEO 1530 microscope (Zeiss, Germany) at a working distance of 3–6 mm and accelerating field of 2–3 kV and transmission electron microscope (TEM) (Tecnai G12, Netherlands). The powder composition was measured using energy-dispersive X-ray spectroscopy (EDS). The elemental contents of the solutions and samples were analyzed using inductively coupled plasma (ICP) emission spectroscopy (IRIS intrepid XSP, Thermo Electron Corporation, Japan). Powder X-ray diffraction (XRD) (Rint-2000, Rigaku, Japan) using Cu Kα radiation was employed to identify the crystalline phase of the synthesized materials. The specific surface area of the produced materials (S_BET) was calculated from N₂ adsorption isotherms using the Brunauer–Emmett–Teller equation in the range of relative pressures from 0.02 to 0.3. XPS analysis was performed using an Al Kα source (Thermo, USA).

14 mm diameter electrodes with typical mass loadings of 1.95–2 mg cm⁻² were tested in 2025-type coin cells versus metallic Li. The produced electrodes were composed of 80 wt% Li₄Ti₅O₁₂ (or TiO₂) powders, 10 wt% carbon black, and 10 wt% poly(vinylidene fluoride) as the binder. After mixing in N-methyl pyrrolidinone, the slurry was uniformly cast on a thin copper foil and dried in vacuum for 12 h at 120 °C. Note that an electrolyte of LiPF₆ in carbonate (EC : EMC : DMC = 1 : 1 : 1 by volume) was used, and a polypropylene separator was used. Cells were assembled in a dry argon-filled glove box and were charged and discharged at room temperature over voltage ranges of 1.0–2.5 V and 1.0–3.0 V for Li₄Ti₅O₁₂ and TiO₂ versus Li/Li⁺, respectively.

3. Results and discussion

3.1 Li₄Ti₅O₁₂ microspheres

The XRD pattern of the precursor of Li₄Ti₅O₁₂ has no peaks, indicating no appearance of crystalline phases (Fig. 2(a)). The as-prepared Li₄Ti₅O₁₂ is single-phase spinel lithium titanium oxide (cubic phase, space group Fd3̄m) in accordance with spinel Li₄Ti₅O₁₂ (JCPDS card no. 49-0207) and is well crystallized. Thus, well crystallized, single phase Li₄Ti₅O₁₂ can be synthesized at a relatively low temperature (675 °C) by calcining the precursor prepared by this method. Prior to heat treatment, the precursor powders are characterized by a spherical morphology with a diameter size of about 1–5 μm. From Fig. 2(b) and (c), it can be seen that the surface of the precursor, which has been magnified in Fig. 2(d), is smooth. The microspheres particles contain lots of nanoparticles with a size of several nanometers (5 nm), which is attributed to the decomposition of the hydrogen peroxide and ammonia solution in the spray process. Owing to the aqueous solution of lithium titanium peroxide (we added the solution of LiOH into lixivium of the ammonia titanium peroxide, which is obtained from the leach of the hydrolyzed titania residue by using the alkaline-hydrogen peroxide leaching method) in which the metal ions are ion-level homogeneously mixed, we obtained a uniform precursor of Li₄Ti₅O₁₂ microspheres with homogeneous mixing of Li and Ti ions. Owing to the effect of these nanoscale and uniform particles, Li₄Ti₅O₁₂ could be obtained at a relatively low calcination temperature with no impurities of Li₂TiO₃ or TiO₂. There is no obvious circle in the SAED pattern, indicating that no crystalline phases appeared (Fig. 2(d)), which is in accordance with the XRD results of the precursor. In Fig. 2(e) and 3(a) and (f), after calcination at 675 °C for 16 h, the as-prepared Li₄Ti₅O₁₂ also shows a uniform and spherical structure morphology with the same size as the precursor. However, owing to the decomposition of the precursor, the surface of microspheres changes into a rough and porous surface which is constituted of aggregated nanoparticles with a size of about 100–200 nm. The as-prepared Li₄Ti₅O₁₂ has a measured S_BET of 10.3 m² g⁻¹ (Fig. 2(f)), which is consistent with an average Li₄Ti₅O₁₂ particle size of ∼170 nm calculated from BET surface area by assuming Li₄Ti₅O₁₂ density to be 1.68 g cm⁻³ and smooth particle surface. From the EDS pattern of the selective area in the TEM image Fig. 3(f), little impurities of Si have been detected in the synthetic Li₄Ti₅O₁₂. Due to the weak alkaline conditions of leaching, a little Si in the hydrolyzed titania residue is dissolved. From EDS mapping of the Li₄Ti₅O₁₂ (Fig. 3(b)–(e)), Si is well-distributed in the as-produced Li₄Ti₅O_12. Because Si and Fe, which are from natural ilmenite, could be the most likely impurities in the as-produced Li₄Ti₅O₁₂, we performed high-resolution X-ray photoelectron spectroscopy (XPS), including an Fe scan and Si scan on the as-produced Li₄Ti₅O₁₂ to know the impurities (Fig. 3(g)). The XPS results show only Si as the impurity in as-produced Li₄Ti₅O_12. Fig. 3(i) shows the HRTEM images of the as-prepared Li₄Ti₅O₁₂ nanoparticles. As can be seen, there is a lattice fringe with a lattice spacing of about 0.48 nm, corresponding to the (1 1 1) inter planar spacing of Li₄Ti₅O₁₂, which indicates the well-crystallized spinel phase in the nanostructured materials prepared from relatively low temperature heat treatment. The SAED pattern in Fig. 3(i), with several marked rings corresponding to the Li₄Ti₅O₁₂ (1 1 1), (3 1 1), (4 0 0) and (4 4 0) planes, can be indexed to spinel Li₄Ti₅O₁₂ with the cubic space group Fd3̄m.

Fig. 2 (a) XRD patterns of the precursors and synthesized Li₄Ti₅O₁₂ and TiO₂; (b) SEM image of the precursor of Li₄Ti₅O₁₂; (c and d) TEM images of the precursor of Li₄Ti₅O₁₂ (inset: SAED pattern of the precursor of Li₄Ti₅O₁₂.); (e) the SEM image of as-prepared Li₄Ti₅O₁₂; (g) SEM image and SAED (inset) of the precursor of TiO₂; (h) SEM image of synthesized TiO₂. Physisorption isotherms for synthesized (i) TiO₂ and (f) Li₄Ti₅O₁₂.

Fig. 3 SEM (a), EDS mapping (b–e) and TEM (f) images of as-prepared Li₄Ti₅O₁₂; (g) typical XPS spectra of as-produced Li₄Ti₅O₁₂ (the insets show high-resolution Fe_2p and Si_2p spectrums); (h) EDS pattern of selective area in TEM image; (i) HRTEM image and SAED (inset) image of as-produced Li₄Ti₅O₁₂.

The voltage profiles of Li₄Ti₅O₁₂ microspheres at room temperature in the charge–discharge test at various current densities are shown in Fig. 4. The cell was first cycled at C/10 for 3 cycles, C/2 for 5 cycles, 1 C for 50 cycles, 2 C for 100 cycles, 5 C for 100 cycles, 10 C for 100 cycles, 20 C for 100 cycles, and then returned to 0.1 C for 10 cycles. At all C-rates, the charge–discharge curves exhibit a long and flat voltage plateau. In the first discharge process, the as-prepared Li₄Ti₅O₁₂ electrode exhibits a remarkably high initial discharge capacity at 0.1 C-rate, up to 173.1 mA h g⁻¹, which is almost equal to its theoretical capacity (175 mA h g⁻¹). The subsequent Li⁺ extraction, proceeding up to 2.5 V, shows a capacity of 170.2 mA h g⁻¹, with a high coulombic efficiency (ratio of extraction to insertion capacity) of 98.32%. By increasing the C-rate, the cell shows 168.5, 167.2, 163.4, 160.0, 155.4 and 142.7 mA h g⁻¹ at 0.5, 1, 2, 5, 10 and 20 C-rates, respectively. The capacities at all charge–discharge rates are much larger than other Li₄Ti₅O₁₂ and Li₄Ti₅O₁₂/C electrodes made by chemically pure or analytically pure Titanium compounds.^26–44 Moreover, the cycling performance of the as-prepared Li₄Ti₅O₁₂ is shown in Fig. 4(b). As can be seen, the cycling curves of discharge and charge coincided with each other, indicating a high coulombic efficiency in the charge/discharge of this cell. At all the current densities, after 50 or 100 cycles, the as-prepared Li₄Ti₅O₁₂ electrode retains 97.9%, 99.1%, 98.9%, 96.8% and 91.8% of its initial discharge capacities at the 1, 2, 5, 10 and 20 C-rates, respectively. In the last 10 cycles at C/10 rate, the charge capacity of 459^th retains 103.9% of 468^th charge capacity. The excellent electrochemical performance might be attributed to the as-prepared Li₄Ti₅O₁₂ microspheres obtained by this novel method and natural ilmenite. The as-produced Li₄Ti₅O₁₂ microspheres composed of nanoparticles of around 100 nm contained a rough surface and inside pores. These nanosized Li₄Ti₅O₁₂ particles connected with each other forming an interconnected Li₄Ti₅O₁₂ nano-network, which provides a massive interface among the nanoparticles and can be the bridge for electron and Li⁺ transport through a shorter path within and between the nano-particles. Moreover, the inside pores make a larger electrode–electrolyte contact area for Li⁺ insertion/extraction in the rough and porous spherical particles. Therefore, the as-prepared Li₄Ti₅O₁₂ shows an excellent high-rated cycling performance, which is better than our previous Li₄Ti₅O₁₂ microspheres obtained by chemically pure or analytically pure titanium compounds (Table 2).

Fig. 4 (a) Charge–discharge curves of Li₄Ti₅O₁₂ at multiple C-rates in the voltage range of 1.0–2.5 V; (b) cycling performance of Li₄Ti₅O₁₂ at multiple C-rates; (c) the first and 468^th charge–discharge curves of Li₄Ti₅O₁₂ at C/10 rate; (d) cyclic voltammograms of the as-prepared Li₄Ti₅O₁₂ at the scan rate of 0.1 mV s⁻¹ between 2.5 and 1 V after charge–discharge testing; SEM images of as-produced Li₄Ti₅O₁₂ electrodes (e) before and (f) after cycling.

Table 2 Comparison of electrochemical performance between Li₄Ti₅O₁₂ in this work and the previous Li₄Ti₅O₁₂ by the similar method using organic titanium compounds

Samples	Discharge capacities (mA h g^−1) at different C-rates							Cycling performances (%) at different C-rates
	0.1 C	0.5 C	1 C	2 C	5 C	10 C	20 C	2 C	5 C	10 C	20 C
Ti(OC4H9)4-organic^26	174.8	170.5	167.5	165.3	158.2	152.1	111.9	96.9	98.8	89.3	89.0
Ilmenite-this work	173.1	168.5	167.2	163.4	160.0	155.4	142.7	99.1	98.9	96.8	91.8

---

Fig. 4(c) shows the first and 468^th charge–discharge curves of Li₄Ti₅O₁₂ at C/10 rate. After a total of 450 cycles at miscellaneous current densities, the cell returns to C/10 and still retains a remarkably high discharge capacity of 168.8 mA h g⁻¹ which is 97.6% of the initial discharge capacity at 0.1 C-rate. As can be seen in Fig. 4(d), all of the five cycles have a sharp pair of cathodic/anodic peaks centered at 1.50 V and 1.63 V, corresponding to the lithium insertion/extraction in the spinel Li₄Ti₅O₁₂ lattice, which are in accordance with the plateaus of the discharging–charging curves. The measured value of the ratio for peak currents i_pa/i_pc is nearly 1, and the integral voltammetric area of the discharging–charging branches is almost equal. Moreover, after 471 cycles, 5 cyclic voltammograms coincided with each other, indicating very perfect coulombic efficiency. After comparing the SEM images of the electrodes before and after cycling in Fig. 4(e) and (f), the nano-micron structure is still maintained and stable after ∼500 cycles, which is due to the “zero-strain” during charge and discharge. As can be seen from the results in Fig. 4, this Li₄Ti₅O₁₂ shows a very steady-going structure, high capacities and high cycling performance even after hundreds of cycles at high C-rates.

3.2 TiO₂ nanowires

Fig. 2(a) shows the XRD patterns of the precursor and as-prepared TiO₂. The precursor shows weak (1 0 1) and (2 0 0) diffraction peaks of anatase TiO₂. After calcination at 300 °C for 3 h, the as-prepared TiO₂ with high crystallinity is well ascribed to the (1 0 1), (1 1 0), (2 0 0), (2 1 1), (2 0 4), (1 1 6) and (2 1 5) diffraction peaks of anatase TiO₂. This indicates that the structure of the precursor is easy to change into anatase TiO₂ nanowires. ICP analysis reveals that the content percent of SiO₂ and TiO₂ in synthetic TiO₂ are 0.32% and 99.68%, respectively. Other elements like Mg, Fe, Al, Mn and Na are not detected by AAS and ICP in the as-prepared TiO₂. In our earlier work, the content percent of SiO₂ and TiO₂ in synthetic TiO₂ are 2.19% and 97.81%, respectively, and the content percent of TiO₂ in nanosheets TiO₂ is 99.31%.^24,25 Therefore, the addition of NaOH before the boiling process can decrease the content percent of SiO₂ in the as-prepared TiO₂. NaOH can react with silicon in the leaching solution to form the SiO₃²⁻ ion, which can exist stably in the alkaline solution. We also detected elements on the surface of samples by EDS. As shown in Fig. 5(d), only little Si is detected in as-prepared TiO₂, which is consistent with the ICP results.

Fig. 5 (a) SEM image of as-prepared TiO₂; (b and c) TEM images of anatase TiO₂; (d) EDS pattern of as-prepared TiO₂; (e) SAED image of anatase TiO₂; (f) HRTEM image and FFT (inset) image of selected area from as-prepared TiO₂.

When boiling the sodium titanium peroxide solution, the H₂O₂ is gradually decomposed. Due to the interfacial tension, van der Waals attractive forces, and other factors, the titanate nucleus would aggregate together to form “multi-nuclei.” Subsequently, the titanate “multi-nuclei” gradually grow up and assemble into a micro spherical shape along with H₂O₂ decomposition. The precursor is calcined at 300 °C for 3 h to synthesize anatase TiO₂. As shown in Fig. 2(g)–(i), the morphology images of the precursor and as-prepared TiO₂ are the similar echinus-like structure with a diameter about 0.5–1.5 μm, which shows a high BET SSA of 20.2 m² g⁻¹. The as-prepared TiO₂ maintains the morphology of the precursor, which is attributed to the low calcined temperature. The enlarged image of the particles of the as-prepared TiO₂ in Fig. 5(a) shows that the spherical particles or microspheres are composed of aggregated nanowires, which is consistent with the TEM images (Fig. 5(b) and (c)). As shown in Fig. 5(b) and (c), the morphology images of the as-prepared TiO₂ are the echinus-like structure with aggregated nanowires. These nanowires are acicula-shaped with a diameter in the range of 14.4 nm and a length of about 200–500 nm. From Fig. 2(g), there is no obvious circle in the inset SAED pattern, indicating that there is no appearance of crystalline phases, which is in accordance with the XRD results of the precursor. However, the SAED pattern in Fig. 5(e), with several marked rings corresponding to TiO₂ (1 0 1), (0 0 4), (2 0 0) and (2 1 1) planes, can be indexed to anatase TiO₂ (JCPDS card no. 65-5714). Fig. 5(f) shows the HRTEM images of the as-prepared TiO₂ nanoparticles. As can be seen, there is a lattice fringe with a lattice spacing of about 3.51 Å, corresponding to (1 0 1) inter planar spacing of TiO₂, which indicates the well-crystallized anatase phase in the nanostructured materials. The inset SAED patterns are obtained by Fourier transformation in a selected area of the HRTEM images. The inset FFT image in Fig. 5(f) also gives a sharp spot, corresponding to the (1 0 1) plane of anatase TiO₂.

Fig. 6(a) shows the initial three potential-capacity profiles of anatase TiO₂ at the current density of 20 mA g⁻¹. In these three cycles, the charge–discharge curves exhibit a long and flat voltage plateau at about 1.77 and 1.88 V for discharging and charging, respectively. In the first discharge process, anatase TiO₂ shows 291.9 mA h g⁻¹ at a current density of 20 mA g⁻¹. The subsequent Li⁺ extraction of anatase TiO₂ shows a capacity of 248.2 mA h g⁻¹. It displays a relative low coulombic efficiency (ratio of extraction to insertion capacity) and a large irreversible capacity. However, in the subsequent two cycles, the discharge–charge curves gradually coincided with each other, indicating a lower irreversible capacity and a higher coulombic efficiency. The fade of the curve between the first and subsequent cycles shows the large irreversible process after the insertion of Li into the TiO₂ anatase structure. This phenomenon is tentatively ascribed to the side reaction of trace surface adsorbed water because of the large specific surface area. The trace water in the anatase TiO₂ could react irreversibly with lithium forming Li₂O inside the inter layers or on the surface, which is the major reason for the larger capacity loss. The discharging–charging tends to be stabilized in the following cycles, because the binding water is consumed gradually during the first several cycles.^45,46 Fig. 6(b) shows the CV pattern of the anatase TiO₂ electrodes. Three cathodic peaks of lithium insertion can be found in the voltammogram near 1.50, 1.55 and 1.71 V vs. Li⁺/Li; moreover, three anodic peaks of lithium extraction at 1.55, 1.67 and 2.03 V vs. Li⁺/Li are observed. The sharp pair of peaks (1.71 and 2.03 V) appears to be the lithium insertion/extraction in the anatase TiO₂ lattice, which is in accordance with the plateaus of the discharging–charging curves in Fig. 6(a); however, the other two pairs at the lower potential region are characteristic of the capacitive charging process, which may represent a different mechanism of lithium intercalation into the TiO₂ host.⁴⁵

Fig. 6 (a) Initial three potential-capacity profiles of anatase TiO₂ nanowires at a charge–discharge current density of 20 mA g⁻¹. (b) Representative cyclic voltammogram plot of anatase TiO₂ nanowires electrodes at 0.1 mV s⁻¹.

Fig. 7(a) shows the discharge and charge capacities of TiO₂ at different current densities. The cell was first cycled at 20 mA g⁻¹ for 3 cycles, 100 mA g⁻¹ for 5 cycles, 200 mA g⁻¹ for 20 cycles, 400 mA g⁻¹ for 50 cycles, 1000 mA g⁻¹ for 100 cycles, 2000 mA g⁻¹ for 100 cycles, and then returned to 20 mA g⁻¹ for 10 cycles. At all current densities, the charge–discharge curves exhibited obvious and flat voltage plateaus. The initial discharge capacities of TiO₂ are 291.9, 230.9, 222.6, 202.7, 177.3 and 142.6 mA h g⁻¹ at the current densities of 20, 100, 200, 400, 1000 and 2000 mA g⁻¹, respectively. Moreover, the cycling performance of TiO₂ is shown in Fig. 7(b). As can be seen, TiO₂ shows a very good cycling performance. After 20 cycles, the TiO₂ retains 96.8% of the initial charge capacity at the current density of 200 mA g⁻¹. At the current density of 400 mA g⁻¹, TiO₂ retains 95.6% of the initial charge capacity after 50 cycles. After 100 cycles, TiO₂ retains 91.9% and 86.9% of the initial charge capacities at the current densities of 1000 and 2000 mA g⁻¹, respectively. Compared to the previous TiO₂ nanowires prepared by same method using organic titanium compounds, the produced TiO₂ nanowires in this work show a better rate performance as can be seen in Table 3. (We cannot compare the cycling performance due to the different cycling programs.) Fig. 7(c) shows the second and 288^th charge–discharge curves of the anatase TiO₂ at a current density of 20 mA g⁻¹. After a total of 270 cycles at 200, 400, 1000 and 2000 mA g⁻¹, the cell returns to 20 mA g⁻¹ and still exhibits a remarkably high discharge capacity of 227.3 mA h g⁻¹, which is 89.3% of the second discharge capacity at 20 mA g⁻¹. As can be seen from the results in Fig. 7, the anatase TiO₂ prepared by natural ilmenite shows a very steady-going structure, a high capacity and a high cycling performance even after hundreds of cycles at high current densities.

Fig. 7 (a) Charge–discharge curves of anatase TiO₂ for multiple current densities in the voltage range of 1.0–3.0 V. (b) Cycling performance of anatase TiO₂ at multiple current densities. (c) Charge–discharge curves for the second and 288^th cycles for TiO₂ at 20 mA g⁻¹.

Table 3 Comparison of electrochemical performance between TiO₂ nanowires in this work and the previous TiO₂ nanowires by the similar methods using organic titanium compounds

Samples	Discharge capacities (mA h g^−1) at different current densities				Cycling performances (%) at different current densities
	200 mA g^−1	400 mA g^−1	1000 mA g^−1	2000 mA g^−1	200 mA g^−1	400 mA g^−1	1000 mA g^−1	2000 mA g^−1
Ti(OC4H9)4-organic^47	221	195	117	90	82 (50 cycles)	96 (50 cycles)	96 (50 cycles)	94 (50 cycles)
Ilmenite-this work	222.6	202.7	177.3	142.6	96.8 (20 cycles)	95.6 (50 cycles)	91.9 (100 cycles)	86.9 (100 cycles)

---

4. Conclusions

In summary, we have reported the successful preparation of uniform Li₄Ti₅O₁₂ microspheres and TiO₂ nanowires from ilmenite via easily replicated chemical and thermal methods. Both of the materials are directly produced from the lixivium of an ammonia titanium peroxide solution, which is the metallurgical-middle product of natural ilmenite. Both of the synthesized materials demonstrate an electrochemical performance that exceeds previously reported data for Li₄Ti₅O₁₂ and TiO₂ nanowires made by chemically pure or analytically pure titanium compounds. The Li₄Ti₅O₁₂ microspheres particles demonstrate excellent high rate cycling and a stable performance. The echinus-like structure of the aggregations of the acicula-shaped TiO₂ nanowires also shows good electrochemical performance. The cost of natural ilmenite is competitive with the market price for the same chemicals prepared from primary resources. We believe that this method is a simple, efficient and economical way for both ilmenite utilization and materials preparation.

Acknowledgements

Support from the Hunan Provincial Innovation Foundation for Postgraduates (CX2012A004), the China Scholarship Council (201206370083), the Young Scholarship Award for doctoral candidate funded by Ministry of Education (1343-76140000019), and the National Basic Research Program of China (2014CB643406) are acknowledged.

Notes and references

1. N. Nitta and G. Yushin, Part. Part. Syst. Charact., 2014, 4, 317.
2. L. W. Ji, Z. Lin, M. Alcoutlabi and X. W. Zhang, Energy Environ. Sci., 2011, 4, 2682.
3. M. Armand and J. M. Tarascon, Nature, 2008, 451, 652.
4. A. S. Aricò, P. Bruce, B. Scrosati, J. M. Tarascon and W. V. Schalkwijk, Nat. Mater., 2005, 4, 366.
5. P. G. Bruce, B. Scrosati and J. M. Tarascon, Angew. Chem., Int. Ed., 2008, 47, 2930.
6. H. Yang, H. J. Bang, K. Amine and J. Prakash, J. Electrochem. Soc., 2005, 152, A73.
7. H. Yang, S. Amiruddin, H. J. Bang, Y. K. Sun and J. Prakash, J. Ind. Eng. Chem., 2006, 12, 12.
8. H. Yang and X. D. Shen, J. Power Sources, 2007, 167, 515.
9. M. Lu, H. Cheng and Y. Yang, Electrochim. Acta, 2008, 53, 3539.
10. R. Fong, U. von Sacken and J. R. Dahn, J. Electrochem. Soc., 1990, 137, 2009.
11. T. Ohzuku, A. Ueda and N. Yamamoto, J. Electrochem. Soc., 1995, 142, 1431.
12. S. I. Pyun, S. W. Kim and H. C. Shin, J. Power Sources, 1999, 81, 248.
13. J. Gao, C. Y. Jiang, J. R. Ying and C. R. Wan, J. Power Sources, 2006, 155, 364.
14. Y. G. Guo, Y. S. Hu and J. Maier, Chem. Commun., 2006, 2783.
15. M. Wagemaker, W. J. H. Borghols and F. M. Mulder, J. Am. Chem. Soc., 2007, 129, 4323.
16. K. Zaghib, M. Simoneau, M. Armand and M. Gauthier, J. Power Sources, 1999, 81, 300.
17. S. Ma and H. Noguchi, J. Electrochem. Soc., 2001, 148, 589.
18. D. Peramunage and K. M. Abraham, J. Electrochem. Soc., 1998, 145, 2609.
19. H. Qiao, L. F. Xiao and L. Z. Zhang, Electrochem. Commun., 2008, 10, 616.
20. B. L. He, B. Dong and H. L. Li, Electrochem. Commun., 2007, 9, 425.
21. L. Wu, X. H. Li, Z. X. Wang, X. J. Wang, L. J. Li, J. Fang, F. X. Wu and H. J. Guo, Powder Technol., 2010, 199, 293.
22. L. Wu, X. H. Li, Z. X. Wang, H. J. Guo, X. J. Wang, F. X. Wu, J. Fang, Z. G. Wang and L. J. Li, J. Alloys Compd., 2010, 506, 271.
23. X. J. Wang, X. H. Li, Z. X. Wang, L. Wu, P. Yue, H. J. Guo, F. X. Wu and T. T. Ma, Powder Technol., 2010, 204, 198.
24. F. X. Wu, X. H. Li, Z. X. Wang, L. Wu, H. J. Guo, X. H. Xiong, X. P. Zhang and X. J. Wang, Int. J. Miner. Process., 2011, 98, 106.
25. F. X. Wu, X. H. Li, Z. X. Wang, H. J. Guo, L. Wu, X. H. Xiong and X. J. Wang, Powder Technol., 2011, 213, 192.
26. F. X. Wu, X. H. Li, Z. X. Wang, H. J. Guo, Z. J. He, Q. Zhang, X. H. Xiong and P. Yue, J. Power Sources, 2012, 202, 374.
27. F. X. Wu, Z. X. Wang, X. H. Li, H. J. Guo, P. Yue, X. H. Xiong, Z. J. He and Q. Zhang, Electrochim. Acta, 2012, 78, 331.
28. H. Yan, Z. Zhu, D. Zhang, W. Li and Q. Lu, J. Power Sources, 2012, 219, 45.
29. R. Cai, S. M. Jiang, X. Yu, B. T. Zhao, H. T. Wang and Z. P. Shao, J. Mater. Chem., 2012, 22, 8013.
30. Y. J. Bai, C. Gong, Y. X. Qi, N. Lun and J. Feng, J. Mater. Chem., 2012, 22, 19054.
31. B. H. Li, C. P. Han, Y. B. He, C. Yang, H. D. Du, Q. H. Yang and F. Y. Kang, Energy Environ. Sci., 2012, 5, 9595.
32. G. Y. Liu, H. Y. Wang, G. Q. Liu, Z. Z. Yang, B. Jin and Q. C. Jiang, J. Power Sources, 2012, 220, 84.
33. H. F. Ni and L. Z. Fan, J. Power Sources, 2012, 214, 195.
34. H. G. Jung, S. T. Myung, C. S. Yoon, S. B. Son, K. H. Oh, K. Amine, B. Scrosati and Y. K. Sun, Energy Environ. Sci., 2011, 4, 1345.
35. J. Wolfenstine and J. L. Allen, J. Power Sources, 2008, 180, 582.
36. X. Li, M. Z. Qu, Y. J. Huai and Z. L. Yu, Electrochim. Acta, 2010, 55, 2978.
37. C. M. Shen, X. G. Zhang, Y. K. Zhou and H. Li, Mater. Chem. Phys., 2002, 78, 437.
38. K. C. Hsiao, S. C. Liao and J. M. Chen, Electrochim. Acta, 2008, 53, 7242.
39. S. H. Ju and Y. C. Kang, J. Phys. Chem. Solids, 2009, 70, 40.
40. D. Yoshikawa, Y. Kadoma, J. M. Kim, K. Ui, N. Kumagai, N. Kitamura and Y. Idemoto, Electrochim. Acta, 2010, 55, 1872.
41. J. Z. Chen, L. Yang, S. H. Fang and Y. F. Tang, Electrochim. Acta, 2010, 55, 6596.
42. Y. F. Tang, L. Yang, Z. Qiu and J. S. Huang, Electrochem. Commun., 2008, 10, 1513.
43. X. Guo, H. F. Xiang, T. P. Zhou, W. H. Li, X. W. Wang, J. X. Zhou and Y. Yu, Electrochim. Acta, 2013, 109, 33.
44. Y. Wang, H. B. Rong, B. Z. Li, L. D. Xing, X. P. Li and W. S. Li, J. Power Sources, 2014, 246, 213.
45. J. R. Li, Z. L. Tang and Z. T. Zhang, Electrochem. Solid-State Lett., 2005, 8, A316.
46. Y. F. Wang, M. Y. Wu and W. F. Zhang, Electrochim. Acta, 2008, 53, 7863.
47. F. X. Wu, Z. X. Wang, X. H. Li and H. J. Guo, J. Mater. Chem., 2011, 21, 12675.

---