TiO₂ Nanoparticles Prepared by Sol-Gel Method for Anode Application in Lithium-Ion Batteries

Abstract

TiO 2 nanoparticles are prepared via sol-gel method using titanium tetraisopropoxide (TTIP) as a precursor under refluxing and controlled pH. It is found that pure anatase phase is obtained with pH 10. Further characterization studies are carried out on pure nanoparticle anatase phase by XRD, SEM, and transmission electron microscope (TEM). Their electrochemical performances as anode material in lithium-ion batteries are investigated by cyclic voltammetry, galvanostatic cycling, and rate capability measurements. A high discharge capacity of 321 mAh/g (vs. 335 mAh/g theoretical) is achieved at 1C rate. After the first galvanostatic charge/discharge cycle, voltage profiles show plateaus at 1.75 and 1.95 V versus Li at discharge and charge, respectively. High Coulombic efficiency (>99%) is maintained after 300 cycles, which makes anatase TiO 2 nanoparticles prepared by sol-gel method, a very promising material for anode application in lithium rechargeable batteries.

1. Introduction

Nanostructured titania (NST) ${\mathrm{TiO}}_2$ is a very promising alternative anode material compared with the commercialized lithium titanate $\mathrm{L}{\mathrm{i}}_4{\mathrm{Ti}}_{15}{\mathrm{O}}_{12}$ and graphite materials for lithium-ion battery application in particular when fast charging is required [1, 2, 3]. Fast charging while keeping long battery operational life is one of the most challenging tasks in lithium-ion batteries nowadays. Both mobile electronics and electro-mobility sectors need fast charging to compensate for relatively low energy density which impacts the operation time and driving range. One serious option to cope with charging time limitations resides in developing new materials able to store large amounts of lithium in a very short time. Nanomaterials, including NST, are among the most promising in this regard.

Good performances of NST result from their outstanding characteristics such as:

1. A low volumetric change upon lithium-ion intercalation and de-intercalation providing a high crystal structure stability and long cycle life

2. A fast diffusivity of lithium ions within the $\mathrm{Ti}{\mathrm{O}}_2$ structure providing fast charging and high power discharge

3. A practical low operating voltage

4. An enhanced safety

Several methods have been used for the synthesis of NST, including hydrothermal [4, 5, 6, 7, 8, 9, 10], sol-gel [5, 6, 7, 8, 9, 10, 11], hydrolysis [12, 13], atomic layer deposition [14], and anode deposition [15], for the most reported in the literature. According to the synthesis method and to the post-thermal treatment conditions, NST of different morphologies are obtained such as nanorods [4, 17], nanotubes [5, 6, 8, 9, 14, 17], nanofibers/nanowires [11, 12, 13, 14, 15, 16, 17, 18, 19, 20], nanoparticles [6, 7, 10], nanocomposites [12, 21], nanofilms, and nanosheets [15, 22].

Several polymorphs of $\mathrm{Ti}{\mathrm{O}}_2$ are known such as tetragonal rutile and anatase, orthorhombic brookite, and monoclinic $\mathrm{Ti}{\mathrm{O}}_2$ (B) [23]. Anatase and $\mathrm{Ti}{\mathrm{O}}_2$ (B) have open-channel structure and show high cycle capacity compared to rutile $\mathrm{Ti}{\mathrm{O}}_2$ [24]. Anatase has edge-sharing $\mathrm{Ti}{\mathrm{O}}_6$ octahedra so that planar double chains are formed with a tetragonal body-centered space group (I4₁/amd) [25].

Li-ion diffusion in anatase occurs along the octahedral interstitial sites up to $\mathrm{L}{\mathrm{i}}_{0.5}\mathrm{Ti}{\mathrm{O}}_2$ composition. Then the tetragonal structure reversibly transforms to the orthorhombic phase [26, 27, 28, 29, 30]. Anatase $\mathrm{Ti}{\mathrm{O}}_2$ anode material reversibly uptakes 0.5 Li per $\mathrm{Ti}{\mathrm{O}}_2$ formula with lithium insertion/extraction taking place at relatively low voltage and showing fast kinetics [29, 30]. However, 0.5 Li per $\mathrm{Ti}{\mathrm{O}}_2$ corresponds to half of theoretical capacity of 335 mAh/g [31, 32, 33]. Enhancing practical capacity in NST should make this material more attractive for anode application in lithium-ion batteries [34].

When wet chemistry route is used to prepare NST, the medium pH is of a significant importance in the crystal structure and morphology of the final product [32]. Syntheses of TiO₂ anatase were mostly performed in acidic media usually at pH greater than 4.5 [33]. To the authors’ knowledge, there has been no report on sol-gel preparation of NST in alkaline pH.

In this chapter, we show that unlike in acidic pH, which lead to mixed phases, pure anatase NST is obtained via sol-gel at pH 10. NST prepared at different pH are analyzed by XRD, transmission electron microscope (TEM), and SEM. The electrochemical properties of pure anatase phase NST are investigated by cyclic voltammetry and galvanostatic cycling, including rate capability tests in view of its application as anode material in lithium-ion batteries.

2. Experimental

2.1 Sample preparation

NST are prepared via sol-gel method using titanium tetraisopropoxide (TTIP), distilled water, and 2-propanol. Hydrochloric acid and sodium hydroxide are mixed in 24 mL of 2-propanol to fix the pH to 1, 4, 10, and 13 and to control the hydrolysis process of the solution. 24 mL of TTIP is then slowly added to the mixture, followed by refluxing at 70°C to obtain the sol. 10 mL of distilled water is added to the solution and left for 1 h under stir heating.

The obtained gel is dried at 70°C in air and then calcined at 500°C in air to produce NST powders.

2.2 Characterization

NST are characterized by X-ray diffractometry (XRD) using a PR-XPERT diffractometer (PW3064) with CuK_α radiation (λ = 1.5406 Å), operating at 45 kV and 40 mA in 2θ angular range of 10–70°. The morphology and particle size were examined by transmission electron microscope (TEM, JEOL, JEM-2010) and by field-emission scanning electron microscopy (FESEM, JEOL JSM-7600F, 5 kV).

2.3 Electrochemical measurements

The anode is prepared by mixing NST prepared at pH 10 as active material, acetylene carbon black as electronic conductor, and polyvinylidene fluoride (PVDF) as a binder in the weight ratio of 80:10:10 with N-methyl-2-pyrrolidone (NMP) to form a slurry. The slurry was coated on a copper foil by doctor blade and then vacuum dried at 110°C for 6 h. The anode loading is ~0.8 mg cm⁻². Lithium half-cells are assembled in a glove box under ultra-dry argon atmosphere (<1 ppm ${\mathrm{H}}_2\mathrm{O}$ and ${\mathrm{O}}_2$ ). Lithium metal was used as the counter and reference electrode. The electrolyte consisted of 1.0 M $\mathrm{LiP}{\mathrm{F}}_6$ solution in ethylene carbonate (EC) and diethyl carbonate (DEC) at a volumetric ratio of 1:1. The separator is microporous polyethylene. Li/NST half-cells are investigated by cyclic voltammetry at 0.2 mV/s sweeping rate and by galvanostatic cycling at 1C–20 C rates, all between 1.0 V and 3.0 V versus Li.

3. Results and discussion

3.1 Physical characterizations

Figure 1 shows the XRD patterns of ${\mathrm{TiO}}_2$ nanoparticles prepared at pH 1, 4, 10, and 13 and calcined at 500°C. X-ray patterns of anatase, rutile, and brookite are shown for reference.

Figure 1.

XRD patterns of titania nanoparticles made at different pH. Anatase, rutile, and brookite spectra are also shown for reference.

The sample prepared at pH 1 consists mostly of the rutile phase. All other samples are anatase with brookite impurities at pH 4 and 13. TiO₂ nanoparticles synthesized at pH 10 consist of highly crystalline anatase phase (00-004-0477). Table 1 summarizes the XRD data of TiO₂ nanoparticles synthesized at pH 10 including diffraction peak 2θ, interlayer spacing “d,” and hkl indexes. Table 2 gives the crystallographic characteristics of ${\mathrm{TiO}}_2$ anatase phase prepared at pH 10 as derived from the XRD data Rietveld refinement. Average crystallite sizes of nanoparticles synthesized at different pH are reported in Table 3 using Scherrer’s equation. The particle size is in ~20–30 nm range with no obvious trends versus pH.

2θ (º)	d (nm)	(hkl)
25.31	0.3518	(101)
36.98	0.2430	(103)
37.82	0.2378	(004)
38.61	0.2331	(112)
48.05	0.1893	(200)
53.92	0.1700	(105)
55.09	0.1637	(221)
62.13	0.1494	(213)
62.70	0.1481	(204)
62.80	0.1479	(116)

Table 1.

Indexation of XRD peaks on TiO₂ nanoparticles obtained at pH 10.

Phase	α β γ	Lattice parameters
Anatase; (Tetragonal I41)	α = β = γ = 90°	a = b = 3.783	c = 9.5190

Table 2.

Characteristics of anatase crystal structure.

pH	Phase	D (nm)
1	Anatase-rutile	32.0
4	Anatase-brookite	21.3
10	Anatase	26.2
13	Anatase-brookite	21.3

Table 3.

Crystallite size of TiO₂ nanoparticles synthesized at different pH.

Figure 2 shows TEM images of $\mathrm{Ti}{\mathrm{O}}_2$ nanoparticles prepared at pH 10 at different magnifications. Figure 2a and b shows primary particles with relatively narrow size distribution. High-resolution TEM of Figure 2c shows the polycrystalline nature of the ${\mathrm{TiO}}_2$ nanoparticles with well-resolved layers such as those assigned to the orientation corresponding to an interlayer of about 0.36 nm as indicated by an arrow.

Figure 2.

TEM images at different magnifications of TiO₂ nanoparticles prepared at pH 10. (a, b) Image analysis enables size distribution to be achieved. HR-TEM image of (c) shows polycrystalline structure with some preferred layer orientation such as with interlayer spacing of about 3.6 nm.

TEM image analysis allows particle-size distribution to be determined as shown in Figure 3. The particle-size distribution is built on a discrete set of clearly identifiable grains, and, therefore, it is not necessarily representative of the overall description of the sample. This is especially true for grain sizes close to the distribution tails where the probability of finding a very large or a very small grain under the TEM objective is rather low. The average size of about 17 nm is slightly lower than the one estimated from the XRD data.

Figure 3.

Grain size distribution for the TiO₂ at pH 10 from the TEM image analysis.

SEM images of NST pH 10 are displayed in Figure 4a and b. High magnification Figure 4a shows primary particles with narrow size distribution aggregate into secondary particles of about 150–600 nm in size, whereas lower magnification Figure 4b shows agglomerated particles remain basically submicron in size.

Figure 4.

SEM images of TiO₂ nanoparticles synthesized at pH 10.

3.2 Electrochemical characterizations

Figure 1ashows cyclic voltammograms of the Li/NST of pH 10 half-cell at 0.2 mV/s sweeping rate for 5 cycles. In the first cycle, two reduction peaks appear at 2.15 V and at 1.75 V, respectively, whereas only one re-oxidation peak is present at 2.10 V. In the following 4 cycles, only the reduction peak at 1.75 V and re-oxidation peak at 2.10 V appeared. The first reduction peak at 2.15 V accounts for about 50% of the total discharge capacity of about 320 mAh/g, which is close to theoretical capacity of 335 mAh/g of ${\mathrm{TiO}}_2$ corresponding to 1 Li/ ${\mathrm{TiO}}_2$ ratio. The second reduction peak at 1.75 V accounts for the second half of the total capacity and is matched by the re-oxidation peak at 2.1 V. This situation remained almost unchanged during the four following cycles, that is, the capacity of the reduction peak at 1.75 V matches the one of the re-oxidation peaks at 2.1 V of about 140 mAh/g.

The data shown above are confirmed during the two first galvanostatic cycles of the Li/NST cell displayed in Figure 5b. The first discharge profile shows two semi-plateaus at about 2.15 and 1.75 V, respectively, whereas the charge profile shows a main plateau at about 2.0 V. The first discharge semi-plateau accounts for nearly half of the total discharge capacity. The second half discharge capacity is matched by the charge capacity.

Figure 5.

Electrochemical tests performed on Li/TiO₂ (pH 10) half-cells. (a) Cyclic voltammograms (CVs) at a scan rate of 0.2 mV/s for the first 5 cycles within a potential range from 1 to 3.0 V (vs. Li/Li⁺), (b) discharge-charge capacity of the first and second cycles at a specific current of 175 mA/g (1C), and (c) cycle capacity profile at 1C rate for 50 cycles.

Cyclic voltammetry and galvanostatic data suggest a two-step electrode reaction process:

1. Irreversible formation of $\mathrm{L}{\mathrm{i}}_{0.5}\mathrm{Ti}{\mathrm{O}}_2$ at ~ 2.15 V average voltage.

2. Reversible formation of $\mathrm{L}{\mathrm{i}}_1\mathrm{Ti}{\mathrm{O}}_2$ from $\mathrm{L}{\mathrm{i}}_{0.5}\mathrm{Ti}{\mathrm{O}}_2$ at ~ 1.75 V. During re-oxidation $\mathrm{L}{\mathrm{i}}_{0.5}\mathrm{Ti}{\mathrm{O}}_2$ is formed from $\mathrm{L}{\mathrm{i}}_1\mathrm{Ti}{\mathrm{O}}_2$ at about 2.1 V.

The formation of lithium-rich $\mathrm{L}{\mathrm{i}}_1\mathrm{Ti}{\mathrm{O}}_2$ is supported by both theoretical and experimental data [30, 31, 32, 33, 34, 35]. In the anatase phase, orthorhombic $\mathrm{L}{\mathrm{i}}_{0.5}\mathrm{Ti}{\mathrm{O}}_2$ and distorted rock-salt structure Li₁TiO₂ in which lithium cation occupy the octahedral sites have been stabilized due to kinetic effects. Metastability has been reported in Li-poor $\mathrm{Ti}{\mathrm{O}}_2$ [30]. Accordingly, our data are consistent with reversible lithium intercalation in the stoichiometry range 0.5 < x < 1 in anatase $\mathrm{L}{\mathrm{i}}_{\mathrm{x}}\mathrm{Ti}{\mathrm{O}}_2$ , which corresponds to a theoretical capacity of 167.5 mAh/g. The cycle capacity profile in Figure 5c shows a plateau at about 120 mAh/g at 350 mA/g rate. A lower capacity may be due to limitation in lithium uptake to $\mathrm{L}{\mathrm{i}}_{0.85}\mathrm{Ti}{\mathrm{O}}_2$ and $\mathrm{L}{\mathrm{i}}_{0.7}\mathrm{Ti}{\mathrm{O}}_2$ as reported in [36, 37].

The intercalation mechanism of Li ions into anatase has been studied in a variety of experimental and theoretical works [38]. The lithium-ion intercalation can be described by the following Eq. (1):

$$\mathit{Ti}{O}_2+x{\mathit{Li}}^{+}+x{e}^{-}\to L{i}_x\mathit{Ti}{O}_2$$ E1

$\mathrm{Ti}{\mathrm{O}}_2$ maximum theoretical capacity is 335 mAh/g [39] which corresponds to x = 1 and to the complete reduction of ${\mathit{Ti}}^{4+}$ to ${\mathit{Ti}}^{3+}$ . This makes $\mathrm{Ti}{\mathrm{O}}_2$ a highly competitive alternative to graphite anodes [40].

Practical capacity of anatase $\mathrm{Ti}{\mathrm{O}}_2$ is about 200 mAh/g associated with 0.6 Li mole reversible intercalation at 1.78 V versus Li⁺/Li [41]. It has been demonstrated that $\mathrm{L}{\mathrm{i}}_1\mathrm{Ti}{\mathrm{O}}_2$ composition can be obtained from anatase $\mathrm{Ti}{\mathrm{O}}_2$ at high temperature [30] or with particle size below 7 nm [42].

The discharge-charge profiles at current rates ranging from 1.0C to 20C and corresponding rate capability profiles are depicted in Figure 6a and b, respectively. It is worth noticing that after 10 cycles at 20C rate where capacity decreased to about 50 mAh/g, the capacity at 1C rate was recovered during the following 10 cycles to about 137 mAh/g. This indicates a high stability of the anatase $\mathrm{Ti}{\mathrm{O}}_2$ nanoparticles based anode toward high rate cycling.

Figure 6.

Rate capability (1C–20C) tests performed on Li/TiO₂ (pH 10) half-cells. (a) Discharge-charge voltage profiles, (b) discharge-charge capacity profile, and (c) cycle capacity and Coulombic efficiency profiles at 2C rate.

Figure 6c shows the cycle capacity profile at 2C rate for 300 cycles together with Coulombic efficiency. The latter stabilizes close to unity after the first 5 cycles, whereas the capacity gradually dropped to 82.4 mAh/g after 300 cycles indicating the high capacity retention and good cycling stability.

4. Conclusion

TiO₂ nanoparticles are synthesized by simple sol-gel method at different pH values and found to have particle size of 10–30 nm. XRD measurements show that the sample prepared at pH 10 crystallized to pure anatase phase. The TEM and the SEM images confirm the formation of nanoparticles with diameters of 10–30 nm. The $\mathrm{Ti}{\mathrm{O}}_2$ nanoparticles prepared at pH 10 exhibit initial capacity of 321 mAh/g and a remarkable high Coulombic efficiency close to unity after 300 cycles at 2C rate. However, the first cycle efficiency is relatively low perhaps due to inherent metastability of Li-poor phases at the early stage of lithium intercalation. More work has currently been performed to improve the first cycle efficiency, therefore making anatase nanoparticles based anode more practical in rechargeable lithium batteries with fast charge and high lithium storage capacity. Among possible routes considered to improve the NTO initial cycle efficiency is the incorporation of a few amounts of lithium during sol-gel process so as to achieve a pre-lithiated phase making it easier to add more lithium during the first cathodic process.