Li-intercalation property of mesoporous anatase-TiO2 synthesized by bicontinuous microemulsion-aided process

Solid State Ionics 176 (2005) 2361 – 2366 www.elsevier.com/locate/ssi

Li-intercalation property of mesoporous anatase-TiO2 synthesized by bicontinuous microemulsion-aided process Isamu Moriguchi a,b,*, Ryoji Hidaka a, Hirotoshi Yamada a, Tetsuichi Kudo a a

Department of Applied Chemistry, Faculty of Engineering, Nagasaki University, 1-14, Bunkyo-machi, Nagasaki, 852-8521, Japan b PREST, Japan Science and Technology Agency, 4-1-8 Honcho Kawaguchi, Saitama, Japan Received 2 December 2004; received in revised form 8 February 2005; accepted 9 February 2005

Abstract Mesoporous materials of anatase-TiO2 with different surface area and mesopore size were successfully synthesized by bicontinuous microemulsion-aided processes. The electrochemical lithium-intercalation into the mesoporous anatase-TiO2 was dependent on the surface area, crystallite size and mesopore size. The increase in surface area of mesoporous TiO2 resulted in an increase in the lithium-intercalation capacity. A mesoporous TiO2 with large mesopore size showed a relatively large lithium-intercalation capacity even at high charging rate. D 2005 Elsevier B.V. All rights reserved. PACS: 81.05.Rm; 84.60.Dn; 82.70.Kj Keywords: Mesoporous; Titanium dioxide; Lithium-intercalation; Bicontinuous microemulsion

1. Introduction There is growing interest in electrical/electrochemical energy storage devices with both high power and high energy densities from the viewpoint of possible application as auxiliary power sources for electric and/or hybrid electric vehicles [1,2]. Although lithium-ion batteries are attractive power storage devices with high energy density, but its power density is generally low due to a large polarization at high charging– discharging rate. The large polarization is thought to be ascribable to slow lithium diffusion in active material solid, increase in resistance of electrolyte and electric resistance of active materials with increasing the charging – discharging rate. Therefore, it is of importance to design and fabricate a nanostructured electrode which provides interconnected nanopaths for electrolyte ions transport and electron conduction in a quasi three-dimensional electrode interface. As such potential electrode materials, mesoporous

* Corresponding author. 1-14 Bunkyo-machi, Nagasaki 852-8521, Japan. Tel./fax: +81 95 819 2669. E-mail address: [email protected] (I. Moriguchi). 0167-2738/$ - see front matter D 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.ssi.2005.02.029

materials are much attractive because of the large surface area which decreases the current density per unit surface area, the thin wall shortened lithium-diffusion length in the solid phase, and mesopores which enable the electrolyte ions transport smoothly. As for TiO2, which is one of candidates for negative electrode materials, electrochemical lithium-intercalation properties of nanoparticulate and mesoporous films have been investigated by Kavan et al. [3,4]. However, the correlation between the nanoporous structure and intercalation capacity is still unclear. Recently we have succeeded in a synthesis of mesoporous anatase-TiO2 by a bicontinuous microemulsion-aided process, which is composed of a sol – gel reaction in a bicontinuous microemulsion formed in a water/surfactant/oil ternary system with a balanced hydrophilicity and lipophilicity, and heating the obtained gel to remove organics [5]. In the present work, mesoporous materials of anatase-TiO2 with various surface area and different pore size were synthesized by the bicontinuous microemulsion-aided process, and their electrochemical Liintercalation properties were studied in detail to investigate a correlation between mesoporous structure and lithiumintercalation property.

2362

I. Moriguchi et al. / Solid State Ionics 176 (2005) 2361 – 2366

2. Experimental

Anatase TiO2

A bicontinuous microemulsion was prepared by mixing didodecyldimethyammonium bromide (DDAB), hexane (C6) or tetradecane (C14) as an oil phase, and aqueous HCl (4  10 3 mol dm 3) with the wt.% of 34, 36 and 30, respectively. The composition of DDAB/oil/water used here is in the bicontinuous microemulsion region of phase diagram reported by Allen et al. [6]. The mixture was stirred for 30 min at room temperature until a transparent solution was obtained. Titanium tetrabutoxide (TTB) was added dropwise to the stirred bicontinuous microemulsion with a weight ratio of DDAB/TTB = 3.4:1.4. After further stirring for 30 s, the mixture solution was aged for gelation at 5 -C, 10 -C or 25 -C for 1 day in N2 atmosphere. The obtained gel was filtered off, washed thoroughly with hexane, dried in an oven at 110 -C for 12 h, and finally calcined at 300 – 380 -C for 6 h in air. The synthetic conditions were listed in Table 1. As shown in Table 1, the samples are denoted as TiO2(oil/T g/Tc), where oil, T g and Tc indicate kind of oil used as the oil phase of bicontinuous microemulsion (C6 or C14), gelation temperature (5 – 25 -C) and calcination temperature (300 – 380 -C), respectively. For example, TiO2(C6/25/350) means a TiO2 sample obtained at the condition of 25 -C gelation temperature and 350 -C calcination by using the bicontinuous microemulsion of C6/ DDAB/aq.HCl. In addition, a mesoporous TiO2 sample (TiO2-acac(C14/25/350)) was also synthesized by the same manner using acetylacetone-stabilized TTB (the molar ratio of acetylacetone/TTB = 3) as a Ti source at the same synthetic condition of TiO2(C14/25/350). 2.2. Instrumental analysis The carbon residue in the samples was estimated by elementary analysis (Perkin-Elmer 2400II analyzer). X-ray diffraction (XRD) patterns of samples were obtained on a Rigaku RINT-2200 diffractometer using Cu-Ka radiation. The crystallite size of TiO2 in the samples was estimated from the FWHM of XRD peaks by using Scherrer’s equation. The morphology of the porous structure was observed by trans-

(c)

Intensity / arb.unit

2.1. Synthesis of mesoporous TiO2

(b)

(a) 20

30

40

2θ / degree (CuKα)

50

60

Fig. 1. XRD patterns of (a) TiO2(C6/25/300), (b) TiO2(C6/25/350), and (c) TiO2(C6/25/380).

mission electron microscopy (TEM, JEOL JEM-100). Adsorption –desorption isotherms of N2 were measured at 77 K (Micromeritics Co. Ltd. Gemini 2370). The specific surface area and mesopore diameter distribution of samples were respectively analyzed by BET and BJH methods from the N2 adsorption branch. In addition, the surface area originating from mesopores was estimated from T-plot analysis [7]. Lithium insertion/extraction property of the TiO2 samples were investigated at room temperature by cyclic voltammetry and galvanostatic technique (Hokuto Denko, HZ-3000) using a sealed three-electrode cell equipped with metallic lithium counter and reference electrodes. A mixture of mesoporous TiO2 (5 mg), acetylene black (Denka Co. Ltd.) as an electron conductive additives (20 mg), and poly(tetrafluoroethylene) (PTFE) as a binder (2 mg) was pressed onto a nickel mesh and was used as a working electrode: the apparent electrode area and volume were ca. 1– 1.5 cm2 and ca. 0.1 cm3, respectively. Electrolyte was a 1 mol dm 3 solution of LiClO4 in PC + DME (1:1 by volume) (Kishida Chemical Co. Ltd.). To minimize the effect of the IR drop associated with the electrolyte resistance, the tip of a capillary in connection to the reference electrode was placed as close as possible to the working electrode. The lithium insertion/extraction measurement was carried out between 3.6 and 1.4 V vs. Li/Li+ for three cycles at a constant current density ranging from 0.168 to 6.72 A g 1 (1– 40 C).

Table 1 Synthetic conditions of bicontinuous microemulsion-aided process for mesoporous TiO2

3. Results and discussion

Samples

Ti source/Oil

Gelation temperature (-C)

Calcination temperature (-C)

3.1. Structural characterization of mesoporous TiO2

TiO2(C6/25/300) TiO2(C6/25/350) TiO2(C6/25/380) TiO2(C14/25/350) TiO2(C14/10/350) TiO2(C14/5/350) TiO2-acac(C6/25/350)

TTB/Hexane

25 25 25 25 10 5 25

300 350 380 350 350 350 350

TTB/Tetradecane

TTB(acac)/Hexane

The TiO2(C6/25/300) was slightly brownish, and TiO2(C6/ 25/350) and TiO2(C6/25/380) were colorless, meaning that almost all the organics were removed by the calcination above 350 -C for 6 h. The amount of carbon remained in these samples was estimated from elementary analysis to be below 10.7, 2.7 and 0.3 wt.%, respectively. On X-ray diffraction (XRD) measurements (Fig. 1), broad peaks assignable to

(i)

(dV / dD) / 10-1 cm3 g-1 nm-1

Amount of N2 / 102 cm3 g-1

I. Moriguchi et al. / Solid State Ionics 176 (2005) 2361 – 2366

(a)

2

(b) 1

(c) 0

0.2 0.4 0.6 0.8 1.0

Relative pressure, p/p0

(ii)

2363

(a)

(a) 1

(b) 50 nm

(c) 0

2

4

6

8

10

(b)

Pore diameter, D / nm

Fig. 2. (i) N2 adsorption – desorption isotherms and (ii) BJH pore size distributions of (a) TiO2(C6/25/300), (b) TiO2(C6/25/350), and (c) TiO2(C6/ 25/380). The closed and opened symbols indicate adsorption and desorption processes, respectively.

20 nm

anatase-TiO 2 were observed on TiO 2(C6/25/350) and TiO2(C6/25/380) although all the samples showed no distinct XRD peak at 2h/degree < 10. This indicates that the TiO2 samples did not possess an ordered mesoporous structure like as in micelle-templated mesoporous silica and TiO2 crystalline phase produced with heating above 350 -C. N2 adsorption –desorption isotherms and BJH pore size distribution of these samples are compared in Fig. 2(i) and (ii). The isotherms can be classified into type IV and the inflection characteristic of capillary condensation into mesopores is observed for all the TiO2(C6/25/300 – 380) samples at 0.2 < p/p 0 < 0.5, confirming the presence of mesopores. The parameters associated with the porous structure such as BET surface area, mesopore volume, pore size and crystallite size of samples were summarized in Table 2. With increasing the calcination temperature, the BET surface area and mesopore volume of samples decreased, and the mean mesopore diameter increased due to the disappearance of small size mesopores below 3 nm in diameter by the sintering. Fig. 3 shows TEM images of a precursor gel heated at 110 -C, TiO2(C6/25/300) and TiO2(C6/25/380). Fibrous continuous nanochannels with ca. 2 – 5 nm width, which consisted of TiO2-gel wall (black lines in the image) and surfactant (bright parts), were observed for the precursor sample (Fig. 3a). This means

(c)

50 nm Fig. 3. TEM images of (a) a gel precursor heated at 110 -C, (b) TiO2(C6/25/ 300), and (c) TiO2(C6/25/380).

that TiO2-gel formation occurred preferentially at the water/oil or water/surfactant interface in the bicontinuous microemulsion. On the other hand, a disordered mesoporous structure consisted of continuous nanochannels with 2 –4 nm width was confirmed on TiO2(C6/25/300) (Fig. 3b). For TiO2(C6/25/380), the observed mesoporous structure included nanocrystallites with the size of 5– 8 nm, which was almost consistent with that determined from the FWHM of XRD peaks of anatase-TiO2. These results suggest that fibrous nanochannels in the precursor, which probably reflect the organized structure of bicontin-

Table 2 The parameter associated with the porous structure of mesoporous TiO2 Samples

BET surface area (m2 g 1)a

Pore volume (cm3 g 1)

Pore size range (nm)

Mean pore size (nm)

Crystallite size (nm)

TiO2(C6/25/300) TiO2(C6/25/350) TiO2(C6/25/380) TiO2(C14/25/350) TiO2(C14/10/350) TiO2(C14/5/350) TiO2-acac(C14/25/350)

402 189 127 167 128 262 218

0.40 0.16 0.21 0.23 0.19 0.22 0.83

2–4 2–4 2–5 2–5 2–6 2–4 10 – 25

2.4 2.8 3.2 2.6 3.0 2.4 17

– 5 7 9 9 8 14

a

(–) (–) (104) (120) (124) (245) (–)

Values in parentheses indicate the mesopore surface area determined from T-plot analysis.

I. Moriguchi et al. / Solid State Ionics 176 (2005) 2361 – 2366

600 (dV / dD) / cm3 g-1 nm-1

uous microemulsion, changed into the mesoporous structure accompanying the shrinkage of mesopore with heating (removal of organics) and the disappearance of small size mesopores with sintering. The porous structural character of 350 -C-heated TiO2 samples was compared in order to investigate the effect of synthetic conditions such as aging temperature, kinds of oil and Ti source (Table 2). Irrespective of the same calcinations temperature, the TiO2(C14/25/350) showed a clearer XRD pattern of anatase-TiO2 and lower BET surface area in comparison with those of TiO2(C6/25/350), suggesting that the gelation in the bicontinuous microemulsion of C14 –oil system might proceed more than in C6 –oil system or the C14 residue in the gel might cause an increase in local heating temperature to accelerate the sintering during the calcination treatment. Among the TiO2(C14/5 – 25/350) samples, TiO2(C14/5/350) possessed the highest BET surface area and a narrow pore size distribution (Fig. 4). The aging temperature of 5 -C is lower than the melting point of C14 (5.5 -C) and the freezing of the bicontinuous microemulsion solution was confirmed visually, thus the frozen C14 – oil phase suppresses the structural change of fluid bicontinuous microemulsion with the gelation, and the interconnected nanochannel structure would be reflected precisely on the porous structure of TiO2(C14/5/350). The effect of Ti-source was also investigated. N2 adsorption – desorption isotherms and BJH pore size distribution of TiO2-acac(C6/25/350) sample were shown in Fig. 5. A clear inflection due to the capillary condensation was observed around p/p 0 = 0.8 –0.9 and the TiO2-acac(C6/ 25/350) sample has higher surface area and larger mesopore size than those of TiO2(C6/25/350), might indicating the gelation speed affects on the mesoporous structure. The acetylacetone-stabilized TTB is more stable against the hydrolysis and condensation reaction than TTB, thus the microemulsion structure might be changeable more during the prolonged gelation time and it would cause the formation of larger mesopores. There would be also a possibility of changing the microemulsion structure with the co-existence of acetylacetone. These will be clarified with a

Amount of N2 / cm3 g-1

2364

500 400 300

0.04 0.03 0.02 0.01

0

200

0

10

20

30

40

50

Pore diameter / nm

100 0

0

0.2

0.4

0.6

0.8

Relative pressure, P/P0 Fig. 5. N2 adsorption – desorption isotherms of TiO2-acac(C6/25/350); the inset shows the BJH pore size distribution. The closed and opened symbols indicate adsorption and desorption processes, respectively.

further study in future. TEM observation showed that the TiO2-acac(C6/25/350) sample was a mesoporous material composed of TiO2 nanocrystallites with the size around 15 nm (Fig. 6). These results indicate a possibility of controlling mesoporous structure of anatase-TiO2 by the choice of synthetic conditions in bicontinuous microemulsion-aided process. It is noteworthy that disordered mesoporous TiO2 with relatively high surface area could be obtained by the present method and the mesoporous structure was retained to some extent even if TiO2 nanocrystallites produced in the mesoporous phase. 3.2. Li-intercalation into mesoporous TiO2 The lithium insertion/extraction to anatase-TiO2 proceeds according to the following reversible reaction: TiO2 + xLi + xe = Lix TiO2, the maximum x for the reversible reaction at room temperature is reported to be 0.5 [8], which corresponds to a capacity of 168 mAh g 1. Fig. 7 shows galvanostatic lithium insertion/extraction curves of TiO2(C6/ 25/300) and TiO2(C6/25/380) at the charging rate of 1 C (= 0.168 A g 1). With increasing the calcination temperature, the plateau regions around 1.7 V (insertion process)

Pore volume / cm3 g-1 nm-1

0.10

(a) 0.08 0.06

(b) 0.04

(c)

0.02 0

1

2

3

4

5

6

7

8

9 10

Pore diameter / nm Fig. 4. BJH pore size distributions of (a) TiO2(C14/5/350), (b) TiO2(C14/10/ 350), and (c) TiO2(C14/25/350).

1

100 nm Fig. 6. TEM image of TiO2-acac(C6/25/350).

I. Moriguchi et al. / Solid State Ionics 176 (2005) 2361 – 2366

0.6

Capacity / mA h g-1 100

150

3.0 2.5

5C 0.4

10 C 0.3

2.0

0.2

1.5

0.1

0.0

0.2

0.4

0.6

∆x in LixTiO2

1C

0.5

b c

a

3.5

200

0.0 50

100

150

200

Mesopore surface area / Fig. 7. Galvanostatic Li-insertion/extraction curves at 1 C of (a) TiO2(C6/ 25/300) and (b) TiO2(C6/25/380).

and 2.0 V (extraction process) appeared. Since the plateaus are related to the phase transition between tetragonal and orthorhombic with Li-intercalation into anatase-TiO2 [9], the appearance of the plateau with confirmed again the formation of anatase-TiO2 phase in mesoporous samples. The lithium-intercalation capacity was determined by subtracting the electric double layer capacity, which was estimated from the charging plateau region around 3– 3.5 V vs. Li/Li+ in cyclic voltammogram, from the capacity of third extraction curve on galvanostatic measurements since the insertion capacity was larger than that of the extraction process probably due to the irreversible over-insertion into TiO2 or a decomposition of electrolyte solution. Fig. 8 shows lithium-intercalation capacities of mesoporous TiO2 obtained by the calcination above 350 -C as a function of charging rate. The capacity decreased steeply with increasing the charging rate for all the samples, suggesting that a polarization due to electrolyte resistance and/or electric resistance of TiO2 increased with increasing the charging rate. In Fig. 9, the capacities at the constant charging rate Current density / A g 0.0

0.5

1.0

were plotted against the specific surface area originating from mesopores, which was estimated by T-plot analysis of N2 adsorption isotherms. Since TiO2 crystallite size (7– 9 nm) and mesopore size (2– 5 nm) were almost the same in these samples, the increase in the lithium-intercalation capacity is ascribable mainly to increasing the mesopore surface area of TiO2. It can be concluded that the increase in mesopore surface area of active materials promotes electrochemical lithium-intercalation reaction probably due to a suppression of the polarization per unit surface area. In order to investigate an effect of pore size on the lithium-intercalation, the charging rate-dependence of intercalation capacity was compared between TiO2(C6/25/350) and TiO2-acac(C6/25/350) samples (Fig. 10). At low charging rate below 10 C, the capacity of TiO2(C6/25/350) was larger than that of TiO2-acac(C6/25/350). With increasing the charging rate, TiO2(C6/25/350) showed a steep decrease in the capacity more than for TiO2-acac(C6/25/ Current density / A g -1

-1

0

1

2

4

3

5

6

200

0.6

1.5

200

TiO2(C6 /25/350)

0.5

0.3

100

0.2 50 0.1

150

LiX in Li X TiO 2

0.4

Capacity / mA h g-1

150

∆x in LixTiO2

300

g-1

Fig. 9. Effect of mesopore surface area on the lithium intercalation capacity of TiO2(C14/5 – 25/350) and TiO2(C6/25/380): the symbols are the same as in Fig. 8.

0.5

0

250

m2

0.4 100

0.3

TiO2-acac(C6 /25/350)

0.2

50

Capacity / mA h g -1

50

∆x in Li xTiO2

Potential / (V vs Li/Li+)

0

2365

0.1 2

4

6

8

10

0

C-rate

0 0

10

20

30

0 40

C-rate Fig. 8. The charging rate-dependency of lithium intercalation capacity of TiO2(C14/5 – 25/350) and TiO2(C6/25/380): (h) TiO2(C14/5/350), (6) TiO2(C14/10/350), (g) TiO2(C14/25/350), (‚) TiO2(C6/25/380).

Fig. 10. The charging rate-dependency of lithium intercalation capacity of TiO2(C6/25/350) (?) and TiO2-acac(C6/25/350) (>).

2366

I. Moriguchi et al. / Solid State Ionics 176 (2005) 2361 – 2366

350). As a result, TiO2-acac(C6/25/350) showed a lithiumintercalation capacity larger than that of TiO2 (C6/25/350) at high charging rate above 10 C. The high capacity of TiO2(C6/25/350) at the low charging rate in this case would be due to the small crystallite size that means the shortened diffusion length in solid phase since the surface area of TiO2(C6/25/350) is smaller than that of TiO2-acac(C6/25/ 350). At high charging rate, an increase in electrolyte resistance in the TiO2-acac(C6/25/350) electrode would be suppressed more than that in the TiO2(C6/25/350) electrode, that is, mesoporous materials with large pore size are favorable to a smooth transport of electrolyte ions. These results are supported by the previously reported study on lithium-intercalation into colloidal crystal-templated porous TiO2 electrodes, in which macroporous TiO2 electrodes showed relatively large lithium-intercalation capacities at high charging rate [10,11]. An evaluation of cycle performance of the present mesoporous systems are now in progress.

4. Conclusion Mesoporous materials of anatase-TiO2 with different surface area and mesopore size were successfully synthesized by the bicontinuous microemulsion-aided process. The electrochemical lithium-intercalation into the mesoporous anatase-TiO2 was dependent on the surface area, crystallite size and mesopore size. The increase in surface area of mesoporous TiO2 resulted in an increase in the lithiumintercalation capacity presumably due to decreasing polarization per unit surface area. The intercalation property of steep decrease in capacity with increasing charging rate, which is generally observed, was improved on a mesoporous TiO2 with a relatively large pore size. These results indicate that it is important to design and fabricate a mesoporous electrode with high surface area and interconnected nanopaths, the size of which is effective to electrolyte ions transport. The bicontinuous microemulsion-

aided process would be one of the promising processes for the optimization of mesoporous electrode structure. Here we focused on the study concerning a correlation between mesoporous structure of active material and the polarization accompanied with electrochemical lithium-intercalation reaction. The study is now being extended to the development of porous electrodes with both ion transport nanochannels and electron conducting paths.

Acknowledgements The study made use of instruments (elementary analysis, XRD, and TEM) in the Center for Instruments Analysis of Nagasaki University. This work was partly supported by Grant-in-Aid for Scientific Research from Ministry of Education, Culture, Science, Sports, and Technology of Japan, and Nippon Sheet Glass Foundation.

References [1] B. Scrosati, Nature 373 (1995) 557. [2] R.F. Nelson, J. Power Sources 91 (2001) 2. [3] L. Kavan, M. Graztel, J. Rathousky, A. Zukal, J. Electrochem. Soc. 143 (1973) 394. [4] L. Kavan, J. Rathousky, M. Gratzel, V. Shklover, A. Zukal, J. Phys. Chem., B 104 (2000) 12012. [5] I. Moriguchi, Y. Katsuki, H. Yamada, T. Kudo, T. Nishimi, Chem. Lett. 33 (2004) 1102. [6] M. Allen, D.F. Evans, D.J. Mitchell, B.W. Ninham, J. Phys. Chem. 91 (1987) 2320. [7] J.C.P. Broekhoff, B.G. Linsen, Physical and Chemical Aspects of Adsorbents and Catalysis: Chap. 2, Academic Press, New York, 1970. [8] H. Lindstrom, S. Sodergrem, A. Solbrand, H. Rensmo, J. Hjelm, A. Hagfeldt, S.E. Lindquist, J. Phys. Chem., B 101 (1997) 7717. [9] R. van de Krol, A. Goossens, E.A. Meulkenkamp, J. Electrochem. Soc. 146 (1990) 3150. [10] H. Yamada, T. Yamato, I. Moriguchi, T. Kudo, Solid State Ion. 174 (2004) 195. [11] H. Yamada, T. Yamato, I. Moriguchi, T. Kudo, Chem. Lett. 33 (2004) 1548.