Spectroelectrochemical Properties and Lithium Ion Storage in Self

Layer-by-layer (LbL) nanocomposite films from TiO2 nanoparticles and tungsten-based oxides (WOxHy), as well as dip-coating films of TiO2 nanoparticles...
0 downloads 0 Views 2MB Size
Download PDF
pubs.acs.org/Langmuir © 2009 American Chemical Society

Spectroelectrochemical Properties and Lithium Ion Storage in Self-Assembled Nanocomposites from TiO2 Tiago Facci and Fritz Huguenin* Departamento de Quı´mica, Faculdade de Filosofia, Ci^ encias e Letras de Ribeir~ ao Preto, Universidade de S~ ao Paulo, 14040-901 Ribeir~ ao Preto (SP), Brazil Received September 2, 2009. Revised Manuscript Received November 12, 2009 Layer-by-layer (LbL) nanocomposite films from TiO2 nanoparticles and tungsten-based oxides (WOxHy), as well as dip-coating films of TiO2 nanoparticles, were prepared and investigated by electrochemical techniques under visible light beams, aiming to evaluate the lithium ion storage and chromogenic properties. Atomic force microscopy (AFM) images were obtained for morphological characterization of the surface of the materials, which have similar roughness. Cyclic voltammetry and chronoamperometry measurements indicated high storage capacity of lithium ions in the LbL nanocomposite compared with the dip-coating film, which was attributed to the faster lithium ion diffusion rate within the self-assembled matrix. On the basis of the data obtained from galvanostatic intermittent titration technique (GITT), the values of lithium ion diffusion coefficient (DLi) for TiO2/WOxHy were larger compared with those for TiO2. The rate of the coloration front in the matrices was investigated using a spectroelectrochemical method based on GITT, allowing the determination of the “optical” diffusion coefficient (Dop) as a function of the amount of lithium ions previously inserted into the matrices. The values of DLi and Dop suggested the existence of phases with distinct contribution to lithium ion diffusion rates and electrochromic efficiency. Moreover, these results aided a better understanding of the temporal change of current density and absorbance during the ionic electro-insertion, which is important for the possible application of these materials in lithium ion batteries and electrohromic devices.

Introduction Titanium dioxide has been extensively investigated and applied in several devices because of properties such as high chemical stability, nontoxicity, and low cost. Particularly in the field of electrochemistry, anatase TiO2 electrodes have been used in several applications like solar cells,1,2 sensors,3 photoelectrochemical degradation,4 and electrocatalysis.5 Moreover, this metal oxide is also relevant for electrochromic devices and lithium ion batteries because of its capacity to store charge, together with a relatively high insertion/deinsertion rate.6-9 Among the TiO2 polymorphs, rutile, brookite, TiO2-B (bronze), and anatase have been investigated as electrodes for lithium storage. However, host matrices formed from bulk anatase TiO2 are generally considered to be most electroactive among them.7 In electrochromism, changes in absorption (ΔA) take place in the visible region of the electromagnetic spectrum under an external electric field.10 Among the transition metal oxides, WO3 has been the most studied as electrochromic material *Corresponding author. E-mail: [email protected]. (1) Grabulosa, A.; Beley, M.; Gros, P. C.; Cazzanti, S.; Caramori, S.; Bignozzi, C. A. Inorg. Chem. 2009, 48, 8030. (2) O0 Regan, B.; Gr€atzel, M. Nature 1991, 353, 737. (3) Wen, D.; Guo, S. J.; Zhai, J. F.; Deng, L.; Ren, W.; Dong, S. J. J. Phys. Chem. C 2009, 113, 13023. (4) Kim, C.; Kim, J. T.; Kim, K. S.; Jeong, S.; Kim, H. Y.; Han, Y. S. Electrochim. Acta 2009, 54, 5715. (5) Li, P. Q.; Zhao, G. H.; Cui, X.; Zhang, Y. G.; Tang, Y. T. J. Phys. Chem. C 2009, 113, 2375. (6) Cant~ao, M. P.; Cisneros, J. I.; Torresi, R. M. J. Phys. Chem. 1994, 98, 4865. (7) Yang, Z.; Choi, D.; Kerisit, S.; Rosso, K. M.; Wang, D.; Zhang, J.; Graff, G.; Liu, J. J. Power Sources 2009, 192, 588. (8) Exnar, E.; Kavan, L.; Huang, S. Y.; Gr€atzel, M. J. Power Sources 1997, 68, 720. (9) Wagemaker, M.; Kentgens, A. P. M.; Mulder, F. M. Nature 2002, 418, 397. (10) Grandqvist, C. G. Handbook of Inorganic Electrochromic Materials; Elsevier: Amsterdam, 1995. (11) White, C. M.; Gillaspie, D. T.; Whitney, E.; Lee, S.-H.; Dillon, A. C. Thin Solid Films 2009, 517, 3596.

Langmuir 2010, 26(6), 4489–4496

because it exhibits high electrochromic efficiency.11-13 However, other metal oxides have also been investigated, such as those formed from Mo, Ir, Nb, V, Ni, and Ti, aiming at enhancing electrochromic efficiency and, reducing the time needed for the materials to change color, among other aspects.12,14 In the case of TiO2, there is a reversible coloration change from white to blue associated with the Ti4þ reduction taking place concomitantly with lithium ion insertion in the host matrix. Depending on the preparation method, films of anatase TiO2 can display high electrochromic efficiency and low response time,15 allowing its use in smart windows. In the case of lithium batteries, there are some additional advantages to the application of TiO2 as negative electrode: the stability of electrolytic solutions is guaranteed and the risk of explosion is decreased compared with carbon and metallic lithium. This makes TiO2 a suitable candidate for use in 2 V lithium battery, which can be applied in microelectronic and consumer devices with photovoltaic recharging. However, several parameters can still be changed to improve its electrochemical performance, which depends on the details of electrode nanoarchitecture. Gao et al. reported high charge capacity (340 mA h g-1) for the anatase nanotubes;16 Armstrong et al. prepared TiO2-B nanowires with high specific capacity (305 mA h g-1) and with rate of lithium ion insertion/deinsertion rate higher than that obtained in the case of anatase nanoparticles and bulk TiO2-B;17 (12) Granqvist, C. G. Sol. Energy Mater. Sol. Cells 2007, 91, 1529. (13) Deb, S. K. Appl. Opt. Suppl. 1969, 3, 192. (14) Monk, P. M. S.; Mortimer, R. J.; Rosseinsky, D. R. Eds. Electrochromism: Fundamentals and Applications; VCH: Weinheim, Germany, 1995. (15) Tebby, Z.; Babot, O.; Toupance, T.; Park, D.-H.; Campet, G.; Delville, M.-H. Chem. Mater. 2008, 20, 7260. (16) Gao, X. P.; Lan, Y.; Zhu, H. Y.; Liu, J. W.; Ge, Y. P.; Wu, F.; Song, D. Y. Electrochem. Solid State Lett. 1999, 2, 184. (17) Armstrong, A. R.; Armstrong, G.; Canales, J.; Garcia, R.; Bruce, P. G. Adv. Mater. 2005, 17, 862.

Published on Web 12/01/2009

DOI: 10.1021/la903301c

4489

Article

Facci and Huguenin

Jiang et al. reported that, in the case of nanocrystalline rutile electrode, the charge capacity of the first cycle was increased more than 3-fold when the particle size decreased from 300 nm to 15 or 30 nm.18 Considering that reactions take place in the entire volume of the electrode, and that the internal surface area in these nanomaterials is enhanced compared with the conventional oxide electrodes, significant changes in the electrochemical properties of TiO2 polymorphs are expected. In the particular case of batteries and electrochromic devices, the limiting-rate step of the electrochemical processes is normally lithium ion diffusion, and an enhancement in the internal surface area can decrease the ionic diffusion pathway within the host matrices. So the charge capacity of the lithium batteries and the response time of electrochromic devices can be increased and reduced, respectively.7,8 Another way of improving the properties of TiO2 is the formation of hybrid materials. Composites of titanium- and tungsten-based oxide can be an alternative in a number of cases: photoelectrochemical anticorrosion system,19 photocatalyst for energy store,20 photoinduced electron storage,21 and sensor.22 Moreover, reversibility, charge capacity, optical density, and lifetime can be enhanced in materials based on TiO2, WO3, and K6[P2W18O62] 3 14H2O.23-26 An approach to manufacturing these composites is to employ the layer-by-layer (LbL) technique, which allows high control of the thickness, uniformity, and nanoarchitecture of thin-films with new electrochemical and chromogenic properties.27,28 An intimate contact between the components in the self-assembled composite can change the interaction forces between these components and lithium ion electro-inserted in the composite,29-31 which can enhance lithium ion mobility and, consequently, increase the charge capacity and decrease the response time. In this work, our interest by on the evaluation of the properties associated with the lithium ion electro-insertion into LbL films of WOxHy and TiO2 nanoparticles, since these materials can be an alternative to TiO2 in electrochromic devices and 2 V lithium batteries. Charge capacity, electrochromic efficiency, lithium ion diffusion rate and the rate of coloration front were analyzed by means of spectroelectrochemical data.

Experimental Section The colloidal dispersion of anatase TiO2 was prepared by hydrolysis of tetra-n-butyl titanate from Du-Pont (Tyzor). A mixture of 30 mL organic titanate and 30 mL 2-propanol was slowly added (15 min) to 300 mL deionized water under vigorous (18) Jiang, C.; Honma, I.; Kudo, T.; Shou, H. Electrochem. Solid State Lett. 2007, 10, A127. (19) Tatsuma, T.; Saitoh, S.; Ohko, Y.; Fujishima, A. Chem. Mater. 2001, 13, 2838. (20) Tatsuma, T.; Saitoh, S.; Ngaotrakanwiwat, P.; Ohko, Y.; Fujishima, A. Langmuir 2002, 18, 7777. (21) Zhao, D.; Chen, C. C.; Yu, C. L.; Ma, W. H.; Zhao, J. C. J. Phys, Chem. C 2009, 113, 13160. (22) Faia, P. M.; Ferreira, A. J.; Furtado, C. S. Sens. Actuators B 2009, 140, 128. (23) Benoit, A.; Paramasivan, I.; Nah, Y.-C.; Roy, P.; Schmuki, P. Electrochem. Commun. 2009, 11, 728. (24) Livage, J.; Guzman, G. Solid State Ionics 1996, 84, 205. (25) Huguenin, F.; Zucolotto, V.; Carvalho, A. J. F.; Gonzalez, E. R.; Oliveira, O. N. Chem. Mater. 2005, 17, 6739. (26) Liu, S.; Xu, L.; Gao, G.; Xu, B.; Guo, W. Mater. Chem. Phys. 2009, 116, 88. (27) Galiote, N. A.; Carvalho, A. J. F.; Huguenin, F. J. Phys. Chem. B 2006, 110, 24612. (28) Huguenin, F.; Nart, F. C.; Gonzalez, E. R.; Oliveira, O. N. J. Phys. Chem. B 2004, 108, 18919. (29) Huguenin, F.; Torresi, R. M. J. Phys. Chem. C 2008, 112, 2202. (30) Huguenin, F.; Ticianelli, E. A.; Torresi, R. M. Electrochim. Acta 2002, 47, 3179. (31) Varela, H.; Huguenin, F.; Malta, M.; Torresi, R. M. Quim. Nova 2002, 25, 287.

4490 DOI: 10.1021/la903301c

stirring. Then, 2 mL 70% nitric acid was added to this mixture (under stirring). The resulting solution was stirred for 2 h at room temperature. The mixture was then heated to 80 °C under stirring for 4 h, giving rise to a stable and transparent (slightly cloudy) TiO2 suspension (pH = 2).27 The effective diameter of the colloidal TiO2 particles was 17.3 nm, as determined by dynamic light scattering (DLS) with a BI-200SM system (Brookhaven Instruments Corp.). The dispersion was cooled, displayed no changes in its transparency, and remained stable for at least 12 months. The WOxHy dispersion was prepared as follows: WO3 3 2H2O (Vetec) and n-butylamine (Riedel-de H€aen) were added to 50 mL of deionized water at a 10:1 molar ratio. An NH4OH aqueous solution was added, to adjust pH at 6.5, thereby dissolving the precipitate. The pH was adjusted again at 2.0, and the obtained dispersion was maintained at 10 °C for 3 days and passed through a poly(vinylidene fluoride) (PVDF) membrane filter (Gelman) with a pore diameter of 200 nm. Here, 15-bilayer LbL films from TiO2/WOxHy were assembled onto a fluorine doped tin-oxide (FTO) coated glass purchased from Flexitec (Curitiba, Brazil). The glass had a sheet resistance e20 Ω and a geometrical area of 1 cm2. The layers were obtained via ionic attraction of oppositely charged materials, by alternate 1 min immersions of the FTO substrate into the polycationic (TiO2) and anionic (WOxHy) dispersions. After each layer deposition, the substrates were rinsed in HCl solution (pH = 2) for 30 s, and they were then dried under nitrogen flow at room temperature. These films were heated at 150 °C for 12 h. The dip-coating method was used for preparation of films from TiO2, with an immersion and submersion rate of 20 mm/min. For each deposited layer, the FTO substrate was kept immersed into the TiO2 dispersion for 1 min, and the drying time was 2 min. This process was repeated, and the dip-coating films consisting of 15 layers were submitted to the same thermal treatment used in the case of LbL films. The anatase phase in the films prepared by the colloidal dispersion was detected by X-ray diffraction (XRD) performed in the grazing incidence mode. The XRD pattern of thin films, obtained after casting the TiO2 dispersions onto the glass substrate, was recorded on a Siemens D5005 diffractometer using monochromatic Cu KR radiation. The diffractogram displayed two diffraction peaks (with low intensity) at 2θ values of 25° and 48°, which are characteristic of (101) and (200) of anatase TiO2 nanoparticles, respectively.32 On the other hand, no diffraction peaks were observed for the LbL films due to their small thickness. Thus, a mixture of TiO2 and WOxHy dispersions (1:0.5, w/w) was prepared, and the XRD pattern of a thin film obtained after casting this mixture onto the glass substrate displayed diffraction peaks similar to those obtained in the case of anatase TiO2. This suggests that the structure of crystalline anatase nanoparticles does not changed in the presence of the amorphous phase of WOxHy in the LbL film of TiO2/WOxHy. Film thickness was analyzed by specular reflectance using the Nanocalc 2000 program coupled with a single channel 2048 pixel CCD spectrophotometer with halogen lamp as light source. The thickness values measured for the dip-coating films of TiO2 and the LbL films of TiO2/WOxHy were 50 ( 4 nm and 70 ( 4 nm, respectively. The geometrical area of both films was 1 cm2. Atomic force microscopy (AFM) height images were obtained on a digital Shimadzu microscope. In order to investigate the Ti/W ratio in the TiO2-WO3 films (7.43 ( 0.6 mol of Ti for each mol of W), energy dispersive spectrometry (EDS) analysis was performed on five different regions of the samples. The measurements were acquired in an IXRF (model 500 Digital Processing) spectrometer coupled to a Zeiss (model EVO 50) scanning electron microscope (SEM). For the electrochemical and spectroelectrochemical experiments, the TiO2 or TiO2/WOxHy films with an area of 1 cm2 (32) Ghicov, A.; Tsuchiya, H.; Hahn, R.; Macak, J. M.; Mun~oz, A. G.; Schmuki, P. Electrochem. Commun. 2006, 8, 528.

Langmuir 2010, 26(6), 4489–4496

Facci and Huguenin

Article

Figure 2. In situ UV-vis spectra for the (a) TiO2 and (b) TiO2/ WOxHy films at (;Δ;) þ1.0, (;2;) -1.25, (;O;) -1.5, (;b;) -1.6, (;0;) -1.7, and (;9;) -1.8 V vs Ag/Agþ.

Figure 1. AFM height images of the (a) dip-coating film from TiO2 and (b) 15-bilayer LbL film from TiO2/WOxHy. Scale bar = 100 nm. deposited on FTO were used as the working electrode, a platinum sheet with an area of 10 cm2 was used as the counter electrode, and Ag/Agþ in propylene carbonate (PC) and LiClO4/PC 0.5 mol L-1 was used as the quasi-reference electrode. A LiClO4/PC electrolytic solution (0.5 mol L-1) was used in all the electrochemical experiments, which were carried out using an Autolab PGSTAT30 potentiostat/galvanostat. Chromogenic analysis was carried out concomitant with the electrochemical experiments by means of an microprocessor-controlled solid-state light source (WPI, Inc.). Plastic fiber optical cables up to 1 mm in diameter were used to deliver light from the instrument to a PDA1 photodiode amplifier (WPI, Inc.). An USB4000 spectrophotometer (Ocean Optics Inc.) with a xenon light source was also employed, to obtain the electromagnetic spectrum at several applied potentials. For the transmission experiments, the films were placed in a cell made of optical glass, where light beams at fixed wavelengths were transmitted across the film during the electrochemical experiments.

Results and Discussion Evaluation of the morphology and surface roughness of the films is important for comparison of the properties associated with lithium ion insertion/deinsertion, such as specific capacity and diffusion rate, once this guest can access electroactive sites both on the surface and in the interior of the host matrices. Figure 1 shows the AFM height images for the (a) TiO2 and (b) TiO2/WOxHy films. These images reveal the presence of roughly spherical colloidal particles with low polydispersivity and ca. 25 nm of diameter, which is close to that determined for the TiO2 colloidal dispersion employed in the preparation of both Langmuir 2010, 26(6), 4489–4496

films. Moreover, there is formation of aggregates with ca. 200 nm diameter, probably due to evaporation of water molecules from the colloidal dispersion, thereby contributing to an increase in the interaction forces between the nanoparticles. The values of smoothing surface roughness (Ra) of the TiO2 and TiO2/WOxHy films are close: 16 and 14 nm, respectively, indicating their similar roughness and surface areas. Furthermore, these results also demonstrate that the difference in the electrochemical and spectroelectrochemical properties associated with the lithium ion diffusion process will depend on the properties associated with the interior of the matrices, and will not be influenced by the surface area of them. Figure 2 presents the absorption change (ΔA) spectra for (a) TiO2 and (b) TiO2/WOxHy films in 0.5 mol L-1 LiClO4/PC at þ1.0, -1.25, -1.5, -1.6, -1.7, and -1.8 V vs Ag/Agþ. These materials display an absorption band in the visible region between 550 and 780 nm when reduced, which can be assigned to absorption by nearly free electrons.33 Comparing the absorption change spectra, the profiles of the TiO2 and TiO2/WOxHy films are different at the same potentials. Additionally, a red shift for the absorbance band maximum is observed for TiO2/WOxHy compared with TiO2, which suggests enhancement in electronic mobility. As also seen for other materials formed from titanium and tungsten oxides,34 this result indicates that the presence of WOxHy in the self-assembled material changes the electronic distribution of TiO2. Figure 3a depicts the potentiodynamic profile of the current density (j) for the TiO2 film in 1.0 mol L-1 LiClO4/PC at several scan rates (v). These cyclic voltammograms demonstrate the lithium ion insertion process during the negative potential scan (33) Enright, B.; Fitzmaurice, D. J. Phys. Chem. 1996, 100, 1027. (34) He, Y.; Wu, Z.; Fu, L.; Li, C.; Miao, Y.; Cao, L.; Fan, H.; Zou, B. Chem. Mater. 2003, 15, 4039.

DOI: 10.1021/la903301c

4491

Article

Facci and Huguenin

Figure 3. Potentiodynamic profile of (a) j and (b) ΔA for the TiO2 film; potentiodynamic profile of (c) j and (d) ΔA for the TiO2/WOxHy film. Scan rates: 5, 10, 20, 40, 50, 60, 80, and 100 mV s-1. The arrow points to increasing scan rates from 5 to 100 mV s-1.

Figure 4. Charge/discharge curves for (b) the TiO2 and (O) TiO2/ WOxHy films at 50 μA.cm-2.

and the lithium ion deinsertion process during the positive potential scan. The voltammetric peaks indicate the presence of the phase anatase of the TiO2 nanoparticles for both materials.15 The reduction and oxidation current peak potentials change as a function of the scan rate. In spite of the small thickness of the films, there is slow lithium ion diffusion into the host matrices, as will be shown below. Measurements of absorbance changes were taken simultaneously with these voltammograms at 660 nm (Figure 3b). The increase in the values of ΔA at lower scan rates (longer time) is associated with a larger amount of Ti3þ and inserted lithium ion. Moreover, these data also reveal a high chemical reversibility, once negligible difference is observed in the absorbance change before and after the voltammetric cycle, thus indicating that the lithium ions inserted during the negative potential scan are deinserted in the positive potential scan. Figure 3 also shows the potentiodynamic profile of (c) the current density and (d) the absorbance change at 660 nm for the TiO2/WOxHy film, in 1.0 mol L-1 LiClO4/PC at several scan rates. The current peak potentials change as a function of the scan 4492 DOI: 10.1021/la903301c

Figure 5. Plot of the values of (b) q/qv=10 mV/s and (O) ΔA/ ΔAv=10 mV/s ratios, and (;;) fitting data for (a) the TiO2 and (b) TiO2/WOxHy films.

rate and the ionic insertion/deinsertion process is highly reversible, as also observed in the case of the TiO2 film. However, note that the j and ΔA values for the self-assembled electrode are larger than those achieved in the case of the TiO2 film. In spite of the difference in the thickness (and volume) of the films and the probable variation in the amount of electroactive sites, the values of current density and inserted charge (q) normalized by volume Langmuir 2010, 26(6), 4489–4496

Facci and Huguenin

Article

Figure 7. (a) Potential and (b) absorbance changes as a function of the time during (O) and after (;O;) a constant current pulse of 20 μA cm-2 for 15 s, for 10-bilayer LbL films from TiO2/WOxHy. Figure 6. (a) Complex plane representation of the capacitance and (b) Nyquist diagram obtained by applying a dc potential of -1.7 V for the (b) TiO2 and (O) TiO2/WOxHy film. The inset of Figure 6b shows the transmission line that can be used to represent the impedance data (figure adapted from ref 39). Rs is the electrolytic solution resistance, Cdl is the double-layer capacitance, r is the resistance associated with lithium ion transport, rct is a resistance associated with charge transfer at the electrode/solution interface, and Cb is the capacitance associated with the charging of lithium ions into the bulk sites.

are also larger for the self-assembled TiO2/WOxHy film. On the basis of the voltammograms at 50 mV s-1, the oxidation current peak normalized by the volume was 18.9 ( 0.4 and 22.0 ( 0.6 A cm-3 for the TiO2 and TiO2/WOxHy films, respectively, and the oxidation charge normalized by the volume was 255 ( 10 and 300 ( 10 C cm-3 for the TiO2 and TiO2/WOxHy films, respectively. Figure 4 shows the charge/discharge curves for the TiO2 and TiO2/WOxHy films in 1.0 mol L-1 LiClO4/PC at a constant charge/discharge current of 50 μA cm-2. The charge inserted in TiO2 and TiO2/WOxHy films were 3.3 and 5.5 mC cm-2, respectively. After 100 charge/discharge cycles, the charge decreases by ca. 3% for the TiO2 film, and by 4% for the TiO2/ WOxHy film. The charge storage capabilities of the TiO2 and TiO2/WOxHy films were 660 ( 20 and 785 ( 20 C cm-3 in the first discharge, respectively. As will be shown below, these differences in the specific capacity of the electrodes observed in the cyclic voltammetry and charge/discharge curves are also associated with the rate of lithium ion diffusion into the host matrices. This enhancement in charge capacity has also been observed for composites formed from nanostructured anatase and rutile TiO2. High capacity to store lithium ions of nanostructured electrodes from anatase TiO2/RuO2, anatase TiO2/funcionalized graphene, and rutile TiO2/funcionalized graphene have been Langmuir 2010, 26(6), 4489–4496

attributed to shortened ionic diffusion pathway and facilitated electron transport.35,36 Figure 5 demonstrates how q/qv=10 mV/s (q for different scan rates normalized with respect to q at 10 mV s-1) and ΔA/ ΔAv=10 mV/s (ΔA for different scan rates normalized with respect to ΔA at 20 mV s-1) vary as a function of the scan rate (v) for the (a) TiO2 and (b) TiO2/WOxHy films. First, these data indicate that the lithium ions accessed not only sites on the surface of the electrodes, but were also inserted into the host matrices. The absorbance change and the injected charge decreased from 56.6  10-3 and 370 mC cm-3 at 10 mV s-1 to 28.4  10-3 and 204 mC cm-3 at 100 mV s-1 for the TiO2 electrode, respectively. In the case of the TiO2/WOxHy electrode, the absorbance change and the injected charge decreased from 73.4  10-3 and 484 mC cm-3 at 10 mV s-1 to 34.2  10-3 and 239 mC cm-3 at 100 mV s-1. The outer sites of the host matrices can be estimated using the voltammetric charge obtained at several scan rates, according to the method proposed by Trasatti et al.37 However, another methodology was used here, once lithium ion diffusion and migration occur in the host matrices during the cyclic voltammetry. The q/qv=10 mV/s ratio as a function of the scan rate was fitted with the empirical exponential function 0.45355 þ 0.53922 exp(-v/59.54895) þ 0.84333 exp(-v/4.48227) for TiO2, and 0.40671 þ 0.46831 exp(-v/58.66685) þ 0.51732 exp(-v/10.41437) for TiO2/WOxHy. As the electronic mobility is much larger than the ionic mobility, the lithium ions access the outer sites at a higher scan rate. Extrapolating these curves to an infinite scan rate, the value of q/qv=10 mV/s multiplied by qv=10 mV/s gives the (35) Guo, Y.-G.; Hu, Y.-S.; Sigle, W.; Maier, J. Adv. Mater. 2007, 19, 2087. (36) Wang, D.; Choi, D.; Li, J.; Yang, Z.; Nie, Z.; Kou, R.; Hu, D.; Wang, C.; Saraf, L. V.; Zhang, J.; Aksay, I. A.; Liu, J. ACS Nano 2009, 3, 907. (37) Ardizzone, S.; Fregonara, S.; Trasatti, S. Electrochim. Acta 1990, 35, 263.

DOI: 10.1021/la903301c

4493

Article

Facci and Huguenin

Figure 8. Potential changes as a function of the time square root during constant current pulses of 15 μA cm-2 for the (a) TiO2 and (b) TiO2/ WOxHy films. Equilibrium potentials as a function of the inserted charges for the (c) TiO2 and (d) TiO2/WOxHy films. Plots of absorbance changes as a function of time during constant current pulses of 15 μA cm-2 for the (e) TiO2 and (f) TiO2/WOxHy films. The sequence of current pulses, from the first to the seventh, is represented by the following order of symbols: 9, 0, b, O, 2, Δ, and 1.

charge associated with the outer sites, which was ca. 0.84 and 1.38 mC cm-2 for TiO2 and TiO2/WOxHy, respectively. Employing the same procedure for the ΔA/ΔAv=10 mV/s ratio versus scan rate, the empirical exponential functions 0.43911 þ 0.58113 exp(-v/41.70783) þ 0.97567 exp(-v/4.46011) and 0.39711 þ 0.52133 exp(-v/48.08787) þ 0.60302 exp(-v/ 8.24817) were obtained for TiO2 and TiO2/WOxHy, respectively. Also, by extrapolation of these curves to an infinite scan rate, the value of ΔA/ΔAv=10 mV/s multiplied by qv=10 mV/s allows estimation of the faradaic charge associated with the outer sites, which was ca. 0.81 and 1.35 mC cm-2 for TiO2 and TiO2/WOxHy, respectively. On the basis of these values and the total charge (sum of faradaic and capacitive charges) associated with outer sites, the capacitive charge and the integral capacitance (q/ΔE) were 4494 DOI: 10.1021/la903301c

determined for both electrodes: 0.03 mC cm-2 and 11 μF cm-2, respectively. Figure 6a shows the complex diagram of capacitance in which the TiO2 and TiO2/WOxHy films were subjected to dc potentials of -1.7 V, with 5 mV of superimposed ac amplitude. The beginning of a semicircle at high frequency is associated with loading of the electrical double-layer. By extrapolating this semicircle at low frequency and intercepting the real axis, the values of double-layer capacitance can be determined, which are close to 10 μF cm-2. This value of differential capacitance is close to the integral capacitance, indicating the validity of the employed method. Thus, the current and charge are almost totally associated with the faradaic process involving the Ti4þ/Ti3þ electro-reduction/oxidation and the lithium ion Langmuir 2010, 26(6), 4489–4496

Facci and Huguenin

electro-insertion/deinsertion, although the films investigated in this work are thin. This allows the direct determination of the electrochromic efficiency (η = ΔA/q), molar absorptivity (ε = Fη), and “optical diffusion coefficient”, according to the methods employed in this work. The values of η are 31 and 25 cm2 C-1 for TiO2 and TiO2/WOxHy at 50 mV s-1, respectively, which are close to those reported in the literature.15,38 The ε values are 3.0  106 and 2.4  106 cm2.mol-1 for TiO2 and TiO2/WOxHy at 50 mV s-1, respectively. Figure 6b shows the Nyquist diagrams for the TiO2 and TiO2/ WOxHy films. There is a straight line between 2.5 Hz and 60 mHz, which is associated with the lithium ion semi-infinite diffusion process, which can also be observed in the complex diagram of capacitance of Figure 6a. At lower frequencies, there is a curvature in this straight line, which can be associated with ion trapping, as mentioned in the literature for intercalation electrodes.25,39,40 Inset of Figure 6b shows the transmission line that can be used to represent the impedance data associated with trapping effects, according to Bueno et al.39 These host matrices are assumed to comprise regions separated by energy barriers of different heights, and therefore the lithium ion takes longer to overcome some of the barriers (trapping) than others (diffusion). However, there is a tendency toward the formation of a straight line almost parallel to the imaginary axis at the lowest frequencies in the case of the TiO2/WOxHy film, which is associated with finite diffusion processes, thus indicating a higher mobility of the charge carrier. Figure 7 displays (a) the potential and (b) the absorbance changes as a function of time under a monochromatic radiation at 660 nm, during a current pulse of 20 μA cm-2 (15 s) and later at open circuit potential, for the TiO2 electrode. The potential change is associated with the Ti4þ reduction and lithium ion semi-infinite diffusion into the matrix.27,41 It is noteworthy that the absorbance increases and decreases as a function of time during and after current pulse, respectively, and tends to the initial value. This behavior is attributed to the dependence of absorptivity on lithium ion concentration in the host matrix, probably due to changes in the chemical environment associated with the ionic concentration gradient formed during the current pulse. Lithium ion accumulation close to the film/electrolytic solution interface is attributed to its low mobility, and ionic distribution in the host matrix tends to be uniform at equilibrium. Considering that the coloration wave accompanies the lithium ion diffusion during the ionic insertion/deinsertion process into the host matrices,6 the spectroelectrochemical experiment shown above can be used to calculate the lithium ion diffusion coefficient (DLi) and the “optical” diffusion coefficient (Dop), on the basis of difference in absorptivity between the transient and equilibrium states.41 This procedure was repeated at several degrees of charging for both films: Figure 8a and 8b show the potential as a function of the time square root for the TiO2 and TiO2/WOxHy films, respectively; potential as a function of the charge injected at the equilibrium state for the TiO2 and TiO2/WOxHy films, respectively; absorbance temporal changes during each current pulse for the TiO2 and TiO2/WOxHy films, respectively (some plots are not shown for better visualization). On the bais of the dE/dt0.5 and dA/dt slopes of each of the pulses and on the values of dE/dq and dA/dq similar to that at the steady-state, one can determine the DLi and Dop values, according to eq 1 and 2, (38) Wang, C.-M.; Lin, S.-Y.; Chen, Y.-C. J. Phys. Chem. Solids 2008, 69, 451. (39) Bueno, P. R.; Leite, E. R. J. Phys. Chem. B 2003, 107, 8868. (40) Bisquert, J. Electrochim. Acta 2002, 47, 2435. (41) Rezende, A. R.; Bizeto, M. A.; Constantino, V. R. L.; Huguenin, F. J. Phys. Chem. C 2009, 113, 10868.

Langmuir 2010, 26(6), 4489–4496

Article

Figure 9. Logarithm of (b) DLi and (O) Dop values as a function of the injected charge for the (a) TiO2 and (b) TiO2/WOxHy films.

respectively .41,42 2

DLi

 32 dE 46 6 7 dq 7 7 ¼ 6iL 4 5 dE π pffi d t 2

Dop



3 dA 2 6 6 iL dq 7 7 ¼ 6pffiffiffiffiffi 7 4 tπ dA 5 dt

ð1Þ



ð2Þ

where L is the thickness of the films and i is the current pulse. The values of dE/dt0.5 and dA/dt were determined for times longer than 4 s, which has shown a linear relation between potential and the time square root, guaranteeing semi-infinite diffusion into the host matrix only. Figure 9 shows the DLi and Dop values as a function of the injected charge for the (a) TiO2 and (b) TiO2/WOxHy films. The tendency of DLi values to decrease as a function of the amount of lithium ions is attributed to the interactions between them, as observed in other insertion electrodes.29 Note that the DLi values are larger for the TiO2/WOxHy film compared with the TiO2 film for low inserted charge, which indicates that the presence of WOxHy enhances ionic mobility. Although the presence of WOxHy did not provoke significant changes in the chromogenic properties, the ionic diffusion rate increased in the electrode

(42) Weppner, W.; Huggins, R. A. J. Electrochem. Soc. 1977, 124, 1569.

DOI: 10.1021/la903301c

4495

Article

Facci and Huguenin

short times, indicating the highest lithium ion rate into phase, which contributes more significantly to absorbance change. Moreover, the difference between these slopes is smaller for the TiO2/WOxHy film for short times compared with the TiO2 film, which is also observed for the difference between the DLi and Dop values in the case of a small amount of injected charge. However, other experiments are still necessary for a better understanding of the influence of local lithium ion environment on the transport properties in these materials. Examples of such experiments can be done by nuclear magnetic resonance (NMR) and X-ray absorption spectroscopy (XAS).30,44 Concerning future applications in electrochromic devices, other measurements are essential for better evaluation of the use of these electrodes. For example, thicker films containing a higher amount of electroactive sites must be analyzed, aiming at investigating the charge capacity, electrochromic efficiency, and transport properties under the contrast ratio desired for these applications.

Conclusions

Figure 10. Temporal profile of (;b;) charge and (;O;) absorbance change normalized by respective maximum values during measurements of potential step (between 1.0 V and -1.7 V) for the (a) TiO2 and (b) TiO2/WOxHy films. qmax = 1.55 mC cm-2 and ΔAmax = 45  10-3 for the TiO2 film; and qmax = 3.11 mC cm-2 and ΔAmax = 85  10-3 for the TiO2/WOxHy film.

composed of WOxHy, contributing to an increase in the charge capacity compared with the TiO2 electrode. The fastest ionic diffusion process in the nanocomposite can be associated with the increase in electronic mobility, as suggested by the spectroelectrochemical results. As inserted lithium ion and electron are transported simultaneously to maintain electroneutrality within the host matrices, the electron accelerates the ionic motion. Moreover, based on the electrochemical impedance spectroscopy data shown above, the trapping effect can also be one of the factors responsible for the difference in diffusion coefficients. It is noteworthy that the Dop values of optical diffusion coefficient are larger compared with lithium ion diffusion coefficient (except for charge density close to zero in the case of TiO2/ WOxHy). A possible explanation would be the coexistence of two phases, one of which contributes more significantly to the absorbance changes.43 This is corroborated by measurements of potential step for TiO2 (Figure 10a) and TiO2/WOxHy (Figure 10b), which show the temporal profile of q/qmax and ΔA/ΔAmax (current and absorbance change normalized by current maximum and absorbance maximum change, respectively) under a potential step from þ1.0 to -1.7 V. Note that the slope of the ΔA/ΔAmax curve is larger than that of the q/qmax curve for (43) van Driel, F.; Decker, F.; Simone, F.; Pennisi, A. J. Electroanal. Chem. 2002, 537, 125.

4496 DOI: 10.1021/la903301c

Dip-coating films of TiO2 and self-assembled films from TiO2/ WOxHy were prepared, aiming at their possible application as lithium ion insertion electrodes in batteries and electrochromic devices. On the basis of AFM images, the roughness factors for the materials were determined, which revealed that their real surface areas are similar. So the surface of the electrodes is not responsible for the difference in the electrochemical and electrochromic properties evaluated in this work. The DLi values were obtained for several amounts of lithium ion previously inserted into the host matrices. There was an increase in the ionic diffusion rate in the TiO2/WOxHy nanocomposite compared with the TiO2 film, due to the presence of WOxHy in the self-assembled structure, which contributed to enhancing its charge storage capacity. Aiming at a better understanding of the coloration wave rate associated with the lithium ion electro-insertion in the films based on TiO2, “optical” diffusion coefficient was determined by the GITT-based spectroelectrochemical method. The Dop values were higher than the DLi values, except when q ≈ 0 in TiO2/ WOxHy, suggesting the presence of two phases with distinct contribution to absorbance changes and lithium ion diffusion rate. Measurements of potential step corroborate these data, which reveal that the coloration wave rate is higher than the electric current associated with ionic insertion into these host matrices. Although this work is associated with fundamental aspects, it is important to mention that other experiments are still necessary for a better evaluation of the use of these materials in lithium batteries and electrochromic devices. Moreover, novel synthetic routes for film preparation and the production of TiO2based nanocomposites can also be an alternative to improving the charge storage and electrochromic properties of these materials. Acknowledgment. We are grateful to FAPESP and CNPq for financial support. We are also grateful to Prof. Jose Maurı´ cio Rosolen, Prof. Rogeria Rocha Gonc-alves (DQ/FFCLRP/USP), and Prof. Antonio J. F. Carvalho (UFScar) for the availability of their laboratories and material preparation. (44) Holland, G. P.; Yarger, J. L.; Buttry, D. A.; Huguenin, F.; Torresi, R. M. J. Electrochem. Soc. 2003, 150, A1718.

Langmuir 2010, 26(6), 4489–4496