Redox Couple - American Chemical Society

Nov 2, 2006 - and Central Research Laboratories, Idemitsu Kosan Co., Ltd., 1280 Kami-izumi, Sodegaura,. Chiba 299-0293, Japan. ReceiVed: July 14, 2006...
0 downloads 0 Views 331KB Size
J. Phys. Chem. B 2006, 110, 23321-23328

23321

Bulk Heterojunction Photoelectrochemical Cells Consisting of Oxotitanyl Phthalocyanine Nanoporous Films and I3-/I- Redox Couple Katsuyoshi Hoshino,*,† Yusuke Hirasawa,† Sang-Kook Kim,† Tetsuo Saji,‡ and Jun-ichi Katano§ Faculty of Engineering, Chiba UniVersity, 1-33 Yayoi-cho, Inage-ku, Chiba 263-8522, Japan, Department of Chemistry and Materials Science, Tokyo Institute of Technology, Ohokayama, Meguro, Tokyo 152-8550, Japan, and Central Research Laboratories, Idemitsu Kosan Co., Ltd., 1280 Kami-izumi, Sodegaura, Chiba 299-0293, Japan ReceiVed: July 14, 2006; In Final Form: September 20, 2006

Photoelectrochemical cells based on oxotitanylphthalocyanine (TiOPc) films and an I3-/I- redox couple have been constructed. The TiOPc films were prepared on an indium-tin oxide coated glass plate (ITO) by the micellar disruption method and characterized by their unique nanoporous structure. A photocurrent action spectrum for input radiation directed through the ITO/TiOPc film, film-thickness dependence, and morphological investigation revealed that the cells consisted of a bulk heterojunction formed between the nanoporous TiOPc films and the liquid I3-/I- electrolyte, resulting in a larger short-circuit current (Jsc ) 2.1 mA/cm2), open-circuit voltage (Voc ) 0.11 V), fill factor (ff ) 0.31), and hence a larger energy conversion efficiency (η ) 0.13% for an incident white-light intensity of 53 mW/cm2) than the bilayer structure composed of the vaccum-evaporated TiOPc compact film and the I3-/I- electrolyte (Jsc ) 0.16 mA/cm2, Voc ) 0.018 V, ff ) 0.27, and η ) (1.5 × 10-3)%).

1. Introduction Significant progress in organic photovoltaic cells has been realized by the use of an interpenetrating network of the donor and acceptor materials.1 Such progress is considered to be due to the introduction of the donor-acceptor heterojunction throughout the film bulk that functions as a dissociation site for the photogenerated excitons. This is because the limiting factor in achieving high-energy conversion efficiency is the short exciton diffusion length in organic materials, and therefore, the bulk donor-acceptor heterojunctions which minimize exciton quenching while maximizing the absorption of light may be preferred over the dual donor-acceptor heterojunction2 cells. On the other hand, organic photoelectrochemical cells (wet photovoltaic cells)3 have attracted much interest because highenergy conversion efficiencies are expected if the transparent electrolyte solutions are used as a front electrode despite the ease of their fabrication. However, the wet photovoltaic cells of dual organic-electrolyte structure would have the similar drawbacks to the dry dual donor-acceptor photovoltaic cells, i.e., higher exciton quenching efficiency and lower absorption of light at the junction site.4 If we prepare a porous film structure of semiconducting materials, an interpenetrating semiconductor-electrolyte interface is created throughout the film bulk and high-energy conversion efficiency is expected. Two decades ago, one of the authors developed a unique film-formation technique called the “Micellar Disruption (MD)” method.5 The details of its film-formation mechanism, procedures, materials used, etc. have appeared elsewhere.6 The MD method is based on the dispersion of film-forming materials in an aqueous redox-active * To whom correspondence should be addressed. E-mail: k_hoshino@ faculty.chiba-u.jp. Fax: +81-43-290-3478. † Chiba University. ‡ Tokyo Institute of Technology. § Idemitsu Kosan Co., Ltd.

surfactant system and the subsequent deposition of the materials on the substrate electrode in a film form by the electrode redox reactions of the surfactants. This method is characterized by its wide applicability to organic materials because any materials can be used if they are dispersed in the surfactant system and film formation is carried out under ordinary pressure and temperature, and by the unique nanoporous structure of the resulting films. Though the MD method has good film-forming features, its technological application is limited to the preparation of color filters for liquid crystal displays7 except for its use in the photoelectrochemical cell, ITO/metal-free phthalocyanine (H2Pc), I3--I-/ITO.8 This photoelectrochemical cell showed a relatively high-energy conversion efficiency (η ) 0.06% for an incident light power of 6 mW/cm2), and therefore, the MD method was considered to be promising for the preparation of photovoltaic cells; however, no strategy has been taken further for using highly efficient photoconducting materials and understanding the mechanistic aspects of the relatively high efficiency. This paper reports the preparation and characterization of the unique nanoporous films of oxotitanylphthalocyanine (TiOPc) and copper phthalocyanine (CuPc) by the MD method and their application to wet photovoltaic cells based on the phthalocyanineI3-/I- redox couple. TiOPc is known as a highly photosensitive pigment and is promising for use in electrophotographic photoreceptors because of its high sensitivity, high stability, and nontoxicity.9 Maximum energy conversion efficiencies of 0.13 (for incident white-light power ) 53 mW/cm2) and 0.0094% (for incident light power ) 55 mW/cm2) are obtained for the devices with use of ITO/TiOPc, I3--I-/ΙΤΟ and ITO/CuPc, I3-I-/ΙΤΟ structures, respectively. These values are respectively 87- and 260-fold higher than those of analogous wet cells having the TiOPc and CuPc photocathodes prepared by the vacuum evaporation (VE) method. The mechanistic study reveals that

10.1021/jp0644659 CCC: $33.50 © 2006 American Chemical Society Published on Web 11/02/2006

23322 J. Phys. Chem. B, Vol. 110, No. 46, 2006

Hoshino et al.

TABLE 1: Characterization of the TiOPc and CuPc Dispersions pigment

dw (µm)

[pigment] (mM)

tind (s)

TiOPc CuPc

0.050 0.088

12.2 6.6

110 3.8

the better performance of the former photocathodes is based on their unique structure in which the interpenetrating phthalocyanine film/liquid I3--I- redox couple is created. 2. Experimental Section 2.1. Materials. As the film-forming materials, TiOPc (Dainichiseika Kogyo Co.) and β-type CuPc (β-CuPc, BASF Japan Co.) were used. The TiOPc powder was a mixture of R and β polymorphs, which was confirmed by X-ray diffraction, and the mole ratio of the former to the latter was 3:2. The surfactant used to disperse the pigments in water was (11-ferrocenylundecyl)polyoxyethyleneglycol (FPEG; Dojin Chemical Co.).10 The films were deposited on a transparent indium-tin oxide coated glass plate (ITO; Kuramoto Co., 15 Ω/sq) that was ultrasonicated in acetone for 10 min and then allowed to dry at ambient temperature before use. 2.2. Preparation of Pigment Dispersions. The preparations of the phthalocyanine dispersions11 and the film-formation procedures6 have already been described elsewhere, but will be outlined here. The TiOPc and CuPc dispersions were prepared by the following successive steps: (1) 60 mM TiOPc or 10 mM CuPc was added to a stirred FPEG micellar solution (2 mM, 50 mL), (2) the suspension was sonicated for 20 min with ice cooling (600 W, Nihonseiki Co. model US-600), (3) the resultant dispersion was allowed to stand for 24 h at room temperature and then the supernatant was separated, and (4) to the supernatant, LiBr (0.1 M) was added as a supporting electrolyte, and the mixture was stirred for ca. 20 min to give a test dispersion, which involved FPEG-pigment aggregates, FPEG monomers, and FPEG micelles. The concentration of the pigment in the test dispersion, [pigment], was determined as follows. By the same method described above, 1 mM of pigment was completely dispersed in the FPEG (2 mM) micellar solution containing 0.1 M LiBr, and the resultant dispersion was diluted to appropriate concentrations. Their visible absorption spectra gave the calibration curves between the absorption maximum (744 nm for TiOPc and 709 nm for β-CuPc) and the pigment concentration. On the other hand, the test dispersion was also diluted to an appropriate concentration and its UV-vis spectrum was recorded. The concentration of the pigment in the test dispersion was determined by using the above calibration curve. The values of [pigment] thus determined are listed in Table 1 for the TiOPc and CuPc dispersions. 2.3. Film Formation Procedures. The electrodeposition of the phthalocyanine films on the ITO was carried out by controlled potential electrooxidation of the test dispersions, using a three-electrode cell consisting of two compartments and a potentiostat (BAS Co., model ALS 750A). The ITO (15 mm × 35 mm × 1.1 mm) and the counter Pt plate (10 mm × 20 mm × 0.3 mm) were immersed in the dispersion (ca. 15 mL) in the main compartment, the distance between the two plates being ca. 1 cm. The area of the ITO exposed to the dispersion was typically 15 mm × 25 mm. In the auxiliary compartment (ca. 5 mL) was immersed an agar bridge connected to a reference Ag/AgCl electrode (BAS Co., model RE-1B). The potential of the ITO was maintained at +0.50 V by considering the halfwave potential of FPEG (+0.28 V vs Ag/AgCl).5 The obtained

Figure 1. Schematic illustration of the phthalocyanine photoelectrochemical cell employed in the present study.

films were washed with distilled-deionized water and ethanol after being dried under air for 1 day. The mechanism of the film formation is speculated to be as follows:6a (i) The free FPEG surfactants (FPEG monomer + FPEG micelle) diffuse to the ITO electrode surface and are oxidized to their cations (FPEG+). The concentration of the free surfactants in the vicinity of the electrode decreases to less than the critical micelle concentration (cmc). (ii) The surfactants adsorbed on the pigment particles are desorbed from the pigment surface to satisfy the adsorption equilibrium. This desorption leads to the deposition of the pigment on the electrode, which occurs effectively when the concentration of the free surfactants decreases to less than the cmc. (iii) After the electrode is covered with the pigment film, the free surfactant diffuses in the film due to the existence of small spaces among the particles in the film, eventually reaches the electrode surface, and then is electrolyzed. The concentration of the free surfactants in the vicinity of the film is kept at less than the cmc, so that the film continues to grow to a respectable thickness. The cmc value of FPEG was previously determined by the dye solubilization (8 µM) and the surface tension methods (10 µM). 2.4. Characterization of Dispersions and Films. The particle size (distribution) of the pigment in the test dispersion was measured with a centrifugal particle analyzer (Shimazu Co., model SA-CP4L). The values of the weight-averaged particle diameter (dw) are summarized in Table 1. Scanning electron micrographic (SEM, Topcon model ABT-32) and atomic force micrographic measurements (AFM, Seiko instruments model Nanopics 1000) were done to observe the morphology of the films. The thicknesses of the films prepared by the MD method were spectrophotometrically determined: The absorption coefficients of the films were calculated from the slope of the film thickness versus the absorbance, where the film thicknesses were measured by the SEM observations. The values of the absorption coefficients thus determined are as follows: TiOPc, 2.44 × 104 (λmax ) 665 nm) and 2.37 × 104 cm-1 (λmax ) 750 nm); CuPc, 2.77 × 104 (λmax ) 612 nm) and 2.82 × 104 cm-1 (λmax ) 709 nm). The thicknesses of the films prepared by the VE method were monitored and controlled by a quartz oscillator thickness monitor (ULVAC CRTM-6000) during the evaporation. 2.5. Photoelectrochemical Measurements. Photoelectrochemical measurements were carried out with the three-electrode arrangement (Figure 1) in which the potential is referred to an Ag/AgCl reference electrode. When the potential is referenced to the equilibrium potential of the I3-/I- redox system, the data correspond to those measured with the two-electrode arrangement. The photovoltaic data presented in this study are based on the two-electrode arrangement. The photovoltaic measure-

Bulk Heterojunction Photoelectrochemical Cells

J. Phys. Chem. B, Vol. 110, No. 46, 2006 23323

Figure 2. Dependence of thickness on the square root of the deposition time, t1/2, for TiOPc (a) and CuPc films (b).

ments were carried out by using an electrochemical analyzer (BAS Co., model ALS 750A) under the illumination of white light from a 500-W xenon lamp (Ushio Electric Co., UI-502Q). Its intensity was measured by a pyranometer (Eiko Co., model MS-601). An action spectrum was obtained by plotting the external quantum efficiency (or collection efficiency) against the wavelength of incident light passed through a set of interference filters and an IR-cut filter (Toshiba IRA-25S). The intensity of the monochromatic light was measured by a silicon photodiode (Advantest Co., model TQ8210). 3. Results and Discussion 3.1. Film Formation of TiOPc and CuPc by the MD Method. Successful formation of phthalocyanine films and their photoelectrochemical cells should require the knowledge of their deposition behavior. Hence, the growth of the films was investigated in some detail. The dependencies of film thickness on the deposition time (electrolysis time), t, for the TiOPc and CuPc films were investigated. Though not shown here, the thicknesses increased with t, but their increase was followed by a decrease in the deposition rate. This deposition behavior can be expected for the film-formation mechanism, which involves the diffusion of pigment-FPEG aggregates to the ITO, desorption of FPEG from the pigment particles, and its oxidation at the ITO, followed by the deposition of the pigment particles to form the film. If we assume that the film formation is limited by the diffusion, the flux of the aggregates at the ITO surface, J(t), is written as12

J(t) )

D1/2[pigment] π1/2t1/2

(1)

where D and [pigment] are the diffusion coefficient and the concentration of the aggregates in water (pigment concentration), respectively. Equation 1 is based on Fick’s first and second laws, and involves the assumptions that the dispersion is homogeneous before the deposition experiment, that the regions sufficiently distant from the ITO are unperturbed by the experiment, and that [pigment] ) 0 at the ITO for t > 0. The integral of eq 1 from t ) 0 gives the coverage of the deposited film, Γ(t):

Γ(t) )

2D1/2[pigment]t1/2 π1/2

(2)

Figure 3. Current density-potential curves for ITO/TiOPc(MD, 580 nm), I3--I-/ΙΤΟ (a) and ITO/TiOPc(VE, 470 nm), I3--I-/ΙΤΟ (b) cells in the dark (dotted curve) and under illumination of white light (solid curve). Light intensity, I: (a) 53 mW/cm2; (b) 51 mW/cm2. Sweep rate of potential: 20 mV/s. Redox potential of I3--I- measured: 0.30 V vs Ag/AgCl.

Figure 2 show the plots of thickness versus t1/2 for the TiOPc (a) and CuPc (b) films. Though the thickness is plotted on the ordinate instead of Γ, both of the plots gave straight lines independent of the kind of pigments. This demonstrates not only that the film formation is controlled by the diffusion, but also that the packing density of the films is kept constant during their growth. Another notable feature of the plots in Figure 2 is a positive intercept on the t1/2 axis. The value of t derived from the intercept stands for an induction period (tind) during which the step (i) described in section 2.3 and ref 6a may occur. The values of tind are given in the last column in Table 1. tind is an indication of the dispersion stability to electrooxidation and may be dependent on the chemistry of the dispersion, e.g., the concentration of free FPEG (FPEG monomers and micelles), the rate of desorption of FPEG from the FPEG-pigment aggregate, etc. The chemistry of the aggregate stabilization coupled with the redox reactions may be far from simple, and a full account of tind would require a detailed description of the equilibrium among the FPEG monomers, FPEG micelles, and FPEGpigment aggregates, and the kinetics of adsorption-desorption of the FPEG on and from the pigment surface, which are not yet clear. 3.2. Photovoltaic Properties of Photoelectrochemical Cells. The configurations of the phthalocyanine photoelectrochemical cells are schematically depicted in Figure 1. The cell consists of a thin film phthalocyanine electrode, an ITO counter electrode, and an aqueous I3-/I- electrolyte. The cell body was made of Teflon with a circular window (diameter: 10 mm). The size of the electrodes was 15 mm × 35 mm with the

23324 J. Phys. Chem. B, Vol. 110, No. 46, 2006

Hoshino et al.

Figure 4. Dependence of Jsc (a), Voc (b), ff (c), and η (d) of the ITO/TiOPc(MD, 580 nm), I3--I-/ITO cell on incident white-light intensity, I.

illuminated area (the area exposed to the electrolyte solution) being 0.78 cm2. The area of the phthalocyanine films was 15 mm × 25 mm. The cell was always illuminated through the ITO/phthalocyanine side by a 500-W xenon lamp. Contact among the phthalocyanine electrode, the cell body, and the counter ITO electrode was ensured by the application of pressure, using stainless fittings and O-rings (see Figure 1). The current density (J)-potential (V) characteristics in the dark and under white-light illumination (Figure 3) were measured on the two types of cells with the TiOPc layer prepared by the MD method, ITO/TiOPc(MD, 580 nm), I3-I-/ΙΤΟ (a), and the VE method, ITO/TiOPc(VE, 470 nm), I3-I-/ΙΤΟ (b). The numerical values designated in parentheses indicate the film thickness of the TiOPc. The light intensities (I) employed for the former and the latter illumination were 53 and 51 mW/cm2, respectively. The dotted and solid curves indicate the dark current and photocurrent, respectively. In the case of the present three-electrode system, the short-circuit photocurrent (Jsc) is the photocurrent at the equilibrium potential of the I3-/I- redox couple (Eeq ) 0.30 V vs Ag/AgCl),13 and the open-circuit photovoltage (Voc) is the potential difference between the potential at which the photocurrent-potential curve intersects the abscissa and the Eeq. The larger dark current flowed when the ITO/TiOPc was biased positively with respect to Eeq than when it was biased negatively, suggesting positivecharge (hole) transport from the TiOPc layer into the I3--Isolution. Additionally, the cells showed larger photoresponse under cathodic potential with respect to Eeq, which means negative-charge (electron) transport from the TiOPc layer into

TABLE 2: Performances of the Cells Shown in Figure 3 photocathode in the cell Jsca (mA/cm2) Voca (v) ITO/TiOPc(MD)b ITO/TiOPc(VE)c ITO/CuPc(MD)d ITO/CuPc(VE)e

2.1 0.16 0.52 0.033

0.11 0.018 0.034 0.0023

ffa

ηa (%)

0.31 0.27 0.29 0.24

1.3 × 10-1 1.5 × 10-3 9.4 × 10-3 3.7 × 10-5

a Jsc, short-circuit photocurrent; Voc, open-circuit photovoltage; ff, fill factor; η, energy conversion efficiency. b Intensity of white light I ) 53 mW/cm2. c I ) 51 mW/cm2. d I ) 55 mW/cm2. e I ) 51 mW/ cm2.

the I3--I- solution. These features indicate that the TiOPc layer functions as a p-type semiconductor in the I3--I- solution. The cell with the ITO/TiOPc(MD) electrode demonstrated much better performances compared with that with the ITO/TiOPc(VE) electrode. In the former cell, larger cathodic photocurrents flowed as the potential became cathodic and saturated at potentials more cathodic than Eeq, while smaller photocurrents were observed with the latter cell. The cell performances are summarized in Table 2. On the basis of the data on Jsc, Voc, and fill factor (ff), the energy conversion efficiencies (η) for the white-light illumination were calculated to be 0.13% and (1.5 × 10-3)% for the former and the latter photocathodes, respectively. The η value of the cell with the ITO/TiOPc(MD) photocathode was 87 times that of the cell with the ITO/TiOPc(VE) photocathode. A similar tendency was observed for the performances of the ITO/CuPc, I3--I-/ITO cells. The photovoltaic parameters derived from the J-V curves under illumination, though the curves are not shown here, are listed in Table

Bulk Heterojunction Photoelectrochemical Cells

J. Phys. Chem. B, Vol. 110, No. 46, 2006 23325

Figure 5. Visible absorption spectrum of TiOPc film (solid curve) and photocurrent action spectrum (IPCE, open circles) for the TiOPc(MD, 580 nm)-based photoelectrochemical cell illuminated through the ITO/TiOPc side.

Figure 7. SEM (a and b) and AFM images (c and d) for the TiOPc(MD) film (a and c) and the TiOPc(VE) film (b and d). Scale bars in the SEM images: 500 nm. The particle size seen in image a reflects that in the TiOPc dispersion (see Table 1).

Figure 6. Schematic illustration of the nanoporous TiOPc(MD) film permeated by an I3-/I- aqueous solution (a) and the Schottky-like barrier formed at the TiOPC grain/solution interface (b). The abbreviations in the figure are as follows: C.B., conduction band; V.B., valence band.

2. In this case also, the cell with a photocathode prepared by the MD method, ITO/CuPc(MD, 650 nm), I3--I-/ITO, exhibited larger Jsc, Voc, and ff, and, therefore, a larger η value than the cell prepared by the VE method, ITO/CuPc(VE, 600 nm), I3-I-/ITO. However, energy conversion efficiency, η, decreased with the increasing intensity of the white light, I: The η value of the cell with the ITO/TiOPc(MD) photocathode was 0.081% for I ) 100 mW/cm2. Figure 4 shows the dependences of Jsc (a), Voc (b), ff (c), and η (d) on I. The values of Voc and ff depend little on I, but Jsc shows a stronger light-intensity dependence: The Jsc value is proportional to I at lower light intensities (up to ca. 50 mW/cm2), but the Jsc tends to deviate downward with increasing I beyond ca. 50 mW/cm2. This

indicates that the decrease in η at higher light intensities is attributed primarily to a limitation of Jsc. This limitation may be explained in terms of the increase in the rate of recombination of charge carriers and the lowering of the quantum efficiency.14 3.3. Spectral Response and Film-Thickness Dependence. To understand the better photovoltaic performances of the ITO/ pigment(MD) than the ITO/pigment(VE), the spectral response and film-thickness dependence were investigated for the ITO/ TiOPc(MD), I3--I-/ΙΤO cell. The incident monochromatic photons relative to the current conversion efficiency (IPCE), defined as the number of electrons generated per number of incident photons, was calculated from the photocurrents by means of the following equation:4

IPCE )

1240Jsc λI

(3)

where Jsc is the short-circuit current (A/cm2), λ the excitation wavelength (nm), and I the intensity of monochromatic light (W/cm2). Figure 5 shows the action spectra obtained upon monochromatic light illumination from the ITO/TiOPc side, together with the visible absorption spectrum of the TiOPc film on the ITO. The shape of the absorption spectrum agreed well with the superposition of the spectra of R-TiOPc and β-TiOPc films15 which were prepared by vacuum evaporation and subsequent treatment with tetrahydrofuran and xylene, respectively. The action spectrum is very similar to the absorption spectrum. If the TiOPc film is so compact that the I3-/I- redox

23326 J. Phys. Chem. B, Vol. 110, No. 46, 2006

Hoshino et al.

Figure 8. Dependence of Jsc (a), Voc (b), ff (c), and η (d) of the ITO/TiOPc(MD), I3--I-/ITO cell on the thickness of the TiOPc(MD) film, d. I ) 53 ( 3 mW/cm2.

couple cannot penetrate into it, this similarity would be obtained in the case where the thickness of the TiOPc film is smaller than the exciton diffusion length (lex). However, the thickness of the TiOPc film (580 nm) is much larger than the lex for phthalocyanine compounds (e.g., lex ) 30 nm for zinc phthalocyanine,16 lex < 24 nm for metal-free phthalocyanine,17 and lex < 30 nm for copper phthalocyanine1b), indicating that the porous structure of the TiOPc film allows for the penetration of the I3-/I- liquid electrolyte to form the TiOPc/I3- -I- bulk heterojunction. This junction formation prevents the incident light from being filtered out by the TiOPc film before it gets to the site of exciton dissociation (filter effect) and increases the contact area between the TiOPc and I3-/I- liquid electrolyte, resulting in the better photovoltaic performance of the ITO/ TiOPc(MD), I3--I-/ΙΤΟ cell compared with the ITO/TiOPc(VE), I3--I-/ΙΤΟ cell. The above model is schematically illustrated in Figure 6a. As described in section 1, the pigment films prepared by the MD method are characterized by their unique nanoporous structure.5,6 Indeed, scanning electron micrographs (SEM) and atomic force micrographs (AFM) of the TiOPc(MD) (images a and c in Figure 7) and TiOPc(VE) films (images b and d in Figure 7) revealed a much higher degree of porosity of the former film compared with the latter. Such a porous structure is expected to form a bulk heterojunction of TiOPc/I3--I- when immersed in the aqueous I3--I- electrolyte solution. When this

bulk heterojunction is illuminated by light, excitons (paired circles in Figure 6a) are photogenerated within a distance lex from the interface of TiOPc/I3--I- due to the interpenetrating network, resulting in an increase in the exciton diffusion efficiency defined as the fraction of the photogenerated excitons that reaches the TiOPc/I3--I- interface before recombination, which in turn induces a higher efficiency of the charge photogeneration. The photogenerated electrons (filled circles in Figure 6a) and holes (open circles) at the TiOPc/I3--I- interface are transported through the channels of the I3--I- solution and the TiOPc grains, respectively, because the injection of electrons from the electrolyte to the TiOPc and that of holes from the TiOPc to the electrolyte may be blocked by the presence of the Schottky-like barrier at the TiOPc/I3--I- interface (Figure 6b).1d If the above model is reasonable for the performance of the ITO/TiOPc(MD) photocathode, loss of light energy due to the filter effect should be hardly observed for input radiation directed through the ITO/TiOPc side even when the thickness of the TiOPc film is increased. Hence, the photovoltaic properties were investigated as a function of the TiOPc film thickness, d. Figure 8 shows the dependences of the Jsc (a), Voc (b), ff (c), and η values (d) on d. Voc and ff vary slightly with increasing d, but Jsc and η show a stronger dependence. The latter two parameters increase at first, but then taper off after d reaches ca. 200 nm. This behavior can be interpreted as follows. If the penetration depth of light (lp) is tentatively taken as the

Bulk Heterojunction Photoelectrochemical Cells distance at which the light intensity falls to 1/e of its initial value, lp is calculated to be ca. 180 nm for light around the absorption peaks (λ ) 650-800 nm, see Figure 5 and section 2.4) of the TiOPc film. This lp value agrees well with the d value at the saturation point of the curves a and d in Figure 8, indicating that Jsc and η are limited by the absorption of light in the thickness region below ca. 200 nm. The nearly constant values of η and Jsc above d ≈ 200 nm suggest that the absorption of effective light ranging from 650 to 800 nm is achieved in the TiOPc films and that the bulk-generated excitons effectively reach the photoactive TiOPc/I3--I- interface. This situation should occur in the bulk heterostructure and may be distinct from the case for the compact bilayer (or planar) junction18 in which the fraction of the photogenerated excitons which contribute to photocurrent generation would decrease with increasing film thickness due to the filter effect. A possible alternative explanation of the behaviors in Figures 5 and 8 is the molecular quenching mechanism proposed by Splan, Massari, and Hupp.19 They prepared an ITO/porous film of porphyrin square dye/I3--I-/Pt cell and measured its shortcircuit photoresponse, photocurrent action spectrum, and the IPCE values plotted as a function of light absorbed. These data, combined with the measurement of the dye excited-state lifetimes, led to the proposal of the quenching mechanism in which the porphyrin dye forms the complex with I3-, dye + I3- h [dye: I3-], and quenching occurs via the complex, [dye: I3-] + hν f dye+ + 3I-. The captured electron is transported by I- through the solution to the Pt electrode. On the other hand, the dye-localized hole reaches the dye/ITO interface by hopping and is compensated by the electron from the Pt electrode, resulting in the regeneration of the neutral form of the dye. This model explains the cathodic direction of the cell photocurrent in Figure 3 (solid curves). In the conventional dye-sensitized solar cells, the I3--I- redox couple is exclusively employed as a dye generator because of the unique ability of I3- to resist recombination with the electron injected into the conduction band of TiO2.20 Such a unique property of I3- is in conflict with the above exciton diffusion model in which the reaction of the photogenerated electron with I3- is assumed (see Figure 6b), whereas the conflict is resolved by assuming the complex formation between the TiOPc and I3-, [TiOPc: I3-], in the molecular quenching model. Additionally, the latter model also explains well the spectral response (Figure 5) and film-thickness dependence (Figure 6) and leads to the same conclusion as the exciton diffusion model that a bulk heterojunction is formed between the TiOPc porous film and the liquid I3--I- electrolyte. Implicit in the bulk heterojunction structure is the drawback of the ITO/pigment(MD), I3--I-/ΙΤΟ cells, i.e., only small photovoltages were produced (see Table 2). The potential of the bare ITO electrode is governed by the I3-/I- redox couple, while the filmed ITO electrode is controlled by the mixed potential for the I3-/I- and the pigment+/0 redox couples.19 This indicates that only small potential differences are generated when the I3-/I- redox reaction is dominant at the filmed ITO electrode. To increase the value of Voc, the contribution of the I3-/I- redox reaction at the filmed ITO electrode should be suppressed. For example, use of the electrode at which the I3-/I- couple shows poor reversibility and/or blocking of the direct contact between the electrode and I3-/I- couple by interposing a thin pigment compact layer may be considered. 4. Concluding Remarks TiOPc and CuPc films were prepared on the ITO-coated glass plate by the micellar disruption method, and their deposition

J. Phys. Chem. B, Vol. 110, No. 46, 2006 23327 behavior was investigated (section 3.1). The investigation revealed that the deposition is controlled by the diffusion of the pigment particles from the dispersion bulk to the ITO surface and that the packing density of the films is kept constant during their growth. Next, the films were employed as photocathodes in structurally simple photoelectrochemical cells, ITO/TiOPc, I3--I-/ΙΤΟ. The energy conversion efficiencies of the cells were ca. 2 orders of magnitude higher than those with the vacuumevaporated phthalocyanine layers under white-light irradiation of ca. 50 mW/cm2 (section 3.2). The measurement of the action spectrum, the film-thickness dependence on the energy conversion efficiency, and morphological investigations showed evidence for the formation of a bulk heterojunction between the nanoporous phthalocyanine layers and the I3--I- redox aqueous solution (section 3.3), the formation likely being the major cause of the relatively high energy conversion efficiency compared with that of the planar junction cell. In view of the fact that interpenetrating organic heterostructures have aroused great interest due to their high light-to-electricity conversion efficiency, our interpenetrating organic pigment/liquid electrolyte structures presented here may also be promising for use in thinfilm solar cells, electrochromic displays, and so forth. Studies to optimize the material conditions such as the kinds of pigment, electrolyte solution, and counter electrode are now under way, the details of which will be reported in a separate paper. Supporting Information Available: AFM and SEM images of the CuPc films prepared by the MD and VE methods. This material is available free of charge via the Internet at http:// pubs.acs.org. References and Notes (1) (a) Uchida, S.; Xue, J.; Rand, B. P.; Forrest, S. R. Appl. Phys. Lett. 2004, 84, 4218-4220. (b) Heutz, S.; Sullivan, P.; Sanderson, B. M.; Schultes, S. M.; Jones, T. S. Sol. Energy Mater. Sol. Cell 2004, 83, 229245. (c) Peumans, P.; Uchida, S.; Forrest, S. R. Nature 2003, 425, 158162. (d) Hiramoto, M.; Fujiwara, H.; Yokoyama, M. J. Appl. Phys. 1992, 72, 3781-3787. (e) Hiramoto, M.; Fujiwara, H.; Yokoyama, M. Appl. Phys. Lett. 1991, 58, 1062-1064. (2) Tang, C. W. Appl. Phys. Lett. 1986, 48, 183-185. (3) (a) Ilatovskii, V. A.; Dmitriev, I. B.; Komissarov, G. G. Russ. J. Phys. Chem. 1978, 52, 63-65. (b) Kampas, F. J.; Yamashita, K.; Fajer, J. Nature 1980, 284, 40-42. (c) Jaeger, C. D.; Fan, Fu-Ren. F.; Bard, A. J. J. Am. Chem. Soc. 1980, 102, 2592-2598. (d) Skotheim, T.; Lundstro¨m, I.; Delahoy, A. E.; Kampas, F. J.; Vanier, P. E. Appl. Phys. Lett. 1982, 40, 281-283. (e) Loutfy, R. O.; McIntire, L. F. Sol. Energy Mater. 1982, 6, 467-479. (f) Skotheim, T.; Petersson, L.-G.; Ingana¨s, O.; Lundstro¨m, I. J. Electrochem. Soc. 1982, 129, 1737-1741. (g) Chen, J.; Burrell, A. K.; Campbell, W. M.; Officer, D. L.; Too, C. O.; Wallace, G. G. Electrochim. Acta 2004, 49, 329-337. (h) Tsekouras, G.; Too, C. O.; Wallace, G. G. Electrochim. Acta 2005, 50, 3224-3230. (i) Murray, P. S.; Ralph, S. F.; Too, C. O.; Wallace, G. G. Electrochim. Acta 2006, 51, 2471-2476. (4) Adi, M.; Yohannes, T.; Solomon, T. Sol. Energy Mater. Sol. Cells 2004, 83, 301-310. (5) Saji, T. Chem. Lett. 1988, 693-696. (6) (a) Saji, T.; Hoshino, K.; Ishii, Y.; Goto, M. J. Am. Chem. Soc. 1991, 113, 450-456. (b) Hoshino, K.; Saji, T. J. Am. Chem. Soc. 1987, 109, 5881-5883 (7) Kurata, H. In Technologies on LCD color filter, components & chemicals; Watanabe, J., Ed.; CMC: Tokyo, Japan, 1998; pp 59-68. (8) Harima, Y.; Yamashita, K.; Saji, T. Appl. Phys. Lett. 1988, 52, 1542-1543. (9) (a) Ohaku, K.; Nakano, H.; Kawara, T.; Yokota, S.; Takenouchi, O.; Aizawa, M. Electrophotography 1986, 25, 258-263. (b) Enokida, T.; Kurata, R.; Seta, T.; Katsura, H. Electrophotography 1988, 27, 533-538. (c) Fujimaki, Y.; Tadokoro, H.; Oda, Y.; Yoshioka, H.; Momma, T.; Moriguchi, H.; Watanabe, K.; Kinoshita, A.; Hirose, N.; Itami, A.; Ikeuchi, S. J. Imaging Technol. 1991, 17, 202-206. (d) Yamaguchi, S.; Sasaki, Y. J. Phys. Chem. B 2000, 104, 9225-9229. (10) (a) Yokoyama, S.; Kurata, H.; Harima, Y.; Yamashita, K.; Hoshino, K.; Kokado, H. Chem. Lett.. 1990, 343-346. (b) Takeoka, Y.; Aoki, T.; Sanui, K.; Ogata, N.; Watanabe, M. J. Electroanal. Chem. 1997, 438, 153158. (c) Gallardo, B. S.; Gupta, V. K.; Eagerton, F. D.; Jong, L. I.; Craig,

23328 J. Phys. Chem. B, Vol. 110, No. 46, 2006 V. S.; Shah, R. R.; Abbott, N. L. Science 1999, 283, 57-60. (d) Aydogan, N.; Abbott, N. L. Langmuir 2001, 17, 5703-5706. (11) (a) Hoshino, K.; Kurasako, H.; Inayama, T.; Kokado, H. J. Electroanal. Chem. 1996, 406, 175-185. (b) Ishida, N.; Sibuya, T.; Kitamura, T.; Hoshino, K. Langmuir 2003, 19, 2458-2465. (12) (a) Bard, A. J.; Faulkner, L. R. In Electrochemical Methods; John Wiley & Sons: New York, 1980; Chapter 5. (b) Delahay, P. In New Instrumental Methods in Electrochemistry; Interscience: New York, 1966; Chapter 3. (13) Skotheim, T.; Lundstro¨m, I.; Delahoy, A. E.; Kampas, F. J.; Vanier, P. E. Appl. Phys. Lett. 1982, 40, 281-283. (14) Skotheim, T. A.; Inganas, O. J. Electrochem. Soc. 1985, 132, 21162120.

Hoshino et al. (15) Saito, T.; Sisk, W.; Kobayashi, T.; Suzuki, S.; Iwayanagi, T. J. Phys. Chem. 1993, 97, 8026-8031. (16) Kerp, H. R.; Donker, H.; Koehorst, R. B. M.; Schaafsma, T. J.; van Faassen, E. E. Chem. Phys. Lett. 1998, 298, 302-308. (17) Popovic, Z. D.; Zamin, J. Chem. Phys. Lett. 1981, 80, 135-138. (18) (a) Minami, N.; Watanabe, T.; Fujishima, A.; Honda, K. Ber. Bunsen-Ges. Phys. Chem. 1979, 83, 476-481. (b) Loutfy, R. O.; Sharp, J. H. J. Chem. Phys. 1979, 71, 1211-1217. (19) Splan, K. E.; Massari, A. M.; Hupp, J. T. J. Phys. Chem. B 2004, 108, 4111-4115. (20) Oskam, G.; Bergeron, B. V.; Meyer, G. J.; Searson, P. C. J. Phys. Chem. B 2001, 105, 6867-6873.