Photoinduced Electron Transfer in Thin Layers Composed of

Apr 18, 2008 - ... in an aqueous solution containing a sacrifice electron donor. The most efficient photocurrent generation was observed in the gold/C...
0 downloads 0 Views 219KB Size
Download PDF
5608

Langmuir 2008, 24, 5608-5614

Photoinduced Electron Transfer in Thin Layers Composed of Fullerene-Cyclic Peptide Conjugate and Pyrene Derivative Shigekatsu Fujii, Tomoyuki Morita, and Shunsaku Kimura* Department of Material Chemistry, Graduate School of Engineering, Kyoto UniVersity, Kyoto-Daigaku-Katsura, Nishikyo-ku, Kyoto 615-8510, Japan ReceiVed January 26, 2008. ReVised Manuscript ReceiVed February 20, 2008 A bilayer structure was constructed on gold by Langmuir–Blodgett deposition of a fullerene (C60)-cyclic peptidepoly(ethylene glycol) (PEG) conjugate and thereafter a pyrene derivative from the air/water interface. The cyclic peptide moiety acts as a scaffold to prevent the fullerenes from self-aggregation and accordingly makes the monolayer homogeneous and stable. In addition to this gold/C60-cyclic peptide-PEG/pyrene bilayer, a pyrene monolayer, a gold/C60-PEG conjugate/pyrene bilayer (lacking the peptide scaffold), and a gold/pyrene/C60-cyclic peptide-PEG bilayer (with the opposite order of layers) were also prepared, and their anodic photocurrent generation were studied in an aqueous solution containing a sacrifice electron donor. The most efficient photocurrent generation was observed in the gold/C60-cyclic peptide-PEG/pyrene bilayer. It is considered that the C60 unit acts not only as sensitizer but also as an electron acceptor facilitating the electron transfer from the excited pyrene unit to gold, and that the fullerene layer suppresses quenching of the excited pyrene unit by energy transfer to gold. Furthermore, the cyclic peptide scaffold helps the fullerenes disperse without aggregation in the membrane and seems to protect their redox properties or inhibit self-quenching of their excited state. It is thus concluded that a bilayer structure with desired orientation of functional units is important for efficient photoinduced electron transfer and that a cyclic peptide scaffold is useful to locate hydrophobic functional groups properly in a thin layer.

Introduction For a few decades, molecular devices have been of great interest for a wide range of applications.1–4 One of the particular goals is to mimic natural photosynthesis, in which light energy harvested by the antenna pigments is efficiently converted into electrochemical energy.5–7 The core process is photoinduced electron transfer in the reaction center from a photoexcited porphyrin dimer to an electron acceptor and simultaneous electron donation from a donor to the dimer (photoinduced charge separation). To mimic this process, donor–acceptor systems in which either the donor or the acceptor works as a sensitizer as well, and donor-sensitizer-acceptor systems have been extensively studied.8–12 Fullerenes discovered by Kroto and co-workers13 have made a great impact on the design of photoinduced electron transfer systems.14–16 Fullerene C60 has an excellent ability as an electron acceptor because of its low unoccupied molecular orbitals and small reorganization energy in an electron transfer * Phone:+81-75-383-2400.Fax: +81-75-383-2401. E-mail: shun@ scl.kyoto-u.ac.jp. (1) Lehn, J. M. Angew. Chem., Int. Ed. Engl. 1988, 27, 89–112. (2) Tans, S. J.; Verschueren, A. R. M.; Dekker, C. Nature 1998, 393, 49–52. (3) Joachim, C.; Gimzewski, J. K.; Aviram, A. Nature 2000, 408, 541–548. (4) Lindsey, J. S. New J. Chem. 1991, 15, 153–180. (5) Wasielewski, M. R. Chem. ReV. 1992, 92, 435–461. (6) Meyer, T. J. Acc. Chem. Res. 1989, 22, 163–170. (7) Brabec, C. J.; Sariciftci, N. S.; Hummelen, J. C. AdV. Funct. Mater. 2001, 11, 15–26. (8) Kuciauskas, D.; Liddell, P. A.; Hung, S. C.; Lin, S.; Stone, S.; Seely, G. R.; Moore, A. L.; Moore, T. A.; Gust, D. J. Phys. Chem. B 1997, 101, 429–440. (9) Carbonera, D.; Di Valentin, M.; Corvaja, C.; Agostini, G.; Giacometti, G.; Liddell, P. A.; Kuciauskas, D.; Moore, A. L.; Moore, T. A.; Gust, D. J. Am. Chem. Soc. 1998, 120, 4398–4405. (10) Di Valentin, M.; Bisol, A.; Agostini, G.; Moore, A. L.; Moore, T. A.; Gust, D.; Palacios, R. E.; Gould, S. L.; Carbonera, D. Mol. Phys. 2006, 104, 1595–1607. (11) Imahori, H.; Hagiwara, K.; Aoki, M.; Akiyama, T.; Taniguchi, S.; Okada, T.; Shirakawa, M.; Sakata, Y. J. Am. Chem. Soc. 1996, 118, 11771–11782. (12) Polese, A.; Mondini, S.; Bianco, A.; Toniolo, C.; Scorrano, G.; Guldi, D. M.; Maggini, M. J. Am. Chem. Soc. 1999, 121, 3446–3452. (13) Kroto, H. W.; Heath, J. R.; Obrien, S. C.; Curl, R. F.; Smalley, R. E. Nature 1985, 318, 162–163.

process.14,15,17–20 Moreover, C60 itself can act as a sensitizer upon irradiation of a light by which C60 is excited. C60 has often been combined with a sensitizing electron donor with a large absorbance for a more efficient collection of photons.21–23 With this combination, more photons are collected by the donor, and the electron emitted from the excited donor is stabilized by the C60 unit, leading to efficient charge separation. For a practical photovoltaic application, this donor-C60 combination needs to be immobilized on a metal substrate, which functions as a working electrode.14,24,25 For efficient photocurrent generation, it is essential to control the orientation of the donor-C60 system to the surface normal. There are two ways to do this. One is onestep immobilization of a donor-C60 dyad with a linker group at either end which can attach to a metal surface (monolayer strategy).26–28 The other is stepwise deposition of a donor layer and a C60 layer (bilayer strategy).29 In this study, we chose the bilayer strategy considering the flexibility in the system design. The order of layers and the number of layers can be easily changed, (14) Yu, G.; Gao, J.; Hummelen, J. C.; Wudl, F.; Heeger, A. J. Science 1995, 270, 1789–1791. (15) Williams, R. M.; Zwier, J. M.; Verhoeven, J. W. J. Am. Chem. Soc. 1995, 117, 4093–4099. (16) Imahori, H.; Sakata, Y. AdV. Mater. 1997, 9, 537–546. (17) Echegoyen, L.; Echegoyen, L. E. Acc. Chem. Res. 1998, 31, 593–601. (18) Imahori, H.; Sakata, Y. Eur. J. Org. Chem. 1999, 2445–2457. (19) Guldi, D. M. Chem. Soc. ReV. 2002, 31, 22–36. (20) Imahori, H.; Hagiwara, K.; Akiyama, T.; Aoki, M.; Taniguchi, S.; Okada, T.; Shirakawa, M.; Sakata, Y. Chem. Phys. Lett. 1996, 263, 545–550. (21) Gust, D.; Moore, T. A.; Moore, A. L. Acc. Chem. Res. 2001, 34, 40–48. (22) Guldi, D. M.; Prato, M. Acc. Chem. Res. 2000, 33, 695–703. (23) Thomas, K. G.; Biju, V.; Guldi, D. M.; Kamat, P. V.; George, M. V. J. Phys. Chem. B 1999, 103, 8864–8869. (24) Eckert, J. F.; Nicoud, J. F.; Nierengarten, J. F.; Liu, S. G.; Echegoyen, L.; Barigelletti, F.; Armaroli, N.; Ouali, L.; Krasnikov, V.; Hadziioannou, G. J. Am. Chem. Soc. 2000, 122, 7467–7479. (25) Imahori, H.; Norieda, H.; Yamada, H.; Nishimura, Y.; Yamazaki, I.; Sakata, Y.; Fukuzumi, S. J. Am. Chem. Soc. 2001, 123, 100–110. (26) Tkachenko, N. V.; Vuorimaa, E.; Kesti, T.; Alekseev, A. S.; Tauber, A. Y.; Hynninen, P. H.; Lemmetyinen, H. J. Phys. Chem. B 2000, 104, 6371– 6379. (27) Guldi, D. M.; Zilbermann, I.; Anderson, G. A.; Kordatos, K.; Prato, M.; Tafuro, R.; Valli, L. J. Mater. Chem. 2004, 14, 303–309.

10.1021/la800269w CCC: $40.75  2008 American Chemical Society Published on Web 04/18/2008

Photoinduced Electron Transfer in Thin Layers

Langmuir, Vol. 24, No. 10, 2008 5609

were analyzed by surface pressure–area (π-A) isotherms; the thin layers prepared on gold were characterized by absorption spectroscopy, fluorescence spectroscopy, ellipsometry, and cyclic voltammetry, and their photocurrent generation was investigated. The objectives of this work are to prepare various well-defined layer structures at a molecular level for clarification of the relationship of their function (photocurrent generation) with those structures and to demonstrate the availability of a cyclic peptide as a scaffold for functional units having difficulties to be organized into a molecularly regular thin layer because of aggregation tendency.

Experimental Section

Figure 1. Chemical structures of R-Pyr, C60-PEG, and cyclo8-C60PEG.

and further substitution of either layer is also facile. Langmuir– Blodgett (LB) technique is a suitable method to prepare such bilayer films by transferring a monolayer prepared at the air/ water interface onto a solid substrate.30–32 In this study, we report the photocurrent generation by bilayer systems composed of pyrene and fullerene C60. Both units are chemically modified to be amphiphilic so that monolayers can be prepared at the air/water interface. The pyrene derivative (R-Pyr; Figure 1) has a long hydrophobic alkyl chain at one end and a hydrophilic diethylene glycol chain at the other end. Two types of C60 derivatives are used: one has a hydrophilic poly(ethylene glycol) (PEG) chain (C60-PEG; Figure 1) and the other has the PEG chain and a cyclic octapeptide (cyclo8-C60PEG; Figure 1). It is well-known that C60 tends to form selfaggregation because of strong intermolecular π-π interaction. In the previous study, we demonstrated that a cyclic peptide scaffold prevents the hydrophobic functional groups including C60 from aggregation formation and helps them disperse homogeneously on the water surface.33,34 For efficient anodic photocurrent generation, the monolayer of the C60 derivative was first transferred on a gold substrate and then the monolayer of the pyrene derivative was transferred over the C60 monolayer (Au/cyclo8-C60-PEG or C60-PEG/R-Pyr). In an aqueous solution of a sacrifice electron donor (triethanolamine; TEOA), anodic photocurrent generation will be promoted by stepwise electron transfer reactions with the C60 unit as an electron acceptor for the excited pyrene unit. For control experiments, Au/R-Pyr monolayer and Au/R-Pyr/cyclo8-C60-PEG bilayer were also examined to assess the importance of the C60 presence and the layer arrangement. The monolayers at the air/water interface (28) Vuorinen, T.; Kaunisto, K.; Tkachenko, N. V.; Efimov, A.; Lemmetyinen, H. Langmuir 2005, 21, 5383–5390. (29) Conoci, S.; Guldi, D. M.; Nardis, S.; Paolesse, R.; Kordatos, K.; Prato, M.; Ricciardi, G.; Vicente, M. G. H.; Zilbermann, I.; Valli, L. Chem.sEur. J. 2004, 10, 6523–6530. (30) Diederich, F.; Jonas, U.; Gramlich, V.; Herrmann, A.; Ringsdorf, H.; Thilgen, C. HelV. Chim. Acta 1993, 76, 2445–2453. (31) Cardullo, F.; Diederich, F.; Echegoyen, L.; Habicher, T.; Jayaraman, N.; Leblanc, R. M.; Stoddart, J. F.; Wang, S. P. Langmuir 1998, 14, 1955–1959. (32) Maggini, M.; Karlsson, A.; Pasimeni, L.; Scorrano, G.; Prato, M.; Valli, L. Tetrahedron Lett. 1994, 35, 2985–2988. (33) Fujii, S.; Morita, T.; Umemura, J.; Kimura, S. Thin Solid Films 2006, 503, 224–229. (34) Fujii, S.; Morita, T.; Umemura, J.; Kimura, S. Bioconjugate Chem. 2007, 18, 1855–1859.

Materials. Cyclo8-C60-PEG and C60-PEG (Figure 1) were synthesized according to the literature.34 R-Pyr was synthesized as follows. O2N-Pyr-COCH3 (200 mg) synthesized according to the literature35 was dissolved in pyridine (10 mL), and the solution was stirred at 70 °C; a NaClO aqueous solution (3 mL) was added, and the solution was stirred for 3 min. The solution was cooled by ice, and pH was adjusted to around 2 with 1 N hydrochloric acid. The precipitated product (O2N-Pyr-COOH) was filtered (195 mg). O2NPyr-COOH (127 mg) and NH2(CH2CH2O)2CH2CH2COOt-Bu (100 mg) were dissolved in DMF and the solution was stirred at 0 °C. O-(7-Azabenzotriazol-1-yl)-1,1,3,3-tetramethyluronium hexafluorophosphate (HATU; 179 mg) and N,N-diisopropylethylamine (DIEA; 198 µl) were added to the solution, and the solution was stirred at 0 °C for a while and, thereafter, at room temperature for 12 h. The solvent was removed, and the residue was purified by a silica gel column with chloroform/methanol as eluant changing the mixture ratio successively from 100/0, 40/1, and 3/1, and further purified by another silica gel column with chloroform/diethyl ether (7/1). The purified product (O2N-Pyr-CONH(CH2CH2O)2CH2CH2COOt-Bu) was dissolved in dichloromethane and hydrogenated under H2 atmosphere in the presence of 10% Pd/C (50 mg). After reaction, the Pd/C was removed by filtration, and the filtrate was condensed to afford the product (H2N-Pyr-CONH(CH2CH2O)2CH2CH2COOt-Bu, 54 mg). Stearic acid (34 mg) and H2N-Pyr-CONH(CH2CH2O)2CH2CH2COOt-Bu (54 mg) was dissolved in dichloromethane; HATU (47 mg) and DIEA (55 µl) were added, and the solution was stirred at room temperature for 12 h. After the reaction, chloroform was added to the solution, washed successively with pure water, 4% NaHCO3 aqueous solution, pure water, and 4% KHSO4 aqueous solution. The organic phase was dried over anhydrous MgSO4; the solvent was removed, and the residue was purified by a silica gel column with chloroform/ methanol changing the mixture ratio from 100/1 to 25/1 to afford the product (C17-CONH-Pyr-CONH(CH2CH2O)2CH2CH2COOtBu, 42 mg). TLC: Rf (chloroform/methanol ) 10/1) ) 0.77, Rf (chloroform/methanol ) 20/1) ) 0.36. 1H NMR (400 MHz, CDCl3): δ (ppm) 0.81 (3H, t, CH3(CH2)13CH2CH2CH2CO), 1.20–1.35 (26H, m, CH3(CH2)13CH2CH2CH2CO), 1.27 (9H, s, COOC(CH3)3), 1.43 (2H, q, CH3(CH2)13CH2CH2CH2CO), 1.77 (2H,q,CH3(CH2)13CH2CH2CH2CO),2.28(2H,t,NHCH2CH2OCH2CH 2OCH 2CH 2), 2.54 (2H, t, CH 3(CH 2) 13CH 2CH 2CH 2CO), 3.50–3.57 (6H, m, NHCH2CH2OCH2CH2OCH2CH2), 3.68–3.72 (4H,m,NHCH2CH2OCH2CH2OCH2CH2),6.92(1H,brs,NHCH2CH2OCH2CH2OCH2CH2), 7.59, 7.75, 7.87, 8.14 (8H, m, Pyr-H), 8.87 (1H, brs, NH-Pyr). MS (FAB, matrix; nitrobenzylalcohol) m/z: 742.5 (calcd for C46H66N2O6 [(M + H)+] m/z 742.49). The t-Bu group of C17-CONH-Pyr-CONH(CH2CH2O)2CH2CH2COOt-Bu was removed by treatment with trifluoroacetic acid; the solvent was removed, and the residue was washed with diisopropyl ether to afford the final product (R-Pyr). Thin Layer Preparation. A USI FSD-110 film balance controller with a trough (100 × 189 mm2) was used to prepare the monolayers of R-Pyr, cyclo8-C60-PEG, and C60-PEG on the water surface at 25.0 ( 1.0 °C. A chloroform solution of each compound was prepared (35) Yoshida, K.; Kawamura, S.; Morita, T.; Kimura, S. J. Am. Chem. Soc. 2006, 128, 8034–8041.

5610 Langmuir, Vol. 24, No. 10, 2008 (0.5–1.0 mM) and spread by a microsyringe onto the water subphase. The subphase was ultrapure water of a resistivity 18 MΩ cm. Compression of the monolayer at a rate of 10 cm2/min was started 15 min after spreading the chloroform solution to allow complete evaporation of the solvent. The compressed monolayer was transferred by vertical dipping onto a gold substrate under keeping the surface pressure constant (the pressure fluctuation is ca. 0.1 mN/m) at a rate of 10 mm/min. Two types of gold substrates were prepared by the vacuum deposition method; one is on glass for ellipsometry and electrochemical study, and the other is on quartz for absorption and fluorescence spectroscopy. Chromium (300 Å) as an adhesive layer and thereafter gold (2000 Å) were deposited on a glass substrate, while chromium (10 Å) and gold (50 Å) were deposited similarly on a quartz substrate. The degree of vacuum during the deposition was in the order of 10-4 Torr, and the thickness of each metal layer was monitored by a quartz oscillator. Before the metal deposition, the glass and quartz substrates were thoroughly washed by H2SO4, water, and methanol. Spectroscopy. The absorption spectra of the compounds in chloroform were recorded at a concentration of 1.0 × 10-5 M on a Shimadzu UV-2450PC spectrometer. The absorption spectroscopy on thin layers prepared on gold/quartz was performed with an attachment for substrate samples. The molecular area of R-Pyr in the monolayer was determined by using an equation, molecular area (nm2) ) 2353/(NAA360 × 10-17), where 353, NA, and A353 are the molar extinction of the pyrene unit at 353 nm, the Avogadro’s number, and the absorbance of the monolayer at 360 nm, respectively. The absorption peak is shifted from 353 nm in solution to 360 nm on a substrate. 353 was obtained from various concentrations of R-Pyr solutions in chloroform to be 27 000. The fluorescence spectrum of R-Pyr in chloroform was recorded at 1.0 × 10-6 M by a Jasco FP-6600 fluorometer. The slit widths were 3 nm for excitation and emission. The spectra of the thin layers containing R-Pyr prepared on gold/quartz were recorded on the same machine with an incident angle of 10° in a substrate holder. The slit widths were 10 nm for excitation and emission. The excitation wavelength was 360 nm for both solution and substrate measurements. All of these spectroscopic measurements were performed at room temperature. Ellipsometry. Determination of the thickness of the thin layer on gold by ellipsometry was carried out by a MIZOJIRI DHA-OLX/S autoellipsometer at room temperature. A helium-neon laser of 632.8 nm was used as the incident light, and the incident angle was set at 65°. The thickness of the thin layer was calculated automatically by using an equipped program. In the calculation, the complex optical constant of the monolayer was initially set to be 1.50 + 0.00i. The layer thickness and the extinction coefficient were the free parameters. The measurements were performed on more than 5 different spots on the substrate and the thickness was obtained as the average. Electrochemical Experiments. Electrochemical experiments were performed by a BAS 604 Voltammetric analyzer with a threeelectrode system with the thin layer-modified gold substrate as the working electrode, Ag/AgCl in 3 M NaCl as the reference electrode, and a platinum wire as the auxiliary electrode. Ultrapure water of a resistivity 18 MΩ cm was used for preparation of solutions. All of the solutions were deaerated by passing N2 gas for 15 min prior to the measurements. All of the electric potentials reported here were measured with respect to the reference electrode. To examine the packing of the layers, cyclic voltammetry was performed in a 1 mM K4[Fe(CN)6] and 1 M KCl aqueous solution at a scan rate of 0.05 V/s. Photocurrent generation measurements were carried out in the same system. The supporting electrolyte was a 0.1 M Na2SO4 aqueous solution. The thin layer-modified electrode was photoirradiated with a JASCO Xe lamp (500 W) equipped with a monochromator. The concentration of a sacrifice electron donor, triethanolamine (TEOA), was 50 mM. The intensity of irradiating light at 353 nm was evaluated to be 1.2 × 1014 photons/s by the potassium ferrioxalate actinometry.36 The potential of the gold substrate was set to be 0 V versus the Ag/AgCl reference. The quantum yield for photocurrent generation was calculated by dividing the number of flowed electrons by the number of photons absorbed

Fujii et al.

Figure 2. π-A Isotherms of R-Pyr (dotted line), cyclo8-C60-PEG (solid line), and C60-PEG (dashed line) at the air/water interface.

by the layer per unit time. The latter value was calculated from the absorbance of the layer at 353 nm and the light intensity. All of the measurements were done at room temperature.

Results and Discussion π-A Isotherms. The R-Pyr, cyclo8-C60-PEG, and C60-PEG monolayers were prepared at the air/water interface by a Langmuir trough. Figure 2 shows the π-A isotherms of the respective compounds. The isotherms of cyclo8-C60-PEG and C60-PEG show a surface pressure from a large area compared with that of R-Pyr, because of the long-range interactions among the flexible PEG chains. After compression to a certain surface pressure, each monolayer was expanded to check whether hysteresis is comprised between the compression and expansion processes. Up to 20 mN/m for the R-Pyr and cyclo8-C60-PEG monolayers and up to 15 mN/m for the C60-PEG monolayer, each monolayer showed a reversible behavior, and there was no hysterisis observed (data not shown). This result indicates that each compound forms a stable monolayer and does not collapse up to the corresponding surface pressure. By extrapolation of the linear part in the R-Pyr monolayer to zero surface pressure, the mean molecular area was estimated to be approximately 0.70 nm2. This molecular area roughly agrees with the area of a pyrene ring, suggesting that the pyrene unit take horizontal orientation to the water surface. Generally, it is difficult to prepare a stable monolayer from pyrene derivatives at the air/water interface because of the bulky property,37,38 and there are only few reports on the successful preparation.39 It is suggested that intermolecular association among the long alkyl chains and appropriate amphiphilicity given by the hydrophilic ethylene glycol chain should help the compound form a stable monolayer. Since the other monolayers (cyclo8-C60-PEG and C60-PEG) did not exhibit a clear linear region in the isotherms, determinations of the mean molecular areas were not done. Thin Layer Formation on Substrates. An R-Pyr monolayer was transferred onto a substrate at a surface pressure of 20 (36) Hatchard, C. G.; Parker, C. A. Proc. R. Soc. London, Ser. A 1956, 235, 518–536. (37) Dutta, A. K.; Misra, T. N.; Pal, A. J. Langmuir 1996, 12, 459–465. (38) Bhattacharjee, D.; Mukherjee, K.; Misra, T. N. J. Phys. Chem. Solids 2000, 61, 751–756. (39) Zhai, C. X.; Wu, Y. Q.; Bu, W. F.; Li, W.; Wu, L. X. Colloids Surf., A 2005, 257–58, 325–327.

Photoinduced Electron Transfer in Thin Layers

Langmuir, Vol. 24, No. 10, 2008 5611

Figure 3. Absorption spectrum of the Au/R-Pyr monolayer (solid line) with spectrum of R-Pyr in chloroform (dotted; left panel), and spectra of the Au/cyclo8-C60-PEG monolayer (solid) and Au/C60-PEG monolayer (dashed) with spectrum of cyclo8-C60-PEG in chloroform (dotted; right panel). The solution spectra were recorded at 1.0 × 10-5 M, and they are scaled down for comparison with the substrate spectra.

mN/m, while the cyclo8-C60-PEG and C60-PEG monolayers were transferred at a surface pressure of 10 mN/m. The monolayers were transferred only at upstroke. The corresponding molecular areas from the isotherms are 0.53 nm2, 1.27 nm2, and 0.77 nm2 for the R-Pyr, cyclo8-C60-PEG, and C60-PEG monolayers, respectively. Two types of gold substrates were used; one was a thin gold (50 Å)/quartz substrate for absorption and fluorescence spectroscopy, and the other was a thick gold (2000 Å)/glass substrate for ellipsometry and electrochemical studies. Six different thin layers were prepared for spectroscopy: Au/R-Pyr monolayer, Au/cyclo8-C60-PEG monolayer, Au/C60-PEG monolayer, Au/ R-Pyr/ cyclo8-C60-PEG bilayer, Au/cyclo8-C60-PEG/R-Pyr bilayer, and Au/C60-PEG/R-Pyr bilayer. The notation of Au/ A/B bilayer represents that the component A was first transferred on gold, and then the component B was transferred on the A-covered surface. For fluorescence spectroscopy, R-Pyr monolayer and R-Pyr/cyclo8-C60-PEG bilayer were also prepared on quartz for reference, and they are referred to quartz/R-Pyr monolayer and quartz/R-Pyr/cyclo8-C60-PEG bilayer, respectively. On the other hand, for photocurrent generation, Au/R-Pyr monolayer, Au/R-Pyr/cyclo8-C60-PEG bilayer, Au/cyclo8-C60-PEG/R-Pyr bilayer, and Au/C60-PEG/ R-Pyr bilayer were prepared. Ellipsometry was performed for the above-mentioned samples except for the Au/R-Pyr/ cyclo8-C60-PEG bilayer, and cyclic voltammetry was carried out for the Au/cyclo8-C60-PEG/R-Pyr and Au/C60-PEG/RPyr bilayers. Spectroscopy. The absorption spectra of the Au/R-Pyr, Au/ cyclo8-C60-PEG, and Au/C60-PEG monolayers on gold are shown in Figure 3 along with the spectra of the corresponding compounds in chloroform. The R-Pyr monolayer showed characteristic absorption bands of a pyrene derivative at 290 and 360 nm (Figure 3 left). The pyrene absorption bands are a little broadened and red-shifted (ca. 7 nm) compared with those of the reference R-Pyr spectrum in chloroform solution, but there is no indication of strong interaction among the pyrene moieties in the ground state. Similarly, for the monolayers with the C60 unit, characteristic absorptions of a fullerene C60 were observed at approximately 270 and 320 nm (Figure 3 right). These results show that the monolayers at the air/water interface were successfully transferred on a substrate with intact electronic

Figure 4. Subtracted absorption spectra of the Au/R-Pyr/cyclo8-C60PEG bilayer (dotted), Au/cyclo8-C60-PEG/R-Pyr bilayer (solid), and Au/C60-PEG/R-Pyr (dashed) bilayers by the spectrum of the cyclo8C60-PEG or C60-PEG monolayer.

structures of the pyrene or the C60 unit. The absorbance ratio at 270 nm in the Au/cyclo8-C60-PEG and Au/C60-PEG monolayers is about 2:3, which roughly agrees with the ratio of the surface densities at transfer, 1/1.27:1/0.77 (nm-2), suggesting quantitative transfer of the monolayers. The absorption spectra of the bilayers were the sum of the absorptions of the R-Pyr monolayer and the cyclo8-C60-PEG or C60-PEG monolayer. Figure 4 shows the subtracted absorption spectra of the Au/R-Pyr/cyclo8-C60-PEG, Au/cyclo8-C60-PEG/R-Pyr, and Au/C60-PEG/R-Pyr bilayers by the absorption of the C60 layers. The subtracted spectra coincide with the absorption of the R-Pyr monolayer, and the similar absorption intensities indicate the transfer of the R-Pyr monolayer is quantitative. The molecular area of R-Pyr was calculated from the absorbance at 360 nm in the Au/cyclo8-C60-PEG/R-Pyr subtracted spectrum to be 0.83 nm2. This value agrees well with the mean molecular area of R-Pyr in the monolayer at the air/ water interface (0.70 nm2). Fluorescence spectroscopy was performed to study the interaction among the pyrene units in the excited state, quenching

5612 Langmuir, Vol. 24, No. 10, 2008

Fujii et al.

Figure 5. Fluorescence spectra of the quartz/R-Pyr monolayer (A), quartz/R-Pyr/cyclo8-C60-PEG bilayer (B), Au/R-Pyr monolayer (C), Au/RPyr/cyclo8-C60-PEG bilayer (D), and Au/cyclo8-C60-PEG /R-Pyr bilayer (E; left), and the spectrum of the Au/R-Pyr monolayer (solid line) with the spectrum of R-Pyr in chloroform (dotted line; right).

Figure 6. Cyclic voltammograms of bare gold (dotted line), the Au/ cyclo8-C60-PEG/R-Pyr (solid line) and Au/C60-PEG/R-Pyr (dashed line) bilayers in a 1 mM K4[Fe(CN)6] and 1 M KCl aqueous solution at a scan rate of 50 mV/s.

Figure 7. Photocurrents generated by the Au/R-Pyr monolayer and the Au/R-Pyr/cyclo8-C60-PEG, Au/cyclo8-C60-PEG/R-Pyr, and Au/C60PEG/R-Pyr bilayers, upon photoexcitation with a 353 nm light in a 0.1 M Na2SO4 aqueous solution. Switching of the irradiation light is shown at the bottom.

of excited pyrene by gold, and photoinduced electron transfer from pyrene to the C60 unit. Quenching of pyrene excited-state occurs by electron transfer and energy transfer, but energy transfer on gold should be dominant because of the strong absorption of the Au substrate in the pyrene emission region. The left panel of Figure 5 shows the fluorescence spectra of the quartz/R-Pyr monolayer (A), quartz/R-Pyr/cyclo8-C60-PEG bilayer (B), Au/ R-Pyr monolayer (C), Au/R-Pyr/cyclo8-C60-PEG bilayer (D), and Au/cyclo8-C60-PEG /R-Pyr bilayer (E). The broadness of spectra is partially because of the wider slit width for substrate measurements. The pyrene unit in the thin layer showed a slight excimer emission around 480–550 nm, but monomer emission was dominant (Figure 5 right), indicating that there is no strong interaction among the pyrene units in the layer in the excited state. The intensities of pyrene emission were 33.8, 7.4, 14.2, 4.8, and 6.8 in an arbitrary unit for quartz/R-Pyr monolayer (A), quartz/R-Pyr/cyclo8-C60-PEG bilayer (B), Au/R-Pyr monolayer (C), Au/R-Pyr/cyclo8-C60-PEG bilayer (D), and Au/cyclo8C60-PEG/R-Pyr bilayer (E), respectively. The fluorescence intensity can be expressed by using the rates of intrinsic decay of excited state (kd), photoinduced electron transfer from pyrene

to C60 (ket), and quenching with gold (kq), and a certain constant (c). For example, 1/33.8 ) ckd for quartz/R-Pyr monolayer and 1/7.4 ) c(kd + ket) for the quartz/R-Pyr/cyclo8-C60-PEG bilayer, yielding ket ) 3.57kd. Similarly, kq was estimated to be 1.38kd from 1/14.2 ) c(kd + kq). The emission intensity for the Au/ R-Pyr/cyclo8-C60-PEG bilayer, in which both electron transfer and quenching with gold occur, was calculated by 1/c(kd + ket + kq) to be 5.7, which is close to the experimental value (4.8). Similarly, the intensity for the Au/cyclo8-C60-PEG/R-Pyr bilayer, in which quenching with gold should be suppressed by separation of the pyrene unit from gold, was calculated to be 7.4 (same as that of the quartz/R-Pyr/cyclo8-C60-PEG bilayer), which again agrees with the experimental value (6.8). Thus, the observed emission intensities in the various layer systems are consistently explained by considering the photochemical processes of electron transfer from the excited pyrene to the C60 unit and quenching of the excited pyrene with gold, implying the efficient photoinduced electron transfer from the excited pyrene to the C60 unit and the suppressive role of the C60 layer against the quenching with gold due to the widening the spacing between the pyrene and gold.

Photoinduced Electron Transfer in Thin Layers

Figure 8. Oxidation and reduction potentials of pyrene and C60 in the ground state and in the excited state, and the oxidation potential of TEOA vs Ag/AgCl reference.

Ellipsometry and Cyclic Voltammetry. The thicknesses of the monolayers and bilayers were measured by ellipsometry. The thicknesses of the Au/R-Pyr, Au/cyclo8-C60-PEG, and Au/ C60-PEG monolayers were 21, 20, and 19 Å, respectively. These values are reasonable for the molecular structures. On the other hand, the thicknesses of the Au/cyclo8-C60-PEG/R-Pyr and Au/ C60-PEG/R-Pyr bilayers were 46–68 and 34–44 Å, roughly double the monolayer thickness, suggesting successful preparation of a bilayer structure. The large variation was observed in these bilayer data, suggesting that excessive transfer of the second layer may occur because of stronger interaction between the organic layers compared with that between the bare substrate surface and the organic layer. The packing of these bilayers were examined by cyclic voltammetry in a solution containing redoxactive ferrocyanide ions. Figure 6 shows the results with a reference cyclic voltammogram of a bare gold substrate. Redox peaks of ferricyanide/ferrocyanide were observed in the bare gold substrate. In contrast, no redox peak was observed in the bilayer-modified substrate. This result indicates that the bilayers are well-packed and do not allow the redox ions to approach to the gold surface. Photocurrent Generation. Anodic photocurrent generation by the thin layers was studied in water upon photoexcitation with a 353 nm light in the presence of TEOA at a potential of 0 V. The results of the Au/R-Pyr monolayer and the Au/RPyr/cyclo8-C60-PEG, Au/cyclo8-C60-PEG/R-Pyr, and Au/C60PEG/R-Pyr bilayers are shown in Figure 7. All of the layers generated an anodic photocurrent in response to photoirradiation. Figure 8 shows the oxidation and reduction potentials of pyrene and C60 units in the ground state and excited state and the oxidation potential of TEOA.25,40–43 All of the possible electron transfer processes from TEOA to gold according to the redox potentials are summarized in Scheme 1, which contains excitation of pyrene and C60 units and electron (anion radical)-transport and hole (cation radical)-transport processes. In the Au/R-Pyr monolayer, there are two possible processes for anodic photocurrent generation. One is electron transfer from the excited pyrene unit to gold and following electron donation from TEOA to the pyrene cation radical (Scheme 1, 1-1), and the other is electron transfer from TEOA to the excited pyrene unit and subsequent electron transfer from the pyrene anion radical to gold (1-2). It is difficult to determine which is a major process because the rates of electron transfer from TEOA to pyrene and deactivations of radical species are not known. However, since the pyrene unit is close to gold, quenching of the pyrene excited state by energy transfer to gold (40) Zhang, W.; Shi, Y. R.; Gan, L. B.; Huang, C. H.; Luo, H. X.; Wu, D. G.; Li, N. Q. J. Phys. Chem. B 1999, 103, 675–681.

Langmuir, Vol. 24, No. 10, 2008 5613

should suppress the photocurrent generation as suggested by fluorescence spectroscopy. On the other hand, the Au/R-Pyr/cyclo8-C60-PEG bilayer generated about twice larger photocurrent than the Au/R-Pyr monolayer did. This bilayer configuration has four processes for photocurrent generation (Scheme 1). The two processes initiated by pyrene excitation (2-1 and 2-2 which are essentially the same as 1-1 and 1-2, respectively) might be disturbed by the intervening C60 layer between the pyrene unit and TEOA. In this bilayer, the C60 unit can serve as a sensitizer because of its absorption at 353 nm. The absorbances of the R-Pyr and cyclo8-C60-PEG layers at 353 nm are roughly the same, approximately 0.01 (Figures 3). Therefore, the half of the absorbed photons were used for the pyrene units to excite, and the other half is for the C60 units. The excited C60 unit sensitizes electron transfer from TEOA to gold via C60 cation radical (2-3) or anion radical (2-4), but considering low reducing ability of the excited C60 or low stability of the cation radical, the process via C60 anion radical should be dominant. Since the C60 unit is separated from the gold surface by the pyrene layer, quenching of the excitation energy should be less effective compared with the case of the pyrene excitation. Thus, a stable C60 anion radical is considered to be formed with a high efficiency, which should contribute to the electron transfer to gold. The observed twofold enhancement is therefore explained by this more efficient process in the bilayer (2-4) over the processes in the R-Pyr monolayer where photoenergy is dissipated by quenching with gold (1-1 and 1-2). Inversion of the order of the layers showed further enhancement of photocurrent generation. The photocurrent generated by the Au/cyclo8-C60-PEG/R-Pyr bilayer is about five times larger than that by the Au/R-Pyr/cyclo8-C60-PEG bilayer. The quantum yield for this photocurrent generation was 2.2%. This bilayer system has up to eight processes for photocurrent generation (Scheme 1). The process via pyrene excitation without C60 contribution (3-1 and 3-2) should occur more efficiently than 1-1 and 1-2 in the R-Pyr monolayer or than 2-1 and 2-2 in the R-Pyr/cyclo8-C60-PEG bilayer because the pyrene unit is separated from gold and is not subjected much to quenching with gold as indicated by fluorescence spectroscopy. But the longer distance between the pyrene unit and the gold also should slow down the electron transfer between them. However, this bilayer structure allows the C60 unit to act as an effective electron acceptor which accelerates the electron transfer from pyrene to gold. Process 3-3 indicates an electron hopping process from the photoexcited pyrene to the C60 unit and successively to gold, and process 3-4 shows the photoinduced charge separation between pyrene and C60. These processes should occur favorably because both problems derived from quenching with gold and longer electron transfer distance are solved. Furthermore, the processes started by the C60 photoexcitation (3-5, 3-6, 3-7, and 3-8) can also contribute to the photocurrent generation especially via the C60 anion radical for its stability (3-6 and 3-8). However, these processes might be less efficient than those by the pyrene photoexcitation because of the quenching of the C60 unit by gold. To study the effect by the cyclic peptide scaffold, the Au/ C60-PEG/R-Pyr bilayer was also examined. Its photocurrent efficiency was about half of that of the Au/cyclo8-C60-PEG/ R-Pyr bilayer despite the 1.5 times higher density of C60 units. Our previous work showed that the cyclic peptide scaffold suppresses aggregation of the C60 units and helps them disperse molecularly in the layer. The isotherms prepared from different volumes of the solutions coincided with each other for cyclo8-C60-PEG, while the isotherms of C60-PEG

5614 Langmuir, Vol. 24, No. 10, 2008

Fujii et al.

Scheme 1. Available Photophysical Processes for Anodic Photocurrent Generation

were found to be volume-dependent, suggesting aggregation formation of the C60 units in the C60-PEG monolayer. One possible explanation for the decreased photocurrent is that aggregation of the C60 units may lower the reducing ability of the anion radical and accordingly slow down electron transfer from the C60 anion radical to gold, resulted in suppression of the photocurrent generation. Another possibility is that the aggregation may cause self-quenching of the C60 units and accordingly suppress electron transfer via the C60 excitation although these processes might be minor because of quenching with gold as mentioned above.

Conclusions Photocurrent generation by thin layers composed of the amphiphilic C60 derivatives and the pyrene derivative was investigated. Three types of compounds were synthesized, the C60-cyclic peptide-PEG conjugate, the C60-PEG conjugate lacking a cyclic peptide scaffold, and the pyrene derivative carrying diethylene glycol and a long alkyl chain. Monolayer formation at the air/water interface from each compound was analyzed by the π-A isotherm, and then the monolayer was successfully transferred on a solid substrate by the vertical dipping method. Bilayers were prepared by the successive transfer of the monolayers. Absorption spectroscopy and ellipsometry confirmed quantitative transfer of the monolayers with intact electronic structures of the C60 and the pyrene units, and cyclic voltammetry in a solution containing a redox species showed that the bilayers have a well-packed structure. (41) Li, J. Y.; Peng, M. L.; Zhang, L. P.; Wu, L. Z.; Wang, B. J.; Tung, C. H. J. Photochem. Photobiol., A 2002, 150, 101–108. (42) Webster, R. D.; Heath, G. A. Phys. Chem. Chem. Phys. 2001, 3, 2588– 2594. (43) Meerholz, K.; Heinze, J. J. Am. Chem. Soc. 1989, 111, 2325–2326.

Fluorescence spectroscopy indicated quenching of the pyrene excited state with gold, suggesting photoinduced electron transfer from the pyrene unit to the C60 unit. Anodic photocurrent generations by these thin layers were studied in an aqueous solution containing a sacrifice electron donor. The gold/C60-cyclic peptide-PEG/pyrene bilayer system generated the largest photocurrent, followed by the gold/C60-PEG/pyrene bilayer, gold/pyrene/C60-cyclic peptide-PEG bilayer, and pyrene monolayer. The results suggest the following three points. First, the C60 unit acts not only as a sensitizer but also as an electron acceptor from the excited pyrene to promote photocurrent generation. Second, the order of the pyrene and C60 layers is important. A gold/C60/pyrene structure is favorable for anodic photocurrent generation because the photoinduced electron transfer from pyrene to C60 has the same direction as that of the overall electron transfer, and also the intervening C60 layer can suppress quenching of the excited pyrene unit with gold. Third, the cyclic peptide scaffold suppresses aggregation of the C60 units. The aggregation should suppress photocurrent generation probably by the change of C60 redox properties or self-quenching. This study shows the importance of an ordered thin layer structure with respect to the different chromophores for efficient photoelectrochemical function and the availability of a cyclic peptide as a scaffold for fullerenes to be applied for functional materials. Acknowledgment. This work is partly supported by Grantin-Aids for Young Scientists B (16750098), for Exploratory Research (17655098), and for Scientific Research B (15350068), and 21st century COE program, COE for a United Approach to New Materials Science, from the Ministry of Education, Culture, Sports, Science, and Technology, Japan. LA800269W