ARTICLE pubs.acs.org/JPCA
Elongation of Lifetime of the Charge-Separated State of FerroceneNaphthalenediimide[60]Fullerene Triad via Stepwise Electron Transfer Mustafa Supur,† Mohamed E. El-Khouly,*,†,|| Jai Han Seok,‡ Kwang-Yol Kay,*,‡ and Shunichi Fukuzumi*,†,§ †
)
Department of Material and Life Science, Graduate School of Engineering, Osaka University, ALCA, Japan Science and Technology Agency (JST), Suita, Osaka 565-0871, Japan ‡ Department of Molecular Science and Technology, Ajou University, Suwon 443-749, Korea § Department of Bioinspired Science, Ewha Womans University, Seoul 120-750, Korea Department of Chemistry, Faculty of Science, Kafr ElSheikh University, Kafr ElSheikh, Egypt
bS Supporting Information ABSTRACT: Photoinduced electron-transfer processes of a newly synthesized rodlike covalently linked ferrocenenaphthalenediimide[60]fullerene (Fc NDIC60) triad in which Fc is an electron donor and NDI and C60 are electron acceptors with similar first one-electron reduction potentials have been studied in benzonitrile. In the examined FcNDIC60 triad, NDI with high molar absorptivity is considered to be the chromophore unit for photoexcitation. Although the free-energy calculations verify that photoinduced charge-separation processes via singlet- and triplet-excited states of NDI are feasible, transient absorption spectra observed upon femtosecond laser excitation of NDI at 390 nm revealed fast and efficient electron transfer from Fc to the singletexcited state of NDI (1NDI*) to produce Fc+NDI•C60. Interestingly, this initial charge-separated state is followed by a stepwise electron transfer yielding Fc+NDIC60•. As a result of this sequential electron-transfer process, the lifetime of the chargeseparated state (τCS) is elongated to 935 ps, while Fc+NDI• has a lifetime of only 11 ps.
’ INTRODUCTION Nature harbors numerous biological mechanisms inspiring the scientists, among which photosynthesis, as an inclusive paradigm for the conversion of solar energy into a chemical potential by an electron-transfer sequence that is started by an excited state and ultimately yields efficient charge separation,16 has a particular position because its mimicry concerns a wide spectrum of scientific disciplines spanning from optoelectronics to biology-related fields. An essential aspect of the natural process is the multistep electron transfer throughout a series of redox-active units resulting in an incremental distance between the initial and final electron-transfer components to stabilize the radical ion pair against the charge recombination, which would waste the excitation energy as heat. Diverse structural motifs, i.e., triads, tetrads, etc. have been developed in order to mimic this multistep electron-transfer process over the components with different redox potentials.7,8 Within this context, photoinduced electron-transfer processes of a newly synthesized triad (Figure 1), ferrocenenaphthalenediimide[60]fullerene (FcNDIC60), have been examined in which NDI and C60 entities act as electron acceptors. NDIs are compact electron-deficient aromatic compounds with high stability.9 Among electron acceptors and visible light harvesting molecules, NDIs have attracted much attention due to their large conjugate planes and special electronic properties, which have been utilized in artificial photosynthetic systems for solar energy r 2011 American Chemical Society
conversion1012 or as an electron reservoir in the design of catenane and rotaxane supramolecular switches.13 Therefore, they show great tendency to form n-type over p-type semiconductor materials with high electron mobility.1416 Besides NDIs, [60]fullerene has also been frequently used as an electron acceptor in artificial photosynthetic systems1719 because of the three-dimensional structure, comparable reduction potential to NDIs, strong absorption in the UVvisible region, and more importantly, small reorganization energy in electron-transfer reactions.2022 There have been many reports on electron donor acceptor arrays containing NDI or C60 as an electron acceptor unit. However, there has been only one report on multicomponent electron donoracceptor arrays containing both NDI and C60 in which integration of silicon phthalocyanine (SiPc) with NDI and C60 through the axial position forms the SiPc (NDI)2(C60)2 pentad.23 In this case, the photoexcitation of the pentad resulted in the formation of the triplet excited state (3SiPc*) because of the low lying triplet energy of SiPc (1.26 eV) as compared with the charge-separated state, SiPc•+ (NDI)2•(C60) or SiPc•+(NDI)2(C60)2•. Thus, it is required to use a stronger electron donor than SiPc to obtain the charge-separated state as the lowest energy state rather than Received: October 8, 2011 Revised: November 3, 2011 Published: November 23, 2011 14430
dx.doi.org/10.1021/jp209668w | J. Phys. Chem. A 2011, 115, 14430–14437
The Journal of Physical Chemistry A
ARTICLE
Figure 1. Molecular structures of the investigated compounds.
the triplet-excited state in electron donoracceptor arrays that contain both NDI and C60. In the newly prepared FcNDIC60 triad 1, ferrocene has been selected as the electron-donating moiety with a low oxidation potential, in order to adjust the energy level of the charge-separated state below the singlet and triplet states of both NDI and C60 in polar benzonitrile (PhCN). FcNDI and NDIC60 dyads have also been investigated electrochemically, computationally, and spectroscopically for the elucidation of photoinduced events of FcNDIC60 triad (Figure 1).
’ EXPERIMENTAL SECTION Materials and Instruments. Reagents and solvents were purchased as reagent grade used without further purification. All reactions were performed using dry glassware under nitrogen atmosphere. Analytical TLC was carried out on a Merck 60 F254 silica gel plate, and column chromatography was done on a Merck 60 silica gel (230400 mesh). NMR spectra were recorded on a Varian Mercury-400 (400 MHz) spectrometer with a TMS peak as the reference. MALDI-TOF-MS spectra were recorded with a Kratos Compact MALDI I (Shimadzu). Elemental analyses were performed with a Perkin-Elmer 2400 Analyzer. IR spectra were recorded by using a ThermoNicolet NEXUS 670 FTIR spectrometer in the range of v = 500 4000 cm1 at room temperature. Samples were prepared as KBr pellets at different concentrations. Steady-state absorption measurements were recorded on a Shimadzu UV-3100PC spectrometer or a Hewlett-Packard 8453 diode array spectrophotometer at room temperature. Electrochemical measurements were performed on an ALS630B electrochemical analyzer in deaerated PhCN containing tetra-n-butylammonium hexafluorophosphate (TBAPF6, 0.10 M) as the supporting electrolyte at 298 K. A conventional three-electrode cell was used with a platinum working electrode (surface area of 0.3 mm2) and a platinum wire as the counter electrode. The Pt working electrode was routinely polished with BAS polishing alumina suspension and rinsed with acetone before use. The measured potentials were recorded with respect to an SCE reference electrode. All electrochemical measurements were carried out under an atmospheric pressure of argon.
Density functional theory (DFT) calculations were performed on a COMPAQ DS20E computer. Geometry optimizations were carried out using the Becke3LYP functional and 6-311G basis set,24 with the unrestricted HartreeFock (UHF) formalism as implemented in the Gaussian03 program, revision C.02. Graphical outputs of the computational results were generated with the Gauss View software program (ver. 3.09) developed by Semichem, Inc. Femtosecond transient absorption spectroscopy experiments were conducted using an ultrafast source, Integra-C (Quantronix Corp.), an optical parametric amplifier, TOPAS (Light Conversion Ltd.), and a commercially available optical detection system, Helios, provided by Ultrafast Systems LLC. The source for the pump and probe pulses were derived from the fundamental output of Integra-C (780 nm, 2 mJ/pulse and fwhm = 130 fs) at a repetition rate of 1 kHz. Seventy-five percent of the fundamental output of the laser was introduced into TOPAS, which has optical frequency mixers resulting in a tunable range from 285 to 1660 nm, while the rest of the output was used for white light generation. Typically, 2500 excitation pulses were averaged for 5 s to obtain the transient spectrum at a set delay time. Kinetic traces at appropriate wavelengths were assembled from the timeresolved spectral data. All measurements were conducted at 298 K. The transient spectra were recorded using fresh solutions in each laser excitation. Solutions were deoxygenated by Ar purging for 15 min prior to the measurements. Preparation of N-(4-Ferrocenylphenyl)-naphthalene-1,8dicarboxyanhydride-4,5-dicarboximide (6). To anhydrous DMF (80 mL) was added naphthalene dianhydride (5, 3.0 g, 11.2 mmol) and refluxed for 2 h until the solution color was changed to black (solution A). To solution A was very slowly added a solution of 4-ferrocenylaniline25 (0.70 g, 2.53 mmol) in toluene (150 mL) by dropping over 3 days. The mixture was refluxed for 24 h, and then the solution was cooled to room temperature. The solvent was removed under reduced pressure, and the residue was dissolved in dichloromethane (200 mL). After the trituration of the solution, the unsolved impurities were removed by filtering. The solvent was removed under reduced pressure and the residue was purified by flash column chromatography over silica gel with dichloromethane to give compound 6 (1.25 g, 94.0%) in a black solid. TLC (dichloromethane): Rf = 0.50. 1 H NMR (CDCl3) δ: 8.86 (s, 4H), 7.65 (d, J = 8.8 Hz, 2H), 7.22 14431
dx.doi.org/10.1021/jp209668w |J. Phys. Chem. A 2011, 115, 14430–14437
The Journal of Physical Chemistry A
ARTICLE
Scheme 1. Synthesis of Compounds 3 and 4a
a (a) 4-Ferrocenylaniline, DMF, toluene, reflux, 24 h, 94.0%. (b) N-octylamine, pyridine, reflux, 72 h, 53.6%. (c) 4-(Noctyl-3,4-fullero-pyrrolidin-2yl)phenylamine, Zn(OAc)2, toluene, reflux, 48 h, 10.0%.
(d, J = 8.8 Hz, 2H), 4.64 (d, J = 9.0 Hz, 2H), 4.38 (d, J = 9.0 Hz, 2H), 4.11 (s, 5H). Anal. Calcd for C30H17NO5Fe: C, 68.33%; H, 3.25%; N, 2.66%. Found: C, 68.09%; H, 3.51%; N, 2.48%. Preparation of N,N0 -(4-Ferrocenylphenyl)-(octyl)-naphthalene1,8:4,5-bis(dicarboximide) (3). To a solution of compound 6 (0.20 g, 0.38 mmol) in pyridine (50 mL) was added N-octylamine (0.10 g, 0.77 mmol), and the mixture was refluxed for 3 days. After cooling to room temperature, the solvent was removed under reduced pressure, and the residue was purified by column chromatography over silica gel with dichloromethane/acetone (150:1 v/v) to give compound 3 (0.13 g, 53.6%) as a black solid. TLC [dichloromethane/hexane 3:1 (v/v)]: Rf = 0.30. 1H NMR (CDCl3) δ: 8.80 (s, 4H), 7.66 (d, J = 8.8 Hz, 2H), 7.24 (d, J = 8.8 Hz, 2H), 4.68 (d, J = 9.0 Hz, 2H), 4.38 (d, J = 9.0 Hz, 2H), 4.12 (m, 2H), 4.06 (s, 5H), 1.801.20 (m, 12H), 0.90 (m, 3H). IR (KBr) 2957, 2921, 2851, 2871, 1709, 1666, 1581, 1462, 1453, 1377, 1347, 1245, 1195, 1106, 1081, 1031, 977, 892, 838, 817, 770, 730, 713 cm1. Mass (MALDITOF): m/z for C38H34N2O4Fe calcd, 638.53; found, 638.35. Anal. Calcd for C38H34N2O4Fe: C, 71.48%; H, 5.37%; N, 4.39%. Found: C, 71.27%; H, 5.56%; N, 4.28%. Preparation of N,N0 -(4-Ferrocenylphenyl)-4-(Noctyl-3,4fullero-pyrrolidin-2-yl)-naphthalene-1,8:4,5-bis(dicarboximide) (4). Compound 6 (0.030 g, 0.057 mmol), 4-(Noctyl-3,4fullero-pyrrolidin-2-yl)phenylamine23 (0.07 g, 0.072 mmol), and zinc acetate (0.10 g, 0.55 mmol) were added to toluene (30 mL), and the mixture was refluxed for 2 days. After cooling to room temperature, the solvent was removed under reduced pressure. The residue was purified by column chromatography over silica gel with toluene, then with dichloromethane/hexane (3:1 v/v), and then with dichloromethane to give compound 4 (8.4 mg, 10.0%) in a black solid. TLC [dichloromethane/hexane (3:1 v/v)]: Rf = 0.30. 1H NMR (CDCl3) δ: 8.81 (s, 4H), 8.01 (br, s, 2H), 7.68 (d, J = 8.8 Hz, 2H), 7.43 (d, J = 8.8 Hz, 2H), 7.12 (d, J = 8.4 Hz, 2H), 5.20 (s, 1H), 5.18 (d, J = 10.2 Hz, 1H), 4.68 (s, 2H), 4.38 (s, 2H), 4.17 (d, J = 10.2 Hz, 1H), 4.08 (s, 5H), 3.40 (m, 2H), 1.801.20 (m, 12H), 0.90 (m, 3H). 13C NMR (CDCl3) δ: 163.00, 154.50, 153.00, 148.58, 147.84, 147.49, 147.01, 146.27, 146.15, 146.09, 145.75, 145.55, 145.12, 144.73, 144.41, 144.21, 144.01, 143.05, 142.59, 142.08, 141.58, 131.99, 131.27, 130.82, 129.85, 128.75, 128.37, 127.61, 78.10, 69.10, 68.20, 53.00, 32.40, 29.50, 29.00, 28.20, 27.10, 23.40, 14.00. IR
(KBr) 2958, 2920, 2850, 1711, 1677, 1580, 1462, 1347, 1261, 1199, 1168, 1093, 1058, 1020, 982, 948, 906, 872, 802, 766, 729, 688, 657 cm1. Mass (MALDITOF): m/z for C106H41N3O4Fe calcd, 1476.35; found, 1477.15. Anal. Calcd for C106H41N3O4Fe: C, 86.24%; H, 2.80%; N, 2.85%. Found: C, 86.59%; H, 3.01%; N, 2.58%.
’ RESULTS AND DISCUSSION Synthesis and Characterization. Preparation of compounds 3 and 4 was performed by following the steps depicted in Scheme 1. Each step of the reaction sequence proceeded smoothly to give a good or moderate yield of product as described in the Experimental Section in detail. A subsequent reaction of 4-ferrocenylaniline25 with commercially available naphthalene dianhydride (5) was performed to yield the corresponding naphthalene monoimide 6 in 94%. The reaction between 6 and N-octylamine was performed in pyridine to give compound 3 in 54% yield. In the following step, naphthalene monoimide 6 was treated with 4-(Noctyl-3,4-fullero-pyrrolidin-2-yl)phenylamine, whose synthesis is described elsewhere,23 in the presence of Zn(OAc)2 to give compound 4 in 10% yield. Reference compounds 1 and 2 were synthesized in a similar fashion as described previously.23 Compound 4 and precursor compounds are soluble in common organic solvents. 1H NMR and/or 13C NMR analyses mainly confirmed the structure and the purity of compound 4 and the precursor compounds (see Supporting Information, Figures S1S4). The matrix-assisted laser desorption ionization time-of-flight (MALDI-TOF) mass spectra provided a direct evidence for the structures of 3 and 4 (Supporting Information, Figures S5 and S6). Further confirmation of corresponding compounds was obtained from the steady-state UVvis measurements as shown in the forthcoming sections. Steady-State Spectral Studies. As shown in Figure 2, reference NDI and Fc-NDI dyad have identical steady-state absorption spectra with the vibronic transitions of NDI core at 344, 362, and 382 nm, indicating no electronic interaction between ferrocene and NDI in benzonitrile (PhCN). The NDI absorption patterns of FcNDIC60 triad at 362 and 382 nm do not change significantly compared to those of individual NDI and Fc-NDI, also suggesting a weak coupling between NDI and C60 in the ground state. The broad absorption at around 430 nm 14432
dx.doi.org/10.1021/jp209668w |J. Phys. Chem. A 2011, 115, 14430–14437
The Journal of Physical Chemistry A is the contribution from the C60 moiety. Hence, the selective photoexcitation of the NDI entity of the FcNDIC60 triad from NDI around the absorption peak at 382 nm can be achieved in order to monitor the photochemical events by using timeresolved transient spectroscopy. Upon excitation at 380 nm excitation light, NDI fluoresces at 389 nm corresponding to the singlet-excited state energy of 3.22 eV.26 After a very fast intersystem crossing, NDI gives phosphorescence with a triplet-excited state energy of 2.05 eV.27 Steady-state fluorescence measurements displayed remarkable fluorescence quenching of the Fc1NDI*C60 and Fc1NDI* relative to that of the NDI reference 1 (Supporting Information, Figure S7). The quenching process may involve the energy transfer from 1NDI* (3.22 eV) to 1C60* (1.75 eV)19e and/or the electron transfer from ferrocene to the electron acceptor NDI. Computational Studies. To gain insight into the molecular and electronic structure of compounds 3 and 4, computational studies were performed by using density functional theory (DFT) calculation methods at the B3LYP/6-311G level. The structures were optimized to a stationary point on the Born Oppenheimer potential surface. As shown in Figure 3 (also see Supporting Information, Figure S8), the center-to-center distances between ferrocene and the NDI core (RFc‑NDI) and between ferrocene and the C60 moiety (RFc‑C60) were calculated to be 11.4 and 23.6 Å, respectively. Ferrocene with the adjacent benzene ring has an orientation almost perpendicular to the
Figure 2. Steady-state absorption spectra of compounds 1, 3, and 4 in PhCN.
ARTICLE
NDI entity. In both triad 4 and dyad 3, the orbital distribution of HOMO was located on the ferrocene entities with a small orbital coefficient on the adjacent benzene ring due to the aligned orientation. However, the orbital distribution of LUMO, was realized on C60 and LUMO+1 on NDI body entities for triad 4. This suggests that upon excitation of the NDI unit, possible charge separation is first to take place between electron-donating ferrocene and the NDI unit as an electron acceptor. This may also suggest a sequential electron transfer from NDI to C60. In addition, the RFcC60 value, which is two times longer than RFcNDI, may have a significant effect on charge-separated state lifetime in this case. The absence of LUMOs on the ferrocene unit and HOMO on both NDI and C60 moieties proposes weak or no charge transfer interactions between the electron donor and acceptors in the ground state in triad 4 and dyad 3. Electrochemical Studies and Electron-Transfer Driving Force. Cyclic voltammetry (CV) and differential pulse voltammetry (DPV) measurements have been performed in order to clarify the electrochemical properties of 4 together with reference compounds in deaerated PhCN containing TBAPF6, (0.10 M) as a supporting electrolyte. The first one-electron oxidation potential
Figure 4. Cyclic voltammograms of the FcNDIC60 triad with reference compounds in deaerated PhCN containing 0.10 M TBAPF6 (sweep rate: 0.1 mV/s).
Figure 3. Frontier molecular orbitals of FcNDIC60 triad obtained by using the ab initio B3LYP/6-311G method. 14433
dx.doi.org/10.1021/jp209668w |J. Phys. Chem. A 2011, 115, 14430–14437
The Journal of Physical Chemistry A
ARTICLE
Figure 5. (left) Differential absorption spectra obtained upon femtosecond flash photolysis (λex = 390 nm) of 3 in deaerated PhCN at the indicated time intervals. (right) Decay profile of the NDI radical anion at 480 nm, monitoring Fc to NDI charge-separation dynamics.
of the reference ferrocene was obtained at 470 mV vs SCE (Figure 4).28 FcNDI dyad undergoes the one-electron oxidation at 500 mV vs SCE leading to the formation of a ferrocenium ion (Fc+) with an anodic shift of 30 mV relative to the reference ferrocene. On reductive scans, the NDI unit displays the oneelectron reduction peak at 610 mV vs SCE.29 When turning to the FcNDIC60 triad, we can observe the similar oxidation properties originating from the ferrocene moiety, which exhibits the one-electron oxidation at 500 mV vs SCE. The first reduction of FcNDIC60 takes place at 600 mV vs SCE, showing only a 10 mV cathodic shift relative to FcNDI and NDIC60 dyads. However, the first reduction of triad 4 appears to be a twoelectron process when its wave area is compared to that of the first one-electron oxidation corresponding to Fc/Fc+. This results from the overlapped first one-electron reduction potentials of individual NDI and C60. First one-electron reduction peaks could not be separated even by DPV measurements at different scan rates (Supporting Information, Figure S9). In this case, the orientation of the electron acceptors with regard to ferrocene and the photosensitizer unit chosen selectively may determine the behavior of the transferred electron. The free-energy change of photoinduced electron transfer (ΔGCS) in eV was calculated according to the equation below:30 ΔGCS ¼ Eox Ered ΔE00 þ ΔGs
ð1Þ
where Eox is the first one-electron oxidation potential, Ered is the first one-electron reduction potential, ΔE00 is the energy of the 00 transition energy gap between the lowest excited state and the ground state of the photosensitizer moiety, and ΔGs refers to the static Coulomb energy in PhCN.31 On this basis, upon photoexcitation of NDI, 1NDI* provides an exothermic driving force for photoinduced electron transfer (ΔGCS) in triad 4 and dyad 3 to be 2.17 eV. In all cases, the exothermic free-energy change for charge recombination (ΔGCR) is computed to be 1.05 eV in PhCN. Hence, the largely positive values of electrontransfer driving forces from various excited states of photosensitizer moieties of triad 4 and dyad 3 indicate the potential of photoinduced electron-transfer processes in these systems. Time-Resolved Transient Absorption Spectral Studies. In order to elucidate the excited-state photochemical events of triad
4 and dyad 3, femtosecond transient spectroscopy has been utilized. In the femtosecond transient absorption spectra of the FcNDI dyad in deaerated PhCN recorded after an excitation at 390 nm (Figure 5), very fast quenching of the singlet state of NDI followed by the fast formation of transient traits (kCS = 6.0 1012 s1) with the maxima at 480 and 610 nm has been monitored. This new transient absorption is assigned to the radical anion of NDI, by comparison to the steady-state absorption spectra29 of NDI• obtained chemically (Supporting Information, Figure S10),32 declaring the photoinduced electron transfer from ferrocene to the excited states of NDI although the formation of Fc+ was not observable due to its low molar extinction coefficient (ε), which is covered by the radical anion of NDI at around 615 nm.33 The decay rate of NDI• at 480 nm is fitted to a first-order plot where the charge-recombination rate constant (kCR) is found to be 9.1 1010 s1, giving a lifetime for the charge-separated state (τCS) of 11 ps in polar PhCN (Figure 5). Despite the fact that possible electron transfer via the triplet-excited state of NDI is energetically feasible, exceedingly fast formation of charge separation is most likely to hinder the intersystem crossing, which is reported to take place within a few tens of picoseconds.34 Another evidence for charge separation via 1NDI* may be the relatively short τCS implying the radical-ion pair is a singlet character (1RIP) rather than a triplet one (3RIP). The fast charge separation (kCS = 6.0 1012 s1) with a large driving force (ΔGCS = 2.17 eV) would afford an unreasonably large reorganization of photoinduced electron transfer (λ ≈ 2 eV) according to the Marcus theory of electron transfer.35 It has often been suggested that photoinduced electron-transfer reactions being so exergonic result in the formation of an electronically excited species (at lower driving force), which is faster than the (highly inverted) reaction directly to ground-state.3638 In the present case as well, rapid electron transfer from Fc to 1NDI* may be made possible by producing the excited state of either Fc+ or NDI•, which has a much smaller driving force as compared with that producing the ground states. The produced excited state may rapidly be converted to the ground state. Femtosecond differential absorption spectra of the Fc NDIC60 triad, obtained by using 390 nm laser pulses that mainly excite the NDI entity, revealed the formation of 14434
dx.doi.org/10.1021/jp209668w |J. Phys. Chem. A 2011, 115, 14430–14437
The Journal of Physical Chemistry A
ARTICLE
state Fc+NDIC60• (τCS2) as 935 ps (kCR2 = 1.1 109 s1). By this way, the lifetime of Fc+NDIC60• is 85-fold elongated with respect to that of Fc+NDI• indicating the effect of C60 with its spherical shape and low reorganization energy on stabilizing the final charge-separated state (Fc+NDIC60•). Seemingly, the Fc+NDI•C60 and Fc+NDIC60• states have the same energy levels. The small reorganization energy of [60]fullerene and the long distance between Fc+ and C60• may be the key parameter to explain this sequentially. The intramolecular photoinduced events of FcNDI and FcNDIC60 are compiled in the qualitative energy level diagram (Figure 7 and Supporting Information, Figure S11). Excitation of the NDI chromophore in dyad and triad affords very fast formation of charge-separated states (
