15074
J. Phys. Chem. B 2009, 113, 15074–15082
Structural Dependence on Excitation Energy Migration Processes in Artificial Light Harvesting Cyclic Zinc(II) Porphyrin Arrays Min-Chul Yoon,† Sung Cho,† Pyosang Kim,† Takaaki Hori,‡ Naoki Aratani,‡ Atsuhiro Osuka,*,‡ and Dongho Kim*,† Spectroscopy Laboratory for Functional π-Electronic Systems and Department of Chemistry, Yonsei UniVersity, Seoul 120-749, Korea, and Department of Chemistry, Graduate School of Science, Kyoto UniVersity, Sakyo-ku, Kyoto, 606-8502, Japan ReceiVed: May 20, 2009; ReVised Manuscript ReceiVed: August 30, 2009
A series of covalently linked cyclic porphyrin arrays CNZ that consist of N/2 of meso-meso directly linked zinc(II) porphyrin dimer subunits Z2 bridged by 1,3-phenylene spacers have been prepared by AgI-promoted oxidative coupling reaction. We have investigated the excitation energy migration processes of CNZ in toluene by using femtosecond transient absorption anisotropy decay measurements by taking 2Z2 composed of two Z2 units linked by 1,3-phenylene as a reference molecule. On the basis of the excitation energy transfer rate determined for 2Z2, we have revealed the excitation energy hopping rates in the cyclic arrays CNZ by using a regular polygon model. The number of excitation energy hopping sites Nflat calculated by using a regular polygon model is close to the observed Nexpt value obtained from the transient absorption anisotropy decays for C12Z-C18Z with circular and well-ordered structures. On the other hand, a large discrepancy between Nflat and Nexpt was found for smaller or larger arrays (C10Z, C24Z, and C32Z). In the case of C10Z, m-phenylene linked 2Z2 motif with the interchromophoric angle of 120° is not well suited to make a cyclic pentagonal array C10Z based on planar pentagonal structure. This geometrical factor inevitably causes a structural distortion in C10Z, leading to a discrepancy between Nexpt and Nflat values. On the contrary, the presence of highly distorted conformers such as figure-eight structures reduces the number of effective hopping sites Nexpt in large cyclic arrays C24Z and C32Z. Thus, our study demonstrates that not only the large number of porphyrin chromophores in the cyclic arrays CNZ but the overall rigidity and three-dimensional orientation in molecular architectures is a key factor to be considered in the preparation of artificial light harvesting porphyrin arrays. I. Introduction Recently, a variety of artificial molecular assemblies as biomimic models of natural photosynthetic systems have been synthesized.1 It is well-known that excitation energy hopping (EEH) processes in light-harvesting complexes mainly depend on the strength of exciton coupling governed by interconnection length and relative configuration between adjacent pigments.2 Determination of the factors to influence the energy migration efficiency and respective rates in naturally occurring lightharvesting array systems is an important aim in photosynthesis research, although it turns out to be difficult due to the large size and complexity of the molecular assembly. Nevertheless, understanding the relationship between structural and electronic properties of natural and artificial molecular assemblies in EEH processes is indispensable for molecular design toward the realization of molecular photonic devices, such as photovoltaic solar-energy conversion, charge storage, and optical sensing.3 In this regard, numerous research activities on the construction of various multidimensional porphyrin arrays have been attempted to mimic light-harvesting complexes such as LH1 and LH2 systems in nature.4 To prepare porphyrin based molecular assemblies for an artificial light harvesting antenna system, various strategies have * To whom correspondence should be addressed. E-mail: dongho@ yonsei.ac.kr (D.K.);
[email protected] (A.O.). † Yonsei University. ‡ Kyoto University.
been employed such as covalently5 and noncovalently6 as well as dendritically linked arrays.7 Among them, covalently linked molecular arrays have strong advantages of robust framework and precise control in the spatial arrangement and connecting spacers in spite of their synthetic difficulty. For these advantages, we have recently investigated the excitation energy transfer rates and mechanisms in various covalently linked cyclic porphyrin arrays with different constituent subunits.8 Along this line, wellordered systems possessing relatively planar and rigid threedimensional structures were mainly considered because the EEH processes in flat polygon systems are suitable for a simple exciton hopping model as proposed by Fleming and co-workers.9 On the basis of a regular polygon model, for example, we obtained the fast EEH time of 4.0 ps between two adjacent pigments, meso-meso directly linked zinc(II) porphyrin dimers linked by 1,3-phenylene spacers in hexameric cyclic arrays.8d As the interchromophoric distance between adjacent subunits becomes larger, the EEH time becomes slower (35 ps) for a giant array consisting of six zinc(II) porphyrin tetramers as subunits in spite of their enlarged transition dipole moments.8a According to these results, it is inferred that the EEH processes should be strongly dependent upon the interchromophoric distance rather than other parameters governing Fo¨rster-type incoherent energy transfer processes. In contrast, orientation factors between transition dipole moments of subunits would be more effective in the EEH processes in the cases of
10.1021/jp904729y CCC: $40.75 2009 American Chemical Society Published on Web 10/07/2009
Migration Processes in Zinc(II) Porphyrin Arrays CHART 1: Molecular Structures of Z2, 2Z2, and CNZ, Where N is the Number of Zinc(II) Porphyrin Monomers
heterogeneous and disordered molecular array systems such as dendrimer and nonplanar macrocyclic array systems.7 In this work, we have prepared a series of cyclic zinc(II) porphyrin arrays CNZ such as C10Z, C12Z, C16Z, C18Z, C24Z, and C32Z,8a which are composed of the same constituent subunits, meso-meso directly linked zinc(II) porphyrin dimers (Z2) bridged by 1,3-phenylene spacers (N is the number of zinc(II) porphyrin monomers) in order to study the EEH rates depending on the orientation between the constituent Z2 subunits (Chart 1). According to the results from the 1H NMR and timeresolved fluorescence lifetime measurements, it is inferred that the number of conformational isomers increased as the molecular size became larger. Femtosecond transient absorption anisotropy (TAA) decay measurement provided detailed depolarization processes including excitation energy migration in the lowest singlet excited states of cyclic porphyrin arrays. We have determined the number of effective hopping sites, which converses as molecular size increases, from the observed depolarization time based on the irregular polygon model giving rise to detailed information on the conformational heterogeneity in larger cyclic arrays. Finally, we have focused our attention on the relationship between EEH processes and three-dimensional structures of cyclic macromolecules with structural analyses by ab initio quantum mechanical calculations. II. Experimental Section Sample Preparation. Synthetic procedures based on intramolecular Ag(I)-promoted cyclization reaction and characterizations are described elsewhere.8a HPLC grade toluene was purchased from Aldrich and used without further purification as solvents. All measurements were carried out at ambient temperature (23 ( 2 °C). Steady-State Absorption and Emission. Steady-state absorption spectra were acquired with an UV-vis spectrometer (Scinco, F4100, and Varian, Cary 5000). Steady-state fluorescence spectra of the samples were recorded on a fluorescence spectrometer (Hitachi, FL2500) and corrected with the comparison of the well-known chromophores such as rhodamine and coumarin dyes. Time-Resolved Emission. Time-resolved fluorescence was detected using a typical time-correlated single-photon-counting (TCSPC) technique. As an excitation light source, we used a homemade cavity-dumped Ti:sapphire oscillator which provides a high repetition rate (200-400 kHz) of ultrashort pulses (100
J. Phys. Chem. B, Vol. 113, No. 45, 2009 15075 fs at full width at half-maximum (fwhm)) pumped by a continuous wave (cw) Nd-YVO4 laser (Coherent, Verdi). The output pulse of the oscillator was frequency-doubled by a 1 mm thickness of a second harmonic crystal (β-barium borate, BBO, CASIX). The detection system consisted of a microchannel plate photomultiplier (MCP-PMT, Hamamatsu, R3809U-51) with a thermoelectric cooler (Hamamatsu, C4878), a time-to-amplitude converter (TAC, EG&G Ortec, 457), two constant fractional discriminators (EG&G Ortec, 584 (signal) and Canberra, 2126 (trigger)), and two wide-band preamplifiers (Philip Scientific (signal) and Mini-Circuit (Hamamatsu, (trigger)). A personal computer with a multichannel analyzer (MCA, Canberra, PCA3) was used for data storage and signal processing. The overall instrumental response function was about 60 ps (fwhm). A vertically polarized pump pulse by a Glan-laser polarizer was irradiated to samples, and a sheet polarizer, set at an angle complementary to the magic angle (54.7°), was placed in the fluorescence collection path to obtain polarization-independent fluorescence decays. Femtosecond Transient Absorption Anisotropy Decay. A dual-beam femtosecond time-resolved transient absorption (TA) spectrometer consisted of two independently tunable homemade optical parametric amplifiers (OPA) pumped by a regeneratively amplified Ti:sapphire laser system (Spectra-Physics, HurricaneX) operating at 5 kHz repetition rate and an optical detection system. The OPA was based on noncollinearly phase-matching geometry, which was easily color-tuned by controlling optical delay between white light continuum seed pulses (450-1400 nm) and visible pump pulses (400 nm) produced by using a sapphire window and BBO crystal, respectively. The generated visible OPA pulses had a pulse width of ∼35 fs and an average power of 10 mW at 5 kHz repetition rate in the range of 500-700 nm after a fused-silica prism compressor. Two OPA pulses were used as the pump and probe pulses, respectively, for TA measurement. The probe beam was split into two parts. The one part of the probe beam was overlapped with the pump beam at the sample to monitor the transient (signal), while the other part of the probe beam was passed through the sample without overlapping the pump beam to compensate the fluctuation of probe beam. The time delay between pump and probe beams was carefully controlled by making the pump beam travel along a variable optical delay (Newport, ILS250). To obtain the time-resolved transient absorption difference signal at specific wavelength, the monitoring wavelength was selected by using a narrow interference filter (fwhm ∼ 10 nm). By chopping the pump pulses at 47 Hz, the modulated probe pulses as well as the reference pulses were detected by two separate photodiodes (New Focus, Femtowatt Photoreceiver). The modulated signals of the probe pulses were measured by a gated integrator (SRS, SR250) and a lock-in amplifier (EG&G, DSP7265) and stored in a personal computer for further signal processing. The polarization angle between pump and probe beam was set to magic angle (54.7°) in order to prevent polarization-dependent signals. In general experimental conditions, time resolutions of less than 50 fs were achieved. For time-resolve transient absorption anisotropy (TAA) measurement, both I| and I⊥ signals were collected simultaneously by combination of polarizing beam splitter cube and dual lock-in amplifiers as the following equation:10
r(t) ) (I| - I⊥)/((I| + 2I⊥) where I| and I⊥ represent TA signals with the polarization of the pump and probe pulses being mutually parallel and
15076
J. Phys. Chem. B, Vol. 113, No. 45, 2009
Figure 1. (a) Steady-state absorption and (b) fluorescence spectra with photoexcitation at 550 nm of Z2, 2Z2, and CNZ in toluene. The absorption spectra are normalized to the high-energy Soret bands at 420 nm, and fluorescence spectra are normalized to their respective maximum intensities.
perpendicular, respectively. The pump pulse was set to vertical polarization and that of probe pulse was set to 45° with respect to the pump pulse by using Glan-laser polarizers and half-wave plates. After the probe pulse passes through the sample cell, it was split by a polarizing beam splitter cube and then detected by two separate photodiodes. Two gated integrators and two lock-in amplifiers record the signal simultaneously within a single scan. As a standard, anisotropy measurement showed a clean single-exponential decay with reorientational relaxation times of 122.1 ( 0.3 ps and the initial anisotropy r(0) value of 0.39 ( 0.02 for rhodamine 6G dye in methanol, which are wellmatched in other reference.11 For all TAA measurements, the wavelength of the pump pulse was set to 570 nm with an average power of less than 40 µW and a thin absorption cell with a path length of 500 µm was used to eliminate additional chirping. Computational Methods. Quantum mechanical calculations were carried out with the Gaussian03 program suite.12 Geometry optimizations were performed by density functional theory (DFT) method with Becke’s three-parameter hybrid exchange functionals and the Lee-Yang-Parr correlation functional (B3LYP), employing a basis set consisting of 6-31G(d) for all atoms in 2Z2. In the case of C12Z, a semiempirical PM3 Hamiltonian was used for geometry optimization of groundstate structure. In all calculations, peripheral meso-aryl substituents were excluded to reduce computational cost. III. Results UV-Vis Absorption and Fluorescence Spectra. Figure 1 shows the steady-state absorption and fluorescence spectra of Z2, 2Z2, and CNZ in toluene. Zinc(II) porphyrin monomers have two distinct absorption bands which correspond to intense B- (Soret) and relatively weak Q-bands appearing at 400 and 550 nm, respectively.13 Through meso-meso direct linkage,
Yoon et al. strong exciton coupling between zinc(II) porphyrin monomers in Z2 leads to split B-band: a high-energy B-band remaining at the position (416 nm) similar to that of zinc(II) tetraphenylporphyrin monomer (ZnIITPP) and low-energy B-band at 451 nm resulting from J-type exciton coupling along the long molecular axis.4b Both B- and Q-bands in Z2 are broader than those in ZnIITPP, which are caused by rotational flexibility along the meso-meso direct linkage.14 As the number of Z2 moieties increases from C10Z to C32Z, the exciton splitting energies in B-bands which correspond to energy differences between highand low-energy B-bands become larger. On the other hand, Q-bands exhibit negligible red shift with slight intensification (Table 1 and Figure S1 in the Supporting Information). The overall spectral features in the absorption spectra seem to be determined by the constituent subunit Z2. This result indicates that additional exciton coupling in CNZ should be accumulated as the number of Z2 subunits increases in CNZ. Furthermore, the peak positions of low-energy B-bands in CNZ converge into a single position, which is similar to the spectral feature observed in directly linked linear arrays Z2-Z128.15 However, in this case, such convergence of low-energy B-bands does not come from conformational flexibility as seen in linear porphyrin arrays but from exciton couplings among nonunidirectional transition dipole moments in circular geometries.16 As compared to the absorption spectra, the fluorescence spectra of CNZ with photoexcitation at Q-bands are similar to each other except for Z2 and 2Z2 (Figure 1b). According to the similar features in both absorption and fluorescence spectra in the Q-band region of CNZ, we could assume that the structural environment between two adjacent Z2 remains relatively the same in all cyclic arrays CNZ. In other words, three-dimensional orientation of 2Z2 moieties seems to be maintained in all cyclic arrays. All parameters obtained by the steady-state spectroscopic measurements are summarized in Table 1. Time-Resolved Fluorescence Decay. To investigate the excited-state dynamics of CNZ, we have performed timeresolved fluorescence decay measurements using TCSPC technique. The fluorescence temporal profiles of CNZ monitored at 625 nm with photoexcitation of high-energy B-band at 400 nm in toluene are depicted in Figure 2, and their fitting parameters are listed in Table 2. Although CNZs are photoexcited at higher electronic S2 states, the observed decay times can be regarded as the S1-state lifetimes since the internal conversion process from S2 to S1 state in Z2 and CNZ is beyond the temporal resolution of TCSPC. While the fluorescence decays of C10Z and C12Z show single-exponential decay, the other larger cyclic arrays exhibit double-exponential decays. Fast components (
