Long-Distance Sequential Charge Separation at Micellar Interface

Dec 31, 2014 - from the core to the hydrophilic interface leads to long-distance ET with a low ... scattering is ∼7 nm (Figure S1 in Supporting Info...
1 downloads 0 Views 991KB Size
Download PDF
Letter pubs.acs.org/JPCL

Long-Distance Sequential Charge Separation at Micellar Interface Mediated by Dynamic Charge Transporter: A Magnetic Field Effect Study Tomoaki Miura,*,† Kiminori Maeda,‡ Hisao Murai,§ and Tadaaki Ikoma†,∥,⊥ †

Department of Chemistry, Niigata University, 2-8050 Ikarashi, Nishi-ku, Niigata 950-2181, Japan Department of Chemistry, Graduate School of Science and Engineering, Saitama University, 255 Shimo-okubo, Sakura-ku, Saitama 338-8570, Japan § Department of Chemistry, Shizuoka University, 836 Oya, Suruga-ku, Shizuoka 422-8017, Japan ∥ Center for Instrumental Analysis, Niigata University, 2-8050 Ikarashi, Nishi-ku, Niigata 950-2181, Japan ⊥ Core Research for Evolutionary Science and Technology, Japan Science and Technology Agency, 4-1-8 Honcho, Kawaguchi 332-0012, Japan ‡

S Supporting Information *

ABSTRACT: Construction of photogenerated long-lived charge-separated states is crucial for light-energy conversion using organic molecules. For realization of cheap and easy-tomake long-distance electron transfer (ET) systems, we have developed a supramolecular donor(D)−chromophore(C)−acceptor(A) triad utilizing a micellar interface. Alkyl viologen (A2+) is adsorbed on the hydrophilic interface of Triton X-100 micelle, which bears D units in the hydrophobic core. Excited triplet state of a hydrophobic flavin C entrapped in the supercage gives rise to primary ET from D, which is followed by the secondary ET from C−• to A2+ to give the long-lived (>10 μs) charge-separated state with negligible yield of escaped C−•. Analysis of magnetic field effect reveals that diffusion of C−• from the core to the hydrophilic interface leads to long-distance ET with a low charge recombination yield of ∼20%. This novel concept of “dynamic charge transporter” has important implications for development of photon-energy conversion systems in solution phase.

I

studies that the dynamic diffusion of the guest molecules is the cause of efficient CR and escape. One of the authors recently succeeded in fixing a viologen derivative as an A molecule on the hydrophilic interface of a nonionic micelle and observed a singlet-born RP.12 However, the lifetime of the RP is not very long as ∼10 ns, which is due to the diffusion-induced escape and recombination at the hydrophilic interface. In this Letter, we report preparation of low-cost, environmentally friendly, and easy-to-make sequential ET triads by utilizing the Aadsorbed micelle and riboflavin (RF or Vitamin B2) derivatives as chromophores (C). We take advantage of the diffusional dynamics of the C molecule as a “dynamic charge transporter” to realize the long-distance CS without generation of the escaped free radical as has been confirmed by detailed analysis of optically detected MFE. D, C, and A molecules used in this study were all commercially available, of which the structures are given in Chart 1.

t is important to create electron donor (D)−acceptor (A) systems that exhibit a long-lived charge-separated state for the realization of efficient photon energy conversion. Recent development of organic synthesis has allowed us to optimize the structure of covalently linked donor−acceptor (D−A) systems for the long-distance charge separation (CS).1,2 A number of multiarray sequential CS systems inspired by the natural photosynthesis have been reported, some of which have successfully achieved superlong CS lifetimes of ∼10 μs to a few milliseconds.3−5 Supramolecular D−A complexes are also attracting increasing attention as an alternative for the covalent systems because of low cost and the ease with which the D and A units are assembled.6,7 Micelles are one of the most classical examples of supramolecules. Entrapment of D and A molecules in the supercage enhances the CS efficiency. Electron transfer (ET) mechanisms in such systems have been extensively studied by focusing on spin dynamics of the in-cage charge-separated states, or radical pairs (RPs), by means of optically detected magnetic field effect (MFE) or time-resolved EPR.8−11 These studies indicate that the charge recombination (CR) and escape of the generated RP usually occur within less than a few microseconds, which is not sufficient for the subsequent conversion reactions to occur. It is widely accepted from these © 2014 American Chemical Society

Received: November 26, 2014 Accepted: December 31, 2014 Published: December 31, 2014 267

DOI: 10.1021/jz502495u J. Phys. Chem. Lett. 2015, 6, 267−271

Letter

The Journal of Physical Chemistry Letters Chart 1

Dioctadecyl viologen dibromide (OdVBr2) was solubilized in the Triton X-100 (TX) dispersion in tetrahydrofuran (THF)/ water mixed solvent to prepare the viologen-adsorbed micellar cage by the previously reported method.12 The alkylphenoxy group of the TX molecule acts as a D unit. Following removal of THF by air blowing, water-insoluble riboflavin tetrabutylate (RFTB) or water-soluble riboflavin (RF) was added as the C to construct TX−RFTB/RF−odV2+ D−C−A2+ complexes. The total concentration of TX, odVBr2, and RF/RFTB were 87 mM, 1.8 mM, and 0.18 mM, respectively. As reference D−C systems, solutions of RFTB/RF in TX micelle without odV2+ were also prepared (TX−RFTB/RF).8 The average number of flavin molecules per one micelle is less than unity. The previous study12 indicates that the concentration of the micelles without odV2+ adsorption is small at [odV2+] = 1.8 mM. Details on materials and spectroscopic methods are provided in the Supporting Information. All the spectroscopic measurements were carried out at room temperature. The sample solutions were deoxygenated by Ar purging before and during TA measurements. In this Letter, results for RFTB systems are mainly described, and those for RF systems are given in the Supporting Information. Diameter of the supercage estimated from dynamic light scattering is ∼7 nm (Figure S1 in Supporting Information). It is expected, as in the case of regular micelles, that the alkyl chains of TX assemble to form a hydrophobic core (∼3−5 nm diameter), which is surrounded by the palisade layer consisting of hydrated polyoxyethylene (POE) chains (∼1−2 nm thickness). Thus, D units distribute at the boundary between the core and the palisade layer. On the other hand, the viologen moiety is considered to be fixed at the hydrophilic outer layer of the supercage. Considerable quenching of fluorescence (FL) is observed in TX−RFTB−odV2+ and TX−RFTB systems (Supporting Information Figure S2), which indicates CS from TX to the excited singlet flavin (1F*). Figure 1a and b, respectively, show nanosecond transient absorption (nsTA) spectra in TX−RFTB and TX−RFTB− odV2+ systems upon photoexcitation by third harmonics of the Nd:YAG laser (λ = 355 nm). In the absence of odV2+, the excited triplet state of the flavin (3F*) observed at λ ∼ 600−700 nm decays within a few hundreds of nanoseconds, whereas the flavin anion radical (F−•) is observed at later times with a long lifetime (>10 μs). This indicates CS from 3F* to give the RP 3 [TX+• F−•] as has been reported previously.8 In the previous study, F−• observed at the late times has been assigned to the free radical escaped to the bulk water due to the diffusion from the hydrophobic region to the outer layer as evidenced by the pH-dependent slow (∼microseconds) protonation kinetics at acidic conditions.

Figure 1. nsTA spectra without external magnetic field for (a) TX− RFTB and (b) TX−RFTB−odV2+ systems at indicated times following excitation by the third harmonics of Nd:YAG laser (λ = 355 nm).

Spectra with odV2+ within ∼600 ns are similar to those observed without odV2+. However, the spectral shape changes drastically at later times, where absorption bands at λ = 400 and 620 nm grow accompanying further recovery of the bleaching. The whole spectrum at t = 5 μs is assigned to the viologen monocation radical (odV+•).13 The spectral change clearly indicates secondary ET from F−• to odV2+; namely, the present system actually works as a D−C−A2+ triad. Energy levels of excited states and RPs (Supporting Information) indicate that the ET processes are energetically possible. It is surprising that the yield of escaped free F−• without reducing odV2+ is negligible as evidenced by the absence of F−• TA signal at the late times. This indicates that the rate of escape to the bulk water, which is observable only without odV2+, is much smaller than that of the secondary ET. Figure 2 shows the TA kinetics at λ = 400 and 710 nm, at which odV+• and 3F* are mainly observed, respectively. The decay rate for 3F* corresponds to the CS rate from 3F* (kCST = 2.7 × 106 s−1) because the natural triplet lifetime is very long (>a few microseconds).13 odV+• signal shows a gradual rise of ∼600 ns, which reflects the secondary ET. The yield of odV+• is increased about 10% by an applied magnetic field of 250 mT. The positive MFE is typical of triplet-born RPs that recombine 268

DOI: 10.1021/jz502495u J. Phys. Chem. Lett. 2015, 6, 267−271

Letter

The Journal of Physical Chemistry Letters

Figure 2. TA kinetics for the TX−RFTB−odV2+ system at λ = 710 nm without magnetic field (top, orange), λ = 400 nm without magnetic field (bottom, blue) and λ = 400 nm with a magnetic field of 250 mT (bottom, red). The green trace indicates the subtracted MFE kinetics. Black dashed line indicates fitting by the sequential ET model described in the text.

to the singlet ground state via the field-sensitive singlet−triplet (S−T) mixing.14 Thus, the singlet-born RP indicated from the FL quenching seems to be so short-lived (≪ 10 ns), due to spin-allowed CR, that it does not contribute to the secondary ET.12 The overall ET processes are summarized as follows; CS from the 3F* gives the primary RP [TX+• F−•] or [D+•−C−•− A2+], which is followed by generation of the secondary RP [TX+• odV+•] or [D+•−C−A+•]. The kinetics of odV+• at 0 mT is fitted by a simple sequential reaction model ignoring reactions of the primary RP other than the secondary ET (Figure 2, for detail, see Supporting Information). The obtained rate constants for generation and decay of the secondary RP are kET2 = 1.8 × 106 s−1 and kdecay ≪ 1 × 105 s−1, respectively. The kinetics are almost unchanged by double dilution of the solution (Supporting Information Figure S7), which confirms that all the ET reactions occur in the same supercage. Because TX and odV2+ assemble by the strong hydrophobic interaction, the escape rate for the secondary RP is considered to be very small as surfactant exchange rates for typical micelles (102−104 s−1). Thus, we have succeeded in creating a long-lived CS state, whose lifetime is comparable to those of the covalently linked systems. The electron is transferred from the core part to the hydrophilic outer layer over the long distance of >1 nm by the C molecule, which is considered to be the reason for the long lifetime. The longdistance charge transportation is considered to be related to the diffusion of C−• toward the hydrophilic outer layer as reported previously in the TX−RFTB system. To confirm this idea and obtain detailed ET dynamics, we analyze MFE in detail as described in the following sections. Figure 3a shows the magnetically affected reaction yield (MARY) spectra in the TX−RFTB system with and without odV2+. The spectral shape is explained by the hyperfine mechanism and slight contribution of the relaxation mechanisms.14−16 The shape and amplitude of the MARY spectrum are unaffected by addition of odV2+. This indicates that the MFE is mostly due to CR of the primary RP even in the presence of odV2+; namely, the recombination of the secondary RP is negligible. Thus, the MFE on the secondary RP directly

Figure 3. (a) MARY spectra at 5 μs in the TX−RFTB (blue) and TX−RFTB−odV2+ (red) systems. Dashed line indicates the simulation for the red spectrum. (b) Kinetics of the field-sensitive RP in the TX− RFTB−odV2+ system at 0 mT detected by delay shifting experiments of nanosecond-magnetic-field switching (black). Colored lines indicate fitting by the sequential kinetic model.

reflects the dynamics of the primary RP as described in the following sections. Fitting of the MARY spectrum by double-Lorentzian function gives a high-field saturation MFE value of 11.4%. The ratio between the rate constants for spin-selective CR (kCRS) and non-spin-selective decay (kd) can be calculated from the MFE amplitude according to the simple MFE scheme with a steady state approximation of S−T mixing (for detail, see Supporting Information) as kd k CRS

∼1

(1)

Note here that the kd corresponds to the transition rate from the CR-active state to any CR-inactive states including the secondary RP, which has turned out to be CR-inactive. The CR yield for the primary RP can be calculated from the ratio as ΦCR (0 mT) =

k CRS/4 ∼ 0.2 (k CRS/4) + kd

(2)

Although this yield is based solely on the triplet-born RP component, small ΦCR is ideal in terms of energy conversion. In order to determine the rate constants, we examined delayshifting experiments with nanosecond magnetic field switching, which directly monitors the kinetics of the field-sensitive RP (Figure 3b, for detail, see Supporting Information).17 However, the apparent decay rate at 0 mT is almost the same with the decay rate of the precursor triplet state (kCST). This observation indicates that the intrinsic decay rate of the RP at 0 mT (kRP0) is much larger than kCST. A simple kinetic fitting gives a smallest limit as kRP0 > 2 × 107 s−1. On the other hand, appearance of MFE requires enough lifetime for the hyperfine-induced S−T 269

DOI: 10.1021/jz502495u J. Phys. Chem. Lett. 2015, 6, 267−271

Letter

The Journal of Physical Chemistry Letters mixing of ∼10 ns, from which one can roughly estimate that kRP0 < 1 × 108 s−1.14 Thus, the range of kRP0 is 2 × 107 s−1 < kRP0 < 1 × 108 s−1

distance for hydrophilic RF. The second difference indicates greater contribution of the relaxation mechanism, which might be due to a longer RP lifetime. The diffusion rate of C−• might be smaller in the RF system due to hydrogen bonding of OH groups. These results indicate that the transportation dynamics can be tuned by substituents on the transporter molecule.8 The ΦCR value of 20% is low, but some reported covalent systems reach ΦCR ∼ 0 for intermediate CS states.18 In order to inhibit the energy-wasting CR process, we are currently trying to control molecular arrangement and diffusion using nonionic vesicles (niosomes) as possibly more rigid architectures.19

(3)

The relationship kRP0 = kd + (kCRS/4) gives the range for the rates 1.6 × 107 s−1 < k CRS ∼ kd < 6 × 108 s−1

(4)

The MARY spectrum can be simulated by the reported modified-Liouville eq (Supporting Information) with parameters consistent with eq 4 (kCRS = 2.2 × 107 s−1 and kd = 2 × 107 s−1) as shown in Figure 3a.16 It is noteworthy that kRP0 obtained from the MFE analysis (eq 3) is at least 10 times larger than kET2 = 1.8 × 106 s−1 obtained from the TA analysis. This means that the fieldsensitive (CR-active) primary RP is not the main precursor of the secondary RP; namely, there exists a CR-inactive primary RP, which is the main precursor of the secondary RP. This finding is the clear evidence of the in-cage diffusion of C−• as the mechanism of long-distance charge migration. As shown in Scheme 1,the CR-active pseudocontact RP [D+•−C−•∥A2+] is

■

ASSOCIATED CONTENT

S Supporting Information *

Details about chemicals used in this study, steady-state and time-resolved fluorescence, energetics of ET, nsTA and MFE data in RF systems, data of nanosecond field switching and details about calculation and simulations. This material is available free of charge via the Internet at http://pubs.acs.org

■

AUTHOR INFORMATION

Corresponding Author

Scheme 1. ET and Molecular Diffusion Dynamics in the TX(D)−RFTB(C)−odV2+, the A2+, Systema

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors are grateful to Prof. Taku Hasobe for permission to use equipment in his laboratory at Keio University, Japan. This work was financially supported by a Grant-in-Aid for Young Scientists (B, No. 25810009), a grant for the Promotion of Niigata University Research Projects, a CREST grant from JST, and a grant from the Network Joint Research Center for Materials and Devices. K.M. was supported by Grant-in-Aid for Research Activity Start-up (No. 26888004).

■

Left side of ∥ indicates the region near the hydrophobic core, whereas the right side indicates the hydrophilic outer layer. a

REFERENCES

(1) Gray, H. B.; Winkler, J. R. Long-Range Electron Transfer. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 3534−3539. (2) Albinsson, B.; Martensson, J. Long-Range Electron and Excitation Energy Transfer in Donor-Bridge-Acceptor Systems. J. Photochem. Photobiol., C 2008, 9, 138−155. (3) Imahori, H.; Guldi, D. M.; Tamaki, K.; Yoshida, Y.; Luo, C.; Sakata, Y.; Fukuzumi, S. Charge Separation in a Novel Artificial Photosynthetic Reaction Center Lives 380 ms. J. Am. Chem. Soc. 2001, 123, 6617−6628. (4) Gust, D.; Moore, T. A.; Moore, A. L. Molecular Mimicry of Photosynthetic Energy and Electron Transfer. Acc. Chem. Res. 1993, 26, 198−205. (5) Greenfield, S. R.; Svec, W. A.; Gosztola, D.; Wasielewski, M. R. Multistep Photochemical Charge Separation in Rod-Like Molecules Based on Aromatic Imides and Diimides. J. Am. Chem. Soc. 1996, 118, 6767−6777. (6) Fujitsuka, M.; Shimakoshi, H.; Hisaeda, Y.; Majima, T. Reorganization Energy of Supramolecular Donor-Acceptor Dyad of Octaethylporphyrin Isomers and Axial-Coordinated Acceptor: Experimental and Computational Study. J. Photochem. Photobiol., A 2011, 217, 242−248. (7) Gayathri, S. S.; Wielopolski, M.; Perez, E. M.; Fernandez, G.; Sanchez, L.; Viruela, R.; Orti, E.; Guldi, D. M.; Martin, N. Discrete Supramolecular Donor−Acceptor Complexes. Angew. Chem., Int. Ed. 2009, 48, 815−819. (8) Horiuchi, M.; Maeda, K.; Arai, T. Dynamic Process of the Photochemically Produced Flavin Radicals in a Neutral Micelle

3

initially populated by CS from the D unit to C* near the core region. The diffusion of C−• toward the hydrophilic outer sphere gives rise to the CR-inactive separated primary RP [D+•∥C−•−A2+] with the rate nearly equal to kd (107−108 s−1). Further diffusion of C−• in the outer sphere induces the secondary ET with the A2+, which resides in this region, with the rate nearly equal to kET2. Thus, it has been demonstrated in the present system that the long-distance charge migration from D to A2+ is gated by the interphase transportation of C−•; namely, the C molecule acts as a “dynamic charge transporter”. The low CR yield of 20% attributes to generation of the CR-inactive primary RP due to in-cage diffusive separation. Diffusion of guest molecules in amphiphilic assemblies is usually a cause of efficient CR or escape, but one can take advantage of such dynamics by carefully arranging the molecular units. The dynamic transporter can be a new concept for creating solution-based photon energy conversion systems. Qualitatively similar results are obtained when the RF is used as C (Supporting Information). The major differences from the RFTB system are (1) smaller kCST and (2) nonsaturation behavior of MARY spectrum at the high fields. The first difference can simply be explained by longer initial D−C 270

DOI: 10.1021/jz502495u J. Phys. Chem. Lett. 2015, 6, 267−271

Letter

The Journal of Physical Chemistry Letters Studied by a Magnetic Field Effect. Chem. Phys. Lett. 2004, 394, 344− 348. (9) Goez, M.; Henbest, K. B.; Windham, E. G.; Maeda, K.; Timmel, C. R. Quenching Mechanisms and Diffusional Pathways in Micellar Systems Unravelled by Time-Resolved Magnetic-Field Effects. Chem.Eur. J. 2009, 15, 6058−6064. (10) Choudhury, S. D.; Basu, S. Exploring the Extent of Magnetic Field Effect on Intermolecular Photoinduced Electron Transfer in Different Organized Assemblies. J. Phys. Chem. A 2005, 109, 8113− 8120. (11) Lebedeva, N. V.; Schmidt, R. D.; Concepcion, J. J.; Brennaman, M. K.; Stanton, I. N.; Therien, M. J.; Meyer, T. J.; Forbes, M. D. E. Structural and Ph Dependence of Excited State Pcet Reactions Involving Reductive Quenching of the Mlct Excited State of [RuII(Bpy)2(Bpz)]2+ by Hydroquinones. J. Chem. Phys. A 2011, 115, 3346−3356. (12) Miura, T. Supramolecular Control of the Spin-Dependent Dynamics of Long-Lived Charge-Separated States at the Micellar Interface as Studied by Magnetic Field Effect. J. Phys. Chem. B 2013, 117, 6443−6454. (13) Montalti, M.; Credi, A.; Prodi, L.; Gandolfi, M. T. Handbook of Photochemistry, 3rd ed.; CRC Press, Taylor & Francis Group: Boca Raton, FL, 2006. (14) Steiner, U. E.; Ulrich, T. Magnetic-Field Effects in ChemicalKinetics and Related Phenomena. Chem. Rev. 1989, 89, 51−147. (15) Hayashi, H.; Nagakura, S. Theoretical-Study of Relaxation Mechanism in Magnetic-Field Effects on Chemical-Reactions. Bull. Chem. Soc. Jpn. 1984, 57, 322−328. (16) Maeda, K.; Miura, T.; Arai, T. A Practical Simulation and a Novel Insight to the Magnetic Field Effect on a Radical Pair in a Micelle. Mol. Phys. 2006, 104, 1779−1788. (17) Suzuki, T.; Miura, T.; Maeda, K.; Arai, T. Spin Dynamics of the Radical Pair in a Low Magnetic Field Studied by the Transient Absorption Detected Magnetic Field Effect on the Reaction Yield and Switched External Magnetic Field. J. Phys. Chem. A 2005, 109, 9911− 9918. (18) Kodis, G.; Liddell, P. A.; de la Garza, L.; Moore, A. L.; Moore, T. A.; Gust, D. Photoinduced Electron Transfer in Π-Extended Tetrathiafulvalene-Porphyrin-Fullerene Triad Molecules. J. Mater. Chem. 2002, 12, 2100−2108. (19) Mandal, S.; Banerjee, C.; Ghosh, S.; Kuchlyan, J.; Sarkar, N. Modulation of the Photophysical Properties of Curcumin in Nonionic Surfactant (Tween-20) Forming Micelles and Niosomes: A Comparative Study of Different Microenvironments. J. Phys. Chem. B 2013, 117, 6957−6968.

271

DOI: 10.1021/jz502495u J. Phys. Chem. Lett. 2015, 6, 267−271