Excited State Charge Redistribution and Dynamics in the Donor-π

Sep 10, 2013 - ABSTRACT: Chromophores containing a donor-π-acceptor (D-π-A) motif have been shown to exhibit many interesting photophysical properti...
0 downloads 0 Views 2MB Size
Article pubs.acs.org/JPCB

Excited State Charge Redistribution and Dynamics in the Donor-πAcceptor Flavin Derivative ABFL Raymond F. Pauszek, III,† Goutham Kodali,†,∥ Stuart T. Caldwell,‡ Brian Fitzpatrick,‡ Nada Y. Zainalabdeen,‡,⊥ Graeme Cooke,‡ Vincent M. Rotello,§ and Robert J. Stanley*,† †

Department of Chemistry, Temple University, Philadelphia, Pennsylvania 19122, United States Glasgow Centre for Physical Organic Chemistry, WestCHEM, School of Chemistry, College of Science and Engineering, University of Glasgow, Glasgow, G12 8QQ, United Kingdom § Department of Chemistry, University of Massachusetts at Amherst, Amherst, Massachusetts 01003, United States ‡

S Supporting Information *

ABSTRACT: Chromophores containing a donor-π-acceptor (D-π-A) motif have been shown to exhibit many interesting photophysical properties. The lowest electronic transition of a flavin derivative containing this motif, azobenzylflavin (ABFL), has previously been shown to be highly sensitive to solvent environment and hydrogen bonding ligands. To better understand this sensitivity, we have investigated the excited state charge redistribution and dynamics of ABFL in a low-dielectric, non-hydrogen bonding solvent by steady-state Stark and femtosecond optical transient absorption spectroscopies. The Stark measurements reveal the difference dipole moment, Δμ⃗ 01, between the ground and first excited states to be 22.3 ± 0.9 D. The direction of Δμ⃗ 01 in the molecular frame was assigned with the aid of TD-DFT and finite field calculations, verifying the hypothesis that electron density moves from the diethylaniline donor to the flavin acceptor in the excited state. The magnitude of the difference dipole moment was used to estimate the hyperpolarizability of ABFL, β0 = 720 × 10−30 esu. Subsequent excited state decay via charge recombination was shown to take place in a few picoseconds. The data was best fit to a kinetic model composed of a sub-picosecond internal conversion step from S2→S1, followed by a 5 ps decay to the ground state. A competing process involving formation of an additional long-lived state from S1 was also observed. Cyclic voltammetry shows one oxidation and two reduction waves and is completely reversible. This analysis lays the groundwork for developing new flavin dyads with the desired excited electronic state properties for applications such as nonlinear optical devices, molecular electronics applications, or dye-sensitized solar cells.



INTRODUCTION

absorption spectra are a sum of the donor and acceptor absorption spectra.6 The conjugated azo bridge of ABFL allows for mixing of the electronic states of the aniline and flavin moieties, and thus these can no longer be considered two distinct chromophores. In addition, this transition shows a significant solvatochromic shift in going from toluene to ethanol, on the order of 50 nm. The second optically accessible transition exhibits only weak solvatochromism and significantly lower extinction. Based on these observations, the S0→S1 transition was assigned as an ICT transition, while the S0→S2 band has been tentatively assigned to a (localized) π−π* transition. While the presence of strong charge transfer character is promising for various molecular electronic applications, ultimately the dynamics of the excited state of

Push−pull chromophores, in which an electron donor (D) and acceptor (A) are connected through a conjugated polyene chain, show promise in applications such as nonlinear optics (NLO) and solar energy conversion.1 These D-π-A motifs often exhibit low-lying intramolecular charge transfer (ICT) states, which give rise to large hyperpolarizabilities, solvatochromism, and optical nonlinearities.2−4 Recently, a flavin derivative 8-[[p-[bis(ethyl)amino] phenyl]azo]isobutylflavin (ABFL, Figure 1), was synthesized by Cooke and Rotello.5 ABFL consists of a diethylaniline moiety attached through an azo bridge to the C(8) position of N(10)-isobutylflavin (Fl) to form a D-π-A molecule with the inherent molecular recognition capabilities of flavin via hydrogen bonding at the pyrimidine moiety. Previous spectroscopic characterization of ABFL revealed the presence of a strong absorption band significantly red-shifted from the S0→S1 transition of Fl in organic solvents.5 This is in stark contrast to similar donor−acceptor flavin dyads containing nonconjugated linkages to the flavin, whose © 2013 American Chemical Society

Special Issue: Michael D. Fayer Festschrift Received: June 28, 2013 Revised: September 10, 2013 Published: September 10, 2013 15684

dx.doi.org/10.1021/jp406420h | J. Phys. Chem. B 2013, 117, 15684−15694

The Journal of Physical Chemistry B

Article

Figure 1. Neutral (left) and charge separated (right) forms of ABFL.

Stark Spectrometer. The experimental setup of the Stark spectrometer has been discussed in detail previously, with some modifications.9 Briefly, light from a 300 W Xe arc lamp (Oriel 69911/68945) is focused onto the entrance slit of a 1/8 m monochromator (CV Instruments CM110) with a 2 nm bandpass. The beam is collimated, passed through a depolarizer followed by a Glan-Taylor polarizer, and then focused onto the sample immersed in liquid nitrogen held inside a dual chamber cryostat (JANIS Research, based on the design of Andrews).10 The transmitted light is then collected and focused onto a Si photodiode. The resulting photocurrent is amplified with a current preamplifier (Stanford Research Systems SR570). The sample cuvette was constructed from 100 Ω cm−2 resistivity 0.7 mm thick Corning 1737 boro-aluminosilicate glass slides coated on one side with indium tin oxide (Delta Technologies #CB-90IN-S107). The conductive sides of the slides were placed facing each other, separated by 55 μm thick kapton spacers. The entire assembly was clipped onto a coldfinger. The cuvette was filled with solution via capillary action immediately prior to submersion in liquid nitrogen to prevent the formation of air bubbles within the sample, which can short-circuit the cuvette. The angle (χ) between the applied electric field and the polarization of the incident light was varied by rotating the cuvette between 45° and 90° with respect to the incident beam. A lock-in amplifier (Stanford Research Systems SR830) was used to generate a sinusoidal AC electric field at a frequency ω = 3.5 kHz, which was amplified by a high voltage amplifier (TREK 609 × 10−6) to field strengths typically around 3 × 105 V/cm and applied to the sample via electrical leads connected to the cuvette with alligator clips. The high frequency of the applied field dramatically reduces 1/f noise, resulting in singlescan spectra with sufficient signal-to-noise for analysis. Phasesensitive detection was performed at the second harmonic of the field frequency. Since the field-modulated change in transmittance is negligible without lock-in detection, the zerofield light intensity can be obtained directly from the preamp output using an analog-to-digital converter (16-bit resolution). Low-temperature absorption spectra were recorded on the same instrumental setup, with an optical chopper operating at ∼6 kHz to modulate the unpolarized probe beam. The output of the photodiode was recorded at ω using the lock-in amplifier. Both reference (I0) and sample (I) intensities were recorded under the same conditions and absorption spectra were calculated according to the Beer−Lambert law, A = log I0/ I. Analysis of Stark Spectra. Stark spectra were analyzed according to the procedure of Liptay.11 For an immobilized, isotropically oriented sample, the Stark spectrum can be described as a linear combination of the zeroth, first, and

the chromophore will determine the rate of electron transfer to secondary substrates within a given device. In this study we have set out to verify this assignment and further characterize the excited state properties of the ICT transition of ABFL. We have used Stark spectroscopy7 to determine both the direction and magnitude of the difference dipole moment Δμ⃗ 01 between the ground and first excited state. This assignment was made with the aid of simulated Stark spectra calculated with TD-DFT and finite field methods. We have used femtosecond optical transient absorption spectroscopy to characterize the decay of the excited state within the first few nanoseconds after optical excitation. Efficient charge separation is a necessary first step in photovoltaic devices, but the charge-separated state must live long enough so that the electron−hole pair can be separated further to perform useful work. As will be shown, not all dyads that form charge separated states with high yield are useful in devices. However, understanding the detailed dynamics following charge separation is a critical step in designing more successful chromophores for solar energy conversion. On the other hand, rapid charge recombination in D-π-A chromophores with large hyperpolarizabilities can be a desirable feature of molecules used for NLO applications. Ultrafast deactivation of the charge separated state back to the ground electronic state can effectively compete with excited state chemical reactions, which may otherwise lead to photodegradation, limiting the lifetime of the NLO device. In addition, ABFL has been shown to self-assemble in aqueous solution, which may facilitate its use in nonlinear optical devices.8 Cyclic voltammetry measurements show that ABFL can be reversibly oxidized and reduced (in aprotic solvents). This redox-switching may also be a useful property to exploit from an applications standpoint.



EXPERIMENTAL SECTION Materials. ABFL was prepared as described previously. 2Methyltetrahydrofuran (MTHF), stabilized by 1% butylated hydroxytoluene, ethanol, toluene, and naphthalene were purchased from Sigma-Aldrich and used as received. Tetrabutylammonium hexafluorophosphate, dichloromethane, and ferrocene were purchased from Sigma. Steady State Spectroscopy. Ground state absorption spectra were measured on an HP 8452A photodiode array UV−vis spectrometer. Steady state emission and excitation spectra were measured on a photon technology international (PTI) fluorimeter. Time resolved emission spectra were measured with an N2 laser pumped dye laser as the excitation source (PTI), using 3 mM coumarin 540A in ethanol as the gain medium (Exciton). 15685

dx.doi.org/10.1021/jp406420h | J. Phys. Chem. B 2013, 117, 15684−15694

The Journal of Physical Chemistry B

Article

are focused into a 2 mm quartz cuvette at an angle of ∼3°. The pump beam diameter is ∼5 times the probe beam diameter in order to ensure that spatial overlap of the two beams is maintained as the delay stage is scanned. The polarization of the pump beam is set to the magic angle relative to the probe beam by use of a half-wave plate. After passing through the sample, the pump beam was blocked. The probe beam was focused into a 50 μm fused-silica optical fiber (Ocean optics), dispersed in a spectrograph (DK240) and detected with a CCD camera cooled to −15 °C (Andor). The slots of a 10-slot optical chopper blade were taped off with black electrical tape so that the probe beam could be chopped at 100 Hz and the pump at 50 Hz. This step was necessary since the regenerative amplifier is most stable when operating at 1 kHz, but the maximum readout rate of the CCD detector is 100 Hz. The difference absorption spectrum was calculated as the ratio in intensity of pairs of probe pulses from the unpumped and pumped sample, respectively. A stir bar was placed in the cuvette to ensure sample replacement between pump shots. Data was acquired for 400 shots at each time delay, and the full time interval (−20 ps to 3.5 ns in a logarithmic spacing) was scanned 20 times to obtain adequate signal-to-noise. The instrumental response function (IRF) was measured by the Kerr effect generated in a solution of saturated naphthalene in toluene. Lensing occurs only while the pump and probe beams are temporally overlapped in the solution, causing the WLC to refract away from the detector, resulting in a positive signal. The Kerr lens effect is greatest close to the pump wavelength, and is less efficient toward the red edge of the probe spectrum. Data for the IRF was collected between −2 and 2 ps, measured in 0.1 ps time steps. Transient Absorption Spectra Analysis. Transient absorption spectra were analyzed by the target analysis method for a photocycle according to van Stokkum, et al.,13 using MatLab software written in-house. The spectral chirp and temporal width of the IRF was determined by fitting a single Gaussian function to the measured Kerr lens effect at each wavelength. The maximum of the first derivative of each Gaussian, as a function of wavelength, was then fit with a second order polynomial to give the best estimate of the zero time for each wavelength, while the average width of the fitted gaussians was used as the initial guess of the IRF width. Initial guesses for the spectral chirp, IRF width, and kinetic rate constants for the target kinetic model were supplied to a nonlinear regression algorithm. The computed decay profiles of transient species were calculated as a convolution of the exponential decay matrix with the IRF at the appropriate zero time for each wavelength. This method avoids the possibility of introducing artifacts into the fit by interpolation of the data to a single time zero. The steady state absorption spectrum of the sample was also supplied, and the ground state bleach (GSB) calculated as the sum of the transient concentrations at each time point. This computed bleach was subtracted from the data prior to the calculation of the decay spectra. Decay profiles were fit to the corrected data matrix to calculate species-associated decay spectra (SADS). Since the ground state bleach signal has been removed, and in the absence of measurable fluorescence from ABFL, the SADS are the (model-dependent) absorption spectra of the transient species after photoexcitation. Finally, the decay profiles and the computed GSB and SADS were used to construct the simulated TA spectrum. The sum of residuals between the experimental

second derivatives of the unperturbed absorption spectrum, weighted by coefficients Aχ, Bχ and Cχ: ⎧ ε(ν)̃ Bχ d ⎛ ε(ν)̃ ⎞ Δε(ν)̃ + = (fc |F |⃗ )2 ⎨A χ ⎟ ⎜ ν̃ ν̃ 15ch dν ̃ ⎝ ν ̃ ⎠ ⎩ +

d2 ⎛ ε(ν)̃ ⎞⎫ ⎜ ⎟⎬ 30c 2h2 dν 2̃ ⎝ ν ̃ ⎠⎭ Cχ

(1)

where ε(ν̃) is the unperturbed extinction as a function of wavenumber, |F⃗| is the magnitude of the applied electric field in V/m, fc is the local field factor due to the attenuation of the applied electric field by the solvent cavity, h is Planck’s constant, and c is the speed of light. The Aχ coefficient is related to the transition polarizability and hyperpolarizability, but is generally negligible in frozen samples. The Bχ coefficient is → ⎯ → ⎯ → ⎯ α⎯ =→ α⎯ −→ α⎯ ) of the related to the difference polarizability (Δ→ 1

0

molecule, Bχ ≈

→ ⎯⎞ → ⎯ → ⎯ ⎛3 5 1 α⎯ + (3 cos2 χ − 1)⎜ m⃗ ·Δ→ α⎯ ·m⃗ − TrΔ→ α⎯ ⎟ TrΔ→ ⎝2 ⎠ 2 2 (2)

while the Cχ term is related to the difference dipole (Δμ⃗ = μ⃗ 1 − μ⃗ 0) and ζ, the angle between Δμ⃗ and m⃗ , Cχ = |Δμ ⃗ |2 {5 + (3 cos2 χ − 1)(3 cos2 ζ − 1)}

(3)

The low temperature absorption spectrum is first fitted to a sum of Gaussian functions to obtain an analytically differentiable model of the spectrum. This process is necessary since numerical derivatives of the experimental spectrum are generally quite noisy. Stark spectra measured at 4 or 5 different values of χ are fit simultaneously with the absorption spectrum. → ⎯ → ⎯ The parameters TrΔ→ α⎯ , m⃗ ·Δ→ α⎯ ·m⃗ , Δμ⃗, and ζ, along with the Gaussian parameters are supplied as variable parameters for the nonlinear fit. These electro-optical parameters are used to calculate the values of Bχ and Cχ for each angle χ according to eq 2 and 3, respectively. The Aχ initial guesses are also supplied, but must be introduced directly for each spectrum since the related physical properties cannot be deconvolved from these coefficients in this type of analysis. The simulated Stark spectrum is calculated according to eq 1 and fit simultaneously with the absorption spectrum. Goodness of fit is judged by an analysis of residuals and the invariance of the final electrooptical properties with respect to the initial choice of Gaussian function parameters. The fitting procedure was repeated for multiple data sets. Fitted parameters with their uncertainty are the average and standard deviation from these fits. The difference polarizabilities and difference dipole moments are corrected for the local field factor with fc = 1.6, calculated as described previously.12 Transient Absorption Spectrometer. The laser spectrometer has been described in detail previously.6 Briefly, the fundamental beam of a Ti:sapphire laser operating at 1 kHz is split into two arms. The ∼398 nm pump beam is generated by frequency doubling the fundamental with a type-I BBO crystal and attenuated to a power of approximately ∼700 nJ. The temporal position of the pump pulse relative to the probe is adjusted by a delay stage with a maximum delay of ∼4 ns. The white light continuum (WLC) probe pulse is generated by focusing a few microjoules of the 795 nm fundamental beam onto a CaF2 window. The collimated pump and probe beams 15686

dx.doi.org/10.1021/jp406420h | J. Phys. Chem. B 2013, 117, 15684−15694

The Journal of Physical Chemistry B

Article

possible. Additional environmental effects of solvent aromaticity or hydrogen bonding were also undesirable. Organic solvents suitable for Stark spectroscopy are limited, due to the requirement of optically clear glass formation at 77 K. MTHF was found to fit all of these criteria. The possibility of aggregation in MTHF was checked by measuring both Stark and low temperature absorption spectra at concentrations ranging from 30−500 μM (see Supporting Information (SI) Figure 1). The Stark spectra are especially sensitive to any changes in absorption features due to the derivative nature of the signal. Within the uncertainty of the measurement, all spectra in this concentration range were identical in amplitude. Additionally, the band shape of spectra normalized to the largest feature is not dependent on sample concentration. These observations suggest that ABFL exists primarily as the monomer under these conditions. The absorption spectra of the ICT transition of 100 μM ABFL at 298 and 77 K are shown in Figure 2a. The room

and simulated spectra were minimized with a built-in MatLab nonlinear regression algorithm (lsqnonlin). To estimate the errors in the fitted rate constants, a Monte Carlo simulation was run in which a normal distribution of 1000 initial guesses was generated, centered at the fitted parameter with a standard deviation of 50% of the fitted parameter. The fitting routine was run with each initial guess, holding the IRF parameters constant. Since there is no mathematical constraint on the shape or amplitude of the SADS, the previously calculated decay spectra were kept constant in this simulation and used to construct the simulated TA spectrum. White noise with an amplitude of that in the measured TA spectrum was artificially added to the decay spectra to more accurately simulate the experiment. MATLAB’s nlparci function was used to calculate the 95% confidence interval for each iteration. Fitted rate parameters are reported as the average result of this simulation. The estimated error is also reported as the mean 95% confidence interval of the simulation. The standard deviation of the 1000 simulated parameters is also reported. TD-DFT/Finite Field Calculations. For TD-DFT calculations, geometry-optimized 7,8,10-trimethyl isoalloxazine in the oxidized form14 was used as a starting structure and modified at the N(10) position to an isobutyl group and the C(8) position to an azo-diethylaniline group, respectively, using Chemcraft software (http://www.chemcraftprog.com). The initial geometry was optimized at AM1 and subsequently at B3LYP/3-21G (d,p) level of theory using C1 symmetry. The excited state energies, transition moments, and static dipole moments were calculated using TD-DFT at B3LYP/6-31G (d,p) level of theory. A polarizable continuum model (PCM) for chloroform developed by the Tomasi group15 was used to model the solvent dielectric response. This solvent model was chosen due to the similar dielectric constant of chloroform (4.9) compared with MTHF (7.6).16 Difference dipole moments, Δμ⃗ 0n, were calculated using the finite-field hexapole method17,18 using an external electric field of 0.001 au propagating along the ±X, ±Y, and ±Z directions in the center of mass coordinates, where Δμ⃗ 0n is defined as the vector difference between the static dipole moments μ⃗n and μ⃗ 0. Chemcraft was used to generate and visualize the excited state dipole moment vectors and HOMO/LUMO figures. All TDDFT and finite field calculations were performed using Gaussian 03, revision E.01.19 Cyclic Voltammetry. Cyclic voltammetry experiments were performed using a CH Instruments 440A electrochemical workstation. All measurements were carried under a nitrogen atmosphere. A solution of electrochemical grade tetrabutylammonium hexafluorophosphate (TBAPF6) dissolved in dry dichloromethane (DCM) (0.1 M) was employed as the supporting electrolyte. A platinum disc working electrode, a platinum wire counter electrode, and a silver wire pseudoreference electrode were used. The half-wave potentials are referenced to ferrocene (Fc) (internal reference) with the Fc/Fc+ redox couple adjusted to 0.0 V.

Figure 2. (a) Room-temperature and low-temperature UV−vis absorption spectrum of ABFL in MTHF. (b) Second derivative of LT absorption spectrum. (c) Stark spectra of 100 μM ABFL at various angles χ.

temperature spectrum shows a broad, featureless band centered at 550 nm. In contrast, the low temperature spectrum shows clear vibronic structure with a maximum at ∼600 nm, and spacing of ∼1000 cm−1 between vibronic features. The second derivative of the low temperature absorption spectrum is shown in Figure 2b. Since the room temperature spectrum is broad and featureless, its second derivative is negligible (not shown). The Stark spectra of ABFL taken at various angles of χ are shown in Figure 2c. The Stark spectrum is primarily second derivative in nature, and thus the difference dipole term is expected to dominate. The dependence of the Stark spectra on the angle χ shows that ζ must be smaller than the magic angle. The absorption spectrum was fitted to 9 Gaussian functions, and the corresponding derivatives were fitted simultaneously to Stark spectra at all values of χ measured for each sample. A representative fit to the Stark spectrum of 280 μM ABFL at χ ≈ 53° is shown in Figure 3a, along with the associated low-



RESULTS Low Temperature Absorption and Stark Spectra of the ABFL Monomer. The choice of solvent was crucial for obtaining spectra of the ABFL monomer. It was previously shown that ABFL is able to form molecular nanowires in water.8 While this process is not observed in organic solvents, the presence of dimers and other aggregate species may be 15687

dx.doi.org/10.1021/jp406420h | J. Phys. Chem. B 2013, 117, 15684−15694

The Journal of Physical Chemistry B

Article

The fitted difference dipole moment, Δμ⃗ was found to be 22.3 D, which is consistent with extensive charge redistribution upon excitation to the lowest electronic state. For comparison, the lowest (π−π*) transition of Fl in MTHF is only 1.6 D. The angle, ζ, between Δμ⃗ and the transition dipole moment, m⃗ , is 16°, indicating that charge redistribution occurs primarily along the direction of the transition dipole moment. TD-DFT calculations show that the lowest optically accessible transition of ABFL is predominantly of HOMO → LUMO character. These orbitals (Figure 4) are not localized on

Figure 3. (a) Comparison of experimental and fitted Stark spectrum. (b) Derivative components of the fit. (c) Comparison of experimental and fitted low temperature absorption spectrum.

temperature absorption spectrum in Figure 3c. While data was collected down to 300 nm, the presence of overlapping absorption bands to higher electronic states with weaker oscillator strengths complicates quantitative analysis. A single set of parameters produced an adequate fit over the spectral range from 500 to 750 nm, indicating the presence of a single electronic transition. The zeroth, first, and second derivative components of the fit are shown in Figure 3b. The Stark spectrum is dominated by the second derivative component, indicating a large Δμ⃗, while the zeroth derivative component is negligible. The first derivative component is much larger compared to underivatized flavin in organic solvents,20 indicating that the more extensively conjugated ABFL has a larger difference polarizability. Indeed both of these properties are expected for the ICT transition of D-π-A compounds. The fitting procedure was repeated for three data sets at 280 and 130 μM. The corresponding results were averaged to obtain error estimates for the fitted parameters, shown in Table 1.

Figure 4. Calculated HOMO and LUMO orbitals of ABFL.

either the donor or acceptor moieties, in agreement with the hypothesis that the conjugated azo bridge allows for extensive mixing of the electronic states of the flavin and aniline. ABFL, therefore, cannot be thought of as a simple dyad of two chromophores. Finite field calculations were used to construct simulated Stark spectra for ABFL in vacuum and chloroform. Figure 5a,b show the field induced shift of the ICT absorption band due to an external electric field applied along the long (X) and short (Y) axes of the plane of ABFL, respectively. The position of the absorption band is most dramatically perturbed by an electric field applied along the X axis, which corresponds roughly to the direction of the transition dipole moment. This observation agrees well with experimentally determined value of ζ. An electric field applied along the Y (Figure 5b) or Z (perpendicular to the plane of ABFL, data not shown) axes show only minimal shifts. The band shape of the simulated Stark spectrum (Figure 5c) are similar to the measured Stark spectrum of ABFL, while the simulated spectrum in chloroform more closely matches the spectral position of the features due to the red-shift expected upon solvation. TD-DFT calculations complement the analysis of Stark spectra if no experimental values for the ground state or transition dipole moments are available. Using the value of the ground state dipole moment in chloroform, which is closest in dielectric to MTHF, along with the experimental value for Δμ⃗ and ζ, the excited state dipole moment vector can be calculated (Table 2). There exist an infinite number of possible vectors that satisfy ζ = 16°; however, the optimized geometry of ABFL is shown to be planar, and thus the difference dipole is expected to lie within the plane of the molecule. This limits the direction

Table 1. Fitted Parameters for the Stark Spectra of ABFL in MTHF ( f = 1.6) → ⎯⎯ (Å3) TrΔ→ α

→ ⎯⎯ ·m (Å3) m⃗ ·Δ→ α ⃗

Δμ⃗ (D)

ζ (deg)

1259 ± 81

1149 ± 72

22.3 ± 0.9

16.2 ± 1.9

→ ⎯ α⎯ is an indication The trace of the polarizability tensor, TrΔ→ of the overall magnitude of the difference polarizability of the molecule. The value for ABFL is 2 orders of magnitude larger than that measured for N(10)-isobutyl-flavin (Fl) in MTHF.20 In addition, the projection of the polarizability onto the → ⎯ α⎯ , indicating transition dipole moment is comparable to TrΔ→

that the molecule is polarizable primarily along the direction of the transition dipole moment. 15688

dx.doi.org/10.1021/jp406420h | J. Phys. Chem. B 2013, 117, 15684−15694

The Journal of Physical Chemistry B

Article

upon photoexcitation to the lowest optically accessible electronic state. Transient Absorption of the ABFL Monomer. Experimental difference spectra at selected delay times are shown in Figure 7. A strong ground state bleach of the S0→S1 at 550 nm

Figure 7. Selected time points of the ABFL transient absorption spectrum.

is observed, while the expected ground state bleach of the S0→ S2 transition being pumped at 400 nm is not within the spectral range of the probe pulse. The minimum of the GSB is initially located at 550 nm, corresponding to the maximum of the ground state absorption spectrum and is slightly red-shifted at later time points. In addition, a strong ESA band appears to the red of the GSB feature located at ∼650 nm. This ESA feature decays completely within 20 ps after excitation. The GSB feature also decays rapidly within 20 ps, and then remains constant throughout the duration of the scan. Assuming no dissociation of the molecule upon excitation, the GSB signal gives a general description of the total concentration of transient species at any given time point. It is noted, however, that this is only true under the assumption that no overlapping positive ESA signal is present in the same spectral region. While this assumption is almost certainly not completely true, the GSB signal does provide a reasonable estimation of the total decay of the system to be used as a guide for rejecting physically unrealistic kinetic models. As an initial step to determining an appropriate model for use in the global target analysis, fitting was performed with a simple sequential kinetic model with increasing number of rate constants. Judging by the decrease in the sum of residuals, the minimum number of states was found to be three. While this model gave an adequate fit to the data, the resulting transient concentration decays were not consistent with the GSB decay (see SI Figures 2−3). In addition, incorporation of the ground state absorption spectrum into the model did not yield reasonable results for the SADS. A second type of model (shown schematically in Figure 8), consisting of a fast internal conversion (IC) step from S2 to S1, followed by a branched decay either from S1 directly to the ground state or to an additional state, which decays with a lifetime longer than the temporal range of the instrument. This model gives an excellent fit and agrees well with the interpretation that most ABFL decays to the ground state within 20 ps, while a small fraction remains in the photocycle for longer than 4 ns. The additional fast rate constant in this

Figure 5. Peak shifts calculated by finite field methods for an external electric field applied along the long axis (a) and short axis (b) of ABFL. (c) Simulated Stark spectra of ABFL in chloroform.

Table 2. Comparison of Calculated and Fitted Experimental Results theory (chloroform)

experiment (MTHF)

μ⃗0

18.73 D

-

μ⃗1

36.89 D

41.0 D

Δμ01 ⃗

18.16 D

22.3 D

ζ



16°

of Δμ⃗ to four possibilities. The choice of the direction of the difference dipole moment vector was made by comparison with the calculated result. The relevant vectors from both theory and experiment are summarized in Figure 6. This determination supports the hypothesis that electron density moves from the aniline donor to the pyrimidine moiety of the flavin acceptor

Figure 6. Comparison of the calculated (blue) and experimental (red) difference dipole moments of ABFL. The four possible vectors for the experimental result, based on the angle ζ and the transition dipole moment (green) are shown, with the assigned vector as a solid arrow, in agreement with TD-DFT/finite field calculations. 15689

dx.doi.org/10.1021/jp406420h | J. Phys. Chem. B 2013, 117, 15684−15694

The Journal of Physical Chemistry B

Article

Figure 8. Kinetic target for transient absorption analysis.

model is needed in order to transfer population into the longlived state from S1. The fitted kinetic parameters are summarized in Table 3. The error of the fit was determined

Figure 10. Calculated SADS of ABFL.

Finally, cyclic voltammetry measurements show reversible redox processes occur for ABFL in organic solvents (Figure 11). In particular, two reduction waves were observed, which

Table 3. Fitted Rate Constants from Transient Absorption

k1 k2 k3 k4

rate constant (ps‑1)

standard deviation (Monte Carlo)

95% confidence interval (mean)

lifetime (ps)

1.9177 0.1464 0.0492 4 ns

using a Monte Carlo procedure (see SI Figure 4). The simulated concentration decay curves are shown in Figure 9.

Figure 11. Cyclic voltammogram of ABFL recorded in DCM (0.5 mM). Scan rate = 0.1 V s−1.

Figure 9. Calculated transient decay profiles of ABFL. The magnitude of the experimental ground state bleach is plotted as open circles, and the total transient concentration as the solid black line. The temporal evolution of states A, B, and C are shown. The time axis is linear for the first 20 ps and logarithmic for later times.

we attribute to the formation of the radical anion states of the azo21 and flavin22 moieties, and an oxidation wave which we attribute to the azo group.21 These oxidized and reduced forms of ABFL will likely have dramatically different optical properties from the charge-separated (yet formally neutral) state of ABFL investigated in the current work. This opens up further possibilities for use of ABFL-like chromophores in redox switchable molecular electronic devices.

The circles show the magnitude of the GSB signal, while the red line shows the overall fit at 550 nm. The fitted IRF width was found to be ∼270 fs. The decay components of the three states are also shown. The corresponding SADS are shown in Figure 10. The spectrum of state A shows a strong absorption peaked at 675 nm and a weaker band centered at 460 nm. The spectrum of state B is similar. Here, the red band has decreased to about half intensity and blue-shifted to ∼650 nm. The blue band has redshifted to ∼500 nm. There is only one broad spectral feature for the state C, spanning 600−450 nm. According to this model, the quantum yield of formation of state C is 25%.

DISCUSSION Analysis of the Stark spectra of the ABFL monomer in MTHF revealed that the lowest optically accessible transition results in difference dipole moment of 22.3 D, indicative of a charge separated excited state. This value corresponds to charge separation of roughly 4.5 Å (1 Åe = 4.8 D). The direction of charge redistribution was found to be along the long axis of the molecule, with electron density moving from the diethylaniline to the flavin. The trace of the difference polarizability was found to be 1259 Å3. These results show a dramatic increase in excited state charge redistribution compared to simple flavin. Our previous work on oxidized N(10)-isobutyl-flavin in MTHF found that the S0→S1 transition has a difference dipole of 1.6 D



15690

dx.doi.org/10.1021/jp406420h | J. Phys. Chem. B 2013, 117, 15684−15694

The Journal of Physical Chemistry B

Article

and ζ = 58°, while the S0→S2 transition has a much larger difference dipole moment of 4.9 D and ζ = 30°. The difference polarizabilities were found to be 9 and 63 Å3, respectively. The large increase in polarizability of ABFL compared to flavin is attributed to the extended conjugated system of the D-π-A chromophore. Push−pull molecules are also known to exhibit large hyperpolarizabilities, β, a desirable property for NLO devices.1,23 Under the two-state approximation, the dominant nonresonant component of the β0 tensor can be related to the electronic properties of the molecule as24−26

β0 = 6Δμge⃗

conversion to the ICT state, with charge separation (CS) taking place with a lifetime of 500 fs. Decay of this state occurs through two pathways. The fast decay pathway is hypothesized to be direct charge recombination (CR) resulting in repopulation of the ground state with a lifetime of 5 ps. The fast CR rate suggests that ABFL would not be suitable in photovoltaic applications since the electron−hole pair does not live long enough to perform useful work. However, the existence of the alternative pathway to a long-lived state suggests a possible route of optimization to achieving useful ABFL-like chromophores for this purpose. The identity of the longer lived (>4 ns) transient species is more ambiguous. The possibility of excited triplet formation was explored through steady state and time-resolved emission spectroscopies. The steady state emission spectrum shows an extremely small signal at ∼680 nm, while the nanosecond timeresolved emission spectrum of ABFL shows a decay component on the order of 15 ns and is therefore assigned to fluorescence from an excited singlet state (see SI Figure 5). This emission is extremely weak, however, and is below the detection limit of our TA Instruments. Additionally, increased emission from a deoxygenated sample or at cryogenic temperature was not observed. While this does not rule out the possibility of triplet formation, at this time this hypothesis cannot be substantiated. The most plausible explanation for the long-lived state C is an alternative conformation of the molecule. The majority of ABFL is predicted to be in the trans conformation by the TDDFT optimized ground state geometry. For many azobenzene derivatives, a photoinduced trans→cis isomerization is experimentally observable by a decrease in absorbance upon irradiation of the sample.31 Subsequent thermal cis→trans isomerization is marked by the gradual increase of absorption of the course of seconds to minutes. This behavior is not observed for ABFL, however, which may be due to an increased rate of cis→trans thermal isomerization relative to many of the azobenzene derivatives studied to date. Temps, et al., have studied the push−pull azobenzene derivative Disperse Red 1, which also does not show the typical bleaching characteristics due to isomerization.32 This process is attributed to a lowered energy barrier between the two isomers in the electronic ground state, causing back isomerization to the trans isomer on the order of milliseconds to seconds. The TA spectrum of this chromophore also shows a long-lived (ns) bleach, which was attributed to formation of a cis product. In order to determine whether multiple conformations are present, the TA spectrum of ABFL was fit with kinetic models involving initial populations of multiple conformations. Such conformational subpopulations have been used to explain excited state lifetimes with orders of magnitude differences, such as the “stacked/unstacked” conformational distribution of FAD in aqueous solution.33−35 This attempt to isolate separate “fast” and “slow” decay pathways for different isomers of ABFL did not produce physically realistic rate constants for SADS and were thus rejected. Again, these results do not rule out the possibility of multiple conformations of ABFL with different decay kinetics; however, the TA data alone cannot support this. The excited state dynamics of several donor−acceptor flavin dyads containing nonconjugated linkages at the N(3) or N(10) positions have been investigated previously by ultrafast spectroscopies. The lack of conjugation between the donor and acceptor chromophores prevents mixing of their electronic states, and the UV−vis spectra of these dyads are generally a linear combination of the spectra of the two components. The

m⃗ ge2 2 Ege

(4)

where Ege is the (mean) energy of the transition and m⃗ ge is approximated from the integrated area of the absorption band. Using this expression and Δμ⃗ 01 obtained from analysis of the Stark data for the ICT band of ABFL β0 was found to be 720 × 10−30 esu. For comparison, the β0 value for retinal, the chromophore in rhodopsins and the benchmark for biologically inspired NLO molecules, varies from ∼150−300 × 10−30 esu in organic solvents depending on the solvent dielectric.27 The hyperpolarizability of ABFL is also comparable with other push−pull polyenes that have been studied with β0 ranging from ∼20−750 × 10−30 esu.28 Previous computational studies have suggested a method of optimizing β0, which combined with these results, can be applied toward the synthesis of new ABFL derivatives to give tunable “designer” push−pull flavins.24 Interestingly, this simple flavin-dyad has as large a β0 as the more sophisticated D-porphyrin-A construct from the Therein group ([5-[[4′-(dimethylamino)phenyl]ethynyl]-15-[(4″-nitrophenyl)-ethynyl]-10,20-diphenylporphinato]zinc(II)).25 In that case, the donor is a dimethylaniline moiety, the azo-linker is replaced by a porphyrin molecule, and a nitrophenyl group is the acceptor. A significant difference between these two systems is the nature of the linker molecule and its effect on the electronic coupling between donor and acceptor. In the porphyrin example, the absorption spectrum is similar to an unperturbed porphyrin. By contrast, the strongest absorption band in ABFL is an ICT band, and no apparent features from the individual donor and acceptor moieties are evident. This suggests that the ABFL dyad has significantly higher electronic coupling, leading to enhanced charge separation (Δμ = 22 D vs 15 D), and therefore an enhanced β0. The nature of the flavin as acceptor may also play an important role. Substitution at the C(8) position of the flavin often leads to large changes in its electronic structure. For example, the transition dipole moment of 8-NH2-flavin is calculated to be significantly larger than for lumiflavin.29 8-Mercaptoflavin has an intense red-shifted absorption band,30 reminiscent of ABFL. Thus, orbital overlap between the flavin and azo linker may be significantly larger than between a porphyrin and acetylenic groups. This suggests that flavin-based systems may have significant NLO properties that have yet to be explored. The femtosecond transient absorption spectrum revealed dynamics of the excited state after excitation at 400 nm. At this wavelength, the initial state being pumped is the π−π* S2 state, which is shown by TD-DFT calculations to be localized on the flavin moiety. A branched kinetic model was necessary to obtain physically realistic concentration decay profiles in agreement with the observed ground state bleach of the system. The initially populated π−π* state undergoes internal 15691

dx.doi.org/10.1021/jp406420h | J. Phys. Chem. B 2013, 117, 15684−15694

The Journal of Physical Chemistry B

Article

× 10−30 esu, which suggests that modified flavins may have useful NLO properties. The excited state decay dynamics of ABFL was measured by femtosecond transient absorption spectroscopy. The simplest model that best fits the data suggests three transient states. The initial subpicosecond component corresponds to internal conversion from the pumped π−π* state to the ICT state. This state decays via two channels: charge recombination on the order of 7 ps, or formation of a long-lived >4 ns state with a quantum yield of approximately 25%. By comparison with other donor-flavin dyads, the extended conjugated system of ABFL leads to increased extinction, large red-shifts in the absorption band position, and dramatically increased charge recombination kinetics. These results indicate that while the initial photoinduced charge separation of ABFL is adequate for applications in solar energy conversion, the charge recombination process occurs on a time scale that prevents subsequent longer range charge separation to perform useful work. This variety of interesting excited state properties may be adjusted through perturbation of the donor moiety or addition of ligands that take advantage of the molecular recognition capabilities of the isoalloxazine ring. We therefore believe that ABFL represents an important backbone structure for building many chromophores that will have suitable electronic properties for photovoltaic and redox-switching applications in addition to robust and highly tunable NLO properties that should find use in a variety of molecular electronic applications.

conjugated azo bridge between the donor and acceptor moieties of ABFL does, however, allow for efficient mixing of the molecular orbitals of the two chromophores, resulting in electronic states and corresponding absorption spectra completely different from the parent molecules. The extended π system of ABFL also leads to significantly different CR dynamics compared to other donor−flavin dyads. For instance, a flavin−oligothiophene dyad, in which the donor was attached at the N(3) position, shows that charge recombination occurs in 435 ps.6 Fluorescence upconversion measurements were used to determine the charge-separated lifetime of a pyrene−flavin dyad to be 150 ps.36 These flavin dyads also exhibit quenched fluorescence compared to the parent molecules. This has been interpreted as evidence of electron transfer to a nonfluorescent charge separated state. The weak emission of ABFL may be due to the formation of a largely delocalized nonfluorescent charge separated state; however, efficient charge recombination from the ICT state leads to rapid repopulation of the ground state, effectively outcompeting the emission process. An aniline−flavin dyad, similar to ABFL, was characterized by Fukuzumi et al. and showed dramatically different excited state dynamics.37 This dyad, in which a dimethylaniline donor was attached at the N(10) position of flavin by a single C−N bond (DMA-Fl), exhibits a charge separated lifetime of an astounding 2.1 ms. The transient absorption spectrum of DMAFl shows an ESA feature peaked at 680 nm, which was assigned to the absorption of a dimethylaniline radical cation, suggesting a single electron transfer from the aniline to the flavin. The appearance of radical ion spectra of the two chromophores is not necessarily expected for ABFL, however, due to the conjugation between donor and acceptor. The differing extent of electronic delocalization in ABFL and Fl-DMA also leads to differences in geometry between these dyads. The optimized geometry of ABFL is completely planar, which allows for overlap of the π orbitals along the conjugated path. However, in DMA-Fl, the plane of the dimethylaniline ring is found to be perpendicular to that of the isoalloxazine. This lack of conjugation also leads to differences in the electronic structure. The lack of conjugation between donor and acceptor in DMA-Fl is seen in the localization of electron density in the HOMO and LUMO states on the dimethylaniline and flavin, respectively.



ASSOCIATED CONTENT

* Supporting Information S

Monte Carlo error estimation, concentration dependence of Stark spectra, alternative kinetic target models, and timeresolved emission. This material is available free of charge via the Internet at http://pubs.acs.org.



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Present Addresses ∥

Department of Biochemistry and Biophysics, University of Pennsylvania School of Medicine, 1004 Stellar-Chance Building, 422 Curie Boulevard, Philadelphia, Pennsylvania 19104-6059, United States. ⊥ Engineering of Fuel and Energy Department, Technical College, Kirkuk Foundation of Technical Education, Kirkuk, Iraq.



CONCLUSIONS The charge redistribution of ABFL upon vertical excitation from the ground to the lowest excited state was studied by Stark spectroscopy. The large difference dipole moment of 22.3 D agrees well with the hypothesis that the lowest excited state of this D-π-A flavin derivative is an intramolecular charge transfer state. In addition, the small value of ζ indicates that the charge redistribution occurs roughly along the same direction as the transition dipole moment, which was determined by TDDFT calculations to be along the long axis of the molecule. Finite field calculations are in close agreement with both the experimentally determined magnitude and direction of the difference dipole moment. This evidence supports the hypothesis that charge transfer occurs from the diethylaniline donor to the flavin acceptor. The measured value of the difference polarizability is 2 orders of magnitude higher than that of flavin alone, as expected for this type of push−pull molecular system. This corresponds to a large value of β0 = 720

Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS R.F.P. and R.J.S. were supported in part by NSF Grant CHE0847855. V.M.R. and G.C. acknowledge NSF Grant CHE1025889. G.C. thanks the EPSRC. N.Y.Z. thanks MOHESR for a scholarship. R.P. and R.S. thank Dr. C. J. Martoff for use of his time-resolved fluorimeter. We wish to thank Prof. David Beratan for useful discussions.



ABBREVIATIONS ABFL, azobenzylflavin; CCD, charge coupled device; CS, charge separation; CR, charge recombination; CV, cyclic voltammetry; D-π-A, donor-π-acceptor; DCM, dichloromethane; DMA-Fl, dimethylaniline-flavin; ESA, excited state 15692

dx.doi.org/10.1021/jp406420h | J. Phys. Chem. B 2013, 117, 15684−15694

The Journal of Physical Chemistry B

Article

absorption; FAD, flavin adenine dinucleotide; Fc, ferrocene; Fl, N(10)-isobutylflavin; FMN, flavin mononucleotide; GSB, ground state bleach; HOMO, highest occupied molecular orbital; IC, internal conversion; ICT, intramolecular charge transfer; IRF, instrument response function; ISC, intersystem crossing; LUMO, lowest unoccupied molecular orbital; MTHF, 2-methyltetrahydrofuran; NLO, nonlinear optics; SADS, species associated difference spectrum; TA, transient absorption; (TD)-DFT, (time dependent) density functional theory; WLC, white light continuum



(17) Locknar, S. A.; Peteanu, L. A.; Shuai, Z. Calculation of Ground and Excited State Polarizabilities of Unsubstituted and Donor/ Acceptor Polyenes: A Comparison of the Finite-Field and SumOver-States Methods. J. Phys. Chem. A 1999, 103 (14), 2197−2201. (18) Cave, R. J.; Burke, K.; Castner, E. W., Jr. Theoretical Investigation of the Ground and Excited States of Coumarin 151 and Coumarin 120. J. Phys. Chem. A 2002, 106 (40), 9294−9305. (19) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Montgomery, J. J. A.; Vreven, T.; Kudin, K. N.; Burant, J. C.; et al. Gaussian 03, revision E.012004; Gaussian, Inc.: Wallingford, CT, 2004. (20) Stanley, R. J.; Siddiqui, M. S. A Stark Spectroscopic Study of N(3)-Methyl, N(10)-Isobutyl-7,8-Dimethylisoalloxazine in Nonpolar Low-Temperature Glasses: Experiment and Comparison with Calculations. J. Phys. Chem. A 2001, 105 (49), 11001−11008. (21) Chen, Y.; Harrison, W. T. A.; Imrie, C. T.; Ryder, K. S. Pyrrole and polypyrrole-based liquid crystals containing azobenzene mesogenic groups. J. Mater. Chem. 2002, 12 (3), 579−585. (22) Niemz, A.; Imbriglio, J.; Rotello, V. M. Model Systems for Flavoenzyme Activity: One- and Two-Electron Reduction of Flavins in Aprotic Hydrophobic Environments. J. Am. Chem. Soc. 1997, 119 (5), 887−892. (23) Blanchard-Desce, M.; Ledoux, I.; Lehn, J. M.; Malthete, J.; Zyss, J. Push−Pull Polyenes and Carotenoids - Synthesis and Non-Linear Optical-Properties. J. Chem. Soc., Chem. Commun. 1988, 11, 737−739. (24) Marder, S. R.; Beratan, D. N.; Cheng, L.-T. Approaches for Optimizing the First Electronic Hyperpolarizability of Conjugated Organic Molecules. Science 1991, 252 (5002), 103−106. (25) Karki, L.; Vance, F. W.; Hupp, J. T.; LeCours, S. M.; Therien, M. J. Electronic Stark Effect Studies of a Porphyrin-Based Push−Pull Chromophore Displaying a Large First Hyperpolarizability: StateSpecific Contributions to β. J. Am. Chem. Soc. 1998, 120 (11), 2606− 2611. (26) Willetts, A.; Rice, J. E.; Burland, D. M.; Shelton, D. P. Problems in the Comparison of Theoretical and Experimenal Hyperpolarizabilities. J. Chem. Phys. 1992, 97 (10), 7590−7599. (27) Lewis, A.; Khatchatouriants, A.; Treinin, M.; Chen, Z. P.; Peleg, G.; Friedman, N.; Bouevitch, O.; Rothman, Z.; Loew, L.; Sheres, M. Second-harmonic generation of biological interfaces: Probing the membrane protein bacteriorhodopsin and imaging membrane potential around GFP molecules at specific sites in neuronal cells of C. elegans. Chem. Phys. 1999, 245 (1−3), 133−144. (28) Blanchard-Desce, M.; Alain, V.; Midrier, L.; Wortmann, R.; Lebus, S.; Glania, C.; Kramer, P.; Fort, A.; Muller, J.; Barzoukas, M. Intramolecular charge transfer and enhanced quadratic optical nonlinearities in push−pull polyenes. J. Photochem. Photobiol., A 1997, 105, 115−121. (29) Choe, Y.-K.; Nagase, S.; Nishimoto, K. Theoretical study of the electronic spectra of oxidized and reduced states of lumiflavin and its derivative. J. Comput. Chem. 2007, 28 (4), 727−739. (30) Massey, V.; Ghisla, S.; Moore, E. G. 8-Mercaptoflavins as ActiveSite Probes of Flavoenzymes. J. Biol. Chem. 1979, 254 (19), 9640− 9650. (31) Bandara, H. M. D.; Burdette, S. C. Photoisomerization in different classes of azobenzene. Chem. Soc. Rev. 2012, 41 (5), 1809− 1825. (32) Bahrenburg, J.; Rottger, K.; Siewertsen, R.; Renth, F.; Temps, F. Sequential photoisomerisation dynamics of the push−pull azobenzene Disperse Red 1. Photochem. Photobiol. Sci. 2012, 11 (7), 1210−1219. (33) Stanley, R. J.; MacFarlane, A. W., IV Ultrafast Excited State Dynamics of Oxidized Flavins: Direct Observations of Quenching by Purines. J. Phys. Chem. A 2000, 104 (30), 6899−6906. (34) Radoszkowicz, L.; Huppert, D.; Nachliel, E.; Gutman, M. Sampling the Conformation Space of FAD in Water−Methanol Mixtures through Molecular Dynamics and Fluorescence Measurements. J. Phys. Chem. A 2010, 114 (2), 1017−1022. (35) Brazard, J.; Usman, A.; Lacombat, F.; Ley, C.; Martin, M. M.; Plaza, P. New Insights into the Ultrafast Photophysics of Oxidized and

REFERENCES

(1) Marder, S. R., Sohn, J. E., Stucky, G. D., Eds. Materials for Nonlinear Optics: Chemical Perspectives; ACS Symposium Series; American Chemical Society: Washington, DC, 1991; Vol. 455. (2) Nguyen, P.; Lesley, G.; Marder, T. B.; Ledoux, I.; Zyss, J. Secondorder nonlinear optical properties of push−pull bis(phenylethynyl)benzenes and unsymmetric platinum bis(phenylacetylide) complexes. Chem. Mater. 1997, 9 (2), 406−408. (3) Katti, K. V.; Raghuraman, K.; Pillarsetty, N.; Karra, S. R.; Gulotty, R. J.; Chartier, M. A.; Langhoff, C. A. First examples of azaphosphanes as efficient electron donors in the chemical architecture of thermally stable new nonlinear optical materials. Chem. Mater. 2002, 14 (6), 2436−2438. (4) Bublitz, G. U.; Ortiz, R.; Marder, S. R.; Boxer, S. G. Stark Spectroscopy of Donor/Acceptor Substituted Polyenes. J. Am. Chem. Soc. 1997, 119 (14), 3365−3376. (5) Jordan, B. J.; Pollier, M. A.; Ofir, Y.; Joubanian, S.; Mehtala, J. G.; Sinkel, C.; Caldwell, S. T.; Kennedy, A.; Rabani, G.; Cooke, G.; Rotello, V. M. ‘Lock and key’ control of optical properties in a push− pull system. Chem. Commun. 2008, 14, 1653−1655. (6) Yu, X.; Eymur, S.; Singh, V.; Yang, B. Q.; Tonga, M.; Bheemaraju, A.; Cooke, G.; Subramani, C.; Venkataraman, D.; Stanley, R. J.; Rotello, V. M. Flavin as a photo-active acceptor for efficient energy and charge transfer in a model donor−acceptor system. Phys. Chem. Chem. Phys. 2012, 14 (19), 6749−6754. (7) Boxer, S. G. Stark Realities. J. Phys. Chem. B 2009, 113 (10), 2972−2983. (8) Jordan, B. J.; Ofir, Y.; Patra, D.; Caldwell, S. T.; Kennedy, A.; Joubanian, S.; Rabani, G.; Cooke, G.; Rotello, V. M. Controlled SelfAssembly of Organic Nanowires and Platelets Using Dipolar and Hydrogen-Bonding Interactions. Small 2008, 4 (11), 2074−2078. (9) Kodali, G.; Stanley, R. J. Excited State Chemistry of Flavoproteins. In Handbook of Flavins and Flavoproteins; Hille, R., Miller, S., Palfey, B., Eds.; de Gruyter: Berlin, 2013. (10) Andrews, S. S.; Boxer, S. G. A liquid nitrogen immersion cryostat for optical measurements. Rev. Sci. Instrum. 2000, 71 (9), 3567−3569. (11) Liptay, W. Dipole Moments and Polarizabilities of Molecules in Excited Electronic States. In Excited States; Lim, E. C., Ed.; Academic Press, Inc.: New York, 1974; Vol. 1, pp 129−229. (12) Premvardhan, L.; Peteanu, L. Electroabsorption measurements and ab initio calculations of the dipolar properties of 2-(2′hydroxyphenyl)-benzothiazole and -benzoxazole: Two photostabilizers that undergo excited-state proton transfer. Chem. Phys. Lett. 1998, 296, 521−529. (13) van Stokkum, I. H. M.; Larsen, D. S.; van Grondelle, R. Global and target analysis of time-resolved spectra. Biochim. Biophys. Acta, Bioenerg. 2004, 1657 (2−3), 82−104. (14) Kodali, G.; Siddiqui, S. U.; Stanley, R. J. Charge Redistribution in Oxidized and Semiquinone E. coli DNA Photolyase upon Photoexcitation: Stark Spectroscopy Reveals a Rationale for the Position of Trp382. J. Am. Chem. Soc. 2009, 131 (13), 4795−4807. (15) Cossi, M.; Barone, V.; Cammi, R.; Tomasi, J. Ab initio study of solvated molecules: A new implementation of the polarizable continuum model. Chem. Phys. Lett. 1996, 255 (4−6), 327−335. (16) Reichardt, C. Solvents and Solvent Effects in Organic Chemistry; VCH Verlagsgesellschaft mbH: Weinheim, Germany, 1988. 15693

dx.doi.org/10.1021/jp406420h | J. Phys. Chem. B 2013, 117, 15684−15694

The Journal of Physical Chemistry B

Article

Reduced FAD in Solution. J. Phys. Chem. A 2011, 115 (15), 3251− 3262. (36) Shirdel, J.; Penzkofer, A.; Prochazka, R.; Shen, Z.; Strauss, J.; Daub, J. Absorption and emission spectroscopic characterisation of a pyrene−flavin dyad. Chem. Phys. 2007, 331 (2−3), 427−437. (37) Murakami, M.; Ohkubo, K.; Fukuzumi, S. Inter- and Intramolecular Photoinduced Electron Transfer of Flavin Derivatives with Extremely Small Reorganization Energies. Chem.Eur. J. 2010, 16 (26), 7820−7832.

15694

dx.doi.org/10.1021/jp406420h | J. Phys. Chem. B 2013, 117, 15684−15694