Article pubs.acs.org/JPCA
Femtosecond and Temperature-Dependent Picosecond Dynamics of Ultrafast Excited-State Proton Transfer in Water−Dioxane Mixtures Adilson A. Freitas,*,† Frank H. Quina,*,‡ and António A. L. Maçanita† †
Centro de Química Estrutural, Instituto Superior Técnico, Universidade de Lisboa, 1649-004 Lisbon, Portugal Instituto de Química, Universidade de São Paulo, CP 26077, São Paulo 05513-970, Brazil
‡
S Supporting Information *
ABSTRACT: Synthetic flavylium salts like the 7-hydroxy-4-methylflavylium (HMF) cation have been used as prototypes to study the chemistry and photochemistry of anthocyanins, the major group of water-soluble pigments in the plant kingdom. In this work, a combination of fluorescence upconversion with femtosecond time resolution and time-correlated single photon counting (TCSPC) with picosecond time resolution have been employed to investigate in details the excited-state proton transfer (ESPT) of HMF in water and in binary water/1,4-dioxane mixtures. TCSPC measurements as a function of temperature provide activation parameters for all of the individual rate constants involved in the proton transfer, including those for dissociation and recombination of the geminate excited base−proton pair (A*···H+) that can be detected in the water/dioxane mixtures (but not in water). Unlike the other rate constants, the deprotonation rate constant kd shows a non-Arrhenius dependence on temperature in both water and water/dioxane mixtures. At low temperatures kd is close to the dielectric relaxation rate of the solvent with a barrier of ca. 8 kJ mol−1, suggesting that the solvent reorganization is the rate-limiting step. At higher temperatures (>30 °C) the proton transfer process is nearly barrierless and solvent-dependent. Fluorescence upconversion results in H2O, D2O, and water/dioxane mixtures confirm the two-step model for the ESPT of HMF and provide additional details of the early events prior to the onset of proton transfer, attributed to conformational relaxation and solvent reaccommodation around the initially formed excited state. The results are consistent with DFT calculations that indicate that charge redistribution occurs after rather than prior to the onset of the ESPT process.
■
INTRODUCTION Anthocyanins are the major class of water-soluble pigments found in nature.1−7 The flavylium cation (AH+) is the stable form of anthocyanins in solution at pH < 2 (Scheme S1 in Supporting Information) and the lowest excited singlet state exhibits the properties of a superphotoacid (pKa* < 0). At higher pH values, AH+ is in prototropic equilibrium with the quinonoidal base (A) and both species have absorption maxima in the spectral region of 400−600 nm, complementary to the absorption spectra of chlorophylls. Nucleophilic attack by a water molecule at position C-2 of the flavylium cation form of natural anthocyanins forms the hemiketal (B),8 which is in tautomeric equilibrium with the open form, the Z-chalcone (CZ) and can in turn isomerize to the corresponding Echalcone (CE). The chemical transformations of anthocyanins in the excited state are as diverse as in ground state and include adiabatic excited-state proton transfer (ESPT) between AH+* and A*, photoinduced ring-opening of B to give CZ, and photoisomerization of CE and CZ. Anthocyanins are believed to play an important role in the photoprotection of plants against an excess of deleterious solar radiation.9,10 The photoprotection mechanisms include ESPT from AH+* to water11,12 in uncomplexed anthocyanins and even more highly efficient internal conversion to the ground state of anthocyanins complexed with colorless organic molecules (copigments).4,10 In this work we present the results of a study of the ESPT of the 7-hydroxy-4-methylflavylium (HMF) cation, a synthetic © XXXX American Chemical Society
analogue possessing the basic 7-hydroxyflavylium cation chromophore present in natural anthocyanins. Time-correlated single photon counting (TCSPC) measurements with picosecond time resolution show that the ESPT process (Scheme 1) of AH+* in water11,12 gives the free base A*, whereas in water/dioxane mixtures13 the geminate pair intermediate (A*··· H+) can be detected. However, the time scale of these Scheme 1. ESPT Reaction of the 7-Hydroxy-4methylflavylium (HMF) Cation in Water/1,4-Dioxane and Micellar Solutions
Special Issue: Current Topics in Photochemistry Received: April 29, 2014 Revised: June 2, 2014
A
dx.doi.org/10.1021/jp504189m | J. Phys. Chem. A XXXX, XXX, XXX−XXX
The Journal of Physical Chemistry A
Article
the maximum per cycle) until a total of 5 × 103 counts had been collected at the maximum. The instrument response function (IRF) measured at 814 fs/channel was 17−19 ps (fwhm). Individual and global analyses of the fluorescence decays were carried out with the Sand program21 by fitting the experimental data with a multiexponential function convoluted with the IRF. The goodness of fit was judged by the analysis of the reduced chi square test, autocorrelation functions and weighted residuals. To fit the femtosecond fluorescence decays, the long picosecond decay constants were fixed. Computational Methodology. All calculations were performed with the Gaussian 03 package.22 Geometry optimizations of HMF in vacuum were performed without any constraint with the mPW1PW91 hybrid functional23 using the 6-31+G(d,p) basis set for the ground state and the TDmPW1PW91/6-31+G(d,p) level for the first excited singlet state. The implicit solvent for calculations in water was described by the Integral Equation Formalism for the Polarizable Continuum Model,24 IEFPCM, and the optimized geometries were obtained at the mPW1PW91/6-31G(d) level for the ground state and at the TD-mPW1PW91/6-31G(d) level for the excited state. The united atom topological model,25 UA0, was used to build the molecular cavity. Harmonic frequency calculations confirmed that all stationary points found were minima on the electronic potential energy surface (no imaginary frequencies present). The selection of the mPW1PW91 hybrid functional was based on our previous study26 showing that the electronic transitions of flavylium cations and quinonoidal bases calculated with this functional were in better agreement with experimental results than those carried out with B3LYP. The atomic charges of the flavylium cation in the ground and lowest excited singlet states were obtained by natural population analysis27 (NPA) and the CHELPG procedure.28 Both atomic charge calculation schemes were carried out employing the geometries optimized in the condensed phase at the mPW1PW91/6-31G(d) level and in the gas phase at the mPW1PW91/6-311+G(2d,2p) level of theory.
measurements does not provide information on the subpicosecond events involved in the proton transfer step itself. A common feature of models14−18 for the short-time events involved in proton transfer to water is the description of the proton displacement in terms of a two-step cooperative process. The first step is the reorientation of a water molecule so as to direct its unshared electron pair toward the proton to be transferred, with a small activation barrier. The second step is then the nearly barrierless translational movement of the proton to this neighboring oriented water molecule capable of accommodating the change in coordination number. Fluorescence upconversion measurements with femtosecond time resolution show that the early events of the proton transfer can indeed be attributed to a two-step process in both water and water/1,4-dioxane mixtures (with the initial step occurring on the time scale of ca. 1 ps). These upconversion results, together with the temperature dependences of the individual kinetic steps in Scheme 1 determined by TCSPC, provide a consistent picture of the ESPT process of anthocyanins.
■
EXPERIMENTAL SECTION Materials and Sample Preparation. The syntheses of 7hydroxy-4-methylflavylium chloride (HMF) and 7-methoxy-4methylflavylium chloride (MMF) have been previously reported.11 The solvent 1,4-dioxane (minimum 99.5%, Riedelde Haën) was used as received. The pH adjustments were made by appropriate dilutions of NaOH (p.a., Pronalab) and HClO4 (60% concentrated, p.a., Merck) solutions. The concentration of HClO4 was fixed at 0.02 mol L−1 in the studies in which it was necessary to ensure that HMF was protonated in the ground state. The water utilized for the preparation of all solutions was deionized in an Elgastat UHQPS system (Elga). The pH of the solutions was measured at 293 K using a Crison BasiC 20 pH meter combined with a Mettler Toledo InLab 423 Ag/AgCl microelectrode. Measurements. Fluorescence spectra were measured on a SPEX Fluorolog 2. The UV−vis absorption spectra were collected on a Beckman DU-70 spectrophotometer. The concentration of HMF was typically 1 × 10−5 mol L−1. The fluorescence decays were measured by upconversion and timecorrelated single photon counting (TCSPC) techniques. The femtosecond upconversion setup and measurement procedures have been described elsewhere.19 The polarization of the probe beam was set to the magic angle (54.7°) with respect to the gate pulse by a Berek’s variable wave plate and focused on a 1 mm thick rotating quartz cell. A cutoff filter (50% transmittance at 455 nm) was used to remove the excitation light from the fluorescence emitted by the sample and the upconversion signal was generated by focusing the fluorescence beam and the gating beam into a second BBO crystal (0.5 mm thick) with type I phase matching conditions. The upconverted signal was filtered by a band-pass filter (250−400 nm), dispersed in a double monocromator (CDP 2202D) and detected by a photomultiplier tube, all interfaced and controlled by a computer. The upconversion signal was accumulated for 0.5 s at each delay step of the translational stage of the gating pulse (step size of 50 fs). The cross-correlation of the excitation and gating pulses was obtained using the Raman scattering from water. The TCSPC setup utilizing a Millennia Xs/Tsunami pumping system from Spectra Physics has also been described elsewhere.20 The laser excitation pulse and sample fluorescence emissions were collected at the magic angle using automatic alternate measurements of pulse and sample (1 × 103 counts in
■
RESULTS Absorption and Steady-State Fluorescence Spectra. In the ground state, HMF exhibits two forms, the flavylium cation (AH+) and the corresponding quinonoidal base (A), with a pKa = 4.45 in water.11 Addition of 1,4-dioxane (Dx) reduces the pKa to 3.913 at a dioxane mole fraction of xDx = 0.2. The absorption spectra of HMF at acidic pH (0.02 mol L−1 added [HClO4], where essentially only the AH+ form is present) in water and in Dx/water mixtures with xDx = 0.2 and xDx = 0.5 are presented in Figure S1 of the Supporting Information. The addition of 1,4-dioxane induces a small bathochromic shift of the absorption maximum of AH+ from 415 nm in water to 423 and 424 nm at xDx = 0.2 and xDx = 0.5, respectively. The corresponding steady-state fluorescence spectra of HMF, measured at [HClO4] = 0.02 mol L−1 in water and the same two H2O/Dx mixtures with excitation at 415 nm, are shown in Figure S2 of the Supporting Information. Fluorescence excitation spectra collected at the emission maxima of AH+* and A* are identical to the absorption spectrum of AH+ in all cases, indicating that A* is formed exclusively via adiabatic deprotonation of AH+* in the excited state.11,13 The addition of Dx results in a progressive increase in the fluorescence emission intensity of the acid form, AH+*, reflecting a decrease in the efficiency of proton transfer and B
dx.doi.org/10.1021/jp504189m | J. Phys. Chem. A XXXX, XXX, XXX−XXX
The Journal of Physical Chemistry A
Article
Figure 1. Fluorescence upconversion profiles of HMF in (a) D2O pD = 1.42; (b) H2O; and water/Dx mixtures with (c) xDx = 0.2, (d) xDx = 0.3, and (e) xDx = 0.5. All measurements were performed at 293 K and samples (b)−(e) contained [HClO4] = 0.02 mol L−1.
results of global fits of the fluorescence decays of HMF in water and in the H2O/Dx mixtures as a function of temperature. Note that τ2 exhibits the smallest variation with T among the decay times measured in water/Dx mixtures, despite the dispersion present. The sum ∑3j=1A2,j of the pre-exponential coefficients at the emission wavelength of A* is near zero after subtraction of the contribution of AH+* to the emission13 at 630 nm (Figure S2, Supporting Information) and is consistent with the formation of A* in the excited state exclusively via adiabatic deprotonation of HA+*. The increase in the χ2 values of the double-exponential fits of the AH+* signal in water as the temperature increases suggests that the decay time of AH+* above ca. 300 K (
