Excited-State Proton Transfer from Quinone ... - ACS Publications

69978, Israel. J. Phys. Chem. A , 2014, 118 (10), pp 1832–1840. DOI: 10.1021/jp412428a. Publication Date (Web): February 10, 2014. Copyright © ...
2 downloads 4 Views 1MB Size
Download PDF
Article pubs.acs.org/JPCA

Excited-State Proton Transfer from Quinone-Cyanine 9 to Protic Polar-Solvent Mixtures Ron Simkovitch, Shay Shomer, Rinat Gepshtein, Doron Shabat, and Dan Huppert* Raymond and Beverly Sackler Faculty of Exact Sciences, School of Chemistry, Tel Aviv University, Tel Aviv 69978, Israel S Supporting Information *

ABSTRACT: Steady-state and time-resolved emission techniques were used to study the excited-state proton-transfer (ESPT) process of quinone cyanine 9 (QCy9) in solvent mixtures. We found that the ESPT rate from QCy9 in water/methanol mixtures is independent of the mixture composition and the rate constant is kPT ∼ 1013 s−1. In ethanol/ trifluoroethanol (TFE) mixtures the ESPT rate strongly depends on the solvent-mixture composition. We observe two ESPT rates rather than one over a wide range of solvent-mixture compositions. The average ESPT rate decreases as the mole fraction of TFE increases.

■

The ESPT rate of weak photoacids (pKa* > 0), in methanol and ethanol is smaller than in water by about 3 orders of magnitude. Because the radiative lifetime of the singlet excited state of photoacids and other organic compounds is on the order of 5−10 ns, the efficiency of the proton-transfer process of weak photoacids in alcohols is very small. For 8hydroxypyrene-1,3,6-trisulfonic acid (HPTS), whose pKa* ∼ 0.4 the ESPT rate in water is 2 × 1010 s−1 (ref 16 and 17), the ESPT rate of HPTS in methanol is estimated to be 107 s−1. Because the radiative rate of the protonated form is 2 × 108 s−1, the ESPT efficiency is less than 5%. For several weak photoacids, the ESPT rate was measured in water/methanol mixtures.18,19 It was found that in the solvent mixtures, the rate decreases as the methanol concentration is increased. A literature review of pKa data of weak acids indicates that the dependence on solvent composition is mainly due to localized counterion stability in water-rich solutions and to proton stability in methanol-rich solutions. Quantitative agreement with literature compilations of proton-transfer free energy was obtained.19 These observations challenge another interpretation, namely that a minimal cluster of four water molecules is required to accept a proton in the ESPT process in water/ methanol mixtures.20−22 Strong photoacids (pKa* < −2) are also capable of transferring a proton at a rapid rate (τPT < 100 ps) to alcohols.23−25 It is expected that the ESPT rate in water/ methanol mixtures will depend on the solvent-mixture composition over a much smaller range than that found for weak photoacids. In QCy9, the strongest photoacid, the ESPT rate is the same in water and in methanol. In this study, we

INTRODUCTION Photoacids are aromatic organic molecules that are weak acids in their ground electronic state, but of acidity greater by many orders of magnitude in their first excited electronic state. Thus, photoexcitation to the excited state, by short UV−vis laser pulses, enables one to follow the photoprotolytic processes. Excited-state intermolecular proton transfer (ESPT) from the acidic group of the excited photoacid to a nearby solvent molecule1−13 is widely researched. In a recent study14 we reported on the photoprotolytic properties of QCy9 (shown in Scheme 1). Scheme 1. Molecular Structure of QCy9

The phenol hydroxyl transfers a proton to the solvent when the QCy9 is in the excited state. We found that QCy9 in water is a superphotoacid with pKa* ≈ −8.5; an even more remarkable finding is that it exhibits a very large ESPT rate constant, kPT ≈ 1 × 1013 s−1. This is the largest kPT value reported in the literature up to now. In a more recent study, we extended the photoacidity study of QCy9 in other protic solvents. We measured the ESPT rates of QCy9 also in methanol and ethanol. For QCy9, the strongest reported photoacid up to now, we find that the ESPT rate is nearly the same for water, methanol, and ethanol,15 namely, kPT ≈ 1013 s−1. © 2014 American Chemical Society

Received: December 19, 2013 Revised: February 6, 2014 Published: February 10, 2014 1832

dx.doi.org/10.1021/jp412428a | J. Phys. Chem. A 2014, 118, 1832−1840

The Journal of Physical Chemistry A

Article

measured the ESPT rate of QCy9 in water/methanol mixtures. We find that the ESPT rate constant is independent of the solvent-mixture composition. In addition, we measured the ESPT rate of QCy9 in ethanol/ trifluoroethanol mixtures. Trifluoroethanol (TFE) differs from ethanol in its hydrogen-bonding properties. Water, methanol, and ethanol accept and donate hydrogen bonds with nearby molecules in neat solvent and in solvent mixtures of protic solvents with similar properties. The TFE molecule can donate a hydrogen bond (HBD) but cannot accept a hydrogen bond (HBA).26 It was found that strong photoacids cannot transfer a proton in the excited state to TFE, which lacks the HBA property, whereas, in ethanol, which can accept a hydrogen bond from the hydroxyl group of the photoacid, the ESPT rate is large. In the current study, we found two ESPT rates rather than one, over a wide range of solvent-mixture compositions. The average ESPT rate decreases as the mole fraction of TFE increases.

Scheme 2. Photoprotolytic Cycle

Proton dissociation, with an intrinsic rate constant, kPT, leads to the formation of an ion-pair RO−*···H3O+ that subsequently forms an unpaired RO−* and a solvated proton, which diffuses into the bulk of the solution. The proton and RO−* may recombine via reversible (adiabatic) recombination with a rate constant, ka, and re-form the excited acid, ROH*. In general, back-protonation may also proceed by an irreversible (nonadiabatic) pathway, involving fluorescence quenching of the RO−* by a proton with a rate constant, kq, forming the groundstate ROH. 1-Naphthol and its derivatives are known to exhibit considerable fluorescence quenching of the deprotonated form, RO−*, in acidic aqueous solutions. Removal of an ion-pair from the contact radius, a, to infinity is described by the transient numerical solution of the DebyeSmoluchowski equation (DSE).27,28 The motion of the transferred proton in water near the photoacid depends strongly on the electrical potential existing between it and the deprotonated form. The diffusion-assisted geminate recombination of RO−* with the proton can be quantitatively described with the use of a numerical solution of the DSE under the initial and boundary conditions of the photoprotolytic process. In addition, the fluorescence lifetimes of all excited species are considered, with 1/k0 = τROH for the acid, and 1/k0′ = τRO− for the conjugate base. Generally, k0′ and k0 are much smaller than both the proton-reaction and the diffusion-controlled rate constants. The amplitude of the long-time fluorescence tail of ROH* depends on the intrinsic rate constants, ka and kPT, on the proton-diffusion coefficient, DH+, and on the electrical potential between RO−* and the proton. Trifluoroethanol/Ethanol Mixtures. Figure 1 shows the steady-state-emission spectra of QCy9 in trifluoroethanol/ ethanol mixtures. TFE donates a hydrogen bond but cannot accept one.26 The steady-state fluorescence of weak photoacids (pKa* > −2) in TFE show only the ROH emission band. In neat TFE, the

■

EXPERIMENTAL SECTION The fluorescence up-conversion technique was employed in this study to measure the time-resolved emission of QCy9 in several solvent mixtures at room temperature. The laser used for the fluorescence up-conversion was a cavity-dumped Ti:sapphire femtosecond laser (Mira, Coherent), which provides short, 150 fs, pulses at about 800 nm. The cavity dumper operated with a relatively low repetition rate of 800 kHz. The up-conversion system (FOG-100, CDP, Russia) operated at 800 kHz. The samples were excited by pulses of ∼8 mW on average at the SHG frequency. The time response of the up-conversion system was evaluated by measuring the relatively strong Raman−Stokes line of water, shifted by 3600 cm−1. It was found that the full width at half-maximum (fwhm) of the signal is 340 fs. Samples were placed in a rotating optical cell to avoid degradation. We found that, during our 5 min time-resolved measurements in a cell rotating at a frequency of 10 Hz, the degradation of the sample was marginal and had no effect on the decay of the signal profile. The QCy9 samples were excited in their protonated ROH form (Scheme 1). We estimate that the error in the determination of the excited-state proton-transfer rate constant is ±10%. For rates greater than 1013 the error is larger ±20%. Measurements of time-correlated single-photon counting (TCSPC) were performed using the same laser as a light source, and in the same setup. The TCSPC detection system was based on a Hamamatsu 3809U photomultiplier and an Edinburgh Instruments TCC 900 computer module for TCSPC. The overall instrument response was approximately 40 ps (fwhm) where the excitation pulse energy was reduced to about 10pJ by neutral-density filters. The steady-state emission and absorption spectra were recorded by a Horiba Jobin Yvon FluoroMax-3 spectrofluorometer. Experiments were carried out on samples at concentrations of about 1 mM or less.

■

RESULTS

Reversible and Irreversible Photoprotolytic Cycles of Photoacids. Excitation of a photoacid solution of pH lower than its ground-state pKa, generates a vibrationally relaxed, electronically excited ROH molecule (denoted by ROH*) that initiates a photoprotolytic cycle (Scheme 2).

Figure 1. Steady-state-emission spectra of QCy9 in trifluoroethanol/ ethanol mixtures. 1833

dx.doi.org/10.1021/jp412428a | J. Phys. Chem. A 2014, 118, 1832−1840

The Journal of Physical Chemistry A

Article

Figure 2. Time-resolved emission of QCy9 in neat ethanol and TFE at 480, 500, and 520 nm.

Figure 3. Time-resolved emission of (a) QCy9 ROH form measured at 520 nm and (b) RO− form at 680 nm in neat ethanol, neat TFE, and several mixtures of these solvents.

steady-state emission of QCy9 consists of two emission bands: an intense and broad ROH band with a peak at ∼480 nm and a weak band of relative intensity ∼0.1 with a peak at 670 nm, assigned to the emission of RO−. As the ethanol content is increased, the ROH-emission intensity decreases and the RO− band intensity increases. In TFE-rich solutions, the RO− band position shifts slightly to the red as the ethanol portion in the mixture increases. The RO− band-shift is about 22 nm between neat TFE and neat ethanol. The ROH band shifts slightly to the blue in ethanol-rich solutions. Figure S1 (Supporting Information) shows the time-resolved emission of QCy9 in neat TFE measured at several wavelengths over the spectral region of 440−660 nm, displayed on a semilogarithmic scale. The fluorescence was measured by the fluorescence up-conversion technique with a system-response function of ∼300 fs full-width half-maximum. The ROH form of QCy9 was excited at 387 nm by a 150 fs laser pulse at a repetition rate of 800kHz. As seen in the figure, the signals

measured at 520, 540, and 560 nm show moderate decay with an average time constant of about τ≈350 ps. At shorter wavelengths (λ < 480 nm), the fluorescence decay also consists of shorter decay-time components which we assign to solvation dynamics that, in time, shift the ROH band to the red. At long wavelengths (λ > 660 nm), the signal shows a distinct rise with a time constant of about 25 ps followed by a long time decay (τ ≈ 600 ps). Figure 2 shows a comparison of the time-resolved emission of QCy9 in neat ethanol and TFE at 480, 500 and 520 nm. As seen in the figure, the average fluorescence-decay rate in ethanol at short wavelengths (480−520 nm), is much larger than in TFE. We can explain this large difference in fluorescence decay by the fact that the ESPT process is ultrafast in ethanol, with a time constant of less than 100 fs, whereas in TFE, the ESPT process is very slow if it exists at all. We estimate that in TFE, the ESPT time constant, τPT ≥ 100 ps, but the proton geminate recombination rate is even larger. 1834

dx.doi.org/10.1021/jp412428a | J. Phys. Chem. A 2014, 118, 1832−1840

The Journal of Physical Chemistry A

Article

Figure 4. Time-resolved emission of the ROH (460−520 nm) and RO− (680, 700 nm) forms of QCy9 in neat methanol, water, and several water/ methanol mixtures.

Information) provide the fitting parameters of the signals shown in Figures S2−S5 (Supporting Information). Panels a and b of Figure 3 show the time-resolved emission of the ROH form of QCy9, measured at 520 nm and the RO− form measured at 680 nm in neat ethanol, neat TFE and several mixtures of these solvents. The fluorescence decay measured at 520 nm follows the conversion of the ROH form to RO− by the ESPT process. For ethanol, the ESPT rate constant is ultrafast (1013 s−1) whereas for TFE, the ESPT rate is rather slow. In ethanol/TFE mixtures, the ESPT rate falls as the TFE content increases. The rise time seen in the signal measured at 680 nm increases with increasing TFE content in the solvent mixture and fits the decay time measured at 520 nm. TCSPC Measurements. Figure S7a and Figure S7b (Supporting Information) show the time-resolved emission signals of the ROH and RO− forms, respectively, of QCy9 in TFE/ethanol mixtures of low ethanol concentration. The average ROH fluorescence-decay time, τav, decreases as the mole fraction of ethanol increases. In neat TFE, τav = 0.43 ns, whereas in TFE/ethanol of χethanol = 0.29, τav = 0.13 ns. We explain these results as arising from a greater proton-transfer rate as the mole fraction of ethanol increases. The RO− timeresolved emission signals of QCy9 in these solvent mixtures show a signal rise followed by an exponential decay. The signal rise time depends on the composition of the solvent mixture and qualitatively fits the decay time of the ROH signal. This is a

The net result of this excited-state quasi-equilibrium on the time-resolved emission of the ROH form is a small amplitude of a relatively short time component followed by a large amplitude nonexponential long-time fluorescence tail with an asymptotic lifetime of the RO−.16,17 Figures S2−S6 (Supporting Information) show the timeresolved emission of QCy9 in five ethanol/TFE mixtures, in which the mole fraction of ethanol was 0.16, 0.34, 0.54, 0.76, and 0.87. Each figure shows the fluorescence up-conversion signal over the spectral range of 440−700 nm, covering the two emission bands of QCy9. In the ethanol-rich mixtures, the signals at short wavelengths (λ < 560 nm), consist of a largeamplitude short-time decay component of 100−150 fs followed by smaller-amplitude long-time decay components. The amplitude of the short-time decay component decreases as a function of χTFE. When the TFE content is larger than χTFE ≥ 0.16, a second time component of τ ≥ 3 ps is noticeable. The amplitude of this time component increases with χTFE. The signals at long wavelengths (λ ≥ 680 nm), show a distinctive rise in the signal with a time component that is similar to the decay time of the short-wavelength component. We attribute the fast-decay components of the ROH signal to the ESPT process taking place in the excited ROH form of QCy9 in these solvent mixtures. We fit the experimental results shown in Figures S2−S5 (Supporting Information) with three or four exponential decay functions. Tables S2−S6 (Supporting 1835

dx.doi.org/10.1021/jp412428a | J. Phys. Chem. A 2014, 118, 1832−1840

The Journal of Physical Chemistry A

Article

Equation 1 indicates that the amplitude of the long fluorescence depends on many parameters. If we assume that a, kr, and kPT are nearly independent of the solvent composition, then the amplitude ratio of the fluorescence tail in methanol and water at time t can be expressed as

consequence of the ESPT process that takes place in the TFE/ ethanol mixtures. The RO− emission lifetime depends to a small degree on the solvent composition. The larger the ethanol fraction in the mixture, the shorter the excited-state lifetime. A similar phenomenon of strong dependence of the RO− excitedstate lifetime of quinone cyanine dyes on the protic solvent has been observed. The lifetime of the RO− form of QCy9 in water, methanol and ethanol is 60, 150, and 220 ps, respectively. The RO− lifetime also exhibits an isotope effect of about 2 in these neat solvents. Figure S8a and Figure S8b in the Supporting Information show the TCSPC emission signals of the ROH and RO− bands on a semilogarithmic scale. In mixtures with large TFE content (χTFE > 0.5), the ESPT rate decreases and the average kPT values are smaller than 1010 s−1. The average values of kPT measured by the TCSPC technique for ethanol/TFE mixtures are given in Table S1 (Supporting Information). Water/Methanol Mixtures. Figure 4 shows the timeresolved emission of the ROH and RO− forms of QCy9 measured at several wavelengths, in neat methanol, water, and several water/methanol mixtures. In a previous study,15 we found that the ESPT rate of QCy9 in water and methanol is about the same and is ultrashort, kPT = 1013s−1. Figure 4 shows that the ESPT rate is ultrafast not only in water or methanol but also in water/methanol mixtures and that the rate is the same for all solvent mixtures compositions. Figure S10 (Supporting Information) shows the ROH timeresolved emission signals shown on a semilogarithmic scale with a time window extended to about 100 ps. The ROH signals show a weak, nonexponential, long fluorescence tail that follows the large-amplitude ultrashort decay component attributed to the ESPT process. This long fluorescence tail is explained by the geminate-recombination (GR) model. The solvated proton may recombine with the excited RO−* to reform the ROH*. The GR process depends on several parameters, such as the forward and reverse intrinsic rate constants, kPT and kr, as well as the Coulomb attraction potential and the ROH and RO− excited-state radiative lifetimes. Because the QCy9 RO− radiative lifetime is solventdependent, the average decay time of the long fluorescence tail is also solvent-dependent, as seen in Figure S10 (Supporting Information). The RO− decay time in neat methanol and pure water is 120 and 60 ps, respectively. In the water mixtures the RO− decay times decrease as the water content increases. Another factor that determines the amplitude of the long fluorescence is the effective dielectric constant of the solvent mixture. The dielectric constants of water and methanol are 78 and 32, respectively. In water/methanol mixtures the effective dielectric constant follows the sum of the product of the mole ratios of the individual components with their dielectric constants. An asymptotic expression for the long-fluorescence amplitude and its unique, nonexponential time dependence is IfROH(t ) ∼

πa 2k r exp[−V (a)] d /2

2kPT(πD)

+ I methanol /I H2O = exp[(εH2O/εmethanol)/(D Hmethanol /D HH+2O)]3/2

= 11/(0.45)3/2 ≈ 36

However, we find experimentally, a ratio of only 4 of the amplitudes of the long fluorescence tails. One explanation for the smaller-than-expected amplitude of the long fluorescence tail in methanol is that the effective negative charge on the phenolate oxygen is much smaller than one electronic unit. The strong photoacidity of QCy dyes arises from the large charge redistribution of their RO− form. The two picolinium moieties receive their electronic charge density at the expense of the phenolate oxygen. When the proton is far from the RO−, the effective charge of the RO− molecule is not that of the phenolate, but of the total charge of the RO−, which also includes the positive picolinium charge, and so the total charge of the RO− is z = 1 rather than −1. Thus the proton is repelled from RO− rather than attracted to it. In such a case, it is expected that a low dielectric constant will decrease the GR probability and the tail amplitude will decrease rather than increase as expected for an attraction potential. Another explanation is that the intrinsic sphere radius increases because it also includes one solvent molecule. The water molecule is much smaller than methanol and thus an increase in the radius, a, causes a decrease in V(a) in eq 1. Another explanation is that the intrinsic recombination rate also decreases when water is replaced by methanol. Amplitude of the Long ROH Fluorescence Tail. As seen in Figures 4 and S10 (Supporting Information) the amplitude of the long fluorescence tail of the ROH form of QCy9 in the time-resolved emission measurement, depends on the mixtures and their composition. The amplitude of the tail is given by eq 1, and its dependence on the proton diffusion coefficient is as D−3/2. The proton mobility in water/methanol mixtures was measured by Erdey−Grúz by electrochemical means.29 It was found that the excess proton diffusion coefficient decreases with increasing mole fraction of methanol in the water/methanol mixture. At about χMeOH ≈ 0.7 it reaches a rather broad minimum and at higher methanol content it increases. The minimum value of DH+ is about a quarter that of water and about half that of neat methanol. The amplitude of the long fluorescence tail depends on both the dielectric constant and the diffusion coefficient. As seen in the figures, the largest longfluorescence-tail amplitude is achieved not in neat methanol with the smallest dielectric constant, but in solvent mixtures of high methanol content in which the effective diffusion coefficient is rather small. Main Findings. 1. Excited-state proton transfer (ESPT) from QCy9, whose pKa* is −8.5, to polar solvents with low hydrogen-bondaccepting properties, occurs at a low rate. 2. Over a wide range of ethanol/trifluoroethanol (TFE) compositions, we observe two ESPT rates rather than one. This indicates that two types of solvent clusters surround the QCy9 phenol. The average ESPT rate of

t −d /2 (1)

where a is the reaction-sphere radius, kPT and kr are the intrinsic ESPT and GR rates occurring on the reaction sphere, −V(a) = RD/a is the potential at the reaction sphere, D is the mutual diffusion coefficient between RO− and H3O+, and d is the diffusion-space dimension. This expression predicts a nonexponential fluorescence decay that fits a power law of t‑d/2. 1836

dx.doi.org/10.1021/jp412428a | J. Phys. Chem. A 2014, 118, 1832−1840

The Journal of Physical Chemistry A

■

Article

hydroxy-1,3,6-pyrenetrisulfonate (HPTS), a weak photoacid, whose pKa* ∼ 1.3, and its photoprotolytic properties in water/ methanol mixtures. Figure S11 in the Supporting Information shows the normalized steady-state emission spectra of HPTS in water/methanol mixtures. In pure water, the relative intensity of the ROH* band, whose peak is at ∼440 nm, with respect to that of the RO−* band is 0.045. When methanol is added to water, the relative intensity of the ROH* band increases. At high concentrations of methanol, namely, at molar ratios of χMeOH ≈ 0.4, the intensity of the ROH* band of HPTS is stronger than that of the RO−* band. In neat methanol, the intensity of the RO−* emission band is very low and, if it exists, it is because commercial methanol (HPLC grade) contains ∼0.15% water. Figure S12 (Supporting Information) shows the timeresolved emission of the protonated ROH* form of HPTS, measured by the TCSPC technique in several neutral-pH water/methanol mixtures. As seen in the figure, the decay time increases considerably with increase in the methanol concentration. The average lifetime (including the GR long fluorescence tail) increases from 220 ps in pure water to about 4 ns (approximately the radiative lifetime) in neat methanol. When χMeOH ≈ 0.7, the radiative rate limits the accurate determination of kPT, because kPT ≪ kr. Figure S13 (Supporting Information) shows the TCSPC signal of the RO−* form of HPTS in several water/methanol mixtures. In mixtures containing more than 0.1 mol ratio of water, the signal measured at 530 nm at early times has a distinct rise-time that nicely fits the decay times of the ROH* signal, shown in Figure S12 (Supporting Information). The exponential long-time radiative decay of the signal, 5.3 ns, is invariant under change in the methanol concentration in mixtures containing more than 0.1 mol ratio of water. Figure 5 shows the plot of kPT of HPTS versus the mole fraction of methanol in the water/methanol mixture.

QCy9 in ethanol/TFE mixtures increases with increasing ethanol content of the mixture. 3. The average ESPT rate from QCy9 in these mixtures increases with a complex dependence on the mixture composition by about 4 orders of magnitude, from ∼109 s−1 in neat TFE to 1013 s−1 in neat ethanol. 4. The ESPT rates of QCy9 in pure water and methanol solvents are about the same, kPT ∼ 1013 s−1. 5. The ESPT rate of QCy9 in water/methanol mixtures is independent of the mixture composition and is constant at kPT ∼ 1013 s−1.

DISCUSSION In this work we studied the dependence of the excited-state proton-transfer (ESPT) rate of QCy9 in two types of solvent mixtures. One mixture consists of two protic solvents, each capable of accepting a hydrogen bond (HBA), like water, methanol, and ethanol; the second type of mixture is composed of a solvent that accepts a hydrogen bond and a cosolvent like trifluoroethanol that is polar with negligible or small hydrogenbond-accepting capability. The results of previous studies reveal that weak photoacids, with pKa* > 0.4, are incapable, within the lifetime of the excited state (a few nanoseconds), of transferring a proton to protic solvents like methanol or ethanol, even though they transfer a proton with a time constant as short as 50 ps to water. In a recent article, we qualitatively explained the dependence of the ESPT rate constant on both the solvent and the strength of the photoacid, by using the adiabatic and nonadiabatic proton-transfer formalism.30 The ESPT rates of the weak acids are determined by the nonadiabatic protontransfer-rate expression. Both the pre-exponential factor and the exponential activation-energy factor of the rate-constant expression depend on the photoacidity strength (pKa*) and the solvent properties. With increasing strength of the photoacid, the pre-exponential factor increases and the activation energy decreases; therefore, the value of the ESPT rate constant increases. The ESPT process of superphotoacids with pKa* > −2 is an adiabatic rather than a nonadiabatic process. The preexponential factor is determined by a characteristic frequency, ws/2p, rather than by the coupling matrix element (HAB) as in the nonadiabatic PT-rate equation. The activation energy is small because it depends on the value of ΔG0 of the ESPT reaction, which is negative for strong photoacids. For QCy9, the activation energy is rather small and therefore the PT rate constant reaches the value of the pre-exponential factor. The intermolecular-vibration-assisted proton-tunneling theory of Goldanski et al.31 provides a reasonable explanation for the value of 1013 s−1 as the limiting value of the ESPT rate constant. The intermolecular frequency of two nearby water molecules, which modulates the O−H---O distance is on the order of 1013 s−1. The same frequency probably also exists between a photoacid hydroxyl and a nearby protic-solvent molecule (water or small monol, like methanol or ethanol). In the asymptotic limit, where the activation energy approaches zero and applies to the ESPT of strong photoacids like QCy9, the PT rate constant, according to the adiabatic proton-transfer theory, will be equal to the prefactor ws/2p ≈ 1013 s−1, determined by the intermolecular oxygen−oxygen vibration frequency and is nearly independent of the protic solvent. ESPT of HPTS (Weak Photoacid) in Water/Methanol Mixtures. In this subsection we provide the ESPT results of 8-

Figure 5. Proton-transfer rate constant kPT of HPTS derived from both TCSPC data analysis and from steady-state emission with the use of eq S1 (Supporting Information), versus the mole fraction of methanol in the water/methanol mixture.

We used the reversible geminate-recombination model and the user-friendly graphic SSDP program of Krissinel and Agmon32 to fit the data. The values were derived from the best fit to the TCSPC signals of the ROH* (black squares). As seen in the figure, the proton-transfer rate, calculated with the use of the reversible-geminate-recombination model, decreases nearly exponentially as a function of the mole fraction. The figure also shows the values of kPT, calculated from eq S1 in the 1837

dx.doi.org/10.1021/jp412428a | J. Phys. Chem. A 2014, 118, 1832−1840

The Journal of Physical Chemistry A

Article

Table 1. Fitting Parameters for the Fluorescence of the ROH Form of QCy9 in Ethanol/TFE Mixtures, Measured at 520 nm composition

a1a

τ1a (fs)

a2b

τ2b (fs)

a3c

τ3c (ps)

a4d

τ4d (ps)

0.16 0.34 0.54 0.76

0.52 0.45 0.32 0.16

100 130 150 150

0.1 0.1 0.1 0.04

700 600 600 600

0.068 0.19 0.24 0.29

3.0 3.6 4 7.0

0.092 0.1 0.18 0.24

15.2 25 30 65

a

This time component is assigned to the ESPT to an ethanol solvent cluster (ROH---ethanol). bThis short time component is also found in pure water, methanol, and ethanol, and we attribute it to rapid solvent rearrangement. cThis time component is assigned to the ESPT to a TFE solvent cluster (ROH---TFE). dThis long time component is assigned as the geminate-recombination component.

Supporting Information and the ratio of the ROH/RO− fluorescence bands intensities, taken from the steady-stateemission spectra shown in Figure S12. (Supporting Information) Poton-Transfer Rate of QCy9 in TFE/Ethanol Mixtures. The fluorescence up-conversion data in TFE/ethanol mixtures indicates that there are two ESPT rates, an ultrafast rate of about 6 × 1012 to 1 × 1013 s−1 and a much smaller rate which is less than 3 × 1011 s−1. The relative amplitude of the fast ESPT rate decreases as χTFE increases. In addition, the ultrafast data indicate that the amplitude of the GR long fluorescence tail is of the order of 0.25 in neat ethanol (Figure 3) and it increases further with χTFE. After the ESPT process takes place, the transient ROH* and RO−* populations tend to form a quasi* = [RO−*]t[H+]t/[ROH*]t. The value of equilibrium with Keq K*eq depends on the solvent mixture composition and decreases when χTFE increases. When K*eq is small, [RO−*]t is small, whereas the [ROH*]t is large. In such a case, quantitative analysis of the time-resolved emission data is rather difficult because the fluorescence decay of the ROH signal results from three major contributors with similar amplitudes (Table 1). Figure 6 provides a plot of log kPT versus the mole fraction of TFE.

fluorescence up-conversion data, the TCSPC and the steadystate data arise from the large contribution of the ROH long fluorescence tail originating from the geminate-recombination rate. Table 1 provides the fitting parameters of a fourexponential-function fit to the time-resolved emission measured at 520 nm by the fluorescence up-conversion technique. As seen in Table 1 and Figure 3, the fluorescence decay of the ROH form of QCy9 measured at 520 nm indicates that the ESPT in TFE/ethanol mixtures occurs on two time scales, an ultrafast ESPT on a scale of 100−150 fs and a long ESPT that occurs at τPT ≥ 3 ps. At χTFE = 0.54 and higher, the main time component is the long one. The TCSPC measurements in TFE-rich mixtures show that the average ESPT rate decreases the larger the TFE content. Robinson and co-workers20 explained the ESPT rate dependence on the water/methanol mixture composition of weak photoacids by suggesting that a proton is transferred only to a unique water structure that exists in the mixture and contains four water molecules that form H9O4+, Eigen’s protonated water structure.33 In a similar line of reasoning we suggest that the proton transfer from QCy9 dissolved in ethanol/TFE mixtures occurs to two types of solvent clusters. We explain the two time scales of the ESPT process by two solvent configurations around the QCy9 phenol. In one solvent cluster, an ethanol molecule is hydrogen-bonded to the phenol OH and in the second configuration, a TFE molecule is hydrogen-bonded to the phenol. The ultrafast ESPT occurs in the cluster in which the ethanol is hydrogen-bonded to the phenol. In this configuration, the PT rate is large because the proton acceptor (the ethanol) has hydrogen-bond-accepting capabilities (HBA), whereas the slow ESPT occurs in solvent clusters in which TFE, which lacks HBA, is bound to the phenol. The two types of solvent clusters coexist over a large range of solvent compositions and thus the time-resolved emission is composed of short and longer-time ESPT components. We found that the short time component is almost independent of the solvent composition, whereas for the longer one, the decay time depends to a greater extent on the solvent composition. The average decay time, which also includes the geminate recombination time component, increases as χTFE increases because the amplitudes of the slow ESPT and the GR time components increase with χTFE.

Figure 6. ESPT rate constant kPT as a function of the mole fraction of TFE.

The fluorescence up-conversion data provide both the ultrafast and the slow ESPT rates. The steady-state fluorescence provides the time-integrated (steady-state) fluorescence intensity ratio IRO−/IROH. The plot also shows the average of three time constants (1/τav) obtained from the TCSPC data. Two of these are the ESPT rate constants and the third is the large contribution of the GR long fluorescence decay that depends on the emission lifetimes of both ROH and RO−. The steadystate intensity ratio IRO−/IROH in Figure 6 was multiplied by ∼109, the radiative rate of RO− (eq 1) and provides an estimate for the average ESPT rate. The large differences between the

■

SUMMARY AND CONCLUSIONS

In the current work, we studied the rate of excited-state proton transfer (ESPT) of quinone cyanine 9 (QCy9), a very strong photoacid with pKa* ∼ −8.5 in two solvent mixtures. Steadystate and time-resolved emission techniques were employed to measure the rate of proton transfer. In previous studies14 of QCy9, we found that the ESPT rate constant to water is ∼1013 s−1, the largest reported rate. In a recent study, we found that this high ESPT rate is also maintained in methanol and 1838

dx.doi.org/10.1021/jp412428a | J. Phys. Chem. A 2014, 118, 1832−1840

The Journal of Physical Chemistry A

Article

ethanol.15 The ESPT rate of weak photoacids, with pKa* > 0 in water, is smaller than 2 × 1010 s−1 and in methanol or ethanol it is lower by more than 3 orders of magnitude. Because the excited-state radiative lifetime is only a few nanoseconds, the ESPT efficiency is low and in many photoacids, the ESPT process is not detected at all. In water/methanol mixtures the ESPT rate of weak photoacids gradually decreases as the methanol content of the mixture is raised. In the current study we measured the ESPT rate of QCy9 in water/methanol mixtures. We found that the rate is independent of the methanol content of the solvent mixture. This finding is expected, because the ESPT rates in pure water and methanol are similar, kPT ∼ 1013 s−1. The ROH timeresolved emission signals of QCy9 in water, methanol and their mixtures also show a distinctive nonexponential long fluorescence tail characteristic of a reversible photoacid. The fluorescence tail arises from the finite probability that the solvated proton may recombine with the excited RO−* to reform the excited ROH*. The intensity of the fluorescence tail depends on many parameters. The dielectric constant and the proton diffusion coefficient of the solvent mixture are the important parameters. Equation 1 provides the relative amplitude and the time-dependence of the fluorescence tail. The relative amplitude of the fluorescence tail in neat methanol and methanol-rich mixtures is much smaller than expected from the differences in the dielectric constants of water, methanol and water/methanol mixtures. We interpret this result as arising from the redistribution of the electronic charge of the RO−* form, which decreases the effective charge on the phenolate oxygen of QCy9. We also measured the ESPT rate of QCy9 in ethanol/ trifluoroethanol (TFE) mixtures. TFE can donate a hydrogen bond (HBD) but cannot accept a hydrogen bond (HBA), whereas water and alcohols like methanol and ethanol can both donate and accept a hydrogen bond. It was found that the ESPT process in TFE does not take place in weak and strong photoacids whose pKa* > −2. This phenomenon is explained by the fact that a hydrogen bond between the hydroxyl hydrogen of the photoacid and the oxygen of the solvent is required for an ESPT process to occur and TFE does not form such an HBA. We found, over a wide range of solvent compositions, two ESPT rates rather than one. We explain this phenomenon by the existence of two types of solvent mixture clusters surrounding the QCy9 phenol. In one cluster, the ethanol forms a hydrogen bond with the phenol hydroxyl group, and this cluster enables ultrafast proton transfer at rates approaching 1013 s−1. In the second cluster, a TFE molecule donates, but does not accept, a hydrogen bond and therefore the ESPT rate for this type of cluster is slow. Thus, the average ESPT rate gradually decreases as the mole ratio of TFE is raised.

■

water/methanol mixtures (including emission spectra and TCSPC signal). This information is available free of charge via the Internet at http://pubs.acs.org

■

AUTHOR INFORMATION

Corresponding Author

*D. Huppert: e-mail, [email protected]; phone, 972-36407012; fax, 972-3-6407491. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported by grants from the James-Franck German-Israeli Program in Laser-Matter Interaction and by the Israel Science Foundation.

■

REFERENCES

(1) Ireland, J. F.; Wyatt, P. A. Acid-Base Properties of Electronically Excited States of Organic Molecules. Adv. Phys. Org. Chem. 1976, 12, 131−221. (2) Gutman, M.; Nachliel, E. The Dynamic Aspects of Proton Transfer Processes. Biochem. Biophys. Acta 1990, 1015, 391−414. (3) Tolbert, L. M.; Solntsev, K. M. Excited-State Proton Transfer: From Constrained Systems to “Super” Photoacids to Superfast Proton Transfer. Acc. Chem. Res. 2002, 35, 19−27. (4) Rini, M.; Magnes, B. Z.; Pines, E.; Nibbering, E. T. Real-Time Observation of Bimodal Proton Transfer in Acid-Base Pairs in Water. J. Science 2003, 301, 349−352. (5) Mohammed, O. F.; Pines, D.; Dreyer, J.; Pines, E.; Nibbering, E. T. Sequential Proton Transfer Through Water Bridges in Acid-Base Reactions. J. Science 2005, 310, 83−86. (6) Tran-Thi, T. H.; Gustavsson, T.; Prayer, C.; Pommeret, S.; Hynes, J. T. Primary Ultrafast Events Preceding the Photoinduced Proton Transfer from Pyranine to Water. Chem. Phys. Lett. 2000, 329, 421−430. (7) Agmon, N. Elementary Steps in Excited-State Proton Transfer. J. Phys. Chem. A 2005, 109, 13−35. (8) Spry, D. B.; Fayer, M. D. Charge Redistribution and Photoacidity: Neutral Versus Cationic Photoacids. J. Chem. Phys. 2008, 128, 0845081−084508-9. (9) Siwick, B. J.; Cox, M. J.; Bakker, H. J. Long-Range Proton Transfer in Aqueous Acid−Base Reactions. J. Phys. Chem. B 2008, 112, 378−389. (10) Mohammed, O. F.; Pines, D.; Nibbering, E. T. J.; Pines, E. BaseInduced Solvent Switches in Acid−Base Reactions. Agnew. Chem. Int. Ed. 2007, 46, 1458−1461. (11) Mondal, S. K.; Sahu, K.; Sen, P.; Roy, D.; Ghosh, S.; Bhattacharyya, K. Excited State Proton Transfer of Pyranine in a γcyclodextrin Cavity. Chem. Phys. Lett. 2005, 412, 228−234. (12) Prasun, M. K.; Samanta, A. Evidence of Ground-State ProtonTransfer Reaction of 3-Hydroxyflavone in Neutral Alcoholic Solvents. J. Phys. Chem. A 2003, 107, 6334−6339. (13) Bhattacharya, B.; Samanta, A. Excited-State Proton-Transfer Dynamics of 7-Hydroxyquinoline in Room Temperature Ionic Liquids. J. Phys. Chem. B 2008, 112, 10101−10106. (14) Simkovitch, R.; Karton-Lifshin, N.; Shomer, S.; Shabat, D.; Huppert, D. Ultrafast Excited-State Proton Transfer to the Solvent Occurs on a Hundred-Femtosecond Time-Scale. J. Phys. Chem. A 2013, 117 (16), 3405−3413. (15) Simkovitch, R.; Shomer, S.; Gepshtein, R.; Roth, M. E.; Shabat, D.; Huppert, D. Comparison of the Rate of Excited-State Proton Transfer from Photoacids to Alcohols and Water. J. Photochem. Photobiol. A 2014, 277, 90−101. (16) Pines, E.; Huppert, D.; Agmon, N. Geminate Recombination in excited-state proton-transfer Reactions: Numerical Solution of the Debye−Smoluchowski Equation with Backreaction and Comparison with Experimental Results. J. Chem. Phys. 1988, 88, 5620.

ASSOCIATED CONTENT

* Supporting Information S

Proton-transfer rate constant calculation from fluorescence intensity, properties of the ESPT of QCy9 in the ethanol/TFE mixture (including time-resolved emission signals and fluorescence up-conversion signals), time correlated single photon counting, fluorescence up-conversion fitting tables of QCy9 in TFE/ethanol mixtures, properties of the ESPT of QCy9 in a water/methanol mixture (fluorescence upconversion signals), and properties of the ESPT of HPTS in 1839

dx.doi.org/10.1021/jp412428a | J. Phys. Chem. A 2014, 118, 1832−1840

The Journal of Physical Chemistry A

Article

(17) Agmon, N.; Pines, E.; Huppert, D. Geminate Recombination in proton-transfer Reactions. II. Comparison of Diffusional and Kinetic Schemes. J. Chem. Phys. 1988, 88, 5631. (18) Presiado, I.; Erez, Y.; Huppert, D. Excited-State Intermolecular Proton Transfer of the Firefly’s Chromophore d-Luciferin. 2. Water− Methanol Mixtures. J. Phys. Chem. A 2010, 114, 9471−9479. (19) Agmon, N.; Huppert, D.; Masad, A.; Pines, E. Excited-State Proton Transfer to Methanol-Water Mixtures. J. Phys. Chem. 1991, 95 (25), 10407−10413. (20) Lee, J.; Griffin, R. D.; Robinson, G. W. 2-Naphthol: A Simple Example of Proton Transfer Effected by Water Structure. J. Chem. Phys. 1985, 82, 4920−4925. (21) Robinson, G. W.; Thistlethwaite, P. J.; Lee, J. Molecular Aspects of Ionic Hydration Reactions. J. Phys. Chem. 1986, 90 (18), 4224− 4233. (22) Moore, R. A.; Lee, J.; Robinson, G. W. Hydration Dynamics of Electrons from a Fluorescent Probe Molecule. J. Phys. Chem. 1985, 89 (17), 3648−3654. (23) Perez-Lustres, J.; Rodriguez-Prieto, F.; Mosquera, M.; Senyushkina, T.; Ernsting, N.; Kovalenko, S. Ultrafast Proton Transfer to Solvent: Molecularity and Intermediates from Solvation-and Diffusion-Controlled Regimes. J. Am. Chem. Soc. 2007, 129, 5408− 5418. (24) Carmeli, I.; Huppert, D.; Tolbert, L.; Haubrich, J. Ultrafast Excited-State Proton Transfer from Dicyano-Naphthol. Chem. Phys. Lett. 1996, 260, 109−114. (25) Huppert, D.; Tolbert, L. M.; Linares-Samaniego, S. Ultrafast Excited-State Proton Transfer from Cyano-Substituted 2-Naphthols. J. Phys. Chem. A 1997, 101, 4602−4605. (26) Kamlet, M. J.; Abboud, J. L. M.; Abraham, M. H.; Taft, R. W. Linear Solvation Energy Relationships. 23. A Comprehensive Collection of the Solvatochromic Parameters, π*, α, and β, and Some Methods for Simplifying the Generalized Solvatochromic Equation. J. Org. Chem. 1983, 48 (17), 2877−2887. (27) Debye, P. Reaction Rates in Ionic Solutions. Trans. Electrochem. Soc. 1942, 82, 265−272. (28) Collins, F. C.; Kimball, G. E. Diffusion-Controlled Reaction Rates. J. Colloid Sci. 1949, 4, 425−437. (29) Erdey-Grúz, T.; Inzelt, G.; Fodor-Csányi, P. Self-Diffusion Coefficients of Water in Methanol-Water Mixtures. Acta Chem. Ac. Sci. Hung. 1973, 77 (2), 173. (30) Borgis, D.; Hynes, J. T. Curve Crossing Formulation for Proton Transfer Reactions in Solution. J. Phys. Chem. 1996, 100 (4), 1118− 1128. (31) Goldanskiĭ, V. I.; Trakhtenberg, L. I.; Flërov, V. N. Tunneling phenomena in chemical physics: Routledge: xxx, 1989. (32) Krissinel, E. B.; Agmon, N. J. Comput. Chem. 1996, 17, 1085. (33) Eigen, M.; De Maeyer, L. Self-Dissociation and Protonic Charge Transport in Water and Ice. Proc. R. Soc. London, Ser. A 1958, 247, 505−533.

1840

dx.doi.org/10.1021/jp412428a | J. Phys. Chem. A 2014, 118, 1832−1840