Article pubs.acs.org/JPCB
Ultrafast Hydrolysis of a Lewis Photoacid Joseph D. Henrich,†,§ Scott Suchyta,‡,∥ and Bern Kohler*,† †
Department of Chemistry and Biochemistry, Montana State University, Bozeman, Montana 59717, United States Department of Chemistry, The Ohio State University, Columbus, Ohio 43210, United States
‡
S Supporting Information *
ABSTRACT: This study explores the concept that electronic excitation can dramatically enhance Lewis acidity. Specifically, it is shown that photoexcitation transforms an electron-deficient organic compound of negligible Lewis acidity in its electronic ground state into a potent excited-state Lewis acid that releases a proton from a nearby water molecule in 3.1 ps. It was shown previously (Peon et al. J. Phys. Chem. A 2001, 105, 5768) that the excited state of methyl viologen (MV2+) is quenched rapidly in aqueous solution with the formation of an unidentified photoproduct. In this study, the quenching mechanism and the identity of the photoproduct were investigated by the femtosecond transient absorption and fluorescence upconversion techniques. Transient absorption signals at UV probe wavelengths reveal a long-lived species with a pH-dependent lifetime due to reaction with hydronium ions at a bimolecular rate of 3.1 × 109 M−1 s−1. This species is revealed to be a charge-transfer complex consisting of a ground-state MV2+ ion and a hydroxide ion formed when a water molecule transfers a proton to the bulk solvent. Formation of a contact ion pair between MV2+ and hydroxide shifts the absorption spectrum of the former ion by a few nm to longer wavelengths, yielding a transient absorption spectrum with a distinctive triangle wave appearance. The slight shift of this spectrum, which is in excellent agreement with steady-state difference spectra recorded for MV2+ at high pH, is consistent with an ion pair but not with a covalent adduct (pseudobase). The long lifetime of the ion pair at neutral pH indicates that dissociation occurs many orders of magnitude more slowly than predicted by the Smoluchowski−Debye equation. Remarkably, there is no evidence of geminate recombination, suggesting that the proton that is transferred to the solvent is conducted at least several water shells away. Although the hydrolysis mechanism has yet to be fully established, evidence suggests that the strongly oxidizing excited state of MV2+ triggers the proton-coupled oxidation of a water molecule. The observed kinetic isotope effect of 1.7 seen in D2O vs H2O is of the magnitude expected for an ultrafast concerted proton−electron transfer reaction. The ultrafast hydrolysis seen here may be a general excited-state quenching mechanism for electronically excited Lewis acids and other powerful photooxidants in aqueous solution.
1. INTRODUCTION There is great interest in organic compounds that are stronger acids and bases in an excited electronic state than in the ground state.1,2 Time-resolved spectroscopic study of photoacids and photobases has added to the understanding of excited-state proton transfer2−9 and revealed principles important for applications that use these compounds.10 The photoacids and photobases that have been the focus of experimental and theoretical work to date are Brønsted acids and bases that lose or accept a proton upon electronic excitation. In an alternative framework for understanding acidity due to Lewis,11 an acid can be viewed as an electron pair acceptor. Mulliken broadened the Lewis concept and defined an acid as a species that accepts any amount of charge from a donor.12,13 In this paper, we propose that a light pulse can enhance the acidity of a Lewis acid enough to release a proton from a nearby water molecule on an ultrafast time scale. A spectroscopic investigation is presented of the excited-state dynamics of a compound that supports this hypothesis and acts as a Lewis photoacid. © 2014 American Chemical Society
Our focus is on hydrolysisthe process by which Lewis acids such as metal ions,14 halogens,15 and carbocations16 lower the pH of an aqueous solution even though they lack acidic protons. Instead of protolysis, these species undergo hydrolysis by accepting electron density from a coordinated water molecule triggering the release of a proton to the bulk solvent. The hydrolysis reaction of a generic Lewis acid, L+, can be written as L+ + H 2O ⇌ L−OH + H+
(1)
In reaction 1 (and hereafter), the hydronium ion, H3O+, is denoted H+. The hydrolysis product L−OH could represent a Special Issue: Photoinduced Proton Transfer in Chemistry and Biology Symposium Received: October 31, 2014 Revised: December 14, 2014 Published: December 15, 2014 2737
DOI: 10.1021/jp510953e J. Phys. Chem. B 2015, 119, 2737−2748
Article
The Journal of Physical Chemistry B
2. EXPERIMENTAL SECTION 2.1. Transient Absorption. The laser setup used for twocolor pump−probe measurements is described elsewhere.26 Some TA measurements were made with a newer laser spectrometer that uses a commercial amplified titanium:sapphire laser system (Libra, Coherent) to generate 3 mJ, 85 fs pulses centered at 800 nm at a 1 kHz pulse repetition rate. Forty percent of the output was split off to pump an optical parametric amplifier (Light Conversion), the output of which was sent down a motorized delay stage (Newport), and used as the UV probe beam. A λ/2 wave plate placed before a polarizer was used to control the probe beam intensity. The polarizer was set at 54.7° with respect to the polarization of the pump pulse. The probe was adjusted to have the same intensity at the detector for each wavelength between 240 and 320 nm to avoid nonlinear effects from the detector. All femtosecond TA experiments were performed with a pump wavelength of 266 nm. Pump pulses were produced by third harmonic generation using 20% of the original 800 nm Libra output. An optical chopper (New Focus) was placed in the pump path and referenced to the laser repetition rate, blocking two out of three pulses to limit the pump exposure to the sample. The pump pulse spot size was 5 times larger at the sample than the probe spot size. The probe was passed through a single grating monochromator and detected by a photomultiplier tube. The photomultiplier current was converted to a voltage that was measured by a lock-in amplifier (Stanford Research Systems) synchronized with the optical chopper. The instrument response was roughly 200 fs (fwhm). 2.2. Fluorescence Upconversion. A custom-built apparatus was used to measure ultrafast fluorescence decays using the fluorescence upconversion technique.27 In this setup, ∼150 fs gate pulses were generated using 5% of the 800 nm fundamental output from an amplified titanium−sapphire ultrafast laser system (CPA-1000, Clark-MXR). The remaining 95% was used to generate excitation pulses at the third harmonic of the laser fundamental (266 nm center wavelength) for exciting the sample fluorescence. The gate pulses were delayed with respect to the pump pulses by means of a corner cube retro-reflector mounted on a motorized stage. The pump pulse polarization was set at the magic angle (54.7°) relative to the gate pulse. Fluorescence was collected by a parabolic mirror (25.4 mm diameter, f = 50 mm) and filtered using a 295 nm long pass filter to block residual pump light. A second, identical parabolic mirror focused the emission on a type I BBO crystal cut at 45°. The gate pulse was focused to a slightly larger spot size and overlapped with the emission spot. The BBO crystal was tuned to the proper phase-matching angle, and the resulting upconverted signal was collected with a collimating lens (50.8 mm diameter, f = 50 mm) and passed through a focusing lens ( f = 200 mm) onto the entrance slit of a double grating monochromator. A lock-in amplifier measured the resulting signal from a photomultiplier tube mounted at the exit plane of the monochromator. An optical chopper placed in the gate beam modulated the emission and generated the reference signal for the lock-in amplifier. 2.3. Steady-State Absorption Spectra. Spectra of samples held in 1 mm fused silica cuvettes were recorded using a UV/vis spectrophotometer (PerkinElmer Lambda 25). Absorption spectra were recorded as a function of pH at constant MV2+ concentration by preparing a stock solution of
covalent adduct with a hydroxide ion, but it could also represent a less tightly bound complex, as will be discussed below. It is well-known that photoexcitation increases the Brønsted acidity of many hydroxyaromatic compounds by many orders of magnitude,2 but the notion that electronic excitation can increase the strength of a Lewis acid has not been discussed to our knowledge. In aqueous solution, the hydrolysis of a photoexcited Lewis acid, L+*, could proceed at a much greater rate than ground state hydrolysis L+* + H 2O ⇌ L−OH + H+
(2)
Reaction 2 has been written as a reversible photoreaction in analogy with the common situation for photoacids,2 although this may not be the case. Lewis photoacidity differs from the concept of using light to switch a compound between ground-state forms that differ in their Lewis acidity.17,18 Examples of photoswitchable Lewis acids include a catecholborane that undergoes a change in coordination number of boron upon photoisomerization of an azo ligand17 and a compound that uses a photochromic reaction to modulate the Lewis acidity of a boron center.18,19 Instead, our interest is in the Lewis acidity of a molecule that is in an excited electronic state. A Lewis photoacid should abruptly lower the pH of an aqueous solution upon photoexcitation in the manner of a conventional photoacid, except that the proton that is released to the solvent will be sourced by a water molecule in a hydrolysis reaction rather than from a hydroxyl group that is covalently bonded to the acid. Here, we demonstrate that the diquaternary nitrogen compound methyl viologen (MV2+) is a Lewis photoacid that undergoes hydrolysis in an excited singlet state. Methyl viologen (MV2+) is a renowned electron acceptor that upon one-electron reduction is converted to an unusually stable organic radical (MV•+) with a strong violet color in oxygen-free aqueous solution. MV•+ is readily identified by its characteristic absorption spectrum, which has a sharp, intense peak at 390 nm and a broader, somewhat weaker band at 600 nm.20,21 Peon et al.22 showed that electronic excitation of this good ground-state electron acceptor (E0 = −0.45 V vs NHE)23 produces a powerful excited-state oxidant that can oxidize a nearby solvent molecule. As expected, quenching by photoinduced electron transfer (ET) is correlated with the gas-phase ionization potential (IP) of the donor. Thus, photoexcitation of MV2+ in methanol (IP = 10.84 eV)24 at 266 nm forms MV•+ in 12% yield.22 On the other hand, the high IP of acetonitrile (12.2 eV)25 precludes ET in acetonitrile solution and radical formation is not observed. Instead, MV2+ fluoresces with a lifetime of 1.0 ns.22 In aqueous solution, a puzzling result was observed.22 Although water (IP = 12.62 eV)24 has a higher gas-phase IP than acetonitrile, fluorescence from photoexcited MV2+ is strongly quenched and a lifetime of just 3.1 ps was observed in H2O solution (5.3 ps in D2O) in transient absorption (TA) experiments.22 Peon et al.22 noted that TA signals at visible probe wavelengths decay to zero, but a residual positive signal is observed when probing near 300 nm. Here, we show that this UV-absorbing species is a product of excited-state hydrolysis. In particular, the transfer of electron density from a water molecule to the excited state of MV2+ releases a proton to the bulk solvent, leaving behind a hydroxide ion that is complexed with MV2+ in its electronic ground state. 2738
DOI: 10.1021/jp510953e J. Phys. Chem. B 2015, 119, 2737−2748
Article
The Journal of Physical Chemistry B aqueous MV2+ and dividing it into equal volume aliquots. One aliquot was left unaltered and served as the reference solution. Solid NaOH was added to the other aliquots to achieve the desired OH− concentration while maintaining the same MV2+ concentration as in the reference solution. Difference absorption spectra were then computed as the spectrum at each given pH value minus the reference spectrum at neutral pH. 2.4. Chemicals and Sample Handling. Methyl viologen dichloride (Sigma-Aldrich) was used as received. Reagent grade nitric acid (Fisher) and solid NaOH (Mallinckrodt) were used to adjust the pH of aqueous solutions. Water from a laboratory ultrapurifier was used to prepare unbuffered aqueous solutions. D2O (Sigma-Aldrich) was used as received and had an isotopic purity of 99.8%. Samples for all femtosecond time-resolved measurements were held in a home-built cell with CaF2 windows that was spun by a motor at several hundred rpm about an axis normal to the windows in order to avoid sample re-excitation. The cell path length was approximately 1 mm and the MV2+ concentration was 5 × 10−4 M, corresponding to an absorbance at the pump wavelength of 266 nm of approximately unity. All steady-state and time-resolved measurements were performed at room temperature (T = 21 ± 2 °C). 2.5. Curve Fitting. Transient signals were fit to one or more exponentials using nonlinear least-squares fitting routines in the Igor Pro 6.0 program (WaveMetrics Inc.). Fitting functions were convoluted analytically with a Gaussian function of 200 fs fwhm to model the instrument response function for TA measurements and with a Gaussian of 400 fs fwhm for upconversion measurements. Stated uncertainties are twice the standard deviation determined by the fitting program.
Figure 1. (A) UV absorption spectra of MV2+ (5 × 10−4 M) in aqueous solution at neutral pH (black curve) and as a function of added hydroxide ion (colored curves, same legend as in B). (B) Difference absorption spectra calculated from the curves in panel A. Each curve is the absorption spectrum with added hydroxide ion minus the spectrum recorded at neutral pH.
3. RESULTS 3.1. Steady-State Spectroscopy. The absorption spectrum of MV2+ in aqueous solution is unaffected by added nitric acid but shows modest changes at high concentrations of added base. When the pH is increased at constant MV 2+ concentration, the absorption spectrum decreases slightly near the band maximum (λmax = 257 nm)21 and increases in the red tail above 280 nm (Figure 1A). The absorbance increases rapidly below 230 nm due to absorption by the hydroxide ion. Absorption difference spectra calculated by subtracting the spectrum at neutral pH from the spectrum recorded at elevated pH with identical MV2+ concentration are sigmoidal in shape with extrema at 250 and 292 nm (Figure 1B). 3.2. Transient Spectroscopy. TA and fluorescence upconversion signals were recorded for MV2+ dissolved in methanol and in water upon excitation at 266 nm (Figures 2 and 3). In methanol, the emission at 343 nm decays monoexponentially with a lifetime of 280 ± 80 fs (Figure 2, squares). The significant uncertainty arises because the decay time is somewhat shorter than the instrument response function (∼400 fs fwhm) of our fluorescence upconversion apparatus. Nonetheless, the emission signal at 343 nm clearly decays faster than the TA signal recorded for MV2+ in methanol at 600 nm (Figure 2, circles). The TA signal decays with a lifetime of 490 ± 60 fs in good agreement with the lifetime of 430 ± 40 fs reported by Peon et al.22 In H2O, the emission lifetime at 343 nm increases by an order of magnitude (Figure 3, blue circles). The scaled emission signal exhibits identical dynamics to the TA signal at 600 nm, which has a best-fit
Figure 2. Decay traces for MV2+ in methanol: transient absorption (TA) signal at 600 nm (orange circles) and fluorescence upconversion signal at 343 nm emission (blue squares) with best-fit curves. The fit functions were convoluted with Gaussians of fwhm of 200 and 400 fs for the TA and upconversion signals, respectively.
Figure 3. Decay traces for MV2+ in aqueous solution: transient absorption (TA) signal at 600 nm (orange circles) and emission signal at 343 nm (blue circles) with a best-fit curve to the former signal. The emission transient in methanol from Figure 2 is reproduced for comparison (squares). The fit functions were convoluted with Gaussians of fwhm of 200 and 400 fs for the TA and upconversion signals, respectively.
lifetime of 3.14 ± 0.18 ps in agreement with the value of 3.1 ± 0.2 ps reported in ref 22. The bleach signal for photoexcited MV2+ probed at 250 nm decays more slowly in D2O than in H2O, but a long-lived signal 2739
DOI: 10.1021/jp510953e J. Phys. Chem. B 2015, 119, 2737−2748
Article
The Journal of Physical Chemistry B
MV•+ when ion pairs formed between MV2+ and chloride are photoexcited.28,29 Such ion pairs are favored at high chloride ion concentrations, but there is negligible ion pairing between MV2+ and its chloride counterions for the submillimolar viologen concentrations used in our experiments. For nitric acid concentrations greater than 0.07 M (pH 1.15), decays in the long-time signal become apparent, but the shorttime dynamics are unaltered. The long-time signal decays more rapidly with increasing acid concentration consistent with a bimolecular reaction between a hydronium ion (H+) and the photoproduct. The slowly decaying signal is quenched by acid (Figure 6 and Table S1, Supporting Information), and a
is seen in both solvents (Figure 4). Although the TA signal at 600 nm decays completely to the baseline 20 ps after excitation
Figure 4. Transient absorption kinetics of MV2+ excited at 266 nm and probed at 250 nm in D2O (blue circles) and H2O (red circles) with best-fit curves. The time axis switches from linear to logarithmic at 30 ps, as indicated by the dashed line.
of MV2+ in aqueous solution, residual TA signals are observed at UV probe wavelengths from 240 to 300 nm (Figure 5). The
Figure 6. Transient absorption signals (266 nm pump/250 nm probe) for aqueous MV2+ in the presence of 0.01 M (red), 0.125 M (blue), and 0.5 M (black) nitric acid. The linear scaling of the time axis is different before and after the dashed vertical line at 20 ps.
bimolecular rate constant of (3.1 ± 0.4) × 109 M−1 s−1 was determined by fitting a straight line to a graph of the inverse lifetime vs the hydronium ion concentration (Figure 7). Figure 5. Transient absorption kinetics of MV2+ in aqueous solution at the indicated probe wavelengths following excitation at 266 nm. Global fit curves are shown by the solid lines.
signal at 270 nm recovers approximately to zero, while negative and positive residual signals are seen at shorter and longer wavelengths, respectively. A global fit made by assuming that the signal at each probe wavelength decays monoexponentially to a long-lived offset yielded a lifetime of 3.10 ± 0.12 ps. The TA signal level is remarkably constant from 30 ps to 3 ns, the longest delay time accessible in our femtosecond TA measurements. In fact, the lifetime of this UV-absorbing species for a 5 × 10−5 M solution of MV2+ in H2O held in a 1 cm path length cuvette was determined to be 1.54 ± 0.10 μs using a nanosecond TA apparatus with excitation at 266 nm and probing at 290 nm (Figure S1 in the Supporting Information). The same lifetime was recorded in aerated aqueous solution and in an aqueous solution that had been purged by argon gas for 15 min. Femtosecond TA experiments were carried out on MV2+ in water in the presence of added nitric acid. Acidified solutions were prepared to have identical MV2+ concentrations by dividing a stock solution of ∼1 mM methyl viologen in water into equal volume aliquots and adding the same volume of water with the desired amount of nitric acid to each. Nitric acid was chosen because the high photodetachment threshold of the nitrate ion prevents its oxidation by photoexcited MV2+. Hydrochloric acid was not used because MV2+ is reduced to
Figure 7. Quenching rate (k) of the photoproduct as a function of nitric acid concentration (circles). The slope of the best-fit line is 3.1 × 109 M−1 s−1.
A transient difference spectrum of the long-lived species was recorded with the fs TA apparatus 1 ns after 266 nm excitation by scanning the probe wavelength from 240 to 320 nm while recording the TA signal with the lock-in amplifier every 2.5 nm (black circles in Figure 8). This sigmoidal difference spectrum has a maximum at 296 nm and a minimum at 258 nm. The shape is similar to the derivative of the ground state spectrum of MV2+ with respect to wavelength multiplied by −1, suggesting that the absorbing species has an absorption spectrum that closely resembles that of MV2+ but which is shifted slightly to longer wavelengths. The steady-state spectra shown in Figure 1A were analyzed using the method of Nash,30 assuming that one MV2+ ion and one hydroxide ion form a photoproduct or CT complex with a distinct absorption spectrum. A satisfactory fit was obtained by assuming 1:1 stoichiometry and treatment of additional equilibria was not warranted. The equilibrium constant estimated from fits to the spectra in Figure 1A is 0.4 M−1. 2740
DOI: 10.1021/jp510953e J. Phys. Chem. B 2015, 119, 2737−2748
Article
The Journal of Physical Chemistry B
4.2. The Identity of the UV-Absorbing Photoproduct in Aqueous Solution. The long-lived species only absorbs below ∼340 nm (Figures 5 and 8), explaining why it was not detected in earlier TA experiments conducted at longer probe wavelengths.29,37 Prasad et al.37 reported that no TA signals are seen 100 ns after photolysis of aqueous MV2+ at 266 nm between 350 and 600 nm. Ebbesen et al. observed no TA signals between 400 and 700 nm on picosecond time scales for various MV2+ salts when the MV2+ concentration was low enough (
