Article pubs.acs.org/JPCB
Water-Assisted Vibrational Relaxation of a Metal Carbonyl Complex Studied with Ultrafast 2D-IR John T. King, Matthew R. Ross, and Kevin J. Kubarych* Department of Chemistry, University of Michigan, 930 North University Avenue, Ann Arbor, Michigan 48109, United States S Supporting Information *
ABSTRACT: Water is capable of assisting exceptionally rapid vibrational relaxation within dissolved solute species. Although ultrafast dynamics of metal carbonyl complexes have long served as models for vibrational relaxation, all reports to-date have investigated nonaqueous solutions due to the insolubility of the vast majority of metal carbonyl complexes in water. Using the water-soluble complex [RuCl2(CO)3]2, which belongs to a class known as “carbon monoxide (CO) releasing molecules” (CORM), we report the first ultrafast vibrational relaxation measurements of a metal carbonyl complex in water and compare this relaxation with polar organic solvents, namely, methanol. The vibrational relaxation, measured by twodimensional IR (2D-IR) spectroscopy, is an order of magnitude faster in H2O (3.12 ± 0.29 ps) than in methanol (42.25 ± 3 ps). The accelerated relaxation times of the coupled CO units in H2O and D2O is interpreted as resulting from the enhancement of intramolecular relaxation pathways through additional coupling induced by the solvent. In addition, the vibrational lifetime shows a significant isotope dependence: in D2O the relaxation time is 4.27 ± 0.27 ps, a difference of roughly 30%. We interpret these measurements in terms of a nonresonant channel primarily arising from water’s reorientational dynamics, which occur primarily through large angular jumps, as well as a resonant transfer of vibrational energy from the carbonyl bands to the libration-bend combination band. These measurements indicate that metal carbonyls, which are among the strongest IR transitions, are exquisitely sensitive to the presence of water and hold promise as IR analogs of EPR spin labels. relaxation between two vibrational modes |a⟩ and |b⟩ can be described by second-order perturbation theory in the form of Landau−Teller theory,2,3,16,17 1 ∞ k ba = dt ei ωabt ⟨Vab(t )Vba(0)⟩ (1) h −∞
1. INTRODUCTION The relaxation of a vibrationally excited species is of central importance to chemical dynamics and has been a topic of research for many years.1−4 The extensive role that the solvent can play in relaxation processes is still being explored and a complete description that would allow predictions of relaxation rates is not available. One aspect of vibrational relaxation that appears to be essentially universal is that water facilitates rapid relaxation in solutes by typically an order of magnitude over than similarly polar solvents. Several anions have been studied using ultrafast IR pump−probe spectroscopy including CN−,5,6 SCN−,7,8 SeCN−,9 and N3−.10−13 Likewise, the peptide amide-I band is also found to relax more rapidly in water than in, for example, DMSO.14,15 The use of ultrafast spectroscopies have greatly enhanced our understanding of the complex dynamics of liquid water, including the time scales of hydrogen bond rearranging and the mechanism by which hydrogen bond partners switch. For a relaxing molecule, the hydrogen bonding dynamics, together with the large dipole moment of water, endow water with dynamics that result in rapid modulations of the local electric fields which have significant influences on the relaxation rates of excited molecules. The bath dynamics involved in vibrational relaxation can be treated classically for energy differences comparable to kBT, while the bath dynamics that aid in relaxation of modes with larger energy gaps must be treated quantum mechanically. The © 2012 American Chemical Society
∫
where kba is the rate constant for energy relaxation between the two states and is proportional to the bath spectral density at the frequency difference of the two modes (ωab = ωa − ωb), since it is essentially the Fourier transform of ⟨Vab(t)Vba(0)⟩, the equilibrium correlation function of the coupling between the two modes involved. In the weak coupling limit the coupling potential can be expanded to first order in normal mode coordinates, (0) Vab(t ) = V ab (t ) +
∑ j
(0) Vab(t ) = V ab (t ) +
∂Vab(t ) ∂Q j
Q j + ... Q j= 0
(2)
∑ Fj(t )Q j + ... j
(3)
Received: December 29, 2011 Revised: February 16, 2012 Published: February 29, 2012 3754
dx.doi.org/10.1021/jp2125747 | J. Phys. Chem. B 2012, 116, 3754−3759
The Journal of Physical Chemistry B
Article
where Fj(t) is the force exerted on the relaxing normal mode Qj by the bath. This expansion allows the relaxation rate to be written in terms of a force−force correlation function (ffcf). k ba =
1 h
are obtained by stepping the waiting time (t2) between the first two pump pulses and the third probe pulse. 2.2. Sample Preparation. The solutions of CORM-2 in H2O or D2O were made at 1 mM, while the solutions of CORM-2 in methanol were 5 mM. The samples were sonicated for up to two minutes, and were then used directly for the experiments. Both the FTIR and 2D-IR measurements were performed using CaF2 sample cells with a path length of 100 μm defined by a Teflon spacer.
∞
∫−∞ dt eiωabt ⟨Fj(t )Fj(0)⟩
(4)
Additionally, anharmonic coupling between the solvent and the relaxing mode can act to increase the coupling between the solute’s internal modes, and thus alter intramolecular vibrational redistribution.3,18−20 This mechanism of solvent-assisted relaxation highlights the complicated role in which the solvent may influence vibrational energy relaxation and redistribution. This manuscript presents an experimental study demonstrating the significant role that water can play in dictating the vibrational dynamics of a small ruthenium carbonyl dimer ([RuCl2(CO)3]2, denoted “CORM-2”),21 where the observed relaxation of the coupled CO vibrational modes is shown to be significantly accelerated by water. For molecular excitations that result in only small perturbations to the solvation environment, the relaxation of the excited molecule back into an equilibrium state is related to the equilibrium fluctuations of the solvent, as described by the fluctuation−dissipation theorem and quantified by linear response theory.22 Thus, the accelerated vibrational relaxation of a solute in water reflects the equilibrium dynamics of water itself. Here, we use ultrafast two-dimensional infrared (2D-IR) spectroscopy to study the relaxation of CORM-2 in two hydrogen bonding liquids, methanol and water. The vibrational lifetimes are observed to be markedly shorter in water (both H2O and D2O) than in methanol, illustrating water’s exceptional ability to rapidly dissipate energy from an excited vibrational mode. The sensitivity of the carbonyl lifetime to the presence of water establishes this measure as an effective “water sensor”, which will have implications for studying more complex systems, such as proteins and membranes.
3. RESULTS The linear Fourier Transform IR (FTIR) spectra of CORM-2 in methanol and D2O are shown in Figure 1. (Fits of the FTIR spectra are presented in the Supporting Information). The modes located at 2003 and 2073 cm−1 in D2O (and H2O) are analyzed because of their strong signals and easily accessible frequencies. For the 2D-IR experiments in D2O, spectral overlap with the OD stretch at 2400 cm−1 prohibits analysis of the high-frequency band. These vibrational modes are shifted to 1997 and 2068 cm−1 in methanol. That these modes do not simply shift, and indeed exhibit blue shifts in some cases, indicates that water influences the coupling between the CO local units, possibly serving to localize the vibrational modes.
Figure 1. Linear FTIR spectra of CORM-2 (structure shown in inset) in methanol (black) and D2O (red). The low frequency vibrational mode of CORM-2 (∼2000 cm−1) was analyzed for vibrational lifetime throughout this study.
2. EXPERIMENTAL SECTION 2.1. 2D-IR. The experimental details of 2D-IR have been detailed in the literature.23,24 A regeneratively amplified Ti:Sapphire laser pumps a dual stage optical parametric amplification (OPA) and difference frequency generation (DFG) generates IR pulses centered at 2000 cm−1. The IR pulses are split into E1, E2, E3, and ELO with wavevectors k1, k2, k3, and kLO (100 fs, 150 cm−1 bandwidth, 400 nJ/pulse), where the first three pulses are arranged into a box-car geometry at the sample and the fourth pulse is used for heterodyne detection. The pulses are in ZZZZ polarization, where orientational dynamics contribute to the observed spectral dynamics, although these contributions are minimal given the time scales used in these experiments are shorter than expected reorientational times of the molecules in the solvents. In our experiments, we implement chirped pulse upconversion, where the emitted mid-IR signal and reference local oscillator are mixed in a sum-frequency crystal (5% MgO doped LiNbO3) with a highly chirped pulse centered at 800 nm with a 160 ps duration, to allow detection in the visible spectral region using a 0.5 m grating spectrograph equipped with a silicon CCD camera.25,26 The detected spectrum is then collected as the time delay between the first two pulses, τ1, is continuously scanned. A Fourier transform over this time delay provides the excitation frequency axis for the 2D-IR spectrum. Population dynamics, such as spectral diffusion and vibrational relaxation,
The vibrational lifetime is extracted by measuring the time dependent amplitude of absolute-magnitude rephasing 2D-IR spectra. In contrast to transient absorption or absorptive 2D-IR spectra, using rephasing spectra for lifetime information may have complications arising from differences in the decay rates of excited state absorption and ground state bleach. Nevertheless, we find that the extremely low solubility (∼1 mM) of metal carbonyls in aqueous environments leads to pump−probe signals that are below our detection limit. The background-free Fourier transform 2D-IR signal is significantly stronger, however, and the error that is introduced through the excited state absorption and ground state bleach not being resolved is overshadowed by the higher signal-to-noise ratios achievable with the more intense signal. Figure 2 shows the vibrational relaxation extracted from 2D-IR rephasing spectra of CORM-2 in H2O (2003 and 2073 cm−1 modes, ∼1 mM) and in methanol (1997 and 2068 cm−1 modes, ∼5 mM), indicating significantly different relaxation time scales. The low frequency mode of CORM-2 in methanol has a fast relaxation time of 6.16 ± 1.2 ps followed by a slow relaxation time of 47.36 ± 4 ps. The 2068 cm−1 mode of CORM-2 in methanol shows a similar rate of slow relaxation as the low frequency mode 3755
dx.doi.org/10.1021/jp2125747 | J. Phys. Chem. B 2012, 116, 3754−3759
The Journal of Physical Chemistry B
Article
Figure 2. (left) Relaxation of the low frequency (modes centered around 2000 cm−1) and mid frequency (modes centered around 2060 cm−1) vibrational modes of CORM-2 in methanol and in H2O, showing an order of magnitude acceleration in relaxation in water. (right) An energy level depiction of the vibrational relaxation of the CO stretches is also shown. While the water-assisted intramolecular relaxation channel is likely to be the dominant pathway, there is also a resonant relaxation channel from the CO vibrations directly to the νbend + νlibration combination band of H2O. This pathway is only present in H2O since the combination band if D2O is red-shifted to 1550 cm−1.
Figure 3. Linear FTIR spectra of CORM-2 in D2O and H2O (a), where the combination band of water is shadowed with a green filling. This band is centered at 2150 cm−1 for H2O, but is red-shifted to 1550 cm−1 for D2O. The 2D-IR spectrum of CORM-2 in D2O (b) shows a dominant feature at 2003 cm−1. Because the IR pulses used in this experiment have a bandwidth of roughly 100 cm−1 that is centered at 2000 cm−1, the higher frequency bands appear with significantly less signal amplitude. (c) The vibrational lifetimes of CORM-2 in D2O (4.27 ± 0.27 ps) and H2O (3.12 ± 0.29 ps) of the 2003 cm−1 mode show sub 5-ps relaxation, however relaxation in H2O occurs 30% faster than in D2O. The vibrational relaxation data are also shown with confidence bounds of one and two standard deviations of the fit.
(42.25 ± 3 ps), but its initial decay is significantly faster (2.49 ± 0.5 ps). The fast component of the decay is due to intramolecular vibrational redistribution, a process that can be influenced by the solvent. While IVR rates are difficult to predict based on the molecular structure and the solvent, we have recently observed in the metal carbonyl complex Mn2(CO)10 that IVR rates are sensitive to the degree of hydrogen bonding between the solute and the solvent. In a series of linear alcohols, we found that the presence of hydrogen bonds acts to hinder the IVR among energetically neighboring modes.20 In the present case, it is possible that the 1997 cm−1 band is more delocalized than is the 2058 cm−1 band. The delocalized modes would have greater hydrogen bonding opportunities, possibly resulting in slower IVR due to the hydrogen bond defect-induced vibrational exciton trapping found in Mn2(CO)10. The localized vibrational modes, however, would not have as extensive of hydrogen bonding opportunities, allowing the IVR process to proceed with limited interference from the solvent. Although a detailed examination of IVR is beyond the scope of the present manuscript, work is underway to examine this issue fully, including accurate quantum chemical calculations of this somewhat challenging, flexible complex. The vibrational relaxation of CORM-2 in methanol occurs on the 40−50 ps time scale for the two modes studied here. The relaxation of these modes in aqueous solvents, however,
are found to be an order of magnitude faster. The relaxation of the low frequency mode in H2O shows a relaxation time that is an order of magnitude faster at 3.12 ± 0.29 ps. Likewise, the relaxation of the 2073 cm−1 mode of CORM-2 in H2O shows an identical relaxation rate as found for the 2003 cm−1 mode (fitting procedures described in Supporting Information). In addition, we observe an isotope effect on the relaxation in water and in heavy water. The linear FTIR spectra of CORM-2 in H2O and D2O are shown in Figure 3, where the broad feature centered at 2150 cm−1 in the H2O spectrum is due to the bend+libration combination band of H2O.27 This band is red-shifted to 1550 cm−1 in D2O. The relaxation of CORM-2 in H2O and D2O is found to be 3.12 ± 0.29 ps and 4.27 ± 0.27 ps respectively, showing a 30% acceleration of vibrational relaxation in H2O relative to D2O. Similar isotope effect trends have been reported in previous relaxation studies of small molecule anions in aqueous solution. The mechanism for this observed effect, as well as the possible role that resonant energy relaxation plays in the case of H2O, will be discussed below.
4. DISCUSSION 4.1. Water-Assisted Relaxation. The observation of ultrafast vibrational relaxation in water demonstrates an extreme example of solvent-assisted relaxation, where the solvent enhances the internal couplings of the modes available for relaxation. CORM-2 has a large number (3 × 18 − 6 = 48) 3756
dx.doi.org/10.1021/jp2125747 | J. Phys. Chem. B 2012, 116, 3754−3759
The Journal of Physical Chemistry B
Article
to intermolecular, direct resonant energy transfer to the water combination band. Resonant energy transfer through a dipole coupling, Försterlike mechanism, is dependent on both the excited state lifetime and the degree of spectral inhomogeneity of both the donor and accepting modes.8,33 Though the precise inhomogeneity of the water combination band is not yet known directly from 2DIR spectroscopy, there is simulation34 evidence and indirect experimental indications35,36 that the 130 cm−1 broad band is inhomogeneously broadened. Inhomogeneity limits the efficiency of resonant energy transfer due to conformation-specific spectral overlap. On the other hand, spectral diffusion arising from the dynamic sampling of the inhomogenous band, can aid resonant transfer, and taking the ∼1 ps time scale for water as a proxy for the combination band dynamics, it is possible that spectral diffusion will be capable of overcoming the transient spectral mismatch between the carbonyl donor and the combination band acceptor. The trouble with this mechanism is that the overall lifetimes are so much shorter than what has been observed in CN− where there is no competing intramolecular pathway, yet so similar to N3− where there are at least two intramolecular channels to relax the vibrationally excited asymmetric stretching mode. Recent progress in understanding the hydrogen bonding dynamics of water aids in a mechanistic description of our observation of isotope-dependent relaxation. Hamm et al. used isotope substitution of the CN− solute to partially map the low frequency friction spectrum of H2O and D2O, showing that differences in the friction spectrum, arising from different dynamics, also contribute significantly to the observed isotope effect.6 It has recently been proposed that the hydrogen bonding network of water rearranges not through gradual diffusive processes, but instead via large angular jumps between hydrogen bonding partners. The switching events also result in abrupt motions in the dipole moment of water, which drives relaxation of the carbonyl probe by concomitant fluctuations of the local electric field. The jump events have been proposed,37 and later experimentally validated,38 to occur on the sub-ps to ps time scales, similar to the observed relaxation time of the CO oscillator in CORM-2. These switching events involve large excursions of the hydrogen atoms, and thus should be very sensitive to isotope substitution. Indeed, the Debye relaxation time, a reasonable macroscopic proxy for molecular reorientation, shows an isotope dependence: H2O relaxes 26% faster than does D2O. Our observation of a 30% faster relaxation in H2O is, therefore, consistent with the difference in Debye relaxation.39 Similarly, Kerr effect measurements of solvation dynamics of rhodamine 800 in H2O (6.8 ps) and D2O (10 ps) also indicate a 32% speed-up in water,40,41 which was also attributed to water reorientational dynamics, though the concept of a large angular jump mechanism for hydrogen bond switching events had not yet been identified at the time. The influence of small molecules on hydration dynamics has been a topic of extensive research.42−45 From a structural perspective, it has been shown that the hydrogen bonding network of hydrating waters is not disrupted by small molecules.46 Dynamically, however, the observed slowdown of water’s reorientational relaxation near small molecules is generally understood to be an excluded volume effect,47 influenced by the lack of hydrogen bonding partners capable of accepting hydrogen bond switching events. Small molecules, however, do not disrupt the hydrogen bond switching events when in low concentrations, thus this subset of water dynamics
of internal degrees of freedom, many of which may be coupled to the relaxing mode, serving as a bath of accepting modes to enhance the intramolecular relaxation. These internal modes play a dominant role in the relaxation pathway of the vibrational energy, though their full influence is not realized even in the polar solvent methanol due to the relatively weak inherent anharmonic coupling. We have observed slow vibrational relaxation in several metal carbonyl systems even in polar solvents capable of hydrogen bonding. Mn2(CO)10, for example, exhibits ∼70 ps vibrational relaxation in linear alcohols, showing little chain-length dependence.20 Extremely slow, ∼20 ps IVR has been observed in organometallic complexes [C5H5Ru(CO)2]2 dimers and the iron analogue, [C5H5Fe(CO)2]2.28 Similar time scales have been reported for hydrogenase model compounds as well as its photoproducts.29 A remarkable example of virtually nonexistent IVR can be found in Fe(CO)5, where the axial and equatorial modes separated by only 20 cm−1 show no measurable IVR, though energy is transferred by chemical exchange through the Berry pseudorotation mechanism.30 Several aqueous anions have been studied with ultrafast vibrational spectroscopy. Hamm et al. measured vibrational relaxation of CN− in water to occur between 30 and 80 ps (depending on carbon and oxygen isotope),6 considerably longer than the 3−4 ps time scales found for CORM-2. The significance of additional intramolecular degrees of freedom in facilitating vibrational relaxation is highlighted by the 1−3 ps vibrational lifetime of N3−,10−13,31 which has additional accepting modes as well as a Fermi resonant relaxation channel. The relaxing modes experience enhanced intramolecular coupling in the presence of a rapidly fluctuating local electric field produced by water’s large dipole and its fast, abrupt reorientation dynamics. The extensive hydrogen bonding network, which gives rise to many modes of the solvent bath, also provides a large density of low-frequency states capable of accepting energy that has relaxed, intramolecularly, to the lower tiers of the solute molecule. 4.2. Isotope Effect. The observed isotope effect raises questions about the mechanism by which H2O induces faster vibrational relaxation than D2O. A feature of the CORM-2/ H2O spectrum that is not present in the CORM-2/D2O spectrum is the broad bend+libration combination band centered at 2150 cm−1. This combination band is shifted to 1550 cm−1 in D2O. While there is a resonant solvent band present in H2O, there is also a dynamical difference between the two solvents. In CN−, where the only pathway for relaxation is via intermolecular energy transfer, the isotopespecific differences in vibrational lifetime are substantially larger than the 30% difference we measure for CORM-2, and the absolute lifetimes are much longer.6 In particular, 13C15N− has the same IR transition frequency as CORM-2/water (2003 cm−1), yet relaxes with time constants of 31 ± 7 ps (H2O) and 120 ± 6 ps (D2O), corresponding to a 4-fold enhancement by H2O. Meuwly et al. have used nonequilibrium molecular dynamics simulations of CN− in water to study the influence of nonbonded interactions on the relaxation rate.32 By observing the direct excitations of the bend and librational modes after energy is deposited into the CN− mode, it was concluded that the resonant channel indeed provides an open pathway for energy flow. The slow time scale of relaxation is also similar to the resonant intermolecular relaxation recently reported in SCN− clusters.8 It seems, therefore, that the 3−4 ps relaxation of CORM-2 is an order of magnitude too rapid to be attributed 3757
dx.doi.org/10.1021/jp2125747 | J. Phys. Chem. B 2012, 116, 3754−3759
The Journal of Physical Chemistry B
Article
Figure 4. Cartoons depicting hydration of a small molecule at low (left) and high (right) concentrations. At low concentrations the solvation shell surrounding one solute molecule is not influenced by neighboring solute molecules nor their solvation shells. In concentrated solutions, however, water molecules can become confined between several solute molecules, leading to a restriction of water’s dynamics via coupled excluded volume effects.
■
should remain relatively unperturbed. A key feature of this work is that we are studying the water dynamics via a dissolved solute, and hence are able to use very low solute concentrations (Figure 4). In contrast to previous work cited above using molar-level solute concentrations, CORM-2 is present in solution at the millimolar level (∼1 mM for D2O, ∼ 5 mM for methanol). Such a low concentration eliminates any collective influence on the water dynamics arising from neighboring solute species, permitting direct sensitivity to the isolated solute’s hydration shell.
ASSOCIATED CONTENT
* Supporting Information S
Fits to FTIR spectra and experimental fits to lifetime data. This material is available free of charge via the Internet at http:// pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected]. Notes
5. CONCLUSIONS The pronounced difference in vibrational excited state lifetime of excited CO stretching modes in polar organic solvents as opposed to aqueous solution establishes the coupled CO vibrational modes of metal carbonyl complexes as effective water sensors. The ultrafast relaxation observed in water can be explained in terms of the discrete large angular jump dynamics of H2O, which induce fluctuations in the hydrogen bonding network and produce rapid modulations of the local electric field surrounding the relaxing oscillator. The observed isotope effect is rationalized in terms of the isotope dependence of hydrogen bonding jumping events, which contribute to water’s dynamics on the picosecond time scale. Since these events involve large displacements of the hydrogen/deuterium atoms, they should be particularly sensitive to the isotope substitution. This ultrafast isotope effect will prove beneficial for studying the disruption of the dynamics of hydrogen bond rearrangements, where the influence of isotope substitution on the relaxation rates can provide insights into the extent to which this subset of water dynamics is perturbed by a solute. Characterizing the behavior of metal carbonyl complexes in aqueous environments is an important step in using these strong IR absorbers as vibrational probes of more complex systems such as proteins and membranes. The vibrational relaxation of these probes can act as an effective water sensor, and the above results suggest that the isotope dependent relaxation can provide a local window into the angular jump dynamics of hydrating water. Reports that are forthcoming will describe how these probes can be placed on molecules such as proteins to study the influence of hydrating water at the protein−water interface.
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS This work was supported by the National Science Foundation (CHE-0748501) and the Camille & Henry Dreyfus Foundation.
■
REFERENCES
(1) Oxtoby, D. W. Annu. Rev. Phys. Chem. 1981, 32, 77−101. (2) Elsaesser, T.; Kaiser, W. Annu. Rev. Phys. Chem. 1991, 42, 83− 107. (3) Owrutsky, J. C.; Raftery, D.; Hochstrasser, R. M. Annu. Rev. Phys. Chem. 1994, 45, 519−555. (4) Skinner, J. L. Theor. Chem. Acc. 2010, 128, 147−155. (5) Heilweil, E. J.; Doany, F. E.; Moore, R.; Hochstrasser, R. M. J. Chem. Phys. 1982, 76, 5632−5634. (6) Hamm, P.; Lim, M.; Hochstrasser, R. M. J. Chem. Phys. 1997, 107, 10523−10531. (7) Ohta, K.; Tominaga, K. Chem. Phys. Lett. 2006, 429, 136−140. (8) Bian, H.; Wen, X.; Li, J.; Chen, H.; Han, S.; Sun, X.; Song, J.; Zhuang, W.; Zheng, J. Proc. Natl. Acad. Sci. U.S.A. 2011, 108, 4737− 4742. (9) Lenchenkov, V.; She, C.; Lian, T. J. Phys. Chem. B 2006, 110, 19990−19997. (10) Morita, A.; Kato, S. J. Chem. Phys. 1998, 109, 5511−5523. (11) Hamm, P.; Lim, M.; Hochstrasser, R. M. Phys. Rev. Lett. 1998, 81, 5326−5329. (12) Zhong, Q.; Baronavski, A. P.; Owrutsky, J. C. J. Chem. Phys. 2003, 118, 7074−7080. (13) Sando, G. M.; Dahl, K.; Owrutsky, J. C. J. Phys. Chem. B. 2007, 111, 4901−4909. (14) Rubtsov, I. V.; Wang, J. P.; Hochstrasser, R. M. J. Phys. Chem. A 2003, 107, 3384−3396.
3758
dx.doi.org/10.1021/jp2125747 | J. Phys. Chem. B 2012, 116, 3754−3759
The Journal of Physical Chemistry B
Article
(15) DeFlores, L. P.; Ganim, Z.; Ackley, S. F.; Chung, H. S.; Tokmakoff, A. J. Phys. Chem. B 2006, 110, 18973−18990. (16) Grote, R. F.; Hynes, J. T. J. Chem. Phys. 1982, 77, 3736−3743. (17) Egorov, S. A.; Skinner, J. L. J. Chem. Phys. 1996, 105, 7047− 7058. (18) Berg, M.; VandenBout, D. A. Acc. Chem. Res. 1997, 30, 65−71. (19) Bonner, G. M.; Ridley, A. R.; Ibrahim, S. K.; Pickett, C. J.; Hunt, N. T. Faraday Discuss. 2010, 145, 429−442. (20) King, J. T.; Anna, J. M.; Kubarych, K. J. Phys. Chem. Chem. Phys. 2010, 13, 5579−5583. (21) Motterlini, R.; Otterbein, L. E. Nat. Rev. Drug Discovery 2010, 9, 728. (22) Chandler, D. Introduction to Modern Statistical Mechanics; Oxford University Press: New York, 1987. (23) Khalil, M.; Demirdoven, N.; Tokmakoff, A. J. Phys. Chem. A 2003, 107, 5258−5279. (24) Jonas, D. M. Annu. Rev. Phys. Chem. 2003, 54, 425−463. (25) Anna, J. M.; Nee, M. J.; Baiz, C. R.; McCanne, R.; Kubarych, K. J. J. Opt. Soc. Am. B, 27, 382−393. (26) Nee, M. J.; McCanne, R.; Kubarych, K. J.; Joffre, M. Opt. Lett. 2007, 32, 713−715. (27) Ratcliffe, C. I.; Irish, D. E. J. Phys. Chem. 1982, 86, 4897−4905. (28) Anna, J. M.; King, J. T.; Kubarych, K. J. Inorg. Chem. 2011, 50, 9273−9283. (29) Kaziannis, S.; Wright, J. A.; Candelaresi, M.; Kania, R.; Greetham, G. M.; Parker, A. W.; Pickett, C. J.; Hunt, N. T. Phys. Chem. Chem. Phys. 2011, 13, 10295−10305. (30) Cahoon, J. F.; Sawyer, K. R.; Schlegel, J. P.; Harris, C. B. Science 2008, 319, 1820−1823. (31) Olschewski, M.; Knop, S.; Lindner, J.; Voehringer, P. J. Chem. Phys. 2011, 134, 214504. (32) Lee, M. W.; Meuwly, M. J. Phys. Chem. A 2010, 115, 5053− 5061. (33) Woutersen, S.; Bakker, H. J. Nature 1999, 402, 507−509. (34) Yagasaki, T.; Ono, J.; Saito, S. J. Chem. Phys. 2009, 131, 164511. (35) Chieffo, L. R.; Shattuck, J. T.; Pinnick, E.; Amsden, J. J.; Hong, M. K.; Wang, F.; Erramilli, S.; Ziegler, L. D. J. Phys. Chem. B 2008, 112, 12776−12782. (36) Ingrosso, F.; Rey, R.; Elsaesser, T.; Hynes, J. T. J. Phys. Chem. A 2009, 113, 6657−6665. (37) Laage, D.; Hynes, J. T. Science 2006, 311, 832−835. (38) Ji, M.; Odelius, M.; Gaffney, K. J. Science 2010, 328, 1003−1005. (39) Schwartz, B. J.; Rossky, P. J. J. Chem. Phys. 1996, 105, 6997− 7010. (40) Zolotov, B.; Gan, A.; Fainberg, B. D.; Huppert, D. J. Lumin. 1997, 72−4, 842−844. (41) Zolotov, B.; Gan, A.; Fainberg, B. D.; Huppert, D. Chem. Phys. Lett. 1997, 265, 418−426. (42) Rezus, Y. L. A.; Bakker, H. J. Phys. Rev. Lett. 2007, 99, 148301− 148304. (43) Bakulin, A. A.; Liang, C.; Jansen, T. L. C.; Wiersma, D. A.; Bakker, H. J.; Pshenichnikov, M. S. Acc. Chem. Res. 2009, 42, 1229− 1238. (44) Bakulin, A. A.; Pshenichnikov, M. S.; Bakker, H. J.; Petersen, C. J. Phys. Chem. A 2011, 115, 1821−1829. (45) Stirnemann, G.; Hynes, J. T.; Laage, D. J. Phys. Chem. B 2010, 114, 3052−3059. (46) Chandler, D. Nature 2005, 437, 640−647. (47) Laage, D.; Stirnemann, G.; Hynes, J. T. J. Phys. Chem. B 2009, 113, 2428−2435.
3759
dx.doi.org/10.1021/jp2125747 | J. Phys. Chem. B 2012, 116, 3754−3759
