Following the solvent directly during ultrafast excited state proton transfer

Article abstract of DOI:10.1021/ja048639g

Excited state intramolecular proton transfer in 1-chloroacetylaminoanthraquinone is investigated from the perspective of the solvent. Using a new two-dimensional nonlinear optical spectroscopy the solvent response is probed directly as the proton transfer takes place. The measurements indicate that solvent reorganization controls the proton transfer in acetonitrile by dynamically shifting the position of equilibrium in the excited state, even on subpicosecond time scales. Copyright

Full text of DOI:10.1021/ja048639g

Published on Web 06/25/2004
Following the Solvent Directly during Ultrafast Excited State Proton Transfer
Sarah J. Schmidtke, David F. Underwood, and David A. Blank*
Department of Chemistry, UniVersity of Minnesota, Minneapolis, Minnesota 55455
Received March 9, 2004; E-mail: blank@chem.umn.edu
Proton transfer, PT, is an important process in numerous areas
of chemistry. The static effect of a polar solvent environment on
the position of equilibrium has long been appreciated. More recent-
ly, a great deal of attention has been given to understanding the
dynamic role played by the solvent.1-3 Experimentally, studies have
been limited to measuring the rate of the appearance (disappearance)
of the products (reactants). Conclusions concerning the solvent’s
role in the reaction are often inferred from a comparison between
the rate of PT and known time scales associated with the solvent.
Following the dynamics from the perspective of the reactants and
products has certainly provided significant advancement of our
understanding in this important area of chemistry.1,2,4-16
To better understand how the solvent participates in chemical
events it is advantageous to probe the response of the solvent
Figure 1. Local solvent response following excitation of three AQs in
directly. In addition to providing a direct look at the solvent’s role
in a reaction, following a reaction via the solvent response can offer
a new perspective on the dynamics of the reaction itself. For
example, although the spectroscopy of a given reactant and product
are often well characterized, mapping observed resonant transitions
to intermediate, transient reaction structures can be poorly under-
stood due to overlap between transitions on both the reactant and
product side of the transition state.7 In such cases, the solvent can
report on the reaction progress in a way that does not require
detailed knowledge of the spectroscopy of transient intermediates.
We recently developed a new technique to follow the dynamics
from the perspective of the solvent, directly probing the response
of the local solvent environment during a dynamic event in
solution.17 Here we report the application of this approach to a
reactive system in solution for the first time, directly probing the
solvent’s role in excited state intramolecular proton transfer, ESIPT.
In this example we examine the role of the solvent, acetonitrile,
in controlling ESIPT in 1-chloroacetylaminoanthraquinone, CAAQ.
The structures of the normal, 1N, and tautomeric, 1T, forms of
CAAQ are shown in Figure 1. In the ground electronic state, S0,
1N is the stable form. Following excitation to the first excited
electronic state, S1, ESIPT takes place between 1N and 1T. In an
extensive set of investigations, Barbara and co-workers demon-
strated that the ESIPT rate in a series of 1-(acylamino)anthraquino-
nes could be tuned from nearly instantaneous to absent by
substitution of acyl groups with larger or smaller electron affinity,
respectively.7 CAAQ was found to be an intermediate case in
dipolar solvents, with the majority of the static fluorescence
originating from 1N (S1), and a measured appearance time for 1T
(S1) of <300 fs. With higher time resolution, Neuwahl et al.
determined an appearance time of 110 fs for 1T (S1).13
acetonitrile. The solid line is the fit to the CAAQ data described by eq 1,
representing the response of the solvent to the ESIPT dynamics in the S1
state. TFAQ and HPAQ are limiting cases of rapid and complete ESIPT
and no ESIPT, respectively. At the right are the normal and tautomer
structures.
associated with librational motion of the solvent molecules.18 The
instrument response along t is Gaussian in time with fwhm ) 140
fs, and we estimate our time resolution to be 80 fs.
Figure 1 shows the change in the Raman response of the local
solvent environment following resonant excitation of CAAQ. The
data were fit with the sum of three components. To emphasize the
response of the solvent to the reactive dynamics in the S1 state two
of the fitting components, a 16 ps exponential decay due to rotation
of CAAQ and the background measured from a sample of neat
acetonitrile under identical conditions, have been subtracted from
the data presented in Figure 1. The raw data and complete fit are
presented in the Supporting Information. The data in Figure 1 have
been fit as a series of three steps,
(-7.3)A 100 fs8 (17)B 180 fs8(4.5)C 1.1 ps8 (-10)D (1)
Equation 1 was convoluted over the instrument response function,
and the time scales above the arrows and weighting factors in par-
entheses were varied to obtain the best fit to the data in Figure 1.
The measured response reflects the change in the polarizability
of the intermolecular motions in the local solvent environment
following excitation of CAAQ. For reference the measured
responses for two other 1-(acylamino)anthraquinones, TFAQ and
HPAQ, are also presented, see Figure 1. TFAQ undergoes ballistic
and complete PT in the excited state, and HPAQ does not undergo
PT in the excited state.7 These two limiting cases demonstrate that
there is a positive change in the solvent response when undergoing
PT to 1T(S1) and a negative change in the response when exciting
to 1N(S1) in the absence of subsequent PT. This observation can
be explained in terms of the change in the solvent-solute interaction
following solute excitation, with the dipole moment increasing
relative to 1N(S0) upon transition to 1T(S1) and decreasing relative
to 1N(S0) upon transition to 1T(S1).20 The details of this effect will
be addressed elsewhere.23 For the purpose of discussion here we
Details of the experiment can be found in reference 17 and the
Supporting Information. A laser pulse at 400 nm excites CAAQ to
the reactive S1 state at time t ) 0. Three 800 nm laser pulses probe
the change in the electronically nonresonant Raman response of
the solvent molecules as a function of time, t. A delay intrinsic to
the Raman probe, 140 fs between the first two and third probe
pulses, is set near the maximum of the intermolecular nuclear
response of the solvent, a region of the response that is often
9
8620
J. AM. CHEM. SOC. 2004, 126, 8620-8621
10.1021/ja048639g CCC: $27.50 © 2004 American Chemical Society

C O M M U N I C A T I O N S
will limit the assignment of the solvent response to time-dependent
changes in the ratio of 1T/1N. An increase in the response indicates
progress toward the product (1T), and a decrease in the response
indicates progress toward the reactant (1N), depicted vertically on
the right side of Figure 1.
Upon excitation, 1N is placed in the S1 state, which is the starting
point labeled A in eq 1. Impulsive excitation from 1N(S0) to 1N(S1)
at t ) 0 is reflected in the negative weight for A. There is an initial
100 fs rise in the solvent response, consistent with the 110 fs
appearance time measured for 1T(S1).13 This is followed by a de-
crease in the signal with time scales of 180 fs and 1.1 ps. The
decrease in the 1T/1N ratio indicates a shift back toward 1N on the
S1 surface. The fact that the solvent response falls to a negative
value in the first 1-2 ps and remains negative is consistent with
the static fluorescence spectrum, which is dominated by emission
from 1N.7
The effect of the solvent on reactive dynamics is typically
considered in the equilibrium limit, where the solvent responds
rapidly compared to progress along the reaction coordinate, and one
envisions a single minimum free energy path from reactants to pro-
ducts.3 In this limit it is the barrier along the motion of the proton
from the donor to acceptor site that determines the rate of proton
transfer, and for intermolecular PT this can be strongly influenced by
small structural changes in reactants.13 If this limit applied to ESIPT
in CAAQ, then we would expect to observe a single rise in the sol-
vent response as equilibrium is established starting from the initially
prepared 1N(S1). However, the most striking feature of the CAAQ
response in Figure 1 is that following initial progress toward the
product the reaction turns around and proceeds back toward the
reactant. The fact that the reaction initially proceeds from 1N to 1T
and then returns to 1N in the first 1-2 ps clearly indicates that on
the time scale of this reaction there is not a single, solvated reaction
coordinate. Rather the dynamics are dictated by a time-dependent
free energy path between 1N and 1T that evolves as the solvent re-
organizes. We note that in the reported time-resolved stimulated
emission from 1T(S1) there is no evidence of the return to 1N(S1).13
The authors present only the first 600 fs after excitation, and we
speculate that within this time frame the overlapping emission from
different points along the PT, as discussed by Barbara and co-
workers,7 masks the start of the return event in these measurements.
This potential complication is avoided when probing the solvent.
In a detailed theoretical study, Kiefer and Hynes recently
considered the limit where PT is rapid on the time scale of solvent
reorganization, and the reaction rate is determined by dynamic
solvent reorganization.19 In this limit the relative free energy of
the reactants and products becomes time dependent as the solvent
configuration evolves. This limit is represented schematically for
ESIPT in CAAQ in Figure 2. We calculate that in the S1 state 1N
and 1T are nearly isoenergetic based on vertical excitation from
the ground state.20 At t ) 0, CAAQ is excited to the S1 state, but
the solvent is still in the initial equilibrium configuration established
around the ground state, 1N(S0). In the long time limit, static
fluorescence indicates greater stabilization of 1N(S1) than 1T(S1).6,7
This is consistent with the larger change in the magnitude and
direction of the dipole moment going from 1N(S0) to 1N(S1) verses
1T(S1).20 In the first 100 fs there is an increase in the 1T/1N ratio
from zero as equilibrium along the PT coordinate is established.
As the solvent reorganizes, the relative free energy of 1N versus 1T
falls, and equilibrium shifts back toward 1N with a bimodal response
consistent with measured solvation dynamics in acetonitrile.21,22 In
this system the local solvent motions play the lead role in dictating
the dynamics of excited state proton transfer. A more complete
accounting of this work will include the ability of the technique to
Figure 2. Normal and tautomer structures for CAAQ . Schematic
representation of the proton transfer coordinate in the excited S1 state.
Lowering of the energy due to solvent reorganization is depicted relative
to the equilibrium solvent configuration around the electronic state of 1N,
t ) 0.
spread the solvent response into a second time dimension, allowing
a detailed analysis of the specific solvent coordinate(s) that
participate as the PT proceeds from reactants to products and back.23
Acknowledgment. This work was supported by NSF under
Award Number CHE-0211894. This work was supported in part
by the MRSEC Program of the NSF under Award Number DMR-
0212302. We thank the donors of the Petroleum Research Fund
for support of this work. D.B. thanks the Camille and Henry Dreyfus
Foundation, the David and Lucille Packard Foundation, and 3M
Company for their generous support. S.J.S. thanks the NSF for
Fellowship support and the Minnesota Supercomputing Institute
for computing time and resources.
Supporting Information Available: Details of the experiment,
fitting, and calculated values referred to in the text. This material is
available free of charge via the Internet at http://pubs.acs.org.
References
(1) Barbara, P.; Walker, G.; Smith, T. Science 1992, 256, 975-981.
(2) Elsaesser, T.; Bakker, H., Eds. Ultrafast Hydrogen Bonding Dynamics
and Proton Transfer Prosesses in the Condensed Phase; Kluwer Academic
Publishers: Boston, 2002.
(3) Voth, G.; Hochstrasser, R. J. Phys. Chem. 1996, 100, 13034-13049.
(4) Barbara, P.; Trommsdorff, H. P. E. Chem. Phys. 1989, 136, 153-360.
(5) Douhal, A.; Lahmani, F.; Zewail, A. Chem. Phys. 1996, 207, 477-498.
(6) Smith, T.; Zaklika, K.; Thakur, K.; Barbara, P. J. Am. Chem. Soc. 1991,
113, 4035-4036.
(7) Smith, T.; Zaklika, K.; Thakur, K.; Walker, G.; Tominaga, K.; Barbara,
P. J. Phys. Chem. 1991, 95, 10465-10475.
(8) Taylor, C.; El-Bayoumi, M.; Kasha, M. Proc. Natl. Acad. Sci. U.S.A. 1969,
63, 253-60.
(9) Moog, R.; Maroncelli, M. J. Phys. Chem. 1991, 95, 10359-69.
(10) Chou, P.-T.; Liu, Y.-I.; Liu, H.-W.; Yu, W.-S. J. Am. Chem. Soc. 2001,
123, 12119-12120.
(11) Das, K.; English, D.; Petrich, J. J. Am. Chem. Soc. 1997, 119, 2763-
2764.
(12) Herbich, J.; Hung, C.-Y.; Thummel, R.; Waluk, J. J. Am. Chem. Soc.
1996, 118, 3508-3518.
(13) Neuwahl, F.; Bussotti, L.; Righini, R.; Buntinx, G. Phys. Chem. Chem.
Phys. 2001, 3, 1277-1283.
(14) Lochbrunner, S.; Wurzer, A.; Riedle, E. J. Phys. Chem. A 2003, 107,
10580-10590.
(15) Hoefer, T.; Kruck, P.; Elsaesser, T.; Kaiser, W. J. Phys. Chem. 1995, 99,
4380-4385.
(16) Waluk, J. Acc. Chem. Res. 2003, 36, 832-838.
(17) Underwood, D.; Blank, D. J. Phys. Chem. A 2003, 107, 956-61; ibid 9736.
(18) Maroncelli, M.; Kumar, V.; Papazyan, A. J. Phys. Chem. 1993, 97, 13-17.
(19) Kiefer, P.; Hynes, J. J. Phys. Chem. A 2002, 106, 1834-1849.
(20) See Supporting Information
(21) McMorrow, D.; Lotshaw, W. J. Phys. Chem. 1991, 95, 10395-406.
(22) Rosenthal, S.; Xie, X.; Du, M.; Fleming, G. J. Chem. Phys. 1991, 95,
4715-18.
(23) Manuscript in preparation.
JA048639G
9
J. AM. CHEM. SOC. VOL. 126, NO. 28, 2004 8621