вход по аккаунту


Ultrafast Solvation of N-Methyl-6-quinolone Probes Local IR Spectrum.

код для вставкиСкачать
Femtosecond Spectroscopy
DOI: 10.1002/anie.200501397
Ultrafast Solvation of N-Methyl-6-quinolone
Probes Local IR Spectrum**
J. Luis Prez Lustres, Sergey A. Kovalenko,
Manuel Mosquera, Tamara Senyushkina,
Wolfgang Flasche, and Nikolaus P. Ernsting*
Solvation is a matter of central interest in chemistry. It
influences the energetics of chemical reactivity because the
reactants, transition state, and products may be stabilized
differently by the solvent, and it creates frictional forces for
the dynamics. Here we consider charge-transfer reactions in
polar media. Usually the polarity of the environment is in
equilibrium with the charge distribution along the reaction
path. But ultrafast reactions may even ?leave the solvent
behind?, the polarity can no longer stay adjusted, and the
energy barrier is crossed in a nonequilibrated environment.[1]
The time behavior of adjustment, or ?polar solvation
dynamics?,[2?10] is a key to understanding many reactions in
the condensed phase. In liquids at room temperature,
individual solvent molecules respond to charge redistribution
on a timescale of 10?100 femtoseconds (fs). The collective
answer comes slower, and restructuring of solvent shells may
take picoseconds or longer.
Measurements of polar solvation dynamics aim to observe
the process directly, independent of reactive motion, and for
this purpose one needs a suitable molecular probe. Usually
these are solvatochromic dye molecules whose color depends
strongly on the surrounding because the
ground and excited electronic states interact
differently with it. Using the novel probe Nmethyl-6-quinolone (MQ), we show that its
solvation in water and methanol is reproduced
quantitatively, both in magnitude and time
behavior, by the bulk dielectric dispersion e(n) of the liquid,
the connection being continuum theory.[11?16] The surprising
ability of a simple model to describe the dynamics was already
[*] Dr. J. L. Prez Lustres, Dr. S. A. Kovalenko, Dr. T. Senyushkina,
W. Flasche, Prof. N. P. Ernsting
Department of Chemistry, Humboldt University
Brook-Taylor-Strasse 2, 12489 Berlin (Germany)
Fax: (+ 49) 30-2093-5553
Prof. M. Mosquera
Department of Physical Chemistry
University of Santiago de Compostela
15782 Santiago de Compostela (Spain)
[**] This work was supported by the Deutsche Forschungsgemeinschaft,
the Volkswagen Foundation, the Fonds der Chemischen Industrie,
and the Xunta de Galicia. We also thank John E. Bertie (University of
Alberta, Edmonton, Canada) for providing the optical constants of
the investigated solvents.
Supporting information for this article is available on the WWW
under or from the author.
Angew. Chem. Int. Ed. 2005, 44, 5635 ?5639
noted.[8, 11, 13] It is now substantiated with the detection of
solvent oscillations which modulate the emission frequency.
The match of observed, time-resolved solvation with
prediction from continuum theory is at once puzzling and a
chance. For one, the solvation process is microscopic by its
very nature, as shown by numerous molecular dynamics
simulations. The solute disturbs the liquid structure, and
liquid dynamics around the probe is therefore expected to
deviate from the bulk. It is far from trivial to relate the
solvation response of a chromophore to solvent properties
such as e(n), and this problem has been addressed by
statistical theories.[17, 18] Deviations from continuum theory
are termed ?molecularity of solvation?.[14] The new dye does
not exhibit molecularity towards collective solvent motion at
the accuracy of present experiments, and this could lead to
specific applications in characterizing water dynamics in
restricted or biological environments on molecular dimensions. For this purpose MQ has several advantages over other
probes. It is not charged, hence does not polarize the
surrounding more than is needed for measurement. The
solvent polarity couples mainly through dipole moment
change, allowing straightforward evaluation of experiments.
Most importantly, there is no interfering solute motion at
frequencies relevant for aqueous solvation. Finally, the
molecule is small enough to replace natural bases in nucleic
acids or tryptophan in proteins. If continuum theory holds in
such environments, femtosecond solvation measurements will
provide the local IR spectrum.
Water and methanol are used here as benchmark solvents.
Water deserves attention owing to its unique features. It can
stabilize polar reactants within 100 fs and thus is the ?fastest?
solvent known. The interest in water extends from the
properties of the pure solvent to its behaviour in complex
environments[19?21] and especially to biopolymers where
?biological water?[21] is expected to be involved in important
primary reaction steps. Methanol is interesting because of its
amphiphilic character. For both solvents, e(n) is known to high
accuracy, and conclusions about the ability of continuum
theory to describe MQ solvation can be drawn.
Femtosecond solvation experiments are explained in
Figure 1. The molecule is suddenly promoted into the S1
state by a short laser pulse. The ensuing relaxation of the
excited state in terms of the free energy gap DE(t) =
E1(t)E0(t) is central to optical spectroscopy in the condensed
phase. How DE(t) is accessed by time-resolved interaction of
femtosecond optical pulses depends on the nonlinear optical
signal that is detected. The fluorescence band provides
probably the most direct view of relaxation in the excited
state. Its time-resolved dynamic Stokes shift is a linear image
of the solvation process, provided that intramolecular reorganization is either absent because the molecular structures in
S1 and S0 are the same, or else avoided by excitation into S1
without excess energy. Here we monitor stimulated emission
which is strictly related to spontaneous fluorescence.
The continuum theory of polar solvation is also sketched
in Figure 1. Consider first the pure liquid in an external
electric field (not shown) which partially orients the solvent
molecules, generally represented by their dipole moments
(arrows). Continuum theory replaces the microscopic liquid
2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Here n and ncav are the refractive indices of the solvent
and the cavity, respectively, due to electronic response
at optical frequencies. The latter is related to the
polarizability a1 of the solute in the S1 state through the
Clausius?Mosotti equation [Eq. (2)]
n2 1
╝ cav
r3 n2cav ■ 2
Equation (1) is needed below to calculate the timedependent Stokes shift.
The absorption and stationary stimulated emission
spectra of MQ in water are shown in Figure 2 a.
Femtosecond transient absorption was measured with
the pump?supercontinuum?probe technique. Typical
transient spectra are shown in Figure 2 b. They have
three contributions: excited-state absorption (Sn S1,
ESA), stimulated emission (S1!S0, SE), and negative
absorption due to the ground-state hole (S1 S0,
bleach). The time-dependent red shift of the SE band
is easily recognized together with a blue shift of the
ESA band at 350 nm.
The stimulated emission band at delay time t is
extracted from the corresponding transient absorption
spectrum, and the peak frequency n(t) is adopted as a
measure of the free-energy gap. n(t) is the desired
time-resolved Stokes shift. The experiment thus provides a normalized relaxation function [Eq. (3)]
Figure 1. The dynamic Stokes shift of fluorescence (yellow?red band) images the relaxation of the surrounding solvent in the S1 state of the solute. Solvent coordinate Q represents the configurational state of the environment and is treated classically. High-frequency motion along internal coordinate q for bond lengths and angles is quantized in
vibrational states. Before femtosecond excitation (a) the solute is in the equilibrated
ground state S0 and the (blue) absorption band extends over a Franck?Condon progression for upward optical transitions. Immediately after excitation (b) the emission overlaps the absorption band at the electronic absorption origin 00. Partially relaxed E1(Q)
(c) corresponds to solvent configurations that have raised E0(Q). The fluorescence therefore changes from yellow to red as solvation proceeds. After several picoseconds (d) a
new equilibrium is reached for the S1 state. A point dipole in a spherical polarizable
cavity represents the solute, while a continuum with dielectric dispersion e(n) represents
the surrounding liquid.
structure with an average, smooth, dipole density?the polarization P. The latter responds to changes in the field, and from
such measurements at frequencies n in the microwave, farinfrared, and infrared spectral regions one obtains the
polarization susceptibility c(n) = e(n)1, where e(n) is the
dielectric dispersion. For solvation, the field is instead derived
from the charge distribution of the chromophore (Figure 1 a).
It polarizes the surrounding medium, which in turn creates a
reaction field R which stabilizes the solute. Analogous to the
previous consideration, R responds to changes in the solute
charge distribution at frequency n. Equivalently the charge
distribution may be switched from one form into another by
femtosecond excitation S1 S0, and the solvent configuration
and R are thus driven towards a new equilibrium.
Altogether the free energy of solvation can be calculated.
In continuum theory a cavity is carved out of the solution,
with the solute charge distribution inside and a homogeneous
medium characterized by e(n) outside. A simple version is
obtained when the cavity is taken to be spherical with radius r
and the solute charge distribution is approximated by its
dipole moment m at the center. In this case the susceptibility
of the reaction field to changes in the dipole moment[11, 12] is
given by Equation (1).
Sn ­tя n­tяn­1я
cdip ­nя ╝
2­n2cav ■ 2я
3 r3
n2 1
2e­nя ■ ncav 2 n ■ n2cav
Figure 2. Optical spectra of MQ in water (pH 9). a) Stationary absorption (Abs) and stimulated emission (SE). b) Femtosecond transient
absorption recorded with a time resolution of 30 fs after excitation at
430 nm. Induced change of optical density DOD < 0 indicates stimulated emission in the fluorescence region and bleach in the absorption
region where excited-state absorption (ESA) dominates. Transient
spectra between 0.05 and 1 ps are shown in 0.1-ps steps and between
1 and 6 ps in 1-ps steps.
2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2005, 44, 5635 ?5639
Here n(0) is the peak emission frequency immediately after
excitation and n(1) the frequency at a later time; the full
Stokes shift is Dn = n(0)n(1).
To calculate the time-dependent Stokes shift from dielectric dispersion we consider the energy gap DEdip(t) of an ideal
dipolar chromophore in a cavity (Figure 1 a). Equilibrium
solvent motion causes fluctuations dDEdip of the energy gap,
and the classical correlation function is given by Equation (4)
Sdip ­tя hdDEdip ­0я dDEdip ­tяi
hdDEdip ­0я dDEdip ­0яi
where hi is the ensemble average. The same time behavior is
expected, essentially, for the nonequilibrium relaxation of
DEdip(t). The full spectral density of the correlation function
for the energy gap is accessible through the fluctuation?
dissipation theorem [Eq. (5)] (see the Supporting Information):
Cdip ­nя ╝
1 ■ coth
2 kB T
Finally, the quantity defined in Equation (4) is given a
theoretical curve [Eq. (6)]
Sdip ­tя /
dn Cdip ­nя cos­2pntя
which can be compared to experimental Sn(t).
The procedure is illustrated for water in Figure 3.[11, 22, 23]
Our main observation is that the measured relaxation of MQ
in water is identical, within measurement error, with the
calculated dipole solvation relaxation. For the moment let us
treat this as an hypothesis and replace Sn(t) in Equation (3) by
Sdip(t) to obtain Equation (7)
n­tя ╝ Dn Sdip ­tя ■ n­1я
where the full Stokes shift is given by Equation (8)
Dn ╝ cdip ­0яm1 ­m1 m0 я=h c
(in wavenumbers, with c speed of light, h PlanckBs constant,
and molecular dipole moments mi of MQ). In order to prove
the hypothesis the theoretical curve [red line, Eq. (7)] is
mapped onto the measured curves (gray dots, blue, green, and
black lines) in Figure 4, where we summarize our results. For
example, Figure 4 a shows several measurements in water.
The peak emission frequency sweeps precipitously from ca.
19 400 cm1 at t = 0 to 18 200 cm1 at t = 100 fs and further to
equilibrium with n(1) = 16 340 cm1. Optimal values for the
cavity index and Stokes shift are found by simultaneous fit of
all measurements, and they are consistent with the molecular
properties of the probe (see the Supporting Information). The
red curve shows the calculation with ncav = 2.3 and Dn =
Angew. Chem. Int. Ed. 2005, 44, 5635 ?5639
Figure 3. Calculations for the solvation of a molecular dipole by dielectric continuum theory. a) Real (e?) and imaginary (e??) parts of water
dielectric dispersion. Sharp structure results from intramolecular IR
activity and broad bands 1?3 are caused by collective motion. b) Susceptibility c00dip of solute stabilization to solvent motion. The transformation from e enhances high-frequency contributions and reduces
low-frequency contributions. c) The spectral density for dipole solvation Cdip(n) is constructed from the susceptibility. From here the theoretical relaxation function Sdip(t) is reached by Fourier transformation.
3050 cm1. Increasing solute polarizability slows down the
relaxation as already noted from molecular dynamics simulations.[24?26]
The initial relaxation dynamics in water is particularly
interesting. The dynamic Stokes shift across this range is best
seen in the insert of Figure 4 a. Oscillations during the first
picoseconds come from vibrational modes in the excited
chromophore. But around 170 fs a shoulder can be discerned
underneath. Dipolar solvation predicts a weakly underdamped relaxation feature at precisely the same time and
with the same relative amplitude. Indeed, if the intramolecular vibrations are removed by spectral filtering, an underlying relaxation curve emerges that exactly matches the
predicted one. (Deviations before t = 100 fs are caused by a
coherent contribution to transient absorption during pumpand-probe overlap in time[27]). It follows that the oscillation
feature around 170 fs is due to coherent nuclear motion in the
solvent shell. The motion can be traced to a broad band in the
2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
100 fs. A slower stage with strong oscillations?that is,
frequency modulation of the emission band?follows up to
2 ps. As with water, simple continuum theory describes
quantitatively the shape of the relaxation curve.[30] Strong
oscillations correspond to the IR band from CO stretching in
methanol at 1030 cm1, or at 980 cm1 in [D3]methanol
(Figure 4 c).
We conclude: 1) ultrafast excitation of MQ launches and
reports coherent solvent motion in the surrounding liquid,
and 2) solvation dynamics is well described by dipolar change
Dm in a sphere of radius r and refractive index ncav surrounded
by a dielectric continuum. Eventually the dynamic Stokes
shift may be inverted into the dielectric dispersion e(n) in the
vicinity of the probe, and the imaginary part will give the local
IR spectrum ne??(n).
Received: April 22, 2005
Published online: August 1, 2005
Keywords: femtochemistry и fluorescence и IR spectroscopy и
Figure 4. The time-dependent, measured peak position of the stimulated emission band (gray points, blue, green, and black lines) is well
described by continuum theory (red). a) MQ in water. Here high-frequency oscillations correspond to intramolecular modes at 460 and
580 cm1. If they are removed by spectral filtering, an underlying relaxation curve emerges that matches the (red) predicted curve exactly. An
oscillation hump at 170 fs marks the coherent OиииO stretching
period in the water shell. b, c) MQ in methanol and [D3]methanol after
excitation at 540 nm. Fast frequency modulation is caused by coherent
CO stretching motion of solvent molecules (1100 and 930 cm1,
respectively). A hump at 0.25 ps indicates the period for collective
dielectric dispersion of water, around 200 cm1, from intermolecular OиииO stretching of the hydrogen-bonded network.[28, 29]
Measurements for MQ in methanol and [D3]methanol are
analyzed in Figure 4 b and c. In this case the excitation was
tuned to the very red absorption edge to avoid creating
intramolecular wavepackets. The emission peak relaxes from
18 450 (fitted) to 17 320 cm1 (measured) during the first
[1] K. Yoshihara, K. Tominaga, Y. Nagasawa, Bull. Chem. Soc. Jpn.
1995, 68, 696 ? 712.
[2] R. Jimenez, G. R. Fleming, P. V. Kumar, M. Maroncelli, Nature
1994, 369, 471 ? 474.
[3] M. L. Horng, J. A. Gardecki, A. Papazyan, M. Maroncelli, J.
Phys. Chem. 1995, 99, 17 311 ? 17 337.
[4] L. Zhao, J. L. PLrez Lustres, V. Farztdinov, N. P. Ernsting, Phys.
Chem. Chem. Phys. 2005, 7, 1716 ? 1725.
[5] G. R. Fleming, M. Cho, Annu. Rev. Phys. Chem. 1996, 47, 109 ?
[6] M. J. Lang, X. J. Jordanides, X. Song, G. R. Fleming, J. Chem.
Phys. 1999, 110, 5884 ? 5892.
[7] H. BNrsing, S. Kundu, P. VOhringer, J. Phys. Chem. B 2003, 107,
2404 ? 2414.
[8] J. Ruthmann, S. A. Kovalenko, N. P. Ernsting, D. Ouw, J. Chem.
Phys. 1998, 109, 5466 ? 5468.
[9] C. Silva, P. K. Walhout, K. Yokoyama, P. F. Barbara, Phys. Rev.
Lett. 1998, 80, 1086 1089.
[10] P. Changenet-Barret, C. T. Choma, E. F. Gooding, W. F.
DeGrado, R. M. Hochstrasser, J. Phys. Chem. B 2000, 104,
9322 ? 9329.
[11] C. P. Hsu, X. Song, R. A. Marcus, J. Phys. Chem. B 1997, 101,
2546 ? 2551.
[12] C. J. F. BOttcher, P. Bordewijk, Theory of Electric Polarization,
Elsevier, Amsterdam, 1996, Vol. 1, Eq. (4.31).
[13] C. P. Hsu, Y. Georgievskii, R. A. Marcus, J. Phys. Chem. A 1998,
102, 2658 ? 2666.
[14] M. Maroncelli, G. R. Fleming, J. Chem. Phys. 1987, 86, 6221 ?
[15] D. L. Beveridge, G. W. Schnuelle, J. Phys. Chem. 1975, 79, 2562 ?
[16] X. Song, D. Chandler, J. Chem. Phys. 1998, 108, 2594 ? 2600.
[17] F. O. Raineri, H. Resat, B.-C. Perng, F. Hirata, H. L. Friedman, J.
Chem. Phys. 1994, 100, 1477 ? 1491.
[18] S. Roy, B. Bagchi, J. Chem. Phys. 1994, 101, 4150 ? 4155.
[19] N. Nandi, K. Bhattacharyya, B. Bagchi, Chem. Rev. 2000, 100,
2013 ? 2046.
[20] E. O. Potma, W. P. De Boeij, D. A. Wiersma, Biophys. J. 2001,
80, 3019 ? 3024.
[21] S. K. Pal, A. H. Zewail, Chem. Rev. 2004, 104, 2099 ? 2124.
2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2005, 44, 5635 ?5639
[22] M. L. T. Asaki, A. Redondo, T. A. Zawodzinski, A. J. Taylor, J.
Chem. Phys. 2002, 116, 8469 ? 8482.
[23] J. E. Bertie, Z. Lan, Appl. Spectrosc. 1996, 50, 1047 ? 1057.
[24] M. Maroncelli, J. Mol. Liq. 1993, 57, 1 ? 37.
[25] P. V. Kumar, M. Maroncelli, J. Chem. Phys. 1995, 103, 3038 ?
[26] B. D. Bursulaya, D. A. Zichi, H. J. Kim, J. Phys. Chem. 1996, 100,
1393 ? 1405.
[27] A. D. Dobryakov, S. A. Kovalenko, N. P. Ernsting, J. Chem.
Phys., in press.
[28] I. Ohmine, H. Tanaka, Chem. Rev. 1993, 93, 2545 ? 2566.
[29] V. I. Gaiduk, D. S. F. Crothers, Ch. M. Briskina, B. M. Tseitlin, J.
Mol. Struct. 2004, 689, 11 ? 23.
[30] In methanol the relaxation in the ps range is faster than
predicted for a filled cavity; we find ncav 1. This and other
deviations (see the Supporting Information) may indicate
molecularity of solvation.
Angew. Chem. Int. Ed. 2005, 44, 5635 ?5639
2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Без категории
Размер файла
215 Кб
methyl, local, spectrum, solvation, quinolones, probes, ultrafast
Пожаловаться на содержимое документа