Solvent-mediated vibrational energy relaxation from Vaska's complex adducts in binary solvent mixtures

Article abstract of DOI:10.1021/jp400328z

A vibrational pump-probe and FTIR study was performed on two different adducts of Vaska's complex in two different sets of binary solvent mixtures. The carbonyl vibrational mode in the oxygen adduct exhibits solvatochromic shifts of ～10 cm^-1 in either benzyl alcohol or chloroform relative to benzene-d₆, whereas this vibration is nearly unchanged for the iodine adduct for the same three solvents. The width and center frequency of the carbonyl stretch for each adduct are compared to its vibrational lifetime in binary mixtures of benzene-d₆ with either benzyl alcohol or chloroform. In neat solvents, the trends in line width, frequency, and vibrational lifetime are consistent for the two adducts, but complex relationships emerge when the trends in each property are compared as a function of mixed solvent composition. ν_CO is more sensitive to the solvation environment around the trans ligand, whereas the line width and lifetime depend on the environment around the CO group itself. The carbonyl frequency and width vary nonlinearly across the two binary solvent series, indicating preferential solvation. In contrast, the vibrational lifetime changes linearly with solvent composition and is correlated with the mole fraction of chloroform but anticorrelated with the mole fraction of benzyl alcohol. The results are explained by differences in the densities of solvent modes that affect intermolecular relaxation of the carbonyl mode.

Full text of DOI:10.1021/jp400328z

Article
pubs.acs.org/JPCA
Solvent-Mediated Vibrational Energy Relaxation from Vaska’s
Complex Adducts in Binary Solvent Mixtures
Brynna H. Jones, Christopher J. Huber, and Aaron M. Massari*
Department of Chemistry, University of Minnesota − Twin Cities, 207 Pleasant Street SE, Minneapolis, Minnesota 55455, United
States
*
S Supporting Information
ABSTRACT: A vibrational pump−probe and FTIR study was performed on two different
adducts of Vaska’s complex in two different sets of binary solvent mixtures. The carbonyl
−1
vibrational mode in the oxygen adduct exhibits solvatochromic shifts of ∼10 cm in either
benzyl alcohol or chloroform relative to benzene-d , whereas this vibration is nearly
6
unchanged for the iodine adduct for the same three solvents. The width and center
frequency of the carbonyl stretch for each adduct are compared to its vibrational lifetime in
binary mixtures of benzene-d with either benzyl alcohol or chloroform. In neat solvents,
6
the trends in line width, frequency, and vibrational lifetime are consistent for the two
adducts, but complex relationships emerge when the trends in each property are compared
as a function of mixed solvent composition. νCO is more sensitive to the solvation
environment around the trans ligand, whereas the line width and lifetime depend on the
environment around the CO group itself. The carbonyl frequency and width vary
nonlinearly across the two binary solvent series, indicating preferential solvation. In
contrast, the vibrational lifetime changes linearly with solvent composition and is correlated with the mole fraction of chloroform
but anticorrelated with the mole fraction of benzyl alcohol. The results are explained by differences in the densities of solvent
modes that affect intermolecular relaxation of the carbonyl mode.
INTRODUCTION
dependent frequency shifts of νCO leading to its dephasing,
■
making the carbonyl group in VC-O a particularly sensitive
2
Vaska’s complex [bis(triphenylphosphine)iridium(II) carbonyl
chloride, VC] is an organometallic catalyst that has served as a
model system for exploring the mechanistic details of oxidative
probe of the local solvation dynamics in the vicinity of the
active site of the catalyst. However, still unclear is the extent to
which the sensitivity of a static indicator such as νCO translates
to sensitivity in dynamic variables that affect the IR line shape
1
−4
5,6
addition
and reductive elimination reactions, as well as
7
−9
biomimetic O binding to metal centers. VC and its adducts
2
such as vibrational energy relaxation (VER) and pure
contain a carbon monoxide ligand whose vibrational frequency
νCO) is a reporter of the oxidation state of the iridium.
is often cited as a textbook example of metal-to-ligand back-
bonding, whereby νCO is effectively tuned by the extent of
14−18
9
−11
dephasing.
It has been shown that, for certain classes of
(
VC
molecules, a linear correlation exists between the frequency of
the excited vibrational mode and its population relaxation
19
time whereas, in other cases, this relationship is complicated
10
electron back-donation into the CO antibonding π* orbital.
20,21
by interactions with the solvent.
We recently used this spectroscopic handle to report on the
time scales of solvent-shell rearrangement around VC in a series
of solvents that are known to have different rate constants for
In the current work, we performed a detailed study of the
solvent−solute interactions between two VC adducts in two
series of binary solvent mixtures. We focused our attention on
the behavior of the frequency and width of the inhomogeneous
line shape and the rate of VER of νCO for two specific VC
12
oxidative addition. Oxidative addition to form an adduct is
the first step in the catalytic cycle, and this reaction restructures
VC and its surrounding solvent shell. We found that the time
scales on which the solvation shells fluctuate are different for
these solvents around VC and its O adduct (VC-O ), and we
adducts: the dioxygen adduct (VC-O ) and the diiodide adduct
2
(
VC-I ). Both molecules are octahedral iridium complexes with
2
2
2
one chloride, one carbon monoxide, and two triphenylphos-
hypothesized that these dynamic differences play a role in the
catalytic rate. In both this study and a subsequent more
exhaustive FTIR study, we showed that νCO is generally
insensitive to the surrounding solvent for VC and most of its
adducts, including the I adduct (VC-I ), but is distinctly
phine ligands, but the difference in electronegativity of the O
2
−
−1
and I ligands positions their ν values 60−70 cm apart.
CO
Although the molecules are structurally similar, their
2
2
13
Special Issue: Prof. John C. Wright Festschrift
solvatochromic for VC-O2. We also found evidence that, for
VC-O , the CO mode is influenced by the solvent motions
around the O2 ligand. Based on this observation, we
Received: January 10, 2013
Revised: March 24, 2013
Published: March 25, 2013
2
1
2
hypothesized that the O ligand could contribute to the time-
2
©
2013 American Chemical Society
6150
dx.doi.org/10.1021/jp400328z | J. Phys. Chem. A 2013, 117, 6150−6157

The Journal of Physical Chemistry A
Article
frequencies exhibit the extremes of sensitivity and insensitivity
lapped and focused in the sample volume by an off-axis
parabolic reflector. The transmitted probe beam passed through
an analyzing polarizer placed directly after the sample (set to
transmit p-polarized radiation) and was then spectrally resolved
in a 0.32 m monochromator with a 75 line/mm grating and
to the nature of the solvent. The frequency of VC-O can be
2
−1
perturbed by up to 20 cm by the solvent, whereas that of VC-
I2 is largely insensitive to all solvents that we have tested.
13
Hence, these VC adducts provide two cases in which the role of
the solvent in VER can be studied in the presence and absence
of vibrational solvatochromism, providing a way to examine the
relationship between the solute−solvent interactions that
detected with a liquid-N -cooled 64-element mercury cadmium
2
telluride (MCT) linear array detector (InfraRed Associates,
Inc.). The spectral resolution of the detection system was ∼4
−1
perturb ν and those that affect the vibrational lifetime.
cm . The pump−probe signal was obtained as the difference
between the probe transmission with and without the pump
beam present, normalized by the laser spectrum of the probe,
and recorded as a function of the time delay between the pump
and probe pulses. The optical system from the OPA to the
detector was continuously purged with dry air (−100 °F dew
point) during data collection. The data were fit from 2 to 270
CO
EXPERIMENTAL METHODS
Bis(triphenylphosphine)iridium(II) carbonyl chloride (Vaska’s
complex, 99.99%, Sigma-Aldrich), chloroform (CHCl , 99.8%,
■
3
Sigma-Aldrich), benzyl alcohol (BA, ≥99%, Sigma-Aldrich),
perdeuterated (d -)benzene (99.5% D, Cambridge Isotope
6
ps to determine the population relaxation times (lifetime, T1)
Laboratories), and iodine (99.8%, Fisher Scientific) were used
as received.
for the v = 0−1 and 1−2 transitions. The vibrational relaxation
times for these two transitions consistently showed the same
trends, and the 0−1 values are used in the subsequent
discussion.
VC-O solutions were prepared by dissolving 6 mg of Vaska’s
2
complex in 2 mL of the desired solvent mixture (freshly
prepared) and waiting for at least 24 h to allow the oxygen
dissolved in solution to bind. VC-I solutions were prepared by
2
RESULTS AND DISCUSSION
The solvent-subtracted FTIR spectra of the carbonyl stretch of
dissolving 6 mg of Vaska’s complex with 2 mg of iodine in 2 mL
of solvent and sonicating for 1 h in a water bath (vials were
capped and covered with parafilm during sonication). All
solutions were filtered before being placed in the sample cell.
Typical VC concentrations were 4 mM. Solutions were used
within 2 weeks (most within 1 week) of preparation. Solutions
of the mixed solvents were stored in small vials with minimum
headspace to minimize solvent composition changes due to
evaporation. The sample cell had a path length of 150 μm for
pump−probe measurements.
■
VC-O and VC-I in neat benzene-d , CHCl , and BA are
2
2
6
3
shown in panels a and b, respectively, of Figure 1. As expected,
Infrared spectra were collected on a Nicolet 6700 FTIR
spectrometer (Thermo Scientific) with at least 16 scans and a
−1
typical resolution of 1 cm . Path lengths of both 150 and 25
μm were used for FTIR measurements.
The IR pump−probe spectroscopy setup has been described
1
2
previously. Briefly, a regeneratively amplified Ti:sapphire laser
[
Spectra-Physics, 800 nm, 40-fs pulse duration (full width at
half-maximum, fwhm), 30-nm bandwidth (fwhm), 600 mW]
was used to pump an optical parametric amplifier (OPA)
(
Spectra-Physics) at a repetition rate of 1 kHz. The near-IR
signal and idler beams generated by a β-barium borate (BBO, 3
mm thick) crystal in the OPA were difference-frequency-mixed
Figure 1. Solvent-subtracted, baselined, and normalized FTIR spectra
of the CO vibrational mode in (a) VC-O and (b) VC-I in benzene-d
6
in a silver gallium sulfide crystal (AgGaS , 0.5 mm thick) to
2
2
2
(
black), BA (red), and CHCl (blue).
3
generate 3 μJ mid-IR pulses. The mid-IR spectrum of the final
output pulses was centered at the iridium-bound CO symmetric
stretching frequency in VC-O or VC-I with approximately
the νCO value of VC-I is considerably higher than that of VC-
2
2
2
00 cm− of bandwidth (fwhm). The pulses had a temporal
1
O because of the electron-withdrawing strength of the I
−
2
2
1
3,23
duration of 90 fs (fwhm) (∼1.2 times the transform limit) as
measured by three-pulse cross-correlation using the non-
resonant third-order signal from carbon tetrachloride.
ligands.
The VC-O carbonyl peak broadens and exhibits a
2
−1
strong hypsochromic shift from 2003.6 cm in benzene-d
6
(black) to 2014.0 cm− in CHCl (blue) and to 2013.2 cm in
1
−1
3
Approximately 0.3% of the full IR laser power was split from
the parent beam with the uncoated surface of a ZnSe window.
This beam was used as the probe beam, and the remaining
stronger beam was used as the pump. The pump beam was
temporally delayed from the probe and chopped at one-half of
the repetition rate of the pulse train (500 Hz), allowing every
other pump pulse to be blocked from the sample. The probe
beam was p-polarized, and the pump beam was polarized to the
magic angle (∼54.7°) to remove any effects from orientational
relaxation. For consistent results, it was necessary for the pump
and probe beam polarizers to be placed immediately before the
BA (red). The νCO values and linewidths in the latter two
solvents are quite similar. It can be said that CHCl and BA are
3
two solvents that can lead to significant and nearly identical
7
changes in the oxidation state of iridium in VC-O2.
In contrast, VC-I shows very little change in νCO, but the
2
peak width varies considerably between the three solvents,
increasing from benzene-d (black) to BA (red) to CHCl
6
3
(blue). A comparison of the two panels in this figure indicates
that the νCO value of VC-I is nearly insensitive to the change in
2
solvents, which shows that there are no specific solvent−solute
interactions with any of the ligands, including CO, that
influence the extent of metal-to-ligand back-bonding. For VC-
2
2
sample. The pump and probe beams were spatially over-
6
151
dx.doi.org/10.1021/jp400328z | J. Phys. Chem. A 2013, 117, 6150−6157

The Journal of Physical Chemistry A
Article
I2, the average oxidation state of the iridium is the same in all
three solvents. Replacing the iodide ligands with O activates
2
solvatochromism through stronger specific interactions be-
tween this particular ligand and the solvent, as we reported
1
2,13
previously.
DFT calculations showed that these interactions
could be through hydrogen-bond donation to the lone pairs of
1
3
the O ligand. Yet, CHCl and BA have different hydrogen-
2
3
24
bonding strengths, and so, the coincidence of their νCO values
for VC-O likely has some dependence on other nonbonding
2
2
5,26
forces, such as dipole−dipole interactions.
The broadening
of the ν line shapes for both adducts suggests that their
CO
solvation environments in CHCl3 and BA are more
heterogeneous than that of benzene-d , although there might
6
also be contributions to the line shapes from changes in
solvent-shell dynamics and the T1 value that cannot be
2
7,28
determined from the FTIR spectra alone.
With the characteristics of νCO of the two adducts established
in these pure solvents, we systematically varied the solvent
composition from benzene-d to either CHCl or BA. Binary
Figure 2. Carbonyl frequency (νCO) for (a) VC-O and (b) VC-I as a
2
2
function of solvent composition in mixtures of benzene-d
and BA
6
3
6
(
(
black circles, green triangles) and mixtures of benzene-d and CHCl
red squares, blue diamonds).
solvent mixtures introduce a linear variation of the bulk solvent
composition as an independent variable onto which one can
map spectroscopic changes to test for preferential solvation. In
general, three types of preferential solvation can be described
6 3
2
5,26
for binary solvent mixtures.
First, the solvent and solute can
have site-specific interactions such as hydrogen bonding. The
second type of preferential solvation involves nonstructural
solute−solvent interactions, such as dipole−dipole interactions.
Alternatively, the solvent mixture itself might exhibit micro-
heterogeneity in the bulk, in which case solvation would occur
in small pockets or domains of one solvent or the other. In all
cases, the mole fraction of the first solvation shell(s) is different
from the mole fraction of the bulk solvent. If a particular
spectroscopic observable, such as νCO, fwhm, or T in the
1
following measurements, has different values in two neat
solvents, then a linear variation of that observable with the mole
fraction of a binary mixture of the solvents would indicate that
the solute experiences a solvation shell whose properties also
vary linearly, at least from the perspective of the measured
quantity. It is important to note that most properties of solvent
Figure 3. Full width at half-maximum (fwhm) for the carbonyl stretch
of (a) VC-O and (b) VC-I as a function of solvent composition in
2
2
mixtures of benzene-d and BA (black circles, green triangles) and
2
5,26
6
mixtures do not vary linearly.
mixtures of benzene-d and CHCl (red squares, blue diamonds).
6
3
Figure 2a shows νCO for VC-O as the composition of the
2
solvent mixture changes from pure benzene-d to either pure
6
BA (black circles) or pure CHCl (red squares). The ν and
solvents; it thus seems unlikely that the solute is entirely or
even primarily surrounded by BA molecules. What is far more
probable is that any BA molecules present in the solvation shell
have a stronger influence on the carbonyl stretch frequency
than the benzene molecules. Considering that benzyl alcohol
has a much greater dipole moment and hydrogen-bond
3
CO
fwhm values used to construct Figures 2 and 3 are included in
Table 1. The full FTIR spectra are provided in the Supporting
Information. The change in the VC-O ν value displayed in
2
CO
Figure 2a is not linear in mole fraction for either BA or CHCl3
in benzene-d . In both types of mixtures, the nonbenzene
6
solvent has a stronger influence on the carbonyl frequency than
would be anticipated from a linear projection between the pure
solvents. This indicates that there is indeed preferential
solvation of the O ligand by BA and CHCl over benzene-
d6, which decreases the availability of metal electrons for back-
bonding. The solvation shell is heterogeneously distributed
around the ligated oxygen such that the local composition is
not the same as that of the bulk solvent mixtures. Although
there certainly might be preferential solvation of other portions
of the molecule as well, the lack of change in νCO of VC-I2
strength than benzene-d (or chloroform, for that mat-
6
2
4,29−31
ter),
the more prominent deviation from linearity is
thus due to a stronger solute−solvent interaction and/or
2
3
heterogeneous solvent-shell composition, with BA preferen-
32
tially solvating the O ligand. In CHCl , this interaction with
2 3
dioxygen is weaker, even though the solvent readily dissolves
VC-O , leading to a more linear trend in ν as the mole
2
CO
fraction of CHCl increases.
3
The solvent insensitivity of νCO in VC-I is apparent in
2
Figure 2b, where the blue diamonds and green triangles show
the center frequency in the CHCl3 and BA mixtures,
respectively. Because the three pure solvents have roughly the
same νCO values, there is no variation of this parameter across
(Figure 2b) indicates that any change in solvation for the other
ligands is not reflected in a frequency shift of the CO group.
Although neither plot is linear in Figure 2a, the BA mixtures
exhibit a far less linear trend. We observed that VC-O was
the binary mixtures. The lack of a shift in the I complex
2
2
much less soluble in BA than in either of the other two
suggests that there are no specific solvent−solute interactions
6
152
dx.doi.org/10.1021/jp400328z | J. Phys. Chem. A 2013, 117, 6150−6157

The Journal of Physical Chemistry A
Article
Table 1. T , νCO, and fwhm Values for Binary Solvent
Mixtures
Based purely on the FTIR spectra discussed earlier, it is
impossible to say whether the broadening in Figure 3 is due to
changes in the homogeneous line width (arising from the
vibrational lifetime and pure dephasing in the sample) or
differences in the inhomogeneity of the systems (arising from
the variety of solvent−solute configurations). An increase in
inhomogeneity would be a reasonable explanation for the
1
a
νCO (cm−1)
fwhm (cm−1
)
T1 (ps)
mole
fraction
VC-
O2
VC-O2
VC-I2
VC-I2
VC-O2
VC-I2
CHCl in Benzene-d
3
6
0
2003.6
2005.5
2008.3
2010.2
2011.9
2012.5
2013.8
2014.5
2014.0
2068.9
2069.0
2069.6
2069.9
2070.2
2070.4
2069.9
2070.6
2070.0
16.5
17.2
17.8
18.6
19.2
19.8
20.2
20.5
21.2
8.0
9.0
92 (3)
94 (2)
194 (7)
198 (8)
205 (7)
207 (6)
214 (7)
216 (6)
219 (10)
221 (6)
216 (7)
broader peaks in the more polar CHCl and BA (stronger
3
0
0
0
0
0
0
0
1
.1
intermolecular interactions generally mean a wider variety of
environments), but the homogeneous line shape can also be
affected by the solvent. It is in these circumstances that
nonlinear spectroscopies such as vibrational pump−probe
spectroscopy can become useful in differentiating sources of
broadening. The dynamic contributions to the line shape are
.3
10.3
11.9
13.7
14.8
15.7
16.2
16.7
101 (3)
102 (1)
106 (2)
106 (2)
109 (3)
106 (2)
108 (2)
.5
.7
.8
.9
.95
the vibrational relaxation time (lifetime, T ), pure dephasing,
1
14,28
and spectral diffusion.
Here, we measure only T , isolating
1
BA in Benzene-d6
that contribution to the homogeneous line width and providing
insight into its role in line-shape broadening, as well as the rate
and possible mechanism of solvent influence on the relaxation
of vibrational energy.
0
0
0
0
0
0
0
0
1
2003.6
2006.7
2008.5
2009.8
2010.5
2011.3
2011.8
2012.5
2013.2
2068.9
16.5
18.9
8.0
8.5
92 (2)
87 (2)
88 (2)
84 (2)
78 (1)
76 (1)
73 (2)
66 (2)
56 (2)
193 (7)
.05
.1
2069.0
2069.0
2069.0
2069.1
2069.2
2069.5
2069.5
2069.4
189 (13)
188 (13)
181 (12)
174 (11)
165 (9)
19.9
8.8
.2
20.2
9.2
Preferential solvation has a peculiar manifestation in the
dynamics of the solvation shell. Cichos and co-workers showed
that the fluorescence Stokes shift that was characteristic of a
coumarin dye in alcoholic solvents was achieved even at low
concentrations of alcohols in alkane/alcohol mixtures. Time
resolving the fluorescence Stokes shift showed that the time
scale for the solvation process around the excited state was
.3
20.4
9.8
.4
20.2
9.9
.5
20.2
10.3
10.7
11.4
172 (10)
150 (7)
.75
20.5
20.5
124 (10)
Toluene
16.1
1
2003.6
2069.3
7.6
61 (4)
155 (5)
33
notably slower than in neat alcohols. Likewise, Luther and co-
a
Errors in parentheses indicate the standard deviation from the mean.
workers reported that the time-resolved fluorescence Stokes
shift of the same coumarin dye contained an abnormally slow
34
solvation process in acetonitrile/benzene mixed solvents. The
mechanism of this process was elegantly described by Nguyen
and Stratt as originating from the exchange of molecules from
the first solvation shell to re-equilibrate around the excited
−
with any of the ligands, including I and CO, that modulate the
electron density on iridium enough to change the extent of
back-bonding to the CO ligand. Because the metal-to-CO back-
bonding is the primary factor dictating the carbonyl frequency,
this does not imply that the composition of the solvent shell is
the same as that of the bulk.
35
state. Hence, the preferential solvation of a solute by one
component of a binary mixture over the other, which occurs for
the VC adducts studied here (see Figures 2 and 3), could have
an influence on the time-dependent reorganization of the
solvation-shell molecules. Viewing the role of the solvent in
As noted previously, despite the insensitivity of νCO for VC-
I2, the spectral line shapes are not independent of the solvent.
Figure 3 shows the width (fwhm) of the carbonyl stretch of
each adduct in both types of solvent mixtures. Here again,
Figure 3a contains the values for VC-O in BA mixtures (black
circles) and CHCl mixtures (red squares), whereas Figure 3b
contains those for VC-I in BA mixtures (green triangles) and
CHCl mixtures (blue diamonds). In some cases, the FTIR
bands show some asymmetry (see Figure S1, Supporting
Information), so the widths in Figure 3 are fwhm values of the
raw absorbance bands, not Gaussian fit parameters. Both
adducts show broadening as the mole fraction of benzene-d in
the bulk decreases. In the BA mixtures, both VC-O and VC-I2
exhibit greater broadening at low mole fractions than would be
expected from a linear variation in solvent composition. This is
most pronounced for VC-O in BA/benzene-d mixtures (black
36
VER as modulating the force applied to the solute, it is
conceivable that preferential solvation might also affect the CO
relaxation times. In the subsequent measurements, we used
vibrational pump−probe spectroscopy on VC adducts in binary
solvent mixtures to test for this possibility, although two-
dimensional IR spectroscopic studies underway in our
laboratory are necessary to complete the dynamical picture of
the solvation shell.
2
3
2
3
In each solvent mixture, it is possible that the CO peak
consists of two or more populations corresponding to benzene-
6
d
and CHCl
T values. This could result from preferential solvation by
1
or BA and that these subensembles have different
2
6
3
domains that are rich in one of the two solvents and might lead
to a pronounced frequency dependence in T . Alternatively, the
2
6
1
circles), which reach their maximum width at 0.1 mole fraction
of BA and broaden very little at higher mole fractions.
Combining this trend with that of νCO in Figure 2a, it can be
said that the carbonyl peak shifts and broadens at low mole
fractions (<0.1) but then retains its line width and simply
changes frequency over the rest of the range of solvent
νCO peak might reflect a single heterogeneous population with a
single relaxation time. For VC-O , the solvatochromic
2
−1
frequency shifts among the pure solvents (∼10 cm ) allow
for spectral distinction between uniform or binary solvation. To
be certain that, in choosing a particular frequency at which to
report the lifetime, we were not ruling out certain solvation
mixtures. In the CHCl mixtures, both adducts broaden mostly
environments, we compared the T values at many frequencies.
3
1
linearly in relation to the mole fraction. At every solvent
Figure 4 shows the values for the fits at most of the
wavelengths (along with the standard error of the fit) as a
composition, the peak of VC-O is consistently broader than
2
that of VC-I in the same solvent.
function of frequency for the VC-O sample in 0.3 mole
2
2
6
153
dx.doi.org/10.1021/jp400328z | J. Phys. Chem. A 2013, 117, 6150−6157

The Journal of Physical Chemistry A
Article
Figure 4. Vibrational lifetime (T ) as a function of frequency for VC-
1
O in 0.3 mole fraction BA in benzene-d (black circles). Error bars
2
6
represent the standard error of the fit. Overlaid are the pump−probe
spectra for the same sample at delay times of 1 ps (blue) and 100 ps
multiplied by 3 (red). The gray area represents the region averaged in
Figure 5. Representative normalized decays for (a) VC-O and (b)
2
VC-I2 in neat benzene-d6 (black circles), BA (red triangles), and
CHCl3 (blue squares). Symbols represent data points, and lines
illustrate the single-exponential fits to the data.
the reported values for T (see text).
1
fraction of BA in benzene-d . This mole fraction was chosen
because it contains significant amounts of both BA and
6
in the Supporting Information). For both adducts, BA has the
benzene-d and because ν in this mixture is some distance
fastest relaxation time, CHCl has the slowest, and benzene-d
6
CO
3
6
from its value in either pure solvent, giving a high probability
that binary solvation, if present, would be visible. The spread of
the data and the errors in the fits show that there are no
is intermediate (although closer to CHCl3).
It was anticipated that the carbonyl would relax much more
slowly in VC-I than in VC-O in light of literature work on the
2
2
statistical differences in T across most of the 0−1 and 1−2
relationship between vibrational frequency and relaxation
1
19
transitions. The slight increase in T at a few of the higher
time. A high-frequency mode relaxes primarily through
intramolecular channels, whereby anharmonically coupled
accepting modes in the solute become excited. For the
intramolecular process, the solvent bath serves to modulate
these interactions and donates or absorbs additional quanta of
energy that conserve energy in the overall process (solvent-
1
frequencies was also present in the T fits to the pump−probe
1
spectra of the pure solvents. The T value of VC-O in pure BA
1
2
is almost 40 ps faster than that in pure benzene-d (see Table
6
1
). Therefore, if the higher-frequency portion of the carbonyl
peak were indeed from primarily BA-surrounded molecules,
36
then in this particular benzene-d /BA mixture, one would
assisted intramolecular relaxation). If this relaxation process is
6
expect a decrease in T with increasing frequency. We conclude
that the νCO peak is a single inhomogeneous population
inefficient, intermolecular relaxation into solvent accepting
1
38,39
modes can also become a viable relaxation channel.
The
represented by a single T value.
Overlaid in Figure 4 is shown the pump−probe spectrum for
this same sample at delay times of 1 ps (blue) and 100 ps (red,
intuitive picture is that, as long as the relaxation pathway
remains the same, a higher-energy mode should take longer to
relax, as more lower-energy modes must be excited for
relaxation to occur. Hill and co-workers demonstrated a linear
relationship between the T and ν values for a library of
1
×
3). The positive peak represents the bleach from the transient
absorption from the ground state to the first excited state (0−
), and the negative peak represents the signal from the
transition from the first excited state to the second excited state
1
CO
1
metalloporphyrins in which the lifetimes became longer with
19
increasing frequency. However, that trend was not con-
sistently followed when the solvent alone was changed. For a
series of simple chlorinated solvents, they reported only minor
changes in T1 that were not correlated with νCO. They
concluded that VER proceeds primarily through intramolecular
channels from the carbonyl to the π system of the porphyrin
and that the anharmonic coupling of these channels is not
perturbed by the nature of the solvent. When a more
structurally complex dibutylphthalate (DBP) solvent was
(
1−2). The difference between the frequencies of these peaks is
indicative of the anharmonicity of the system. If a binary
population were present, one would expect the shape of the
37
spectrum to change dramatically over this long time delay.
Instead, the shapes are nearly identical, confirming that the
solvent mixture has a collective influence on T for these
1
pump−probe spectral bands. For the remainder of the
discussion, the T values are reported as the average of the
1
five points around the maximum of the positive (0−1) peak,
shown in Figure 4 as the gray area, for each trial (most data
points have at least two trials associated with them). The T1
values for both complexes in all solvents and solvent mixtures
are compiled in Table 1.
used, they observed a 40% decrease in T , which was partly
1
attributed to intermolecular relaxation into the solvent on a
time scale that was competitive with intramolecular relaxation
due to the large density of states of DBP and its strong
19
nonbonding interactions with the porphyrin. It is worth
noting that the relaxation times of their systems are on the
order of 5−10 times faster than those for the molecules we
studied here. Heilweil and co-workers reported dramatic
Representative decays of the bleached 0−1 transition for VC-
O and VC-I in the pure solvents are shown in panels a and b,
2
2
respectively, of Figure 5. The symbols represent the normalized
experimental data, and the lines show the single-exponential fits
changes in T with a change in solvent in their study of
1
from which the T values were obtained. The signal-to-noise
metal hexacarbonyls when only very small changes in ν
1
CO
20
ratios for all data sets in this study are comparable to those of
these decays, and the fits are of similarly high quality (available
occurred. In contrast, King and co-workers found minimal
changes in the CO lifetime in a multicarbonyl system, despite
6
154
dx.doi.org/10.1021/jp400328z | J. Phys. Chem. A 2013, 117, 6150−6157

The Journal of Physical Chemistry A
Article
changes in the number of hydrogen bonds that perturbed the
CO frequencies and hindered intramolecular vibrational
redistribution (IVR).
replaced with CHCl . In the T values for benzene-d /BA
3
1
6
mixtures, unlike the CHCl mixtures, VER is faster as the
3
40
solvent moves toward a higher fraction of BA.
The T values for the VC adducts reported herein are only
If the changes shown in Figure 6 for the BA (CHCl3)
mixtures were due to an increase (decrease) in the rate of IVR
caused by a change in the density of states in the VC species,
one would expect that the molecular structure and, therefore,
1
partly consistent with these studies. All of the VC-I decay
2
times are longer than those of VC-O in the same solvents. Yet,
2
our results in Table 1 illustrate clearly that solvent effects on T1
are not well-correlated with frequency changes for VC adducts.
νCO would be perturbed. However, νCO for VC-I is nearly
2
Comparing the frequencies and T values for the neat solvents,
1
unaffected by the nature of the solvent (Figure 2b). Therefore,
the νCO values of VC-O in BA and CHCl are similar and
2
3
we attribute the decrease in T to a change in the density of
1
higher in energy than those in benzene-d , but the T trend is
6
1
states in the solvent into which intermolecular relaxation
occurs. This is not unexpected because BA is a larger molecule
than benzene and should have more degrees of freedom.
BA < benzene-d < CHCl . Additionally, VC-I shows the same
6
3
2
neat-solvent trend in T when there is virtually no change in
1
center frequency. One explanation for these differences is that
the solvent-assisted intramolecular relaxation process is
Despite the fact that νCO for VC-O shifts to higher energies on
2
going from benzene-d to either BA or CHCl , the rate of
6
3
enhanced by BA but inhibited by CHCl . However, based on
3
carbonyl relaxation in BA increases, contrary to the expectation
the time scales of these T values (up to 200 ps), this channel is
19
1
for a primarily intramolecular relaxation. This is a further
inefficient for these complexes, and intermolecular energy
transfer to the solvent is probable. Because VER becomes faster
as the solvent molecules increase in size, the process is
apparently modulated by intermolecular relaxation into the
manifold of solvent accepting modes, which has a greater
indication that VER is significantly affected by intermolecular
relaxation into solvent modes for these adducts.
To further test this evidence, we also examined the relaxation
times of VC-O and VC-I in toluene. The IR spectrum of
2
2
toluene is similar to that of benzyl alcohol in the high-frequency
38,39
density of states for the larger, more complex species.
range of the spectrum close to the CO stretch (2000−1450
To more closely examine the connection between νCO and
T1, we determined the relaxation times across binary mixtures
of benzene-d with BA or CHCl . Figure 6 shows the T values
−1
cm ) but similar to benzene-d and CHCl in its lower-
6
3
−1
frequency 500−650 cm range (see spectra in the Supporting
Information). This range was chosen to ascertain whether the
relaxation pathway proceeds through the solute vibrations
closest in energy to the carbonyl. In both VC-O and VC-I ,
these vibrational peaks appear at ∼1435 cm (attributed to the
triphenylphosphine ligands), and so, 570 and 635 cm− are the
approximate differences between these vibrations and the CO
6
3
1
2
2
−1
1
stretch for VC-O2 and VC-I , respectively. The resulting
2
lifetimes are given in Table 1 and are much closer to the
lifetimes in BA than either CHCl3 or benzene-d , again
6
supporting the intermolecular model of relaxation. Also shown
in Table 1 are both the center frequency and fwhm of each
adduct in toluene. The νCO value for VC-O in toluene is also
2
very similar to that in neat benzene-d , providing another
6
indication that frequency is not the determining factor in the
value of the relaxation time for these systems.
Returning to Figure 6, we note that the T values for both
adducts in BA mixtures decrease in a linear fashion. In the
1
Figure 6. T normalized by the lifetime of pure benzene-d (T /
1
6
1
T1,benzene) as a function of solvent composition for VC-O in mixtures
2
of benzene-d and BA (black circles), VC-O in mixtures of benzene-d
CHCl mixtures, they increase linearly. This shows that the
3
6
2
6
and CHCl (red squares), VC-I in mixtures of benzene-d and BA
composition of the solvation shell in the vicinity of the CO
varies proportionately with the bulk solvent composition for
both adducts in both mixtures. This is strikingly different from
the behavior of νCO in Figure 2a and suggests that the specific
stronger solvent−solute interactions that are responsible for
3
2
6
(
(
green triangles), and VC-I in mixtures of benzene-d and CHCl
blue diamonds). Error bars represent the standard deviations from
2
6
3
the mean.
changes in ν do not have a noticeable effect on VER for the
normalized by the lifetime in neat benzene-d for VC-O and
CO
6
2
carbonyl. The evidence accumulated in Figure 6 and compared
to Figure 2 indicates that the relaxation of the carbonyl
VC-I as a function of the mole fraction of BA or CHCl . This
2
3
normalization was done to better enable comparison between
vibration in VC-O and VC-I is not sensitive to the same
adducts. Again, black circles represent VC-O in BA mixtures,
2
2
2
influences as the solvatochromic response of νCO. The shape of
νCO as a function of mole fraction of diluent is not reflected in
red squares represent VC-O2 in CHCl3 mixtures, blue
diamonds represent VC-I2 in CHCl3 mixtures, and green
the T plots with the same independent variable. Similar
triangles represent VC-I in BA mixtures. The error bars
1
2
represent the standard deviation of all frequency points
averaged (five for each trial; see Figure 4 and discussion).
proportional changes in T
fraction range regardless of whether νCO of the solute is
sensitive (VC-O ) or completely insensitive (VC-I ) to the
1
are observed over the full mole
Comparing the trends exhibited by each adduct, VC-O and
2
2
2
VC-I in both solvent mixtures show very similar proportional
solvent. The vibrational frequency is tuned by the extent of
back-bonding from the iridium, which is strongly tied to the
2
changes in T with changing solvent composition. The CHCl3
1
mixtures for VC-O (red squares) and VC-I (blue diamonds)
strength of the solvent interactions (or lack thereof) with the
2
2
−
show a monotonic increase in T as benzene-d is progressively
trans ligand (O or I ). VER has a significant contribution from
1
6
2
6
155
dx.doi.org/10.1021/jp400328z | J. Phys. Chem. A 2013, 117, 6150−6157

The Journal of Physical Chemistry A
Article
intermolecular relaxation channels that rely on coupling to the
CO ligand rather than interactions with the trans ligand.
As mentioned previously, the lifetime is one of several
variables that determines the line width of an absorbance. The
this carbonyl group as an exquisitely sensitive probe of the
equatorial solvation environment of the molecule. The linear
spectrum is strongly affected by the solvation of the catalytic
active site when a side-binding ligand such as O is present.
2
homogeneous line shape contains T as well as the pure
Furthermore, because the relaxation time responds entirely to
the environment around the carbonyl, it could be used to
decouple the influences from the front and back sides of the
molecule. This study shows that the dynamic influence of
preferential solvation is not reflected in the VER of the
carbonyl, although it might still be operational in other
dynamical solvation processes. A two-dimensional IR study of
these systems is underway, which will provide further insight
into the influences of the solvent on the carbonyl group and the
contributions of homogeneous and inhomogeneous broadening
to the linear line shape.
1
dephasing time (T *). If the lifetime decreases, with consistent
2
pure dephasing, then the homogeneous line width will increase
as dictated by the uncertainty principle. Despite the fact that, in
solution, the homogeneous peak is usually buried under large
amounts of inhomogeneous broadening, changes to its width
can still contribute to the linear line shape. This could explain
the difference in the ranges of widths covered by the two
adducts in Figure 3. The lifetimes of VC-I as listed in Table 1
2
are always longer than those of VC-O in the same solvents, so
2
the homogeneous width of the VC-O carbonyl stretch is
2
expected to be consistently larger than its counterpart in VC-I2.
The VC-O2 fwhm exhibits this expected trend. Without
ASSOCIATED CONTENT
■
knowing that T * is the same for the two adducts in the
2
S
*
Supporting Information
same solvent, it is impossible to say with certainty that this is
the cause, but there is at least a correlation between the
differences in line widths and the T values for VC-I and VC-
Full FTIR spectra for both adducts in both binary solvent
mixtures; pump−probe decays for both adducts in all solvent
mixtures; and FTIR spectra of neat benzene-d , BA, CHCl ,
and toluene. This material is available free of charge via the
Internet at http://pubs.acs.org.
1
2
6
3
O .
2
Examining the trends in Figures 6 and 3 as the solvent
composition changes, it is clear that something beyond the
change in T is affecting the width. Both CHCl and BA
1
3
AUTHOR INFORMATION
■
*
mixtures exhibit broadening in the fwhm, but only the BA
mixtures show a corresponding decrease in vibrational lifetime.
Corresponding Author
E-mail: massari@umn.edu, Phone: 612-626-8416.
The contribution to the homogeneous line width from T in
1
the CHCl mixtures is actually decreasing with increasing mole
Notes
3
fraction. Furthermore, even in the BA mixtures, the lifetime
shows a gradual, linear change, whereas the width increases
rapidly to its value at 0.1 mole fraction and then plateaus. We
hypothesize that these differences are due to inhomogeneous
broadening and/or dynamical differences induced by a range of
solvent configurations and their interconversion; further
experiments are required to affirm this conjecture.
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The authors gratefully acknowledge funding from the National
Science Foundation under CHE-0847356.
REFERENCES
■
(
1) Atwood, J. Organoiridium Complexes as Models for Homoge-
neously Catalyzed Reactions. Coord. Chem. Rev. 1988, 83, 93−114.
2) Chock, P. B.; Halpern, J. Kinetics of the Addition of Hydrogen,
CONCLUSIONS
■
We measured the FTIR and pump−probe spectra and decays
of two different adducts of Vaska’s complex in two sets of
binary solvent mixtures. The oxygen adduct exhibited
solvatochromism of the carbonyl stretch in the mixtures,
whereas the iodine adduct did not. Although we observed
higher νCO values and longer T values for VC-I complexes
(
Oxygen, and Methyl Iodide to Some Square-Planar Iridium(I)
Complexes. J. Am. Chem. Soc. 1966, 88 (15), 3511−3514.
(3) P’Pool, S.; Klingshirn, M.; Rogers, R.; Shaughnessy, K. Kinetic
Study of the Oxidative Addition of Methyl Iodide to Vaska’s Complex
in Ionic Liquids. J. Organomet. Chem. 2005, 690, 3522−3528.
1
2
(4) Stieger, H.; Kelm, H. Effect of Solvent and Pressure on the Rates
relative to VC-O , the positive linear correlation between these
2
of the Oxidative Addition Reactions of Methyl Iodide and Oxygen to
Chlorocarbonylobis(triphenylphosphine)iridium(I). J. Phys. Chem.
two variables that was reported for a different class of
compounds generally does not hold in this case. The frequency
of the CO stretch is sensitive to the solvation environment of
the O ligand for VC-O , whereas the vibrational lifetime is
1
(
973, 77, 290−294.
5) Geoffroy, G.; Hammond, G.; Gray, H. Photochemical Reductive
2
2
Elimination of Oxygen, Hydrogen, and Hydrogen Chloride from
Iridium Complexes. J. Am. Chem. Soc. 1975, 97, 3933−3936.
(6) Grushin, V. V. Reductive Elimination of Hydrogen Chloride from
Chloro Hydrido Transition Metal Complexes: An Efficient and Simple
Method for Generation of Electron-Rich, Coordinatively Unsaturated,
Reactive Intermediates. Acc. Chem. Res. 1993, 26, 279−286.
sensitive only to the solvation environment of the CO. The
dissimilarities between the T and FTIR trends across the
1
binary solvent mixtures and the similarity of the T values in
1
toluene to those in BA show that the vibrational relaxation for
these VC adducts is strongly affected by an intermolecular
pathway involving the solvent vibrational modes rather than
being dominated by solvent-mediated IVR.
(
7) La Placa, S.; Ibers, J. Structure of IrO Cl(CO)(P(C H ) ) , the
2 6 5 3 2
Oxygen Adduct of a Synthetic Reversible Molecular Oxygen Carrier. J.
Am. Chem. Soc. 1965, 87, 2581−2586.
The width of the CO stretch absorption is sensitive to
(8) Vaska, L.; Chen, L.; Senoff, C. Oxygen-Carrying Iridium
configurational influences that elude T , indicating that there is
1
Complexes: Kinetics, Mechanism, and Thermodynamics. Science
971, 174 (4009), 587−589.
9) Valentine, J. S. Dioxygen Ligand in Mononuclear Group VIII
likely some inhomogeneous component to the broadening.
However, the consistent difference between the widths of VC-
1
(
O and VC-I in the same solvent mixtures is possibly due to
2
2
Transition Metal Complexes. Chem. Rev. 1973, 73 (3), 235−245.
(10) Cotton, F. A. Advanced Inorganic Chemistry, 5th ed.; Wiley-
Interscience: New York, 1988.
homogeneous broadening, because T also exhibits a similar
1
spread. From this and previous studies, a picture is emerging of
6
156
dx.doi.org/10.1021/jp400328z | J. Phys. Chem. A 2013, 117, 6150−6157

The Journal of Physical Chemistry A
Article
(
11) Ray, G. B.; Li, X. Y.; Ibers, J. A.; Sessler, J. L.; Spiro, T. G. How
Quartet.’ New Light on Hydrogen Bond Strength and Directionality. J.
Chem. Soc., Perkin Trans. 2 1992, No. 4, 705−722.
(31) McLellan, A. L. Tables of Experimental Dipole Moments; W.H.
Freeman: San Francisco, 1963; Vol. 1.
Far Can Proteins Bend the FeCO Unit? Distal Polar and Steric Effects
in Heme Proteins and Models. J. Am. Chem. Soc. 1994, 116 (1), 162−
176.
(
12) Jones, B. H.; Huber, C. J.; Massari, A. M. Solvation Dynamics of
Vaska’s Complex by 2D-IR Spectroscopy. J. Phys. Chem. C 2011, 115
50), 24813−24822.
13) Huber, C. J.; Anglin, T. C.; Jones, B. H.; Muthu, N.; Cramer, C.
(32) Li, J.; Zhang, Y.; Carr, P. W. Novel Triangle Scheme for
Classification of Gas Chromatographic Phases Based on Solvatochro-
mic Linear Solvation Energy Relationships. Anal. Chem. 1992, 64 (2),
210−218.
(33) Cichos, F.; Willert, A.; Rempel, U.; Von Borczyskowski, C.
Solvation Dynamics in Mixtures of Polar and Nonpolar Solvents. J.
Phys. Chem. A 1997, 101 (44), 8179−8185.
(
(
J.; Massari, A. M. Vibrational Solvatochromism in Vaska’s Complex
Adducts. J. Phys. Chem. A 2012, 116 (37), 9279−9286.
(
14) Hamm, P.; Hochstrasser, R. M. Structure and Dynamics of
Proteins and Peptides: Femtosecond Two-Dimensional Infrared
Spectroscopy. In Ultrafast Infrared and Raman Spectroscopy; Fayer,
M. D., Ed.; Marcel Dekker: New York, 2001; Vol. 26, pp 273−347.
(34) Luther, B.; Kimmel, J.; Levinger, N. Dynamics of Polar Solvation
in Acetonitrile−Benzene Binary Mixtures: Role of Dipolar and
Quadrupolar Contributions to Solvation. J. Chem. Phys. 2002, 116
(8), 3370−3377.
(35) Nguyen, C.; Stratt, R. Preferential Solvation Dynamics in
Liquids: How Geodesic Pathways through the Potential Energy
Landscape Reveal Mechanistic Details about Solute Relaxation in
Liquids. J. Chem. Phys. 2010, 133, 124503.
(
15) Mathew, N. A.; Rickard, M. A.; Kornau, K. M.; Pakoulev, A. V.;
Block, S. B.; Yurs, L. A.; Wright, J. C. Coherent Multidimensional
Vibrational Spectroscopy of Representative n-Alkanes. J. Phys. Chem. A
2
(
009, 113 (36), 9792−9803.
16) Yurs, L. A.; Block, S. B.; Pakoulev, A. V.; Selinsky, R. S.; Jin, S.;
(36) Kenkre, V. M.; Tokmakoff, A.; Fayer, M. D. Theory of
Wright, J. Spectral Isolation and Measurement of Surface-Trapped
State Multidimensional Nonlinear Susceptibility in Colloidal Quantum
Dots. J. Phys. Chem. C 2012, 116 (9), 5546−5553.
Vibrational Relaxation of Polyatomic Molecules in Liquids. J. Chem.
Phys. 1994, 101 (12), 10618−10629.
(37) Ohta, K.; Tominaga, K. Vibrational Population Relaxation of
(
17) Zhao, W.; Wright, J. C. Doubly Vibrationally Enhanced Four
Hydrogen-Bonded Phenol Complexes in Solution: Investigation by
Wave Mixing: The Optical Analog to 2D NMR. Phys. Rev. Lett. 2000,
4 (7), 1411−1414.
18) Besemann, D. M.; Condon, N. J.; Murdoch, K. M.; Zhao, W.;
Ultrafast Infrared Pump−Probe Spectroscopy. Chem. Phys. 2007, 341
8
(
(
1−3), 310−319.
(
38) King, J.; Ross, M.; Kubarych, K. Water-Assisted Vibrational
Meyer, K. A.; Wright, J. C. Interference, Dephasing, and Vibrational
Coupling Effects Between Coherence Pathways in Doubly Vibration-
ally Enhanced Nonlinear Spectroscopies. Chem. Phys. 2001, 266 (2−
Relaxation of a Metal Carbonyl Complex Studied with Ultrafast 2D-IR.
J. Phys. Chem. B 2012, 116 (12), 3754−3759.
(39) Hamm, P.; Lim, M.; Hochstrasser, R. M. Vibrational Energy
3
(
), 177−195.
Relaxation of the Cyanide Ion in Water. J. Chem. Phys. 1997, 107 (24),
0523−10531.
40) King, J. T.; Anna, J. M.; Kubarych, K. J. Solvent-Hindered
Intramolecular Vibrational Redistribution. Phys. Chem. Chem. Phys.
011, 13 (13), 5579−5583.
19) Hill, J. R.; Dlott, D. D.; Fayer, M. D.; Peterson, K. A.; Rella, C.
1
(
W.; Rosenblatt, M. M.; Suslick, K. S.; Ziegler, C. J. Vibrational
Relaxation of Carbon Monoxide in Model Heme Compounds: 6-
Coordinate Metalloporphyrins (M = Fe, Ru, Os). Chem. Phys. Lett.
2
1
(
995, 244 (3−4), 218−223.
20) Heilweil, E.; Cavanagh, R.; Stephenson, J. Population Relaxation
of CO(v = 1) Vibrations in Solution Phase Metal Carbonyl
Complexes. Chem. Phys. Lett. 1987, 134, 181−188.
(
21) Banno, M.; Ohta, K.; Yamaguchi, S.; Hirai, S.; Tominaga, K.
Vibrational Dynamics of Hydrogen-Bonded Complexes in Solutions
Studied with Ultrafast Infrared Pump−Probe Spectroscopy. Acc. Chem.
Res. 2009, 42 (9), 1259−1269.
(
22) Tan, H. S.; Piletic, I. R.; Fayer, M. D. Polarization Selective
Spectroscopy Experiments: Methodology and Pitfalls. J. Opt. Soc. Am.
B: Opt. Phys. 2005, 22 (9), 2009−2017.
(
23) Vaska, L. Reversible Activation of Covalent Molecules by
Transition-Metal Complexes. The Role of the Covalent Molecule. Acc.
Chem. Res. 1968, 1 (11), 335−344.
(
24) Marcus, Y. The Properties of Organic Liquids That Are Relevant
to Their Use As Solvating Solvents. Chem. Soc. Rev. 1993, 22 (6),
09−416.
25) El Seoud, O. A. Solvation in Pure and Mixed Solvents: Some
Recent Developments. Pure Appl. Chem. 2007, 79 (6), 1135−1151.
26) Silva, P. L.; Trassi, M. A. S.; Martins, C. T.; El Seoud, O. A.
4
(
(
Solvatochromism in Binary Mixtures: First Report on a Solvation Free
Energy Relationship between Solvent Exchange Equilibrium Constants
and the Properties of the Medium. J. Phys. Chem. B 2009, 113 (28),
9
(
512−9519.
27) Hamm, P.; Lim, M. H.; Hochstrasser, R. M. Structure of the
Amide I Band of Peptides Measured by Femtosecond Nonlinear-
Infrared Spectroscopy. J. Phys. Chem. B 1998, 102 (31), 6123−6138.
(
28) Park, S.; Kwak, K.; Fayer, M. D. Ultrafast 2D-IR Vibrational
Echo Spectroscopy: A Probe of Molecular Dynamics. Laser Phys. Lett.
007, 4 (10), 704−718.
2
(
29) Mayer, U.; Gutmann, V.; Gerger, W. The Acceptor Number 
A Quantitative Empirical Parameter for Electrophilic Properties of
Solvents. Monatsh. Chem. 1975, 106 (6), 1235−1257.
(
30) Leahy, D. E.; Morris, J. J.; Taylor, P. J.; Wait, A. R. Model
Solvent Systems for QSAR. Part 3. An LSER Analysis of the ‘Critical
6
157
dx.doi.org/10.1021/jp400328z | J. Phys. Chem. A 2013, 117, 6150−6157