Thermo-solvatochromism of merocyanine polarity probes - What are the consequences of increasing probe lipophilicity through annelation?

Article abstract of DOI:10.1002/ejoc.200700805

The question raised in the title has been answered by comparing the solvatochromism of two series of polarity probes, the lipophilicities of which were increased either by increasing the length of an alkyl group (R) attached to a fixed pyridine-based structure or through annelation (i.e., by fusing benzene rings onto a central pyridine-based structure). The following novel solvatochromic probes were synthesized: 2,6-dibromo-4-[(E)-2-(1- methylquinolinium-4-yl)ethenyl]-phenolate (MeQMBr₂) and 2,6-dibromo-4-[(E)-2-(1-methylacridinium-4-yl)ethenyl)]phenolate (MeAMBr ₂). The solvatochromic behavior of these probes, along with that of 2,6-dibromo-4-[(E)-2-(1-methylpyridinium-4-yl)ethenyl]phenolate (MePMBr ₂) was analyzed in terms of increasing probe lipophilicity, through annelation. Values of the empirical solvent polarity scale [E _T(MePMBr₂)] in kcalmol^-1 correlated linearly with E_T(30), the corresponding values for the extensively employed probe 2,6-diphenyl-4-(2,4,6-triphenylpyridinium-1-yl)phenolate (RB). On the other hand, the nonlinear correlations of E_T(MeQMBr₂) or E_T(MeAMBr₂) with E_T(30) are described by second-order polynomials. Possible reasons for this behavior include: i) self-aggregation of the probe, ii) photoinduced cis/trans isomerization of the dye, and iii) probe structure- and solvent-dependent contributions of the quinonoid and zwitterionic limiting formulas to the ground and excited states of the probe. We show that mechanisms (i) and (ii) are not operative under the experimental conditions employed; experimental evidence (NMR) and theoretical calculations are presented to support the conjecture that the length of the central ethenylic bond in the dye increases in the order MeAMBr₂ > MeQMBr₂ > MePMBr₂. That is, the contribution of the zwitterionic limiting formula predominates for the latter probe, as is also the case for RB, this being the reason for the observed linear correlation between the E_T(MePMBr₂) and the E_T(30) scales. The effect of increasing probe lipophilicity on solvatochromic behavior therefore depends on the strategy employed. Increasing the length of R affects solvatochromism much less than annelation, because the former structural change hardly perturbs the energy of the intramolecular charge-transfer transition responsible for solvatochromism. The thermo-solvatochromic behavior (effect of temperature on solvatochromism) of the three probes was studied in mixtures of water with propanol and/or with DMSO. The solvation model used explicitly considers the presence of three species in the system: bulk solution and probe solvation shell [namely, water (W), organic solvent (Solv)], and solvent-water hydrogen-bonded aggregate (Solv-W). For aqueous propanol, the probe is efficiently solvated by Solv-W; the strong interaction of DMSO with W drastically decreases the efficiency of Solv-W in solvating the probe, relative to its precursor solvents. Temperature increases resulted in desolvation of the probes, due to the concomitant reduction in the structured characters of the components of the binary mixtures. Wiley-VCH Verlag GmbH & Co. KGaA, 2008.

Full text of DOI:10.1002/ejoc.200700805

FULL PAPER
DOI: 10.1002/ejoc.200700805
Thermo-Solvatochromism of Merocyanine Polarity Probes – What Are the
Consequences of Increasing Probe Lipophilicity through Annelation?
Clarissa T. Martins,[a] Michelle S. Lima,[a] Erick L. Bastos,[b] and Omar A. El Seoud*[a]
Keywords: Solvatochromism / Thermo-solvatochromism / Merocyanine dyes / Binary solvent mixtures
The question raised in the title has been answered by com-
paring the solvatochromism of two series of polarity probes,
ture that the length of the central ethenylic bond in the dye
increases in the order MeAMBr2 Ͼ MeQMBr2 Ͼ MePMBr2.
the lipophilicities of which were increased either by increas- That is, the contribution of the zwitterionic limiting formula
ing the length of an alkyl group (R) attached to a fixed pyr-
idine-based structure or through annelation (i.e., by fusing
benzene rings onto a central pyridine-based structure). The
following novel solvatochromic probes were synthesized:
predominates for the latter probe, as is also the case for RB,
this being the reason for the observed linear correlation be-
tween the ET(MePMBr2) and the ET(30) scales. The effect of
increasing probe lipophilicity on solvatochromic behavior
2,6-dibromo-4-[(E)-2-(1-methylquinolinium-4-yl)ethenyl]- therefore depends on the strategy employed. Increasing the
phenolate (MeQMBr2) and 2,6-dibromo-4-[(E)-2-(1-methyl- length of R affects solvatochromism much less than annel-
acridinium-4-yl)ethenyl)]phenolate (MeAMBr2). The solva- ation, because the former structural change hardly perturbs
tochromic behavior of these probes, along with that of 2,6-
dibromo-4-[(E)-2-(1-methylpyridinium-4-yl)ethenyl]phenol-
ate (MePMBr2) was analyzed in terms of increasing probe
lipophilicity, through annelation. Values of the empirical sol-
vent polarity scale [ET(MePMBr2)] in kcalmol–1 correlated
linearly with ET(30), the corresponding values for the exten-
sively employed probe 2,6-diphenyl-4-(2,4,6-triphenylpyridi-
the energy of the intramolecular charge-transfer transition
responsible for solvatochromism. The thermo-solvatochromic
behavior (effect of temperature on solvatochromism) of the
three probes was studied in mixtures of water with propanol
and/or with DMSO. The solvation model used explicitly con-
siders the presence of three “species” in the system: bulk
solution and probe solvation shell [namely, water (W), or-
nium-1-yl)phenolate (RB). On the other hand, the nonlinear ganic solvent (Solv)], and solvent–water hydrogen-bonded
correlations of ET(MeQMBr2) or ET(MeAMBr2) with ET(30)
are described by second-order polynomials. Possible reasons
for this behavior include: i) self-aggregation of the probe,
ii) photoinduced cis/trans isomerization of the dye, and
iii) probe structure- and solvent-dependent contributions of
the quinonoid and zwitterionic limiting formulas to the
aggregate (Solv-W). For aqueous propanol, the probe is ef-
ficiently solvated by Solv-W; the strong interaction of DMSO
with W drastically decreases the efficiency of Solv-W in sol-
vating the probe, relative to its precursor solvents. Tempera-
ture increases resulted in desolvation of the probes, due to
the concomitant reduction in the structured characters of the
ground and excited states of the probe. We show that mecha- components of the binary mixtures.
nisms (i) and (ii) are not operative under the experimental
conditions employed; experimental evidence (NMR) and
theoretical calculations are presented to support the conjec-
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2008)
ergy of desolvation that occurs when the temperature is in-
creased, and 3. the phenomenon involved in solvatochro-
mism (excitation of the probe ground state) serves as a sim-
ple model for other processes.
Introduction
The study of solvatochromic substances or polarity indi-
cators (hereafter referred to as “probes”) in pure liquids
and binary solvent mixtures is important because: 1. the re-
sults obtained shed light on the factors that affect solvation,
2. the study of thermo-solvatochromism (effect of tempera-
ture on solvatochromism) can be used to calculate the en-
Recent results have indicated that the molecular structure
of the probe and the physicochemical properties of the sol-
vent, or of the binary solvent mixture, are most relevant to
solvation (see point 1. above).[1–6] Although desolvation of
reactants and activated complexes contributes to tempera-
ture effects on reaction rates, there is no obvious way to
calculate this contribution to ∆H϶ from the Arrhenius plot.
This desolvation energy is readily calculated for the probe
from its thermo-solvatochromic data (point 2.). The magni-
tude of this energy is sizeable relative to the activation en-
thalpies of many organic reactions (e.g., ranging from 2.1
to 3.7 kcalmol–1 over a 50 ºC range in aqueous alcohols).[4]
[a] Instituto de Química, Universidade de São Paulo,
C.P. 26077, 05513-970, São Paulo, S.P., Brazil
Fax: +55-11-3091-3874
E-mail: elseoud@iq.usp.br
[b] Centro de Ciências Naturais e Humanas, Fundação Universid-
ade Federal do ABC,
09210-170, Santo André, S.P., Brazil
Supporting information for this article is available on the
WWW under http://www.eurjoc.org or from the author.
Eur. J. Org. Chem. 2008, 1165–1180
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1165

C. T. Martins, M. S. Lima, E. L. Bastos, O. A. El Seoud,
FULL PAPER
Examples of point 3. include the almost identical responses polarity scales of the probes depicted in Figure 1 are re-
of the rate constants of pH-independent hydrolyses of acti- ferred to as ET(30), ET(WB), ET(QB), ET(MePM), and
vated esters (carbonate, chloroformate, and perfluorobuty- ET(RPMBr2), respectively.
rate) and of solvatochromic probes to the compositions of
binary solvent mixtures.[7,8] More recently, probes of dif-
ferent lipophilicities have successfully been employed to ex-
plain the phenomenon of “gelation” of lysozyme solutions
in water/tetramethylurea mixtures. Briefly, the enzyme is
preferentially solvated by the organic component of the bi-
nary mixture; the biomacromolecule/solvent hydrophobic
interactions play an important role in the gelation phenom-
enon.[9] These similar responses to solvent compositions are
remarkable because of their distinct origins: attack of water
on an acyl group, disruption of the native conformation of
an enzyme, and excitation of a ground state of a probe,
respectively.
As shown, RB, WB, and QB differ widely in their molec-
ular structures, and thus in physicochemical properties that
are relevant to their solvation. Consequently, quantification
of the effects of a single probe property on its solvation
(e.g., pKa or lipophilicity) is not feasible because these prop-
erties change simultaneously for each of the probes de-
picted. A solution to this problem has recently been intro-
duced by examination of the solvatochromism of SR, where
R = C1 to C8. Members of this series differ in their hydro-
phobic characters, but not in their pKa values in water. The
reason is that the Hammett σpara values for alkyl groups
are practically the same,[11] these groups being attached to
the pyridinium ring where the negligible differences in their
inductive effects are not transmitted to the phenolate oxy-
gen.[5]
Information is obtained from solvatochromic studies by
examining the dependence of an empirical solvent polarity
scale [ET(probe)] defined by Equation (1) on some experi-
mental variable.
An alternative approach to investigation of the effect of
the molecular structure of the probe on its solvatochromism
is to keep the alkyl group fixed (methyl), while changing
the heterocyclic moiety (pyridine, quinoline, and acridine,
respectively). The effect of this structural modification
ET(probe) [kcalmol–1] = 28591.5/λmax [nm]
(1)
This scale converts the electronic transition within the would be expected to be different from that observed for
probe into the corresponding intramolecular charge-trans- SR, because fusion of benzene rings onto the heterocyclic
fer transition energy.[1] Examples of some of the probes that ring should change the energies of the resonance hybrids of
we have studied are shown in Figure 1, along with their pKa the ground and excited states. The following compounds
(in water) and logP values. The latter refers to the partition were employed: 2,6-dibromo-4-[(E)-2-(1-methylpyridinium-
coefficient of a substance between (mutually saturated) 4-yl)ethenyl]phenolate (MePMBr2), 2,6-dibromo-4-[(E)-2-
octan-1-ol and water: logP = log([substance]octan-1-ol
/
(1-methylquinolinium-4-yl)ethenyl]phenolate (MeQMBr2),
[substance]water);[10] hydrophilic and hydrophobic probes are and 2,6-dibromo-4-[(E)-2-(1-methylacridinium-4-yl)ethen-
associated with negative and positive logP values, respec- yl]phenolate (MeAMBr2); see Figure 2 (series SA;
tively. The probes shown in Figure 1 include 2,6-diphenyl- MePMBr2 is common to both series). The last two probes
4-(2,4,6-triphenylpyridinium-1-yl)phenolate (RB), 2,6- were synthesized in the present work, whereas MePMBr2
dichloro-4-(2,4,6-triphenylpyridinium-1-yl)phenolate (WB), was that studied previously.[5] Although these indicators
1-methylquinolinium-8-olate (QB), 4-[(E)-2-(1-methylpyr- have close pKa values in water (the phenolate oxygen of
idinium-4-yl)ethenyl]phenolate (MePM), and 2,6-di- MePMBr2 is 4.2 times more basic than that of MeAMBr2),
bromo-4-[(E)-2-(1-n-alkylpyridinium-4-yl)ethenyl]phenolate their lipophilicities are very different (MePMBr2 is 173.8
(RPMBr2; series SR; R = methyl to n-octyl). The solvent times more soluble in water than MeAMBr2). A number of
Figure 1. The structures and pKa (in water) and logP values of some solvatochromic probes.
1166
www.eurjoc.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2008, 1165–1180

Merocyanine Thermo-Solvatochromism – Effects of Annelation
earlier studies have shown that the electronic structures of
As shown in part (A) of Figure 3 and in Equation (2),
merocyanines can be described as resonance hybrids of qui- values of ET(MePMBr2) correlate linearly with the ET(30)
nonoid (Q) and zwitterionic (ZW) limiting formulas and scale, where r and sd refer to the correlation coefficient and
that any external perturbation, such as solvation, changes standard deviation, respectively.[5]
the relative contributions of the limiting formulas to the
Parts (B) and (C) of Figure 3 show that fusion of one or
two benzene rings onto the heterocyclic ring of MePMBr2
results in deviation from linearity; second-degree polynomi-
als are required to correlate ET(MeQMBr2) or ET(Me-
AMBr2) with ET(30), as shown in Equation (3) and Equa-
tion (4), respectively, where r2 is the correlation coefficient.
Because of the scatter in Figure 3 (C), and the correspond-
ing r2 of the regression, Equation (4) is not intended for use
for the calculation of ET(MeAMBr2) from ET(30); it just
shows that MeAMBr2 exhibits both positive and negative
solvatochromism.
resonance hybrids of the dyes.[12–14,15]
Possible explanations for the nonlinear behavior include:
i. aggregation of the more lipophilic probes leading to sol-
vent-dependent formation of dimers and higher aggregates;
the measured values of λmax being those of dye monomers
or aggregates in solvents of high and low solvent polarity,
respectively,[17–23]
ii. the dye undergoing solvent-dependent cis/trans photo-
isomerization, with values of λmax being those of the trans
isomers in polar solvents and their cis counterparts in non-
polar ones,[24] and
iii. the transition energy involved (ground stateǞexcited
state) essentially reflecting the solvation of the ground state,
for which two limiting formulas – zwitterionic and quino-
noid – may be written. The contributions of these Lewis
structure to the ground states depend on the molecular
structures of the dyes, in particular, the number of rings
present, and the solvent.[25] These explanations are analyzed
below.
Figure 2. Limiting formulas (ZW↔Q), pKa and logP values, and
numbering of carbons of the dyes studied.
Solvatochromic responses in thirty-eight protic and apro-
tic solvents were examined. The correlation of ET(probe) vs.
ET(30) showed interesting behavior: whereas ET(MePMBr2)
correlated linearly with ET(30), a second-degree poly-
nomial was required to describe the same correlation for
ET(MeQMBr2) and/or ET(MeAMBr2), with the latter probe
showing positive and negative solvatochromism. Three pos-
sible explanations were offered and analyzed. We show that
the structure- and solvent-dependent contributions from
the two limiting formulas of the probe (ZW and Q) are
responsible for the nonlinear correlation. Thermo-solva-
tochromism was studied at 10, 25, 40, and 60 ºC, and at
25, 40, and 60 ºC in water/propan-1-ol and water/DMSO,
respectively. As a function of increasing temperature, desol-
vation occurred due to the concomitant decrease of solvent
structure; this effect being larger in the case of water/pro-
pan-1-ol than in that of water/DMSO.
1. Auto-Association of the Probes
In assessment of the effects of auto-association on solva-
tochromic behavior, attention should be paid to the solvent
polarity range employed and to the concentration and the
molecular structure of the probe, in particular its lipophil-
icity and propensity to form hydrogen bonds. The linear
plot in part A of Figure 3 argues against auto-aggregation
of (very lipophilic) RB as the reason for the nonlinear be-
havior. Nevertheless, the enthalpies of dilution of solutions
of RB in some of the alcohols employed in this study (e.g.,
ethanol, propan-1-ol, butan-1-ol, and octan-1-ol) suggest
that this probe aggregates at concentrations greater than
2ϫ10–4 molL–1; Beer’s law, however, is obeyed at lower
concentrations, less than 2ϫ10–5 molL–1, in cells of 1 cm
path length.[17] Since the latter experimental conditions are
usually employed, and as we have routinely checked the va-
lidity of Beer’s law for RB in different solvents, auto-aggre-
gation of that probe can be ruled out; this leaves auto-ag-
gregation of the merocyanine probes as a possible reason.
There is spectroscopic evidence (UV/Vis) that some merocy-
anine dyes, such as 4-[(E)-2-(1-hydroxy-n-hexylpyridinium-
4-yl)ethenyl]phenolate, aggregate in cyclohexane/pyridine
Results and Discussion
Solvatochromism in Pure Solvents
Values of ET(probe) in thirty-eight solvents at 25 °C are mixtures (70:30, v/v) at [probe]Ն5ϫ10–5 molL–1.[18] This
listed in Table 1.
particular probe, however, is prone to aggregation in this
Eur. J. Org. Chem. 2008, 1165–1180
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
1167

C. T. Martins, M. S. Lima, E. L. Bastos, O. A. El Seoud,
Table 1. Solvent polarity scales [ET(probe) in kcalmol–1] at 25 °C for MePMBr2, MeQMBr2, MeAMBr2, and RB.[a]
FULL PAPER
Solvents
water
ET(MePMBr2)[b]
ET(MeQMBr2)[c]
ET(MeAMBr2)[c]
ET(30)[d]
1
65.24
59.59
50.96
63.1
n-Chain alcohols
2
3
4
5
6
7
methanol
ethanol
propan-1-ol
butan-1-ol
hexan-1-ol
octan-1-ol
59.24
56.03
54.88
54.15
53.07
52.26
51.12
48.36
47.30
46.69
45.84
45.32
41.18
40.44
40.11
40.02
39.96
39.97
55.4
51.9
50.7
49.7
48.8
48.1
Branched-chain alcohols, other alcohols, 2-alkoxyethanols
8
9
propan-2-ol
butan-2-ol
2-methylpropan-2-ol
3-methylbutan-1-ol
ethylene glycol
benzyl alcohol
cyclohexanol
2-methoxyethanol
2-ethoxyethanol
2-propoxyethanol
2-butoxyethanol
2-(2-methoxyethoxy)ethanol
53.54
52.29
50.59
53.32
61.27
54.56
52.50
56.97
55.52
54.60
54.10
55.42
46.21
45.37
43.68
46.05
52.94
47.41
45.40
48.45
47.16
46.51
46.12
47.14
40.14
40.04
39.97
40.06
43.51
39.67
39.90
40.40
40.11
40.03
40.04
40.02
48.4
47.1
43.3
49.0
56.3
50.4
47.2
52.0
51.0
50.6
50.0
50.6
10
11
12
13
14
15
16
17
18
19
Chlorinated and aromatic solvents
20
21
22
23
24
25
26
chloroform
dichloromethane
1,2-dichloroethane
chlorobenzene
benzene
toluene
xylenes
46.12
48.13
48.55
41.56
42.59
41.66
41.18
40.82
40.87
40.16
39.95
40.01
44.05
46.21
46.36
46.58
39.1
40.7
41.3
36.8
34.3
33.9
34.3
[e]
–
[e]
–
–
[e]
[e]
[e]
–
–
Polar aprotic solvents
27
28
29
30
31
32
33
34
35
36
37
38
acetone
acetonitrile
N,N-dimethylacetamide
N,N-dimethylformamide
1,3-dimethylimidazolidin-2-one
DMSO
1,4-dioxane
ethyl acetate
ethylene glycol dimethyl ether
nitromethane
pyridine
51.06
53.32
51.98
52.39
51.83
53.41
45.69
42.67
44.27
44.81
45.04
44.72
45.66
41.39
42.25
42.67
46.10
43.19
42.12
40.06
40.54
40.14
40.04
39.97
39.78
46.39
41.72
40.90
40.14
40.04
40.50
42.2
45.6
42.9
43.2
42.8
45.1
36.0
38.1
38.2
46.3
40.5
37.4
[e]
–
48.72
53.14
49.66
47.69
THF
[a] The uncertainty in ET(probe) is 0.15 kcalmol–1. [b] Values taken from ref.[5] [c] Values determined in this work. [d] Values taken from
ref.[1,16] ET(30) refers to the empirical polarity scale of the probe RB. [e] Negligible probe solubility precluded measurement of λmax; hence
calculation of ET(probe) in this solvent.
binary solvent mixture of low polarity and relative permit- employed, lead to erroneous conclusions (of negative and
tivity εr, due to hydrogen bonding between its hydroxy positive solvatochromism).[26,12]
groups, coupled with solvophobic effects.
The key to solving this problem rests on the (chemically
The plot of ET(MePM) vs. ET(30) (see Figure 1) is sim- plausible) argument that solvents of low εr and high vis-
ilar to that shown in Figure 3 (C) (i.e., it shows positive and cosity should promote merocyanine aggregation. Almost
negative solvatochromism).[12,26] As argued elsewhere, the uniquely, 2-ethylhexan-1-ol possess
a large viscosity,
solvatochromism of this dye is always negative; the positive 9.8 mPas at 20 ºC, coupled with low εr (4.4), so solutions
“branch” of the plot is due to dye aggregation. The reason of merocyanines should show aggregation in this solvent.
advanced is that ET(MePM) values for nonpolar solvents This was experimentally verified for 4-[(E)-2-(1-methylnon-
such as chloroform, THF, or 1,4-dioxane were calculated by ylpiridinium-4-yl)ethenyl]phenolate in this solvent. At low
extrapolation from binary mixtures of these solvents with probe concentration (7.4ϫ10–5 molL–1), the dye shows an
relatively high concentrations of methanol; this uncertainty intense peak at 575 nm, and a smaller one at 482 nm, attrib-
in ET(MePM), together with the high dye concentrations uted to probe monomer, and dimer respectively. Increasing
1168
www.eurjoc.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2008, 1165–1180

Merocyanine Thermo-Solvatochromism – Effects of Annelation
Figure 3. Plots of ET(probe) vs. ET(30) for the solvents investigated. The arrows in the graphs refer to: water (1), methanol (2), DMSO
(3), 1,4-dioxane (4), ethyl acetate (5), and benzene (6). Note that MePMBr2 is not soluble in the last two solvents.
ET(MePMBr2) = 10.780 + 0.875ET(30), r = 0.9878, sd = 0.418 (2)
shown in Figure S1 in the electronic supporting infor-
mation. In that case, Beer’s Law was obeyed; the ab-
sorbance equals 0.155 + 2136.450 [probe], r = 0.9986, sd
= 0.162. In summary, under the experimental conditions
employed, the nonlinear behavior shown in parts B and C
of Figure 3 is not due to auto-association either of RB or
of the employed merocyanines.
ET(MeQMBr2) = 61.297 – 1.226ET(30) + 0.019ET(30)2,
r2 = 0.9626, sd = 0.731 (3)
ET(MeAMBr2) = 129.854 – 3.885ET(30) + 0.042ET(30)2,
r2 = 0.8990, sd = 0.853 (4)
Table 2. Concentration ranges employed to test probe auto-aggre-
gation in different solvents at 25 ºC.
the probe concentration to 2.9ϫ10–3 molL–1 increased the
intensity of the latter peak, at the expense of the former
one, and the (overlapped) spectra show a sharp isosbestic
point at about 520 nm.[27] This probe, however, is very lipo-
philic (calculated log P = 4.3;[28] i.e., 166 times more soluble
in octan-1-ol than MeAMBr2) and the concentration range
is larger than that usually employed in solvatochromic stud-
ies; both factors should promote dye auto-aggregation.
The UV/Vis spectra of the employed merocyanines were
inspected as a function of their concentrations in aprotic and
protic solvents (acetone, and ethanol, respectively). No
changes in band shapes, or in λmax, and no isosbestic points
were observed for the probe concentration ranges listed in
Probe
Solvent
Concentration range [molL–1]
MePMBr2 acetone
ethanol
MeQMBr2 acetone
ethanol
MeAMBr2 acetone
ethanol
1.6ϫ10–5 to 2.2ϫ10–6
9.4ϫ10–5 to 1.4ϫ10–6
1.3ϫ10–5 to 8.1ϫ10–7
5.3ϫ10–5 to 2.0ϫ10–6
1.0ϫ10–5 to 1.5ϫ10–6
2.4ϫ10–5 to 1.5ϫ10–6
2.4ϫ10–3 to 4.8ϫ10–5
2-ethylhexan-1-ol
2. Thermal or Photoisomerization of cis/trans Forms
The thermal/photochemical cis/trans isomerization of
Table 2. The same result was observed for solutions of the merocyanine indicators has been the subject of many stud-
most lipophilic probe, MeAMBr2, in 2-ethylhexan-1-ol, as ies.[27,29–34] The photochemical/protolytic isomerization cy-
Eur. J. Org. Chem. 2008, 1165–1180
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
1169

C. T. Martins, M. S. Lima, E. L. Bastos, O. A. El Seoud,
FULL PAPER
Figure 4. Suggested protolytic/photochemical cycle for MePM.[32]
cle is shown in Figure 4.[32] Protonation of the trans isomer tion of time. That is, transǞcis isomerization does not
(I) produces H+trans (II), which undergoes light-induced seem to be operative for any of the probes employed under
isomerization to H+cis (III). This, after deprotonation, pro- our experimental conditions.
duces (IV), which after absorption of light is transformed
3. The Nonlinear Correlation is due to Probe Structure- and
Solvent-Dependent Contributions of the Quinonoid and
Zwitterionic Limiting Formulas to the Ground and Excited
States of the Probe
back into (I); this step is irreversible. It is known that the
protonated species absorb at lower λmax values than their
deprotonated counterparts, due to the disappearance of the
CTC band. In this study, (I), (II), and (III) were charac-
terized by 1H NMR spectroscopy. The calculated λmax value
The coupling constants of the hydrogen atoms of the
for the cis isomer (by peak deconvolution) was found to be merocyanine ethylenic bonds (i.e., those attached to C8–C9)
practically the same as that of its trans counterpart, and so are sensitive to the medium: they decrease as a function
(IV), if it were present, would not be detectable by UV/Vis of decreasing solvent polarity.[35] Indeed, JH8,H9 of MePM
spectroscopy.[33]
decreases in the sequence: 16.12 Hz in [D4]methanol,[24]
cisǞtrans isomerization has been invoked in order to 15.20 Hz in [D6]DMSO,[36] and 14.64 Hz in [D2]dichloro-
explain the dependence on medium composition of the UV/ methane.[24,35] Due to the low solubility of MePMBr2 and
Vis spectra of another merocyanine probe: 2,6-bis(2-meth- MeQMBr2 in CDCl3, we employed 2,6-dibromo-4-[(E)-
ylpropyl)-4-[(E)-2-(1-n-dodecylquinolinium-4-yl)ethenyl]phe- 2-(1-octylpyridinium-4-yl)ethenyl]phenolate (OcPMBr2)[5]
nolate (calculated logP = 11.4;[28] apparent pKa value deter- and 2,6-dibromo-4-[(E)-2-(1-butylquinolinium-4-yl)ethen-
mined in methanol/water, 70:30, is 7.86).[29] No changes in yl]phenolate (BuQMBr2)[37] as models for MePMBr2 and
λmax or spectrum appearance were observed as a function MeQMBr2, respectively. As argued above for SR, the length
of changing the medium composition (methanol/water), up of (R) changes the lipophilicity of the molecule without
to 50% water, v/v. In that binary mixture, however, the spec- changing its energy [i.e., without changing the double bond
trum appearance changed from two sharp peaks at ca. 625 character of the (C8–C9) bond]. This assumption can be
and 675 nm to a broad band with λmax at ca. 630 nm and experimentally tested because the four probes (MePMBr2,
with two shoulders at ca. 575 and 725 nm; the intensity of OcPMBr2, MeQMBr2, and BuQMBr2) are soluble in [D6]-
this band decreased as a function of time. Apparently, the DMSO. We found that JH8,H9 for each pair is the same [i.e.,
probe is (partially) protonated by water at higher water con- independent of the length of (R)]. 1H NMR spectra of
tent, followed by (relatively slow) transǞcis isomerism.[29]
OcPMBr2, BuQMBr2, and MeAMBr2 were recorded in
We decided to test whether the probes employed in this three solvents of decreasing polarity: [D4]methanol, [D6]-
study show similar behavior. The UV/Vis spectra of the DMSO, and CDCl3, and the coupling constants are listed
merocyanines in methanol/water mixtures were examined as in Table 3. For the same probe, values of JH8,H9 decrease as
a function of the volume fraction of water (0.5 and 0.67, a function of decreasing solvent polarity, while for different
respectively) at 25 and 50 ºC, as a function of time (one probes in the same solvent, JH8,H9 values decrease as a func-
hour). In all cases, the spectra remained the same as a func- tion of annelation. Reduction of the coupling constant re-
1170
www.eurjoc.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2008, 1165–1180

Merocyanine Thermo-Solvatochromism – Effects of Annelation
flects a concomitant reduction in the double bond character vi. RM1-CI/COSMO and DFT calculations were used to
of (C8–C9). In other words, the zwitterionic character of calculate λmax for all probes in the gas phase and water, and
the molecule is reduced as a function of decreasing solvent vii. conclusions relating to bond lengths are supplemented
polarity, and of increasing numbers of benzene rings pres- by calculations of bond orders by the semiempirical RM1
ent.
method.
With regard to points iv and v, full conformational analy-
sis of the three probes was carried out by molecular me-
chanics (UFF). Lowest-energy conformers were optimized
without constraints by the semiempirical RM1 and DFT
(B3LYP/6-31+G(d,p)) methods in the gas phase, or in a po-
larizable water continuum, hereafter referred to as water,
by the COSMO treatment.[38,39] Figure 5 shows alternative
views of optimized structures in the gas phase and in water,
predicted by both methods. For additional clarity, alterna-
tive views of MeAMBr2 are shown in Figure S2. Tables S1,
S2, and S3 in the electronic supporting information show
theoretically calculated bond lengths in MePMBr2,
MeQMBr2, MeAMBr2, respectively, along with available X-
ray data for structurally related compounds.[40,41] Table 4
shows theoretically calculated bond lengths for the central
C8–C9 bonds in MePMBr2, MeQMBr2, and MeAMBr2.
Table 3. 1H NMR coupling constants (J [Hz]) of the central C8–C9
bonds of the probes OcPMBr2, BuQMBr2, and MeAMBr2.
Coupling constant, JH8,H9 [Hz]
Probe/solvent
CDCl3
[D6]DMSO
[D4]Methanol
OcPMBr2
BuQMBr2
MeAMBr2
15.0
13.9
13.2
15.8
15.0
13.8
15.9
15.5
15.0
In conclusion, the nonlinear behavior observed in
parts (B) and (C) of Figure 3 is not due to aggregation of
any of the probes, cannot be traced to photoinduced
trans/cis isomerization of the conjugate acids, and is consis-
tent with the dependence of JH8,H9 on the molecular struc-
ture of the probe and the polarity of the solvent. This de-
pendence is due to variable, probe structure- and solvent-
dependent contributions of the two limiting formulas (zwit-
terionic and quinonoid) to the ground and excited states of
the probe.
Table 4. Theoretically calculated bond lengths [Å] for the C8–C9
bonds in MePMBr2, MeQMBr2 and MeAMBr2.
RM1
RM1/COSMO
DFT
DFT/COSMO
These conclusions are supported by theoretical calcula-
tions on the structures and electronic properties of
MePMBr2, MeQMBr2, and MeAMBr2. The strategy em-
ployed was as follows:
MePMBr2 1.426
MeQMBr2 1.430
MeAMBr2 1.447
1.353
1.377
1.422
1.402
1.405
1.416
1.371
1.377
1.389
iv. the geometries of all probes were optimized by semiem-
pirical and DFT methods in gas phase and in water by the
COSMO treatment,[38]
Both methods predict that MePMBr2 and MeQMBr2 are
v. the quality of these optimizations was checked by com- planar in the gas phase and in water. A nonplanar geometry
paring theoretically calculated bond lengths with experi- is predicted for MeAMBr2, presumably in order to reduce
mental ones, where available,
HC8–HC21 steric interactions (see Figures 2 and 5), in agree-
Figure 5. Semiempirical and DFT-optimized geometries of the probes studied in the gas phase and in water.
Eur. J. Org. Chem. 2008, 1165–1180
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
1171

C. T. Martins, M. S. Lima, E. L. Bastos, O. A. El Seoud,
FULL PAPER
ment with X-ray data for Me2ASO4.[41] The use of the Cs ployed.[42] In both cases, COSMO was employed to model
symmetry point group imposes planarity on the acridinium the water effect; the results are listed in Table 5.
moiety of MeAMBr2. We have also imposed this constraint
and repeated the calculation by RM1/COSMO; the energy
Table 5. Single-point energy calculation of electronic excitation
calc
λ
[nm] from DFT-optimized geometries and experimentally
max
calculated was found to be 9.2 kJmol–1 higher than that
previously calculated (no symmetry constraint, Figure S3).
A comparison of theoretically calculated bond lengths
with X-ray-based ones corroborates the correctness of the
approach employed. First let us consider MePMBr2 and its
precursor MePM. The two bromine atoms in the ortho posi-
tions in the phenolate moiety of the former probe should
affect the O1–C2 (ipso) bond length significantly. Therefore,
it is legitimate to exclude the data relating to this bond from
regression analysis. As shown from the r values in Table S1,
the correlation is poor for gas-phase data (rՅ0.67), but is
exp
measured λ
[nm] in water.
max
Probe
RM1-CI
TDDFT
exp
calc
λ
[nm]
λ
[nm], difference (%)[a]
max
max
MePMBr2
MeQMBr2
MeAMBr2
r
438.3
479.8
561.1
497.0 (13)
505.4 (5)
522.6 (–7)
0.9999
482.7 (10)
534.5 (11)
582.9 (4)
0.9791
[a] The difference is given by: (λ –λexp )/λemxapx ϫ100; r refers to the
calc
max max
calc
exp
correlation coefficient of plots of λ
vs. λ
.
max
max
calc
exp
The agreements between λ
and λ
compare favor-
max
max
good for optimization in water (rϾ0.94), with both meth- ably with literature data for merocyanine compounds, where
ods (RM1 and DFT) giving compatible results. Note that differences up to 24% have been reported;[43] significantly,
exp
calc
MePM·3H2O crystals were employed in the X-ray measure- plots of λ
vs. λ
yielded high r values.
max
max
ments;[40] this may have contributed to the agreement be-
tween data calculated for a compound in water (MePMBr2) TDDFT indicate the presence of one dominant transition,
and those for a crystal (MePM). first dipole-allowed π–π* transition. As shown in Figure 6,
For the three probes employed, both RM1-CI and
To the best of our knowledge, there are no crystallo- the ground states in the gas phase show strong quinonoid
graphic data for any acridine-based merocyanine. Therefore character (π-bonds in C5–C8 and C9–C10 of HOMO),
we compared theoretically calculated bonds for the acridin- whereas the corresponding excited states have zwitterionic
ium moiety of MeAMBr2 and the corresponding X-ray- character (π-bonds in C8–C9 of LUMO). A different trend
based ones for Me2AcSO4; the atom numbering of the latter is observed for the same probes in water: the characters of
compound is the same as that employed for the acridinium the ground and excited states depend on the number of
moiety of MeAMBr2 (Figure 2). For the latter probe, ex- rings present. This can be seen by comparing the C8–C9 π-
clusion of the data for C10–C11 and C10–C15 led to excellent bonds of the HOMOs of MePMBr2 and MeQMBr2 (zwit-
correlations, even for the DFT-based data in the gas phase terionic characters) with the C5–C8 and C9–C10 π-bonds of
(Table S3).
MeQMBr2 (quinonoid characters). In the electronically ex-
calc
Calculated wavelengths (λ ) in the gas phase and in cited state, MePMBr2 and MeQMBr2 have quinonoid char-
max
water (see viii. below) were obtained by performing single- acters, whereas MeAMBr2 is predicted to be mainly zwitter-
point energy calculations on DFT-optimized geometries of ionic (see C8–C9 π-bond of the corresponding LUMO).
the three probes employed; adiabatic approximation to the
The above conclusions relating to the characters of the
time-dependent density functional theory (TDDFT), or the central ethenylic bonds of the dyes can be further corrobo-
multi-electron configuration interaction (MECI) were em- rated by calculations of the C8–C9 bond orders.[35] This ad-
Figure 6. HOMOs and LUMOs of the studied merocyanines calculated in water at the RM1-CI/COSMO level from DFT/COSMO-
optimized geometries.
1172
www.eurjoc.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2008, 1165–1180

Merocyanine Thermo-Solvatochromism – Effects of Annelation
ditional calculation is important because interpretation of crease with annelation. This conclusion is in agreement with
the results (in terms of bond character) does not require the dipole moments calculated for the ground and excited
additional information, such as (unavailable) X-ray data for states; see Table S4.
a model for MeQMBr2. Bond orders both in the gas phase
In summary, both experimental data (NMR) and the re-
and in water were calculated by single-point calculations in sults of theoretical calculations, including bond lengths,
the ground states (S0, RM1) and the first electronic excited bond orders, λmax values, and dipole moments, show that
states (S1, RM1-CI) from DFT-optimized geometries.[35] annelation of the basic MePMBr2 structure leads to a de-
The results obtained are presented in Scheme 1, where the crease in the double bond character of the central ethenylic
black bars drawn parallel to the abscissa show the overall group in the ground state, and that the solvent-dependence
change in the calculated orders of the C8–C9 bonds; these of this change is responsible for the observed inversion of
vary between 1 (single, pure quinonoid) and 2 (double, pure solvatochromism, from negative in polar solvents to posi-
zwitterionic). In the gas phase, all probes show strong qui- tive in nonpolar ones.
nonoid characters in the ground state (bond orders 1.17,
1.14, and 1.09 for MePMBr2, MeQMBr2, and MeAMBr2,
respectively). Excitation from S0 to S1 increases their zwit-
terionic characters, and the magnitude of this change de-
Thermo-Solvatochromism in Binary Solvent Mixtures
creases with annelation (bond orders 1.36, 1.32, and 1.26
Thermo-solvatochromism has been studied in propan-1-
for MePMBr2, MeQMBr2, and MeAMBr2, respectively). In ol-W and DMSO-W, over the whole composition range,
water, the ground state of MePMBr2 is predicted to have from pure water to pure solvent. The former organic solvent
the highest zwitterionic character. On excitation, the order (hereafter referred to as PrOH) is the longest-chain n-
of the C8–C9 bond is reduced (from 1.66 to 1.54); the probe alcohol that is miscible with water at any ratio; DMSO was
is still zwitterionic. The same tendency is observed for chosen because it is a highly dipolar aprotic solvent; it is
MeQMBr2: the reduction in C8–C9 bond order is smaller unique in that DMSO-W interactions are stronger than
(from 1.59 to 1.56) than that observed for MePMBr2. On W-W interactions.[44–48]
the other hand, the order of the C8–C9 bond of MeAMBr2
Figure 7 shows the dependence of ET(probe) on solvent
is the smallest among the probes studied, and increases on composition at 25 °C, for five probes in the two binary mix-
going from the ground to the excited state (from 1.41 to tures. Our previous results with BuPMBr2 and OcPMBr2
1.50). That is, the zwitterionic characters of the probes de- are also included in order to compare the effects of increas-
Scheme 1. Bond orders of MePMBr2, MeQMBr2, and MeAMBr2 in water and in the gas phase calculated from RM1//DFT and RM1-
CI//DFT, with and without the COSMO treatment.
Eur. J. Org. Chem. 2008, 1165–1180
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
1173

C. T. Martins, M. S. Lima, E. L. Bastos, O. A. El Seoud,
FULL PAPER
ing probe lipophilicity, as produced by increasing the length
In order to compare the dependencies of ET(probes) on
of (R)[5] or by annelation, respectively. All plots are nonlin- χw, we used Equation (5) to calculate reduced polarity
ear (i.e., probe solvation is non-ideal), which may be attrib- scales.
uted to several factors and/or solute–solvent interaction
mechanisms, as discussed in detail elsewhere.[6] Briefly, the
non-ideal behavior may originate from: dielectric en-
richment of the probe solvation shell in the solvent of
higher relative permittivity εr,[49] preferential solvation of
the probe by one of the solvent components, or micro-
(5)
Figure 7 shows that the effects of increasing probe lipo-
heterogeneity of the binary solvent mixture. The first philicity on its solvatochromic behavior depend on the
mechanism can be ruled out, however, because if dielectric strategy employed. Values of ∆logP are 2.86 and 2.24 for
enrichment were operative, all curves in Figure 7 should lie SR and SA, respectively. That is, although SR covers a
above, not below, the straight line that connects the polarit- wider change of lipophilicity, this structural variable has a
ies of the two pure liquids. The reason is that the εr value much smaller impact on solvatochromism than SA. The
for water (73.36) is larger than those for the two organic reason for this is that increasing the length of (R) does not
solvents (20.45 and 46.45 for PrOH and DMSO, respec- perturb the energy of the intramolecular CTC; annelation,
tively).[50] A large body of experimental data and theoretical on the other hand, has a profound impact on this energy,
calculations shows that the binary mixtures employed are and hence on solvatochromic response.
micro-heterogeneous: there exist microdomains composed
The solvent polarity/temperature/solvent composition
of organic solvent surrounded by water, and of water sol- contours for the three indicators are shown in parts A
vated by organic solvent. The onset and compositions of (PrOH) and B (DMSO) of Figure 8. Considering these re-
these microdomains depend on the pair of solvents. There is sults, the following is relevant:
the possibility of preferential solvation of the (hydrophobic)
viii. We calculated the (polynomial) dependence of
probe in the less polar microdomains, leading to below the ET(probe) on the analytical mole fraction of water, and
line deviation, as shown in Figure 7.[49,51–53] In summary, present the regression coefficients in Table S3; ET(probe) at
non-ideal solvation behavior is not unexpected.
any χW can thus be readily calculated. Note that the degree
Figure 7. Dependence of the reduced empirical solvent polarity parameter [ET(probe)r] on the mole fraction of water (χW) at 25 °C, for
mixtures of water with PrOH and DMSO. The straight lines have been plotted to guide the eye; they represent ideal solvation of the dye
by the mixture, see text for details. The probe symbols are: (ᮀ) MePMBr2, (᭺) MeQMBr2, and (∆) MeAMBr2, (᭹) BuPMBr2, (*)
OcPMBr2.
1174
www.eurjoc.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2008, 1165–1180

Merocyanine Thermo-Solvatochromism – Effects of Annelation
Figure 8. Solvent polarity/temperature/solvent composition contours for MePMBr2, MeQMBr2, and MeAMBr2 in A) PrOH/W, and
B) DMSO/W.
of polynomial employed is that which gave the best data fit, fractionation factor of unity indicates ideal behavior (i.e.,
as indicated by the multiple correlation coefficients r2 and the solvation shell and bulk solvent have the same composi-
sd. The data for PrOH-W, for example, were conveniently tion). The same line of reasoning applies to φSolv-W/Solv
adjusted with a sixth-degree polynomial; for MePMBr2 at (complex solvent displaces organic solvent), and φsolv-W/W
25 ºC the corresponding r2 and sd values are 0.9995 and (complex solvent displaces W), depicted in Equation (10)
0.087, respectively. The quality of our data is demonstrated and Equation (11), respectively.
by these statistical criteria and also by the excellent agree-
ment between calculated and experimentally measured
ET(probe)solv values at all temperatures (see Table S6).
ix. We treated the data obtained according to the follow-
(9)
ing solvation model, employed in previous studies.[2–4]
where (m) represents the number of solvent molecules
whose exchange in the probe solvation shell affects ET (usu-
ally mՅ2). Arguments for the use of 1:1 stoichiometry for
Solv-W have been published elsewhere.[6,54,55] The relevant
point in this model is that it explicitly considers the forma-
tion of hydrogen-bonded (or complex) solvent species Solv-
W. Consequently, the mole fractions employed in all calcu-
lations (except for those in Table S3) are “effective” or
“local” (i.e., not analytical) ones. The equilibrium constants
of Equations (6), (7), and (8) are termed solvent “fraction-
ation factors”.
(10)
(11)
x. Results of application of Equations 6 to 8 to the
thermo-solvatochromic data for MePMBr2 and MeQMBr2
are listed in Table 6. The data for MeAMBr2 have not been
analyzed quantitatively, because they exhibited very pro-
nounced deviation from ideal behavior, as shown in Fig-
ure 7. More significantly, ET(MeAMBr2) is almost indepen-
Solv + W h Solv-W
(6)
(7)
(8)
Probe(Solv)m + m (Solv-W) h Probe(Solv-W)m + m Solv
Probe(W)m + m (Solv-W) h Probe(Solv-W)m + m W
The superscript Bk refers to bulk solvent. In Equation (9), dent of binary mixture composition, up to relatively high
φW/Solv describes the composition of the probe solvation water contents. For PrOH-W, the slopes of plots of
shell, relative to that of bulk solvent. For φW/Solv Ͼ1, the ET(probe)r vs. χW, up to χW = 0.82 (corresponding to
solvation shell is richer in W than bulk solvent, while the 29.1 molL–1 water) are 0.67, 0.52, and 0.10 for MePMBr2,
converse is true for φW/Solv Ͻ1 (i.e., the probe is preferen- MeQMBr2, and MeAMBr2, respectively. This indicates that
tially solvated by the organic solvent). Finally, a solvent MeAMBr2 is very strongly preferentially solvated by the or-
Eur. J. Org. Chem. 2008, 1165–1180
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
1175

C. T. Martins, M. S. Lima, E. L. Bastos, O. A. El Seoud,
FULL PAPER
Table 6. Analysis of thermo-solvatochromic responses of MePMBr2 and MeQMBr2 in solvent/water mixtures, according to Equations 9–
11.
Organic Probe
solvent
T
[°C]
m
φW/Solv φSolv-W/Solv φSolv-W/W ET(probe)Solv
ET(probe)W
ET(probe)Solv-W
r2
χ2
PrOH
MePMBr2
10
25
40
60
10
25
40
60
25
40
60
25
40
60
1.580 0.211
1.359 0.215
1.300 0.233
1.110 0.239
1.683 0.129
1.594 0.144
1.400 0.158
1.302 0.212
0.768 0.342
0.745 0.412
0.703 0.421
0.796 0.287
0.787 0.303
0.783 0.294
71.138
32.546
27.653
13.105
75.026
53.580
31.445
26.522
0.356
0.342
0.242
0.408
0.401
337.147
151.377
118.682
54.833
581.597
372.083
199.019
125.104
1.041
0.830
0.575
1.422
1.323
55.45 [Ϯ 0.06]
54.95 Ϯ0.08
66 [Ϯ0.08]
65.42 [Ϯ0.11]
59.76 [Ϯ0.11] 0.9995 0.0069
59.68 [Ϯ0.25] 0.9990 0.0133
59.35 [Ϯ0.33] 0.9990 0.0142
59.60 [Ϯ0.70] 0.9989 0.0166
51.26 [Ϯ0.11] 0.9991 0.0135
51.08 [Ϯ0.19] 0.9992 0.0103
50.55 [Ϯ0.32] 0.9991 0.013
49.76 [Ϯ0.41] 0.9993 0.0102
57.14 [Ϯ7.00] 0.9999 0.0036
54.05 [Ϯ6.70] 0.9998 0.0044
52.93 [Ϯ8.25] 0.9998 0.0045
46.89 [Ϯ4.82] 0.9999 0.0037
46.25 [Ϯ3.75] 0.9999 0.0025
45.62 [Ϯ4.72] 0.9998 0.0044
54.34 [Ϯ0.09] 65.21 [Ϯ0.12]
53.70 [Ϯ0.10] 64.88 [Ϯ0.13]
47.89 [Ϯ 0.07] 59.55 [Ϯ0.11]
47.44 [Ϯ 0.07] 57.80 [Ϯ0.10]
46.97 [Ϯ 0.08] 58.02 [Ϯ0.11]
46.51 [Ϯ 0.07] 57.51 [Ϯ0.10]
53.33 [Ϯ0.05] 65.25 [Ϯ0.06]
53.12 [Ϯ0.03] 65.13 [Ϯ0.06]
52.74 [Ϯ0.06] 64.81 [Ϯ0.06]
45.82 [Ϯ0.05] 58.59 [Ϯ0.06]
45.69 [Ϯ0.04] 57.97 [Ϯ0.05]
45.49 [Ϯ0.05] 57.46 [Ϯ0.06]
MeQMBr2
DMSO
MePMBr2
MeQMBr2
0.397
1.350
ganic component of the mixture; this behavior has not been cause φPrOH-W/W is related to the difference between hydro-
observed before, even for OcPMBr2, a probe that is more gen bonding plus hydrophobic interactions of PrOH-W vs.
lipophilic than MeAMBr2.[5]
only hydrogen bonding by water, see Equation (8). On the
xi. The fit of the model to our thermo-solvatochromic other hand, hydrogen bonding and hydrophobic interac-
data is shown by values of r2 and χ2, and by the excellent tions contribute to solvation by the two solvent species in-
agreement between experimentally measured and calculated volved in φPrOH-W/PrOH, see Equation (7).[2–4,58,59]
ET(probe)solvent, and ET(probe)W values, respectively. The
xiv. Solvation by aqueous DMSO is different; whereas
results in Table 6 are discussed in terms of their dependence φW/DMSO is less than 1 (i.e., similar to solvation by PrOH-
on the structures of the probe and the solvent (at the same W), values of φDMSO-W/W and φDMSO-W/DMSO are less than,
temperature, T) and on T, for the same probe and solvent. or close to, unity. To the best of our knowledge, this is the
Values of (m) are close to unity, and generally decrease as only binary mixture that is inefficient in displacing its pre-
a function of increasing T. Likewise, for each probe in each cursor components from the probe solvation shell. Consider
solvent, values of φSolv-W/Solv and φSolv-W/W, ET(probe)Solv
,
first the exchange of the pure solvents. Values of φW/DMSO
and ET(probe)W decrease as a function of increasing T. This are Ͻ1, probably because the organic solvent may solvate
probe desolvation is consistent with the known effect of the probe through strong dipole–dipole and hydrophobic
temperature on solvent structure, due to less efficient hydro- interactions, akin to those operative in aqueous DMSO.[48]
gen bonding and dipolar interactions.[53]
The small magnitudes of φDMSO-W/W and φDMSO-W/DMSO
xii. Values of φW/PrOH are Ͻ unity (i.e., water is not ef- may be attributed to the fact that the interaction of DMSO
ficient in displacing the alcohol from the probe solvation with W attenuates the solvation efficiency of the complex
shell). Whereas water and alcohols (ROH) solvate the probe solvent. Several pieces of evidence, including theoretical cal-
1
by hydrogen bonding to its phenolate oxygen, PrOH may culations,[44] IR and H and 13C NMR spectroscopy,[45,46]
further interact with the probe through hydrophobic inter- neutron scattering,[47] and electron-spray mass spec-
actions.[5]
troscopy,[48] have indicated that the DMSO-W interactions
xiii. All φPrOH-W/PrOH and φPrOH-W/W values are Ͼ1, indi- are stronger than W-W interactions. Additionally, a plot of
cating that the probes are preferentially solvated by PrOH- αmixt (acidity of the DMSO-W mixtures) vs. χDMSO shows
W; all φPrOH-W/W values are greater than the corresponding negative deviation from linearity; the corresponding plot for
φPrOH-W/PrOH values, indicating that PrOH-W displaces W βmixt (basicity of the DMSO-W mixtures) shows a positive
more efficiently than PrOH. Since PrOH is more basic than deviation[60] (i.e., aqueous DMSO is less acidic than ex-
water, we can assume that the structure of the complex spe- pected). In other words, DMSO-W aggregate may be re-
cies is given by: Hw–O–H···O(R)HROH (i.e., water is the hy- garded as a deactivated species both in hydrogen bonding
drogen bond donor to PrOH, so that the two hydrogen to the probe phenolate oxygen and in electrostatic interac-
atoms printed in italics are the sites for hydrogen bonding tion with the probe’s positively charged nitrogen, which
with the probe phenolate oxygen). As argued elsewhere, this leads to the small φ observed.
hydrogen bonding partially deactivates Hw toward further
xv. Table 6 shows that as a function of increasing tem-
bonding, this deactivation being greater the stronger the ba- perature, (m), ET(probe)solv, ET(probe)W, φSolv-W/solv, and
sicity of the alcohol.[56,57] Therefore, the efficiency of PrOH- φSolv-W/W decrease, whereas φW/Solv increases (except in one
W in displacing alcohol and/or water from the solvation case). The decrease in polarities of pure solvents can be
shell does not seem to be due to a better H-bonding ability attributed to a decrease in solvent stabilization of the probe
than those of the pure solvents; it is due to hydrophobic ground state, as a result of the concomitant decrease in sol-
interactions. The order φPrOH-W/W Ͼ φPrOH-W/PrOH is be- vent structure, and hydrogen bonding ability.[61,62] Preferen-
1176
www.eurjoc.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2008, 1165–1180

Merocyanine Thermo-Solvatochromism – Effects of Annelation
tial “clustering” of water and solvents as a function of in- cation. Although the lipophilic character of the SR series
creasing temperature means that the strengths of Solv-W changes by a factor of 724, plots of ET(probe) vs. χW are
interactions decrease in the same direction,[52,53,63–66] with a similar, showing only progressive dependence on the lipo-
concomitant decrease in ability to displace both water philicity of the probe. Changing lipophilicity through annel-
and solvent. This explains the decrease in φSolv-W/Solv and ation affects the solvatochromic responses of the three
φSolv-W/W as a function of increasing T.
probes noticeably, because this change has a profound im-
xvi. The effects of probe lipophilicity on the composition pact on the energy of the probe ground and excited states;
of its solvation shell becomes readily apparent when values the latter acquires less zwitterionic character as a function
of φ are compared (Table 6). For both binary mixtures, at on fusion of benzene rings.
comparable temperatures, all φW/Solv values are smaller, and
all φSolv-W/Solv and φSolv-W/W values are larger for MeQMBr2
Calculations
than the corresponding values for MePMBr2. This depen-
dence on probe structure can be analyzed by considering
the solvent/solute interaction mechanisms discussed above;
the relative contribution of hydrophobic interactions in-
creases as a function of increasing probe lipophilicity. This
leads to less efficient displacement of Solv by W, and more
efficient displacement of W and Solv by Solv-W in case of
the more lipophilic probe, MeQMBr2.
Calculation of the Dissociation Constant of Solv-W, Kdissoc
,
Effective
χ
, and Solvent Fractionation Factors
Species
Effective
For the solvents studied, Kdissoc and χ
were avail-
Species
able from previous studies.[5] Calculation of these param-
eters has been discussed in details elsewhere, and will be
addressed here only briefly.[2–4] Knowledge of Kdissoc (calcu-
lated on the basis of dependence of density on the composi-
tion of the binary mixture) allows calculation of the effec-
tive mole fractions of all solvent species present. The probe
solvation micro-sphere is composed of W, Solv, and Solv-
W. Observed ET (EOTbs) is the sum of the polarity of each
component, EWT , ESTolv, ESTolv-W, respectively, multiplied by
the corresponding mole fraction in the probe solvation
Conclusions
Solvation in pure solvents is due to interactions that de-
pend on the properties of the solute: namely, structure, pKa,
and lipophilicity, and the solvent. Evaluation of the relative
importance of these interactions requires the study of
probes of appropriate structures, such as the SA and SR
series. ET(MePMBr2) correlates linearly with ET(30); the
nonlinear correlations of MeQMBr2 and MeAMBr2 are
due to probe molecular structure- and solvent-dependent
contributions of the limiting zwitterionic and quinonoid
structures to the corresponding resonance hybrids. This ex-
planation is in agreement with experimentally obtained
NMR spectroscopic data and with theoretical calculations
of bond lengths, bond orders, and λmax. Thermo-solvatoch-
romism in binary solvent mixtures can be described by a
mechanism based on solvent exchange equilibria between
the species present in solution (W, Solv, and Solv-W com-
plexes, respectively) and their counterparts in the probe sol-
micro-sphere, χWProbe, χ
Equations (12) and (13) can then be solved to obtain
ESTolv-W, and the appropriate solvent fractionation factors,
respectively.
and χ
, respectively.
Probe
Probe
Solv
Solv-W
(12)
The input data to solve Equation (13) include EOTbs, EWT ,
ESTolv, and χ
, along with initial estimates of (m),
Effective
Species
ESTolv-W, and the appropriate φ; values of φSolv-W/W are ob-
tained by dividing φSolv-W/Solv by φW/Solv
.
vation shell. The non-ideal dependence of ET(probes) on Experimental Section
χW is mainly due to preferential solvation of the probe, es-
Materials: All chemicals were purchased from Acros or Merck. The
solvents were purified by the recommended procedures,[67] followed
by storing over activated type 4-Å molecular sieves. Their purities
were established from their densities (DMA 40 densimeter, An-
ton Paar, Graz) and from agreement between their experimentally
pecially by Solv-W; aqueous DMSO is an exception. The
temperature effect on φ has been interpreted in terms of the
structures of water and solvent and their mutual interac-
tions. Temperature increases result in gradual desolvation
of every probe, in all binary mixtures; desolvation energies determined ET(30) and published data.[1,16]
depend on the hydrophobicity of the probe and the solvent.
A salient point of this study is that the consequences of
increasing probe lipophilicity for its solvation depend on
Apparatus: Melting points were determined with an IA 6304 appa-
ratus (Electrothermal, London). Elemental analyses were per-
formed on a Perkin–Elmer 2400 CHN-analyzer (Perkin–Elmer,
the strategy employed in introducing this structural modifi- Wellesley), in the Analytical Center of this Institute. IR and NMR
(13)
Eur. J. Org. Chem. 2008, 1165–1180
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
1177

C. T. Martins, M. S. Lima, E. L. Bastos, O. A. El Seoud,
FULL PAPER
spectra were recorded with a Bruker Victor-22 FTIR spectrometer
(Bruker Optics, Ettlingen), and a Varian Innova 300 NMR spec-
trometer (Varian, Palo Alto). Analysis of the 1H NMR and 13C
NMR spectroscopic data was based on simulation of the 1-D spec-
tra, the DQF-COSY, and HETCOR experiments.[68] A vortex
mixer (ThermoFischer Scientific, Waltham) or a sonication bath
(Laborette 17, Fretsch, Berlin) were employed to accelerate probe
dissolution.
Synthesis of the Probes Employed
2,6-Dibromo-4[(E)-2-(1-methylquinolinium-4-yl-)ethenyl]phenolate
(MeQMBr2): The synthesis of 1,4-dimethylquinolinium iodide was
carried out as described previously for MePMBr2.[5] The reaction
between methyl iodide and 4-methylquinoline [Equation (14)] was
carried out in acetonitrile, followed by removal of the excess of
CH3I and solvent to give a light amber liquid, the purity of which
was established by TLC; ethanol/acetic acid/chloroform eluent
(1:1:18, by volume). The aldehyde, 3,5-dibromo-4-hydroxybenzal-
dehyde, was available from a previous study.[37] Condensation of
the aldehyde with 1,4-dimethylquinolinium iodide in the presence
of piperidine, followed by treatment with KOH [Equation (15)],
gave MeQMBr2 as red–purple crystals; these were washed with hot
(17)
hydroxybenzaldehyde in propan-2-ol, followed by treatment with
KOH and recrystallization from methanol [Equation (17)], gave
dark greenish-blue crystals, which were washed with water and
dried. Yield 70%, decomposes at 209 °C. IR (KBr): ν = 3034
˜
(νC–H), 1601 (νC=C), 1202 (νC–N), 1044 (νC–Br) cm–1. Calculated for
C22H15Br2NO (%): C 56.32, H 3.22, N 2.99; found C 55.38, H 2.93,
1
N 2.99. The H NMR results are listed in Table 7.
Spectrometric Determination of logP, the Partition Coefficient of
the Probe Between Water and Octan-1-ol: The aqueous phase was
a phosphate buffer solution (0.05 molL–1, pH 7.50). Equal volumes
of this buffer and octan-1-ol were agitated for one hour (tube ro-
tator), and the phases separated. A probe solution [5ϫ10–4 molL–1
in (buffer-saturated) octan-1-ol] was prepared, and its absorbance
(AInitial) was recorded. An aliquot of this solution (VOctanol) was
water and dried. Yield 70%, m.p. 270–271 °C. IR (KBr): ν = 3028
˜
(νC–H), 1593 (νC=C), 1213 (νC–N), 1036 (νC–Br) cm–1. Calculated for
C18H13Br2NO (%): C 51.58, H 3.13, N 3.34; found C 50.82, H 3.56,
1
N 3.34. The H NMR results are given in Table 7.
agitated with (octan-1-ol saturated) phosphate buffer (VBuffer
)
Table 7. 1H NMR spectroscopic data for the probes synthesized:
MePMBr2, MeQMBr2, and MeAMBr2.[a]
(14)
δ [ppm]
J [Hz]
(15)
[b]
H4, H6 8.050 (s)
–
H4, H6
–
H8
H9
7.591 (d) J8,9 = 15.0 H8
7.998 (d) H9
7.668 (d)
J8,9 = 13.8
–
MeAMBr2 was synthesized as shown in Equations (16) and (17).
[b]
–
H11
H12
H17
H18
H19
H20
H16
7.933 (d) J11,12 = 6.9 H17, H21 8.459 (d)[c] J17,18/21,22 = 8.1
8.689 (d)
–
H18, H22 7.477 (m)
–
–
–
–
[b]
8.116 (d) J16,18 = 8.4 H19, H23
8.059 (t) J18,19 = 7.5 H20, H24 8.459 (d)[c]
7.779 (t) J19,20 = 8.5 H 16
4.025 (s)
8.974 (d)
4.238 (s)
–
–
[a] At 300 MHz and 25 °C, digital resolution = 0.1 Hz per data
point, solvent [D6]DMSO, reference TMS. The following abbrevi-
ations were employed for peak multiplicities: d = doublet, m =
multiplet, s = singlet, and t = triplet. [b] Chemical shifts are not
listed because of peak overlapping (e.g., peak integration of the
multiplet at 7.880 to 7.841 ppm corresponds to five hydrogen
atoms). [c] These H atoms show the same shifts, peak integration
corresponds to five hydrogens.
(16)
The synthesis of 9,10-dimethylacridinium sulfate from (CH3)2SO4
and 9-methylacridine was carried out as given elsewhere[69] [Equa-
tion (16)]. Excess dimethyl sulfate and solvent were removed by
distillation, and the yellow product was recrystallized from meth-
anol. Condensation of the methyl sulfate salt with 3,5-dibromo-4-
1178
www.eurjoc.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2008, 1165–1180

Merocyanine Thermo-Solvatochromism – Effects of Annelation
at room temperature for 6 h. After phase separation at 25 °C, the
absorbance (AEquilibrium) of the octan-1-ol phase was measured,
timized geometries of MeAMBr2 in the gas phase and in water;
RM1-optimized geometry of MeAMBr2 in water, calculated by im-
posing Cs symmetry point group. Theoretically calculated bond
lengths (in Å) for MePMBr2 and crystallographic data of MePM.
Theoretically calculated bond lengths (in Å) for MeQMBr2. Theo-
retically calculated bond lengths (in Å) for MeAMBr2 and crystal-
lographic data of 9,10-dimethylacridinium methyl sulfate, Me2-
ASO4. Dipole moments of the probes in the ground and excited
state, calculated by RM1 and RM1-CI methods in water (CO-
SMO). Coordinates for optimized geometries for all the probes
studied. Thermo-solvatochromic data for MePMBr2, MeQMBr2
and MeAMBr2 in binary solvent mixtures.
and the partition coefficient was calculated from: logP
=
log [AEquilibrium ϫVBuffer/(AInitial – AEquilibrium)ϫVOctanol]. Values of
logP were found to be: 1.02Ϯ0.01 and 2.08Ϯ0.1, for MeQMBr2
and MeAMBr2, respectively.
Spectrometric Determination of the Apparent pKa Values of the
Probes: The pKa values were calculated from the Henderson–Has-
selbach equation.[70] A methanolic probe solution was added to
potassium hydrogen phthalate buffer (0.05 molL–1) so that the final
concentration of the probe was 5ϫ10–4 molL–1 and the final vol-
ume fraction of methanol was Յ5%. The concentrations of the
probe zwitterionic form were measured at 25 °C, at 480 nm for
MeQMBr2 and 562 nm for MeAMBr2. The apparent pKa values of
the probes were found to be 5.03Ϯ0.02 and 4.52Ϯ0.02 for
MeQMBr2 and MeAMBr2, respectively.
Acknowledgments
We thank the State of São Paulo Research Foundation (FAPESP)
for financial support (grant 2004/15400-5), and research fellow-
ships 04/15677-7 and 07/00684-6, to C. T. M. and E. L. B., the
National Council for Scientific and Technological Research
(CNPq) for a research productivity fellowship to O. A. E. S. (grant
305547/2003-8), and a PIBIC fellowship to M. S. Lima. We also
thank Dr. P. A. R. Pires and Mr. C. Guizzo for their help, and the
LCCA for making computation facilities available to us.
Spectroscopic Determination of ET(Probe) in Pure Solvents and in
Binary Solvent Mixtures: Aliquots of the probe solution in acetone
were pipetted into 2 mL volumetric tubes, followed by evaporation
of the solvent at room temperature, under reduced pressure, over
P4O10. Pure solvents, and/or binary solvent mixtures were added
so that the probe final concentration was 10–5 to 10–6 molL–1.
A Shimadzu UV 2550 UV/Vis spectrophotometer, was used. The
temperature inside the thermostatted cell-holder was controlled to
within Ϯ0.05 °C with a digital thermometer (model 4000A, Yellow
Springs Instruments, Ohio). Each spectrum was recorded twice at
a rate of 140 nmmin–1; the values of λmax were determined from
the first derivative of the absorption spectra. The uncertainties in
ET(MePMBr2), ET(MeQMBr2), and ET(MeAMBr2) are
0.15 kcalmol–1. Thermo-solvatochromism was studied in mixtures,
16 per set, of water with PrOH (from 10 to 60 °C) and DMSO
(from 25 to 60 °C). DMSO mixtures were not studied in 10 °C be-
cause the pure solvent freezes at this temperature.
[1] C. Reichardt, in Solvents and Solvent Effects in Organic Chem-
istry, 3rd ed., VCH, Weinheim, 2003, p. 389.
[2] E. B. Tada, P. L. Silva, O. A. El Seoud, J. Phys. Org. Chem.
2003, 16, 691–699.
[3] E. B. Tada, P. L. Silva, O. A. El Seoud, Phys. Chem. Chem.
Phys. 2003, 5, 5378–5385.
[4] E. B. Tada, P. L. Silva, C. Tavares, O. A. El Seoud, J. Phys. Org.
Chem. 2005, 18, 398–407.
[5] C. T. Martins, M. S. Lima, O. A. El Seoud, J. Org. Chem. 2006,
71, 9068–9079.
Quantum Chemical Calculations: Structures of MePMBr2,
MeQMBr2, and MeAMBr2 were optimized by the RM1 semiem-
pirical method, as implemented in MOPAC2007.[39,71] Geometry
optimization and vertical excitation energy calculations were per-
formed by the multi-electron configuration interaction (MECI) ap-
proach and the RM1 method, hereafter referred to as RM1-CI.
The active space was constructed with five molecular orbitals
(MOs) and two double-filled levels [C.I. = (5,2), 100 configurations
in active space]. Solvent effects on the optical properties were mod-
eled with the conductor-like screening model (COSMO)[38]
[6] O. A. El Seoud, Pure Appl. Chem. 2007, 79, 1135–1151.
[7] O. A. El Seoud, M. I. El Seoud, J. P. S. Farah, J. Org. Chem.
1997, 62, 5928–5933.
[8] F. Siviero, O. A. El Seoud, J. Phys. Org. Chem. 2006, 19, 793–
802.
[9] M. A. Silva, C. T. Martins, E. P. G. Arêas, O. A. El Seoud, Pol.
J. Chem. 2007, 81, 1135–1145.
[10] A. J. Leo, C. Hansch, Perspect. Drug Discovery Des. 1999, 17,
1–25.
[11] C. Hansch, A. J. Leo, R. W. Taft, Chem. Rev. 1991, 91, 165–
195.
[12] H. G. Benson, J. N. Murrell, J. Chem. Soc. Faraday Trans. 2
1972, 68, 137.
[13] A. Botrel, A. Lebeuze, P. Jacques, H. Strub, J. Chem. Soc. Fara-
day Trans. 2 1984, 80, 1235–1252.
[14] E. Buncel, S. Rajagopal, Acc. Chem. Res. 1990, 23, 226–231.
[15] A. Abbotto, L. Beverina, S. Bradamante, A. Facchetti, C.
Klein, G. A. Pagani, M. Redi-Abshiro, R. Wortmann, Chem.
Eur. J. 2003, 9, 1991–2007.
[16] C. Reichardt, in Solvents and Solvent Effects in Organic Chem-
istry, 3rd ed., VCH, Weinheim, 2003, p. 147.
[17] C. Machado, M. D. Nascimento, M. C. Rezende, A. E. Beezer,
Thermochim. Acta 1999, 328, 155–159.
Further optimization was performed at the DFT level, with the
hybrid B3LYP exchange-correlation functional and the 6-
31+G(d,p) basis set as implemented in the Gaussian 03 quantum
mechanical package.[72–74]
Vertical excitation energies were calculated from DFT and RM1
geometries with the adiabatic approximation of the TDDFT by
use of the B3LYP functional and the 6-31+G(d,p) basis set. All
calculations were performed both in gas phase and with the CO-
SMO polarizeable continuum aqueous solvent model coupled to
B3LYP/6-31+G(d,p). All geometry optimizations were performed
without any constraints, except when otherwise indicated. Station-
ary points were confirmed as minima via vibrational frequency cal-
culations. Coordinates for optimized geometries are listed in
Table S5. All calculations were performed at the advanced comput-
ing facilities (LCCA) of the University of São Paulo.
[18] M. Niedbalska, I. Gruda, Can. J. Chem. 1990, 68, 691–695.
[19] R. Palepu, H. Gharibi, D. M. Bloor, E. Wynjones, Langmuir
1993, 9, 110–112.
[20] W. Binanalimbele, R. Zana, Colloid Polym. Sci. 1989, 267, 440–
447.
[21] D. J. Lee, W. H. Huang, Colloid Polym. Sci. 1996, 274, 160–
165.
[22] J. Penfold, E. Staples, I. Tucker, P. Cummins, J. Colloid Inter-
face Sci. 1997, 185, 424–431.
Supporting Information (see also the footnote on the first page of
this article): Absorption spectra of MeAMBr2 at different probe
concentrations in 2-ethyl-1-hexanol; semi-empirical- and DFT-op-
Eur. J. Org. Chem. 2008, 1165–1180
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
1179

C. T. Martins, M. S. Lima, E. L. Bastos, O. A. El Seoud,
FULL PAPER
[23] R. Nagarajan, C. C. Wang, Langmuir 2000, 16, 5242–5251.
[24] M. Tsukada, Y. Mineo, K. Itoh, J. Phys. Chem. 1989, 93, 7989–
7992.
[25] M. C. Rezende, J. Braz. Chem. Soc. 1997, 8, 631–635.
[26] P. Jacques, J. Phys. Chem. 1986, 90, 5535–5539.
[27] J. O. Morley, R. M. Morley, A. L. Fitton, J. Am. Chem. Soc.
1998, 120, 11479–11488.
[28] ACD Labs Software, version V8:19, 1994–2006.
[29] H. Hisamoto, H. Tohma, T. Yamada, K. Yamauchi, D. Sis-
wanta, N. Yoshioka, K. Suzuki, Anal. Chim. Acta 1998, 373,
271–289.
[30] K. J. Dennis, T. Luong, M. L. Reshwan, M. J. Minch, J. Phys.
Chem. 1993, 97, 8328–8335.
[31] J. Catalan, E. Mena, W. Meutermans, J. Elguero, J. Phys.
Chem. 1992, 96, 3615–3621.
[32] E. Hammam, A. M. El Nahas, J. Phys. Chem. A 1998, 102,
9739–9744.
th
[50]
D. R. Lide, CRC Handbook of Chemistry and Physics, 85 ed.,
CRC Press, Boca Raton, 2005.
[51]
[52]
Y. Marcus, Chem. Soc. Rev. 1993, 22, 409–416.
I. Shulgin, E. Ruckenstein, J. Phys. Chem. B 1999, 103, 872–
877.
Y. Marcus, Monatsh. Chem. 2001, 132, 1387–1411.
C. Rafols, M. Roses, E. Bosch, J. Chem. Soc. Perkin Trans. 2
1997, 243–248.
U. Buhvestov, F. Rived, C. Rafols, E. Bosch, M. Roses, J. Phys.
Org. Chem. 1998, 11, 185–192.
B. Kingston, M. C. R. Symons, J. Chem. Soc. Faraday Trans.
2 1973, 69, 978–992.
M. C. R. Symons, Pure Appl. Chem. 1986, 58, 1121–1132.
E. B. Tada, L. P. Novaki, O. A. El Seoud, J. Phys. Org. Chem.
2000, 13, 679–687.
M. S. Antonious, E. B. Tada, O. A. El Seoud, J. Phys. Org.
Chem. 2002, 15, 403–412.
J. Catalan, C. Diaz, F. Garcia-Blanco, J. Org. Chem. 2001, 66,
5846–5852.
R. Zana, M. J. Eliebari, J. Phys. Chem. 1993, 97, 11134–11136.
J. R. Haak, J. B. F. N. Engberts, Recl. Trav. Chim. Pays-Bas
1986, 105, 307–311.
[53]
[54]
[55]
[56]
[57]
[58]
[59]
[60]
[33] U. Steiner, M. H. Abdelkader, P. Fischer, H. E. A. Kramer, J.
Am. Chem. Soc. 1978, 100, 3190–3197.
[34] M. H. Abdelkader, U. Steiner, J. Chem. Educ. 1983, 60, 160–
162.
[61]
[62]
[35] D. Armstron, P. G. Perkins, J. J. P. Stewart, J. Chem. Soc. Dal-
ton Trans. 1973, 838–840.
[36] E. B. Tada, Solvatochromism in solvents and in micellar systems,
Ph.D. Thesis, University of São Paulo, São Paulo, 2004.
[37] C. T. Martins, M. S. Lima, O. A. El Seoud, J. Phys. Org. Chem.
2005, 18, 1072–1085.
[38] F. Eckert, A. Klamt, Aiche J. 2002, 48, 369–385.
[39] G. B. Rocha, R. O. Freire, A. M. Simas, J. J. P. Stewart, J. Com-
put. Chem. 2006, 27, 1101–1111.
[40] D. J. A. Deridder, D. Heijdenrijk, H. Schenk, R. A. Dommisse,
G. L. Lemiere, J. A. Lepoivre, F. A. Alderweireldt, Acta Crys-
tallogr. 1990, 46, 2197–2199.
[41] J. Meszko, A. Sikorrski, O. M. Huta, A. Konitz, J. Blazejowski,
Acta Crystallogr. 2002, 58, 669–671.
[42] E. Runge, E. K. U. Gross, Phys. Rev. Lett. 1984, 52, 997–1000.
[43] M. Guillaume, B. Champagne, F. Zutterman, J. Phys. Chem.
A 2006, 110, 13007–13013.
[44] I. A. Borin, M. S. Skaf, J. Chem. Phys. 1999, 110, 6412–6420.
[45] K. Mizuno, S. Imafuji, T. Ochi, T. Ohta, S. Maeda, J. Phys.
Chem. B 2000, 104, 11001–11005.
[46] S. N. Shashkov, M. A. Kiselev, S. N. Tioutiounnikov, A. M. Ki-
selev, P. Lesieur, Physica B 1999, 271, 184–191.
[47] J. T. Cabral, A. Luzar, J. Teixeira, M. C. Bellissent-Funel, J.
Chem. Phys. 2000, 113, 8736–8745.
[63]
[64]
K. R. Harris, P. J. Newitt, J. Phys. Chem. A 1999, 103, 6508–
6513.
K. Nishikawa, H. Hayashi, T. Iijima, J. Phys. Chem. 1989, 93,
6559–6565.
M. Huelsekopf, R. Ludwig, J. Mol. Liq. 2000, 85, 105–125.
I. Shulgin, E. Ruckenstein, J. Phys. Chem. B 1999, 103, 2496–
[65]
[66]
2503.
[67]
[68]
[69]
[70]
W. L. F. Armarego, C. L. L. Chai, in Purification of Laboratory
Chemicals, 5th ed., Elsevier, New York, 2003, p. 80.
T. D. W. Claridge, in High Resolution NMR Techniques in Or-
ganic Chemistry, Pergamon, New York, 1999, p. 148.
C. Huber, K. Fahnrich, C. Krause, T. Werner, J. Photochem.
Photobiol. A 1999, 128, 111–120.
E. V. Anslyn, D. A. Dougherty, in Modern Physical Organic
Chemistry; University Science Books, Sausalito, 2004, p. 259.
J. J. P. Stewart, MOPAC2007, version 7.176 W, 2007.
M. J. Frisch, G. W. Trucks, G. E. Scuseria, M. A. Robb, J. R.
Cheeseman, J. Montgomery, Gaussian, Inc., Wallington CT,
2004.
[71]
[72]
[73]
[74]
C. T. Lee, W. T. Yang, R. G. Parr, Phys. Rev. B 1988, 37, 785–
789.
[48] D. N. Shin, J. W. Wijnen, J. B. F. N. Engberts, A. Wakisaka, J.
Phys. Chem. B 2001, 105, 6759–6762.
[49] P. G. N. Suppan, in Solvatochromism, Royal Society of Chemis-
try, Cambridge, 1997, pp. 21–67.
A. D. Becke, J. Chem. Phys. 1993, 98, 5648–5652.
Received: August 29, 2007
Published Online: January 15, 2008
1180
www.eurjoc.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2008, 1165–1180