Thermosolvatochromism of merocyanine polarity indicators in pure and aqueous solvents: Relevance of solvent lipophilicity

Article abstract of DOI:10.1021/jo061533e

(Chemical Equation Presented) The following novel solvatochromic probes were synthesized: 2,6-dibromo-4-[(E)-2-(1-alkylpyridinium-4-yl)ethenyl] phenolate, where the alkyl groups are methyl, n-butyl, n-hexyl, and n-octyl, respectively. Solvatochromism of t

Full text of DOI:10.1021/jo061533e

Thermosolvatochromism of Merocyanine Polarity Indicators in Pure
and Aqueous Solvents: Relevance of Solvent Lipophilicity
Clarissa T. Martins, Michelle S. Lima, and Omar A. El Seoud*
Instituto de Qu´ımica, UniVersidade de Sa˜o Paulo, C.P. 26077, 05513-970, Sa˜o Paulo S.P., Brazil
elseoud@iq.usp.br
ReceiVed July 24, 2006
The following novel solvatochromic probes were synthesized: 2,6-dibromo-4-[(E)-2-(1-alkylpyridinium-
4-yl)ethenyl] phenolate, where the alkyl groups are methyl, n-butyl, n-hexyl, and n-octyl, respectively.
Solvatochromism of three of these probes (C1, C4, and C8) was studied in 36 protic and aprotic solvents.
A modified linear solvation energy relationship has been applied to the data obtained at 25 °C. Correlation
of (empirical) polarities with other solvent properties showed more dependence on lipophilicity than
on basicity. A similar conclusion has been reached for a series of other solvatochromic indicators.
Exceptions are those that carry acidic hydrogens, being biased toward solvent basicity. Thermosolvato-
chromism has been studied in mixtures of water with methanol, 1-propanol, acetonitrile, and
DMSO. Thermosolvatochromic data have been treated according to a model that explicitly considers
the presence in bulk solution of three “species”: water, organic component, and solvent-water
hydrogen-bonded aggregate. Solvation by the latter is favored over solvation by either of the two precursor
solvents (aqueous DMSO is an exception). Temperature increase resulted in desolvation of the probes,
due to concomitant decrease of the structures of the component solvents. The above-mentioned modified
solvation equation has been successfully applied to solvatochromism in aqueous methanol and aqueous
1-propanol.
Introduction
employed as a measure of lipophilicity or hydrophobic character;
it refers to the partition coefficient of a substance between
We have been interested in studying the effects of solvents
and binary solvent mixtures on the UV-vis spectra of solva-
tochromic probes or polarity indicators (hereafter referred to as
“probes”),1-5 examples of which are shown in Figure 1, along
with their pKa and log P. The latter property is extensively
n-octanol and water: log
P ) log([substance]n-octanol/
[substance]water).6 The probes shown in Figure 1 include 2,6-
diphenyl-4-(2,4,6-triphenylpyridinium-1-yl) phenolate (RB);
2,6-dichloro-4-(2,4,6-triphenyl pyridinium-1-yl) phenolate
(WB); 1-methylquinolinium-8-olate (QB); and 4-[(E)2-(1-me-
thylpyridinium-4-yl)ethenyl] phenolate (MePM) and 2,6-di-
bromo-4-[(E)-2-(1-butylquinolinium-4-yl)ethenyl] phenolate
(BuQMBr2), respectively. Note that log P is not available for
RB because its solubility in water is negligibly small, ca. 7.2
× 10-6 mol/L.7
* To whom correspondence should be addressed. Fax: +55-11-3091-3874.
(1) Tada, E. B.; Novaki, L. P.; El Seoud, O. A. J. Phys. Org. Chem.
2000, 13, 679-687.
(2) Antonious, M. S.; Tada, E. B.; El Seoud, O. A. J. Phys. Org. Chem.
2002, 15, 403-412.
(3) Tada, E. B.; Silva, P. L.; El Seoud, O. A. J. Phys. Org. Chem. 2003,
16, 691-699.
(4) Tada, E. B.; Silva, P. L.; El Seoud, O. A. Phys. Chem. Chem. Phys.
2003, 5, 5378-5385.
(6) Leo, A. J.; Hansch, C. Perspect. Drug DiscoVery Des. 1999, 17, 1-25.
(7) Reichardt, C. In SolVents and SolVent Effects in Organic Chemistry,
3rd ed.; VCH: Weinheim, 2003; p 389.
(5) Tada, E. B.; Silva, P. L.; Tavares, C.; El Seoud, O. A. J. Phys. Org.
Chem. 2005, 18, 398-407.
10.1021/jo061533e CCC: $33.50 © 2006 American Chemical Society
Published on Web 11/03/2006
9068
J. Org. Chem. 2006, 71, 9068-9079

ThermosolVatochromism of Polarity Indicators
FIGURE 1. Structures, pKa (conjugate acid), and log P of some solvatochromic probes.1-5,7
An empirical solvent polarity scale, ET(probe), is calculated
from the UV-vis or fluorescence spectral data, as shown by
eq 1:
although RB is much more basic than WB, the response of both
probes to solvent “acidity” or hydrogen-bond donation is similar.
Briefly, whereas the ability of the solvent to form hydrogen
bonds with the phenolate oxygen of RB is attenuated as a result
steric hindrance by the two ortho phenyl rings,9 the correspond-
ing ability of WB is enhanced because of lower steric hindrance
around the phenolate oxygen and the additional ability of two
ortho chlorine atoms to form hydrogen bonds.1
ET (probe) ) 28591.5/λmax (nm)
(1)
This equation converts the electronic transition within the probe
into the corresponding intramolecular charge-transfer transition
energy in kcal/mol.7 The solvent polarity scales of the probes
depicted in Figure 1 are referred to as ET(30), ET(33), ET(QB),
ET(MePM), and ET(BuQMBr2), respectively. The dependence
of ET(probe) on solvent properties and, for binary solvent
mixtures, on medium composition shed light on the relative
importance of factors that contribute to solvation, in particular
the pKa and hydrophobic/hydrophilic character of both probe
and solvent. We have also been interested in studying thermo-
solvatochromism (solvatochromism at different temperatures)
because this bears on the effects of temperature on solvation
and, in binary solvent mixtures, effects of temperature on the
composition of the probe solvation coordination shell. An
important application of this research is that the results obtained
may be employed to understand effects of the medium on rate
and equilibrium constants of chemical reactions. For example,
the nonlinear dependences of the rate constants of the pH-
independent hydrolysis of esters of different hydrophobicities
(4-nitrophenyl chloroformate and 4-nitrophenyl heptafluorobu-
tyrate, respectively) on medium composition (acetonitrile-
water) are remarkably similar to the dependence of ET(QB) and
ET(33), respectively, on the same experimental variable. That
is, the more hydrophilic probe, QB, serves as a model for the
more hydrophilic ester (4-nitrophenyl chloroformate), whereas
the more hydrophobic probe, WB, serves as a model for the
perflurobutyrate ester.8 This similar responses to solvent com-
position is very interesting because their origins are distinct,
namely, a chemical reaction and an electronic transition,
respectively.
To address this problem, we have synthesized a series of
novel merocyanine probes, RPMBr2 where R ) alkyl group,
of identical pKa and increasing hydrophobic character (see
Figure 2); these were employed to measure the polarity of 36
protic and aprotic solvents. Values of ET(RPMBr2) were found
to correlate linearly with ET(30), which shows that the probes
synthesized are sensitive to the same solute-solvent interactions
as RB. Application of a modified multiparameter solvation
equation has shown that ET(RPMBr2) are sensitive to solvent
dipolarity/polarizability, acidity, and lipophilicity but are little
dependent on solvent basicity. We have also studied the
thermosolvatochromism of RPMBr2 in mixtures of water (W)
with methanol (MeOH), 1-propanol (PrOH), acetonitrile (MeCN),
and dimethylsulfoxide (DMSO). Values of ET(RPMBr2) were
found to increase as a function of increasing hydrophobicity of
the probe; ET(RPMBr2)/CH2 of the probe alkyl group was found
to be similar to several Gibbs free energies of transfer of, e.g.,
alkylammonium halides from water to binary solvent mixtures.
Temperature increase resulted in a gradual desolvation of the
probes due to temperature-induced solvent structure perturbation.
The modified multiparameter solvation equation was satisfac-
torily applied to MeOA-W and PrOH-W mixtures.
Results and Discussion
Probes Synthesized. A comment on the probes synthesized
is in order. As expected, their pKa’s are identical, since Hammett
σ
para of alkyl groups are similar, e.g., -0.17, -0.16, -0.15, for
methyl, n-butyl, and n-pentyl, respectively;10 these groups are
attached to the pyridinium ring where the small differences in
their inductive effects are not transmitted to the phenolate
oxygen. The solubility of OcPMBr2 in water is very low; this
introduces uncertainties in the determination of its ET in water
and of log P. The problem was solved as follows: ET for this
The preceding paragraph shows that a clearer understanding
of solvation requires the use of probes whose structures are
modified in a systematic manner. The ones shown in Figure 1
differ widely in structure and hence in physicochemical proper-
ties that are relevant to their solvation. Consequently, quantifica-
tion of the effects of a single property on solvation, e.g., pKa or
hydrophobicity, is not feasible because these properties change
simultaneously for each pair of probes depicted. For example,
(9) Coleman, C. A.; Murray, C. J. J. Org. Chem. 1992, 57, 3578-3582.
(10) Hansch, C.; Leo, A.; Taft, R. W. Chem. ReV. 1991, 91, 165-195.
(11) Reichardt, C. Chem. ReV. 1994, 94, 2319-2358.
(8) Siviero, F.; El Seoud, O. A. J. Phys. Org. Chem. 2006, in press.
J. Org. Chem, Vol. 71, No. 24, 2006 9069

Martins et al.
FIGURE 2. Structures, pKa (conjugate acid), and log
P
of the merocyanine dyes synthesized. Groups longer than methyl are
n-alkyl chains. See Experimental Section for determination of pKa and log P.
probe was determined at different temperatures in binary
mixtures of water ([W] ) 40-52 mol/L or øw ) 0.90-0.98,
where ø refers to mole fraction) with the following solvents:
MeOH, acetone (from 10 to 40 °C), PrOH, and MeCN (from
10 to 60 °C). Plots of ET(OcPMBr2) in the different binary
mixtures versus øw of water were found to be linear and
converged to the value in pure water, as shown in Figure SI-1
(Figure 1 of Supporting Information). To check the validity of
this approach, we repeated the experiment with HxPMBr2, at
25 °C, in the same binary mixtures. Note that this probe is
water-soluble; its ET can be directly determined in water.
Both ET(HxPMBr2), i.e., that determined directly and by
extrapolation, agreed within experimental uncertainty, 64.80 (
0.15 kcal/mol.
where SD is the standard deviation. This result is satisfying
and can be explained by the fact that all probes are zwitterionic;
the solvatochromic shift involves a π f π* transition in the
UV-vis region. That is, all merocyanine probes are sensitive
to the same solute-solvent interactions as RB, e.g., Coulombic,
dispersion, and hydrogen bonding. The magnitude of regression
coefficients will be discussed below.
ET(MePMBr2) ) 21.774 + 0.668ET(30), r )
0.9685,SD ) 1.1230 (2)
ET(BuPMBr2) ) 21.547 + 0.667ET(30), r )
0.9727,SD ) 1.0753 (3)
ET(OcPMBr2) ) 22.873 + 0.637ET(30), r )
0.9691,SD ) 1.0112 (4)
The presence of a long chain alkyl group in the probe may,
in principle, lead to its aggregation, especially in water or water-
rich binary mixtures. In water, the UV-vis spectra of MePMBr2,
BuPMBr2, and HxPMBr2 showed no changes in λmax and/or
peak shape as a function of [probe] in the range 10-4-10-3
mol/L. This indicates that no probe aggregation occurs under
our experimental conditions. For OcPMBr2, Beer’s law is
obeyed in the concentration range from 4 × 10-6 to 5 × 10-5
mol/L, so that [OcPMBr2] employed was 3 × 10-5 mol/L. The
low solubility of OcPMBr2 in water also precluded determination
of a reliable log P. Therefore the corresponding value, 2.70 (
0.1, was determined by extrapolation from the (linear) plot of
log P versus number of carbon atoms in the probe alkyl chain,
Nc, for MePMBr2, BuPMBr2, and HxPMBr2, respectively (log
P ) -0.548 + 0.406 Nc; r ) 0.9993, where r is the correlation
coefficient). In summary, this novel series of probes is especially
adequate to determine the relative importance of probe lipo-
philicity to its solvation.
Effects of solvent properties on ET(probe) have been explained
in terms of multiparameter equations, e.g., that of Taft-Kamlet-
Abboud, eq 5, for a single probe in a series of solvents:12-14
ET(probe) ) constant + s (π/solv + dδ) + a Rsolv
+
b âsolv + h (δ2H) (5)
Here ET(probe) is modeled as a linear combination of a
dipolarity/polarizability term [s (π/solv + dδ)], two hydrogen-
bonding terms, in which the solvent is the hydrogen-bond donor
(a Rsolv) and/or the hydrogen-bond acceptor (b âsolv), and a cavity
term (h (δ2H)). The later is redundant when the “Frank-Condon
principle” is obeyed, as in case of the probes studied. The
parameters π/solv, Rsolv, and âsolv are known as solvatochromic
parameters; we use the subscript (solv) so that they are not
confused with the same symbols, e.g., R and â of the Brønsted
catalysis equation. In applying eq 5, care should be exercised
in order to obtain meaningful statistical correlations. For
example, the solvatochromic parameters tested should not
correlate linearly, and a sufficient number of solvents, usually
Solvatochromism in Pure Solvents: A Modified Equation
for Correlating ET(probe) with Solvent Properties. Note:
Details of all calculations performed are given in Calculations
in Supporting Information.
Values of ET(RPMBr2) for 34 organic solvents, including 18
(12) Kamlet, M. J.; Abboud, J. L. M.; Taft, R. W. Prog. Phys. Org.
Chem. 1981, 13, 485-630.
(13) Abraham, M. H.; Grellier, P. L.; Abboud, J. L. M.; Doherty, R.
M.; Taft, R. W. Can. J. Chem. 1988, 66, 2673-2686.
(14) Laurence, C.; Nicolet, P.; Dalati, M. T.; Abboud, J. L. M.; Notario,
R. J. Phys. Chem. 1994, 98, 5807-5816.
protic, 4 chlorinated, and 12 polar aprotic are listed in Table 1.
The corresponding values for H2O and D2O are listed in footnote
a of the same table. ET(RPMBr2) were found to correlate linearly
with the ET(30) scale, as shown in Figure 3 and by eqs 2-4,
9070 J. Org. Chem., Vol. 71, No. 24, 2006

ThermosolVatochromism of Polarity Indicators
TABLE 1. Solvent Polarity, ET(probe) (kcal mol-1, at 25 °C), Based on the Solvatochromic Probes MePMBr2, BuPMBr2, and OcPMBr2,
Respectivelya
solvent
ET(MePMBr2)
ET(BuPMBr2)
ET(OcPMBr2)
Normal-Chain Alcohols
1
2
3
4
5
6
methanol
ethanol
1-propanol
1-butanol
1-hexanol
1-octanol
59.24
56.03
54.88
54.15
53.07
52.26
58.65
55.60
54.40
53.75
52.81
52.08
58.54
55.56
54.40
53.74
52.81
51.99
Branched-Chain Alcohols, Other Alcohols, 2-Alkoxyethanols
7
8
9
2-propanol
2-butanol
2-methyl-2-propanol
3-methyl-1-butanol
1,2-ethanediol
benzyl alcohol
cyclohexanol
2-methoxyethanol
2-ethoxyethanol
2-propoxyethanol
2-butoxyethanol
53.54
52.29
50.59
53.32
61.27
54.56
52.50
56.97
55.52
54.60
54.10
55.42
53.07
53.02
51.90
50.30
52.97
60.70
54.25
52.12
56.28
54.96
54.22
53.69
54.76
52.01
50.40
52.97
60.79
54.28
52.16
56.43
54.99
54.18
53.72
54.76
10
11
12
13
14
15
16
17
18
2-(2-methoxy-ethoxy)ethanol
Chlorinated and Aromatic Solvents
19
20
21
22
chloroform
46.12
48.13
48.55
46.246
48.134
48.522
45.71
46.25
48.15
48.52
45.70
dichloromethane
1,2-dichloroethane
chlorobenzene
Polar Aprotic Solvents
23
24
25
26
27
28
29
30
31
32
33
34
acetone
acetonitrile
N,N-dimethylacetamide
N,N-dimethylformamide
1,3-dimethyl-2-imidazolidinone
DMSO
1,4-dioxane
ethyl acetate
ethylene glycol dimethylether
nitromethane
pyridine
51.06
53.32
51.98
52.39
51.83
53.41
45.69
50.83
52.89
51.59
52.00
51.23
52.68
45.65
47.61
48.32
52.85
49.40
47.40
50.82
52.99
51.52
52.02
51.23
53.02
45.65
47.55
48.30
52.83
49.40
47.44
48.72
53.14
49.66
47.69
THF
a ET(probe) values of water were found to be 65.24, 64.98, and 64.65 kcal/mol for MePMBr2, BuPMBr2 and OcPMBr2, respectively. The last value was
determined by extrapolation; see text for details. ET(probe) values of D2O were found to be 65.57 and 65.06 kcal/mol for MePMBr2, BuPMBr2, respectively
property could not be assessed because of the small number of
probes and solvents tested. Therefore, we examined a possible
modification of eq 5, based on both chemistry and statistics,
namely, by adding a term for solvent hydrophobicity, as shown
by eq 6:
ET (probe) ) constant + s (π/solv + dδ) + a Rsolv
+
b âsolv + p log Psolv (6)
where log Psolv refers to the partition coefficient of the solVent
between n-octanol and water. This lipophilicity term has been
included by analogy to similar linear solvation energy relation-
ships for solutes. For example, solubilities, distribution between
immiscible solvents, and other properties that depend on solute-
solvent interactions have been modeled by equations that
correlate the property of interest with solvatochromic parameters
and molar volume of the solute.15,16. Log Psolv in n-octanol/W
has been employed because more data are available for this
biphasic system; in principle other lipophilicity scales may be
FIGURE 3. Plots of ET(RPMBr2) versus ET(30) for the solvents
investigated. Values of the former ET are from Table 1; the latter are
those published elsewhere.7,11
5 per each solvent property is employed;1-5,12,13 these conditions
are met in the present work.
(15) Kamlet, M. J.; Doherty, R. M.; Abraham, M. H.; Marcus, Y.; Taft,
R. W. J. Phys. Chem. 1988, 92, 5244-5255.
(16) Abraham, M. H. Chem. Soc. ReV. 1993, 22, 73-83.
Although we have shown that zwitterionic probes are sensitive
to solvent lipophilicity,1-5 the relative importance of this
J. Org. Chem, Vol. 71, No. 24, 2006 9071

Martins et al.
FIGURE 4. Structures of additional probes employed in the multicorrelation analysis equation (eq 6). P1, sodium-{4-[4-(4-carboxylatophenyl)-
2.6-diphenyl-1-pyridinio]-2.6-diphenolate; P2, 4-{2,4,6-tris[(4-methasulfonyl)phenyl]-1-pyridinio}-2,6-diphenylphenolate; P3, phenol blue; P4, N-(4-
nitrophenyl) pyrrolidine; P5, N-(4-nitrophenyl) piperidine; P6, pyridazine; P7, 4-nitrophenol; P8, N-methyl-4-nitroaniline; P9, N-ethyl-4-nitro-
aniline; P10, N-isopropyl-4-nitroaniline; P11, N,N-diethyl-4-nitroaniline; P12, 3-methyl-4-nitroaniline; P13, N-ethyl-4-carbomethoxy pyridinium
iodide.18
employed, e.g., those based on partition between water and
dichloroethane; chloroform or heptane. Note that these lipo-
philicity scales are linearly correlated with the n-octanol/W
system.17
and the solvent properties tested (Rsolv, âsolv, etc.) have different
scales. Use of the standardized coefficients, âstatistical, however,
solves this problem (see Calculations in Supporting Informa-
tion);19 these are listed in Table 2.
Because the cavity term was dropped, eq 5 contains three
solvent parameters, whereas eq 6 contains four. Therefore, it is
necessary to determine whether including an additional term is
statistically significant; alternatively, whether âsolv may be
dropped from eq 6, leading to a three-term equation. This
statistical test relies on (1) use of a sufficiently large solvato-
chromic data set; (2) use of a stepwise correlation procedure,
i.e., ET (probe) is correlated with two, three, and four solvent
parameters followed, in each case, by examination of the
goodness of fit. With regard to point 1, we used the solvato-
chromic data of 21 probes, namely, the five probes shown in
Figures 1, three of the probes of Figure 2 (R ) Me, Bu, and
Oc) and the 13 probes, P1, P2, etc., of Figure 4 .18 The probes
selected represent different chemical classes, this is expected
to lead to differences in the contributions of solvation mecha-
nisms (hydrogen bonding; solute dipole-solvent dipole interac-
tions, etc.); the number of solvents tested is large enough to
secure statistically valid correlations, vide supra. Point 1 will
be examined first, based on four solvent parameters.
Several points merit comments:
1. We address the relative magnitudes of âstatistical of basicity
b and lipophilicity p. Although data scatter for the former
property is high, âstatistical of p > âstatistical of b for the probes
synthesized and for MePM, BuQMBr2, RB, WB, P1, P6, P11,
and P13. This indicates that these probes are more sensitive to
solvent lipophilicity than to its basicity, as argued above. The
only probes for which âstatistical of b > âstatistical of p are those
that carry a relatively acidic hydrogen, namely, P7, P8, P9, P10,
and P12, being understandably more sensitive to solvent basicity.
In fact, these probes have been employed to determine âsolv
because their solvatochromic behavior is dominated by their
response to solvent basicity.20
2. With regard to point 2 above, the minimum number of
variables that satisfactorily describes the phenomenon investi-
gated (dependence of ET(probe) on solvent properties) should
be sought.21 Therefore, we examined the dependence of the
multiple-linear regression coefficient (r2) and âstatistical on the
number of solvent properties employed. It is not practical to
list and discuss the results of this stepwise variation for the 21
probes examined, due to the excessive number of regression
equations (84) resulting from, per probe, one equation with two
variables, two equations with three variables, and one equation
with four variables. Therefore, we discuss the results of three
The regression coefficients of eq 6 are listed in Table SI-1
(Table 1 in Supporting Information). They permit a comparison
of the response of different probes to the same solvent property,
e.g., acidity or basicity. They do not permit, however, a direct
comparison of the relative importance of solvent properties to
the solvation of different probes. The reason is that ET(probe)
(19) Hill, T.; Lewicki, P. In Statics Methods and Applications, A
ComprehensiVe Reference for Science, Industry and Data Mining, 1st ed.;
StatSoft: Tulsa, 2006; pp 555-757.
(20) Kamlet, M. J.; Taft, R. W. J. Am. Chem. Soc. 1976, 98, 377-383.
(21) Krygowski, T. M.; Radomski, J. P.; Rzeszowiak, A.; Wrona, P. K.;
Reichardt, C. Tetrahedron 1981, 37, 119-125.
(17) Steyaert, G.; Lisa, G.; Gaillard, P.; Boss, G.; Reymond, F.; Girault,
H. H.; Carrupt, P. A.; Testa, B. J. Chem. Soc,. Faraday Trans. 1997, 93,
401-406.
(18) Novaki, L. P.; El Seoud, O. A. Ber. Bunsenges. Phys. Chem. 1996,
100, 648-655 and references cited therein.
9072 J. Org. Chem., Vol. 71, No. 24, 2006

ThermosolVatochromism of Polarity Indicators
TABLE 2. âstatistical Coefficients of Equation 6 for the Probes Synthesized and Those of Figures 1, 2, and 4
probe
â
statistical (s)
â
statistical (a)
â
statistical (b)
â
statistical (p)
r2
na
MePMBr2
BuPMBr2
OcPMBr2
MePM
BuQMBr2
RB
WB
QB
P1
P2
P3
P4
P5
P6
P7
P8
P9
P10
0.46 (( 0.07)
0.44 (( 0.07)
0.43 (( 0.07)
0.43 (( 0.09)
0.28 (( 0.06)
0.47 (( 0.09)
0.37 (( 0.07)
0.29 (( 0.07)
0.28 (( 0.06)
0.41 (( 0.07)
-0.84 (( 0.18)
-0.70 (( 0.12)
-0.67 (( 0.12)
0.34 (( 0.06)
-0.52 (( 0.09)
-0.72 (( 0.06)
-0.74 (( 0.12)
-0.68 (( 0.05)
-0.86 (( 0.06)
-0.50 (( 0.04)
0.30 (( 0.08)
0.81 (( 0.06)
0.80 (( 0.06)
0.79 (( 0.06)
0.92 (( 0.08)
0.82 (( 0.07)
0.80 (( 0.06)
0.80 (( 0.06)
0.85 (( 0.06)
0.88 (( 0.05)
0.89 (( 0.07)
-0.13 (( 0.06)
-0.34 (( 0.09)
-0.32 (( 0.09)
0.86 (( 0.03)
-0.07 (( 0.06)
-0.22 (( 0.05)
0.14 (( 0.11)
-0.23 (( 0.04)
-0.29 (( 0.05)
-0.15 (( 0.04)
0.85 (( 0.05)
0.02 (( 0.07)
-0.01 (( 0.07)
-0.01 (( 0.07)
-0.06 (( 0.08)
-0.23 (( 0.09)
0.03 (( 0.06)
-0.02 (( 0.09)
0.01 (( 0.07)
-0.08 (( 0.05)
0.08 (( 0.08)
0.09 (( 0.09)
0.23 (( 0.10)
0.13 (( 0.11)
-0.09 (( 0.05)
-0.73 (( 0.07)
-0.54 (( 0.07)
-0.48 (( 0.13)
-0.45 (( 0.05)
0.08 (( 0.07)
0.86 (( 0.04)
-0.02 (( 0.06)
-0.24 (( 0.07)
-0.21 (( 0.07)
-0.22 (( 0.07)
-0.15 (( 0.07)
-0.35 (( 0.08)
-0.06 (( 0.09)
-0.08 (( 0.08)
-0.02 (( 0.10)
-0.14 (( 0.06)
0.08 (( 0.09)
0.02 (( 0.19)
0.12 (( 0.15)
0.12 (( 0.15)
-0.11 (( 0.07)
0 (( 0.12)
0.03 (( 0.09)
0.02 (( 0.18)
0.14 (( 0.08)
0.16 (( 0.09)
0.04 (( 0.05)
-0.15 (( 0.11)
0.9258
0.9334
0.9317
0.9542
0.9011
0.8677
0.9412
0.9415
0.9704
0.9500
0.9459
0.8867
0.8880
0.9902
0.9491
0.9672
0.8748
0.9777
0.9664
0.9833
0.9746
36
35
35
36
36
57
25
24
19
20
23
24
23
20
23
21
21
21
21
20
20
P11
P12
P13
a Number of solvents tested.
TABLE 3. Results Obtained by Stepwise Multilinear Regression Analysis of the Dependence of ET(probe) on Solvent Propertiesa
ET(BuPMBr2)
r2
ET(30)
r2
ET(P7)
r2
ET(probe) ) a Rsolv + s (π/solv + d δ)
0.78 (( 0.06) 0.9140 0.81 (( 0.05)
0.52 (( 0.07) 0.43 (( 0.05)
ET(P7) ) b â solv + s (π/solv + dδ)
-0.74 (( 0.07)
â
statistical (a)
statistical (s)
0.8668
â
â
statistical (b)
statistical (s)
0.9448
0.9448
â
-0.49 (( 0.06)
ET(probe) ) a Rsolv + s (π/ + d δ) + p log Psolv
ET(P7) ) b â solv + s ( π/ + dδ) + p log P solv
âstatistical (a)
âstatistical (s)
âstatistical (p)
0.79 (( 0.05)
0.44 (( 0.06)
-0.21 (( 0.07)
0.9s3o3lv7
0.81 (( 0.05)
0.47 (( 0.09)
0.06 (( 0.09)
0.8673
0.8671
0.8677
â
â
â
statistical (b)
statistical (s)
statistical (p)
-0.74s(o(lv 0.07)
-0.49 (( 0.09)
0 (( 0.11)
ET(probe) ) a Rsolv + s (π/solv + d δ) + b âsolv
ET(P7) ) b âsolv + s (π/solv + d δ) + a R solv
-0.72 (0.07)
âstatistical (a)
âstatistical (s)
âstatistical (b)
0.77 (( 0.07)
0.56 (( 0.06)
0.02 (( 0.07)
0.9139
0.80 (( 0.06)
0.02 (( 0.06)
0.42 (( 0.05)
â
â
â
statistical (b)
statistical (s)
statistical (a)
0.9485
-0.49 (0.06)
-0.07 (0.06)
ET(probe) ) a Rsolv + s (π/solv + d δ) + p log Psolv + b âsolv
ET(P7) ) b âsolv + s (π/solvv + d δ) + p log P solv + a R solv
âstatistical (a)
âstatistical (s)
âstatistical (p)
âstatistical (b)
0.80 (( 0.06)
0.44 (( 0.07)
-0.21 (( 0.07)
-0.13 (( 0.07)
0.9338
0.47 (( 0.09)
0.80 (( 0.06)
0.03 (( 0.06)
-0.06 (( 0.09)
â
â
â
â
statistical (b)
-0.73 (( 0.07)
-0.52 (( 0.10)
-0.01 (( 0.08)
-0.08 (( 0.06)
0.9491
statistical (s)
statistical (p)
statistical (a)
b- ET(probe) data at 25 ïC, taken from the references listed in Table 2.
representative probes, namely, BuPMBr2 (model for merocya-
nines), RB (model for WB and QB), and P7 (probe biased
toward solvent basicity, model for P8 and P12); see Table 3.
Table 3 shows that ET(BuPMBr2) is reasonably correlated
by two solvent properties, namely, Rsolv and s π/solv. Inclusion
of log P resulted in a better fit to the data (increase of r2) relative
to inclusion of âsolv; use of four solvent properties produced no
further improvement. On the other hand, ET(30) and ET(P7) may
of tetra-alkylammonium halides on solvent properties did not
increase the overall correlation coefficient, so âsolv was dropped.23
As expected, only solvent basicity and dipolarity/polarizability
are relevant to solvation of P7.
3. The slopes of eq 2-4 are ca. 0.65, indicating that the
oVerall solvatochromic responses of the probes synthesized are
lower than that of RB. What is relevant, however, is the probe
response to indiVidual properties of the solvent, as given by
the left-hand terms of eq 6. Interestingly, âstatistical of s, a, and
b for the probes synthesized are not different from those of RB;
they are more sensitive to solvent hydrophobicity. On the basis
of these results, they may substitute RB where this is advanta-
geous, e.g., in relatively acidic buffered solutions, where
protonation of its phenolate oxygen leads to loss of solvato-
chromic response;
be conveniently described by two solvent properties, Rsolv
,
π/solv and âsolv, πs/olv, respectively.
It is possible that the lower significance of âsolv is due to
inefficient interactions between the solvent (as electron donor)
and the heterocyclic quaternary nitrogen of the zwitterionic
probe. For example, whereas the (CH3)4N+ ion has no effect
on the structure of water, (n-C4H9)4N+ has a net structure-
enhancing effect due to hydrophobic hydration of the alkyl
groups.22 On the other hand, addition of âsolv to the equation
that describes the dependence of Gibbs free energies of solution
4. Introduction of two bromine atoms in the structure of the
precursor merocyanine, MePM, has resulted in the response
expected, i.e., a decrease in âstatistical of a, due to lower pKa,
(22) El Seoud, O. A. J. Mol. Liq. 1997, 72, 85-103 and references cited
therein.
(23) Taft, R. W.; Abraham, M. H.; Doherty, R. M.; Kamlet, M. J. J.
Am. Chem. Soc. 1985, 107, 3105-3110.
J. Org. Chem, Vol. 71, No. 24, 2006 9073

Martins et al.
FIGURE 5. Dependence of the empirical solvent polarity parameter ET(RPMBr2) on the mole fraction of water, øW, at 25 °C, for mixtures of water
with methanol, MeOH, 1-propanol, PrOH, acetonitrile, MeCN, and dimethylsulfoxide (DMSO), respectively. The straight lines were plotted to
guide the eye and represent ideal solvation of the dye by the mixture; see text for details. Probe symbols: (0) MePMBr2; (O) BuPMBr2; and(4)
OcPMBr2.
and an increase in âstatistical of p, because bromine is more
lipophilic than hydrogen.24
for this solvent and not for others of still higher ꢀ () 78.36,
32.66, 20.45, 46.45, and 35.94 for W, MeOH, PrOH, DMSO,
and MeCN, respectively).
In summary, the Taft-Kamlet-Abboud equation may be
expanded to include solvent lipophilicity, as given by log P (or
any equivalent scale). At least for merocyanines, the basicity
term may be dropped from eq 6, so that solvent polarity is
described in terms of its acidity, dipolarity/polarizability and
lipophilicity. Equation 6 is general, unless the structure of the
probe makes it particularly sensitive (or biased) toward a single
property of the solvent.
Thermosolvatochromism in Binary Solvent Mixtures.
Nonlinear Dependence of ET(RPMBr2) on øW. Figure 5 shows
the dependence of ET(RPMBr2) on solvent composition at 25
°C, for the three probes synthesized, in four binary mixtures.
All plots shown in Figure 5 are nonlinear; this may be attributed
to several factors and/or solute-solvent interaction mechanisms.
Nonideal behavior may originate from dielectric enrichment,
i.e., enrichment of the solvation coordination shell in the solvent
of higher relative permittivity, ꢀ.25 This mechanism may be
rejected, however, because if dielectric enrichment were opera-
tive, all curves of Figure 5 should lie above, not below, the
straight line that connects the polarities of the two pure liquids.
A part of the data of MeCN lies above the line, but there is no
reason to believe that dielectric enrichment is operative only
Another reason for nonideal behavior is preferential solvation
of the probe by a component of the mixture due to solute-
solvent specific interactions, e.g., hydrogen bonding and dipole-
dipole interactions. A large body of experimental data and
theoretical calculations, e.g., of the Buff-Kirkwood integral
functions (that describe W-W, Solv-Solv, and Solv-W
interactions), has shown that the binary mixtures employed are
microheterogeneous; there exist microdomains composed of
organic solvent surrounded by water and of water solvated by
organic solvent. The onset and composition of these micro-
domains depend on the pair of solvents. There exists the
possibility of preferential solvation of the probe in the less polar
microdomains, leading to below-the-line deviation, as shown
in Figure 5.25-28 In summary, nonideal solvation behavior is
not unexpected.
Response of ET(probe) to Lipophilic Character of the
Probe. Figure 5 clearly shows that the deviation from linearity
increases as a function of increasing Nc of the probe, i.e., its
hydrophobic character. Plots (not shown) of ET(solvent) versus
Nc gave slopes of ca. 0.1 kcal/CH2 in water and 0.33 ( 0.18
kcal/CH2 at the points of maximum deviation from linearity
(24) Meylan, W. M.; Howard, P. H. J. Pharm. Sci. 1995, 84, 83-92.
(25) Suppan, P. G. N. In SolVatochromism; The Royal Society of
Chemistry: Cambridge, 1997; pp 21-67 and references cited therein.
(26) Marcus, Y. Chem. Soc. ReV. 1993, 22, 409-416.
(27) Shulgin, I.; Ruckenstein, E. J. Phys. Chem. B 1999, 103, 872-877.
(28) Marcus, Y. Monatsh. Chem. 2001, 132, 1387-1411.
9074 J. Org. Chem., Vol. 71, No. 24, 2006

ThermosolVatochromism of Polarity Indicators
FIGURE 6. Solvent polarity/temperature/solvent composition contours for MePMBr2, BuPMBr2, and OcPMBr2 in MeOH/W.
Probe(W)m + m(Solv-W) a Probe(Solv-W)m +
m W (10)
for the binary mixtures. These values are of the same magnitude
as those calculated for other systems, e.g., Gibbs free energies
of transfer of alkylammonium ions from water to binary solvent
mixtures, at øw ) 0.5; 0.06 and 0.3 kcal/CH2 for MeCN-W
and MeOH-W, respectively;29,30 experimental and theoretically
calculated Gibbs free energies of solvation of alkanes and
n-chain alcohols in water, 0.18 ( 0.05 kcal/CH2 and 0.16 (
0.05 kcal/CH2, respectively.30 Therefore, solute-solvent hy-
drophobic interactions are relevant to ET(probe), as argued
elsewhere.1-5 To our knowledge, this is the first time that
solvatochromic probes have been employed to obtain solvation
energy/CH2. Compared to other methods, the present one has
the merit of versatility and experimental simplicity.
where m represents the number of solvent molecules whose
exchange in the probe solvation coordination shell affects ET
(probe); usually m e 2 (m should not be confused with the total
number of solvent molecules that solvate the probe). The use
of 1:1 stoichiometry for Solv-W is a practical and convenient
assumption because it renders subsequent calculations tractable;
it has been extensively employed by others to describe solva-
tochromism31,32 Mixed solvent species with stoichiometry other
than 1:1 may be treated, to a good approximation, as mixtures
of the 1:1 structure plus excess of a pure solvent. The relevant
point about this model is that it explicitly considers the formation
of hydrogen-bonded (or complex) solvent species Solv-W.
Consequently, the mole fractions employed in all calculations
(except those of Table SI-2) are “effectiVe” not analytical ones.
The equilibrium constants of eqs 8-10 are termed solvent
“fractionation factors, æ”. These are defined on the mole fraction
scale, after rearrangement, as:
Thermosolvatochromism. Thermosolvatochromism of RP-
MBr2 has been studied in mixtures of water with MeOH, PrOH,
MeCN, and DMSO, over the whole composition range. The
solvent polarity/temperature/solvent composition contours for
the three indicators synthesized are shown Figure 6 for MeOH.
The corresponding contours for PrOH, MeCN and DMSO are
shown in Figure SI-2 in Supporting Information.
xPWrobe/xProbe
Solv
Considering these results, the following is relevant:
æW/Solv
)
(11)
(12)
(13)
(xBWk;Effective/xBk;Effective m
)
5. Instead of reporting extensive lists of ET(RPMBr2) and
solvent compositions, we have calculated the (polynomial)
dependence of polarity on the analytical mole fraction of water
and present the regression coefficients in Table SI-2 in Sup-
porting Information. The degree of polynomial employed is that
which gave the best data fit, as indicated by the multiple
correlation coefficients, r2 and SD. For example, the data for
PrOH-W could have been conveniently adjusted with a fifth-
power polynomial. The quality of our data is evidenced by these
statistical criteria and by the excellent agreement between
calculated and experimental ET(RPMBr2)solv, at all temperatures;
see Table SI-2 in Supporting Information.
Solv
xPSroolvb-e W/xProbe
Solv
æSolv-W/Solv
)
Bk;Effective Bk;Effective m
(x
/xSolv
)
Solv-W
xPSroolvb-e W/xWProbe
æSolv-W/W
)
(x
/xBWk;Effective m
)
Bk;Effective
Solv-W
where Bk refers to bulk solvent. In eq 11, æW/Solv describes the
composition of the probe solvation coordination shell, relative
to that of bulk solvent. For æW/Solv > 1, the solvation coordina-
tion shell is richer in W than bulk solvent; the converse is true
for æW/Solv < 1, i.e., the probe is preferentially solvated by the
organic solvent. Finally, a solvent fractionation factor of unity
indicates an ideal behavior, i.e., the solvation coordination shell
and bulk solvent have the same composition. The same line of
reasoning applies to æSolv-W/Solv (complex solvent displaces
organic solvent) and æSolv-W/W (complex solvent displaces W),
depicted in eqs 12 and 13, respectively.
6. We have treated the data obtained according to the
following solvation model:3-5
Solv + W a Solv-W
(7)
Probe(Solv)m + m(W) a Probe(W)m + m Solv (8)
7. Results of application of eqs 11-13 are listed in Table 4.
The fit of the model to our thermosolvatochromic data is shown
by values of (r2) and ø2 and by the excellent agreement between
Probe(Solv)m + m(Solv-W) a Probe(Solv-W)m +
m solv (9)
(29) Hefter, G.; Marcus, Y.; Waghorne, W. E. Chem. ReV. 2002, 102,
2773-2835.
(30) Mendes, C. L. D.; da Silva, C. O.; da Silva, E. C. J. Phys. Chem.
A 2006, 110, 4034-4041.
(31) Rafols, C.; Roses, M.; Bosch, E. J. Chem. Soc. Perkin Trans. 1997,
2, 243-248.
(32) Buhvestov, U.; Rived, F.; Rafols, C.; Bosch, E.; Roses, M. J. Phys.
Org. Chem. 1998, 11, 185-192.
J. Org. Chem, Vol. 71, No. 24, 2006 9075

Martins et al.
TABLE 4. Analysis of Thermosolvatochromic Responses of MePMBr2, BuPMBr2, and OcPMBr2 in Solvent/Water Mixtures, According to
Equations 11-13
organic
solvent
probe
T,°C
m
æW/Solv æSolv-W/Solv æSolv-W/W
ET(probe)Solv
ET(probe)W
ET(probe)Solv-W
r2
ø2
MeOH
MePMBr2
10
25
40
10
25
40
10
25
40
10
25
40
60
10
25
40
60
10
25
40
60
10
25
40
60
10
25
40
60
10
25
40
60
25
40
60
25
40
60
25
40
60
0.971
1.007
0.870
1.264
1.115
1.056
1.289
1.246
1.147
1.580
1.359
1.300
1.110
1.652
1.430
1.341
1.323
1.696
1.618
1.512
1.442
0.994
1.057
1.015
0.978
1.067
1.064
1.064
1.006
1.092
1.140
1.121
1.050
0.768
0.745
0.703
0.805
0.783
0.756
0.911
0.829
0.800
0.567
0.578
0.610
0.524
0.539
0.544
0.474
0.498
0.52
1.867
1.805
1.756
2.618
2.400
2.05
2.676
2.544
2.287
3.293
3.123
2.879
4.996
4.453
3.768
5.646
5.108
4.398
59.77 [( 0.04] 65.96 [( 0.04] 61.32 [( 0.2]
59.27 [( 0.03] 65.27 [( 0.02] 60.99 [( 0.03]
58.55 [( 0.04] 65.17 [( 0.03] 60.67 [( 0.21]
59.31 [( 0.04] 65.26 [( 0.04] 60.90 [( 0.13]
58.70 [( 0.04] 64.96 [( 0.04] 60.44 [( 0.14]
58.10 [( 0.05] 64.73 [( 0.05] 59.98 [( 0.22]
59.05 [( 0.03] 65.23 [( 0.03] 61.08 [( 0.07]
58.55 [( 0.04] 64.62 [( 0.04] 60.47 [( 0.12]
57.97 [( 0.05] 64.51 [( 0.04] 60.02 [( 0.16]
55.45 [( 0.06] 66.00 [( 0.08] 59.76 [( 0.11]
54.95 [( 0.08] 65.42 [( 0.11] 59.68 [( 0.25]
54.34 [( 0.09] 65.21 [( 0.12] 59.35 [( 0.33]
53.70 [( 0.10] 64.88 [( 0.13] 59.60 [( 0.70]
54.99 [( 0.08] 65.30 [( 0.12] 58.70 [( 0.17]
54.54 [( 0.09] 65.03 [( 0.13] 58.56 [( 0.26]
53.99 [( 0.05] 64.79 [( 0.06] 58.14 [( 0.15]
53.42 [( 0.05] 64.46 [( 0.07] 57.69 [( 0.22]
55.08 [( 0.09] 65.30 [( 0.14] 58.31 [( 0.17]
54.54 [( 0.07] 64.71 [( 0.09] 58.00 [( 0.13]
54.02 [( 0.05] 64.53 [( 0.07] 57.82 [( 0.11]
53.41 [( 0.05] 64.28 [( 0.07] 57.37 [( 0.12]
53.52 [( 0.08] 65.99 [( 0.08] 60.14 [( 0.29]
53.39 [( 0.14] 65.36 [( 0.13] 59.58 [( 0.69]
52.93 [( 0.06] 65.21 [( 0.06] 58.90 [( 0.42]
52.53 [( 0.04] 64.85 [( 0.04] 57.67 [( 0.46]
53.40 [( 0.08] 65.30 [( 0.08] 59.46 [( 0.30]
52.94 [( 0.06] 65.00 [( 0.06] 59.02 [( 0.25]
52.60 [( 0.03] 64.75 [( 0.03] 58.48 [( 0.17]
52.11 [( 0.05] 64.46 [( 0.05] 57.33 [( 0.45]
53.27 [( 0.08] 65.25 [( 0.08] 59.49 [( 0.21]
53.05 [( 0.09] 64.73 [( 0.09] 59.08 [( 0.34]
52.74 [( 0.04] 64.53 [( 0.04] 58.40 [( 0.18]
52.34 [( 0.04] 64.27 [( 0.04] 57.13 [( 0.36]
53.33 [( 0.05] 65.25 [( 0.06] 57.14 [( 7.00]
53.12 [( 0.03] 65.13 [( 0.06] 54.05 [( 6.70]
52.74 [( 0.06] 64.81 [( 0.06] 52.93 [( 8.25]
52.95 [( 0.06] 64.93 [( 0.07] 57.22 [( 7.76]
52.66 [( 0.05] 64.77 [( 0.06] 57.33 [( 8.88]
52.46 [( 0.06] 64.43 [( 0.06] 52.60 [( 6.00]
0.9996 0.0022
0.9999 0.0008
0.9998 0.0015
0.9994 0.0031
0.9997 0.0023
0.9995 0.0033
0.9998 0.0011
0.9996 0.0022
0.9996 0.0023
0.9995 0.0069
0.9990 0.0133
0.9990 0.0142
0.9989 0.0166
0.9988 0.0154
0.9988 0.0164
0.9997 0.0042
0.9997 0.0049
0.9984 0.0191
0.9992 0.0093
0.9996 0.0046
0.9997 0.0045
0.9995 0.0075
0.9988 0.0199
0.9998 0.0043
0.9999 0.0018
0.9995 0.0077
0.9998 0.0037
0.9999 0.0011
0.9999 0.0029
0.9995 0.0072
0.9994 0.0099
0.9999 0.0016
0.9999 0.0022
0.9999 0.0036
0.9998 0.0044
0.9998 0.0045
0.9998 0.005
0.9998 0.004
0.9998 0.0043
BuPMBr2
OcPMBr2
MePMBr2
PrOH
0.211
0.215
0.233
0.239
0.200
0.208
0.229
0.233
0.185
0.200
0.219
0.226
1.461
1.494
1.527
1.578
1.446
1.450
1.481
1.512
1.396
1.405
1.444
1.489
0.342
0.412
0.421
0.292
0.303
0.400
0.248
0.258
0.282
71.138
32.546
27.653
13.105
82.601
40.891
36.227
32.890
108.880
98.820
76.662
68.298
27.076
26.781
22.579
16.844
30.535
29.756
28.452
20.600
41.098
37.621
35.521
24.088
0.356
337.147
151.377
118.682
54.833
413.005
196.591
158.197
141.159
588.541
494.100
350.055
302.204
18.533
17.926
14.787
10.674
21.117
20.521
19.211
13.624
29.440
26.777
24.599
16.177
1.041
BuPMBr2
OcPMBr2
MePMBr2
BuPMBr2
OcPMBr2
MeCN
DMSO
MePMBr2
BuPMBr2
OcPMBr2
0.342
0.242
0.357
0.353
0.348
0.514
0.408
0.370
0.830
0.575
1.223
1.165
0.870
2.073
1.581
1.312
53.01 [( 0.05] 64.68 [( 0.06] 57.67 [( 13.09] 0.9998 0.0041
52.46 [( 0.06] 64.52 [( 0.04] 57.92 [( 7.02]
52.45 [( 0.06] 64.24 [( 0.06] 56.48 [( 7.44]
0.9999 0.0023
0.9998 0.0038
experimental and calculated ET(probe)solvent and ET(probe)W,
respectively. The results of Table 4 are discussed in terms of
their dependence on the structures of the probe and the solvent
(at the same temperature, T) and on T, for the same probe and
binary mixture. Values of m are close to unity and generally
decrease as a function of increasing T. Likewise, for each probe
in each solvent, all values of æ, ET(probe)Solv and ET(probe)W
decrease as a function of increasing T. This probe desolvation
agrees with the known effect of temperature on solvent structure
due to less efficient hydrogen bonding and dipolar interactions.28
solvent (log P ) -0.77), more efficiently than PrOH, a weaker
acid (pKa ) 16.1) but more hydrophobic solvent (log P )
0.25).33
9. All æROH-W/ROH and æROH-W/W are greater 1, indicating
that the probes are preferentially solvated by ROH-W; all
æROH-W/W are greater than the corresponding æROH-W/ROH
,
indicating the ROH-W displaces W more efficiently than ROH.
Since the alcohols employed are more basic than water, we can
assume that the structure of the complex species is given by
Hw-O-H‚‚‚O(R)HROH, i.e., water is the hydrogen-bond donor to
alcohol, so that the two hydrogen atoms marked in italic are
the sites for hydrogen bonding with the probe phenolate oxygen.
As argued elsewhere, this hydrogen bonding partially deactivates
Hw toward further bonding; this deactivation is greater for more
basic alcohols.34,35 Therefore, the efficiency of ROH-W in
displacing alcohol and/or water from the solvation coordination
shells does not seem to be due to a H-bonding ability better
than those of the precursor solvents; it is due to hydrophobic
8. Values of æW/ROH are < unity, i.e., water is not efficient
in displacing the alcohol from the probe solvation coordination
shell (see the discussion on the significance of the magnitude
of æ after eq 13). Whereas water and alcohols may solvate the
probe by hydrogen bonding to its phenolate oxygen, ROH may
further solvate the probe by hydrophobic interactions. The
importance of the latter has been discussed above for pure
solvents (eq 6) and may be further corroborated by the fact that
the order observed is æW/MeOH > æW/PrOH, for every probe, in
the temperature range investigated. That is, water displaces
MeOH, a stronger acid (pKa ) 15.5) but less hydrophobic
(33) Barlin, G. B.; Perrin, D. D. Quart. ReV. 1966, 20, 75-&.
(34) Kingston, B.; Symons, M. C. R. J. Chem. Soc., Faraday Trans.
1973, 2, 69, 978-992.
(35) Symons, M. C. R. Pure Appl. Chem. 1986, 58, 1121-1132.
9076 J. Org. Chem., Vol. 71, No. 24, 2006

ThermosolVatochromism of Polarity Indicators
TABLE 5. âstatistical Coefficients of Equation 14, Applied to Solvatochromism of the Probes Synthesized in Aqueous Methanol and 1-Propanol,
at 25 °C
mixture
probe
â
statistical (s)
â
statistical (a)
â
statistical (b)
â
statistical (p)
r2
MeOH/W
MePMBr2
BuPMBr2
OcPMBr2
MePMBr2
BuPMBr2
OcPMBr2
-0.94 (( 0.02)
-1.25 (( 0.08)
-1.29 (( 0.06)
10.52 (( 1.11)
10.67 (( 1.22)
13.23 (( 1.48)
-0.04 (( 0.01)
-0.07 (( 0.02)
-0.08 (( 0.02)
1.13 (( 0.10)
1.24 (( 0.12)
1.43 (( 0.14)
-0.57 (( 0.01)
-0.66 (( 0.04)
-0.63 (( 0.03)
-0.54 (( 0.16)
-0.70 (( 0.17)
-0.84 (( 0.21)
-1.33 (( 0.03)
-1.53 (( 0.11)
-1.59 (( 0.09)
9.07 (( 1.03)
9.30 (( 1.14)
11.87 (( 1.38)
0.9999
0.9999
0.9999
0.9964
0.9978
0.9936
PrOH/W
interactions. This conclusion is corroborated by the above-
discussed order of æW/ROH, and by the fact that æPrOH-W/W
MeOH-W/W, for all probes, at all temperatures; see Table 4. The
to the probe phenolate oxygen and electrostatic interaction with
the probe positively charged nitrogen; this leads to the small æ
observed.
>
æ
order æROH-W/W > æROH-W/ROH is because æROH-W/W is related
to the difference between hydrogen bonding plus hydrophobic
interactions of ROH-W versus only hydrogen bonding by water
(see eq 10). On the other hand, hydrogen bonding and
hydrophobic interactions contribute to solvation by the two
solvent species involved in æROH-W/ROH (eq 9).
10. For mixtures of water with dipolar aprotic solvents
æW/MeCN > 1, whereas æW/DMSO < 1, i.e., water is more efficient
in displacing MeCN than DMSO from the solvation coordination
shell. MeCN can solvate positive centers better than negative
ones, i.e., it interacts less with the probe phenolate oxygen, being
displaced by water, because the latter is capable of solvating
both types of centers effectively8,36 The preceding conclusion
11. Table 4 shows that as a function of increasing temperature,
m, ET(probe)solv, ET(probe)W, æSolv-W/Solv, and æSolv-W/W de-
crease, whereas æW/Solv increases. The decrease in polarities of
pure solvents may be attributed to a decrease of solvent
stabilization of the probe ground state, as a result of the
concomitant decrease of solvent structure, and hydrogen-bonding
ability.43,44 Preferential “clustering” of water and solvents as a
function of increasing temperature means that the strength of
Solv-W interactions decrease in the same direction,27,28,45,46-48
with a concomitant decrease in its ability to displace both water
and solvent. This explains the decrease of æSolv-W/Solv and
æSolv-W/W as a function of increasing T. Plots (not shown) of
ET(probe)Solv versus T gave straight lines whose (negative) slopes
are given by ∆ET(probe)Solvent/degree (cal mol-1 K-1). For PrOH
and MeCN (for which four temperatures were investigated) the
order observed is ∆ET(probe)Solv > ∆ET(probe)W, reflecting the
greater effect of temperature on the structure of the solvent.
Consequently, hydrogen bonding of water with probe ground
state is less susceptible to temperature increase than that of the
organic component. This leads to a measurable “depletion” of
pure solvent in the probe solvation coordination shell, so that
æW/Solv increases as a function of increasing temperature;
12. The results of application of eq 6 to a single probe in a
series of pure solvents raises the question whether the same
approach can be applied to the solvatochromic behavior of a
probe in a series of binary solvent mixtures at a fixed
temperature. We used eq 14, where (more fundamental) pKa mixt
agrees with the following result: æMeCN-W/MeCN > æMeCN-W/W
,
i.e., the complex solvent displaces MeCN from the probe
solvation coordination shell more efficiently than it displaces
water, in agreement with the weak interaction of MeCN with
the probe phenolate oxygen.
Solvation by aqueous DMSO merits a comment: Whereas
æW/DMSO < 1, i.e., similar to solvation by alcohols, values of
æDMSO-W/W and æDMSO-W/DMSO are less than or close to unity.
To our knowledge, this is the first example of a Solv-W species
that is inefficient in displacing its precursor components from
the probe solvation coordination shell. Consider first the
exchange of the pure solvents. Values of æW/DMSO < 1 probably
because the organic solvent may solvate the probe by strong
dipole-dipole and hydrophobic interactions, akin to those
operative in aqueous DMSO.37 The small magnitudes of
æDMSO-W/W and æDMSO-W/DMSO may be attributed to the fact
that the interaction of DMSO with W attenuates the solvation
efficiency of the complex solvent. Evidence showing that
DMSO-W interactions are stronger than W-W interactions
was substituted for Rmixt
:
ET(probe)mixt ) constant + s π/mixt + a pKa mix + b âmixt
+
p log Pmixt (14)
1
include theoretical calculations38 IR; H and 13C NMR;39,40
neutron scattering41 and electron-spray mass spectroscopy37
Additionally, plot of Rmixt versus øDMSO shows negative devia-
tion from linearity; the corresponding plot for âmixt shows a
positive deviation,42 i.e., aqueous DMSO is less acidic than
expected. In other words, DMSO-W aggregate may be
considered as a deactivated species both in hydrogen bonding
This equation has been applied to aqueous MeOH and PrOH;
the regression coefficients are listed in Table SI-3 in Supporting
Information, whereas the âstatistical values are listed in Table 5.
Although the number of probes and solvents tested is small,
we are satisfied with the high correlation coefficients calculated.
An important result of Table 5 is the noticeable increase of
âstatistical of s and p as a function of increasing Nc (in the same
binary mixture) and, for the same probe, on going from MeOH
to PrOH. Therefore, our data indicate that hydrophobic interac-
(36) Gopalakrishnan, G.; Hogg, J. L. J. Org. Chem. 1984, 49, 3161-
3166.
(37) Shin, D. N.; Wijnen, J. W.; Engberts, J. B. F. N.; Wakisaka, A. J.
Phys. Chem. B 2001, 105, 6759-6762.
(38) Borin, I. A.; Skaf, M. S. J. Chem. Phys. 1999, 110, 6412-6420.
(39) Mizuno, K.; Imafuji, S.; Ochi, T.; Ohta, T.; Maeda, S. J. Phys. Chem.
B 2000, 104, 11001-11005.
(40) Shashkov, S. N.; Kiselev, M. A.; Tioutiounnikov, S. N.; Kiselev,
A. M.; Lesieur, P. Physica B 1999, 271, 184-191.
(41) Cabral, J. T.; Luzar, A.; Teixeira, J.; Bellissent-Funel, M. C. J. Chem.
Phys. 2000, 113, 8736-8745.
(42) Catalan, J.; Diaz, C.; Garcia-Blanco, F. J. Org. Chem. 2001, 66,
5846-5852.
(43) Zana, R.; Eliebari, M. J. J. Phys. Chem. 1993, 97, 11134-11136.
(44) Haak, J. R.; Engberts, J. B. F. N. Rec. TraV. Chim. 1986, 105, 307-
311.
(45) Harris, K. R.; Newitt, P. J. J. Chem. A 1999, 103, 6508-6513.
(46) Nishikawa, K.; Hayashi, H.; Iijima, T. J. Phys. Chem. 1989, 93,
6559-6565.
(47) Huelsekopf, M.; Ludwig, R. J. Mol. Liq. 2000, 85, 105-125.
(48) Shulgin, I.; Ruckenstein, E. J. Phys. Chem. B 1999, 103, 2496-
2503.
J. Org. Chem, Vol. 71, No. 24, 2006 9077

Martins et al.
tions are important to solvatochromism, both in pure solvents
and binary solvent mixtures.
Experimental Section
Synthesis of the Probes Employed. RPMBr2 were synthe-
sized according to the following scheme:63-66 The synthesis of
Conclusions
Solvation in pure solvents is due to interactions that depend
on the properties of the solute (structure, pKa, and hydrophobic-
ity) and the solvent, including proton donation/acceptance,
dipolarity/polarizability, and as shown here, lipophilicity. Evalu-
ation of the relative importance of these interactions requires
studying the solvatochromism of probes of adequate structure,
e.g., the series RPMBr2, where the pKa is kept constant while
the hydrophobic character is increased. Thermosolvatochromism
in binary solvent mixtures can be described by a general
mechanism, based on solvent exchange equilibria between the
species present in solution (W, Solv, and Solv-W complexes,
respectively) and their counterparts in the probe solvation
coordination shell. The nonideal dependence of ET(RPMBr2)
on øW is mainly due to preferential solvation of the probe,
especially by Solv-W; aqueous DMSO is an exception.
Temperature effect on æ is rationalized in terms of the structures
of water and solvent and their mutual interactions. Temperature
increase results in a gradual desolvation of eVery probe (i.e.,
decreased stabilization of its ground-state by W, Solv, and
Solv-W), in all binary mixtures; desolvation energies depend
on the hydrophobicity of the probe and the solvent and are
sensitive to the composition of the binary solvent mixture. The
Taft-Kamlet-Abboud equation has been modified by including
a solvent lipophilicity term; the modified equation applies
satisfactorily to pure and binary solvent mixtures. Solvation of
zwitterionic probes seems to be more sensitive to medium
lipophilicity than its basicity.
1-alkyl-4-methylpyridinium iodides from 4-methylpyridine or
n-alkyl iodide was carried out in MeCN, as recommended
elsewhere, followed by removal of the solvent and excess alkyl
iodide, eq 15.63 The light yellow products were either solid,
1,4-dimethylpyridinium iodide, or liquid, other 1-alkyl-4-
methylpyridinium iodides. Their purity was established by TLC
analysis by using ethanol/acetic acid/chloroform eluent (1:1:
18, by volume). The aldehyde 3,5-dibromo-4-hydroxybenzal-
dehye was prepared by the reaction of bromine with 4-hydroxy-
benzaldehye in glacial acetic acid, as given elsewhere.67
Condensation of this aldehyde with 1-alkyl-4-methylpyridinium
iodides in the presence of piperidine, eq 16, followed by
treatment with KOH, and recrystallization from aqueous metha-
nol gave RPMBr2, as red crystals. Table SI-4 in Supporting
Information shows the yields, melting points, elemental analyses,
and relevant IR frequencies for the probes synthesized. Attribu-
1
tions of the H and 13C NMR spectra are listed in Tables SI-5
and SI-6, respectively, in Supporting Information.
Values of æ and its dependence on the components of the
binary mixture and the properties of the probe, in particular its
lipophilicity, may be fruitfully employed to better explain
reactivity data, e.g., the (complex) dependence on medium
composition of rate constants and activation parameters of
different reactions, e.g., spontaneous decarboxylations,49-51
acid-, base-, pH-independent, and enzyme-catalyzed hydrolyzes
of carboxylic and carbonate esters and N-acylimizaoles52-60 and
the dependence of the kinetic order with respect to water on
the composition of the binary mixture8,61-62
Spectroscopic Determination of ET(probe) in Pure Sol-
vents and in Binary Solvent Mixtures. The probes employed
for studying solvatochromism and thermosolvatochromism were
MePMBr2, BuPMBr2, and OcPMBr2. Aliquots of the probe
solution in acetone were pipetted into small volumetric vials,
followed by evaporation of the acetone at room temperature,
under reduced pressure, in the presence of P4O10. The solvent
(or binary mixture) whose polarity is to be determined was
added, the probe was dissolved, and the UV-vis spectrum of
its solution was recorded. The following are relevant experi-
mental data: Temperature control inside the thermostatted cell-
holder, (0.05 °C; final probe concentrations, 2-5 × 10-4 mol/L
for MePMBr2, BuPMBr2, and HxPMBr2 and 3 × 10-5 mol/L
for OcPMBr2, respectively; cuvette path length, 1-4 cm;
number of spectra recorded, 2 at a rate of 120 nm/min; λmax
calculated from the first derivative of the absorption spectrum;
uncertainty in ET(RPMBr2) e 0.15 kcal/mol. The same proce-
dure was repeated for binary solvent mixtures, 16 per set,
prepared by weight at 25 °C. Thermosolvatochromism was
studied in mixtures of water with MeOH (10-40 °C), PrOH
and MeCN (10-60 °C), and DMSO (25-60 °C).
(49) Kemp, D. S.; Paul, K. G. J. Am. Chem. Soc. 1975, 97, 7305-7312.
(50) Grate, J. W.; Mcgill, R. A.; Hilvert, D. J. Am. Chem. Soc. 1993,
115, 8577-8584.
(51) Catalan, J.; Diaz, C.; Garcia-Blanco, F. J. Org. Chem. 2000, 65,
3409-3415.
(52) Huskey, W. P.; Warren, C. T.; Hogg, J. L. J. Org. Chem. 1981, 46,
59-63.
(53) Gopalakrishnan, G.; Hogg, J. L. J. Org. Chem. 1984, 49, 3161-
3166.
(54) Neuvonen, H. J. Chem. Soc., Perkin Trans. 1987, 2, 159-167.
(55) Mata-Segreda, J. F. J. Am. Chem. Soc. 1988, 110, 6224-6230.
(56) Galema, S. A.; Blandamer, M. J.; Engberts, J. B. F. N. J. Org. Chem.
1989, 54, 1227-1229.
(57) Mousa, M. A.; Diefallah, E. H. M.; Dessouki, H. A.; Atwa, S. T.
Thermochim. Acta 1989, 141, 1-8.
(58) El Seoud, O. A.; El Seoud, M. I.; Parah, J. P. S. J. Org. Chem.
1997, 62, 5928-5933.
(63) Davidson, S. J.; Jencks, W. P. J. Am. Chem. Soc. 1969, 91, 225-
234.
(64) Brooker, L. G. S.; Keyes, G. H.; Heseltine, D. W. J. Am. Chem.
Soc. 1951, 73, 5350-5356.
(65) Minch, M. J.; Shah, S. S. J. Chem. Educ. 1977, 54, 709.
(66) Minch, M. J.; Shah, S. S. J. Org. Chem. 1979, 44, 3252-3255.
(67) Martins, C. T.; Lima, M. S.; El Seoud, O. A. J. Phys. Org. Chem.
2005, 18, 1072-1085.
(59) Tian, J. H.; Tan, J.; Schowen, K. B.; Schowen, R. L. Can. J. Chem.
1999, 77, 781-786.
(60) Venkatasubban, K. S.; Bush, M.; Ross, E.; Schultz, M.; Garza, O.
J. Org. Chem. 1998, 63, 6115-6118.
(61) Sorensen, P. E. Acta Chem. Scand. A 1976, 30, 673-679.
(62) Henderson, J. W.; Haake, P. J. Org. Chem. 1977, 42, 3989-3994.
9078 J. Org. Chem., Vol. 71, No. 24, 2006

ThermosolVatochromism of Polarity Indicators
Spectrometric Determination of the Partition Coefficient
of the Probe between n-Octanol and Water. The aqueous phase
was a phosphate buffer solution (0.05 mol/L, pH ) 7.50). Equal
volumes of this buffer and n-octanol were agitated for 1 h (tube
rotator), and the phases were separated. A probe solution, 5 ×
10-4 mol/L in (buffer-saturated) n-octanol was prepared, and
its absorbance, Ainitial, was recorded. An aliquot of this solution,
25 °C, the apparent pKa values of all probes were found to be
5.15 ( 0.05.
Acknowledgment. We thank FAPESP (State of Sa˜o Paulo
Research Foundation) for financial support and a predoctoral
fellowship to C.T.M. and the CNPq (National Council for
Scientific and Technological Research) for a PIBIC undergradu-
ate research fellowship to M.S.L. and a research productivity
fellowship to O.A.E.S. We thank Dr. Erick L. Bastos for
calculating Kdissoc for MeCN-W and DMSO-W and Dr. Paulo
A. R. Pires and C. Guizzo for their help.
Voctanol, was agitated with (n-octanol-saturated) phosphate buffer,
V
buffer, at room temperature, for 2 h. After phase separation at
25 °C, the absorbance, Aequilibrium, of the n-octanol phase was
measured, and the partition coefficient calculated from log P
) log(Aequilibrium × Vbuffer/(Ainitial - Aequilibrium)Voctanol). Values
of log P were found to be -0.16 ( 0.01, 1.12 ( 0.01, and 1.86
( 0.1 for MePMBr2, BuPMBr2, and HxPMBr2, respectively.
Spectrometric Determination of the Apparent pKa Values
of the Probes. The pKa was calculated from the Henderson-
Hasselbalch equation.68 A methanolic solution of each probe
was added to potassium hydrogen phthalate buffer (0.05 mol/
L) so that the final volume fraction of methanol was e0.05
and the final [probe] was ) 5 × 10-4 mol/L. The concentrations
of the zwitterionic forms were measured at 438, 443, and 440
nm for MePMBr2, BuPMBr2, and OcPMBr2, respectively. At
Supporting Information Available: Plots of ET(OcPMBr2)
versus øw, for different binary mixtures; solvent polarity/ temper-
ature/solvent composition contours for MePMBr2, BuPMBr2, and
OcPMBr2 in binary solvent mixtures; results of the application
of eq 6; polynomial dependence of ET(RPMBr2) on the
analytical mole fraction of water in the binary mixture; regression
cofficients of eq 14; yields, elemental analyses, and IR
1
frequencies of the probes synthesized; H and 13C NMR data for
the probes synthesized; calculations of the Dependence of ET(probe)
on the properties of the solvent; calculation of the dissociation
constant of Solv-W, Kdissoc, x Effective, and solvent fractionation
Species
factors. This material is available free of charge via the Internet at
http://pubs.acs.org.
(68) Anslyn, E. V.; Dougherty, D. A. In Modern Physical Organic
Chemistry; University Science Books: Sausalito, 2004; p 259.
JO061533E
J. Org. Chem, Vol. 71, No. 24, 2006 9079