Solvent influence on the photophysical properties of 4-methoxy-N-methyl-1,8-naphthalimide

Article abstract of DOI:10.1016/j.saa.2005.05.008

4-Methoxy-N-methyl-1,8-naphthalimide (1) exhibits considerable solvatochromism and its UV-vis spectral properties have been studied in several polar/non-polar and protic/aprotic solvents, as well as in ethanol-water mixtures. The results reveal a strong influence of the solvent's polarity and its hydrogen-bond donor (HBD) capability on the photophysical properties of 1. For binary ethnaol/water mixtures, preferential solvation models describe the band shifts in the probe's visible absorption spectrum well, but they fail to describe the correponding shifts of the emission maxima. Pseudolinear approximations between solvent composition and molecule's transition energies, E_T, can be used to study the composition of ethanol-water mixtures, simplifying the mathematical treatment for eventual analytical applications.

Full text of DOI:10.1016/j.saa.2005.05.008

Spectrochimica Acta Part A 63 (2006) 220–226
Solvent influence on the photophysical properties of
4-methoxy-N-methyl-1,8-naphthalimide
a,∗
b
c
Gr e´ goire J.-F. Demets , Eduardo R. Triboni , Emiliano B. Alvarez ,
c
b
c
Guilherme M. Arantes , Pedro Berci Filho , M a´ rio J. Politi
a
Departamento de Qu ´ı mica, Faculdade de Filosofia Ci eˆ ncias e Letras de Ribeir a˜ o Preto, Universidade de S a˜ o Paulo,
Av. Dos Bandeirantes 3900, CEP 14040-901, Ribeir a˜ o Preto, S.P., Brazil
b
Instituto de Qu ´ı mica de S a˜ o Carlos, Universidade de S a˜ o Paulo, CP 780, CEP 13560-970, S a˜ o Carlos, S.P., Brazil
c
Instituto de Qu ´ı mica, Universidade de S a˜ o Paulo, CP 26077, CEP 05599-970, S a˜ o Paulo, S.P., Brazil
Received 7 April 2005; received in revised form 9 May 2005; accepted 10 May 2005
Abstract
-Methoxy-N-methyl-1,8-naphthalimide (1) exhibits considerable solvatochromism and its UV–vis spectral properties have been studied in
4
several polar/non-polar and protic/aprotic solvents, as well as in ethanol–water mixtures. The results reveal a strong influence of the solvent’s
polarity and its hydrogen-bond donor (HBD) capability on the photophysical properties of 1. For binary ethnaol/water mixtures, preferential
solvation models describe the band shifts in the probe’s visible absorption spectrum well, but they fail to describe the correponding shifts of
the emission maxima. Pseudolinear approximations between solvent composition and molecule’s transition energies, E
the composition of ethanol–water mixtures, simplifying the mathematical treatment for eventual analytical applications.
2005 Elsevier B.V. All rights reserved.
T
, can be used to study
©
Keywords: Naphthalimide; Solvatochromism and spectrophotometric probe; Ethanol–water mixtures
1
. Introduction
transitions. These variations can differentially stabilize
the ground and/or the excited state in polar and non-polar
solvents. Hydrogen-bond formation also affects the molec-
ular orbitals energy levels considerably, influencing the
molecular energetics in a similar way. By convention, the
solvatochromic effect is said to be negative if blue band shifts
are observed (hypsochromic shifts) with increasing solvent
polarity. Conversely, positive solvatochromism occurs when
red band shifts (bathochromic shifts) are observed in these
circumstances. Generally, if one considers Franck–Condon
excited states, molecules exhibiting a larger permanent
dipole moment in the excited state (∝e > ∝g) display positive
solvatochromism. The opposite is observed in cases when the
permanent dipole moment is larger in the ground state [1].
Several attempts have been made in order to quan-
tify and predict solvatochromic effects and most of the
solvatochromic solvent polarity scales have been built
using empirical solvent parameters. Based on comparisons
between different solvents, Kamlet and Taft have built a scale
of solvatochromic parameters to be used in linear solvation
Several properties of chemical systems depend on the
nature of their solvent. The solvent not only affects reaction
kinetics, but also influences the characteristics of the spectra
of the solute molecules. In some cases more than in others,
this phenomenon is manifested by solvent-induced differ-
ences in the electronic, vibrational and NMR spectra. This
property is known, when talking about spectral effects, as
solvatochromism or, in a more general way, as perichromism
(peri: around) [1]. Many factors influence the spectral
behavior of dissolved molecules, especially the solvent’s
polarity, its temperature [2] and its hydrogen-bond donor
(HBD) or acceptor (HBA) capacity [3–11].
InthecaseofUV–visspectroscopy, strongsolvatochromic
effects are generally observed for dipolar molecules exhibit-
ing large variations in their dipole moment during electronic
∗
Corresponding author. Tel.: +55 16 602 4459; fax: +55 16 633 8151.
E-mail address: demets@ffclrp.usp.br (G.J.-F. Demets).
1
386-1425/$ – see front matter © 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.saa.2005.05.008

G.J.-F. Demets et al. / Spectrochimica Acta Part A 63 (2006) 220–226
221
energy relationships (LSERs). This scale comprises several
solvent parameters, such as their HBD acidity and HBA
basicity (α and β parameters), which estimate the solvent’s
ability to donate an H-atom or an electron pair in a solvent-
methods. This kind of probe could also be used in quality
control in colorless liquor production, such as cacha c¸ a (a
sugar-cane liquor and export product), vodka, rum and gin.
This paper presents some of the solvatochromic properties
of 4-methoxy-N-methyl-1,8-naphthalimide (1) in several sol-
vents and in ethanol–water binary mixtures, and shows how
it could be used as a UV–vis spectrophotometric probe for
analytical purposes.
*
to-solute hydrogen bond. Another parameter, the ␲ index,
measures the solvent’s dipolarity and polarizability and its
ability to stabilize charges by dielectric effects [12–14].
Catal a´ n et al. have also developed a “single parameter
solvent dipolarity/polarizability (SPP) scale”, which uses
the gas phase (SPP = 0.000) and DMSO (SPP = 1.000) as
reference media to fix the scale [15,16].
2. Experimental
It is well known that 1,8-naphthalimide derivatives are
very sensitive to their surroundings and their electronic
emission and absorption spectra are frequently affected by
the nature of the solvent. What is generally observed in
N-alkyl-1,8-naphthalimide photophysics is a decrease in the
fluorescence quantum yield as the solvent’s polarity and its
HBD capability increase. This happens mainly because of
the proximity between the n, ␲ and ␲,␲ excited states,
which controls the excited-state conversion from singlet to
triplet manifold, thus explaining the response in fluorescence
yields.
In addition, the presence of electron-donor/acceptor
groups attached to the naphthalene moiety is also very
important because these substituents can activate or deacti-
vateemittingmodes, affectingtheiremissionspectraconside-
rably.
The presence of electron-donating substituents can also
increase the separation between the singlet (S1) and its
closer triplet excited state, thus decreasing the intersystem
crossing (ISC) and favoring the fluorescence deactivation
channel. Furthermore, N-naphthalimides modified with
electron-donating groups can also allow photo-induced
electron-transfer processes, generating photochemical diads
and they can be part of even larger systems exploring the
electron-transfer phenomena, such as triads and tetrads
2
.1. Synthesis of 4-methoxy-N-methyl-1,8-
naphthalimide (1)
4-Methoxy-N-methyl-1,8-naphthalimide (1) (Fig. 1) was
obtained from 4-nitro-N-methyl-1,8-naphthalimide, which
was synthesized from 4-nitro-naphthalic anhydride and
methylamine by a sonochemistry route [23].
*
*
-Nitro-N-methyl-1,8-naphthalimide (1.0 × 10− mol)
3
4
was dissolved with five equivalents of potassium carbonate
in 35 mL of methanol and the solution was sonicated in an
ultrasonic cleaner bath (Bransonic, 150 W/25 kHz output) in
a 40 mm diameter cylindrical reaction vessel. The reaction
was monitored by TLC [Merck silica gel plates with fluores-
cent indicator (254 nm), using CH2Cl2 as eluent]. At the end
of the reaction, the solvent was removed in vacuo, the residue
was dissolved in CHCl3 and the solution was washed several
times with 10% aqueous sodium hydroxide and water and
dried in vacuum to give pale yellow needles (yield = 97%)
of mp = 197–198 C (in Ref. [24] mp = 197–201 C). The
melting point was measured in an electrothermal melting
point apparatus. H NMR spectra were recorded on a Bruker
AC-200 (in CDCl3). The IR spectrum was registered on a
BOMEM-FTIR MB 102 spectrophotometer, in the range
3
◦
◦
1
−
1
00–4000 cm , in KBr pellets. Capillary GC analysis was
performed on a HP-5890 coupled to an MSD-5970 mass
selective detector.
[17–22].
Solvatochromic compounds can be used as analytical
1
R = 0.21 (CH2Cl2); H NMR (CDCl3): δ = 3.5 (s, 3H,
f
UV–vis spectrophotometric probes [3–11] and, eventually,
for water quantification in binary solvent mixtures. We have
chosen to study ethanol–water mixtures, since hydrated
ethanol has been used as a natural car fuel in Brazil for more
than 20 years, being an alternative to gasoline. Dishonest
distributors adulterate such fuel by adding water to it in
order to increase its volume, thus largely exceeding the
legally allowed water content in automotive ethanol (max.
water content = 6.8 ± 0.4 mass%). This excess water causes
engine misfunctioning and accelerates the wearing off of
ethanol engines. Besides adding extra water, some distrib-
utors also correct the mixture density by means of other
impurities, making the densimeters useless for the control
of water content in the fuel. In this case, solvatochromic
spectrophotometric probes could become an alternative for
fast certification analysis, since they are easier to use than the
traditional Karl–Fischer titration method for precise water-
content determinations and much cheaper than specific NIR
CH3), 4.0 (s, 3H, CH3), 7.0 (d, J = 8.2 Hz, 1H, Ar),
7.6 (t, J = 8.2 Hz, 1H, Ar), 8.5 ppm (m, 3H, Ar); MS:
(
m/z) = 241(M+, 100), 213, 198, 182, 113; IR (KBr):
’
−1
v˙ ˙ /cm = 3021, 2945, 1698, 1658, 1580, 1356, 1254, 1077;
Anal. Calcd. for C14H11NO3: N, 5.81; C, 69.70; H, 4.60.
Found: N, 5.86; C, 69.68; H, 4.63%.
Fig. 1. 4-Methoxy-N-methyl-1,8-naphthalimide.

2
22
G.J.-F. Demets et al. / Spectrochimica Acta Part A 63 (2006) 220–226
2
.2. Measurements
UV–vis spectra were recorded on a Shimadzu UV-
◦
2
401PC spectrophotometer at 25 C. Emission data were
obtained using a Photon Technology International LS1
spectrofluorimeter. All emission spectra were corrected as a
function of the lamp intensity, using the equipment software.
The absorption and emission maxima were calculated using
the spectra first derivatives, at the point were they cross the
x-axis. All solvents are of analytical grade.
3
. Results and discussion
In order to study the solvatochromic properties of 1, a
mother solution was prepared from 2 mg of 1 in 50 mL
−
5
−1
Fig. 3. UV–vis emission spectra of 3.3 × 10 mol L 4-methoxy-1,8-N-
−
4
−1
ethanol (c = 1.65 × 10 mol L ), where it is fairly solu-
ble. All the samples used in this work were prepared from
starting portions of 1 mL of this solution. This amount was
placed in a flask, ethanol was evaporated and 5 mL of sol-
vent was added in order to obtain solutions of imide 1 with
max
methyl-naphthalimide solutions in three solvents (λexc = λabs
solvent: 353 nm for isooctane, 475 nm for water and 363 nm for ethanol).
for each
In the visible region of the spectrum, the first solvent type
generates well-resolved absorption spectra, showing vibronic
bands separated by ꢀλ ≈ 15 nm, which correspond to the
−
5
−1
c = 3.3 × 10 mol L . In this study, 14 different solvents
were used in order to explore the solvatochromic effect in
different media. In this concentration range, 1 was soluble
in all solvents. Polar and non-polar solvents were employed,
but all of them were non-aromatic, in order to eliminate ␲-
type solvent–solute interactions. The chosen solvents present
different hydrogen-bonding abilities and this was also taken
into account in this work. By observing Figs. 2 and 3, one
can easily notice that the solvent exerts a great influence on
both the electronic absorption and emission spectra, changing
their shape, spectral maxima positions and intensities. All the
spectra can be assembled in basically three sets of spectral
patterns. The first one occurs with all non-polar solvents, like
isooctane; another is present in the case of all dipolar non-
HDB solvents, like DMSO and a third pattern can be detected
in hydrogen-bond donor solvents, like alcohols and water.
*
1
*
␲
–␲ S level transitions. The n–␲ band is also visible, but
weaker, around λmax = 294 nm. The second spectral type con-
sists of spectra with a single band above λmax = 350 nm, with
comparable intensities, but without fine structure. In HBD
solvents (water, methanol, ethanol, 1-propanol, 2-butanol and
2
-methyl-1-propanol, some not shown), a much less intense
broadened band, without fine structure appears. These bands
intensities are very close to each other in all solvents, except
for water.
The emission spectra show basically the same behavior.
Well-resolved bands are obtained in non-polar solvents,
0
1
which are assigned to S ← S transitions. In polar solvents,
the bands are broadened and collapsed, when compared
with other solvents. In the case of HBD solvents, like water
and alcohols, this effect is more pronounced than in the case
of all other solvents. The solvent–solute hydrogen-bond
formation is very clear in both absorption and emission
spectra. Positive solvatochromic shifts are observed as
’
−1
between absorption
a difference of ꢀ v˙ ˙ = 1660 cm
maxima (ꢀλmax = 22 nm for water and isooctane) and of
’
−1
ꢀ
˙v˙ = 3230 cm between emission maxima (ꢀλmax = 60 nm
for water and isooctane), as one goes from a polar to a
non-polar solvent, showing that the transition energy
decreases with increasing solvent polarity. The spectral data
are reported in Table 1.
Good linear correlations between Catal a´ n’s SPP parame-
ter and absorption and emission maxima, as well as Stoke’s
shifts are obtained, as can be observed in Fig. 4. It is clear
that in non-HBD solvents, the spectral data fit very well, cor-
relation with SPP: r = 0.996 for Stokes shifts; r = 0.963 for
absorption and r = 0.994 for emission energies, indicating
that the observed solvatochromic effect is mainly due to dipo-
larity/polarizabilty interactions taking place when hydrogen
bonds are not involved. This parameter is quite useful to
Fig. 2. UV–vis absorption spectra of 3.3 × 10− mol L 4-methoxy-N-
5
−1
methyl-1,8-naphthalimide (1) solutions in three solvents of different polarity.

G.J.-F. Demets et al. / Spectrochimica Acta Part A 63 (2006) 220–226
223
Table 1
has already been considered by Catal a´ n et al., who had to
create an additional acidity scale to explain solvatochromism
in these solvents [16]. HBD solvents can easily form hydro-
gen bonds, binding to the molecule’s carbonylic oxygens of
the molecule and, possibly, to the methoxy group. The influ-
ence of this kind of interaction via carbonylic oxygens is very
strong in the spectroscopic properties of the naphthalimide.
This has been proven by Pardo et al., who have studied
the influence of intramolecular hydrogen bonds formed by
substitutes of the imidic nitrogen [25–27], as well as the pro-
tonation of the imidic substituents. All these elements seem
to indicate that the origins of this different solvatochromic
effect in HBD solvents lie in the specific solvent–solute
interactions occurring mainly with the carbonylic oxygens.
Several attempts have been made to correlate these param-
eters with absorption and emission intensities. Once the
experiments in all solvents have been carried out in the same
conditions, we have tried to correlate the emission/absorption
integrals ratio with the scale mentioned above, as if it was a
relative quantum yield, but there is no reliable experimental
relationship between values in this specific case (details in
the supporting information). In addition, lifetimes and flu-
orescence quantum yields should increase with increasing
SPP, as predicted in theory and verified by Pardo et al. [26],
but this fact could not be experimentally verified in our case.
Theoretical calculations (see supporting information)
were carried out in order to explain the stabilization of the
excited state in polar solvents. The results in vacuum, how-
ever, do not reveal appreciable dipole moment variations
during HOMO–LUMO transitions.
UV–vis spectroscopic data for solutions of naphthalimide 1 in 11 solvents
abs
−3
−1
em
Solvent
λ
(nm) ±
ε (mol dm cm
)
λ
(nm) ±
’
max
max
’
−1
−1
)
0
.5 (˙v˙, cm
)
0.5 (˙v˙, cm
H2O
375 (26666)
360 (27777)
364 (27472)
363 (27548)
355 (28169)
355 (28169)
359 (27855)
365 (27397)
363 (27548)
353 (28328)
357 (28011)
2240
9120
8450
8120
11330
8700
8420
7640
9550
8760
9600
462 (21645)
434 (23041)
450 (22222)
444 (22522)
404 (24752)
402 (24875)
430 (23255)
452 (22123)
444 (22522)
402 (24875)
412 (24271)
MeCN
MeOH
EtOH
Cyclohexane
n-Dodecane
THF
DMSO
DMF
Isooctane
CCl4
c (1) = 3.3 × 10− mol L ].
5
−1
[
describe and predict the behavior of 1 in all non-HBD sol-
vents. An increase in the SPP parameter causes a decrease in
the transitions energies, as would be expected for a positive
solvatochromism.
This kind of behavior has been observed by Barros et
al. [22] with other naphthalimides and it can be assigned
to the proximity between the n,␲ and ␲,␲ transition ener-
gies. Increasing medium polarity diminishes the ␲,␲ energy,
increasing its contribution to the mixed n,␲ ; ␲,␲ state, lead-
ing to an enhanced ␲,␲ character in the LUMO energy.
On the other hand, the HBD solvents do not behave in
this way, generating random values when compared to other
solvents. The maxima for water, ethanol and methanol have
also been plotted in Fig. 4 (the cross-marked shapes) and
no correlation with these parameters is apparently possible,
once the variations in absorption and emission data when one
alcohol is substituted for another are not at all predictable.
Data for 1-propanol,2-methyl-1-propanol (isobutanol) and
*
*
*
*
*
*
In order to study the behavior of 1 in ethanol–water
mixtures, the same amount of this imide was dissolved in
11 numbered ethanol–water mixtures with growing ethanol
content. This choice was made for commercial reasons, as
explained in Section 1 of this paper. The absorption and
emission maxima in these different binary solvent mixtures
were measured and are reported in Table 2.
2
-butanol are very similar and are not shown. As can be seen,
alcohols and water must be treated as a special case and this
Fig. 5 shows the absorption maxima energy of 1 as a
function of the composition of the binary mixtures (in %,
v/v) and the molar fraction of water. Volumetric variations
due to mixture contraction have not been taken into account,
once they would affect the transitions intensities rather
than their maxima. At a first glance, one can imagine that
a linear correlation exists between absorption energy and
the mole fraction of ethanol (χ(ethanol) or χ2). In order to
simplify the equations, water will be called solvent 1 and
ethanol will be solvent 2. χ1 and χ2 correspond to their
respective mole fractions. A correlation factor of r = 0.9935
*
can be obtained from this data (ET = Ewater + 0.098(± 0.003)
χ2). However, the emission maxima (Fig. 6) clearly show
that this is not necessarily the case. The emission data
describe an exponential function instead of a straight line.
In fact, experimental data fit quite well into the function
with t = 0.22 (r = 0.99). The
pseudolinear approach, using a single line, is only valid for
absorption maxima, where differences in energy are smaller.
Fig. 4. Absorption (ꢀ) and emission (᭹) maxima and Stokes shifts (ꢁ) of
−
5
−1
3
.3 × 10 mol L 4-methoxy-1,8-N-methyl-naphthalimide (1) solutions
emission
emis
(χ2/(−t))
E
= E
+ e
in different solvents: a, isooctane; b, cyclohexane; c, n-dodecane; d, tetra-
chloromethane; e, tetrahydrofuran; f, ethanol; g, methanol; h, acetonitrile; I,
N,N-dimethylformamide; j, water; k, dimethyl sulfoxide.
water

2
24
G.J.-F. Demets et al. / Spectrochimica Acta Part A 63 (2006) 220–226
Table 2
UV–vis spectroscopic data for naphthalimide (1) solutions in eleven ethanol–water mixtures
max
abs
−3
max
abs
−3
Mixture (% ethanol, v/v)
Water mole fraction χ water
E
(eV) (± 6.8.10
)
E
(eV) (± 6.8.10
)
’
−1
’
−1
(
λmax, nm; ˙v˙, cm
)
(λmax, nm; ˙v˙, cm
)
0
1
2
3
4
5
6
7
8
9
1
(0%)
1
3.310 (375; 26673)
3.312 (375; 26688)
3.317 (374; 26709)
3.319 (374; 26745)
3.330 (373; 26838)
3.339 (372; 26910)
3.350 (370.5; 26990)
3.357 (370; 27048)
3.366 (369; 27122)
3.379 (367; 27233)
3.410 (364; 27480)
2.696 (460; 21720)
2.707 (458; 21815)
2.729 (455; 21992)
(10%)
(20%)
(30%)
(40%)
(50%)
(60%)
(70%)
(80%)
(90%)
0 (100%)
0.966
0.927
0.882
0.827
0.762
0.681
0.578
0.445
0.262
0
2.749 (451.5; 22148)
2.579 (450; 22232)
2.790 (445; 22482)
2.797 (444; 22537)
2.805 (442.5; 22598)
2.812 (441; 22655)
2.821 (440; 22727)
2.831 (438; 22810)
c (1) = 3.3 × 10− mol L ].
5
−1
[
◦
In order to use a linear approximation with emission data, it is
necessaryto use two lines insteadofone. The first linear equa-
cacha c¸ a contains 46 vol.%, which corresponds to 46 GL
(degrees Gay–Lussac). The nominal value for both is 96%
emission
emis.
∗
◦
tion (E
= E
+ 0.393(± 0.023) χ2; r = 0.992)
and 50 GL, respectively. However, this is a simplification
water
would describe the emission energies for mixtures
and these plots should formally be considered as polyno-
mial functions, which is common for other solvent mixtures
[28,29].
with low ethanol content (χ2 < 0.24) and the second
emission
emis
∗
(E
= E
+ 0.052(± 0.003) χ2; r = 0.989) for
χ2=0.24
mixtures with higher ethanol content (χ2 > 0.24).
Several models have been created to explain solva-
tochromism in solvent mixtures and one of the most
employed nowadays is the preferential solvation model
[30–35]. This model assumes that, in a mixture of solvents,
the one with greater affinity for the solute will preferentially
compose the solvation microsphere. This kind of approach
has been well described in several papers [34,35] and the
mathematical treatment of the experimental data becomes
gradually more complicated as more realistic models of the
solvation phenomenon are considered. Simple correlations
between the composition of solvent mixtures and solute
absorption maxima are only obtainable when the two sol-
vents do not interact between themselves strongly, i.e. when
As there are small shifts in the absorption energies,
this pseudolinearity can be used for analytical purposes in
composition-maxima correlations when absorption spectra
are concerned. The stars plotted in Fig. 5 are real samples:
one consists of automotive ethanol, the other is cacha c¸ a. It is
obvious that two samples are not enough to establish any ana-
lytical chemistry statement, but they are enough to show that
a correlation is possible. According to their absorption max-
imum (obtained from the 0 value of the first derived), the gas
station automotive alcohol contains 96 vol.% ethanol, while
Fig. 5. UV–vis absorption maxima of 4-methoxy-1,8-N-methyl-
naphthalimide solutions in ethanol–water mixtures, as a function of their
−
5
−3
composition [c (1) = 3.3 × 10 mol dm ]. The two stars represent real
Fig. 6. Emission maxima of 4-methoxy-1,8-N-methyl-naphthalimide solu-
−
5
−3
samples, the first one is automotive ethanol and the second one is cacha c¸ a
at 96 and 48%, respectively).The dashed line corresponds to the adjusted
curve. Inset: plot of solvent composition in the bulk and in the solvation
sphere (solvent 2 is ethanol, see text below).
tions in ethanol–water mixtures (λexc = 370 nm, [3.3 × 10 mol dm ]) as
a function of the ethanol mole fraction of the mixture and the exponential
function adjusted to experimental data. In gray: linear regressions for data
in two intervals: χ2 < 0.24 and χ2 > 0.24.
(

G.J.-F. Demets et al. / Spectrochimica Acta Part A 63 (2006) 220–226
225
nosynergeticeffectisobserved. Inourcase, ethanolandwater
do interact very strongly [36,37] (type 2 solvents, accord-
ing to the classification of Ros e´ s et al.), but they do not
seemtoformtheso-called“highlypolarhydrogen-bondcom-
plexes” [34] responsible for synergetic effects in the spectra
of binary mixtures. However, these solvents behavior cannot
be explained with simpler models. To describe this system
using the models presented by Ros e´ s et al. or by Skwier-
czynski and Connors, we have to use the m model which is
the most complex. However, it seems to be the only model
that provides good correlation with experimental data.
The m model assumes that equilibria occur between the
solvents that compose the solvation microsphere of a sol-
vatochromic indicator in binary solvent mixtures and these
solvents can be exchanged. The solvatochromic indicator can
be solvated by m molecules of water, ethanol or both. The
equilibria between the solvated forms Eqs. (1) and (2) depend
on the affinity of the solvents for the solute, on the interac-
tions between them and on the composition of the mixtures.
These equilibria are shown in Eqs. (1) and (2).
In this general model, m is the number of solvent molecules
in the solvation microsphere of the solvatochromic indicator
affecting its transition energies.
Adjustment of the experimental data by non-linear regres-
sion with the aid of the Levenberg-Marquardt algorithm
N
(dashed line, Fig. 5) leads to the E
, f2/1 and f12/1
Twater–ethanol
−2
parameters, respectively, corresponding to 0.59 (± 5.10 ),
0.43 (± 0.24) and 5.05 (± 0.77) for m = 3 solvent molecules in
the solvation microsphere. According to these numbers and to
the m preferential solvation model, the solvation microsphere
should be mainly composed of water molecules throughout
almost all the composition range of the mixture. In other
words, the imide should be poorly solvated by ethanol. We
should always remember that this model is oversimplified
[38] and the m number may not be enough to describe the
whole solvation microsphere.
However, emission data cannot be fitted to equations sim-
ilar to Eq. (5), when values of m = 3 and f2/1 = 0.43 are used
and when it is admitted that the solvation sphere does not
change during the electronic transition, as should be the case
of a Franck–Condon transition. These parameters should be
maintained, unless the lifetime of the excited state of 1 is
I(water) + methanol ꢂ I(ethanol) + mwater
(1)
m
m
−
10
long enough (>10
s) to allow configuration relaxation of
m
m
I(water) + ethanol ꢂ I(water–ethanol) + 2 water
the solvation microsphere which seems to be the case.
m
m
2
(2)
The constants for these processes are defined by:
4. Conclusion
sphere sphere
χ
/χ
1
2
The UV–vis absorption and emission spectra intensity,
shape and energy of 1 are strongly solvent-dependent, as it
is the case of some of its analogs [16–18]. The absorption
maxima exhibit linear behavior, in agreement with Catal a´ n’s
f2/1 =
(3)
(4)
m
)
bulk bulk
(
χ
/χ
2
1
sphere sphere
χ
/χ
12
1
f12/1 = ꢀ
*
m
)
SPP parameter, as well as with Kamlet–Taft’s ␲ parameter.
bulk bulk
(χ
/χ
2
1
The exception are HBD solvents such as water and alcohols,
which cause different effects on the spectra. Solvents
forming hydrogen bonds generate significant deviations
from these correlations and they should be treated separately.
In these cases, the full Kamlet–Taft model should be used,
where χ’s represent the molar fractions of solvents 1, 2 and
co-solvent 12, in both the bulk and solvation microsphere.
f constants (Eqs. (3) and (4)) measure if the solute is best
solvated by solvent 2 instead of 1 (f2/1) or by a mixture of
them (f12/1). The normalized transition energies vary as a
function of these parameters and of the molar fraction of
*
including ␲ , ␣ and ␤ parameters.
The positive solvatochromic effect suggests that stabiliza-
tion of the excited states occurs in polar solvents. However,
the semi-empirical calculations in vacuum were not able to
give evidence of great variations in dipole moments during
the electronic transition, which could justify this observation.
The more efficient solvatochromic probes are generally
complex, charged, conjugated molecules that are frequently
unstable or insoluble in some kind of solvent. In spite of
displaying relatively small solvatochromic shifts, imide 1, is
a very simple, small and stable molecule that can be synthe-
sized in large amounts by the sonochemical route. Also, it is
soluble in most solvents, at least sparingly. This imide can be
employed as a cheap solvatochromic probe in ethanol–water
mixtures, using the absorption or emission maxima pseudo-
linear approximation. This is especially useful for automotive
ethanol and liquor industry quality control. In mixed solvents,
the chromophore can be viewed as the center of a cavity,
bulk
solvent 2 (χ2 ), according to expression (5).
ꢀ
m
m
a(χ b2 ulk) + c [(1 − χ2 )χ
bulk bulk
]
2
N
E = 1 −
(5)
T
m
m
)
bulk
bulk
2
(
1 − χ2 ) + f2/1(χ
ꢀ
m
bulk bulk
+
f12/1 [(1 − χ2 )χ2
]
Coefficients a and c depend on f1 and f12/1, as shown
/2
in expressions (6)–(8). Term c provides the value of
N
E
, which is the transition energy of the tran-
Twater–ethanol
sition in a solvation microsphere containing both solvents.
N
N
a = f2/1(E
− E
)
(6)
(7)
(8)
T ethanol
T water
N
N
(E
− E
N
) = 1
T ethanol
T water
N
c = f12/1(E
− E
)
T water ethanol
T water

2
26
G.J.-F. Demets et al. / Spectrochimica Acta Part A 63 (2006) 220–226
surrounded by hydrogen-bond forming molecules. When
two solvents coexist, competitive solvation occurs at dif-
ferent degrees, according to their proportions and these two
solvents take part in the composition of the solvation cavity.
In the case of ethanol and water, the solvents interact with
each other, forming solvent–solvent hydrogen bonds, which
change the bulk structure of the pure solvents. This reflects
in the composition of such cavity. These changes should also
modify the availability of water binding hydrogens, gradually
changing the spectrum of 1. The preferential solvation mod-
els have shown that 1 is best solvated by water and changes
in the solvation sphere occur during decay of its excited state.
[9] H. Langhals, Angew. Chem. 94 (1982) 738;
Angew. Chem. Int. Ed. Engl. 21 (1982) 432–433.
[
[
10] M.J. Minch, S.J. Sadiq Shah, J. Chem. Educ. 54 (1977) 709.
11] E.A. Gastilovich, N.V. Korol’kova, G.A. Val’kova, G.G. Konoplev,
B.V. Ni, Russ. J. Phys. Chem. 66 (1992) 1121–1124.
12] M.J. Kamlet, J.-L.M. Abboud, M. Abraham, R.W. Taft, J. Org.
Chem. 48 (1983) 2877–2887.
[
[
13] M.J. Kamlet, J.-L.M. Abboud, R.W. Taft, J. Am. Chem. Soc. 99
(1977) 6027–6038.
[
[
14] D. Burget, P. Jacques, J. Chim. Phys. 88 (1991) 675–688.
15] J. Catal a´ n, V. L o´ pez, P. P e´ rez, R. Mart ´ı n-Villamil, J.-G. Rodr ´ı guez,
Liebigs Ann. (1995) 241–252.
[16] J. Catal a´ n, V. L o´ pez, P. P e´ rez, Liebigs Ann. (1995) 793–795.
[
[
17] C.M. Lee, W.D. Kumler, J. Org. Chem. 27 (1962) 2055–2059.
18] P. Berci Filho, V.G. Toscano, M.J. Politi, J. Photochem. Photobiol.
A 43 (1988) 51–58.
[
19] F.C.L. Almeida, V.G. Toscano, O. Dos Santos, M.J. Politi, M.G.
Acknowledgements
Neumann, P. Berci Filho, J. Photochem. Photobiol. A 58 (1991)
289–294.
[
[
[
[
[
[
[
20] V. Wintgens, P. Valat, J. Kossanyi, L. Biczok, A. Demeter, T. B e´ rces,
J. Chem. Soc. Faraday Trans. 90 (1994) 411–421.
21] M.A. Rodrigues, S. Brochsztain, T.C. Barros, M.S. Baptista, M.J.
Politi, J. Photochem. Photobiol. A 70 (1999) 35–39.
22] T.C. Barros, G.R. Molinari, P. Berci Filho, V.G. Toscano, M.J. Politi,
J. Photochem. Photobiol. A 76 (1993) 55–60.
23] E.R. Triboni, P. Berci Filho, R.G. De Souza Berlinck, M.J. Politi,
Synth. Commun. 34 (2004) 1989–1990.
The authors would like to thank FAPESP, CNPq and
CAPES agencies for financial support.
Appendix A. Supplementary data
Supplementary data have been deposited with the British
Library at Boston Spa, Wetherby, West Yorks, UK as Sup-
plementary Publication No. SUP. . . (. . . pages). Persons
wishing to obtain copies of the deposited material should
contact Service Enquiries, British Lending Library, Boston
Spa, Wetherby, West Yorks LS23 7BQ, UK citing the SUP
number. Tel.: +44 1937 546 060; fax: +44 1937 546 333;
e-mail: dsc-sutomer-services@bl.uk.
24] W. Adam, X. Qian, C.R. Saha-M o¨ ller, Tetrahedron 49 (1993)
417–422.
25] A. Pardo, E. Martin, J.M.L. Poyato, J.J. Camacho, M.F. Bra n˜ a, J.M.
Castellano, J. Photochem. Photobiol. A 41 (1987) 69–78.
26] A. Pardo, J.M.L. Poyato, E. Martin, J.J. Camacho, D. Reyman,
M.F. Bra n˜ a, J.M. Castellano, J. Photochem. Photobiol. A 46 (1989)
323–328.
[
27] A. Pardo, E. Martin, J.M.L. Poyato, J.J. Camacho, J.M. Guerra, R.
Weigand, M.F. Bra n˜ a, J.M. Castellano, J. Photochem. Photobiol. A
48 (1989) 259–263.
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/j.saa.2005.05.
[
[
[
[
28] C. Lerf, P. Suppan, J. Chem. Soc. Faraday Trans. 88 (1992) 963–
9
69.
29] Y. Migron, Y. Marcus, J. Chem. Soc. Faraday Trans. 87 (1991)
339–1343.
0
08.
1
30] J. Tomasi, R. Bonaccorsi, R. Cammi, F.J. Olivares del Valle, J. Mol.
Struct. (Theochem.) 234 (1991) 401–424.
31] H. Agren, K.V. Mikkelsen, J. Mol. Struct. (Theochem.) 24 (1991)
References
4
25–467.
[
1] C. Reichardt, Chem. Rev. 94 (1994) 2319–2358.
[
[
32] O.W. Kolling, J. Phys. Chem. 95 (1991) 3950–3954.
33] P. Chatterjee, S. Bagchi, J. Chem. Soc. Faraday Trans. 87 (1991)
[2] E.B. Tada, P.L. Silva, O.A. El Seoud, J. Phys. Org. Chem. 16 (2003)
1
–9.
3] S.E. DeBolt, P.A. Kollman, J. Am. Chem. Soc. 112 (1990)
515–7524.
587–589.
[
[
[
[
[
[
34] M. Ros e´ s, C. R a´ fols, J. Ortega, E. Bosch, J. Chem. Soc. Perkin
Trans. 2 (1995) 1607–1615.
35] R.D. Skwierczynski, K.A. Connors, J. Chem. Soc. Perkin Trans. 2
7
[
[
4] T. Fox, N. R o¨ sch, Chem. Phys. Lett. 191 (1992) 33–37.
5] J.E. Dubois, E. Goetz, A. Bienvenue, Spectrochim. Acta A 24 (1964)
(
1994) 467–472.
36] A. Coccia, P.L. Indovina, F. Podo, V. Viti, Chem. Phys. 7 (1975)
0–40.
37] R.C. Pemberton, C.J. Mash, J. Chem. Thermodyn. 10 (1978)
67–888.
38] W.E. Acree Jr., J.R. Powell, S. Tucker, J. Chem. Soc. Perkin Trans.
(1995) 529–531.
1
815–1828.
6] M.A. Kessler, O.S. Wolfbeis, Spectrochim. Acta A 47A (1991)
87–192.
7] C. Reichardt, S. Asharin-Fard, G. Sch a¨ fer, Liebigs Ann. Chem.
1993) 23–34.
[
[
[
3
1
8
(
8] C. Reichardt, S. Asharin-Fard, Angew. Chem. 103 (1991) 614–616;
Angew. Chem. Int. Ed. Engl. 30 (1991) 558–559.
2