Differentiating between dipolarity and polarizability effects of solvents using the solvatochromism of barbiturate dyes

Article abstract of DOI:10.1002/ejoc.200800355

Barbiturate dyes containing either a dipolar or polarizable chromophoric system have been synthesized and their solvatochromism investigated. The effect of specific and non-specific solvent interactions on the position of their UV/Vis absorption bands has been evaluated by using the Kamlet-Taft and Catalan solvent parameter sets. Furthermore, the effect on the solvatochromic behaviour of different substitution patterns on the barbiturate moiety has been examined. Wiley-VCH Verlag GmbH & Co. KGaA, 2008.

Full text of DOI:10.1002/ejoc.200800355

FULL PAPER
DOI: 10.1002/ejoc.200800355
Differentiating Between Dipolarity and Polarizability Effects of Solvents Using
the Solvatochromism of Barbiturate Dyes
Mirko Bauer,[a] Anja Rollberg,[a] Anina Barth,[a] and Stefan Spange*[a]
Keywords: Solvatochromism / Solvent effects / Polarity / Barbiturates / Donor–acceptor systems / Polyenes
Barbiturate dyes containing either a dipolar or polarizable
chromophoric system have been synthesized and their solva-
tochromism investigated. The effect of specific and non-spe-
and Catalán solvent parameter sets. Furthermore, the effect
on the solvatochromic behaviour of different substitution pat-
terns on the barbiturate moiety has been examined.
cific solvent interactions on the position of their UV/Vis ab- (© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
sorption bands has been evaluated by using the Kamlet–Taft
Germany, 2008)
only one parameter. The use of a polarizability correction
term dδ[12] is not suitable for solving the problem, because
it does not allow differentiation of the polarizability effects
between two related solvents, for example, aromatic or halo-
genated hydrocarbons. Another disadvantage of the Kam-
let–Taft parameters is that their determination is not based
on a well-defined reference process, rather they are derived
from an average of measurements using numerous probes.
Therefore, the search for an alternative solvent polarity
scale with well-defined reference processes is required. In
recent years, Catalán and co-workers have introduced three
alternative empirical polarity scales: SA, SB and SPP, which
are formally related to the Kamlet–Taft parameters α, β and
π*, respectively.[13] The advantage of Catalán’s concept is
that each solvatochromic solvent parameter is based on a
pair of well-defined reference homomorph solvatochromic
probes. In 2004, Catalán and Hopf reported an important
fourth scale: the polarizability parameter SP.[14] This novel
scale promises to measure gradual differences in the sur-
rounding’s polarizability by the solvatochromic method.
One can now formulate another multiparameter equation,
analogous to Equation (1), which comprises two param-
eters for specific and two parameters for non-specific inter-
actions [Equation (2)]. It has, however, been pointed out
that the SPP and SP scales are actually not independent of
each other.[14] Thus, it seems reasonable to include only one
of these parameters at a time when using Equation (2).
Introduction
Solvatochromism[1] has been established as a readily ap-
plicable and powerful tool to investigate the influence of
numerous polarity effects[2] of the surroundings of a reac-
tion on reaction rates, equilibria or the physicochemical
properties of molecules. Therefore, the use of solvatoch-
romic probes has been established to measure the polarity
effects of environments comprising not only well-behaved
solvents and room-temperature ionic liquids,[3,4] but also in-
organic surfaces,[5,6] biomolecules
dia.[10]
and organized me-
[7–9]
In recent decades, the Kamlet–Taft approach[11] has suc-
cessfully been applied to separate the influence of non-spe-
cific chemical interactions, including electrostatic effects (di-
polarity/polarizability), from specific interactions, that is,
hydrogen-bonding, which are related to the molecular struc-
ture of a compound. The commonly used simplified Kam-
let–Taft equation applied to the UV/Vis absorption shift
(νmax) of a solute is given by Equation (1),[11] where ν
˜
˜
max,0
is the solute property of a reference system, for example, a
non-polar medium, α describes the HBD (hydrogen-bond
donating) ability, β the HBA (hydrogen-bond accepting)
ability, π* the dipolarity/polarizability of the solvents and
a, b and s are solvent-independent correlation coefficients
that indicate the contribution of the different solvent effects
to the UV/Vis absorption shift.
ν
= ν
+ aα + bβ + sπ*
(1)
˜
˜
max
max,0
ν
= ν
+ aSA + bSB + sSPP + tSP
(2)
˜
˜
max
max,0
One problem of the Kamlet–Taft π* scale is that the di-
polarity and polarizability of the solvent are included in
With this in mind we have designed a range of novel
barbiturate dyes that are structurally related but supposed
to show different sensitivities towards electrostatic effects.
To examine this sensitivity and separate the individual sol-
vation effects, the solvatochromic properties of these dyes
were investigated in detail, and the coefficients of the indi-
vidual interaction contributions were determined by using
[a] Department of Polymer Chemistry, Chemnitz University of
Technology,
Straße der Nationen 62, 09107 Chemnitz, Germany
Fax: +49-371-531-21239
E-mail: stefan.spange@chemie.tu-chemnitz.de
Supporting information for this article is available on the
WWW under http://www.eurjoc.org or from the author.
Eur. J. Org. Chem. 2008, 4475–4481
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4475

M. Bauer, A. Rollberg, A. Barth, S. Spange
FULL PAPER
multiple correlation analysis of both Kamlet–Taft’s Equa-
tion (1) and Catalán’s Equation (2). Related barbituric acid
derivatives have already been widely used in materials sci-
ence [e.g., in materials for non-linear optics (NLO)][15–19]
and analysis.[20–23] The barbiturate moiety also offers the
possibility of a facile variation of its electron-demanding
effects as well as its interaction with the surroundings by
introducing different substituents. In particular, we wanted
to determine whether the free NH functions and those
capped by alkyl groups have an effect on the chromophoric
π-electron system. Furthermore, 2-thiobarbiturates have
been investigated to examine the effects of the more dipolar
and polarizable thiocarbonyl moiety.
Figure 1. Barbiturate dyes studied in this work.
Results and Discussion
Electron-poor retinyl dyes related to 2 have already found
interest as dyes and NLO materials[25] as well as anticancer
drugs.[26] However, no such compounds with barbiturate
functions have been reported so far.
We have used two types of solvatochromic dyes that in
principle can show all kinds of interactions with a solvent
(Figure 1). The barbituric acid moiety, which acts as an
electron-withdrawing substituent, was coupled either with
two 4-(dimethylamino)phenyl groups, which results in a di-
polar structure (1), or with a retinyl group, to yield a polar-
izable all-(E) linear polyene structure (2).
Synthesis
The solvatochromism of one of the merocyanine-type
According to the literature, the compounds to be investi-
dyes, 1b, as well as a related dye has already been studied gated should be readily accessible by aldol condensation of
by Rezende et al.[21a] These dyes show positive solvatochro- retinal or Michler’s ketone with the respective barbituric
mism as the position of the solvatochromic UV/Vis absorp- acid derivatives.[21a,24,27,28]
tion band is shifted to a lower energy with increasing di-
Although this route proved to be suitable for preparing
polarity/polarizability and HBD strength of the solvent. compounds 2a–d (Scheme 1), the method described by Re-
However, Rezende reported only on the synthesis and sol- zende[21a] for 1b, heating of Michler’s ketone and N,N-di-
vatochromism of N,N-dimethyl-substituted barbituric acid methylbarbituric acid at reflux in acetic anhydride/acetic
dyes. The alkylated thiobarbiturate dye 1d is described in a acid, failed for the NH derivatives. Finally, the method of
patent, yet no characterization data is given there.[24a] Sur- choice involved a two-step reaction starting from Michler’s
prisingly, no information on the NH derivatives of 1 were hydrol blue (Scheme 2).[29] The idea for the synthesis of the
found in the literature in spite of a patent reference[24b] that intermediates 3 was inspired by Mayr’s reactivity scale for
names a few related compounds.
nucleophiles (for the anion of barbituric acid, N Ͼ 10 was
Scheme 1. Synthesis of compounds 2a–d.
Scheme 2. Synthesis of compounds 1a,c,d.
4476
www.eurjoc.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2008, 4475–4481

Dipolarity and Polarizability Effects of Solvents
estimated[30a]) and electrophiles (E = –7.0 for Michler’s hy- Table 2. Solvent polarity parameter sets of Kamlet–Taft[34] and Ca-
talán[13,14] for the solvents in Table 1.
Kamlet–Taft
drol blue cation[30b]). The rate constant k for the recombina-
tion reaction, estimated to be about 103 –1 s–1 from Equa-
tion (3),[31] provides a solid background for the retrosynthe-
sis.
Catalán
SB SPP
0.00 0.00 –0.04 0.000 0.056 0.519 0.6164
Solvent
α
β
π*
SA
SP
n-Hexane
Triethylamine
1-Butanol
Ethyl acetate
Tetrahydrofuran
Benzene
Methanol
TFE
Acetonitrile
0.00 0.71
0.84 0.84
0.00 0.45
0.00 0.55
0.00 0.10
0.98 0.66
1.51 0.00
0.19 0.40
0.14 0.000 0.885 0.617 0.6603
0.47 0.341 0.809 0.837 0.6742
0.55 0.000 0.542 0.795 0.6558
0.58 0.000 0.591 0.838 0.7139
0.59 0.000 0.124 0.667 0.7929
0.60 0.605 0.545 0.857 0.6079
0.73 0.893 0.107 0.908 0.5431
0.75 0.044 0.286 0.895 0.6448
1.00 0.072 0.647 1.000 0.8295
logk20 °C = s(N + E)
(3)
The recombination products 3, being sensitive to air and
light, were isolated as slightly coloured solids in reasonable
1
yields. The H NMR spectrum of 3a in [D6]DMSO shows
two doublets at δ = 4.2 and 4.7 ppm, and its 13C NMR
spectrum shows two signals at δ = 52 and 53 ppm, which
shows that this compound exists only in the triketo form.
Remarkably, the thiobarbiturates 3c and 3d are present in
their enol form in [D6]DMSO, as indicated by 13C NMR
peaks at δ = 97 and 93 ppm, respectively, and a single sing-
Dimethyl sulfoxide 0.00 0.76
with the dimethylamino groups (see below). In the cases of
split bands, the UV/Vis band at longest wavelength was
used for the correlation analysis.
1
let at δ ≈ 5.0 ppm in the H NMR spectra. Finally, oxi-
dation of 3a,c,d was accomplished with p-chloranil in water
to yield 1a,c,d as intensely coloured solids.
Solvatochromism
The solvatochromism of the novel dyes was studied in
13–30 solvents for which the solvent parameters of Kamlet–
Taft and Catalán are available (see Tables 1 and 2). The UV/
Vis spectra of these compounds showed no dependence on
the dye concentration, and thus the formation of dimers or
higher aggregates can be neglected.
In aprotic solvents, compounds 1a,c,d show a single ab-
sorption band in the visible region. However, in protic sol-
vents this band is considerably broadened and is split into
two bands in more acidic media (Figure 2). This effect can
probably be attributed to protonation reactions and is in
Figure 2. UV/Vis absorption spectra of 1a in different solvents.
For the retinal-type dyes 2a–d there is also one absorp-
accordance with the observations made by Rezende et al. tion band, yet there is no indication of proton transfer [only
for 1b.[21a] His semiempirical calculations showed that this in 2,2,2-trifluoroethanol (TFE) does a small shoulder in the
protonation is likely to occur at the carbonyl moieties. Our UV/Vis spectrum appear, see Figure 3]. Nevertheless, rea-
solvatochromic investigations indeed confirm that HBD sonable correlations are only obtained when several sol-
solvents interact mainly with the carbonyl moieties and not vents, for example, triethylamine and chlorinated solvents,
Table 1. UV/Vis absorption maxima of compounds 1 and 2 in selected solvents and the extent of the solvatochromic shift.
Solvent
ν
[10–3 cm–1]
˜
max
1a
1b[a]
21.10[c]
1c
1d
19.76
2a
2b
2c
19.61
2d
[b]
[b]
[b]
n-Hexane
Triethylamine
1-Butanol
Ethyl acetate
Tetrahydrofuran
Benzene
Methanol
TFE
Acetonitrile
Dimethyl sulfoxide
21.37[c]
20.16[d]
20.83
20.79
21.28
19.96[c]
23.81[d]
18.87
19.27
19.34
21.46[c]
19.31[e]
20.04
–
19.96[c]
18.55[e]
18.87
19.38
18.02
18.59
18.52
18.59
17.15[e]
16.64[f,g]
18.18
17.86
21.41[d]
20.33
20.88
21.10
19.96
20.49
18.73[g]
20.58
20.33
19.01[d]
19.16
19.57
19.69
18.87
19.27
17.54[g]
19.27
19.38
19.01
20.00
19.84
–
20.16
18.98
19.53
18.66
20.33
19.01
18.05[f]
16.98[f,g]
19.42
18.05[f]
17.06[f,g]
19.53
19.01
17.36[f]
16.58[f,g]
18.55
21.01[d]
19.49
19.01
17.79[g]
19.23
21.23
20.00
19.31
18.25
18.52
∆λ [nm]
˜
123
4481
112
4032
102
3376
96
3163
61
2415
58
1901
67
2337
61
2166
∆ν [cm–1]
[a] Data from ref.[21a] [b] Substance is insoluble. [c] Solvent with the highest hypsochromic shift. [d] Value excluded from correlation. [e]
Broad maximum. [f] Spectrum shows two maxima. [g] Solvent with the highest bathochromic shift.
Eur. J. Org. Chem. 2008, 4475–4481
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
4477

M. Bauer, A. Rollberg, A. Barth, S. Spange
FULL PAPER
are excluded from the correlation analyses. The reason for Table 4. Values of the solvent-independent correlation coefficients
a, b, s and t for the Catalán equation [Equation (2)], wavenumber
the deviation of these solvents is not clear yet and requires
of a reference system (νmax,0), correlation coefficient (r), standard
˜
further investigation. It might for instance be due to a reac-
tion with the solvent leading to differently coloured prod-
ucts (see, for example ref.[32]). As the colour of ethanolic
solutions of these dyes slowly fades over a few days only
when exposed to light, some photoisomerization in certain
solvents cannot be ruled out.[33]
deviation (sd), significance (f) and number of solvents (n) for com-
pounds 1 and 2.
ν
a
b
s
t
r
sd
f
n
˜
max,0
1a 21.41 –2.60 1.35 –2.70
1b 22.77 –2.55 –3.62
–
–
–
–
0.96 0.24 Ͻ0.0001 22
0.98 0.23 Ͻ0.0001 13
0.95 0.22 Ͻ0.0001 23
0.95 0.23 Ͻ0.0001 26
–
1c 20.44 –2.06 0.89 –2.57
1d 21.36 –1.61
2a 22.92 –2.27 1.08
2b 24.12 –2.23
2c 22.37 –2.30 0.99
2d 22.43 –2.03
–
–3.54
–
–
–
–
–3.74 0.92 0.23 Ͻ0.0001 27
–4.57 0.85 0.33 Ͻ0.0001 30
–4.69 0.92 0.25 Ͻ0.0001 30
–4.40 0.90 0.23 Ͻ0.0001 30
–
–
tochromic range of about 3000 cm–1 for the thiobarbitur-
ates, compared with more than 4000 cm–1 for the barbitu-
rates.
For 2a–d the extent of the solvatochromic shift is almost
the same, with ∆ν ≈ 2000 cm–1. Although the conjugated π
˜
system is quite long in these dyes, this solvatochromic range
is rather small, clearly due to the lack of a strong electron
donor. The solvent-independent correlation coefficients are
not systematically different for the thio- and oxobarbitur-
ates, but again the UV/Vis absorption energy is lower for
the thiobarbiturates.
Figure 3. UV/Vis absorption spectra of 2a in different solvents.
The influence of solvent properties on the shift of ν
˜
max
has been evaluated by means of multiple regression analyses
using the solvent polarity parameter sets of Kamlet–Taft[34]
and Catalán.[13,14] For comparison, the data for 1b from
ref.[21a] have also been included in these analyses. The best-
fitting correlation results are summarized in Tables 3 and 4.
Specific Interactions
Specific solvent–solute interactions through hydrogen-
bonding are expressed by the solvent acidity, α or SA, and
the solvent basicity, β or SB. The respective correlation co-
efficients obtained for both solvent parameter sets (Tables 3
and 4) are in good agreement with each other, so both mod-
els are equally suitable for describing these kinds of interac-
tions.
Table 3. Values of the solvent-independent correlation coefficients
a, b and s for the Kamlet–Taft equation [Equation (1)], wave-
number of a reference system (νmax,0), correlation coefficient (r),
˜
standard deviation (sd), significance (f) and the number of solvents
(n) for compounds 1 and 2.
The influence of hydrogen-bonding on the UV/Vis ab-
sorption shift is essentially the same for both types of dyes,
1 as well as 2. This shows that the barbiturate moiety plays
the major role in this regard (Figure 4) and that the struc-
ture of the chromophoric π system is of less importance.
ν
a
b
s
r
sd
f
n
˜
max,0
1a 20.94 –1.62 0.84 –2.21
1b 21.12 –1.58 –2.05
1c 19.73 –1.26 0.49 –1.81
1d 19.65 –1.06 –1.84
2a 20.67 –0.87 1.28 –1.04
2b 21.22 –0.53 –1.00
2c 19.50 –0.77 1.29 –1.33
2d 19.83 –0.48 –1.33
0.97 0.23 Ͻ0.0001 22
0.98 0.20 Ͻ0.0001 13
0.96 0.21 Ͻ0.0001 23
0.98 0.15 Ͻ0.0001 26
0.83 0.33 Ͻ0.0001 27
0.67 0.47 0.0003
0.86 0.31 Ͻ0.0001 30
0.91 0.23 Ͻ0.0001 30
–
–
–
30
–
It can be seen that for 1a–d the magnitude of the coeffi-
cients a, b and s from Equations (1) or (2) of the thiobarbit-
urates are generally lower than those of the barbiturates.
This can be explained by the stronger electron-accepting
ability of the thiocarbonyl group, which enhances the –M
effect of the thiobarbiturate moiety. Thus, the overall di-
polarity of the molecule is increased in the ground state and
decreased in the excited state.[17b,23] As both these states are
approaching the cyanine limit with ideal charge delocaliza-
tion, the push–pull chromophore becomes less susceptible
Figure 4. Possible hydrogen-bonding interactions between solute
and solvent and their effect on the UV/Vis absorption, exemplified
with dye 1a.
As expected, the UV/Vis absorption maxima undergo a
to solvent effects. This explanation is further supported by bathochromic shift with increasing solvent HBD capacity,
the lower UV/Vis absorption energy and the smaller solva- as shown by the negative a terms. This can be explained by
4478
www.eurjoc.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2008, 4475–4481

Dipolarity and Polarizability Effects of Solvents
the interaction of HBD solvents with the barbiturate car-
bonyl moieties, which reduces their electron density and
thus increases the push–pull character of the chromophore.
On the other hand, interaction of the dimethylamino
groups of 1 with HBD solvents, which would lead to a hyp-
sochromic shift and positive a values, clearly is of minor
importance. In contrast, the HBA properties of the sol-
vents, represented by the b coefficient, play only a minor
role. Because of the increase in electron-release due to hy-
drogen-bonding of the NH groups, the UV/Vis absorption
maxima shift to slightly higher energies. For the N-alkylated
derivatives the basicity term is negligible due to the lack of
a suitable HBD moiety.
To examine whether these hydrogen-bonding interactions
are interfered with by steric effects, some bulky alcohols,
namely tert-butyl alcohol and cyclohexanol, were tested.
Comparison of their measured UV/Vis absorption maxima
with those calculated with the correlation coefficients of
Tables 3 or 4 yields a fairly good agreement (see the Sup-
porting Information), so steric effects are negligible in this
respect.
Figure 5. Possible resonance structures contributing to the ground-
and excited-state structures of 1a.
(see the Supporting Information). However, inclusion of the
SP parameter instead of SPP leads to a substantial improve-
ment in the correlation quality. Then a very strong bathoch-
romic shift with increasing solvent polarizability (large
negative t coefficients) can be stated, albeit this effect shows
no marked trend with the substitution pattern. This is in-
dicative of a π–π* excitation in which the excited state is
more polarizable than the ground state. To account for the
dipolarity effects, some additional parameters were in-
cluded in the analyses using the results in Table 4 as the
starting point. Although none of these parameters led to
any improvement for the barbiturates 2a and 2b, some im-
provement was achieved for 2c and 2d. The Kirkwood func-
tion,[1c,35] Km = (εr – 1)/(2εr + 1), was found to give the best
results, leading to Equation (4) for 2d (see the Supporting
Information for all results). Thus, at least the excitation of
the more polar thiobarbiturates seems to exhibit some
charge-transfer character as well.
Non-Specific Interactions
Large differences occur when regarding non-specific in-
teractions. These are expressed by Catalán’s SP parameter
(solvent polarizability) as well as by the π* or SPP param-
eter (both of which represent a combination of the solvent
dipolarity and polarizability).
ν
˜
max·[10–3 cm–1] (2d) = 22.54 – 1.54SA – 3.88SP – 1.49Km
(4)
As expected, the dipolar compounds 1a–d show positive
solvatochromism with regard to π* or SPP (as indicated by
the large negative s terms). This indicates a higher dipole
moment of the electronically excited state compared with
the ground state. Therefore the neutral resonance structure
(I in Figure 5) clearly dominates over the zwitterionic ones
(e.g., II) in the ground state, whereas in the excited state the
reverse case is found. This behaviour is in fact typical for
intramolecular charge-transfer (ICT) processes. Further-
more, with the Catalán model a markedly stronger depen-
dence on the dipolarity/polarizability for the N,N-dialkyl-
ated dyes 1b and 1d relative to the NH derivatives 1a and
1c is indicated. The s values of the Kamlet–Taft model on
the other hand do not depend significantly on the substitu-
r = 0.94, sd = 0.19, f Ͻ 0.0001, n = 29
Interestingly, similar results are obtained when Km is re-
placed by SPP, which leads to Equation (2). It should, how-
ever, also be noted that the error in this additional param-
eter, no matter which one is used, is rather high in compari-
son with SA, SB and SP. Thus, the conclusions drawn from
these last correlation analyses should only be regarded as
an approximation.
Inverted Solvatochromism?
An occasionally described effect when dealing with poly-
tion pattern of the barbiturate moiety. The Catalán values ene dyes is inverted solvatochromism,[18–20,36] that is, the
are thus in better accord with the stronger electron-releas- change from positive to negative solvatochromism with in-
ing effect of the alkyl groups, which leads to a less dipolar creasing solvent polarity. This phenomenon is ascribed to a
chromophore compared to the NH-substituted compounds solvent-induced change in the ground-state structure from
1a and 1c. The inclusion of Catalán’s polarizability (SP) a less dipolar polyene-like structure (non-polar solvents) to
parameter does not lead to any improvement in the corre- a strongly dipolar betaine-like structure (polar solvents),
lation quality. It can therefore be concluded that the solvent the maximum wavelength being at the cyanine limit.
polarizability exerts only a small influence on the UV/Vis
Although at first sight the measured UV/Vis absorption
shifts of 1a–d.
wavelengths of 2a–d also follow the same trend in polar
The retinal-type dyes 2a–d show a somewhat different solvents (e.g., alcohols, acetonitrile), our investigations indi-
behaviour. Kamlet–Taft’s π* parameter suggests a moderate cate a pronounced positive solvatochromism of these com-
influence of the solvent’s dipolarity/polarizability, yet this pounds even in polar solvents. This contrasting behaviour
model yields only a mediocre correlation quality. When may only partially be attributed to the lack of a strong elec-
using Catalán’s SPP parameter, the results are even worse tron-donating moiety. Regarding the good correlation with
Eur. J. Org. Chem. 2008, 4475–4481
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
4479

M. Bauer, A. Rollberg, A. Barth, S. Spange
FULL PAPER
was stirred at room temperature for 96 h. The solid was vacuum-
filtered, washed with water and acetone and dried in vacuo. Owing
to its limited stability, the product was used immediately in the
following reaction without further purification. Light-blue solid
(0.33 g, 30%). M.p. 158–160 °C (decomp.). 1H NMR (250 MHz,
[D6]DMSO): δ = 2.87 (s, 12 H), 4.21 (d, J = 3.9 Hz, 1 H), 4.72 (d,
J = 3.9 Hz, 1 H), 6.71 (d, J = 8.4 Hz, 4 H), 7.08 (d, J = 8.4 Hz, 4
H), 11.08 (br. s, 2 H) ppm. 13C NMR (63 MHz, [D6]DMSO): δ =
40.0, 52.4, 53.3, 112.6, 129.2, 137.3, 150.5, 163.8, 169.8 ppm. IR
Catalán’s SP parameter, it seems more reasonable to con-
sider the solvent polarizability as a major influence on the
solvatochromic behaviour instead of the frequently used di-
polarity-based solvent parameters. In particular, when
rather weak donor–acceptor systems are investigated, these
considerations should be taken into account.
(KBr): ν = 3354, 3045, 1694, 1586, 1366, 822 cm–1. ESI-MS: m/z
˜
Conclusions
(%) = 253.19 (100) [CH(C6H4NMe2)2]+, 381.23 (23) [M + H]+.
By using different solvatochromic dyes it can be shown
that the Kamlet–Taft equation works as well as the Catalán
equation when the highly dipolar barbiturate dyes 1a–d are
considered. A possible contribution of the solvent polari-
zability already is comprised in the π* parameter so that
further differentiation does not improve the correlation
quality. With regard to weakly dipolar but highly polariz-
able chromophores like 2a–d, reasonable results are only
obtained by including Catalán’s SP parameter due to the
unequal contribution of the solvent dipolarity and polariz-
ability.
5-{Bis[p-(dimethylamino)phenyl]methylene}barbituric Acid (1a): A
suspension of 3a (0.20 g, 0.5 mmol) in water (50 mL) was treated
with p-chloranil (0.12 g, 0.5 mmol) and stirred at room temperature
for 144 h. The solid was vacuum-filtered, washed with water and
suspended in dichloromethane. The mixture was filtered and the
filtrate was concentrated and dried in vacuo. Remaining impurities
were finally removed by sublimation in vacuo at 200 °C. Dark-
1
green solid (0.11 g, 57%). M.p. 211 °C. H NMR (250 MHz, [D6]-
DMSO): δ = 3.07 (s, 12 H), 6.68 (d, J = 9.0 Hz, 4 H), 7.09 (d, J =
9.0 Hz, 4 H), 10.31 (br. s, 2 H) ppm. 13C NMR (63 MHz, [D6]-
DMSO): δ = 39.6, 105.4, 110.4, 128.3, 136.2, 150.6, 153.6, 163.4,
175.7 ppm. IR (KBr): ν = 3184, 3083, 1698, 1652, 1586, 1362,
˜
Specific interactions through hydrogen-bonding, ex-
pressed by solvent acidity and basicity, are essentially the
826 cm–1. ESI-MS: m/z (%) = 379.11 (100) [M + H]+.
same for both kinds of dyes and are described equally well (2Z,4Z,6Z,8Z)-5-[3,7-Dimethyl-9-(2,6,6-trimethylcyclohex-1-enyl)-
nona-2,4,6,8-tetraenylidene]barbituric Acid (2a): This compound
has to be prepared and stored under strict exclusion of light. Barbi-
turic acid (0.262 g, 2.05 mmol) was dissolved in ethanol (20 mL)
at 45 °C. After addition of a solution of all-(E)-retinal (0.567 g,
2.00 mmol) in ethanol (20 mL), stirring was continued at 45 °C for
1 h. The mixture was cooled to room temperature and concentrated
in air overnight. The resulting solid was filtered, washed with hot
water and dried in vacuo. Violet solid (0.742 g, 94%). M.p.
by both solvent polarity scales. Thus, they can be attributed
mainly to the barbiturate moiety and are hardly influenced
by the nature of the chromophoric system. On the other
hand, non-specific interactions depend mainly on the di-
polarity and polarizability of the chromophoric system and
are only slightly affected by the substitution pattern of the
barbiturate moiety.
1
Ͼ210 °C. H NMR (250 MHz, CDCl3): δ = 1.05 (s, 6 H), 1.50 (m,
no integration possible due to signal overlap), 1.63 (m, 2 H), 1.75
(d, J = 0.6 Hz, 3 H), 2.05 (m, 5 H), 2.31 (d, J = 0.6 Hz, 3 H), 6.26
(m, 2 H), 6.45 (d, J = 15.8 Hz, 1 H), 6.65 (d, J = 14.8 Hz, 1 H),
7.30 (dd, J = 14.8, J = 11.8 Hz, no integration possible due to
signal overlap), 7.96 (d, J = 13.3 Hz, 1 H), 8.02 (br. s, 1 H), 8.18
Experimental Section
General: Solvents were dried and distilled according to standard
procedures prior to use. Bis[p-(dimethylamino)phenyl]methylium
tetrafluoroborate was prepared according to known procedures.[30]
All other compounds were commercially available and used as re-
ceived. NMR spectra were recorded with a Varian Inova-400 or a
Bruker Avance 250 spectrometer using the solvent residue signal as
the internal standard. The signal designations are given as follows:
s = singlet, d = doublet, t = triplet, q = quartet, m = multiplet,
and br. = broad. Owing to the generally low solubility of 2a–d, no
evaluable 13C NMR spectra of these compounds could be obtained.
IR spectra were recorded in diffuse reflection with a BioRad FT-
IR-Spektrometer FTS 165. Elemental analyses were obtained by
using a Vario EL from Elementaranalysensysteme GmbH (Hanau).
Mass spectra were obtained with a Hewlett–Packard HP 5988a
mass spectrometer. The UV/Vis spectra of freshly prepared solu-
tions (c ≈ 1ϫ10–5 to 1ϫ10–4 molL–1) were measured with an MCS
400 diode-array spectrometer from Carl Zeiss Jena with a resolu-
tion of 1 nm. Multiple-regression analyses were performed with the
statistical program Origin 5.0.[37]
(br. s, 1 H), 8.53 (d, J = 13.3 Hz, 1 H) ppm. IR (KBr): ν = 3213,
˜
3078, 2926, 2862, 2733, 1736, 1676, 1657, 1574, 1516, 1418, 1373,
1339, 1304, 1234, 1184, 961, 812, 790 cm–1. C24H30N2O3 (394.51):
calcd. C 73.07, H 7.66, N 7.10; found C 73.46, H 7.67, N 6.88.
Supporting Information (see also the footnote on the first page of
this article): Characterization data for all new compounds, UV/Vis
absorption maxima of 1a–d and 2a–d in various solvents, multiple-
regression data of the solvatochromic studies.
Acknowledgments
Financial support from the Deutsche Forschungsgemeinschaft
(DFG) is gratefully acknowledged. We also thank the BASF AG,
Ludwigshafen, for kindly providing all-(E)-retinal and Prof. M. An-
tonietti (Max-Planck-Institut für Kolloid- und Grenzflächenfor-
schung, Golm) for hosting A. B. during her diploma thesis.
Typical Procedures
5-{Bis[p-(dimethylamino)phenyl]methyl}barbituric Acid (3a): A sus-
pension of barbituric acid (0.39 g, 3.1 mmol) in water (100 mL) was
treated with KOH (0.16 g, 2.9 mmol) dissolved in water (25 mL).
After stirring for 30 min, bis[p-(dimethylamino)phenyl]methylium
tetrafluoroborate (1.00 g, 2.9 mmol) was added, and the mixture
[1] a) C. Reichardt, Chem. Rev. 1994, 94, 2319–2358; b) P. Suppan,
N. Ghoneim, Solvatochromism, The Royal Society of Chemis-
try, London, 1997; c) C. Reichardt, Solvents and Solvent Effects
in Organic Chemistry, 3rd ed., Wiley-VCH, Weinheim, 2003.
[2] P. Müller, Pure Appl. Chem. 1994, 66, 1151.
4480
www.eurjoc.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2008, 4475–4481

Dipolarity and Polarizability Effects of Solvents
[3] C. Chiappe, D. Pieraccini, J. Phys. Org. Chem. 2005, 18, 275– [21]
a) M. C. Rezende, P. Campodonico, E. Abuin, J. Kossanyi,
Spectrochim. Acta Part A 2001, 57, 1183–1190; b) M. C. Re-
zende, P. Flores, J. Guerrero, L. Villarroel, Spectrochim. Acta
Part A 2004, 60, 1637–1640.
297.
[4] A. Oehlke, K. Hofmann, S. Spange, New J. Chem. 2006, 30,
533–536.
[5] a) S. Spange, A. Reuter, E. Vilsmeier, Colloid Polym. Sci. 1996,
274, 59–69; b) S. Spange, A. Reuter, Langmuir 1999, 15, 141–
150; c) S. Spange, A. Reuter, D. Lubda, Langmuir 1999, 15,
2103–2111; d) S. Spange, Y. Zimmermann, A. Gräser, Chem.
Mater. 1999, 11, 3245–3251; e) S. Spange, E. Vilsmeier, Y. Zim-
mermann, J. Phys. Chem. B 2000, 104, 6417–6428.
[6] a) R. S. Helburn, S. C. Rutan, J. Pompano, D. Mitchern, W. T.
Patterson, Anal. Chem. 1994, 66, 610–618; b) S. C. Rutan, J. M.
Harris, J. Chromatogr. A 1993, 656, 197–215.
[22]
[23]
[24]
S. Nakatsuji, T. Yahiro, K. Nakashima, S. Akiyama, H. Naka-
zumi, Chem. Express 1989, 4, 601–604.
G. U. Bublitz, R. Ortiz, S. R. Marder, S. G. Boxer, J. Am.
Chem. Soc. 1997, 119, 3365–3376.
a) S. Horie, M. Nagata, J. Nakano, H. Sato, DE-3150266, 1982,
Chem. Abstr. 1983, 99, 131329; b) K. Inai, N. Go, H. Miyazaki,
T. Miyaji, JP-03012660, 1991, Chem. Abstr. 1991, 115, 146585.
See selected publications: a) P. E. Blatz, J. H. Mohler, H. V. Na-
vangul, Biochemistry 1972, 11, 848–855; b) T. Baasov, M.
Sheves, J. Am. Chem. Soc. 1985, 107, 7524–7533; c) Y. Kawabe,
H. Ikeda, T. Sakai, K. Kawasaki, J. Mater. Chem. 1992, 2,
1025–1031.
See selected publications: a) N. Acton, A. Brossi, D. L. New-
ton, M. B. Sporn, J. Med. Chem. 1980, 23, 805–809; b) C. C.
Willhite, M. I. Dawson, K. J. Williams, Toxicol. Appl. Pharma-
col. 1984, 74, 397–410; c) E. P. Jaeger, P. C. Jurs, T. R. Stouch,
Eur. J. Med. Chem. 1993, 28, 275–290.
a) A. M. Asiri, Molecules 2000, 5, M170; b) A. M. Asiri, Mole-
cules 2000, 57, 1183–1190; c) A. M. Asiri, Appl. Orgonomet.
Chem. 2001, 15, 905–915.
M. A. Whiteley, J. Chem. Soc. Trans. 1907, 91, 1330–1350.
S. Naya, K. Yoda, M. Nitta, Tetrahedron 2005, 61, 8616–8624.
a) T. Bug, H. Mayr, J. Am. Chem. Soc. 2003, 125, 12980–12986;
b) H. Mayr, T. Bug, M. F. Gotta, N. Hering, B. Irrgang, B.
Janker, B. Kempf, R. Loos, A. R. Ofial, G. Remennikov, H.
Schimmel, J. Am. Chem. Soc. 2001, 123, 9500–9512.
k is the second-order rate constant for an electrophile–nucleo-
phile recombination in –1 s–1, E is the electrophilicity param-
eter, N is the nucleophilicity parameter, and s is a nucleophile-
specific slope parameter (s ≈ 1 can be assumed for such qualita-
tive considerations).
[25]
[26]
[27]
[7] R. W. Sinkeldam, Y. Tor, Org. Biomol. Chem. 2007, 5, 2523–
2528.
[8] F. Jara, C. Mascayano, M. C. Rezende, C. Tirapegui, A. Urzua,
J. Inclusion Phenom. Macrocycl. Chem. 2006, 54, 95–99.
[9] S. Spange, C. Schmidt, H. R. Kricheldorf, Langmuir 2001, 17,
856–865.
[10] L. P. Novaki, O. A. El Seoud, Ber. Bunsen-Ges. Phys. Chem.
1996, 100, 648–655.
[11] a) R. W. Taft, M. J. Kamlet, J. Chem. Soc. Perkin Trans. 2 1979,
1723–1727; b) M. J. Kamlet, T. H. Hall, J. Bodkin, R. W. Taft,
J. Org. Chem. 1979, 44, 2599–2604; c) M. J. Kamlet, J.-L. M.
Abboud, M. H. Abraham, R. W. Taft, J. Org. Chem. 1983, 48,
2877–2887.
[12] δ has only three values: 1.0 for aromatic, 0.5 for polyhalogen-
ated and 0.0 for aliphatic solvents.
[13] a) J. Catalán, V. López, P. Pérez, R. Martín-Villamil, J.-G. Rod-
riguez, Liebigs Ann. 1995, 241–252; b) J. Catalán, V. López, P.
Pérez, Liebigs Ann. 1995, 793–795; c) J. Catalán, C. Díaz, V.
López, P. Pérez, J.-L. G. DePaz, J.-G. Rodriguez, Liebigs Ann.
1996, 1785–1794; d) J. Catalán, C. Díaz, Liebigs Ann./Recueil
1997, 1941–1949; e) J. Catalán, C. Díaz, Eur. J. Org. Chem.
1999, 885–891.
[14] J. Catalán, H. Hopf, Eur. J. Org. Chem. 2004, 4694–4702.
[15] G. Sun, Y.-Q. Qiu, H. Z. Sun, Z.-M. Su, J.-D. Feng, Y. L. Zhu,
Chin. J. Chem. 2004, 22, 425–429.
[16] X. Zhou, A.-M. Ren, J. K. Feng, X. J. Liu, C. C. Shu, Chin. J.
Chem. 2004, 22, 38–46.
[17] a) R. Andreu, M. A. Cerdan, J. Orduna, ARKIVOC 2004, 32–
41; b) R. Andreu, M. J. Blesa, L. Carrasquer, J. Garín, J. Or-
duna, B. Villacampa, R. Alcalá, J. Casado, M. C.
Ruiz Delgado, J. T. López Navarrete, M. Allain, J. Am. Chem.
Soc. 2005, 127, 8835–8845.
[28]
[29]
[30]
[31]
[32]
[33]
K.-D. Nolte, S. Dähne, J. Prakt. Chem. 1976, 318, 643–657.
See, for example: a) Z. He, D. Gosztola, Y. Deng, G. Gao,
M. R. Wasielewski, L. D. Kispert, J. Phys. Chem. B 2000, 104,
6668–6673; b) D. Laage, P. Plaza, M. Blanchard-Desce, M. M.
Martin, Photochem. Photobiol. Sci. 2002, 1, 526–535.
Y. Marcus, Chem. Soc. Rev. 1993, 22, 409–416.
J. G. Kirkwood, J. Chem. Phys. 1934, 2, 351–361.
a) B. W. Domagalska, K. A. Wilk, S. Wysocki, Phys. Chem.
Chem. Phys. 2003, 5, 696–702; b) B. W. Domagalska, K. A.
Wilk, R. Zielinski, J. Photochem. Photobiol. A: Chem. 2006,
184, 193–203.
[34]
[35]
[36]
[18] G. Bourhill, J.-L. Brédas, L.-T. Cheng, S. R. Marder, F. Meyers,
J. W. Perry, B. G. Tiemann, J. Am. Chem. Soc. 1994, 116, 2619–
2620.
[19] M. Blanchard-Desce, V. Alain, P. V. Bedworth, S. R. Marder,
A. Fort, C. Runser, M. Barzoukas, S. Lebus, R. Wortmann,
Chem. Eur. J. 1997, 3, 1091–1104.
[37]
Origin 5.0, MicroCal Software Inc., Northampton, MA, USA,
1997.
[20] A. V. Kulinich, N. A. Derevyanko, A. A. Ishchenko, Russ. J.
Gen. Chem. (Engl. Trans.) 2006, 76, 1441–1457.
Received: April 7, 2008
Published Online: July 25, 2008
Eur. J. Org. Chem. 2008, 4475–4481
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
4481