Analysis of the solvatochromism of 9,9′-biaryl compounds using a pure solvent dipolarity scale

Article abstract of DOI:10.1021/jp211330x

The UV/vis absorption and fluorescence spectra of 9,9′-bisanthracenyl (BA), 9,9′- bisacridinyl (BAC), 9,9′-bicarbazyl (BC), and 9-(9′-anthracenyl)carbazole (C9A) were precisely recorded in solvents that change their solvent dipolarity scale (SdP) values from 0.000 (in 2-methylbutane at 293 K) to 1.294 (in 1-chlorobutane at 77 K). An analysis based (i) on the solvatochromic behavior of the four 9,9′-biaryl compounds in terms of the dipolarity of the solvents used, quantified by means of their SdP values, and (ii) on the behavior of the emission band profiles, allows the conclusion that, whereas the apolar 9,9′-biaryl compounds BA, BAC, and BC need solvents with SdP values larger than 0.74 to give a twisted intramolecular charge-transfer (TICT) process after excitation, for dipolar C9A the TICT mechanism is predominant already at solvent SdP values smaller than 0.23, i.e., near a solvent dipolarity of 0. The accuracy and simplicity achieved by application of the SdP scale in analyzing the solvatochromic behavior of the four 9,9′-biaryl compounds studied make evident the advantages of this solvent scale in comparison to other empirical solvent scales, which not only include the solvent's dipolarity but also consist of a more limited numerical range of values only.

Full text of DOI:10.1021/jp211330x

Article
pubs.acs.org/JPCA
Analysis of the Solvatochromism of 9,9′-Biaryl Compounds Using a
Pure Solvent Dipolarity Scale
́
Javier Catalan*
Departamento de Química Física Aplicada, Universidad Auton
́
oma de Madrid, 28049 Madrid, Spain
Christian Reichardt
Fachbereich Chemie, Philipps-Universitat
̈
, Hans-Meerwein-Strasse, 35032 Marburg, Germany
*
S Supporting Information
ABSTRACT: The UV/vis absorption and fluorescence spectra of 9,9′-
bisanthracenyl (BA), 9,9′- bisacridinyl (BAC), 9,9′-bicarbazyl (BC), and 9-(9′-
anthracenyl)carbazole (C9A) were precisely recorded in solvents that change
their solvent dipolarity scale (SdP) values from 0.000 (in 2-methylbutane at
2
93 K) to 1.294 (in 1-chlorobutane at 77 K). An analysis based (i) on the
solvatochromic behavior of the four 9,9′-biaryl compounds in terms of the
dipolarity of the solvents used, quantified by means of their SdP values, and
(
ii) on the behavior of the emission band profiles, allows the conclusion that,
whereas the apolar 9,9′-biaryl compounds BA, BAC, and BC need solvents
with SdP values larger than 0.74 to give a twisted intramolecular charge-
transfer (TICT) process after excitation, for dipolar C9A the TICT
mechanism is predominant already at solvent SdP values smaller than 0.23,
i.e., near a solvent dipolarity of 0. The accuracy and simplicity achieved by
application of the SdP scale in analyzing the solvatochromic behavior of the
four 9,9′-biaryl compounds studied make evident the advantages of this solvent scale in comparison to other empirical solvent
scales, which not only include the solvent’s dipolarity but also consist of a more limited numerical range of values only.
INTRODUCTION
The dual emission from solutions of certain chromophores,
observed when the solvent's dipolarity changes, has received
values into SdP (solvent dipolarity) and SP (solvent polar-
■
izability) values by means of a substituted polyene chromo-
17
phore (ttbP9).
1
−6
The solvatochromism of the four 9,9′-biaryl compounds
great interest in the last sixty years,
because the new
involved in the TICT process should exhibit a bilinear behavior
emission band generated was ascribed to a second electronic
excited state with a zwitterionic molecular structure exhibiting a
large charge separation, a process that is involved in many
important, naturally occurring photochemical mechanisms such
15
as indicated by Schneider and Lippert, i.e., with one range of
solvent dipolarity dominated by the locally excited (LE)
fluorescence emission, and another range of solvent dipolarity
showing the TICT emission.
7
as vision and photosynthesis. This type of process is also of
8
,9
In this work a wide solvent dipolarity scale will be employed
for the first time, ranging in the liquid phase from SdP = 0.000−
1.045, using for it pure 2-methylbutane and 1-chlorobutane as
limiting solvents. For instance, dipolar solvents such as
formamide, dimethyl sulfoxide, water, and N,N-dimethylforma-
mide have SdP values of 1.006, 1.000, 0.997, and 0.977,
respectively, whereas SdP values of apolar solvents such as 2-
potential interest for producing laser light, storing solar
9
10
energy, developing transistors, and organic conductors or
1
1
even superconductors.
In this connection two phenomena are of importance: (i) the
solvent-induced inverted solvatochromism of meropolymethine
dyes, discovered by Kiprianov et al. and Brooker et al.,
12
13
14
which was recently studied also by our research group, and
ii) the twisted intramolecular charge transfer (TICT) of so-
called tictoid compounds, discovered by Schneider and
1
7
methylbutane and n-hexane amount to 0.000. The TICT
(
̈
process is studied for three molecules with D symmetry (BA,
2d
1
5
16
BAC, and BC) and one with C symmetry (C9A) in the
Lippert. In this work, the TICT mechanism is studied
1
electronic ground state (Scheme 1). The first three molecules
exhibit a zero dipole moment and the fourth a nonzero dipole
with four representative 9,9′-biaryl compounds (Scheme 1), i.e.,
,9′-bisanthracenyl (BA), 9,9′-bicarbazyl (BC), 9,9′-bisacridinyl
9
(BAC), and 9-(9′-anthracenyl)carbazole (C9A) within the full
range of solvent dipolarity, using as its measure the empirical
solvent dipolarity (SdP) scale, obtained in 2009 by splitting the
formerly introduced solvent polarizability/dipolarity (SPP)
Received: November 24, 2011
Revised: April 18, 2012
Published: January 31, 2012
©
2012 American Chemical Society
4726
dx.doi.org/10.1021/jp211330x | J. Phys. Chem. A 2012, 116, 4726−4734

The Journal of Physical Chemistry A
Article
29
Scheme 1. Molecular Structures of the Four 9,9′-Biaryl
Compounds Studied
Rettig et al. studied the solvatochromism of the asymmetric
9,9′-biaryl compound C9A in diethyl ether, tetrahydrofuran, 1-
chlorobutane, 1,2-dichloroethane, n-butyronitrile, acetonitrile,
and they found a linear relationship in the correlation with the
corresponding Δf(ε ,n) values. On the other hand, our research
r
28
group found for C9A in 60 solvents also a linear relationship
with SPP, however, without the data obtained for perfluoro-n-
alkanes; although the standard deviation is quite large, a linear
relationship is clearly seen.
Based on the evidence from the solvatochromism of these
,9′-biaryl compounds, the conclusion could be drawn that an
9
adequate description of their solvatochromic behavior has not
yet been achieved. Their controversial behavior found for the
same range of solvents is still not well understood. For instance,
from 2-methylbutane to dimethyl sulfoxide as solvents, the
solvatochromism of BA yields a bilinear relationship, whereas it
is linear for BAC and BC; all these molecules exhibit D2d
symmetry and are apolar (i.e., without dipole moment).
If the two aryl rings of these molecules in the electronic
ground state are rotated out-of-plane to avoid steric hydrogen/
hydrogen repulsions at the peri-ring positions and the TICT
mechanism consists of rotating both rings toward a coplanar
structure, then the question can be raised, how is it feasible that
BC presents a peri-H/H repulsion that is significantly smaller
than the corresponding peri-H/H repulsion for BAC, which
does not exhibit TICT, or for BAC, which does exhibit TICT?
In our opinion, this controversy can be due to the lack of
sufficient solvatochromic data, which should depend only on
the solvent’s dipolarity. Therefore, the until now available
solvatochromic data seem to be contaminated with specific
solute/solvent interactions. This interpretation can possibly
explain the large standard deviations found in the correlations
studied. On the other hand, there is a lack of an adequate
solvent parameter which can describe the solvent’s dipolarity
only.
1
8
moment (estimated to μ ≈ 2 D ) in their electronic ground
state. The TICT mechanism assumes that these four molecules
adopt an almost orthogonal conformation with respect to the
two rings in the electronic ground state. On photoexcitation,
these molecules dissolved in not very polar solvents, exhibit
structured UV/vis absorption and emission spectra, with a
rotation of the two aryl rings toward a more coplanar
conformation in the excited state with increasing solvent
polarity, from which they emit fluorescence bands without
vibronic structure.
1
5,19
The pioneering work of Schneider and Lippert
the solvatochromism of 9,9′-bisanthracenyl (BA), measured in
0 solvents of different polarity, and the results were
deals with
In this work, both factors are minimized by using a pure
solvent dipolarity parameter, i.e., the recently introduced,
3
interpreted with the famous Lippert Δf(ε ,n) equation. This
r
30
solvatochromically derived solvent dipolarity scale SdP, and
by measuring SdP values larger than 0.39 by means of the pure
solvent 1-chlorobutane, and SdP values smaller than 0.37 by
means of binary 2-methylbutane/1-chlorobutane mixtures.
study shows that 20 of the 30 solvents exhibit a bilinear
behavior of the corresponding correlation lines, whereas for 9
solvents of lowest polarity a linear relationship with a slope of 0
was observed.
2
0
Later on, the same solvatochromic data for BA were drawn
against the solvent dipolarity/polarizability SPP values. The
relationship observed was also bilinear, except for 9 low-polar
solvents, which gave a correlation line with a nonzero slope
meeting the gas-phase value. On the other hand, Acree et al.
studied the behavior of BA in 45 solvents: for 42 solvents, for
which Lippert’s or Bilot-Kawski’s f(ε ,n) function were
available, they obtained such highly dispersive results that the
bilinear relationship suggested by Schneider and Lippert
was only observed clearly when 8 values were omitted. After
this work, our group analyzed the solvatochromism of BA,
measured in 60 solvents, by correlation with the SPP scale: a
bilinear relationship was observed, however, with a significantly
EXPERIMENTAL SECTION
■
The 1-chlorobutane used was of Chromasolv grade (purity
99.8%; bp 77−78 °C) with a moisture content equal to or less
than 0.001%, and 2-methylbutane was of Merck-Uvasol grade
(purity ≥99.5%; bp 30 °C). Samples of the four biaryl
compounds studied were synthesized following procedures
21
22
2
3
24
r
31−33
given in the literature:
9-(9′-anthracenyl)anthracene (9,9′-
1
5,19
bisanthracenyl, BA) was obtained by reduction of anthrone
31
with zinc; 9-(9′-carbazyl)carbazole (9,9′-bicarbazyl, BC) was
25
synthesized by oxidation of carbazole with potassium
32
permanganate and was purified by column chromatography
using n-hexane/trichloromethane (90:10% v/v) as eluent; 9-
(9′-acridinyl)acridine (9,9′-bisacridinyl, BAC) was prepared by
2
6,27
large standard deviation. Durocher et al.
studied the
33
solvatochromism of BC in a small number of solvents and they
concluded that, contrary to the behavior of BA, BC does not
show a TICT mechanism.
Thereafter, our group analyzed the solvatochromism of BC,
measured in 54 solvents, by correlation with the SPP values,
confirming the absence of a bilinear relationship. It is important
to point out that BAC, studied in 48 solvents, does not show
reduction of 9-chloroacridine with zinc and was purified by
column chromatography using n-hexane/ethyl acetate (50:50%
v/v) as eluent; 9-(9′-anthracenyl)-9H-carbazole (C9A) was
obtained by fusion of carbazole with 9-bromoanthracene at 300
25
33
°C for 2.5 h and was purified by column chromatography
using n-hexane/trichloromethane (60:40% v/v) as eluent. All
compounds were colorless crystals with melting points
corresponding to the literature values.
28
any bilinear response in the correlation with the SPP values.
4
727
dx.doi.org/10.1021/jp211330x | J. Phys. Chem. A 2012, 116, 4726−4734

The Journal of Physical Chemistry A
Article
−
1
The temperature of solutions of the four biaryl compounds
in the range 77−293 K were controlled with an Oxford
DN1704 cryostat, purged with dried nitrogen of 99.99% purity,
and equipped with an ITC4 controller interfaced to the
spectrometers. Sample temperatures in the range 293−343 K
were controlled within ±0.1 °C with a Fisons-Haake D8 GH
thermostat.
All UV/vis absorption spectra were recorded with a Cary-5
spectrophotometer at variable temperatures from 77 to 293 K,
using Suprasil quartz cells of 1 cm path length that were fixed to
the cryostat and to a cell holder thermostatted between 293 and
Table 1. Wavenumbers ν̃ (cm ) of the Maxima of the Long-
Wavelength UV/Vis Absorption Band of the Reference
Compounds ttbP9 (0 → 0 Component) and DMANF,
Measured in Five 2-Methylbutane (2-MB)/1-Chlorobutane
(1ClB) Mixtures at 293 K, and the Corresponding SdP
17
Values
mixtures 2-MB/1ClB
%, v/v)
ttbP9 (0 → 0)
DMANF
−1
−1
(
ν
̃
/cm
ν
̃
/cm
SdP values
9
8
7
6
5
0:10
0:20
0:30
0:40
0:50
21872
21825
21795
21745
21715
25171
25004
24734
24707
24472
0.093
0.144
0.228
0.283
0.368
343 K.
Corrected fluorescence spectra were obtained with a
calibrated Aminco-Bowman AB2 spectrofluorometer. The
sensitivity factors for the emission channel, which include not
only those depending on the detector but also those related to
the emission monochromator and the optical arrangement
Photophysics of the 9,9′-Biaryl Compounds. On the
UV/Vis Absorption and Emission of BA, BC, and BAC. Figure
1
brings into focus the UV/vis spectral pattern of these 9,9′-
(
channel emission included), were obtained by using the
correction kit FP-123 from SLM Instruments. This required
mounting a standard lamp in a channel at right angles to the
emission channel in an OL 254 M spectra irradiance lamp from
Optronic Laboratories. The lamp was operated at a constant
voltage supplied by an SP-270 power source. Its light output
was driven into an integrating sphere with a pinhole leading to
the fluorometer emission channel. The conversion factors thus
obtained allowed us to convert the measured spectra into
absolute spectra, which are independent of the instrument used.
RESULTS AND DISCUSSION
■
To complete the recently introduced solvent dipolarity scale
30
SdP for 1-chlorobutane, we start by using a solvent that is
structurally similar to that of 1-chlorobutane, i.e., 2-methyl-
butane with SdP = 0.000 (apolar), to SdP = 0.399 (dipolar), the
latter value corresponding to that of 1-chlorobutane at 343 K,
Figure 1. Emission spectra (λexc = 350 nm) of BA, dissolved in 2-
30
−6
the lowest SdP value measurable with this solvent. That is, we
use 2-methylbutane for SdP = 0.000 and binary 2-
methylbutane/1-chlorobutane mixtures at 293 K for SdP values
between 0.000 and 0.399. Thereafter, the photophysics of the
methylbutane (c = 2 × 10
mol/L), measured at solution
temperatures between 293 and 143 K.
biaryl compounds. It shows the emission band of BA, measured
in 2-methylbutane at different temperatures, with a vibronic
structure at room temperature, which increases significantly in
structure and intensity with decreasing temperature, particularly
the 0 → 0 component. It is important to point out that the
vibronic structure of the emission band of BA (Figure 3) is
significantly smaller than that exhibited by its UV/vis
absorption spectra, measured in 1-chlorobutane at different
temperatures (Figure 2).
This observation indicates that, on photoexcitation of BA, its
molecules relax in this solvent their molecular structure by
decreasing the interplanar angle between the two rings to such
an extent that fosters a significant loss of vibronic structure in
its emission spectrum, accompanied by a small band shift.
The corresponding emission spectra of BA, measured in 1-
chlorobutane at different temperatures (Figure 3), exhibit a
vibronic structure at room temperature, which diminishes with
decreasing temperature, leading eventually to an absence of
vibronic structure at 253 K. When the solution temperature
increases from room temperature (293 K) to 343 K, the
emission band gains in vibronic structure; the same observation
is made with solutions of BA in 2-methylbutane/1-
chlorobutane mixtures rich in 2-methylbutane.
9
,9′-biaryl compounds is analyzed between 293 and 143 K (a
temperature at which 1-chlorobutane is still a liquid), with
emphasis on the photophysics of BA. Subsequently, the
solvatochromism of the symmetrical biaryl compounds BA,
BC, and BAC is explained, as well as the photophysical
behavior of nonsymmetrical C9A. Finally, all data observed are
assessed in terms of the TICT mechanism.
Enlarging the SdP Scale for Completeness. By
employing only one dipolar solvent, i.e., 1-chlorobutane, at
different temperatures, we established a comprehensive solvent
dipolarity scale, ranging from SdP = 0.399 (1-chlorobutane at
30
3
43 K) up to SdP = 1.294 (1-chlorobutane at 77 K). For the
sake of completeness, we have now enlarged the SdP scale from
.000 (2-methylbutane; μ = 0 D) to 0.399 by adding dipolar 1-
0
30
chlorobutane (μ = 1.90 D at 20 °C in CCl4 ) to apolar 2-
methylbutane. Table 1 compiles the long-wavelength UV/vis
absorption maxima of 3,20-di-tert-butyl-2,2,21,21-tetramethyl-
5
,7,9,11,13,15,17,19-docosanonaene (ttbP9; 0 → 0 compo-
nent) and 2-(dimethylamino)-7-nitrofluorene (DMANF) for
the binary mixtures mentioned above as well as the
corresponding SdP values derived from the equations reported
in ref 17. As can be seen, it is sufficient to use a 50:50% (v/v) 2-
methylbutane/1-chlorobutane mixture for completing the SdP
scale adequately.
To sum up, we found two kinds of fluorescence behavior for
solutions of these three 9,9′-biaryl compounds: (i) one with an
4
728
dx.doi.org/10.1021/jp211330x | J. Phys. Chem. A 2012, 116, 4726−4734

The Journal of Physical Chemistry A
Article
Figure 4. UV absorption spectra of BC, dissolved in 1-chlorobutane (c
Figure 2. UV/vis absorption spectra of BA, dissolved in 1-
−6
−
6
≈ 10 mol/L), measured at solution temperatures between 293 and
chlorobutane (c = 2 × 10
mol/L), measured at solution
1
43 K.
temperatures between 293 and 143 K.
Figure 5. Emission spectra (λ = 290 nm) of BC, dissolved in 1-
chlorobutane (c ≈ 10 mol/L), measured at solution temperatures
between 293 and 143 K.
exc
Figure 3. Emission spectra (λexc = 350 nm) of BA, dissolved in 1-
−6
−
6
chlorobutane (c = 2 × 10
mol/L), measured at solution
temperatures between 293 and 143 K.
emission band possessing vibronic structure similar to that
shown by BA solutions in 2-methylbutane (Figure 1), which
can be ascribed to the relaxed molecular structure to be
generated by the Franck−Condon photoexcitation, and (ii)
another emission below 253 K, generated from a new molecular
structure formed in the excited state from the relaxed excited-
state structure, presenting a spectral band envelope without
vibronic structure (Figure 3).
The corresponding UV/vis excitation spectra for the
aforementioned emission spectra match the spectral behavior
shown by the respective UV/vis absorption spectra, either for
the 2-methylbutane or for the 1-chlorobutane solutions, which
exhibit bands with a significant vibronic structure.
Figures 4 and 5 show the UV absorption and emission
spectra of BC, measured in 1-chlorobutane at different
temperatures. The temperature-dependent spectral absorption
behavior of BC resembles that shown by solutions of BA
Figure 6. UV/vis absorption spectra of BAC, dissolved in 1-
chlorobutane (c ≈ 10 mol/L), measured at solution temperatures
between 293 and 143 K.
−
6
(
Figure 2). Therefore, the comments given for the BA spectra
are also valid for the BC spectra.
Figures 6 and 7 show the UV absorption and emission
spectra of BAC, measured in 1-chlorobutane at different
temperatures. Its spectral absorption behavior is equivalent to
that of BA and BC, but the emission behavior of BAC shows
large differences. For instance, the emission spectra of BAC,
measured from 343 to 143 K, do not exhibit vibronic structure
in 1-chlorobutane, as well as its emission spectra measured in 2-
methylbutane/1-chlorobutane mixtures and in pure 2-methyl-
butane. This unique loss of vibronic structure for BAC as
4
729
dx.doi.org/10.1021/jp211330x | J. Phys. Chem. A 2012, 116, 4726−4734

The Journal of Physical Chemistry A
Article
Figure 7. Emission spectra (λ = 345 nm) of BAC, dissolved in 1-
chlorobutane (c ≈ 10 mol/L), measured at solution temperatures
between 293 and 143 K.
exc
−6
Figure 9. UV/vis absorption spectra of C9A, dissolved in 1-
−6
chlorobutane (c ≈ 10 mol/L), measured at temperatures between
2
93 and 143 k.
compared to the significant vibronic structure found for the BA
and BC emission spectra, provides the following information:
(
i) the structural relaxation in the electronic excited state of
BAC is larger than that for BA and BC, which makes at first
sight no sense because all three compounds are symmetrical
9
,9′-biaryl compounds, whereas their absorption behavior is
similar with a significant vibronic structure in all three cases;
ii) the emission band of the acridine chromophore of BAC
(
might be much less structured than that of the other
chromophores anthracene and carbazole, and on photo-
excitation BAC gets relaxed, and their emission spectra do
not show vibronic structure.
To test the last proposition, the emission spectra of
anthracene, carbazole, and acridine were measured in 1-
chlorobutane at 293 K (Figure 8). It can easily be seen that
the emission spectrum of acridine exhibits a much less vibronic
structure than those of anthracene and carbazole.
Figure 10. Emission spectra (λ = 325 nm) of C9A, dissolved in 1-
exc
−6
chlorobutane (c ≈ 10 mol/L), measured at temperatures between
2
93 and 143 K.
Figure 2), and at shorter wavelengths the peaks corresponding
to the carbazole chromophore (see also Figure 4).
The emission spectra of C9A, measured in 1-chlorobutane
from 343 to 143 K, show no vibronic structure. However, for
solutions of C9A in 2-methylbutane and 2-methylbutane/1-
chlorobutane mixtures from 90:10 to 70:30% (v/v), the
emission spectra exhibit a vibronic structure (Figure 11). One
should keep in mind that the emission spectra of anthracene
and carbazole, in particular that of anthracene, show a neat
vibronic structure (Figure 8).
Solvatochromism of the 9,9′-Biaryl Compounds.
Figures 12−15 compile the solvatochromic results obtained
from the UV/vis absorption and emission spectra of BA, BC,
BAC, and C9A, dissolved in media with a wide range of solvent
Figure 8. Normalized emission spectra of anthracene (λexc = 325 nm),
1
7,30
carbazole (λexc = 280 nm), and acridine (λ = 325 nm), dissolved in
exc
dipolarity, that is, from SdP = 0.000 to 1.045.
results the following conclusions can be drawn:
i) The large standard deviations reported in previous
From these
6
1
-chlorobutane (c ≈ 10− mol/L), measured at 293 K.
(
1
5,25
studies
for analogous correlations of solvatochromic
On the UV/Vis Absorption and Emission of C9A. The UV/
data (obtained at 25 °C for a wide range of solvents
vis absorption and emission spectra of C9A, measured in 1-
chlorobutane at different temperatures, are shown in Figures 9
and 10. In Figure 9, at longer wavelengths one can see the
absorption peaks due to the anthracene chromophore (see also
characterized by means of the polarity function
1
5
Δf(ε ,n), and of the dipolarity/polarizability parameters
r
2
5
SPP ) with the SdP scale are absent in Figures 12−15.
The correlations are now statistically much improved.
4
730
dx.doi.org/10.1021/jp211330x | J. Phys. Chem. A 2012, 116, 4726−4734

The Journal of Physical Chemistry A
Article
Figure 11. Emission spectra (λexc = 335 nm) of C9A, dissolved in 2-
methylbutane (SdP = 0.000) and 2-methylbutane/1-chlorobutane
mixtures up to 70:30% (v/v) (SdP = 0.228), measured at 293 K.
Figure 14. Correlation between the absorption and emission
wavenumbers of BAC and the 1-chlorobutane dipolarity parameter
SdP.
3
0
Figure 12. Correlation between the absorption and emission
wavenumbers of BA and the 1-chlorobutane dipolarity parameter
SdP.
Figure 15. Correlation between the absorption and emission
wavenumbers of C9A and the 1-chlorobutane dipolarity parameter
SdP.
30
3
0
(
ii) The absorption and emission data values corresponding
to 2-methylbutane and its mixtures with 1-chlorobutane
are drawn with empty diamonds and empty squares to
demonstrate that they fit adequately into the remaining
data of the SdP scale.
30
(
iii) The data values derived from the absorption-band
maxima indicate that there is only one solvatochromic
pattern throughout the whole SdP scale applied. The
absorption-band maxima shift to the red (to lower
wavenumbers → bathochromic band shift) as the
solvent’s dipolarity increases, leading to correlation
lines with small negative slopes: −267 ± 4 for BA
(
Figure 12), −103 ± 3 for BC (Figure 13), −374 ± 3 for
BAC (Figure 14), and −349 ± 4 for C9A (Figure 15).
iv) The emission data allow the conclusion that, for low
solvent-dipolarity ranges where these molecules do not
undergo a TICT mechanism, the solvatochromic slopes
indicate that the bathochromic band shift is significantly
larger than that observed for the absorption data,
showing the following negative slopes: −770 ± 41 for
BA (Figure 12), −318 ± 33 for BC (Figure 13), and
(
−
435 ± 39 for BAC (Figure 14).
(
v) When at larger SdP values the TICT mechanism comes
into play, the bathochromic shift of the emission bands
increases significantly with the solvent’s dipolarity,
Figure 13. Correlation between the absorption and emission
wavenumbers of BC and the 1-chlorobutane dipolarity parameter
30
SdP.
4
731
dx.doi.org/10.1021/jp211330x | J. Phys. Chem. A 2012, 116, 4726−4734

The Journal of Physical Chemistry A
Article
−
1
leading to the following solvatochromic negative slopes:
with n = 19, r = 0.967, and sd =70 cm .
The TICT solvatochromism of BA, however, follows eq 2:
−
3332 ± 35 for BA (Figure 12), −1643 ± 58 for BC
Figure 13), −7300 ± 447 for BAC (Figure 14), and
2207 ± 48 for C9A (Figure 15).
(
−
̃
ν = −(2308 ± 964)SP − (6560 ± 848)SdP
b
−
+
(1211 ± 363)SA + (951 ± 361)SB
(29090 ± 980)
The data values given in (iv) and (v) foster the existence of
an excited-state relaxation process, implemented in the analysis
of the spectral informations observed, which involves a
significant loss of vibronic structure of the emission bands of
the four compounds studied, in which the basic chromophores
anthracene, carbazole, and acridine are coupled to give the
respective biaryl molecular structure. These results confirm the
(2)
−1
with n = 22, r = 0.925, and sd =302 cm .
Equations 1 and 2 allow us to discount those solute/solvent
interactions which are distinct from solute/solvent dipolarity
interactions. Therefore, the normal emission solvatochromism
for each solvent appears to be equal to (ν + 1782 × SP),
whereas the TICT solvatochromism yields (ν + 2308 × SP +
b
1211 × SA − 951 × SB), thus enabling us to plot these
solvatochromic variations against the corresponding SdP values
of these solvents, as shown in Figure 16. By comparing Figure
34,35
conclusions drawn from femtosecond pulse studies,
which
̃
a
̃
read as follows: “... in nonpolar solvents, the original mutually
orthogonal anthracenyl units relax by torsional motions to a
twist angle of about 70°, and the emitting state does not involve
appreciable CT character”.
The solvatochromic slopes m of the TICT process for the
three symmetrical 9,9′-biaryl compounds BA, BC, and BAC are
quite different from each other; however, they follow a linear
relationship with the difference of the redox potentials (ΔE =
ox
red
ox
E
D
− E , with E = oxidation potential of the electron
A
D
red
donor, and EA = reduction potential of the electron acceptor)
for the electron donor/acceptor chromophores involved (Table
3
±
in ref 28), with the following correlation equation: m = (4697
ox
red
456)·(ED − E ) − 19483 (with n = 3, r = 0.9953, and a
A
standard deviation of sd = 397). The unsymmetrical 9,9′-biaryl
compound C9A does not follow this equation because of its
dipolar character, which is essential for the TICT mechanism,
and that distinguishes its behavior from that of the other three
9
,9′-biaryl compounds which are apolar.
At this point it should be emphasized that the results reached
in this work on changing the temperature from 345 to 77 K of a
biaryl solution in 1-chlorobutane (with the corresponding
variation in medium dipolarity) are physically consistent with
those obtained for biaryl solutions in pure solvents of different
dipolarity at a constant temperature of 298 K. To prove this,
the solvatochromism of BA and C9A (both measured in a wide
Figure 16. Bilinear correlation between the emission wavenumbers of
BA, measured at 298 K and corrected as described in the text, and the
solvent dipolarity parameter SdP.
3
0
25
range of pure solvents at 298 K ) has been checked, taking
into account that BA shows a clear-cut bilinear behavior
1
2 (observed for 1-chlorobutane at different temperatues) with
Figure 16 (obtained for different solvents at 298 K), one can
draw the conclusion that both analyses are consistent, both
showing a bilinear solvatochromic behavior of BA as the solvent
dipolarity changes.
(
Figure 12) and, on the other hand, C9A a linear
solvatochromic behavior (Figure 15).
Among the 60 solvents employed to study the normal
solvatochromism of BA (the so-called a emission in ref 25), we
have chosen 19 solvents with solvent dipolarity values SdP
lower than 0.4 (i.e., perfluoro-n-hexane, 2-methylbutane, n-
pentane, n-hexane, n-heptane, cyclohexane, methylcyclohexane,
n-hexadecane, cis-decalin, triethylamine, p-xylene, o-xylene, tri-
n-butylamine, di-n-butyl ether, toluene, benzene, tetralin,
diethyl ether, and 1,4-dioxane). To describe the TICT
solvatochromism of BA (the so-called b emission in ref 25),
In addition, it should be mentioned that eqs 1 and 2 indicate
that (i) the polarizability influence on the emission from the
TICT excited structure of BA is significantly larger than that on
the emission from the normal excited structure of BA,
according to −2308 ± 964 (eq 2) vs 1782 ± 167 (eq 1),
respectively, and (ii) the dipolarity influence on the TICT
emission (at 298 K) is about 10 times larger than that of the
normal emission, according to −6560 ± 848 (eq 2) vs −651 ±
2
2 solvents with SdP values larger than 0.7 have been selected
1
39 (at 298 K; eq 1) or −770 ± 41 (by changing the
(
i.e., tert-butanol, 1-propanol, 2-propanol, ethanol, methanol,
temperature of a BA solution in 1-chlorobutane; Figure 12),
and about twice than that of the TICT emission recorded by
changing the temperature of a BA solution in 1-chlorobutane,
corresponding to −3332 ± 35 (Figure 12).
The solvatochromic analysis of the C9A emission, measured
in 55 solvents (i.e., 41 solvents as used for the analysis of the
BA solvatochromism plus fluorobenzene, trichloromethane,
ethyl acetate, 1-hexanol, 1-pentanol, 1-butanol, 1-propyl
formate, dibenzyl ether, anisole, chlorobenzene, methyl
benzoate, 1-chlorobutane, tetrahydrofuran, and 1,2-dichloro-
benzene; extracted from ref 25) leads to eq 3:
acetonitrile, propionitrile, n-butyronitrile, benzonitrile, di-
chloromethane, pyridine, propylene carbonate, N,N-diethylace-
tamide, ethylene glycol, N,N-diethylformamide, N-methylimi-
dazole, tetramethylurea, N,N-dimethylformamide, N,N-dime-
thylacetamide, γ-butyrolactone, sulfolane, and dimethyl
sulfoxide).
The normal solvatochromism of BA follows eq 1:
̃
ν = −(1782 ± 167)SP − (651 ± 139)SdP
a
+
(25354 ± 108)
(1)
4
732
dx.doi.org/10.1021/jp211330x | J. Phys. Chem. A 2012, 116, 4726−4734

The Journal of Physical Chemistry A
Article
ν
̃
= −(2370 ± 340)SP − (2393 ± 91)SdP
(25563 ± 236)
with n = 55, r = 0.974, and sd =224 cm .
+
(3)
−1
Equation 3 again allows one to subtract for C9A those
solute/solvent interactions that are distinct from solute/solvent
dipolarity interactions. As a result, the solvatochromism for
̃
each solvent corresponds to (ν + 2370 × SP); a plot of this
solvatochromism against the solvent dipolarity values SdP is
shown in Figure 17. As can be seen, the correlation lines of
Figures 15 and 17 show a consistent linear behavior on
changing the solvent dipolarity.
Figure 18. Emission spectrum (λ = 350 nm) of BA, dissolved in 1-
exc
−
6
chlorobutane (c ≈ 10 mol/L), measured at solution temperatures of
2
93, 143, and 123 K.
almost at the same wavelength positions as observed in the
emission spectrum of BA measured at 293 K. A careful analysis
of the emission spectra of Figure 18 allows the conclusion that,
by decreasing the temperature of the 1-chlorobutane solution of
BA from 293 to 143 K, the dipolarity of the medium increases
30
significantly from SdP = 0.560 to 1.045, and consequently, the
TICT process is now switched on. Although by decreasing the
sample temperature further from 143 to 123 K, the dipolarity of
the medium is also further increased from SdP = 1.045 to
30
1
.162, the TICT process is stopped, and the emission band is
hypsochromically shifted to the blue, now being located at the
peak-maxima positions of the emission spectrum measured at
293 K (Figure 18). In other words, the process occurring in the
excited state in the temperature range 293 to 143 K, yielding an
emission band without vibronic structure and a bathochromic
band shift, involves a distortion of the molecular structure
which is unreachable at 123 K, a temperature at which the
surrounding solvent matrix is so rigid that this distortion is
obviously hindered. The dramatic increase in vibronic structure
in the emission band of BA at 123 K, as compared to its
emission spectrum at 293 K, should be emphasized, which
supports the idea of a relaxation process experienced by BA in
its first electronic excited state before emitting fluorescence.
From the three emission spectra of BA shown in Figure 18,
that at 143 K can be taken as prototype for a TICT process,
that at 293 K as prototype for a locally excited (LE)
fluorescence emission, and that recorded at 123 K as pattern
for a Franck−Condon state emission.
Figure 17. Linear correlation between the emission wavenumbers of
C9A, measured at 298 K and corrected as described in the text, and
the solvent dipolarity parameter SdP.
30
Although solvent polarizability and dipolarity perturb
significantly the TICT emissions of BA and C9A, the TICT
processes generate, in addition, cavity effects that may differ for
solvents such as 1-chlorobutane as compared to solvents of
different molecular structure. These contributions may shift the
TICT emission to the red due to the instability caused by the
solvation of the Franck−Condon ground state as function of
the molecular structure of the solvents employed, which may
help to discriminate both solvatochromic treatments.
On the Inter-ring Torsional Nature of the TICT
Mechanism. The TICT mechanism was originally conceived
as a torsional process which triggers an electronic charge-
transfer between two chromophore fragments of the same
molecule in the excited state, which is modulated by the
polarity of the surrounding medium. However, frequently it was
tried to analyze these mechanisms as generated by an electronic
charge-transfer (CT) process between two chromophore
fragments of the molecule which are not involved in any
torsional rotation process.
Taking into account the evidence mentioned before, a
tentative photophysical diagram can be proposed for all four
9,9′-biaryl compounds studied in this paper (Figure 19).
CONCLUSIONS
The application of the solvent dipolarity scale SdP
■
17,30
is an
To shed some light on this point, the emission spectrum of
BA was measured in 1-chlorobutane at 293, 143, and 123 K
excellent tool for the study of solution processes depending on
the medium dipolarity, as demonstrated in this paper by the
analysis of the solvent- and temperature-dependent photo-
physical behavior of the four 9,9′-biaryl compounds BA, BC,
BAC, and C9A.
(Figure 18). The spectrum at 143 K does not exhibit any
vibronic structure and the emission band is significantly shifted
to the red as compared to the spectrum measured at 293 K.
Therefore, the emission spectrum measured at 143 K should
come from a TICT or CT state. However, by lowering the
sample temperature down to 123 K, the emission spectrum
shows now a significant vibronic structure and sharp peaks
The loss of vibronic structure in the emission bands of these
9,9′-biaryl compounds can be ascribed to a TICT process if
their emission bands exhibit a large bathochromic band shift.
This allows to distinguish between the loss of vibronic structure
4
733
dx.doi.org/10.1021/jp211330x | J. Phys. Chem. A 2012, 116, 4726−4734

The Journal of Physical Chemistry A
Article
(
14) Catalan
pig.2012.03.018.
15) Schneider, F.; Lippert, E. Ber. Bunsen-Ges. Phys. Chem. 1968, 72,
155−1160.
16) Grabowski, Z. R.; Rotkiewicz, K.; Siemiarczuk, A.; Cowley, D. J.;
Baumann, W. Nouv. J. Chim. 1979, 3, 443−454.
17) (a) Catalan, J.; Hopf, H. Eur. J. Org. Chem. 2004, 4694−4702.
b) Catalan, J. J. Phys. Chem. B 2009, 113, 5951−5960.
18) The compound closest in molecular structure to C9A for which
́
, J. Dyes and Pigments, DOI: 10.1016/j.dye-
(
1
(
(
́
(
́
(
the dipole moment is known is 9-methyl-9H-carbazole with μ = 1.99
D; see: Rogacheva, S. S.; Klepikova., S. G.; Lopatinskii, V. P.;
Sukhoroslova, M. M.; Sirotkina, E. E. Izvestiya Tomskogo Politekhni-
cheskogo Instituta 1974, 233, 66−70; Chem. Abstr. 1975, 82, 169553.
(
19) Schneider, F.; Lippert, E. Ber. Bunsen-Ges. Phys. Chem. 1970, 74,
24−630.
20) Catalan
21) Khundkar, L. R.; Zewail, A. H. J. Chem. Phys. 1986, 84, 1302−
311.
22) Tucker, S. A.; Griffin, J. M.; Acree, W. E.; Zander, M.; Mitchell,
R. H. Appl. Spectrosc. 1994, 48, 458−464.
23) (a) Lippert, E. Z. Naturforsch. A 1955, 10a, 541−545.
6
(
(
1
́ ́ ́
, J.; Lopez, V.; Perez, P. J. Fluoresc. 1996, 6, 15−22.
(
Figure 19. Tentative photophysical diagram for the angle dependence
of the S energy of the four 9,9′-biaryl compounds studied (angle =
1
(
interplanar angle between the two aryl rings).
(b) Lippert, E. Z. Phys. Chem. N. F. 1956, 6, 125−128. (c) Lippert,
E. Z. Elektrochem., Ber. Bunsen-Ges. Phys. Chem. 1957, 61, 962−975.
found for BAC: (i) as not caused by a TICT process below
solvent dipolarity values of SdP = 0.74 and (ii) due to a TICT
process for solvent dipolarity values of SdP > 0.74.
(24) Bilot, L.; Kawski, A. Z. Naturforsch. A 1962, 17a, 621−627.
(25) Catalan
Phys. Chem. 1996, 100, 18392−18398.
26) Harvey, P. D.; Zelent, B.; Durocher, G. Spectrosc.: Int. J. 1983, 2,
128−143.
27) Zelent, B.; Harvey, P. D.; Durocher, G. Can. J. Spectrosc. 1984,
9, 23−30.
28) Catalan
Chem. 1998, 1697−1704.
29) Rettig, W.; Zander, M. Ber. Bunsen-Ges. Phys. Chem. 1983, 87,
143−1149.
́
30) Catalan, J.; de Paz, J. L. G.; Reichardt, C. J. Phys. Chem. A 2010,
́ ́ ́
, J.; Dιaz, C.; Lopez, V.; Perez, P.; Claramunt, R. M. J.
́
(
ASSOCIATED CONTENT
Supporting Information
■
(
*
S
2
(
Table S1 with all solvent SdP values available at present. This
information is available free of charge via the Internet at http://
pubs.acs.org
́ ́
, J.; Dιaz, C.; Perez, P.; Claramunt, R. M. Eur. J. Org.
́
(
1
(
AUTHOR INFORMATION
Corresponding Author
E-mail: e-mail:javier.catalan@uam.es.
■
114, 6226−6234.
(31) Bell, F.; Waring, D. H. J. Chem. Soc. 1949, 267−269.
(32) Perkin, W. H.; Tucker, S. H. J. Chem. Soc. 1921, 216−225.
(33) Boyer, G.; Claramunt, R. M.; Elguero, J.; Fathalla, M.; Foces-
*
REFERENCES
■
Foces, C.; Jaime, C.; Llamas-Saiz, A. L. J. Chem. Soc., Perkin Trans. 2
993, 757−766.
34) Jurczok, M.; Plaza, P.; Martin, M. M.; Meyer, Y. H.; Rettig, W.
Chem. Phys. 2000, 253, 339−349.
35) Jurczok, M.; Plaza, P.; Rettig, W.; Martin, M. M. Chem. Phys.
000, 256, 137−148.
(
1) Lippert, E.; Rettig, W.; Bonici
A. Adv. Chem. Phys. 1987, 68, 1−173.
2) Rettig, W.; Baumann, W. In Photochemistry and Photophysics;
́
́ ́ ́
c-Koutecky, V.; Heisel, F.; Miehe, J.
1
(
(
Rabek, J. F., Ed.; CRC Press: Boca Raton, FL, USA, 1992; Vol. VI,
Chapter 3, pp 79−134.
(
2
(
(
3) Rettig, W. Top. Curr. Chem. 1994, 169, 253−299.
4) Zachariasse, K. A.; Grobys, M.; von der Haar, Th.; Hebecker, A.;
Il’ichev, Yu.; Jiang, Y.-B.; Morawski, O.; Ku
Photobiol. A 1996, 102, 59−70.
5) Grabowski, Z. R.; Rotkiewicz, K.; Rettig, W. Chem. Rev. 2003,
̈
hnle, W. J. Photochem.
(
1
(
(
03, 3899−4032.
6) Haidekker, M. A.; Theodorakis, E. A. J. Biol. Eng. 2010, 4, 11.
7) Lueck, H.; Windsor, M. W.; Rettig, W. J. Phys. Chem. 1990, 94,
4
(
550−4559.
8) Wiessner, A.; Hu
Chem. 1995, 99, 14923−14930.
9) Rettig, W. Angew. Chem. 1986, 98, 969−986. Rettig, W. Angew.
Chem., Int. Ed. Engl. 1986, 25, 971−988.
10) Nakaura, M.; Tizuka, M.; Kudo, K. Molecular Electronics and
̈ ̈
ttmann, G.; Kuhnle, W.; Staerk, H. J. Phys.
(
(
Bioelectronics 2001, 12, 51−56.
(
(
11) Saito, G.; Yoshida, Y. Chem. Record 2011, 11, 124−145.
12) (a) Kiprianov, A. I.; Petrun’kin, V. E. Zh. Obshch. Khim. (J. Gen.
Chem. USSR) 1940, 10, 613−619. (b) Kiprianov, A. I. Usp. Khim.
1
960, 29, 1336−1352. Kiprianov, A. I. Russ. Chem. Rev. 1960, 29, 618−
6
26. (c) Kiprianov, A. I. Cvet i Stroenie Cianinovych Krasitelej −
Isobrannye Trudy (Color and Structure of Cyanine Dyes − Collected
Works); Izdatel’stvo Naukova Dumka: Kiev, Ukraine, 1979; Chapter II
(
Solvatochromiya), pp 107ff.
13) Brooker, L. G. S.; Keyes, G. H.; Heseltine, D. W. J. Am. Chem.
Soc. 1951, 73, 5350−5358.
(
4
734
dx.doi.org/10.1021/jp211330x | J. Phys. Chem. A 2012, 116, 4726−4734