Photophysical properties of 3-[2-(N-phenylcarbazolyl)benzoxazol-5-yl] alanine derivatives-experimental and theoretical studies

Article abstract of DOI:10.1039/c2pp25114k

Solvatochromic probes are often used in biophysical studies to obtain information about polarity of the microenvironment. As there is not much natural fluorophores with such properties, there is still need for new synthetic compounds such as 3-(2-benzoxaz

Full text of DOI:10.1039/c2pp25114k

Photochemical &
Photobiological Sciences
View Article Online
View Journal | View Issue
PAPER
Photophysical properties of 3-[2-(N-phenylcarbazolyl)-
benzoxazol-5-yl]alanine derivatives – experimental and
theoretical studies†
Cite this: Photochem. Photobiol. Sci., 2013,
12, 284
Katarzyna Guzow,*a Marlena Czerwińska,a Agnieszka Ceszlak,a Marta Kozarzewska,a
Mariusz Szabelski,‡b Cezary Czaplewski,a Anna Łukaszewicz,b Aleksander A. Kubickib
and Wiesław Wiczka
Solvatochromic probes are often used in biophysical studies to obtain information about polarity of the
microenvironment. As there is not much natural fluorophores with such properties, there is still need for
new synthetic compounds such as 3-(2-benzoxazol-5-yl)alanine derivatives. Among this group of non-
proteinogenic fluorescent amino acids especially interesting are 3-[2-(4-aminophenyl)benzoxazol-5-yl]-
alanine derivatives whose solvatochromism depends on the substituents on the nitrogen atom, as
revealed by our recent studies. To expand them we synthesized two new derivatives with an N-phenylcar-
bazole moiety in position 2 of the benzoxazole ring and studied their photophysical properties in sol-
vents of different polarity and ability to form hydrogen bonds using absorption and steady-state and
time-resolved fluorescence spectroscopy. Applying single parameter and multi-linear correlations with
different solvent parameters, the excited state dipole moments were determined as well as the influence
of solvent parameters on each photophysical property was estimated. Moreover, the geometry of com-
pounds and vertical absorption transition were theoretically calculated (DFT and TD DFT methods). It was
found that the place of substitution of the N-phenylcarbazole part by the benzoxazole unit determines
the character of the electron transition (π–π* or ICT) and thereby the spectral and photophysical proper-
ties of the compounds studied.
Received 23rd April 2012,
Accepted 3rd September 2012
DOI: 10.1039/c2pp25114k
www.rsc.org/pps
application of non-proteinogenic amino acids such as 3-(2-ben-
zoxazol-5-yl)alanine derivatives. Those compounds are fluo-
1 Introduction
The solvatochromic effect is frequently used in biophysical rescent amino acids whose spectral and photophysical
studies of the polarity of the microenvironment of peptides, properties are often sensitive to the ambient medium. More-
proteins and lipid bilayers using intrinsic or extrinsic fluor- over, their solvatochromic effect depends on the substituent in
escent probes.1 Among the proteinogenic aromatic amino position 2.4–12 In the case of 3-[2-(phenyl)benzoxazol-5-yl]-
acids only tryptophan shows an evident solvatochromic effect alanine derivatives, the presence of a substituted amino group
and could be used as such a probe.1–3 An alternative is in position 4 of the phenyl is mainly responsible for their sol-
vatochromism.5,9,10,12 How big is this effect depends on the
amount and type (aliphatic or aromatic) of substituents on
the amino group.9,10,12 This is mainly due to the fact that the
aFaculty of Chemistry, University of Gdańsk, Sobieskiego 18, 80-952 Gdańsk, Poland.
nature of the substituent on the amino group determines the
strength of electron-transfer (dependent on the ionization
potential) and π-donor ability (coupling of electrons with the
residual part of a molecule).13
Taking these facts into account we decided to broaden our
previous studies on solvatochromism of 3-[2-(4-aminophenyl)-
benzoxazol-5-yl]alanine derivatives9,12 synthesizing such
derivatives with N-phenylcarbazole substituents and analyzing
their spectral and photophysical properties. Such studies allow
us to answer the question of how stiffening of substituents on
the amino group affects the solvatochromic effect of that
E-mail: kasiag@chem.univ.gda.pl; Fax: (+48)-(58)-5235472; Tel: (+48)-(58)-5235413
bInstitute of Experimental Physics, Faculty of Mathematics, Physics and Informatics,
University of Gdańsk, Wita Stwosza 57, 80-952 Gdańsk, Poland
†Electronic supplementary information (ESI) available: Emission spectra of exci-
plex of compound 2 in dimethoxyethane, results of the single-parameter linear
N
correlations of photophysical parameters of 1 and 2 with ET as well as their
multi-parameter linear correlations with three-parameter Catalán and Kamlet–
Taft solvent scales, theoretically calculated absorption wavelengths for both com-
pounds studied, their HOMO and LUMO orbitals and orientation of the tran-
sition dipole moments as well as the results of fluorescence anisotropy
measurements of stretched samples in PVA films. See DOI: 10.1039/c2pp25114k
‡Current address (M. Sz.): Department of Physics and Biophysics, University of
Warmia and Mazury in Olsztyn, Oczapowskiego 4, 10-719 Olsztyn, Poland.
284 | Photochem. Photobiol. Sci., 2013, 12, 284–297
This journal is © The Royal Society of Chemistry and Owner Societies 2013

View Article Online
Photochemical & Photobiological Sciences
Paper
compound. Choosing N-phenylcarbazole as a substituent ether 1 : 15 (v/v) as an eluent. A pale yellow solid was obtained
results from its important role in the design of compounds for as a product (0.78 g, 2.88 mmol, 96% yield) which was ident-
optoelectronics (OLED) and the presence of such a moiety in ified by means of 1H NMR (Varian, Mercury 400BB spec-
fluorescent dyes or photosensitive dyes in solar cells14,15 as trometer (400 MHz) in CDCl3), IR spectra (Bruker IFS-66
well as biologically active compounds.16
instrument) and mass spectra (Bruker Biflex III (MALDI-TOF)).
IR (KBr): νmax/cm−1 3049.0 (ArH), 1702.8 (CvO),
1478.9–1510.8 (CvC); 1H NMR: δH (ppm) 7.31–7.49 (6H, m,
C7′H, C6′H, C3′H, C2′H, C5′′H, C3′′H); 7.80 (2H, d, J = 8.0 Hz,
C2′′H, C6′′H); 8.13–8.16 (4H, m, C8′H, C5′H, C4′H, C1′H); 10.12
(1H, s, CHO); MS: m/z = 271.1 (M+).
2 Experimental
2.1 Synthesis
N-Boc-3-[2-(4-(9H-carbazolyl)phenyl)benzoxazol-5-yl]alanine
methyl ester (1) and N-Boc-3-[2-(3-(9-phenyl-9H-carbazolyl))-
benzoxazol-5-yl]alanine methyl ester (2) (Fig. 1) were syn-
thesized according to the procedure published previously4,6–12
based on the oxidative cyclization of the Schiff base obtained
from N-Boc-3-aminotyrosine methyl ester and the appropriate
aldehyde. Moreover, 4-(9H-carbazolyl)benzaldehyde and
9-phenyl-9H-carbazole-3-carbaldehyde were not commercially
available. Because of that the former was obtained by alkyl-
ation of carbazole with 4-bromobenzaldehyde17 whereas the
latter by formylation of N-phenylcarbazole with 1,1-dichloro-
methyl methyl ether and TiCl4.18
2.1.2 9-Phenyl-9H-carbazole-3-carbaldehyde. To a magneti-
cally stirred solution of N-phenylcarbazole (0.5 g, 2.1 mmol) in
1,2-dichlorobenzene (20 ml) cooled in an ice bath 1,1-dichloro-
methyl methyl ether (0.24 ml, 2.66 mmol) and TiCl4 (0.36 ml,
3.28 mmol) were added. The reaction was continued for 1 h in
the ice bath. Then the reaction mixture was allowed to warm
up to the room temperature and poured on ice (200 g) and
concentrated HCl (5 ml). The organic layer was diluted with
dichloromethane, washed with a 5% solution of HCl and water
and dried with anhydrous MgSO4. After evaporation the
residue was purified by column chromatography (Merck, Silica
gel 60, 0.040–0.063 mm) using a mixture of ethyl acetate–
petroleum ether 1 : 15 (v/v) as an eluent, giving the desired
compound as a white solid (0.27 g, 1.03 mmol, 49% yield). The
product was identified by means of 1H NMR (Varian, Unity
plus spectrometer (500 MHz) in CDCl3), IR spectra (Bruker
IFS-66 instrument) and mass spectra (Bruker Biflex III
(MALDI-TOF)).
2.1.1 4-(9H-Carbazolyl)benzaldehyde. To
a magnetically
stirred solution of carbazole (0.50 g, 2.99 mmol) in 1,2-dichloro-
benzene (12 ml) 4-bromobenzaldehyde (1.80 g, 9.73 mmol),
K2CO3 (0.99 g, 7.18 mmol), Cu (0.08 g, 1.41 mmol) and 18-
crown-6 (0.16 g, 0.61 mmol) were added. The mixture was
degassed with argon for about 30 min and then refluxed for
about 3 days. After that time, the reaction mixture was cooled
down, filtered and evaporated giving a brown oil which was
treated with petroleum ether and left overnight with constant
magnetic stirring to elute impurities. The ether layer was
rejected due to the lack of product and the residue was puri-
fied by column chromatography (Merck, Silica gel 60,
0.040–0.063 mm) using a mixture of ethyl acetate–petroleum
IR (KBr): νmax/cm−1 3060.8 (ArH), 1686.1 (CvO),
1436.7–1623.0 (CvC); 1H NMR: δH (ppm) 7.39–7.52 (7H, m,
C4′′H, C6′′H, C3′′H, C2′′H, C5′′H, C1′H, C8′H), 7.67 (2H, t, J =
7.82 Hz, C6′H, C7′H), 7.98 (1H, d, J = 8.06 Hz, C2′H), 8.24 (1H,
d, J = 7.81 Hz, C5′H), 8.70 (1H, s, C4′H), 10.15 (1H, s, CHO);
MS: m/z = 271.1 (M+).
2.1.3 3-(2-Benzoxazol-5-yl)alanine derivatives – general pro-
cedure. A mixture of N-Boc-3-nitro-L-tyrosine methyl ester and
10% palladium on active carbon in MeOH was stirred under a
hydrogen atmosphere at room temperature for about 90 min
(TLC monitoring, Merck silica-gel plates (Kieselgel 60 F254),
(CH2Cl2–MeOH–AcOH 100 : 10 : 1), Rf = 0.9 (N-Boc-3-nitro-L-
tyrosine methyl ester), Rf = 0.72 (N-Boc-3-amino-L-tyrosine
methyl ester)). The catalyst was filtered off and the solvent
evaporated to give a brownish oily product which was dissolved
in anhydrous EtOH and mixed with the solution of the appro-
priate aldehyde (1 equiv.) in anhydrous EtOH. The mixture was
stirred at RT overnight (TLC monitoring (AcOEt–petroleum
ether 1 : 5)). After this time, the solvent was removed by evapor-
ation and the Schiff base obtained (yellow solid) was dissolved
in DMSO and lead tetraacetate (1.5 equiv.) was added. The
mixture was stirred at RT for about 2 h (TLC monitoring
(AcOEt–petroleum ether 1 : 3)) and then dissolved in AcOEt
and washed in turn with a saturated aqueous solution of NaCl
(1×), a 5% solution of NaHCO3 (2×), a saturated aqueous solu-
tion of NaCl (3×), and dried with anhydrous MgSO4. The
solvent was evaporated and the product was isolated by
means of column chromatography (Merck, Silica gel 60,
Fig. 1 Structures of compounds studied.
This journal is © The Royal Society of Chemistry and Owner Societies 2013
Photochem. Photobiol. Sci., 2013, 12, 284–297 | 285

View Article Online
Paper
Photochemical & Photobiological Sciences
0.040–0.063 mm) using a mixture of AcOEt–petroleum ether monochromator (about 12 nm spectral band-width). Fluor-
1 : 5 (v/v) (compound 1) or AcOEt–petroleum ether 1 : 3 (v/v) escence decay data were fitted by the iterative convolution to
(compound 2) as an eluent. The crude products were recrystal- the sum of exponents according to eqn (1):
lized from a mixture of AcOEt–petroleum ether. Compound 1
X
IðtÞ ¼
αi expðꢀt=τiÞ
ð1Þ
was obtained as a white solid with 44% yield. In the case of
compound 2, the additional purification by means of
RP-HPLC was necessary (gradient 50–100% B over 120 min
(A = 0.1% water solution of trifluoroacetic acid, B = 80% of
acetonitrile in A), Kromasil column, C-8, 5 μm, 250 mm long,
i.d. = 20 mm). It was obtained as a white solid with 10% yield.
The purity of the obtained compounds was checked by means
of analytical RP-HPLC (compound 1 – Phenomenex® Jupiter
column (C18, 5 μm, 250 mm long, i.d. = 4.5 mm), compound 2
– Kromasil column (C-8, 5 μm, 250 mm long, i.d. = 4.5 mm))
with detection at λ = 223 nm. The mobile phase was a gradient
running from 0 to 100% of B over 60 min plus 100% of B over
10 min (tR = 67.4 min (1), tR = 66.8 min (2)). The identification
of the products was based on the 1H NMR (Varian, Mercury
400 BB (400 MHz) (compound 1) or Unity 500 plus spec-
trometer (500 MHz) (compound 2)) in CDCl3 and mass spectra
(Bruker Biflex III (MALDI-TOF)).
i
where αi is the pre-exponential factor obtained from the fluo-
rescence intensity decay analysis and τi the decay time of the
i-th component, using a software supported by the manufac-
turer. The adequacy of the exponential decay fitting was
judged by visual inspection of the plots of weighted residuals
2
as well as by the statistical parameter XR and shape of the
autocorrelation function of the weighted residuals and serial
variance ratio (SVR). For compound 2 in dimethoxyethane
additional measurements were made with a picosecond laser
excitation using a streak-camera as a detector.19
The multiple linear regression, approach of Kamlet and co-
workers,20–23 and Catalán and co-workers,24–27 has been used
to correlate UV–Vis absorption and emission energies with an
index of the solvent dipolarity/polarizability which is
a
measure of the solvent’s ability to stabilize a charge or dipole
through nonspecific dielectric interactions (π* or SPP), and
indices of the solvent’s hydrogen-bond donor strength (α or
SA) and hydrogen-bond acceptor strength (β or SB) according
to eqn (2) and (3):
Compound 1: 1H NMR: δH (ppm) 1.44 (9H, s, (CH3)3),
3.19–3.31 (2H, m, CβH2), 3.76 (3H, s, OCH3), 4.67 (1H, d, J =
8.0 Hz, CαH), 5.06 (1H, d, J = 8.0 Hz, NH), 7.16–7.19 (1H, m,
C6H), 7.31–7.34 (2H, m, C3′H, C6′H), 7.42–7.46 (2H, m, C2′H,
C7′H), 7.50–7.54 (4H, m, C7H, C4H, C1′H, C8′H), 7.76–7.78 (2H,
m, C3′′H, C5′′H), 8.15–8.17 (2H, m, C2′′H, C6′′H), 8.47–8.49 (2H,
m, C4′H, C5′H); MS: m/z 562.3 (MH+).
vA;F ¼ v0 þ aπꢁ ꢂ π ꢁ þbα ꢂ α þ cβ ꢂ β Kamlet–Taft equation
ð2Þ
Compound 2: 1H NMR: δH (ppm) 1.46 (9H, s, (CH3)3),
3.22–3.31 (2H, m, CβH2), 3.77 (3H, s, OCH3), 4.68 (1H, q, J = or
5.9 Hz, J = 13.2 Hz, CαH), 5.07 (1H, d, J = 8.3 Hz, NH), 7.14
(1H, dd, J = 2.0 Hz, J = 8.3 Hz, C8′H), 7.38 (1H, t, J = 7.8 Hz,
C6H), 7.44 (1H, d, J = 7.8 Hz, C7H), 7.48–7.56 (6H, m, C3′′H,
C4′′H, C5′′H, C6′H, C7′H, C4H), 7.61 (2H, dd, J = 1.5 Hz, J =
7.3 Hz, C2′′H, C6′′H), 7.67 (1H, t, J = 7.3 Hz, C4′H), 8.28 (1H, d,
J = 7.3 Hz, C2′H), 8.32 (1H, dd, J = 1.5 Hz, J = 8.8 Hz, C5′H), 9.08
(1H, d, J = 1.5 Hz, C1′H); MS: m/z 561.1 (M+).
´
vA;F ¼ v0 þ aSPP ꢂ SPP þ bSA ꢂ SA þ cSB ꢂ SB Catalan equation
ð3Þ
Moreover, a similar correlation was performed using Cata-
lán’s new generalized solvent polarity empirical scale splitting
the SPP parameters for two components: SP – solvent polariz-
ability and SdP – solvent dipolarity. Mathematically, this
scheme is formulated as follows:
2.2 Spectroscopic measurements
The absorption spectra of 1 and 2 in all solvents studied (22
for 1 and 36 for 2) were measured with a Perkin-Elmer
Lambda 40P spectrophotometer, whereas emission spectra
were measured using a Perkin-Elmer LS 50B or a Horiba Jobin
Yvon Fluoromax-4 spectrofluorimeter. Solvents of the highest
available quality (spectroscopic or HPLC grade) were used.
Fluorescence quantum yields (φF) were calculated with
quinine sulphate in 0.5 M H2SO4 (φF = 0.53 0.02) as a refer-
ence and were corrected for different refractive indices of sol-
vents.1 In all fluorimetric measurements, the optical density of
solution does not exceed 0.1.
The fluorescence lifetimes were measured with a time-corre-
lated single-photon counting apparatus Edinburgh CD-900.
The excitation source was a NanoLed N16 (UV led 339 nm)
from IBH. The half-width of the response function of the
apparatus was about 1.0 ns using a Ludox solution as a scatter.
y ¼ y0 þ aSP ꢂ SP þ bSdP ꢂ SdP þ cSA ꢂ SA þ dSB ꢂ SB
ð4Þ
where y denotes a solvent-dependent physicochemical property
in a given solvent and y0 the statistical quantity corresponding
to the value of the property in the gas phase; SP, SdP, SA, SB
represent independent solvent parameters accounting for
various types of solute–solvent interactions; aSP, bSdP, cSA and
dSB are adjustable coefficients that reflect the sensitivity of
physical property y in a given solvent to the various solvent
parameters.28
Based on spectroscopic data the excited state dipole
moment of molecules studied was determined using eqn (5)
and (6) from the Kawski’s theory:29–32
˜υa ꢀ ˜υf ¼ m1f ðεr; nÞ þ const
ð5Þ
ð6Þ
˜υa þ ˜υf ¼ ꢀm2½f ðεr; nÞ þ 2gðnÞꢃ þ const
The emission wavelengths were isolated using
a
286 | Photochem. Photobiol. Sci., 2013, 12, 284–297
This journal is © The Royal Society of Chemistry and Owner Societies 2013

View Article Online
Photochemical & Photobiological Sciences
Paper
where ˜υa and ˜υf are the absorption and emission band functions of the fluorophore in
a
partly ordered PVA
maximum (in wavenumbers), respectively, measured in sol- matrix,33–36,38–40 ten samples of different stretching factors
vents of different relative permittivities εr and different refrac- were prepared using a standard procedure and equipment
tive indices n. The solvent polarity functions f(εr, n) and g(n) developed earlier.37,38
are given by eqn (7) and (8):
2.3 Quantum chemical calculations
ꢀ
ꢁ
2n2 þ 1 ε ꢀ 1 n2 ꢀ 1
f ðεr; nÞ ¼
ꢀ
ð7Þ
ð8Þ
The structures of all calculated compounds were prepared and
initially optimized using the Avogadro program applying the
UFF force field method.41 All calculations were performed at
the DFT level with the hybrid functional pbe0 using TURBO-
MOLE package (version 5.9). The geometry optimization was
carried out without symmetry constraints, using a def2-TZVP
basis set. This procedure was considered satisfactory if the
energy difference between optimized cycles was <1 × 10−6
Hartree and a gradient of <1 × 10−3 au was achieved. The low-
lying excited states were treated within the adiabatic approxi-
mation of time-dependent density functional theory (DFT-RPA)
using a pbe0 functional. The geometry optimization of excited
states was performed using a grid parameter equal to 4. The
convergence of all systems studied was checked by harmonic
vibrational analysis. No imaginary frequencies were observed.
In addition, TD DFT calculation was carried out at the geome-
try of the ground state for the first 20 transitions in order to
simulate the absorption spectrum in the Franck–Condon
region. In order to take into account the influence of the
solvent effects on the geometry and spectral shifts of UV–vis
spectra, all calculations were carried out using the COSMO
model implemented in the TURBOMOLE program. The calcu-
lated absorption energies and intensities were transformed
with the GaussSum program into simulated spectra using
Gaussian functions assuming a constant bandwidth at the
half-height of 4000 cm−1. This value constitutes an average
width for an absorption band observed in the UV–Vis range.
n2 þ 2 ε þ 2 2n2 þ 1
3ðn4 ꢀ 1Þ
gðnÞ ¼
2
2ðn2 þ 2Þ
Using a slope obtained from eqn (5) and (6) and knowing
the ground state dipole moment, the excited state dipole
moment can be calculated using eqn (9) and (10) as well as an
angle ψ between them (eqn (11) and (12)):
2
ðμ2e þ μg2 ꢀ 2μeμg cos ΨÞ
2πε0hca3
ð~μe ꢀ~μgÞ
2πε0hca3
m1 ¼
¼
ð9Þ
ðμ2e ꢀ μ2gÞ
m2 ¼
ð10Þ
2πε0hca3
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
1
2
μe ¼ ðμ2g þ m2hca3Þ
ð11Þ
ð12Þ
ꢂ
ꢃ
1
m1
m2
cos Ψ ¼
ðμ2g þ μ2eÞ ꢀ
ðμ2e ꢀ μ2gÞ
2μeμg
where μe and μg are the dipole moments in the excited and
ground states, respectively, a is Onsager’s interaction radius of
the solute, h is the Planck’s constant (6.626 × 10−34 Js), c is the
velocity of light in a vacuum (3 × 108 ms−1), ε0 is the permittiv-
ity of vacuum (8.85 × 10−12 V C−1
m
−1), thus 2πε0hc =
1.105110440 × 10−35 C2. If the ground and excited state dipole
moments are parallel, based on eqn (9) and (10), eqn (13)–(15)
can be derived:
sffiffiffiffiffiffiffiffiffi
m2 ꢀ m1 hca3
3 Results and discussion
μg ¼
μe ¼
ð13Þ
ð14Þ
2
2m1
Both benzoxazol-5-ylalanine derivatives studied (Fig. 1)
possess as a substituent in position 2 an N-phenylcarbazole
subunit, but differently connected to the benzoxazole moiety –
for N-Boc-3-[2-(4-(9H-carbazolyl)phenyl)benzoxazol-5-yl]alanine
methyl ester (1) by phenyl whereas for N-Boc-3-[2-(3-(9-phenyl-
9H-carbazolyl))benzoxazol-5-yl]alanine methyl ester (2) by car-
bazole. This is probably a reason for observed significant
differences between the compounds studied in their spectral
sffiffiffiffiffiffiffiffiffi
m2 þ m1 hca3
2
2m1
or
m1 þ m2
μe ¼ μ
for m2 > m1
ð15Þ
g m2 ꢀ m1
The angles between absorption and fluorescence transition and photophysical properties.
moments of the compounds studied were calculated using the The absorption spectra of both compounds studied in
Kawski–Gryczyński method33–37 based on fluorescence polariz- selected solvents are shown in Fig. 2, whereas in Fig. 3 their
ation measurements (λex = 330 nm) in stretched poly(vinyl emission spectra are presented. For both compounds the
alcohol) (PVA) films. Each compound was dissolved in a 5% absorption band with a maximum at about 340 nm overlaps
water–methanol mixture of PVA (M ∼ 205 000) at 340 K to with the absorption band assigned to the S2 ← S0 and/or
obtain a homogeneous solution. The samples were left for S3 ← S0 transition of phenylcarbazole.42,43 This band is shifted
several days for evaporation to obtain thin PVA films which to about 305 nm for compound 2 and to about 290 nm for 1,
were then uniaxially stretched. As mechanical stretching gives indicating a bigger overlap with the main absorption band for
the control over the degree of orientation of transition dipole 2. A more pronounced vibrational structure is observed for 1 in
moments through the correlation with respective distribution saturated hydrocarbon solvents. In more polar solvents, the
This journal is © The Royal Society of Chemistry and Owner Societies 2013
Photochem. Photobiol. Sci., 2013, 12, 284–297 | 287

View Article Online
Paper
Photochemical & Photobiological Sciences
Fig. 2 Normalized absorption spectra of 1 (at the top) and 2 (at the bottom)
Fig. 3 Normalized emission spectra of 1 (at the top) and 2 (at the bottom) in
in selected solvents.
selected solvents.
absorption spectra of both compounds shift. Moreover, the
band’s shape changes and its vibrational structure becomes
diffuse. All these changes are more distinct for compound 2. A
bigger shift of the absorption band indicates that the change
of the dipole moment between the ground and Franck–
Condon states is larger for 2 than for 1.
The influence of the solvent is greater in the case of fluor-
escence spectra of both compounds (Fig. 3). However, the
emission maxima of compound 2 are shifted to the red to a
lesser degree than those of compound 1 indicating that 1 has
a larger dipole moment in the excited state than 2. Also, the
vibrational structure of fluorescence spectra is more pro-
nounced for 1 than for 2 as the structure of 1 is probably more
stiffened. It is a result of charge transfer between the substitu-
ent (N-phenylcarbazole) and the benzoxazole moiety causing
close to the fluorescence lifetime of N-phenylcarbazole in air-
saturated solvents,42,43 are observed in polar and protic sol-
vents (6.17 ns in ethylene glycol, 5.45 ns in DMSO), whereas in
hydrocarbon solvents they are shorter (1.30 ns in isooctane).
For 2 the fluorescence quantum yields are also high and close
to unity in polar and protic solvents and only a little lower in
saturated hydrocarbon solvents (Table 2). The fluorescence
lifetimes of 2 are much shorter (from 1.3 ns in chloroform to
2.7 ns in n-pentane) than that for 3-substituted carbazole.44
Both fluorescence quantum yields and fluorescence lifetimes
of 2 only slightly depend on solvent polarity. In most solvents
mono-exponential fluorescence intensity decays are observed
for both derivatives. For 1, multi-exponential fluorescence
intensity decays are observed in THF, 2-methyl-THF, diethyl
ether, n-hexanol, n-heptanol and propylene glycol, whereas for
2 only in dimethoxyethane. In the case of 1, the presence of
additional fluorescence lifetime (shorter one) is caused by
specific interaction between the solute and these solvents. For
aprotic solvents, this component may be connected with for-
mation of a complex with solvent possessing lone electron
pairs at the oxygen atom (charge-transfer type complexes),
whereas for protic ones it is probably a result of specific sol-
vation of 1 by alcohols with suitably long hydrocarbon chains.
In the case of 2, its fluorescence spectrum in dimethoxyethane
possesses a long-wavelength shoulder, probably connected
with formation of an excited state complex with the solvent
increase of the electron density on Cbenzoxazole–CPh and CPh
–
Ncarbazole bonds which become multiple bonds instead of
single. This indicates that the electron-donating property of
the phenylcarbazole substituent depends on the position of
substitution.
The photophysical properties of compounds 1 and 2 are
compiled in Tables 1 and 2, respectively. As shown in Table 1,
the fluorescence quantum yield of 1 is high (close to unity in
cyclohexane) to moderate (0.36 in methanol) and depends on
the solvent polarity. A similar dependence is observed for the
fluorescence lifetimes of 1. The longest fluorescence lifetimes,
288 | Photochem. Photobiol. Sci., 2013, 12, 284–297
This journal is © The Royal Society of Chemistry and Owner Societies 2013

View Article Online
Photochemical & Photobiological Sciences
Paper
Table 1 Spectroscopic and photophysical properties of 1 in all solvents studied
˜ν abs [cm−1
]
˜ν fluo [cm−1
]
φF
τ [ns]
α
χR
Δ˜ν [cm−1
]
kf/108 [s−1
]
knr/107 [s−1
]
2
Solvent
Methanol
Ethanol
1-Butanol
tert-Buthyl alcohol
iso-Propyl alcohol
Amyl alcohol
n-Hexanol
29 438
29 370
29 257
29 257
29 370
29 343
29 189
21 460
22 472
23 256
23 753
23 202
24 450
23 585
0.36
0.63
0.69
0.81
0.59
0.77
0.73
5.34
5.30
4.52
4.03
4.49
3.36
4.11
4.21
0.39
4.00
4.15
0.42
6.17
5.74
5.9
0.71
4.28
5.02
1.37
1.30
2.49
3.32
3.55
0.67
2.78
1.14
2.91
2.80
0.76
2.95
1.84
3.08
4.97
5.45
1.00
1.00
1.00
1.00
1.00
1.00
1.00
0.60
0.40
1.00
0.52
0.48
1.00
1.00
0.68
0.32
1.00
1.00
1.00
1.00
1.00
1.00
0.60
0.40
1.00
0.19
0.81
1.00
0.30
0.70
1.00
1.00
1.00
1.00
1.03
1.17
1.16
1.08
1.17
1.08
2.44
1.10
7979
6897
6001
5504
6167
4893
5604
0.68
1.18
1.53
2.00
1.32
2.29
—
11.97
7.04
6.79
4.76
9.04
6.90
—
n-Heptanol
29 197
23 697
0.65
3.74
1.10
5500
—
—
Ethylene glycol
Propylene
Glycol
29 137
29 214
21 276
21 978
0.52
0.55
1.06
3.25
1.13
7861
7236
1.78
—
6.57
—
Acetone
29 351
29 412
29 412
29 420
29 291
29 257
23 095
22 371
27 624
27 700
25 000
23 310
0.44
0.46
0.94
0.55
0.62
0.53
1.07
1.10
1.15
1.11
1.14
3.98
1.09
6436
7040
1788
1720
4291
5947
1.03
0.91
6.88
4.19
2.49
—
13.06
11.95
4.16
35.00
15.26
—
Acetonitrile
Cyclohexane
Isooctane
1,4-Dioxane
THF
2-Methyl-
THF
29 290
29 446
24 752
24 856
0.52
0.56
1.59
1.00
4539
4570
—
—
—
—
Diethyl
Ether
2.59
1.07
Toluene
Ethyl acetate
DMF
29 188
29 410
29 230
29 104
25 575
24 450
22 830
22 422
0.69
0.54
0.58
0.47
1.14
1.12
1.13
1.10
3613
4962
6400
6682
3.77
1.76
1.17
0.86
16.68
14.90
8.53
DMSO
9.74
molecules (Fig. 1S, ESI†) which was confirmed by time-
The calculations of excited state dipole moment values were
resolved measurements. The fluorescence intensity decays of 2 performed taking into account only solvents in which mono-
in dimethoxyethane measured at 370 nm and 500 nm are bi- exponential fluorescence decays were observed without a divi-
exponential in both cases, however, a negative pre-exponential sion on protic or aprotic ones. The correlations of spectral
factor (the growth of fluorescence intensity) is observed for properties with solvent polarity-polarizability functions accord-
measurement at λ = 500 nm (Table 2). The exciplex formation ing to eqn (5) and (6) are satisfactory for 1 (r = 0.9728 for eqn
with a solvent molecule is also confirmed by time-resolved (5) and r = 0.9682 for eqn (6), Table 3). The calculated excited
spectra (Fig. 2S, ESI†).
state dipole moment is equal to 8.7 D whereas an angle
The solvatochromic properties were applied also to the cal- between the ground and excited state dipole moments is 56.5°.
culations of excited state dipole moments of the compounds For 2, these correlations are much poorer (r = 0.5921 for eqn
studied using eqn (5) to (15), presented in the experimental (5) and r = 0.7514 for eqn (6), Table 3) giving an excited state
section. In these studies we accept the condition 2α/a3 = 1, fre- dipole moment value equal to 5.4 D with parallel orientation
quently adopted in the literature, and solvent polarity func- of the ground and excited state dipole moments. In such a
tions represented by eqn (7) and (8) were used.
case the ground state dipole moment can be calculated from
To calculate the excited state dipole moment the ground eqn (13) and the value of 2.7 D, comparable to the theoretically
state dipole moment is necessary. For both compounds, they calculated value (2.6 D), was obtained. Moreover, based on
were obtained from the theoretical calculation performed by eqn (15), the excited state dipole moment can be calculated
applying a DFT method with a def2-TZVP basis set (TURBO- only from the spectroscopic properties without taking into
MOLE v. 5.9) (2.6 D for 2 and 1.6 D for 1). The smaller value of account the Onsager’s radius and for 2 it is equal to 5.4 D,
the ground state dipole moment for 1 is probably caused by exactly as obtained from eqn (11).
nonplanarity of the N-phenylcarbazole moiety and the benzoxa-
The spectroscopic and photophysical properties of 2 and to
zole ring resulting in weaker interaction between them.13,42 In a less degree those of 1 depend on whether the solvent is
the case of 2, the carbazole subunit can adopt planar orien- protic or aprotic. Thus, the excited state dipole moments were
tation with the rest of the molecule, as it is not separated by the also calculated taking into account only aprotic solvents to
phenyl, yielding a higher ground state dipole moment value.
exclude the influence of hydrogen bond formation between
This journal is © The Royal Society of Chemistry and Owner Societies 2013
Photochem. Photobiol. Sci., 2013, 12, 284–297 | 289

View Article Online
Paper
Photochemical & Photobiological Sciences
Table 2 Spectroscopic and photophysical properties of 2 in all solvents studied
˜νabs [cm−1
]
˜νfluo [cm−1
]
φF
τ [ns]
α
χR
Δ˜ν [cm−1
]
kf/108 [s−1
]
knr/107 [s−1
]
2
Solvent
Methanol
Ethanol
1-Butanol
29 922
29 878
30 544
29 780
30 581
30 553
30 469
30 432
30 460
29 507
29 586
30 516
30 684
30 684
30 331
30 414
30 817
30 826
30 675
30 741
30 760
30 675
30 656
30 600
25 455
25 641
25 940
25 974
26 008
26 247
25 907
26 042
26 110
25 189
25 221
24 752
26 385
26 178
26 008
26 178
26 846
26 846
26 740
26 810
26 810
26 702
26 738
26 178
0.99
1.00
1.10
1.00
1.00
1.00
0.94
0.95
1.00
1.00
1.00
0.82
0.63
0.99
0.93
0.94
0.59
0.59
0.76
0.58
0.86
0.72
0.75
0.70
1.72
1.66
1.54
1.55
1.57
1.54
1.56
1.58
1.55
1.84
1.74
1.86
1.75
1.77
1.32
1.50
2.66
2.25
2.29
2.23
2.22
2.30
2.33
1.20
1.14
2.20
4.27
1.08
3.93
1.96
1.73
1.83
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
0.97
0.03
1.00
0.98
1.24
1.07
0.85
1.02
1.15
1.17
0.96
1.00
1.11
0.94
1.08
1.06
1.13
1.05
1.15
1.00
1.03
0.91
0.96
1.09
1.06
1.00
1.52a
1.09a
4537
4237
4604
3806
4573
4306
4589
4390
4350
4318
4365
5764
4299
4506
4323
4236
3971
3980
3935
3931
3950
3973
3918
4422
5.76
6.02
6.49
6.45
6.37
6.49
6.03
6.01
6.45
5.43
5.75
4.41
3.60
5.59
7.05
6.27
2.21
2.62
3.32
2.61
3.87
3.13
3.22
—
0.58
0
0
0
0
Iso-Butyl alcohol
Iso-Propyl alcohol
Amyl alcohol
n-Hexanol
Cyclohexanol
n-Heptanol
Ethylene glycol
Propylene glycol
Formamide
Acetone
Acetonitrile
Chloroform
Methylene chloride
n-Pentane
3-Methyl-pentane
Cyclohexane
n-Hexane
Isooctane
n-Nonane
n-Decane
Dimethoxyethane
0
3.85
3.16
0
0
0
9.68
21.14
0.56
5.30
4.00
15.41
18.22
10.48
18.83
6.31
12.17
10.73
—
18.49b
−2.73
3.73
1.00
1.00
1.00
1.08b
1.01
0.91
1.20
1,4-Dioxane
THF
2-Methyl-THF
Diethyl
30 553
30 554
30 590
26 350
26 281
26 385
1.00
0.84
0.93
4203
4273
4205
5.10
4.86
5.08
0
9.25
3.82
Ether
30 741
30 544
30 675
30 414
30 404
30 386
30 713
30 497
30 066
26 560
26 008
26 469
26 281
26 350
26 350
26 420
26 008
25 707
0.68
0.55
1.00
0.97
0.78
0.92
0.96
1.00
0.88
2.13
1.66
1.83
1.67
1.66
1.76
1.87
1.73
1.69
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
0.99
0.83
1.36
1.08
1.04
0.97
0.94
1.07
1.10
4181
4536
4206
4133
4054
4036
4293
4489
4359
3.19
3.31
5.46
5.81
4.70
5.23
5.13
5.78
5.21
15.02
27.11
0
1.79
13.25
4.54
2.14
0
Di-isopropyl ether
Propylene carbonate
Benzene
Toluene
o-Xylene
Ethyl acetate
DMF
DMSO
7.10
Measured ata λem = 370 nm, b λem = 500 nm.
Table 3 Ground state dipole moments (μg), the slopes of linear fits (m1) and (m2) obtained from the fitting of Stokes shift (eqn (5)) and sum of absorption and
emission maxima (eqn (6)) versus solvent polarity functions, regression coefficient (r) (representing the quality of fit), excited state dipole moments (μe), the angle
N
between ground and excited state dipole moments (ψ) and changes of dipole moments calculated from the dependence of the Stokes shift on the ET parameter
(eqn (16))
Compound
μg [D]
a [Å]
m1
rm1
m2
rm2
μe [D]
Ψ
Δμ [D]
Comment
1
1.6
6.0
5098 243
5188 215
0.9728
0.9857
−(6784 350)
−(6787 329)
0.9682
0.9806
8.7
8.7
56.5
53.3
6.1
8.3
6.9
2.7
3.9
4.6
All solvents
Aprotic
Protic
All solvents
Aprotic
Protic
2
2.6
6.5
574 136
670 83
0.5921
0.8644
−(1626 245)
−(1175 209)
0.7514
0.7606
5.4
4.8
0.0
34.3
the solute and the solvent. It was found that exclusion of bit lower (μe = 4.8 D), while the angle between dipoles substan-
protic solvents does not change much parameters obtained for tially increases to the value of 34.3° (Table 3).
1 in contrast to compound 2 (Table 3). The quality of fit for 2
To calculate the dipole moment change between the excited
is much better, especially for eqn (5) (r = 0.8644). The calcu- and ground states, the correlation of the Stokes shift with the
lated excited state dipole moment for aprotic solvents is now a microscopic solvent polarity parameter ETN can be applied.45,46
290 | Photochem. Photobiol. Sci., 2013, 12, 284–297
This journal is © The Royal Society of Chemistry and Owner Societies 2013

View Article Online
Photochemical & Photobiological Sciences
Paper
According to the equation proposed by Ravi et al.,45 the 3.2 Multi-linear correlations
problem associated with the Onsager’s radius estimation can
Multi-parameter correlation is preferable and has been applied
successfully to various physicochemical parameters. The most
frequently used solvent scales are those of Kamlet and
Taft,20–23 and Catalán.24–28 As Filarowski et al.47 and Guzow
et al.12 demonstrated, the new four-parameter Catalán solvent
scale has an advantage over other solvent scales. Therefore, in
this manuscript the spectral and photophysical parameters of
the compounds studied in all solvents used will be discussed
based on the four-parameter Catalán solvent scale.28 However,
the correlations obtained using Kamlet–Taft and an old three-
parameter Catalán solvent scale are also presented (Tables 3S–
6S, ESI†) and will be compared with each other.
Table 4 collects the estimated coefficients and correlation
coefficients (r) for the multi-linear regression analysis of
absorption (˜νabs) and fluorescence (˜νfluo) maxima, fluorescence
quantum yield (φF) and Stokes shift (Δ˜ν) of 1 using the four-
parameter Catalán solvent polarity scale (data obtained for
three-parameter Catalán or Kamlet–Taft solvent polarity scales
are collected in Tables 3S and 4S, ESI†). A fair fit was obtained
for the dependence of maxima of absorption (˜νabs) (r = 0.9387)
in which all four coefficients are important. However, the large
value of the aSP coefficient compared to others (bSdP, cSA and
dSB) (Table 4) indicates that the change of (˜νabs) may reflect pri-
marily a change in polarizability of the environment of the
chromophore, whereas the solvent dipolarity (bSdP), because of
its low value and large error, may be disregarded (quality of
the fit decreases a little (r = 0.9378) compared to the original
fit (r = 0.9387)). Solvent acidity and basicity affect the position
of the absorption maxima but to a smaller degree than its
polarizability. For compound 2, the analysis of (˜νabs) according
to eqn (4) (Table 5) indicates that (˜νabs) is affected by all four
parameters, however, the greatest impact is exhibited by polar-
izability of the environment of the chromophore, as for com-
be minimized since a ratio of two Onsager’s radii is involved:
ꢂꢀ
ꢁ
ꢃ
2
ꢄ
ꢅ
3
Δμ
ΔμB
aB
a
Δ˜υ ¼ ˜υa ꢀ ˜υf ¼ 11307:6
ETN þ const
ð16Þ
where ΔμB and aB are the dipole moment change (Δμ = μe −
μg) and Onsager’s radius, respectively for pyridinium
a
N
N-phenolate betaine dye used to determine the ET values
(ΔμB = 9 D and aB = 6.2 Å), whereas Δμ and a are the corre-
sponding quantities for the molecule under study. Calculated
according to eqn (16) dipole moment changes are presented in
Table 3. The excited state dipole moments calculated taking
into account all solvents are comparable to the values obtained
from the Kawski’s theory, however, excited state dipole
moments calculated only for aprotic solvents appear to be
overestimated.
3.1 Single parameter correlation
Absorption maximum (in wavenumber), fluorescence
maximum (in wavenumber), Stokes shift, fluorescence
quantum yield as well as fluorescence lifetime of the com-
N
pounds studied were also correlated with the ET solvent
N
polarity parameter using an equation: y = y0 + b × ET (Tables
1S and 2S (ESI†) for compounds 1 and 2, respectively). Gener-
ally, the values of correlation coefficients are rather small indi-
N
cating the weak or lack of correlation with the ET parameter
regardless of whether it is a whole set of solvents studied, or
after the division on protic and aprotic ones. Moreover, for
protic solvents the correlation coefficient is always lower than
that for the aprotic ones, except for the fluorescence quantum
yield of 1 and the fluorescence lifetime of 2. The increase of
the correlation coefficient for protic or aprotic solvents com-
pared to that for all set of solvents indicates that the partition
of solvents on these two groups in most cases is justified.
pound 1. The influence of solvent acidity and basicity on (˜νabs
)
is bigger for 2 than for 1. Similar dependences reflect the cor-
relation of (˜νabs) with the three-parameter solvent polarity
Table 4 Estimated from eqn (4) coefficients (yo, aSP, bSdP, cSA, dSB), their standard errors and correlation coefficients (r) for the multiple linear regression analysis of
, ν , Δν, φ , k , knr of 1 as a function of the Catalán four-parameter solvent scale
ν
˜
˜
˜
abs fluo
F
f
y
y0
aSP
bSdP
cSA
−(67 46)
−(62 43)
−(2292 595)
−(2347 593)
2225 589
dSB
r
˜ν abs
˜ν fluo
Δ˜ν
30 437 107
30 433 104
28 032 1389
27 144 264
2405 1373
2218 257
0.37 0.25
0.38 0.24
0.35 1.11
1.15 0.21
3.84 1.75
5.23 0.56
1.12 0.35
1.20 0.07
4.25 3.15
−(1552 154)
−(1539 148)
−(1208 2004)
17 36
−(111 44)
−(103 40)
−(258 571)
0.9387
0.9378
0.9623
0.9610
0.9617
0.9614
0.7501
0.7437
0.9644
0.9627
0.9186
0.9073
0.9160
0.9045
0.4985
−(4470 468)
−(4595 409)
4487 462
−(344 1081)
147 565
4518 399
2283 526
0.06 0.11
φF
0.50 0.36
0.48 0.35
1.27 1.60
−(0.37 0.08)
−(0.36 0.08)
3.60 0.37
0.19 0.10
0.20 0.10
−(0.15 0.45)
τ*109
kf × 10−8
2.39 0.47
2.30 0.43
−(0.09 0.76)
3.57 0.33
1.87 2.52
0.19 0.51
−(3.69 0.58)
−(3.20 1.03)
−(0.77 0.11)
−(0.84 0.09)
−(0.73 1.26)
−(1.46 0.71)
−(2.43 1.73)
−(0.17 0.14)
kf/(n2˜υf3luo) × 105
knr × 10−8
−(0.01 0.15)
−(3.41 4.64)
1.03 15.18
0.03 1.88
This journal is © The Royal Society of Chemistry and Owner Societies 2013
Photochem. Photobiol. Sci., 2013, 12, 284–297 | 291

View Article Online
Paper
Photochemical & Photobiological Sciences
Table 5 Estimated from eqn (4) coefficients (yo, aSP, bSdP, cSA, dSB), their standard errors and correlation coefficients (r) for the multiple linear regression analysis of
, ν , Δν, φ , τ, k , knr of 2 as a function of the Catalán four-parameter solvent scale
ν
˜
˜
˜
abs fluo
F
f
y
y0
aSP
bSdP
cSA
−(843 250)
−(1376 105)
−(1356 101)
523 263
dSB
r
˜ν abs
˜ν fluo
32 345 504
28 032 211
28 093 194
4313 528
4003 89
0.27 0.24
0.27 0.23
3.40 0.41
3.40 0.40
−2.22 1.78
0.11 0.50
3.57 1.23
3.79 1.14
−(2484 736)
−(2241 309)
−(2100 291)
−(460 772)
−(327 179)
−(583 75)
−(552 63)
910 188
860 167
0.10 0.09
−(489 199)
0.7384
0.9785
0.9781
0.7744
0.7713
0.6696
0.5899
0.7954
0.7954
0.8067
0.7129
0.6696
0.6349
63 83
Δ˜ν
−(553 208)
−(513 195)
0.11 0.10
533 260
φF
0.69 0.35
0.79 0.33
−(1.79 0.60)
−(1.80 0.58)
8.45 2.61
1.18 0.73
−3.51 1.80
−4.19 1.64
0.17 0.12
0.32 0.10
0.02 0.22
τ × 109
kf × 10−8
−(0.28 0.14)
−(0.28 0.14)
1.11 0.62
0.49 0.17
−0.44 0.43
−(0.42 0.16)
−(0.41 0.15)
1.30 0.71
0.30 0.20
−0.21 0.49
1.47 0.91
0.70 0.26
−1.42 0.65
−1.97 0.53
kf/(n2˜υf3luo) × 105
knr × 10−8
scales (Tables 3S–6S, ESI†), however, the qualities of fits are that the SP parameter does not influence the Stokes shift
much worse. (Table 5). For both compounds, the correlations with three-
The fit of (˜νfluo) as a function of SP, SdP, SA, SB is satisfac- parameter solvent scales give worse fits and indicate a negli-
tory, as judged by the value of r as the quality-of-fit criterion, gible influence of basicity of the solvent (Tables 3S–6S, ESI†).
for both compounds studied (r = 0.9632 and r = 0.9785 for 1
The correlation of the fluorescence quantum yield of 1 with
and 2, respectively). However, contrary to the fit of (˜νabs), the SP, SdP, SA, SB as independent variables gives a rather weak fit
estimated coefficients determining the influence of solvent (r = 0.7501, Table 4). Moreover, it indicates that solvent acidity
polarizability and basicity for 1 are calculated with a consider- does not influence the fluorescence quantum yield whereas
able standard error aSP = −(1208 2004) and dSB = −(258
the solvent polarizability increases while solvent dipolarity
571), higher than the coefficient value, and can be omitted. decreases the fluorescence quantum yield to a similar degree.
The fit according to eqn (4) with (SdP and SA) as independent Also, for 2 a weak correlation of the fluorescence quantum
variables has a correlation coefficient (r = 0.9610) that is nearly yield with solvent polarity parameters (eqn (4)) was obtained
the same as for the original fit (r = 0.9623). Also, it is worth (r = 0.6696), however, in this case solvent dipolarity and basi-
mentioning that the bSdP and cSA coefficient values are negative city have no impact on it. The same conclusion, for both com-
meaning that their changes cause red-shift of the fluorescence pounds, can be drawn from the correlations by applying three-
maximum. Also, they are higher, taking into account their parameter solvent scales (Tables 3S–6S, ESI†).
absolute values, than those calculated for the absorption. This
For both compounds studied, the decay of fluorescence
indicates that the electronic structure of the Franck–Condon intensity can be described as a mono-exponential function,
and relaxed excited state of 1 differ significantly, and the except for the previously discussed solvents, thus, further
relaxed excited state is more polar, which is consistent with a analysis including the dependence of fluorescence lifetime,
higher excited state dipole moment than the ground state. For fluorescence rate constant and radiationless rate constant on
compound 2, the greatest impact on (˜νfluo) is exhibited by solvent polarity parameters can be performed. The correlation
solvent polarizability and dipolarity (almost equally), and according to eqn (3) and (4) of the fluorescence lifetimes of 1
acidity. Because of a small value and a large standard error of is satisfactory as in the case of (˜νfluo) and the same solvent
solvent basicity (dSB), it can be omitted in the correlation parameters have a decisive impact (Tables 4, 3S and 4S, ESI†).
without changing the quality of the fit (r = 0.9781 compared to For 2, the correlation according to eqn (4) is better (Table 5)
r = 0.9785 for the original fit). The values of coefficients influen- than in the case of three-parameter solvent scales (Tables 5S
cing (˜νfluo) and (˜νabs) are comparable indicating that the elec- and 6S, ESI†). However, the solvent acidity can be omitted for
tronic structures of the Franck–Condon and relaxed excited Catalán four-parameter and Kamlet–Taft solvent scales
states of 2 are similar and the relaxed excited state is only (Table 5 and Table 6S, ESI†).
slightly more polar, which is consistent with a small change in
For the mono-exponential fluorescence intensity decay, the
dipole moment values. Application of the three-parameter radiative (kf) and radiationless (knr) rate constants can be cal-
solvent scale gives similar dependences but worse qualities of culated from the measured fluorescence quantum yield (φF)
the correlation for 2 contrary to 1 for which the correlation is and fluorescence lifetime (τ) according to eqn:
very good and all three parameters influence the position of the
kf ¼ φF=τ
ð17Þ
fluorescence maximum significantly (Tables 3S–6S, ESI†).
Correlation of the Stokes shift of 1 with the four-parameter
solvent scale gives the same results as for (˜νfluo) (Table 4). In
the case of 2, such correlation is weak, probably because of the
difficulties in the correct estimation of the (˜νabs), and reveals
and
knr ¼ ð1 ꢀ φFÞ=τ
ð18Þ
292 | Photochem. Photobiol. Sci., 2013, 12, 284–297
This journal is © The Royal Society of Chemistry and Owner Societies 2013

View Article Online
Photochemical & Photobiological Sciences
Paper
Analysis of the influence of the solvent on the radiative and denotes the permittivity of vacuum. The multi-linear Catalán
non-radiative rate constants gives information which of them fit of y = kf/(n2˜υf3luo) according to eqn (4) gives the information
depends on the solvent parameters enabling us to estimate about the influence of solvent parameters on the transition
which path of the excited state deactivation predominates. dipole moment. For 2, a low value of r = 0.7129 indicates that
Moreover, the radiationless rate constant is the sum of the rate there is no clear correlation between kf/(n2˜υf3luo) and the
constant for internal conversion (kIC) and rate constant for Catalán solvent parameters (SP, SdP, SA and SB) (Table 5). The
intersystem crossing (kISC). The value of the first one depends lack of such a correlation was also established by Filarowski
on the energy gap between the ground and excited states. If it et al.47 for BODIPY derivatives. The average value of kf/(n2˜υf3luo
)
correlates with the solvent polarity it indicates that the energy for the set of 33 solvents presented in Table 2 is 1.4 0.6 ×
gap between these states decreases. In the opposite case, the 10−5 s−1 cm3. However, for 1 there is a clear correlation
main path of the excited state deactivation is intersystem cross- between kf/(n2˜υf3luo) and the Catalán solvent parameters (r =
ing to the triplet state.
0.9160, Table 4). More detailed analysis shows that the solvent
The multi-linear analysis of knr according to eqn (4) for 1 dipolarity is a major factor influencing the square of the tran-
revealed that it does not correlate with solvent parameters sition dipole moment as derived from the linear fit of
(Table 4) in contrast to kf which depends mainly on solvent kf/(n2˜υf3luo) as a function of SdP giving an r-value equal to
dipolarity. A linear fit of knr versus (˜νfluo) gives r = 0.1620, 0.9054, just a little smaller than the original fit (0.9160).
which means that the non-radiative rate constant probably does
3.3 Quantum chemical calculations
not depend on the energy gap between the ground and excited
states. Moreover, this lack of correlation may indicate that the Theoretical calculations performed show that both compounds
intersystem crossing is the main radiationless deactivation studied are not flat in the ground state as well as in the excited
pathway and the excited state deactivation proceeds through the state. For compound 1, the 2-phenyl-benzoxazole subunit
triplet state. A mean value of knr is equal to 1.5 0.8 × 108 s−1
.
forms a flat structure, whereas the carbazole moiety is twisted
As both kf and (˜νfluo) are controlled mainly by solvent dipolarity against the rest of the molecule. The dihedral angle is equal to
one can expect a linear dependence of kf versus (˜νfluo). Indeed, a 52.6° and 52.0° in acetonitrile (ε = 37.5) and cyclohexane (ε =
linear dependence of kf = kf0 + α(˜νfluo) is satisfactory (r = 0.9411). 2), respectively. It is bigger than that for N-phenylcarbazole
Thus, the invariance of knr and dependence of kf on the solvent obtained from a semi-empirical calculation (AM1 method, 37.6
properties indicate that the change of the fluorescence degrees43), but a little lower than that obtained at the
quantum yield and the fluorescence decay time with solvent DFT-B3LYP/6-311G(d,p) level (ca. 60°13). In the excited state,
properties is mainly due to the change of the kf rate constant.
A more complicated situation occurs for 2. A multi-linear 2-phenyl-benzoxazole unit regardless of the solvent polarity.
correlation according to eqn (4) reveals that there is no clear A similar situation exists for 2, but in this case a flat system
the carbazole subunit is oriented perpendicularly to the
correlation between knr and Catalán solvent parameters (SP, is formed by benzoxazole and carbazole subunits whereas the
SdP, SA, SB) taking into account the low value of r. Moreover, phenyl ring forms with carbazole a dihedral angle equal to
the fit with all four solvent parameters gives a little better r 57.7° in both solvents. In the excited state this dihedral angle
value (0.6696) than that with only solvent polarizability and decreases to 45.6° and 44.3° in cyclohexane and acetonitrile,
acidity (r = 0.6349). The linear fit of knr versus (˜νfluo) gives r = respectively.
0.5114 meaning that only 25% of knr changes results from the
Applying the TD DFT method and the COSMO module the
changes of the energy gap between the ground and excited vertical absorption transitions for both compounds were calcu-
states caused by increasing solvent polarity while the rest of lated and are presented in the graphic form in Fig. 4 where ver-
the changes are determined by other factors. The multi-linear tical lines whose height is proportional to the oscillator
analysis of kf according to eqn (4) gives a little better fit (r = strength and the theoretical absorption spectra have been
0.7129). The radiative rate constant depends on solvent proper- simulated by considering a constant FWHM of 4000 cm−1. Fur-
ties in about 50% and all Catalán solvent parameters (SP, SdP, thermore, in Fig. 4 experimental absorption spectra measured
SA, SB) have impact on it. The mean value of kf = 5.0 × 108 s−1 in cyclohexane and acetonitrile are also presented. Addition-
is one order higher than knr = 6 × 107 s−1 explaining high ally, all calculated transitions with the oscillator strength
values of the fluorescence quantum yield of 2.
higher than 0.1 and the molecular orbital contribution to the
According to Birks,48 the radiative rate constant kf divided transition are collected in Tables 7S and 8S (ESI)† for 1 and 2,
by the square of the refractive index of a medium and average respectively. Generally, the shape of theoretical absorption
wavenumber of fluorescence in the third power is proportional spectra calculated using the pbe0 functional reproduced the
to the squared transition dipole moment:
experimental spectra.
For 1, both calculated spectra are a little red-shifted in the
long-wavelength part, compared to the experimental ones,
whereas the short-wavelength part of these spectra is very well
reproduced. The long-wavelength absorption transition of 1 is
kf
n2˜υ3fluo 3ε0h
16π3
¼
kΨGjμˆjΨEl2
ð19Þ
where μˆ stands for the dipole moment operator, n is the refrac- mainly the HOMO → LUMO transition (contribution more
tive index of the medium, h is Planck’s constant and ε0 than 98%, Table 7S (ESI)†). The HOMO has a pronounced π
This journal is © The Royal Society of Chemistry and Owner Societies 2013
Photochem. Photobiol. Sci., 2013, 12, 284–297 | 293

View Article Online
Paper
Photochemical & Photobiological Sciences
state dipole moments are comparable (19.9 D in cyclohexane).
The calculated oscillator strengths of emission transition are
small (0.15 × 10−3 in acetonitrile (ε = 37.5) and 0.36 × 10−4 in
cyclohexane (ε = 2)), which is consistent with the transition of
ICT type but does not agree with large fluorescence quantum
yield measured for 1 in acetonitrile (0.46) and cyclohexane
(0.52).
In contrast to the geometrical parameters, which varied
only slightly when different functionals are used, the excitation
energies and theoretical oscillator strengths are sensitive to
the choice of the functional. A crucial step in the theoretical
investigations employing TD DFT is the choice of an approxi-
mate exchange–correlation functional and a reasonable basis
set. Since one cannot expect that one approximate functional
describes excited states of a molecule equally well, especially
when different kinds of transitions are present (π → π* and
CT).49 It is well-known that for a large extended π system,50 CT
transitions are improperly reproduced by DFT calculations.51
However, in such cases the pbe0 functional as well as BNL or
CAM-B3LYP are recommended.52
For compound 2, the first vertical absorption transition at
326.2 nm in acetonitrile (HOMO → LUMO, 85.4% and HOMO
→ LUMO + 1, 8.2%) agrees well with the experimental spec-
trum, however, the second transition at 315.9 nm (HOMO →
LUMO + 1, 71.5%, HOMO − 2 → LUMO, 14.8%, HOMO →
LUMO, 8.5%) is red-shifted, while the third at 288.5 nm
(HOMO − 2 → LUMO, 71.7%, HOMO − 2 → LUMO + 1, 10.5%,
HOMO → LUMO + 1, 9.0%) is blue-shifted according to the
experimental spectrum (Fig. 4). Calculated absorption tran-
sitions, except for the first two absorption transitions, have
small oscillator strengths compared to the experimental spec-
trum. The calculated vertical absorption transitions in cyclo-
hexane, except being blue-shifted for a few nanometers, look
very similar to those calculated for acetonitrile (Fig. 4,
Table 8S, ESI†). One of the reasons which can explain to some
extent the differences between theoretical and experimental
oscillator strengths might be the intensity borrowing phenom-
ena. The analysis of HOMO and LUMO orbitals of 2 (Fig. 4S,
ESI†) reveals that contrary to 1 long-wavelength absorption
transition has a π → π* character.
The optimized structure of 2 in the excited state is similar
to the structure in the ground state. The subunit formed by
benzoxazole and carbazole is flat, while the phenyl substituent
is twisted. The dihedral angle between carbazole and phenyl
rings decreased to about 45.6° in cyclohexane and 44.3° in
acetonitrile. In both solvents the vertical emission transition
occurs exclusively between LUMO and HOMO orbitals (95.9%
in both solvents). The calculated emission transitions for 2 are
blue-shifted for about 20 nm in both solvents, compared to
Fig. 4 Experimental and theoretically calculated absorption spectra as well as
the oscillator strengths of calculated vertical transitions of 1 (at the top) and 2
(at the bottom) in MeCN and cyclohexane.
character localized on the carbazole moiety, whereas the
LUMO has antibonding π character localized on the 2-phenyl-
benzoxazole moiety (Fig. 3S, ESI†). Thus, the HOMO → LUMO
transition has an intramolecular charge-transfer character
(Fig. 3S, ESI†). Contrary to the vertical absorption transition,
the calculated vertical emission transition does not correctly
reproduce the position of the emission spectrum of 1 in both
studied solvents. The transition obtained from the calculation
using the pbe0 functional is blue-shifted in acetonitrile com-
pared to the experimental spectrum (λcal = 403.4 nm, λmax
447 nm), whereas it is red-shifted in cyclohexane (λcal
401.4 nm, λmax = 362 nm).
=
=
The difference in electron density between the ground and
excited states of 1 shows that the emission transition possesses
ICT character (HOMO–LUMO 98.5%) which generates a big
excited state dipole moment (μS1cal = 19.7 D) – much bigger
than that calculated based on the solvatochromic shift (μS1
=
8.7 D, Table 3). The difference between wavelengths of calcu-
lated vertical emission transitions in acetonitrile and cyclo-
hexane is due to the destabilization of the ground and excited
states of 1 in cyclohexane compared to acetonitrile. The
ground and excited state energies in cyclohexane are higher
for about 42.6 kJ mol−1 and 26.5 kJ mol−1, respectively, com-
pared to acetonitrile. Moreover, for both solvents the character
of the transition is the same (ICT) and the calculated excited
the maxima of experimental emission spectra (ε = 2, λcal
=
357 nm, λmax = 374 nm; ε = 37.5, λcal = 361 nm, λmax = 382 nm).
Moreover, the calculated oscillator strength in cyclohexane is
equal to f = 1.035, while for acetonitrile f = 0.915, which agrees
well with large fluorescence quantum yields measured for that
compound (0.76 and 0.99 in cyclohexane and acetonitrile,
respectively). The electron density difference between the
294 | Photochem. Photobiol. Sci., 2013, 12, 284–297
This journal is © The Royal Society of Chemistry and Owner Societies 2013

View Article Online
Photochemical & Photobiological Sciences
Paper
ground and the excited state indicates a small charge transfer, consequence,
a
smaller excited state dipole moment is
which is reflected in the relatively small dipole moment of 2 in observed for 2 which leads to the smaller Stokes shift. The
the excited state (5.95 D in cyclohexane and 9.45 D in aceto- different types of transitions (π → π* for 2 and ICT for 1) result
nitrile, respectively). The excited state dipole moments of 2 are in a higher fluorescence quantum yield observed for 2 than for
much smaller than those calculated for 1. However, the value 1. The diverse charge distribution in the excited state of each
calculated for ε = 37.5 (acetonitrile) is far higher than the value compound suggests that the primary determinant of the
determined from solvatochromic measurements (Table 3), photophysical properties in the multi-linear analysis is dipolar-
while for ε = 2 (cyclohexane) they are comparable.
ity for 1, whereas for 2 – polarizability and dipolarity. More-
Theoretical calculations give information not only about over, as both compounds contain a heterocyclic nitrogen atom
dipole moments but also about the transition dipole moments some influence of solvent basicity and acidity is also recorded.
and their relative orientation. The values of angles between As stated in the literature,12,47 also this case revealed the pre-
absorption and fluorescence transition dipole moments dominance of a four-parameter solvent scale over three-para-
obtained from theoretical calculations are equal to 0.3° for 1 metric ones. A four-parameter solvent scale allows a deeper
and 4.1° for 2 (Fig. 5S and 6S, ESI†) whereas those obtained interpretation of the results due to the breakdown of the
experimentally from fluorescence anisotropy measurements solvent polarizability/dipolarity parameter (π* in the Kamlet–
for stretched samples in PVA films are about 16.8° and 21°, Taft scale or SPP in the old Catalán scale) into two new para-
respectively (Fig. 7S and 8S, ESI†). The experimental values do meters – solvent polarizability (SP) and dipolarity (SdP).
not fit very well to the Kawski–Gryczyński model, especially in
The results of theoretical calculation significantly differ
the case of 2, and should be treated only as approximate from experimental results, especially in the case of excited
values. For compound 2 larger scatter of the experimental state dipole moment values and the influence of solvent
points is observed than for compound 1 which probably polarity on the position of vertical transition. These discrepan-
results from the differences in shapes of those compounds. cies are connected with the simulation of the solvent polarity
The rotational ellipsoid of compound 1 is more elongated (COSMO model) used in the calculations which does not take
than that of compound 2. Thus, compound 2 does not arrange into account specific solute–solvent interactions. However, the
appropriately along the stretching direction. The observed dis- main reason is the method used (DFT) which for compounds
crepancy between theoretical and experimental values may be as those described here with large extended π systems and/or
a result of different environments used in each method (model exhibited charge transfer does not reproduce well experimental
of continuous solvent in theoretical calculations and PVA absorption and emission spectra in contrast to the geometrical
matrix in anisotropy measurements) as well as not appropriate parameters. The calculated geometries of the compounds
model used for the calculation of the angle between transition studied in the ground and excited states help to explain the
dipole moments. However, both methods, experimental and experimental results of solvatochromic studies (the character
theoretical, give higher values of that angle for compound 2.
of the ground and excited states as well as the photophysical
properties).
Both studied compounds show solvatochromism but their
practical application may be limited as their spectral proper-
ties are in the UV range. However, changes in the carbazole
4 Conclusions
The analysis of photophysical properties of the compounds substituent such as introduction of the electron-donating
studied revealed that they are determined by the place of the groups or extension of the aromatic ring will probably help to
N-phenylcarbazole substitution by the benzoxazole moiety. overcome this limitation by shifting the spectra of such deriva-
Introduction of carbazole in position 4 of the phenyl ring tives bathochromically. Another possibility is alkylation of the
(compound 1) leads to a compound with ICT type excited state benzoxazole nitrogen which not only changes the spectral
with perpendicular orientation of the carbazole subunit in properties of the compound but also should significantly
relation to a flat phenylbenzoxazole subunit. This in turn leads improve its solubility in water.
to a large excited state dipole moment and a large Stokes shift,
similarly as in the case of 3-[2-(4-diphenylaminophenyl)ben-
zoxazol-5-yl]alanine – the analogue without stiffening.9 For
this compound, the conformational freedom of phenyl substi-
Acknowledgements
tuents on the nitrogen atom as well as different coupling of This work was supported by the Polish Ministry of Science and
non-bonding lone pair of the nitrogen atom with aromatic Higher Education under grant DS-8441-4-0132-12.
parts of the molecule suggest that the excited state dipole
moment is higher. Moreover, in this case the influence of the
character of the solvent is more pronounced.
Notes and references
A substitution in position 3 of the carbazole core leads to
the compound with a little higher ground state dipole
moment, however, with much smaller charge redistribution in
1 J. R. Lakowicz, Principles of Fluorescence Spectroscopy,
Kluwer Academic/Plenum Publishers, New York, 2nd edn,
1999.
the excited state in comparison to compound 1. As
a
This journal is © The Royal Society of Chemistry and Owner Societies 2013
Photochem. Photobiol. Sci., 2013, 12, 284–297 | 295

View Article Online
Paper
Photochemical & Photobiological Sciences
2 M. R. Eftink and C. A. Ghiron, Exposure of tryptophanyl 15 D. Kim, J. K. Lee, S. O. Kang and J. Ko, Molecular engineer-
residues in proteins. Quantitative determination by fluo-
rescence quenching studies, Biochemistry, 1976, 15, 672–
680.
3 J. R. Lombardi, Solvatochromic shifts reconsidered: field-
induced mixing in the nonlinear region and application to
indole, J. Phys. Chem. A, 1999, 103, 6335–6338.
4 K. Guzow, M. Szabelski, J. Malicka and W. Wiczk, Synthesis
ing of organic dyes containing N-aryl carbazole moiety for
solar cell, Tetrahedron, 2007, 63, 1913–1922.
16 C. Ito, M. Itoigawa, A. Sato, C. M. Hasan, M. A. Rashid,
H. Tokuda, T. Mukainaka, H. Nishino and H. Furukawa,
Chemical constituents of Glycosmis arborea: three new car-
bazole alkaloids and their biological activity, J. Nat. Prod.,
2004, 67, 1488–1491.
of a new, highly fluorescent amino acid derivative: N-[(tert- 17 Y. Liu, M. Nishiura, Y. Wang and Z. Hou, π-Conjugated aro-
butoxy)carbonyl]-3-[2-(1H-indol-3-yl)benzoxazol-5-yl]-L-alanine
methyl ester, Helv. Chim. Acta, 2001, 84, 1086–1092.
5 A. Rzeska, J. Malicka, K. Guzow, M. Szabelski and
matic enynes as a single-emitting component for white
electroluminescence, J. Am. Chem. Soc., 2006, 128, 5592–
5593.
W. Wiczk, New highly fluorescent amino-acid derivatives: 18 M. V. Skorobogatyi, A. A. Pchelintseva, A. L. Petrunina,
substituted 3-[2-(phenyl)benzoxazol-5-yl]-alanines: syn-
thesis and photophysical properties, J. Photochem. Photo-
biol., A, 2001, 146, 9–18.
6 K. Guzow, M. Szabelski, J. Malicka, J. Karolczak and
W. Wiczk, Synthesis and photophysical properties of 3-[2-
I. A. Stepanova, V. L. Andronova, G. A. Galegov,
A. D. Malakhov and V. A. Korshun, 5-Alkynyl-2′-deoxyuri-
dines, containing bulky aryl groups: evaluation of struc-
ture-anti-HIV-1 activity relationship, Tetrahedron, 2006, 62,
1279–1287.
(pyridyl)benzoxazol-5-yl]-L-alanine derivatives, Tetrahedron, 19 A. A. Kubicki, P. Bojarski, M. Grinberg, M. Sadownik and
2002, 58, 2201–2209.
B. Kukliński, Time-resolved streak camera system with
solid state laser and optical parametric generator in
different spectroscopic applications, Opt. Commun., 2006,
263, 275–280.
7 K. Guzow, K. Mazurkiewicz, M. Szabelski, R. Ganzynkowicz,
J. Karolczak and W. Wiczk, Influence of an aromatic substi-
tuent in position 2 on photophysical properties of benzoxa-
zol-5-yl-alanine derivatives, Chem. Phys., 2003, 295, 119– 20 M. J. Kamlet, J. L. M. Abbound and R. W. Taft, An examin-
130.
ation of linear solvation energy relationships, Prog. Phys.
Org. Chem., 1982, 13, 485–623.
8 K. Guzow, M. Szabelski, J. Karolczak and W. Wiczk, Solvato-
chromism of 3-[2-(aryl)benzoxazol-5-yl]alanine derivatives, 21 M. J. Kamlet and R. W. Taft, Solvatochromic comparison
J. Photochem. Photobiol., A, 2005, 170, 215–223.
9 K. Guzow, M. Milewska and W. Wiczk, Solvatochromism of
method. 1. β-Scale of solvent hydrogen-bond acceptor
(HBA) basicities, J. Am. Chem. Soc., 1976, 98, 377–383.
3-[2-(4-diphenylaminophenyl)benzoxazol-5-yl]alanine methyl 22 R. W. Taft and M. J. Kamlet, Solvatochromic comparison
ester. A new fluorescence probe, Spectrochim. Acta, Part A,
2005, 61, 1133–1140.
method. 2. α-Scale of solvent hydrogen-bond donor (HBD)
acidities, J. Am. Chem. Soc., 1976, 98, 2886–2894.
10 K. Guzow, J. Zielińska, K. Mazurkiewicz, J. Karolczak and 23 M. J. Kamlet, J. L. M. Abboud and R. W. Taft, Solvatochro-
W. Wiczk, Influence of substituents in the phenyl ring on
photophysical properties of 3-[2-(phenyl)benzoxazol-5-yl]-
mic comparison method. 6. π* Scale of solvent polarities,
J. Am. Chem. Soc., 1977, 99, 6027–6038.
alanine derivatives, J. Photochem. Photobiol., A, 2005, 175, 24 J. Catalán, V. López, P. Pérez, R. Martin-Villamil and
57–68.
J. G. Rodriguez, Progress towards a generalized solvent
polarity scale The solvatochromism of 2-(dimethyl-
amino)-7-nitrofluorene and its homophorm 2-fluoro-7-
11 K. Guzow, D. Szmigiel, D. Wróblewski, M. Milewska,
J. Karolczak and W. Wiczk, New fluorescence probes based
on 3-(2-benzoxazol-5-yl)alanine skeleton – synthesis and
–
nitrofluorene, Liebigs Ann., 1995, 241–252.
photophysical properties, J. Photochem. Photobiol., A, 2007, 25 J. Catalán and C. Díaz, A generalized solvent acidity scale:
187, 87–96.
the solvatochromism of o-tert-butylstilbazolium betaine dye
and its homophorm o,o′-di-tert-butylstilbazolium betaine
dye, Liebigs Ann. Recl., 1997, 1941–1949.
12 K. Guzow, A. Ceszlak, M. Kozarzewska and W. Wiczk, Influ-
ence of substituents on the nitrogen atom of 3-[2-(4-amino-
phenyl)benzoxazol-5-yl]alanine
derivatives
on
their 26 J. Catalán, C. Díaz, V. López, P. Pérez, J. L. G. de Paz and
photophysical properties – solvatochromic studies, Photo-
chem. Photobiol. Sci., 2011, 10, 1610–1621.
13 O. Kwon, S. Barlow, S. A. Odom, L. Beverina,
J. G. Rodriguez, A generalized solvent basicity scale: the sol-
vatochromism of 5-nitroindoline and its homophorm
1-methyl-5-nitroindoline, Liebigs Ann., 1996, 1785–1794.
N. J. Thompson, E. Zojer, J.-L. Bredas and S. R. Marder, 27 J. Catalan and H. Hopf, Empirical treatment of the induc-
Aromatic amines:
strengths, J. Phys. Chem. A, 2005, 109, 9346–9352.
14 V. Perciuraite, S. Grigalevicius, J. Simokaitiene and
a
comparison of electron-donor
tive and dispersive component of solute-solvent inter-
actions: the solvent polarizability (SP) scale, Eur. J. Org.
Chem., 2004, 4694–4702.
J. V. Grazulevicius, Indolyl-substituted carbazole deriva- 28 J. Catalán, Toward a generalized treatment of the solvent
tives as amorphous electroactive materials for opto-
electronics, J. Photochem. Photobiol., A, 2006, 182,
38–42.
effect based on four empirical scales: dipolarity (SdP, a new
scale), polarizability (SP), acidity (SA), and basicity (SB) of
the medium, J. Phys. Chem. B, 2009, 113, 5951–5960.
296 | Photochem. Photobiol. Sci., 2013, 12, 284–297
This journal is © The Royal Society of Chemistry and Owner Societies 2013

View Article Online
Photochemical & Photobiological Sciences
Paper
29 A. Kawski, in Progress in Photochemistry and Photophysics, 42 A. Sarkar and S. Chakravorti, A solvent-dependent lumines-
ed. J. F. Rabek, CRC Press, Boca Raton, USA, 1992, vol. 5.
30 A. Kawski and P. Bojarski, in Chromic Materials, Phenomena
cence study on 9-phenylcarbazole, J. Lumin., 1998, 78,
205–211.
and their Technological Applications, ed. P. R. Somani, 43 S. M. Bonesi and R. Erra-Balsells, Electronic spectroscopy of
Applied Science Innovations, 2009.
carbazole and N- and C-substituted carbazoles in homo-
31 A. Kawski, B. Kukliński and P. Bojarski, Dipole moment
geneous media and in solid matrix, J. Lumin., 2001, 93, 51–74.
of aniline in the excited S1 state from thermochromic 44 G. E. Johnson, Fluorescence quenching of carbazoles,
effect on electronic spectra, Chem. Phys. Lett., 2005, 415,
251–255.
32 A. Kawski, Ground- and excited-state dipole moments of
J. Phys. Chem., 1980, 84, 2940–2946.
45 M. Ravi, A. Samanta and T. P. Radhakrishnan, Excited state
dipole moments from an efficient analysis of solvatochro-
mic stokes shift data, J. Phys. Chem., 1994, 98, 9133–9136.
6-propionyl-2-(dimethylamino)naphthalene
determined
from solvatochromic shifts, Z. Naturforsch., 1999, 54a, 46 S. Kumar, V. C. Rao and R. C. Rostogi, Excited-state dipole
379–381.
moments of some hydroxycoumarin dyes using an efficient
solvatochromic method based on the solvent polarity para-
meter, ETN, Spectrochim. Acta, Part A, 2001, 57, 41–47.
47 A. Filarowski, M. Kluba, K. Cieślik-Boczula, A. Koll,
A. Kochel, L. Pandey, W. M. de Borggewaeve, M. van der
Auweraer, J. Catalán and N. Boens, Generalized solvent
scales as a tool for investigating solvent dependence of
spectroscopic and kinetic parameters. Application to fluor-
escent BODIPY dyes, Photochem. Photobiol. Sci., 2010, 9,
996–1008.
33 A. Kawski and Z. Gryczyński, On the determination of tran-
sition-moment directions from emission anisotropy
measurements, Z. Naturforsch., 1986, 41a, 1195–1199.
34 A. Kawski and Z. Gryczyński, Determination of the tran-
sition-moment directions from photoselection in partially
oriented systems, Z. Naturforsch., 1987, 42a, 808–812.
35 Z. Gryczyński and A. Kawski, Relation between the emis-
sion anisotropy and the dichroic ratio for solute alignment
on stretched polymer films, Z. Naturforsch., 1987, 42a,
1396–1398.
48 J. B. Birks, Photophysics of Aromatic Molecules, Wiley-Inter-
science, London, 1970.
36 Z. Gryczyński and A. Kawski, Directions of the electronic
transition moments of dioxydo-p-terphenyl, Z. Naturforsch., 49 A. Dreuw and M. Head-Gordon, Single-reference ab initio
1988, 43a, 193–195.
methods for the calcualtion of excited states of large mole-
37 A. Synak and P. Bojarski, Transition moment directions of
cules, Chem. Rev., 2005, 105, 4009–4037.
selected carbocyanines from emission anisotropy and 50 Z.-L. Cai, K. Sendt and J. R. Reamers, Failure of density-
linear dichroism measurements in unaxially stretched
polymer films, Chem. Phys. Lett., 2005, 416, 300–304.
38 J. Michl and E. W. Thulstrup, Spectroscopy with Polarized
Light, VCH Publ. Inc., New York, 1986.
39 Y. Tanizaki, T. Kobayashi and N. Ando, The relation
between the deformation of substrate of PVA sheet by
stretching and the stretch ratio-RS, Bull. Chem. Soc. Jpn.,
1959, 32, 1362–1363.
functional theory and time-dependent density-functional
theory for large extended π systems, J. Chem. Phys., 2002,
117, 5543–5549.
51 A. Dreuw and M. Head-Gordon, Failure of time-dependent
density functional theory for long-range charge-transfer
excited states: the zincbacteriochlorin-bacteriochlorin and
bacteriochlorophyll-spheroidene complexes, J. Am. Chem.
Soc., 2004, 126, 4007–4016.
40 Y. Tanizaki, Correction of relation of optical density ratio to 52 J. Andzelm, B. C. Rinderspacher, A. Rawlett, J. Dougherty,
stretch ratio on dichroic spectra, Bull. Chem. Soc. Jpn.,
1965, 38, 1798–1799.
41 http://avogadro.openmolecules.net/
R. Baer and N. Govind, Performance of DFT methods in
the calculation of optical spectra of TCF-chromophores,
J. Chem. Theor. Comput., 2009, 5, 2835–2846.
This journal is © The Royal Society of Chemistry and Owner Societies 2013
Photochem. Photobiol. Sci., 2013, 12, 284–297 | 297