Solvatochromic properties of 3,6-di-tert-butyl-8H-indolo[3,2,1-de]acridin- 8-one

Article abstract of DOI:10.1039/c4ra00787e

The spectral and photophysical properties of newly synthesized 3,6-di-tert-butyl-8H-indolo[3,2,1-de]acridin-8-one were studied. To determine the contribution of specific and nonspecific interactions of this compound with solvents, the solvatochromic metho

Full text of DOI:10.1039/c4ra00787e

RSC Advances
View Article Online
View Journal | View Issue
PAPER
Solvatochromic properties of 3,6-di-tert-butyl-8H-
indolo[3,2,1-de]acridin-8-one†
Cite this: RSC Adv., 2014, 4, 19310
Marlena Czerwinska, Małgorzata Wierzbicka, Katarzyna Guzow, Irena Bylinska
´
´
*
and Wiesław Wiczk
The spectral and photophysical properties of newly synthesized 3,6-di-tert-butyl-8H-indolo[3,2,1-de]-
acridin-8-one were studied. To determine the contribution of specific and nonspecific interactions of
this compound with solvents, the solvatochromic methods were used. It was found that bathochromic
shifts of absorption and fluorescence spectra were mostly caused by polarizability/dipolarity interactions
of the solute with solvent. The excited state dipole moment, determined based on a solvatochromic
method, is two times higher than that in the ground state. Also, the spectral and photophysical
properties of the compound studied correlate with the ENT solvent polarity scale. The radiative rate
constant does not depend on the solvent properties, while the non-radiative rate constant decreases
with increasing solvent polarity. Such a decrease of the non-radiative rate constant with the lowering of
the energy gap between the S0 and S1 states indicates the participation of the triplet state in the
deactivation of the excited state. After photoexcitation hydrogen bond energy increases in both the
locally excited Franck–Condon and the relaxed excited state, however, the change of hydrogen bond
energy is three times higher in the relaxed excited state. The hydrogen bond energy change measured in
Received 27th January 2014
Accepted 10th April 2014
DOI: 10.1039/c4ra00787e
www.rsc.org/advances
five alcohols correlates with the Kamlet–Taft or Catalan solvent acidity parameter.
´
It is commonly acknowledged that solvent-dependent UV-Vis
1 Introduction
spectral band shis can arise from either general and/or
specic solvent effects.9 The rst effect results from isotropic
interactions of the chromophore dipole moment with the
reaction eld induced in the surrounding solvent. Specic
effects result from the short-range anisotropic interactions
between the chromophore with one or more solvent molecules
in its rst solvation shell, an important example of which is the
formation of hydrogen bonds. It is thus of interest to quantify
the relative contributions from these two effects.10,11 The elec-
tronic transition in molecules, involving intramolecular charge
transfer, is very sensitive to the nature of the microenvironment
around the solute and the spectral parameters can be used to
study solute–solvent interactions at the microscopic level.12–15 In
the case of different electron densities in the electronic ground
and excited state of a light-absorbing molecule, its dipole
moment varies in these two states. Thus, a change of the solvent
affects the ground and excited state differently. The solvent-
dependent photophysical characteristic can be investigated by
steady-state absorption and emission spectroscopy as well as
time-resolved uorimetry. From these experiments, the position
of the spectra maxima, the Stokes shi, uorescence lifetime
and the rate constants of radiative and non-radiative deactiva-
tion of compound studied are determined12–15. The normalized
It is well known that natural and synthetic acridines/acridones
possess antitumor activity1–3 and DNA photo-damaging ability.4
Their biological activity is mainly attributed to the planarity of
these aromatic compounds, which enables intercalation within
the double-stranded DNA structure, thus interfering with the
cellular machinery. Besides biological activity, acridone and its
derivatives exhibit a high uorescence quantum yield and
characterized by a high photostability. For this reason, are used,
among others, as sensors for uorescence detection of DNA.5–8
Studied compound 3,6-di-tert-butyl-8H-indolo[3,2,1-de]acridin-
8-one belongs to the acridone group comprising many potential
drugs or uorescent probes. To determine whether it is suitable
as a uorescent sensor the factors inuencing photophysical
properties should be determined. To establish what type of
interactions are important ones in the solute–solvent interac-
tions, we performed a detailed study of the environmental
impact on spectral and photophysical properties of 3,6-di-tert-
butyl-8H-indolo[3,2,1-de]acridin-8-one.
´
´
Faculty of Chemistry, University of Gdansk, Wita Stwosza 63, 80-308 Gdansk, Poland.
E-mail: wieslaw.wiczk@ug.edu.pl
† Electronic supplementary information (ESI) available: Includes correlations of
spectral and photophysical parameters with Reichardt's solvent polarity
parameter ENT, luminescence spectra at 77 K as well as the correlation of the
difference and the sum of absorption and emission maxima versus Bilot–Kawski
polarity functions. See DOI: 10.1039/c4ra00787e
ENT polarity scale,16,17 Bilot–Kawski solvent polarizability func-
23
tions,18–21 and Kamlet–Ta22 and Catalan multi-parameters
´
solvent scales are used to describe the solvent effect on the
19310 | RSC Adv., 2014, 4, 19310–19320
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
RSC Advances
Jupiter® (C-18, 5 mm, 250 mm long, i.d. ¼ 4.6 mm, gradient 50–
100% B over 30 minutes, l ¼ 223 nm, 245 nm, 300 nm, tR ¼ 28.3
min.). The product was identied by means of the 1H and 1H–1H
COSY NMR (Bruker AVANCE III spectrometer (500 MHz) in
CDCl3), IR spectrum (Bruker IFS-66 instrument) and mass
spectrum (Bruker Biex III (MALDI-TOF)). m.p. 237–241 ꢁC; MS:
m/z ¼ 382.2 (M + H)+; IR (KBr): nmax (cmꢂ1) 2867.7–2953.9 CH;
1
1648.0 C]O; 1442.6–1484.0 C]C; 1144.6–1264.4 C(CH3)3; H
NMR: d (ppm) 1.537 (s, 9H, (CH3)3 (a)); 1.581 (s, 9H, (CH3)3 (b));
7.497–7.529 (m, 1H, C11H); 7.714 (dd, 1H, C3H, J ¼ 2.0 Hz, J ¼
8.8 Hz); 7.920–7.955 (m, 1H, C10H); 8.247–8.264 (m, 2H, C2H,
C8H); 8.495 (d, 1H, C9H, J ¼ 8.5 Hz); 8.532 (d, 1H, C5H, J ¼
2.0 Hz), 8.551 (d, 1H, C6H, J ¼ 2.0 Hz); 8.757 (dd, 1H, C12H, J ¼
1.5 Hz, J ¼ 8.0 Hz). In 2D NMR 1H–1H COSY spectrum the
following correlation signals were observed: the coupling of
C11H (d ¼ 7.5 ppm) with C10H (d ¼ 7.9 ppm) and C12H (d ¼ 8.8
ppm), the coupling of C10H (d ¼ 7.9 ppm) with C9H (d ¼ 8.5
ppm), the coupling of C3H (d ¼ 7.7 ppm) with C2H (d ¼
8.2 ppm). For hydrogen atoms C5H, C6H and C8H no correlation
signals were observed indicating on the absence of their
coupling with other hydrogen atoms present in the molecule.
Fig. 1 Scheme of 3,6-di-tert-butyl-8H-indolo[3,2,1-de]acridin-8-one
synthesis.
spectral and photophysical properties. The changes of the
hydrogen bond aer excitation are also addressed.10,24–28 In this
work, applying the solvatochromic methods the inuence of
solvents on the spectral and photophysical properties and
hydrogen bond energy changes aer photoexcitation as well as
multi-linear correlations of newly synthesized 3,6-di-tert-butyl-
8H-indolo[3,2,1-de]acridin-8-one (3, Fig. 1), are presented.
2.2 Spectroscopic measurements
2 Experimental
The UV/Vis absorption spectra in all 26 solvents studied were
measured using a Perkin-Elmer Lambda 40P spectrophotom-
eter, whereas emission spectra were measured using a Horiba-
Jobin Yvon FluoroMax-4 spectrouorimeter. Solvents of the
highest available quality were used. Fluorescence quantum
yields (F) were calculated with quinine sulphate in 0.5 M
H2SO4 (F ¼ 0.53 ꢃ 0.02) as reference and were corrected for
different refractive indices of solvents.31 In all uorimetric
measurements, the optical density of the solution does not
exceed 0.1. Luminescence spectra were measured in quartz
capillaries at liquid nitrogen temperature using a low-
temperature accessory. The uorescence lifetimes were
measured with a time-correlated single-photon counting
apparatus Edinburgh CD-900 equipped with NanoLed N16 (UV
LED l ¼ 339 nm) from IBH as the excitation source. The half-
width of the response function of the apparatus, measured
using a Ludox solution as a scatter, was about 1.0 ns. The
emission wavelengths were isolated using a monochromator.
Fluorescence decay data were tted by the iterative convolu-
2.1 Synthesis
Synthesis of 3,6-di-tert-butyl-8H-indolo[3,2,1-de]acridin-8-one
was performed according to the scheme (Fig. 1).
To obtain 3,6-di-tert-butyl-9H-carbazole (compound 1) and
9-phenyl-3,6-di-tert-butyl-9H-carbazole (compound 2) literature
procedures were applied.29 Both compounds were puried by
column chromatography (Merck, Silica gel 60, 0.040–0.063 mm,
eluent–CH2Cl2–petroleum ether 1 : 10 (v/v) (compound 1),
CH2Cl2–petroleum ether 1 : 5 (v/v) (compound 2)) and then
crystallized from boiling petroleum ether. Compound 1 was
obtained with 46% yield whereas compound 2 with 29% yield.
The title compound 3 was obtained applying procedure
described in ref. 30. To a stirred solution of 9-phenyl-3,6-di-tert-
butyl-9H-carbazole (compound 2) (0.09 g, 0.253 mmol) in 1,2-
dichlorobenzene (5 ml) cooled in an ice bath, 1,1-dichlor-
omethyl methyl ether (0.033 ml, 0.365 mmol) and TiCl4
(0.046 ml, 0.419 mmol) were added. The reaction was continued
for 1 hour in the ice bath. Then the reaction mixture was
allowed to warm up to the room temperature and poured on ice
(14.05 g) and concentrated HCl (0.63 ml). The organic layer was
diluted with chloroform, washed with a 5% solution of HCl (1ꢀ)
and water (3ꢀ) and dried with anhydrous Na2SO4. Aer evapo-
tion to the sum of exponents according to eq.:
X
IðtÞ
aiðꢂt=siÞ
(1)
i
ration the residue (yellow solid) was puried by column chro- where ai is the pre-exponential factor obtained from the uo-
matography (Merck, Silica gel 60, 0.040–0.063 mm) using a rescence intensity decay analysis and si the decay time of the i-th
mixture of CH2Cl2–petroleum ether 1 : 2 (v/v) as an eluent. component, using a soware supported by the manufacturer.
Additional purication was performed using RP-HPLC (Kro- The adequacy of the exponential decay tting was judged by
masil column, C-8, 5 mm, 250 mm long, i.d. ¼ 20 mm; linear visual inspection of the plots of weighted residuals as well as by
2
gradient 50–100% B over 120 minutes, A ¼ 0.1% water solution the statistical parameter cR and shape of the autocorrelation
of triuoroacetic acid, B ¼ 80% of acetonitrile in A) giving function of the weighted residuals and serial variance ratio
desired compound as a yellowish-orange solid (0.064 g, (SVR).
0.168 mmol, 66% yield). The purity of the obtained compound
Linear and multi-parametric correlations were performed
was checked by RP-HPLC using analytical column Phenomenex using Origin v. 9 soware.
This journal is © The Royal Society of Chemistry 2014
RSC Adv., 2014, 4, 19310–19320 | 19311

View Article Online
RSC Advances
Paper
More substantial changes are observed in emission spectra
(Fig. 3). In non-polar solvents the emission spectra possess a
maximum in the range 24 450–23 580 cmꢂ1 with weakly marked
vibrational structure. An increase of solvent polarity results in
bathochromic shi of the spectrum and loss of its vibrational
structure. Like for absorption spectra, a major bathochromic
shi of emission band is observed in alcohols. The larger
proton-donor properties of alcohol, the greater shi is observed
(position of the emission band maximum is located at
22 200 cmꢂ1 for 2-propanol whereas at 19 800 cmꢂ1 for HFIP).
The spectral and photophysical properties of compound
studied are gathered in Table 1.
Solvent properties inuences uorescence quantum yield
and uorescence lifetime of the studied compound (Table 1).
Larger uorescence quantum yields are observed in alcohols
(up to 0.69 for HFIP) and polar solvents whereas for non-polar
solvents these values are relatively small (about 0.01 for hexane).
Similarly, uorescence lifetimes are longer for protic and polar
solvents (about 6.7 ns for HFIP) and much shorter for non-polar
solvents (0.98 ns for hexane). It should be noted that in all
solvents studied the uorescence intensity decay is mono-
exponential, except ethyl ether in which uorescence intensity
decay of 3,6-di-tert-butyl-8H-indolo[3,2,1-de]acridin-8-one is bi-
exponential indicating on its specic interaction with this
solvent.
3 Results and discussion
3.1 Spectral properties
Absorption spectra of the title compound in selected solvents
are presented in Fig. 2, whereas the emission spectra in Fig. 3.
In non-polar solvents, saturated hydrocarbons and 1-chlor-
oalkanes, absorption spectrum possesses a maximum at about
24 700–23 580 cmꢂ1. The spectra in these solvents have small
half-width and the most intense is the rst vibronic band with
poorly separated vibrational structure on the short-wavelength
side of the spectrum. The increase of solvent polarity causes
bathochromic shi of the spectrum and the simultaneous loss
of the vibrational structure. In protic solvent, the long-wave-
length absorption band takes bell shape form and the position
of its maximum shis with increasing proton-donor capacity of
the solvent. However, the bathochromic shi of the absorption
band is not substantial and ranges from 24 690 cmꢂ1 for
cyclohexane to 24 100 cmꢂ1 in iPrOH. A larger bathochromic
shi is observed for TFE (2,2,2-triuoroethanol) (23 700 cmꢂ1
and HFIP (1,1,1,3,3,3-hexauoro-2-propanol) (22 990 cmꢂ1).
)
3.2 Correlation with Reichardt's ENT solvent polarity scale
parameter
The various models of solvation, several empirical solvent
polarity parameters have been proposed to characterize and
rank empirically the polarity of a solvent.16,17,22,23 One of the
most popular solvent polarity scale is the ET(30) scale, based on
Vis spectroscopic band shis of a negatively solvatochromic
standard probe,16,17 or the normalized ENT parameter dened
according to eqn (2):
ETðsolventÞ ꢂ ETðTMSÞ
ETN
¼
(2)
Fig. 2 Absorption spectra of 3,6-di-tert-butyl-8H-indolo[3,2,1-de]-
acridin-8-one in selected solvents.
ETðwaterÞ ꢂ ETðTMSÞ
where TMS denotes tetramethylsilane. They are mostly used to
correlate the absorption or emission transition energies and
Stokes shis.13–15,32,33
As mentioned above, the maxima of absorption spectra
shi to the red with increasing solvent polarity. The plot of
dependence of wavenumber of maximum of absorption
spectrum versus ENT (Fig. S1 ESI†) reveals rather week corre-
lation (r2 ¼ 0.8812, N ¼ 21). Main reason for the lack of good
correlation is large deviations observed for alcohols, espe-
cially for TFE and HFIP, but also for other solvents such as
toluene or DMSO. This indicates on a specic solute–solvent
interactions. Separate analyses of the data for the protic and
non-protic solvents do not improve the correlation. For non-
protic solvents correlation coefficient is r2 ¼ 0.7349 for 16
solvents whereas for protic ones r2 ¼ 0.6745 (N ¼ 5) (Fig. S1
ESI†). A better correlation with Reichardt's solvent polarity
parameter (ENT) (r2 ¼ 0.9554, N ¼ 21) was obtained for the
position of emission maximum (Fig. S2 ESI†). However, the
Fig.
3 Emission spectra of 3,6-di-tert-butyl-8H-indolo[3,2,1-de]-
acridin-8-one in selected solvents.
19312 | RSC Adv., 2014, 4, 19310–19320
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
RSC Advances
Table 1 Spectroscopic and photophysical properties of 3,6-di-tert-butyl-8H-indolo[3,2,1-de]acridin-8-one in solvents used
Nr
Solvent
nabs [cmꢂ1
]
nem [cmꢂ1
]
Dn [cmꢂ1
]
4
s [ns]
kf ꢀ 10ꢂ7 [sꢂ1
]
knr ꢀ 10ꢂ7 [sꢂ1
]
1
2
3
4
5
6
7
8
MeOH
EtOH
iPrOH
TFE
HFiP
Hexane
24 038
24 038
23 981
23 889
23 202
24 752
24 752
24 630
24 610
24 691
24 331
24 501
24 510
24 510
24 510
24 510
24 510
24 390
24 331
24 450
24 450
24 510
24 630
21 598
21 882
22 222
20 790
20 000
24 450
24 331
24 253
24 301
24 331
23 640
23 584
23 584
23 697
23 753
23 780
23 864
23 148
22 831
23 364
23 640
23 471
23 302
2440
2156
1759
3019
3202
302
421
277
309
360
691
926
926
813
757
730
646
0.469
0.445
0.396
0.532
0.689
0.010
0.010
0.034
0.047
0.010
0.251
0.164
0.127
0.112
0.103
0.056
0.034
0.226
0.253
0.229
0.149
0.134
0.272
5.71
5.20
4.59
6.09
6.69
0.98
0.98
0.85
0.82
0.91
1.12
1.47
1.37
1.11
1.07
0.96
0.78
2.28
3.21
1.52
1.35
1.17
0.36 (a ¼ 0.46)
1.44 (a ¼ 0.54)
1.41
8.28
8.56
8.64
8.73
10.30
0.98
1.04
4.02
5.79
1.11
22.42
11.14
9.29
10.09
9.59
5.85
4.38
9.92
7.87
15.06
11.04
11.48
—
9.30
10.67
13.15
7.69
4.65
101.06
101.00
113.62
116.16
108.78
66.87
56.89
63.70
80.00
83.87
98.31
123.82
33.94
23.29
50.73
63.03
73.99
—
Isooctane
Hexadecane
Squalene
Cyclohexane
Toluene
1-Chloropropane
1-Chlorobutane
1-Chlorohexane
1-Chlorooctane
1-Chlorodecane
1-Chlorohexadecane
Acetone
Acetonitrile
Dioxane
THF
MeTHF
Et2O
9
10
11
12
13
1`4
15
16
17
18
19
20
21
22
23
1242
1500
1086
810
1039
1328
24
25
26
AcOEt
DMF
DMSO
24 510
24 272
24 096
23 584
22 831
23 256
926
1441
840
0.134
0.325
0.388
9.48
11.18
11.49
61.44
23.18
18.10
2.91
3.38
experimental points for non-protic solvents are grouped in
three clusters of points which give bad correlation coefficient
(r2 ¼ 0.7645, N ¼ 16) with ENT , in opposite to protic solvents
(r2 ¼ 0.9823, N ¼ 5). The Stokes shi correlates well with the
ENT parameter aer the rejection of the points corresponding
to ethyl ether and DMSO, (r2 ¼ 0.9829, N ¼ 19) and there is no
reason to analyse separately protic and non-protic solvents
(Fig. S3 ESI†). Also, the dependence of uorescence quantum
yield on ENT parameter does not give rise to a division of
solvents into two groups, however, the correlation coefficient
is rather small (r2 ¼ 0.8848, N ¼ 21) due to the large deviation
from a straight line observed for polar solvents (Fig. S4 ESI†).
The dependence of uorescence lifetime can be described as a
straight-line dependence on ENT parameter with good corre-
lation coefficient (r2 ¼ 0.9594, N ¼ 20, Fig. S5 ESI†). However,
it can be seen that the uorescence decay time values
measured in alcohols clearly differ in a systematic way from
this relationship. A separate analysis of uorescence lifetime
dependence on ENT for protic and non-protic solvents (Fig. S5
ESI†) gave a worse value of correlation coefficient for non-
protic solvents (r2 ¼ 0.8858, N ¼ 15), however, a better
correlation for protic solvents (r2 ¼ 0.9817, N ¼ 5).
kf ¼ F/s
knr ¼ (1 ꢂ F)/s
(3)
(4)
Therefore, an analysis of these rate constants as a function of
ENT was performed to receive information which of them and
how depend on the solvent parameters, enabling to estimate
which path of the excited state deactivation predominates.
It was found that the radiative rate constant (kf) poorly depends
on ENT (r2 ¼ 0.2319, N ¼ 15) (Fig. S6 ESI†) even aer exclusion of
the data corresponding to hydrocarbon solvents. Very large devi-
ation from correlation line of kf values in hydrocarbon solvents
may indicate that in these solvents the uorescence occurs from a
different state than in the other ones.
In the case of dependence of non-radiative rate constant (knr)
on ENT parameter, separate analyses for the protic and non-protic
solvents have to be performed (Fig. S7 ESI†). In both cases, the
increase of solvent polarity causes a decrease of non-radiative
rate constant as indicated by negative slope ꢂ(3.63 ꢃ 0.27), (r2 ¼
0.9334, N ¼ 15) for non-protic solvents and ꢂ(1.89 ꢃ 0.19), (r2 ¼
0.9695, N ¼ 5) for protic ones. Thus, the solute–solvent inter-
actions with polar environment decrease the non-radiative rate
constant of 3,6-di-tert-butyl-8H-indolo[3,2,1-de]acridin-8-one.
Moreover, this process is slower in alcohols than in non-protic
solvents. The strong increase of knr upon solvent polarity can be
rationalized if knr can be attributed mainly to internal conver-
sion (the energy gap lowering).15,29 In the opposite case, the
main path of the excited state deactivation is intersystem
crossing to the triplet state.30
Both uorescence quantum yield and uorescence lifetime
are complex function of uorescence and non-radiative rate
constant. For the mono-exponential uorescence intensity decay,
the radiative (kf) and non-radiative (knr) rate constants can be
calculated from the measured uorescence quantum yield (F)
and uorescence lifetime (s) according to eqn (3) and (4):
This journal is © The Royal Society of Chemistry 2014
RSC Adv., 2014, 4, 19310–19320 | 19313

View Article Online
RSC Advances
Paper
According to Englman and Jortner34 the weak coupling limit spectrum (Fig. S9 ESI†). In a weakly polar and polar solvents
(the relative horizontal displacement of the two potential energy (MeTHF, C3H7Cl, DMF) well separated uorescence and phos-
surfaces is small) reveals an exponential dependence of the phorescence spectra are observed (Fig. 5 and S10 ESI†).
transition probability on the energy gap DE. The dependence of
It is worth mentioning that the phosphorescence spectrum
non-radiative rate constant on the energy gap between two with maximum at 20 400 cmꢂ1 and well-resolved vibrational
electronic states can be roughly estimated from equation:35
structure practically does not depend on solvent polarity,
contrary to the uorescence spectrum. Additionally, the ratio of
phosphorescence to uorescence intensity decreases with an
increase of solvent polarity (Fig. 5). To explain such a relation-
ship the interdependence of the internal conversion and
intersystem crossing rate constants should be examined. Radi-
ationless rate constant (knr) is the sum of internal conversion
(kIC) and intersystem crossing (kISC) rate constants. The
observed decrease of knr with decrease of (S1 ꢂ S0) energy gap
can be rationalized assuming decrease of intersystem rate
constant. One can assume that the excited state lies over two (n,
p*) and (pp*) states. Two scenarios are possible. In the rst
one, a decrease of the energy gap between S1 and S0 states at the
same time lowers the energy of S1(pp*) state below the T(n, p*)
state so that it is located between triplet states. According to El-
Sayed rule,36 the rate constant of intersystem crossing (kISC) is
greater for S1(pp*) / T(n, p*) than for S1(pp*) / T(pp*)
transition. Thus, as a result of the decrease of the energy of
excited singlet, the intersystem crossing between S1(pp*) /
T(pp*) state takes place resulting in the decrease of intersystem
rate constant and therefore an increase in the uorescence
intensity and decrease in the phosphorescence intensity is
observed. Based on such a mechanism the differences in the
photophysics of 2-nitrouorene with that of 2-diethylamino-7-
nitrouorene was explained.37 A second possible mechanism
may result from the reduction of Franck–Condon factor in the
S1 / T1 intersystem crossing process which is the result of the
increase of the ICT participation in the S1 state (Fig. S11 ESI†).
In both of these explanation one assumes no effect or little
effect of the solvent polarity on the internal conversion rate
constant. Based on collected data, the second mechanism
seems to be more probable.
max
fluo
log(kIC) z 12 ꢂ 2n
(5)
where: kIC is the rate constant of internal conversion expressed
in sꢂ1 and n
expressed in mmꢂ1. As can be seen from the
max
uo
formula (5) the bathochromic shi of uorescence spectrum
causes the increase of the non-radiative rate constant. In the
case studied, a bathochromic shi of emission spectra is
observed with increase of solvent polarity, however, the
decrease of the non-radiative rate constant is recorded. An
inverse relationship between non-radiative rate constant (knr)
and the environment polarity and thereby decreasing the energy
gap between relaxed excited state and ground state (Fig. 4) is in
opposition to that founded in literature.15,29 The plot of ln(knr) as
a function of energy gap between S1 and S0 states (DE(S1 ꢂ S0) ¼
max
n
) gives two straight lines (Fig. 4).
uo
A slope is equal to (1.58 ꢃ 0.22) ꢀ 10ꢂ3, (r2 ¼ 0.8125, N ¼ 14,
without hydrocarbon solvents which form
a
separate
cluster of points) for non-protic solvents and (4.33 ꢃ 0.44) ꢀ
10ꢂ4, (r2 ¼ 0.7969, N ¼ 5) for protic ones.
To further investigate the dependence of non-radiative rate
constant on solvent polarity, luminescence spectra of studied
compound were measured in several solvents. In non-polar
methylcyclohexane (MeCx), apart from weak uorescence
spectrum shied to the red compared to the spectrum
measured at room temperature, a strong phosphorescence
spectrum is observed (Fig. S8 ESI†). This indicates that the main
deactivation route of the excited state is the intersystem
crossing to the triplet state, hence, low uorescence quantum
yield and short uorescence lifetime observed in saturated
hydrocarbon solvents. The situation is different in protic
solvents. In methanol glass only weak phosphorescence is
visible as a shoulder on the red-side of the uorescence
Thus, it seems that efficient intersystem crossing to a triplet
state determines photophysical properties of 3,6-di-tert-butyl-
8H-indolo[3,2,1-de]acridin-8-one (low uorescence quantum
Fig. 4 Correlations between the logarithm of the non-radiative rate Fig.
5 The luminescence spectra of 3,6-di-tert-butyl-8H-indolo
max
fluo
constant ln(knr) and DE(S1 ꢂ S0) ¼ n
energy gap for 3,6-di-tert- [3,2,1-de]acridin-8-one (normalized to the same fluorescence inten-
butyl-8H-indolo[3,2,1-de]acridin-8-one in studied solvents.
sity) in C3H7Cl, MeTHF and DMF measured at 77 K.
19314 | RSC Adv., 2014, 4, 19310–19320
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
RSC Advances
yield and short uorescence lifetime) in non-polar and weakly (6.5%) have a minor inuence. The same conclusion can be
polar solvents.
drown from Kamlet–Ta correlation. Two parameters, solvent
dipolarity/polarizability (p*) and solvent acidity (a), inuence
equally (in about 50%) the bathochromic shi of absorption
spectra, while solvent basicity can be omitted because of low
value and large error exceeding the value of the parameter (the
quality of t decreases a little to r2 ¼ 0.9139 comparing to the
original t r2 ¼ 0.9231). Good quality of ts were obtained in
multi-linear correlations of the position of uorescence
3.3 Multi-parametric correlations
The multiple linear regression, approach of Kamlet and
coworkers22 and Catalan, has been used to correlate UV-Vis
23
´
absorption and emission energies14,15,38 as well as another
photophysical properties with an index of the solvent dipolarity/
polarizability which is a measure of the solvent's ability to
stabilize a charge or dipole through nonspecic dielectric
interactions (p* or SP and solvent dipolarity SdP in a new
2
2
´
maximum (r ¼ 0.9510 for Catalan and r ¼ 0.9731 for Kamlet–
Ta correlation). In both cases, solvent basicity has minor
inuence on bathochromic shi of uorescence spectrum
(cSB z 12%), whereas the solvent acidity (dSA z 74%) and
solvent dipolarity (bSdP z 26%) have major inuence on it.
´
generalized solvent polarity Catalan scale), and indices of the
solvent's hydrogen-bond donor strength (a or SA) and
hydrogen-bond acceptor strength (b or SB) according to eqn (6)
and (7):
´
Solvent polarizability in the Catalan equation may be dis-
regarded because of its large error without substantial
decrease of the quality of t (r2 ¼ 0.9500 compared to the
original t r2 ¼ 0.9510). A similar results were obtained from
Kamlet–Ta correlation (r2 ¼ 0.9731). A bathochromic shi of
uorescence spectrum is caused mainly by solvent dipolarity/
polarizability (35%) and solvent acidity (55%), while solvent
basicity has a minor inuence (10%). It is worth mentioning
that the correlation parameters for both absorption and emis-
sion spectra are negative meaning that their changes cause red-
shi of the spectra. Additionally, correlation parameters
obtained for uorescence spectra are larger, taking into account
their absolute values, than those calculated for absorption
spectra. This indicates that the electronic structure of the locally
excited Franck–Condon and relaxed excited states of 3,6-di-tert-
butyl-8H-indolo[3,2,1-de]acridin-8-one differ signicantly, and
the relaxed excited state is more polar.
y ¼ y0 + ap*p + baa + cbb
(6)
or
y ¼ y0 + aSPSP + bSdPSdP + cSASA + dSBSB
(7)
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; p*, a, b, SP, SdP,
SA, SB represent independent solvent parameters accounting
for various types of solute–solvent interactions; ap*, ba, cb and
aSP, bSdP, cSA and dSB are adjustable coefficients that reect the
sensitivity of physical property y in a given solvent to the various
solvent parameters.
´
The results of multi-linear correlation according to Catalan
Correlation of the Stokes shi with the four-parameter
solvent scale (r2 ¼ 0.9311) reveals that the increase of solvent
polarizability decreases the Stokes shi (a negative value of
adjustment parameter) in opposite to the remaining three
parameters. The biggest contributions to the changes of Stokes
shi have solvent polarizability (41%) and solvent acidity (37%),
whereas the solvent dipolarity (15%) and solvent basicity (7%)
have minor inuence on a Stokes shi. A good correlation of
the Stokes shi was also obtained using Kamlet–Ta equation
(r2 ¼ 0.9241). In this case, two parameters, solvent polariz-
ability/polarity (41%) and solvent acidity (59%), play a crucial
and Kamlet–Ta equations are presented in Tables 2 and 3.
Both solvent polarity scales predict a bathochromic shi of
absorption spectra with increase of solvent polarity (all coeffi-
cients in eqn (6) and (7) are negative). A fair t was obtained in
2
2
´
both cases (r ¼ 0.9190 for Catalan and r ¼ 0.9231 for Kamlet–
´
Ta correlation, respectively). In the Catalan equation all four
coefficients are important. However, the large values of aSP
(51.5%) and dSA (33%) coefficients compared to the others
indicate that the changes of nabs may reect primarily a change
in polarizability and acidity of the environment of the chro-
mophore, whereas the solvent dipolarity (9%) and basicity
Table 2 Estimated coefficients (y0, aSP, bSdP, cSB, dSA), their errors and correlation coefficients (r2) for multi-linear correlation analysis of nabs, nfluo
f, s, Dn, kf, knr of compound as a function of the Catalan four-parameter solvent scale (eqn (7))
,
´
y0
aSP
bSdP
cSB
dSA
r2
N
nabs [cmꢂ1
nuo [cmꢂ1
]
]
25 601 ꢃ 257
ꢂ(1355 ꢃ 376)
608 ꢃ 1087
—
ꢂ(2007 ꢃ 978)
0.566 ꢃ 0.273
0.113 ꢃ 2.378
—
5.262 ꢃ 2.010
3.945 ꢃ 1.412
ꢂ(0.426 ꢃ 0.652)
—
ꢂ(229 ꢃ 84)
ꢂ(172 ꢃ 109)
ꢂ(450 ꢃ 316)
ꢂ(451 ꢃ 309)
318 ꢃ 284
ꢂ(873 ꢃ 119)
ꢂ(2711 ꢃ 343)
ꢂ(2711 ꢃ 343)
1777 ꢃ 308
0.9190
0.9510
0.9500
0.9311
0.8879
0.9368
0.9368
0.4552
0.4389
0.9017
0.8987
20
20
23 836 ꢃ 743
24 248 ꢃ 101
1774 ꢃ 668
ꢂ(973 ꢃ 243)
ꢂ(945 ꢃ 232)
752 ꢃ 218
Dn [cmꢂ1
]
20
20
19
f
ꢂ(0.347 ꢃ 0.186)
0.626 ꢃ 1.160
0.702 ꢃ 0.205
ꢂ(3.020 ꢃ 1.382)
ꢂ(2.130 ꢃ 0.986)
1.345 ꢃ 0.447
1.056 ꢃ 0.057
0.122 ꢃ 0.061
1.105 ꢃ 0.508
1.108 ꢃ 0.487
0.488 ꢃ 0.429
0.581 ꢃ 0.243
ꢂ(0.630 ꢃ 0.139)
ꢂ(0.641 ꢃ 0.136)
0.163 ꢃ 0.079
1.404 ꢃ 0.686
1.407 ꢃ 0.659
ꢂ(0.158 ꢃ 0.580)
—
0.497 ꢃ 0.086
5.280 ꢃ 0.719
5.258 ꢃ 0.534
0.616 ꢃ 0.607
—
s ꢀ 109 [ns]
kf ꢀ 10ꢂ8 [sꢂ1
]
19
19
knr ꢀ 10ꢂ9 [sꢂ1
]
ꢂ(0.286 ꢃ 0.188)
ꢂ(0.511 ꢃ 0.197)
ꢂ(0.299 ꢃ 0.183)
ꢂ(0.429 ꢃ 0.149)
This journal is © The Royal Society of Chemistry 2014
RSC Adv., 2014, 4, 19310–19320 | 19315

View Article Online
RSC Advances
Paper
Table 3 Estimated coefficients (y0, ap*, ba, cb), their errors and correlation coefficients (r2) for multi-linear correlation analysis of nabs, nfluo, f, s,
Dn, kf, knr of compound as a function of the Kamlet–Taft solvent scale (eqn (6))
y0
ap*
ba
cb
r2
N
nabs [cmꢂ1
]
24 720 ꢃ 50
24 718 ꢃ 50
22 224 ꢃ 94
24 205 ꢃ 97
478 ꢃ 113
ꢂ(491 ꢃ 119)
ꢂ(505 ꢃ 90)
ꢂ(492 ꢃ 48)
ꢂ(489 ꢃ 45)
ꢂ(1758 ꢃ 92)
ꢂ(1719 ꢃ 97)
1258 ꢃ 110
1216 ꢃ 108
0.238 ꢃ 0.023
2.917 ꢃ 0.177
ꢂ(1.812 ꢃ 1.735)
—
ꢂ(20 ꢃ 107)
—
ꢂ(306 ꢃ 203)
0.9231
0.9139
0.9731
0.9693
0.9241
0.9160
0.9306
0.9638
0.4704
0.4070
0.9441
20
20
20
nuo [cmꢂ1
]
ꢂ(1122 ꢃ 227)
ꢂ(1334 ꢃ 184)
630 ꢃ 272
Dn [cmꢂ1
]
231 ꢃ 244
—
0.097 ꢃ 0.052
1.683 ꢃ 0.403
ꢂ(4.370 ꢃ 3.965)
—
497 ꢃ 115
852 ꢃ 218
f
0.027 ꢃ 0.02
0.764 ꢃ 0.188
3.509 ꢃ 1.850
3.321 ꢃ 1.834
10.444 ꢃ 0.444
0.248 ꢃ 0.058
0.944 ꢃ 0.454
15.023 ꢃ 4.463
11.103 ꢃ 3.251
ꢂ(6.786 ꢃ 1.073)
20
19
19
s ꢀ 109 [ns]
kf ꢀ 10ꢂ7 [sꢂ1
]
knr ꢀ 10ꢂ8 [sꢂ1
]
ꢂ(3.110 ꢃ 0.417)
ꢂ(3.627 ꢃ 0.953)
19
role. Because of a small value and large error, solvent basicity adjustment coefficients negative. The larger decrease of knr is
can be omitted without deterioration of the quality of t (r2 ¼ caused by solvent polarizability/polarity (44% and 50% for
0.9160 whereas for original t r2 ¼ 0.9241).
Catalan and Kamlet–Ta correlation, respectively), solvent
´
The correlation of the uorescence quantum yield with acidity (36% and 23%) and solvent basicity (20% and 27%). It is
´
SP, SdP, SA, SB as independent variables gives a rather good t worth mentioning that solvent polarizability in Catalan solvent
(r2 ¼ 0.8879, Table 2). Moreover, it indicates that all solvent scale has no impact on knr. Moreover, a good correlation
parameters inuence the uorescence quantum yield. Solvent between knr and solvent acidity parameter (a and SA) for alco-
acidity (37%) and solvent dipolarity (42%) increases the uo- hols were found (Fig. S12 and S13 ESI†). Thus, the invariance of
rescence quantum yield to a similar degree, while solvent kf and dependence of knr on the solvent properties indicate that
dipolarity and basicity have only a little impact on it (9% and the change of the uorescence quantum yield and the uores-
12%, respectively). The same conclusion arises from the Kam- cence decay time with solvent properties is mainly due to the
let–Ta correlation (r2 ¼ 0.9306). Solvent polarizability/dipo- change of the knr rate constant.
larity and solvent acidity affect almost equally (42% and 41%,
respectively) while solvent basicity in a much lesser extent work, Reichardt's empirical one is a useful and easy to apply,
(17%). but because it is a one-parameter scale, it does not reect the
Among three different solvent polarity scales used in this
A good t was also obtained for the uorescence lifetime complexity of solute–solvent interactions. However, it allows to
2
´
correlation with both Catalan (r ¼ 0.9386) and Kamlet–Ta easily determine whether the solvent studied shows a deviation
(r2 ¼ 0.9638) equations. The largest impact on the increase of from linearity due to specic interactions. Its additional
uorescence lifetime has solvent acidity (68% and 53% for advantage is possibility to estimate the dipole moment change
Catalan and Kamlet–Ta correlation, respectively) whereas in the excited state based on the Stokes shi dependence on ENT.
´
solvent dipolarity contributes with 14% and 17%, respectively. Another two solvent polarity scales are more useful as they are
Moreover, the contribution of solvent basicity is different for multi-parametric. They allow to estimate the impact of various
these two solvent scale. For Kamlet–Ta scale solvent basicity solvent parameters on studied quantity. For this reason a four-
´
contribution (30%) is almost twice of that of Catalan scale (18%). parameter scale has advantage over the Kamlet–Ta scale
The decay of the uorescence intensity of 3,6-di-tert-butyl- because it allows to assess independently the impact of both the
8H-indolo[3,2,1-de]acridin-8-one can be described as a mono- polarizability and dipolarity, however, both offer comparable
exponential function, except for ethyl ether solution, thus, quality of the t. In the case of multi-parametric solvents scales,
further analysis including the dependence of uorescence life- better quality of the ts are obtained than for Reichardt's due to
time, uorescence rate constant and radiationless rate constant the greater number of variables. Thus, in order to evaluate
on solvent polarity parameters can be performed. Analysis of which scale used gives a better quality of the t, the adjusted r2
the inuence of the solvent on the radiative and non-radiative rather than coefficient of determination (r2) should be applied.
rate constants gives information which one depends on the The explanation of adjusted r2 is almost the same as r2 but it
solvent parameters enabling to estimate which path of the penalizes the statistics as extra variables included in the model.
excited state deactivation predominates. The multi-linear
analysis of kf according to eqn (6) and (7) revealed that it does
2
3.4 Specic interactions
´
not correlate with solvent parameters (r ¼ 0.4389 for Catalan
and r2 ¼ 0.4070 for Kamlet–Ta correlation, respectively) A variety of solvent interactions with solute may be responsible
(Tables 2 and 3), however, it depends exclusively on solvent for the solvent effect on the properties of solute. The solvent
polarizability/polarity. In contrast to kf, non-radiative rate effect can be split into two different type of interactions: specic
2
´
constant knr gives good correlations (r ¼ 0.8987 for Catalan and involving charge-transfer interactions, acid–base interactions,
r2 ¼ 0.9441 for Kamlet–Ta equation, respectively) with all p–p interactions and hydrogen bonding, while those
19316 | RSC Adv., 2014, 4, 19310–19320
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
RSC Advances
interactions which arise from the solvent acting as a dielectric approximation that the polarizability of the solute is equal
continuum are referred to as non-specic interactions. The 2a/4p30a3 ¼ 1 is as follows:
classication of solvents with respect to their polarity/polariz-
ꢂ
ꢃ
2n2 þ 1 3 ꢂ 1 n2 ꢂ 1
fBKð3; n2Þ ¼
ꢂ
(10)
ability and hydrogen bonding ability are based on the Kamlet–
n2 þ 2 3 þ 2 n2 þ 2
´
Ta or Catalan solvent polarity scale. Solvents capable to form
hydrogen bond have been classied as hydrogen-bond accep-
tors, e.g. possessing ability to accept hydrogen atom from solute
If the polarizability of the solute is neglected (a ¼ 0), Lippert–
´
(b parameter in Kamlet–Ta scale or SB parameter in Catalan
Mataga in the form below is used:
scale), and hydrogen-bond donor, e.g. possessing ability to
donate hydrogen atom to solute to form hydrogen bond
3 ꢂ 1
n2 ꢂ 1
fLMð3; n2Þ ¼
ꢂ
(11)
23 þ 1 2n2 þ 1
´
(a parameter in Kamlet–Ta scale or SA parameter in Catalan
scale). The protic solvents mostly used in solvatochromic
studies have both hydrogen-donor and hydrogen-acceptor
properties. In solvatochromic studies, the shi of maximum of
absorption or emission spectra depends on both non-specic
interactions and solute–solvent hydrogen-bonds formation.
In a series of publications, Maciejewski and co-workers10,25–27
gave a method to extract the contribution of specic interac-
tions between solute and hydrogen-bonding solvents based on
solvatochromic data. The rst step in this method requires
determination of the contribution of non-specic interactions
based on solvatochromic shi of absorption and emission
spectra of compound studied. The solvent used for the esti-
mation of non-specic interactions are 1-chloro-n-alkanes.10
These solvents interact only non-specically because they do
not have p-electron, and charge-transfer character, and do not
form hydrogen bond. The straight-line obtained from sol-
vatochromic studies of solute in 1-chloro-n-alkanes describes
the effect of non-specic interactions on position of maximum
of absorption and emission spectra not only in these solvents
but also in other solvents that make hydrogen bond. The
distance between nabs or nuo value and the straight line
describes the changes in the energy of hydrogen bond in
Franck–Condon excited state (from nabs dependence) or relaxed
excited state (from nuo dependence) in a given solvent accord-
ing to equation (eqn (8)):10
In the papers published by Maciejewski et al.10,25–27 dedi-
cated to the calculation of energy changes of hydrogen bonds
in the excited state, the Lippert–Mataga polarizability func-
tion was used. We decided to use the Bilot–Kawski polariz-
ability function. It is worth noting that the function selection
has practically no effect on the calculated values of the
hydrogen bond energy changes, however, Bilot–Kawski func-
tion has a wider range of changes, which allows a more
accurate approximation of nonspecic interactions by a linear
approximation. The calculated changes of energy of hydrogen
bond aer excitation to excited (locally excited Franck–Con-
don and relaxed) state of 3,6-di-tert-butyl-8H-indolo[3,2,1-de]-
acridin-8-one in ve used alcohols are presented in Table 4
and Fig. 6 and 7. As can be seen from aforementioned gures
max
a good correlation were obtained for the dependence of n
abs
max
and n
versus solvent polarizability function. The low value
uo
max
of the slop of n
vs. fBK(3, n2) indicates that the dipole
abs
moment of compound studied is nearly the same in S0 and
FC
S1 state, however, in the relaxed excited state the dipole
moment increases (higher negative slope of the dependence
max
uo
of n
vs. fBK(3, n2)). Also, like for coumarins,24 blue shi is
observed for saturated hydrocarbons in both absorption and
emission spectra. The deviation of these points from the line
formed by 1-chloro-n-alkanes corresponds to the energy of
interactions between permanent dipole moment of 1-chloro-
n-alkane solvents and permanent and induced dipole
moment of 3,6-di-tert-butyl-8H-indolo[3,2,1-de]acridin-8-one,
while saturated hydrocarbon solvents can interact only
through dispersion interaction.26
abs
shift
max
abs
max
abs
Dn ðH-bondÞ ¼ n ð f ð3; n2ÞÞ ꢂ n ðexpÞ
(8)
fluo
max
fluo
max
fluo
Dn ðH-bondÞ ¼ n ð f ð3; n2ÞÞ ꢂ n ðexpÞ
shift
where: nmabasx(exp) and n (exp) are experimental positions of the
max
uo
maxima of absorption and uorescence spectra of the solute in
the solvent forming the hydrogen bond, nmabasx( f (3, n2)) and
max
uo
n
( f (3, n2)) are the values of predicted maxima of absorption
and uorescence spectra of the solute in this solvent based on
the non-specic interaction only (from straight line of sol-
vatochromic plot) and f(3, n2) is the orientation polarizability
function. The orientation polarizability function for a spherical
cavity radius can be calculated according to various formulas.11
For a linear Stark effect and neglecting solute polarizability
(a ¼ 0) the MacRae function is used (eqn (9)):
Table 4 Hydrogen bond energy changes after photoexcitation of 3,6-
di-tert-butyl-8H-indolo[3,2,1-de]acridin-8-one in alcohols calculated
based on solvatochromic plot of absorption and emission spectra
Kamlet–Ta (a)
´
and Catalan (SA)
solvent H-bond
donating parameter
DE (H-bond) [cmꢂ1
]
3 ꢂ 1 n2 ꢂ 1
ꢀ
ꢁ
FC
relax
fMR 3; n2
¼
ꢂ
(9)
Solvent
S1
S1
a
SA
3 þ 2 n2 þ 2
iPrOH
EtOH
MeOH
TFE
522
496
463
612
1300
1193
1491
1744
2570
3421
0.76
0.86
0.98
1.51
1.96
0.28
0.40
0.61
0.90
1.01
Solvent polarizability function for spherical cavity with the
radius (a) obtained from Bilot–Kawski theory in which linear
and quadratic Stark effect was taken into account and with
HFIP
This journal is © The Royal Society of Chemistry 2014
RSC Adv., 2014, 4, 19310–19320 | 19317

View Article Online
RSC Advances
Paper
relaxed excited state (a larger changes in the position of
emission spectra) than for the locally excited Franck–Condon
state.
Moreover, the quality of t is better using Kamlet–Ta
´
solvent acidity parameter (a) than Catalan parameter (SA).
3,6-Di-tert-butyl-8H-indolo[3,2,1-de]acridin-8-one can form
two types of hydrogen bonds. One with lone-pair of nitrogen
atom and two hydrogen bonds with lone-pairs of oxygen
atom. In the excited state intramolecular charge transfer
from the amino group to the carbonyl group occurs. This
results in a change of the dipole moment as well as the
change in the energy of hydrogen bonds. A hydrogen
bond formed by a lone pair of nitrogen atom is weakened
whereas those formed by a carbonyl group strengthens. Aer
max
abs
Fig. 6 The solvatochromic plot of the n
dependence on Bilot–
Kawski polarizability function fBK(3, n2) in: (1) MeOH, (2) EtOH, (3)
iPrOH, (4) TFE, (5) HFIP, (6) hexane, (7) iso-octane, (8) hexadecane, (9)
squalene, (10) cyclohexane, (11) toluene (12) 1-chloro-propane, (13) solvent relaxation, the strength of hydrogen bond formed by
1-chloro-butane, (14) 1-chlorohexane, (15) 1-chloro-octane, (16)
1-chloro-decane, (17) 1-chlorohexadecane. The correlation equation
as well as the hydrogen bond energy changes in locally excited
Franck–Condon state for alcohols studied are also shown.
oxygen atom increases and simultaneously increases the
dipole moment in the relaxed excited state. A smaller bath-
ochromic shi of absorption spectra compared to the emis-
sion spectra indicates that dipole moment in the excited
Franck–Condon state is smaller than in the relaxed excited
state.
max
fluo
Fig. 7 The solvatochromic plot of the n
dependence on Bilot–
Kawski polarizability function fBK(3, n2) in: (1) MeOH, (2) EtOH, (3)
iPrOH, (4) TFE, (5) HFIP, (6) hexane, (7) iso-octane, (8) hexadecane, (9)
squalene, (10) cyclohexane, (11) toluene (12) 1-chloro-propane, (13)
1-chloro-butane, (14) 1-chlorohexane, (15) 1-chloro-octane, (16)
1-chloro-decane, (17) 1-chlorohexadecane. The correlation equation
as well as the hydrogen bond energy changes in relaxed excited state
for alcohols studied are also shown.
Fig.
8 The hydrogen bond energy changes in locally excited
Franck–Condon excited state (dashed line) and relaxed excited
state (solid line) as a function of Kamlet–Taft solvent acidity
parameter (a).
Moreover, red shi of absorption and emission spectrum
observed for toluene can results from interactions with
permanent dipole moment of solvent, however, the staking
interaction with solute cannot be excluded. Because of that,
in the correlation of spectroscopic properties with solvent
polarizability function only 1-chloro-n-alkanes were taken
into account. Moreover, the maxima of absorption and uo-
rescence spectra measured in alcohols are below the line
dened by 1-chloro-n-alkanes which determines the non-
specic interactions. Thus, the energy of hydrogen bond
increases in the excited state in such a way that in the relaxed
excited state the hydrogen bond is about three times stronger
than that in the locally excited Franck–Condon state
(Table 4). Additionally, the changes of hydrogen bond energy
Fig.
Franck–Condon excited state (dashed line) and relaxed excited
state (solid line) as a function of Catalan solvent acidity parameter
9 The hydrogen bond energy changes in locally excited
´
follow the Kamlet–Ta or Catalan solvent acidity parameter
´
(a or SA) (Fig. 8 and 9), while the quality of t is better for the (SA).
19318 | RSC Adv., 2014, 4, 19310–19320
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
RSC Advances
linearity is observed for solvents with a quadrupole moment
and with a single bond dipole moment, and for solvents with
bond dipole centered at a single atom. These solvents appear
more polar locally to optical probe molecule than expected on
the basis of bulk dielectric constant.39 The correlation of the
Stokes shi with the microscopic solvent polarity parameter ETN
can be applied to calculate the dipole moment change between
the excited and ground state. According to the equation
proposed by Ravi et al.40 and Kumar et al.,41 the problem asso-
ciated with the Onsager's radius estimation can be minimized
since a ratio of two Onsager's radii is involved according to
eqn (16):
3.5 Calculations of dipole moment in excited state
According to Bilot–Kawski theory,18–21 the change of dipole
moment in the excited state can be calculated, basing on the
spectroscopic data, from following equations:
max
abs
max
fluo
~n ꢂ ~n ¼ m1 fBKð3; n2Þ þ const
(12)
max
max
fluo
~n þ ~n ¼ ꢂm2½ fBKð3; n2Þ þ 2gðnÞꢄ þ const
abs
where f (3, n2) is given by eqn (9) while
3ðn4 ꢂ 1Þ
gðnÞ ¼
(13)
2
2ðn2 þ 2Þ
"
#
ꢂ
ꢃ
2
Using a slope obtained from eqn (12) and knowing the
ground state dipole moment (mg), the excited state dipole
moment can be calculated using equation,
ꢅ
ꢆ
3
Dm
aB
a
max
abs
max
fluo
n
ꢂ n
¼ 11307:6
ETN þ const
(16)
DmB
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
where DmB and aB are the dipole moment change (Dm ¼ me ꢂ mg)
and Onsager's radius, respectively, for a pyridinium N-pheno-
late betaine dye used to determine the ENT values (DmB ¼ 9 D and
me ¼ mg þ m2ba3
(14)
2
where: a is Onsager's interaction radius of the solute and b ¼
2p30hc ¼ 1.10511044 ꢀ 10ꢂ35 C2 is the universal constant. If
ground and excited state dipole moments are not parallel, an
angle j between them can be calculated according to eqn (15):
˚
aB ¼ 6.2 A), whereas Dm and a are the corresponding quantities
for the molecule under study. The slope of dependence of
Stokes shi vs. ENT parameter is equal to 2736 (Fig. S3 ESI†)
which according to eqn (16) gives Dm ¼ 4.8 D and thereby me ¼
9.1 D. This value is larger than that calculated basing on Bilot–
Kawski theory, however, in calculation according to eqn (16) all
solvents studied were used including alcohols.
ꢄ
ꢇ
ꢅ
ꢆ
ꢅ
ꢆ
1
m1
m2
2
2
2
2
cos 4 ¼
mg þ me
ꢂ
me ꢂ mg
(15)
2mgme
To calculate the excited state dipole moment, the ground
state dipole moment is necessary. It was obtained from theo-
retical calculation performed applying DFT method with a def2-
TZVP basis set (TURBOMOLE v. 6.1) and is equal to mg ¼ 4.32 D.
The calculation of excited state dipole moment value was per-
formed taking into account only saturated hydrocarbons and 1-
chloro-n-alkanes. The correlation of spectral properties with
Bilot–Kawski solvent polarizability functions according to eqn
4 Conclusions
The observed solvatochromism of 3,6-di-tert-butyl-8H-indolo
[3,2,1-de]acridin-8-one is a result of the increase of the excited
state dipole moment. Therefore the solvent-dependent spec-
troscopic and photophysical properties were analyzed using
´
the Reichard's as well as Kamlet–Ta and Catalan solvent
(12) are satisfactory (r2 ¼ 0.9623, m1 ¼ 1063 cmꢂ1 and r2
¼
scales. The analysis of these relationships leads to the
conclusion that the solvent polarity and solvent hydrogen-
bonding donor ability have major inuence on measured
values. The radiationless deactivation rate constant value
decreases with a decrease of the DE(S1 ꢂ S0) energy gap not
only for molecule but also for hydrogen-bonded complex. It
means that the increase of solvent polarity causes the
decrease of S1–T intersystem rate constant and thereby
decreases the intersystem crossing which seems to be the
main route of deactivation of the excited state.
Additionally, the spectroscopic properties of this compound
were investigated in protic and non-protic solvents and the
results were discussed in terms of hydrogen-bonding. It was
experimentally found that 3,6-di-tert-butyl-8H-indolo[3,2,1-de]-
acridin-8-one forms hydrogen bonds which are overall
strengthened in the locally excited Franck–Condon and relaxed
excited states compared to that of the ground state.
0.9853, m2 ¼ 1763 cmꢂ1, respectively) (Fig. S14 and S15 ESI†).
The calculated excited state dipole moment using Onsager's
˚
cavity radius equal to a ¼ 6.63 A is equal to me ¼ 8.35 D whereas
an angle between the ground and excited state dipole moment
is 36ꢁ. Thus, the excited state dipole moment is about two times
larger than in the ground state. Apart from the rigid structure of
3,6-di-tert-butyl-8H-indolo[3,2,1-de]acridin-8-one a meaningful
change of the angle between ground and excited state dipole
moments is observed indicating on the substantial modica-
tion of charge distribution in the relaxed excited state. As
mentioned above, the excited state dipole moment was calcu-
lated taking into account saturated hydrocarbons and 1-chloro-
n-alkanes only. In Fig. S14 and S15 ESI,† the best t to the eqn
(12) considering all solvents, except alcohols (dashed line), are
also shown. The quality of t is of course worse, however, the
slope differs only slightly from the value of slope obtained for
saturated and 1-chloro-n-hydrocarbons resulting in a small
change of the calculated dipole moment compared to that
calculated for saturated and 1-chloro-hydrocarbons solvents.
The solvatochromic plot of nmabasx and nmuaox versus fBK(3, n2) for all
studied solvents are shown in Fig. S16 ESI.† Omitting the data
for alcohols, it should be noted that the largest deviation from
Acknowledgements
This work was nancially supported by the grant for Young
Scientists no. 538-8443-B046-13 from the University of Gdansk.
´
This journal is © The Royal Society of Chemistry 2014
RSC Adv., 2014, 4, 19310–19320 | 19319

View Article Online
RSC Advances
Paper
21 L. Bilot and A. Kawski, Z. Naturforsch., A: Phys. Sci., 1963, 18,
10 and 256.
Notes and references
22 M. J. Kamlet, J.-L. M. Abbound and R. W. Ta, Prog. Phys.
Org. Chem., 1982, 13, 485.
1 P. Belmont, J. Bosson, T. Godet and M. Tiano, Anti-Cancer
Agents in Medicinal Chemistry, 2007, 7, 139.
´
´
23 J. Catalan, J. Phys. Chem. B, 2009, 113, 5951.
2 G. Cholewinski, K. Dzierzbicka and A. M. Kołodziejczyk,
24 W. Zhao, L. Pang, W. Bian and J. Wang, ChemPhysChem,
2008, 9, 1593.
Pharmacol. Rep., 2011, 63, 305.
´
3 A. F. Valdes, Open. Med. Chem. J., 2011, 5, 11.
25 E. Krystkowiak, K. Dobek and A. Maciejewski, Photochem.
Photobiol. Sci., 2013, 12, 446.
4 P. Yang, Q. Yang, L. Tong and X. Li, J. Photochem. Photobiol.,
B, 2006, 84, 221.
´
26 E. Krystkowiak, K. Dobek, G. Burdzinski and A. Maciejewski,
5 Y. Hagivara, T. Hasegawa, A. Shoji, M. Kuwahara, H. Ozaki
and H. Sawai, Bioorg. Med. Chem., 2008, 16, 7013.
6 C. Gao, S. Li, X. Lang, H. Liu, F. Liu, C. Tan and Y. Jiang,
Tetrahedron, 2012, 68, 7920.
Photochem. Photobiol. Sci., 2012, 11, 1322.
27 E. Krystkowiak and A. Maciejewski, Phys. Chem. Chem. Phys.,
2011, 13, 11317.
28 F. C. Chipen, A. Mishra and G. Krishnamoorthy, Phys. Chem.
Chem. Phys., 2012, 14, 8775.
7 B. Qiu, L. Guo, Z. Chen, Y. Chi, L. Zhang and G. Chen,
Biosens. Bioelectron., 2009, 24, 1281.
29 Y. Liu, M. Nishiura, Y. Wang and Z. Hou, J. Am. Chem. Soc.,
2006, 128, 5592.
30 M. V. Skorobogatyi, A. A. Pchelintseva, A. L. Petrunina,
I. A. Stepanova, V. L. Andronova, G. A. Galegov,
A. D. Malakhov and V. A. Korshun, Tetrahedron, 2006, 62,
1279.
31 J. R. Lakowicz, Principles of Fluorescence Spectroscopy, Kluwer
Academic/Plenum Publishers, New York, 2nd edn, 1999.
32 M. Baruah, W. Qin, C. Flores, J. Hoens, R. A. L. Valle,
D. Beljonne, M. Van der Auweraer, W. M. De Borggrave
and N. Boens, J. Phys. Chem. A, 2006, 110, 5998.
33 J. Jayabarathi, V. Thanikachalam and K. Jayamoorthy,
Photochem. Photobiol. Sci., 2013, 121, 761.
8 S. Cosnier and P. Mailley, Analyst, 2008, 133, 984.
9 P. Suppan and N. Ghoneim, Solvatochromism, The Royal
Society of Chemistry, Cambridge, U.K., 1997.
10 E. Krystkowiak, K. Dobek and A. Maciejewski, J. Photochem.
Photobiol., A, 2006, 184, 250.
11 G.-J. Zhao and K.-L. Han, Acc. Chem. Res., 2012, 45, 404.
12 K. Guzow, M. Milewska and W. Wiczk, Spectrochim. Acta, Part
A, 2005, 61, 1133.
13 K. Guzow, A. Ceszlak, M. Kozarzewska and W. Wiczk,
Photochem. Photobiol. Sci., 2011, 10, 1610.
´
14 K. Guzow, M. Czerwinska, A. Ceszlak, M. Kozarzewska,
M. Szabelski, C. Czaplewski, A. Łukaszewicz,
A. A. Kubicki and W. Wiczk, Photochem. Photobiol. Sci.,
2013, 12, 284.
34 R. Englman and J. Jortner, Mol. Phys., 1970, 18, 145.
35 Postgraduate Chemistry Series, Photochemistry of Organic
´
15 A. Filarowski, M. Kluba, K. Cieslik-Boczula, A. Koll,
´
Compounds. From Concept to Practice, ed. P. Klan and
A. Kochel, L. Pandey, W. M. De Borggrave, M. van der
´
J. Wirz, J. Wiley & Sons Ltd., UK, 2009.
36 N. J. Turro, V. Ramamuthry and J. C. Scaiano, Principles of
Molecular Photochemistry. An Introduction, University
Science Book, Sausalito, California, 2009, ch. 5.
37 R. Lopez-Artega, A. B. Stephansen, C. A. Guarin, T. I. Solling
and J. Peon, J. Phys. Chem. B, 2013, 117, 9947.
38 X. Liu, J. M. Cole and K. S. Low, J. Phys. Chem. C, 2013, 117,
14731.
Auweraer, J. Catalan and N. Boens, Photochem. Photobiol.
Sci., 2011, 9, 996.
16 C. Reichardt, Chem. Rev., 1994, 94, 2319.
17 C. Reichardt and E. Harbusch-Gornert, Liebigs Ann. Chem.,
¨
1983, 721.
18 A. Kawski, in Progress in Photochemistry and Photophysics, ed.
J. F. Rabek, CRC Press, Boca Raton, USA, 1992, vol. 5,
pp. 1–47.
39 I. Renge, J. Phys. Chem. A, 2010, 114, 6250.
40 M. Ravi, A. Samanta and T. P. Radharkrishnan, J. Phys. Chem.
A, 1994, 98, 9133.
41 S. Kumar, V. C. Rao and R. C. Rostogi, Spectrochim. Acta, Part
A, 2001, 57, 41.
19 A. Kawski and P. Bojarski, in Chromic Materials, Phenomena
and Their Technological Applications, Applied Science
Innovations, ed. P. R. Somani, 2009, pp. 114–163.
20 L. Bilot and A. Kawski, Z. Naturforsch., A: Phys. Sci., 1962, 17,
621.
19320 | RSC Adv., 2014, 4, 19310–19320
This journal is © The Royal Society of Chemistry 2014