Synthesis, photophysical characterization, CASSCF/CASPT2 calculations and CT-DNA interaction study of amino and azido benzazole analogues

Article abstract of DOI:10.1016/j.molliq.2019.111938

This work describes the synthesis and photophysical investigation of amino and azido benzazoles. The amino derivatives were obtained by condensation reaction between ortho-substituted anilines and p-aminobenzoic in polyphosphoric acid. The respective azides were synthesized by reaction of diazonium salts from the previously prepared amines with sodium azide. These compounds present absorption maxima in the UV-A region, nm ascribed to fully spin and symmetry electronic transitions. All compounds presented a main fluorescence emission in the UV-A to the violet region with a relatively large Stokes shift. The latter related to a solvent dependence. The amino derivatives presented higher values to the fluorescence quantum yields in despite of the azido analogues. DFT, TD-DFT and multiconfigurational calculations (SA-CASSCF and MS-CASPT2) were performed in order to investigate the photophysical features of these molecules, mainly on the azide derivatives, where the main interest was the investigation of the intrinsic fluorescence quenching present in these compounds. In this sense, it was observed that the weak fluorescence emission observed in the azide compounds could be related to the dissociative character of the S₁ state, which reaches a conical intersection point between S₁/S₀ states, and through this point, goes back to the ground state by a nonradioactive decay. In addition, the DNA binding assays by UV–Vis absorption and fluorescence emission methodologies indicated that the benzazoles presented strong interaction with CT-DNA, which could be attributed to π-stacking and/or intermolecular hydrogen-bonding. Docking was also performed to better understand the observed interaction.

Full text of DOI:10.1016/j.molliq.2019.111938

Journal Pre-proof
Synthesis, photophysical characterization, CASSCF/CASPT2 calculations and CT-
DNA interaction study of amino and azido benzazole analogues
Eduarda S. Gil, Cláudia B. da Silva, Pablo A. Nogara, Carolina H. da Silveira, João
B.T. da Rocha, Bernardo A. Iglesias, Diogo S. Lüdtke, Paulo F.B. Gonçalves, Fabiano
S. Rodembusch
PII:
S0167-7322(19)34500-3
DOI:
https://doi.org/10.1016/j.molliq.2019.111938
MOLLIQ 111938
Reference:
To appear in: Journal of Molecular Liquids
Received Date: 9 August 2019
Revised Date: 26 September 2019
Accepted Date: 14 October 2019
Please cite this article as: E.S. Gil, Clá.B. da Silva, P.A. Nogara, C.H. da Silveira, Joã.B.T. da
Rocha, B.A. Iglesias, D.S. Lüdtke, P.F.B. Gonçalves, F.S. Rodembusch, Synthesis, photophysical
characterization, CASSCF/CASPT2 calculations and CT-DNA interaction study of amino and
azido benzazole analogues, Journal of Molecular Liquids (2019), doi: https://doi.org/10.1016/
j.molliq.2019.111938.
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition
of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of
record. This version will undergo additional copyediting, typesetting and review before it is published
in its final form, but we are providing this version to give early visibility of the article. Please note that,
during the production process, errors may be discovered which could affect the content, and all legal
disclaimers that apply to the journal pertain.
© 2019 Published by Elsevier B.V.

Synthesis, photophysical characterization, CASSCF/CASPT2 calculations
and CT-DNA interaction study of amino and azido benzazole analogues
Eduarda S. Gil,a† Cláudia B. da Silva,a† Pablo A. Nogara,b Carolina H. da
Silveira,c João B. T. da Rocha,b Bernardo A. Iglesias,c Diogo S. Lüdtke,a Paulo
F. B. Gonçalves,*,a Fabiano S. Rodembusch*,a
aInstituto de Química, Universidade Federal do Rio Grande do Sul, UFRGS. Av.
Bento Gonçalves 9500, 91501-970, Porto Alegre, RS, Brazil. E-mail:
rodembusch@iq.ufrgs.br (Fabiano S. Rodembusch) and paulo@iq.ufrgs.br
(Paulo F. B. Gonçalves)
bLaboratório de Bioquímica Toxicológica, Departamento de Bioquímica e
Biologia Molecular, Universidade Federal de Santa Maria, 97105-900, Santa
Maria, RS, Brazil.
cLaboratório de Bioinorgânica e Materiais Porfirínicos, Universidade Federal de
Santa Maria, 97105-900, Santa Maria, RS, Brazil.
†Both authors contributed equally to this work
Abstract
This work describes the synthesis and photophysical investigation of
amino and azido benzazoles. The amino derivatives were obtained by
condensation reaction between ortho-substituted anilines and p-aminobenzoic
in polyphosphoric acid. The respective azides were synthesized by reaction of
diazonium salts from the previously prepared amines with sodium azide. These
compounds present absorption maxima in the UV-A region, nm ascribed to fully
spin and symmetry electronic transitions. All compounds presented a main
fluorescence emission in the UV-A to the violet region with a relatively large
Stokes shift. The latter related to a solvent dependence. The amino derivatives
presented higher values to the fluorescence quantum yields in despite of the
azido analogues. DFT, TD-DFT and multiconfigurational calculations (SA-
CASSCF and MS-CASPT2) were performed in order to investigate the
photophysical features of these molecules, mainly on the azide derivatives,
where the main interest was the investigation of the intrinsic fluorescence
1

quenching present in these compounds. In this sense, it was observed that the
weak fluorescence emission observed in the azide compounds could be related
to the dissociative character of the S1 state, which reaches a conical
intersection point between S1/S0 states, and through this point, goes back to the
ground state by a nonradioactive decay. In addition, the DNA binding assays by
UV-Vis absorption and fluorescence emission methodologies indicated that the
benzazoles presented strong interaction with CT-DNA, which could be
attributed to π-stacking and/or intermolecular hydrogen-bonding. Docking was
also performed to better understand the observed interaction.
1. Introduction
Benzazoles are a class of heterocyclic compounds with numerous
applications in pharmaceuticals and materials science [1]. These scaffolds
consist of the fusion of benzene and the azole moiety [2]; this latter constituted
by five atoms of which two are heteroatoms in positions 1 and 3, and one of
them always being the nitrogen atom [3]. These compounds can exhibit
interesting photophysical properties, such as fluorescence emission with large
Stokes shift and high photostability [4-9]. In this regard, several methods of
preparing benzazoles are described in the literature, including for instance
intramolecular cyclization, catalyzed cyclization of o-haloanilides and
condensation of o-substituted anilines with carboxylic acids, acyl esters and
chlorides, amides, nitriles or even aromatic aldehydes [10-19]. Among this class
of compounds, several benzazole derivatives have been described with
inhibitory activity in various enzymatic reactions allowing its investigation to treat
or even prevent diseases [20-23]. In this context, several benzothiazole
derivatives have also shown antibacterial, anticancer, anti-oxidant, antiviral,
anti-tuberculosis, anti-diabetic and anti-inflammatory activities [24-29]. In
addition, the particular photophysical properties, as well as their planar
structures, make them suitable for use as fluorescent probes to study biological
mechanisms in protein and DNA [30-35]. In this context, another class of
molecules able to act as DNA probes are the azide based compounds, through
the DNA hybridization. In this approach, a single-stranded DNA probe (ssDNA)
is immobilized on a surface and exposed to a sample containing the
2

complementary target sequence characteristics, which is captured by the
formation of a double-stranded DNA (dsDNA). This event, so-called
hybridization, is then translated into a signal. A variety of transduction
techniques can be used to observe this process, including the optical one
[36,37].
Thus, the present study presents the synthesis of benzazole
heterocycles containing amino and azido groups, as well as their photophysical
study and the application to study the interaction with CT-DNA. The
photophysical characterization and the interaction study were complemented
with theoretical calculations and docking, respectively.
2. Experimental
2.1. Materials and methods
All reagents were used as received and the solvents were purchased
from commercial sources and purified according to the literature [38]. The
reactions were monitored by Thin Layer Chromatography (TLC) using Silica Gel
60 F254. Fourier transform infrared (FTIR) spectra in the range 600-4000 cm-1
and in KBr pellets were recorded on a Varian model 640-IR with a resolution of
4 cm-1. Hydrogen and carbon nuclear magnetic resonance spectra in CDCl3 (1H
13
and C NMR) were recorded in CDCl3 at 400 and 100 MHz, respectively. The
chemical shifts (δ) are reported in parts per million (ppm) relative to TMS (0.00
ppm), and the coupling constants J are reported in hertz (Hz). Melting points
were obtained on a QUIMIS model Q340S and are uncorrected. In the
photophysical studies, spectroscopic-grade solvents were used. ESI-QTOF-MS
measurements were performed in the positive ion mode (m/z 50-2000 range).
The UV-Vis spectra were recorded on a Shimadzu UV-2450 spectrophotometer,
and the steady-state fluorescence spectra were measured on a Shimadzu
spectrofluorometer model RF-5301PC. The quantum yields of fluorescence
(ΦFL) were measured at 25ºC using solutions with absorbance intensity lower
than 0.05 (optical dilute regime). In these experiments the emission spectra
were obtained using slits of 1.5 nm/1.5 nm (Exc/Em). Quinine sulfate in H2SO4
1N was used as the quantum yield standards [39].
3

2.2. Synthesis
General procedure for the synthesis of amino benzoxazoles 4-5
An equimolar mixture of o-aminophenol (1) (2.4 g, 22.0 mmol) or o-
aminothiophenol (2) (2.7 mL, 22.0 mmol) and p-aminobenzoic acid (3) (3.0 g,
22.0 mmol) in polyphosphoric acid (20 mL) was stirred at 170°C for 5 h. After
cooling, the mixture was poured onto ice under vigorous stirring. The solution
was neutralized with NaHCO3. The precipitate was filtered, washed with water
and dried at 60oC. The solid was purified by column chromatography using
dichloromethane as eluent [40].
2-(4'-Amino)benzoxazole (4) [41]. Yield: 69% (1.7 g). Melting point: 150-152ºC.
FTIR (KBr, cm-1): 3497 (νas N-H), 3325 (νs N-H), 3000 (νarom C-H), 1620 and
1470 (νarom C=C). 1H NMR (400 MHz, CDCl3) δ 8.10-8.04 (m, 2H), 7.76-7.70 (m,
1H), 7.57-7.52 (m, 1H), 7.36-7.28 (m, 2H), 6.79-6.75 (m, 2H), 4.21 (S, 2H). APT
(101 MHz, CDCl3) δ 163.8, 150.5, 149.7, 142.4, 129.4, 124.3, 124.2, 119.9,
116.8, 114.7, 110.2.
2-(4'-Amino)benzothiazole (5) [42]. Yield: 63% (1.4 g). Melting point: 161-163ºC.
FTIR (KBr, cm-1): 3497 (νas N-H), 3325 (νs N-H), 3000 (νarom C-H), 1620 and
1
1470 (νarom C=C). H NMR (400 MHz, CDCl3) δ 7.87 (dd, J=13.9 and 7.9 Hz,
3H), 7.46 (t, J=7.6 Hz, 3H), 7.34 (t, J=7.6 Hz, 1H), 6.34 (d, J=2.0 Hz, 1H), 6.28
(dd, J=8.4 and 2.1 Hz, 1H), 4.25 (s, 2H). APT (101 MHz, CDCl3) δ 169.6, 159.8,
150.9, 131.9, 130.0, 126.4, 121.4, 121.3, 108.5, 107.2, 101.7.
General procedure for the synthesis of azido benzoxazoles 6-7
In a 50 mL round bottom flask under an ice bath containing 2.0 mmol of
the benzazole precursors 4-5 (2.0 mmol) and a HCl:H2O (3:1) solution (15 mL)
it was added dropwise a NaNO2 (0.166 g, 2.4 mmol) solution (1mL). The
reaction mixture was stirred under ice bath for 30 min. After this time, NaN3 was
added (0.260 g, 4.0 mmol). The reaction mixture was stirred at room
temperature for 2h. After this time, the solution was neutralized with NH4OH,
extracted with ethyl acetate (3x20 mL) and concentrated under reduced
pressure to yield the respective product.
4

2-(4'-Azido)benzoxazole (6) [41]. Yield: 71% (0.453 g). Melting point: 128-
130oC. FTIR (ATR, cm-1): 3077 (νarom C-H), 2077 (ν-N=N=N), 1513 and 1452
(νarom C=C).1H NMR (400 MHz, CDCl3) δ 8.28-8.22 (m, 1H), 7.83-7.74 (m, 1H),
7.62-7.55 (m, 1H), 7.43-7.32 (m, 1H), 7.20-7.12 (m, 1H). APT (75 MHz, CDCl3)
δ 150.8, 143.4, 142.1, 129.2, 125.1, 124.6, 123.8, 119.9, 119.5, 110.5. HRMS
(ESI-TOF) m/z: [M+H]+ Calcd for C13H8N4O 237.0776; Found 237.0782.
2-(4'-Azido)benzothiazole (7). Yield: 78% (0.431 g). Melting point: 154-156oC.
FTIR (ATR, cm-1): 3046 (νarom C-H), 2108 (ν-N=N=N), 1602 and 1482 (νarom
1
C=C). H NMR (400 MHz, CDCl3) δ 8.12-8.06 (m, 1H), 7.91 (d, J=7.6 Hz, 1H),
7.52-7.49 (m, 1H), 7.44-7.37 (m, 1H), 7.16-7.12 (m, 1H). APT (75 MHz, CDCl3)
δ 167.1, 154.3, 142.8, 135.3, 130.6, 129.0, 126.4, 125.2, 123.1, 121.6, 119.5.
HRMS (ESI-TOF) m/z: [M+H]+ Calcd for C13H8N4S 253.0548; Found 253.0549.
2.3. CT-DNA binding study
For calf-thymus DNA (CT-DNA) interaction studies with amino and azido
derivatives UV-Vis absorption measurements were performed at room
temperature in Tris-HCl buffer at pH 7.2, using DMSO stock solution of
derivatives (10-4 M). The DNA pair base concentrations of low molecular weight
CT-DNA were determined using the molar extinction coefficient of 13,200
M−1·cm−1 (per base pair) at absorption maximum of 260 nm. Compound
solutions in DMSO (2%)/Tris-HCl buffer mixture were titrated with increasing
concentrations of CT-DNA (ranging from 0-100 ꢀM). Changes in absorption
spectra of derivatives were acquired and observed in the wavelength range of
290-500 nm. The intrinsic binding constants (Kb) of derivatives 4-5 and 6-7 were
calculated according to the decay of the absorption bands of compounds using
the following Equation (1) through a plot of [DNA]/(εa − εf ) versus [DNA],
[DNA]/(εa − εf) = [DNA]/(εb − εf) + 1/Kb(εb − εf) (1)
where [DNA] is the concentration of DNA in the base pairs, εa is the extinction
coefficient (Aobs/[compound]), εb and εf are the extinction coefficients of free and
5

fully bound forms, respectively. In the plots of [DNA]/(εa − εf) versus [DNA], Kb is
given by the slope/interception ratio.
In EB-DNA competitive assays, steady-state emission fluorescence
analysis was recorded and the amino and azido derivatives were dissolved in
DMSO (10-4 M range) and competitive studies were performed through the
gradual addition of the stock solution of the derivatives to the quartz cuvette (1.0
cm path length) containing ethidium bromide (EB, 2.0x10-7 M) and CT-DNA
(2.0x10-5 M) in a Tris-HCl pH 7.2 buffer solution. The concentration of
derivatives ranged from 0 to 100 ꢀM. In these studies an excitation wavelength
of 510 nm was used and the respective fluorescence emission spectra were
recorded in the range of 550-800 nm, 5 min after each addition of the derivative
solution in order to allow incubation to occur. The fluorescence quenching
Stern-Volmer constants (KSV) of compounds were calculated according to the
decay of the emission bands of EB-DNA using the following Equation (2)
through a plot of F0/F versus [DNA],
F0/F = 1+ kqτ0[Q] = 1 + KSV[Q] (2)
where F and F0 are the fluorescence intensities in the presence and absence of
a quencher, respectively [43]. KSV, kq, τ0 and [Q] denote Stern-Volmer
quenching constant, quenching rate constant, lifetime of EB-DNA adducts
(23.0x10-9 s) [44] and the concentration of quencher, respectively. According to
Equation (2), the Stern-Volmer constants (KSV) were calculated from the slope
and kq is equal KSV/τ0.In order to quantify the displacement, the concentration of
the heterocycle derivatives at which EB fluorescence decreases by 50% (in this
case, assumed to be 50% displacement of EB) is calculated [45]. The values of
apparent DNA-binding constant (Kapp) were calculated using the Equation (3):
KEB[EB] = Kapp[compound] (3)
where KEB (4.94 × 105 M−1) is the DNA-binding constant of ethidium bromide,
[EB] is the concentration of EB (1.50x10-6 M), and [compound] is the
concentration of the derivative used to obtain 50% reduction in fluorescence
6

emission intensity of EB. The standard Gibbs free-energy (∆Gº) of benzazole-
DNA complex was calculated from the values of binding constant (Kb) using the
following Equation (4):
ΔGꢀ = −RT ln ꢁꢂ
(4)
2.4. Docking
The 3D structure of DNA was obtained from the Protein Data Bank
(http://www.rcsb.org/pdb/) with the code: 423D [46]. The Chimera 1.8 software
was used to remove waters, ions, and other molecules, and add hydrogens to
the DNA [47]. The studied benzazoles were built in the software Avogadro 1.1.1
[48], following the semi-empirical PM6 geometry optimization using the program
MOPAC2012 (http://OpenMOPAC.net) [49]. The derivatives and DNA in the
pdbqt format were generated by AutoDockTools, where the ligands were
considered flexible (with PM6 charges), and the DNA rigid (with Gasteiger
charges) [50]. AutoDock Vina 1.1.1 program was used for the blind docking
[51], using a gridbox of 603 Å (coordinates: x=-7.914, y=52.208, z=-0.176), an
exhaustiveness of 50 and 1.0 Å of grid spacing. The results from docking were
and analyzed using the Accelrys Discovery Studio 3.5 software [52], and the
conformer with the lowest binding free energy was selected as the best model
of interaction.
2.5 Theoretical calculations
Quantum chemical calculations have been carried out by DFT and TD-
DFT, CASSCF and CASPT2 methods in order to evaluate the photophysical
and electronical properties of the amino and azido derivatives. The ground state
geometries were optimized in DFT (Density Functional Theory) with the CAM-
B3LYP functional and cc-pVDZ basis set [53]. Frequency calculations of the
optimized structures were performed at the same level to confirm that the
optimized structures are stationary points. The calculations of the vertical
transitions were carried out using TDDFT with the CAM-B3LYP functional and
jun-cc-pVTZ basis function. The CAM-B3LYP functional was employed because
it presents the attributes of the B3LYP functional and long-range correction,
which is crucial for a good description of excited states. The solvent effect was
7

considered in DFT and TDDFT calculations with PCM (Polarizable Continuum
Model), and the solvents considered were acetonitrile (CH3CN),
dichloromethane (DCM) and ethanol (EtOH) [54]. Natural Transition Orbitals
(NTO) was calculated because the molecules 4 and 6 showed more than one
contribution to the first transition, hindering their qualitative description. NTOs
drastically simplify the situation by providing a compact representation of the
transition density matrix [55]. All the calculations described above were
computed using the Gaussian 16 package [56]. In order to verify if the
molecules present charge transfer, the analysis with the DCT descriptor [57]
were carried out with the Multiwfn program [58].
Multiconfigurational calculations were performed for azido benzoxazole 6
in order to investigate the lack of fluorescence of that molecule. The minima
points and conical intersection (CoIn) point were optimized with an active space
including 12 electrons and 12 orbitals (Figure 7) and state-averaged over three
states (SA3-CASSCF(12,12)) with ANO-S-VDZP basis set. The orbitals
included in the active space were six occupied and six virtual orbitals, wherein it
were taken account the last six occupied orbitals (six occupied π orbitals) and
the first six unoccupied orbitals (six virtual π orbitals). This active space was
maintained during the subsequent geometry optimizations. Single point
calculations using MS5-CASPT2(12,12) with ANO-L-VDZP basis set were
performed for CASSCF geometries in order to improve the electronic
correlation. The Minimal Energy Point (MEP) coordinate for molecule 6 was
construct at SA3-CASSCF/cc-pVDZ level and energy corrected at MS5-
CASPT2/ANO-L-VDZP level (CASPT2//CASSCF protocol). Baker criteria was
used as geometry convergence criteria, i.e., 1.0 ꢃꢄꢅ and 3.0 ꢃꢄꢆ with respect to
the energy change and the norm of the gradient, respectively [59].
In all the CASSCF calculations, a level shift of 1.0 a.u. was used, and in
the CASPT2 calculations. The IPEA modified H0 with the shift parameter of 0.25
was also used [60,61]. The multiconfigurational (SA-CASSCF and MS-
CASPT2) calculations were carried out with OpenMolcas package [62] and all
orbitals were rendered with IboView program [63].
8

3. Results and Discussion
3.1. Synthesis
The amino benzazoles 4-5 were obtained by a condensation reaction
between ortho-substituted anilines 1-2 and the p-aminobenzoic (3) in
polyphosphoric acid (PPA) at 170oC for 5h as presented in Scheme 1 [64,65].
The progress of the reaction was monitored by TLC. The obtained precipitate
was neutralized with NaHCO3, filtered and dried at 60oC. The purification was
performed by column chromatography using dichloromethane as eluent. The
respective azides 6-7 were synthesized using the reaction of formation of
diazonium salts from the previously prepared amines. The first step is the amino
group diazotization at 0oC by reaction with sodium nitrite (NaNO2) in aqueous
hydrochloric acid. After the formation of the diazonium salt, sodium azide
(NaN3) was added to produce the respective azide. After 2h under stirring, the
final solution was neutralized with NH4OH, extracted with ethyl acetate and
concentrated under reduced pressure to yield the azides 6-7, which were used
without further purification.
Scheme 1. Preparation of the amino (4-5) and azido benzazole (6-7)
derivatives.
3.2. Photophysical properties
The photophysical investigation of the synthesized compounds was
carried out using dichloromethane (DCM), ethanol (EtOH) and acetonitrile
(ACN). The relevant data from the electronic ground and excited states
characterization are summarized in Table 1. Figure 1 shows the UV-Vis spectra
of the amino (4-5) and azido benzazole (6-7) derivatives.
9

Figure 1. UV-Vis absorption spectra in solution of the amino benzazoles (a) 4
and (b) 5 and the respective azido derivatives (c) 6 and (d) 7, where DCM:
dichloromethane, EtOH: ethanol ACN: acetonitrile at a concentration ~10-6 M for
the amino and ~10-5 M for the azido derivatives.
It can be observed that all studied compounds present absorption in the
UV-A region. In both set of compounds, the heteroatom plays an important role,
in which the sulfur derivatives present the absorption maxima located at higher
wavelengths. This behavior was already observed in similar benzazoles and is
related to the better electron delocalization allowed for this more electronegative
heteroatom [5,6,8]. As already observed for benzazole analogues containing
amino and azido moieties located at the same position, the azido derivatives
10

presented higher HOMO-LUMO energy gap, due to their lower electronic
conjugation if compared to the amino ones [5].
In addition, the molar absorptivity coefficient values (ε) obtained by the
UV-Vis spectroscopy allowed the obtention of the respective oscillator strengths
0
(fe), as well as the theoretical rate constants for emission (ke ) by applying a
simplified version of the Strickler-Berg relations presented in Equations (4) and
(5), respectively [66]. In Equation (5), ꢇ
̅
ꢈ is the wavenumber (energy in 1/λ units)
0
of the maximum of the absorption band. From ke , the pure radiative lifetime τ0
0
was also calculated, defined as 1/ke [67]. All these data are also summarized in
Table 1.
ꢉ ≈ 4.3ꢋ10ꢄꢌ ꢍ ꢎꢏꢇ
̅
(4)
ꢊ
ꢐꢊꢈ ≈ 2.88ꢋ10ꢄꢌꢇꢑ
̅
ꢈ ꢍ ꢎꢏꢇ (5)
̅
The obtained molar absorptivity coefficient (ε) values ~105 M-1·cm-1, as
well as the calculated radiative rate constants (ke ) (~108 s-1) indicate for the
0
amino (4-5) and azido (6-7) compounds, spin and symmetry allowed electronic
transitions, which could be related to 1ππ* transitions. Despite the higher values
to ε, small values to the oscillator strength were found (<0.3). This result can be
related to the size of the transition dipole which seems to be smaller than
usually expected for compounds presenting fully electronic transitions. An
almost constant radiative lifetime indicates that after the radiation absorption the
molecular hybrids populate the same excited state.
11

Table 1. Photophysical data of the amino (4-5) and azido (6-7) derivatives by UV-Vis absorption spectroscopy, where εmax
is the molar extinction coefficient at the absorption maxima (x105 M-1·cm-1), λabs is the absorption maxima (nm), fe is the
0
calculated oscillator strength, ke is the calculated radiative rate constant (108 s-1) and τ0 is the calculated pure radiative
lifetime (ns).
εmax
λabs
(nm)
317
327
320
333
343
338
317
316
316
327
325
325
ꢔꢓꢕ
(108 s-1)
τ0
Compound
4
Solvent
ꢒꢓ
(x105 M-1·cm-1)
(ns)
6.33
5.09
4.67
5.52
5.06
4.10
7.31
7.41
7.23
6.92
6.14
6.18
Dichloromethane
Ethanol
0.81
0.16
0.21
0.22
0.20
0.23
0.28
0.14
0.13
0.14
0.15
0.17
0.17
1.58
1.96
2.14
1.81
1.98
2.44
1.37
1.35
1.38
1.45
1.63
1.62
1.02
Acetonitrile
Dichloromethane
Ethanol
1.09
0.64
5
6
7
1.43
Acetonitrile
Dichloromethane
Ethanol
1.05
0.51
0.32
Acetonitrile
Dichloromethane
Ethanol
0.62
0.59
0.79
Acetonitrile
0.55
12

The fluorescence emission curves presented in Figure 2 were obtained
by exciting the compounds at the absorption maxima (Table 1). The relevant
data is summarized in Table 2. All compounds presented a main fluorescence
emission maximum in the UV-A to the violet region. In this sense, it can be
observed that changes from amino to the azido moiety seem do not affect the
emission maxima position. Once more, but with a higher magnitude, the sulfur
compounds presented the emission maxima redshifted in contrast to the
oxygenated analogues. All compounds presented relatively large Stokes shift
(~5000 cm-1). Based on the chemical structure of these compounds and their
relative rigidity, as well as the absence of a clear charge transfer character, the
observed values can be probably related to organic solute-solvent interaction
[68].
It is worth mentioning that the most significant difference between these
compounds could be observed in the fluorescence quantum yield values. While
derivatives containing the amino group show significantly higher values (~0.5),
the azido derivatives presented values ten times smaller. As already observed
in the literature, usually azide compounds are non-fluorescent, related to a
photoinduced electron transfer (PET) mechanism ascribed to the presence of
the -N3 group. However, recent results indicate that a different mechanism than
PET can tailor the fluorescence emission in such compounds [5]. In this sense,
based on the theoretical results presented in this study and a study recently
presented in the literature [69], the very weak fluorescence emission observed
in these compounds could be related to the dissociative character of the S1
state of these compounds. For a deeper discussion please see the theoretical
calculation section.
13

Figure 2. Steady-state fluorescence emission spectra in solution of the amino
benzazoles (a) 4 and (b) 5 and the respective azido derivatives (c) 6 and (d) 7,
where DCM: dichloromethane, EtOH: ethanol ACN: acetonitrile at a
concentration ~10-6 M for the amino and ~10-5 M for the azido derivatives.
Exc/Em slits 3.0 nm/3.0 nm. Pictures of compounds 4 and 5 are also presented
under normal light and UV-radiation (365 nm).
14

Table 2. Photophysical data from steady-state fluorescence emission
spectroscopy of the amino ( ) and azido ( ) derivatives, where λem is
4
-
5
6-7
the emission maxima, ∆λST is the Stokes shift (nm/cm-1) and φFL is the
fluorescence quantum yield.
λem
(nm)
371
382
376
393
410
401
364
380
368
392
415
390
∆λST
Compound
4
Solvent
φFL
(nm/cm-1)
54/4592
55/4403
56/4654
60/4585
67/4764
63/4648
47/4073
64/5330
52/4472
65/5071
90/6673
65/5128
Dichloromethane
Ethanol
0.58
0.51
0.48
0.50
0.40
0.47
0.05
0.04
0.08
0.06
0.02
0.07
Acetonitrile
Dichloromethane
Ethanol
5
6
7
Acetonitrile
Dichloromethane
Ethanol
Acetonitrile
Dichloromethane
Ethanol
Acetonitrile
3.3 Theoretical calculations
The theoretical study has as its main goal the understanding how the
intrinsic fluorescence quenching of the azido derivatives in comparison to their
amino analogues. In this section, the photophysics features were investigated
by TD-DFT and by multiconfigurational calculations. The discussion will be
carried out for the amino benzoxazole 4 and the respective azide 6 as model
compounds, since the sulfur analogues presented similar photophysical
features. All the calculations resulting from these sulfur derivatives can be found
in the ESI. Figure 3 shows the optimized geometries for the ground and first
excited states of 4 and 6 in acetonitrile calculated with TD-DFT. Similar
geometries were obtained in dichloromethane and ethanol (data not shown, see
ESI).
15

Figure 3. Optimized geometries of amino and azido benzazoles 4 (top) and 6
(bottom), respectively, calculated with CAM-B3LYP/cc-pVDZ level in
acetonitrile.
It is possible to see that the geometries of the ground state are very
similar of those in the first excited state for both molecules, indicating that these
compounds do not present considerable geometry changes during the first
electronic transition. Natural Transition Orbitals (NTO’s) of the first electronic
transition for the amino benzazole 4 and the respective azide 6, including the
solvent effect, are presented in Figure 4. These orbitals do not show a spatial
separation in the transition, showing that there is only a local excitation.
Furthermore, it is possible to confirm by analyzing the orbital coefficients that
1π→ π* transitions are occurring in both structures.
16

Figure 4. NTO’s involved in the first transition for 4 (top) and 6 (bottom)
computed at the CAM-B3LYP/jun-cc-pVTZ level in acetonitrile, dichloromethane
and ethanol.
Table 3 shows the vertical transition analysis for amino benzazole 4 and
the respective azido derivative 6. It can be observed from the oscillator strength
values, that the S1 state for 4 is a bright state, and the S2, S3, S4 and S5 states
are dark states, due to their small oscillator strengths. In this way, it can be
predicted that the excitation undergoes from S0 to S1, once S1 presents the
highest oscillator strength, i.e., the transition dipole moment is the highest for
the S0→S1 transition. On the other hand, in the azide 6, the state which
17

presents the highest oscillator strength is the S2, and S1, S3, S4 and S5 are the
dark ones. Thus, it can be assumed for this derivative that the excitation goes
from S0 to S2. Taking the Kasha's rule into account, the azide derivative is
excited to the S2 state, and then, goes down to the S1 state via internal
conversion, due to the very close energies of both states.
Table 3. Absorption wavelength, oscillator strength (ƒosc) and the main
configuration of the lowest excited states calculated using CAM-B3LYP/jun-cc-
pVTZ for amino and azido benzazoles 4 and 6, respectively.
Compound
State
S1
Character
¹ππ*
λabs
ƒosc
298.2
266.5
242.6
233.4
226.1
305.2
299.3
258.5
244.1
234.9
1.4820
0.0220
0.0680
0.0013
0.0241
0.0003
1.2733
0.0009
0.0074
0.0271
S2
¹ππ*
4
S3
¹ππ*
S4
¹ππ*
S5
¹ππ*
S1
¹ππ*
S2
¹ππ*
6
S3
¹ππ*
S4
¹ππ*
S5
¹ππ*
It can be found in the literature that the observed fluorescence quenching
in molecules containing the azido moiety is due to Photoinduced Charge
Transfer (PET) [70,71]. In order verify if the azido derivative 6 also presents
PET, the overlap integral between the centroids of charges (C+ and C−) before
and after transition, obtained from the charge transfer diagnostic index (DCT),
were computed in dichloromethane and ethanol. After the transition, the
electronic density tends to be relocated considering the donor and acceptor
groups in the molecule. The results of DCT analysis are shown in Figure 5. For
both solvents, the centroids of charges before and after transition are localized
in the same region with no displacement, excluding the charge transfer chance
for this azide.
18

Figure 5. DCT overlap integral between centroids of charges obtained before
and after the transition computed with the Multiwfn Program for azide 6 in
dichloromethane (left) and ethanol (right).
Table 4 shows the theoretical absorption and emission wavelengths, the
absorption and emission oscillator strengths and the dipole moment of the
ground and excited states for molecules 4 and 6 in acetonitrile, dichloromethane
and ethanol. The theoretical values of absorption and emission are relatively
close to the experimental ones. It can be observed that the solvent does not
affect the absorption and emission wavelengths, as experimentally observed.
As mentioned, the amine 4 presents a bright state at S1 and the azide 6
at the S2 for the ground state geometry. For this reason, the excitation energy of
S0 to S1 was considered for the azide 4 and S0 to S2 energy for the molecule 6.
Despite the S1 state is not a bright state for the molecule 6, considering the
Kasha’s rule, it is possible to conclude that this state is quickly reached by
internal conversion, and due to that, this state was optimized in order to obtain
the theoretical emission values. As expect, the emission oscillator strength of
amine 4 was high, explaining the high fluoresce emission experimentally
observed. On the other hand, the emission oscillator strength of azide 6 was
also high, an unexpected behavior due to the low fluoresce emission observed
experimentally for this molecule. When the first excited state geometry was
taken into account, the orbitals involved in the S1→S0 transition for molecule 6
were the HOMO and LUMO, and for the S2→S0 transition, the
HOMO→LUMO+1 prevails. On the other hand, when the geometry of the
ground state was taken into account, the orbitals involved in the S0→S1
transition for molecule 6 were the HOMO and LUMO+1 and for the S0→S2
transition, the HOMO→LUMO prevails. These results demonstrate an inversion
between S1 and S2 states when the first excited state of molecule 6 has its
19

geometry relaxed. The orbitals involved in the S0→S1 and S0→S2 transitions of
molecule 6 for the ground state geometry and first excited state geometry are
shown in Figure 6. This figure clearly shows the inversion of the S1 and S2
states, wherein for the ground state geometry, the S1 is the dark state and the
S2 the bright state, while for the first excited state geometry, the S2 is the dark
state and the S1 is the bright state.
20

Table 4. Calculated photophysical data (CAM-B3LYP / jun-cc-pVTZ) for amine 4 and azide 6 in dichloromethane, ethanol and
acetonitrile, where λabs and λem are the absorption and emission wavelengths (nm), µS0 is the ground state dipole moment (Debye)
and µS1 is the first excited state dipole moment (Debye). The parameters λabs, ƒabs and µS0 were obtained from the ground state
optimized geometry and the λem, ƒem and µS1 were obtained from the first excited state optimized geometry.
ƒem
Molecule
Solvent
ƒabs
µS0
µs1
λem
λabs (S0→S1)
298.52
(S1→S0)
Dichloromethane
Ethanol
1.1735
1.1541
1.1482
4.2426
4.4120
4.4362
374.15
380.04
381.13
1.4573
8.0159
8.6454
8.7587
4
Molecule
6
298.41
1.50361
1.5117
ƒem
Acetonitrile
298.23
Solvent
ƒabs
µS0
µs1
λabs (S0→S2)
300.33
λem
(S1→S0)
Dichloromethane
Ethanol
1.2991
1.2792
1.2733
1.6788
1.7098
1.7156
383.11
385.88
386.76
1.5618
2.0216
1.9816
1.9646
299.55
1.6001
1.6074
Acetonitrile
299.32
21

As mentioned, the amine 4 presents a bright state at S1 and the azide 6
at the S2 for the ground state geometry. For this reason, the excitation energy of
S0 to S1 was considered for the azide 4 and S0 to S2 energy for the molecule 6.
Despite the S1 state is not a bright state for the molecule 6, considering the
Kasha’s rule, it is possible to conclude that this state is quickly reached by
internal conversion, and due to that, this state was optimized in order to obtain
the theoretical emission values. As expect, the emission oscillator strength of
amine 4 was high, explaining the high fluoresce emission experimentally
observed. On the other hand, the emission oscillator strength of azide 6 was
also high, an unexpected behavior due to the low fluoresce emission observed
experimentally for this molecule. Taking the first excited state geometry into
account, the orbitals involved in the S1→S0 transition for molecule 6 were the
HOMO and LUMO, and for the S2→S0 transition, the HOMO→LUMO+1
prevails. These results demonstrate an inversion between S1 and S2 states
when the first excited state of molecule 6 has its geometry relaxed. The orbitals
involved in the S0→S1 and S0→S2 transitions of molecule 6 for the ground state
geometry and first excited state geometry are shown in Figure 6. This figure
clearly shows the observed inversion of the S1 and S2 states, wherein for the
ground state geometry, the S1 is the dark state and the S2 the bright state, while
for the first excited state geometry, the S2 is the dark state and the S1 is the
bright state.
22

Figure 6. Kohn-Shan orbitals involved in the first and second transition for
molecule 6 computed at the CAM-B3LYP/jun-cc-pVTZ level in acetonitrile for S0
geometry (top) and S1 geometry (bottom).
TDDFT results did not provide sufficient answers to explain to explain the
lack of florescence of molecule 6. In order to complement the computational
photophysical study, multiconfigurational calculations were also applied. In
these calculations, it was used an active space including 12 electrons and 12
orbitals as shown in Figure 7.
23

Figure 7. Active space of the azido benzazole 6 calculated with SA(4)-
CASSCF(12,12)/ANO-S-VZDP.
Table 5. Vertical excitation energies (∆Evert), oscillator strength (ƒosc) and
character of the first four singlet states of azido derivative 6 with geometry
optimized at SA3-CASSCF/ANO-S-VDZP and energy corrected at MS5-
CASPT2/ANO-L-VDZP levels of theory.
State
S1
Character
¹ππ*
∆Evert (eV)
4.280
ƒosc
0.000520
1.108350
0.002934
0.000334
S2
¹ππ*
4.282
S3
¹ππ*
5.662
S4
¹ππ*
6.715
The vertical transitions presented in Table 5 show that the S1 and S2
states energies are very close, and we can consider that are degenerate. In
addition, the S3 state is about 1.4 eV higher than S2 and S1, and S4 state is
about 2.5 eV higher than S2 and S1. Similar to the results obtained from TD-
DFT, the oscillator strength of the S2 state stands out from the other states and
the S1 and S2 state presented a quite similar energy. In this sense, it can be
supposed that the excitation undergoes from S0 to S2. The S0→S2 vertical
24

transition of molecule 6 had only one major contribution involving the f and g
orbitals (Figure 7), this is a HOMO→LUMO transition, as well as observed in
the TD-DFT results. Figure 8 shows the orbitals involved in the S0→S1 and
S0→S2 transitions.
Figure 8. Orbitals involved in the first and second transition for molecule 6
computed at the CASSCF/ANO-L-VDZP level.
The orbital coefficients of CASSCF and TD-DFT involved in the S0→S1
and S0→S2 vertical transitions of molecule 6 (Figure 6 and 8, respectively) are
similar, in other words, the dark state and the bright state computed at
CASCF/CASPT2 and TDDFT are very comparable. However, the CASPT2
predicts almost an absorption maximum of 290 nm (4.28 eV) for the bright state.
On the other hand, the TDDFT value is 299 nm (4.15 eV) for the bright state
and experiment result is 325-327 nm (3.81-3.79 eV). It is important to
emphasize that the TD-DFT calculations take into account the solvent effects,
while CASPT2 does not. Furthermore, we do not have gas-phase UV spectrum
in order to compare with the theoretical results and take into account the solvent
effects in CASPT2 calculations demands a lot of computational efforts.
Moreover, take into account the similarity of the TDDFT and CASPT2 results,
we can infer that the systems are been describing suitably with the active space
employed.
By analyzing the multiconfigurational calculations, we can
predict that the S2 state is populated after irradiation and, by internal
conversion, it populates the S1 state almost instantly. For these reasons, we
performed geometry optimization of the S1 state for molecule 6, but this
25

calculation reached geometry wherein the S0 and S1 presented energies quite
similar. In this way, a Minimal Energy Crossing Point (MECP) calculation was
performed and thus, two points were obtained: the first one is related to the
ground state minimum and the second one is related to the conical intersection
(CoIn) point. No minimum in the S1 surface close to the FC region was
achieved, and because of that, it goes directly to the global minimum in the
surface: the CoIn point. In order to support these results, a MEP coordinate
from the S1 state was constructed with CASPT2//CASSCF protocol, i.e., all
points were optimized at SA3-CASSCF(12,12)/cc-pVDZ level and energy
corrected at MS5-CASPT2(12,12)/ANO-L-VDZP level (Figure 9).
Figure 9. Potential energy curve for N1-N2 bond reaction coordinate construct
by MEP for molecule 6 at SA3-CASSCF/cc-pVDZ level, energy corrected at
MS5-CASPT2/ANO-L-VDZP level (CASPT2//CASSCF protocol).
Figure 9 shows that S1 state is dissociative, since its curve
presented no barrier to overcome. Thus, azide
6 reaches the CoIn and
quickly relaxes back to the ground state. It explains the observed intrinsic
quenching of fluorescence in this compound, since it decays by a CoIn,
26

which is a nonradioactive process. Furthermore, it is possible to verify that
the electronic excitation of this system involves the occupation of an
antibonding orbital on the azide group, leading to a increasing in the bond
length N-N2. Due to this change, the azide
6 reaches to a CoIn point and
returns to the ground state. The deactivation mechanism of this system
involves degenerate points, such as the CoIn point. Here, it is important to
emphasize once again that previous studies in the literature indicated that
the fluorescence quenching in azido-based molecules were related to
photoinduced charge transfer mechanism (PET). In contrast to the
obtained results from TDDFT, the multiconfigurational methodology
presented a notable difference between the geometries of the ground
state and the global minimum of the first excited state, which is the conical
intersection point. The main difference in the geometry occurs in the azide
group, with increasing of the bond length between N1 and N2 in almost
0.4 Ǻ at the conical point and the decreasing in the angle among the three
nitrogen atoms. In order to understand why the S1 state in
CASSCF/CASPT2 was dissociative while in TDDFT was not. It was
carried out a geometry optimization of the S2 state of molecule
6 at CAM-
B3LYP/cc-pVDZ level. This calculation reached a very similar geometry to
the CoIn geometry as computed at CASSCF/CASPT2 level, as it is
possible to see in the Figure 10. It was performed a single point
calculation of that geometry and it presented a ∆Evert of 0.57 eV (2171
nm) between S0 and S1. Despite the low ∆Evert, this value is very high to
be considered
a
CoIn, however, degenerate points present
multiconfigurational character, which cannot be properly described with
monoconfigurational methods as TDDFT. In this way, it is necessary to
use multiconfigurational methods in order to properly describe this
phenomenon. The multiconfigurational methodologies (CASSCF and
CASPT2) allowed explaining such particular photophysical behavior of
this
derivative.
Thus,
with
the
combination
between
the
multiconfigurational and TDDFT results, we can conclude that the dark
state involved in the excitation between the HOMO and the π orbitals,
which involves the azido group (LUMO+1 of TDDFT calculations and the
27

LUMO+2 of CASPT2 calculations), is dissociative. When the geometry of
the first excited state was relaxed at CAMB3LYP/cc-pVDZ level, it was
observed the inversion of the S1 and S2 states. Due to this inversion, the
S1 does not reach the dissociative path, because the dark state becomes
the S2, which are the dissociative one.
Figure 10. S2 relaxed geometry of azido 6 calculated with CAM-B3LYP/cc-
pVDZ level in acetonitrile.
The Figure 11 presents a qualitative deactivation pathway to the excited
state of azide 6, as well as, the orbitals involved in this process. In our
proposal, the bright state (S2) is populated after irradiation and by internal
conversion the dark state (S1) is reached. The dark state is dissociative and
reaches a CoIn and returns to the ground state by a non-radioactive process.
28