The impact of the heteroatom in a five-membered ring on the photophysical properties of difluoroborates

Article abstract of DOI:10.1016/j.dyepig.2019.04.026

A series of novel BF₂ complexes, bearing a five-membered heterocyclic ring (with X = NMe, O, and S), were synthesized and characterized with a focus on the influence of atom exchange on the photophysical properties of both unsubstituted and dim

Full text of DOI:10.1016/j.dyepig.2019.04.026

Accepted Manuscript
The impact of the heteroatom in a five-membered ring on the photophysical properties
of difluoroborates
Anna Grabarz, Beata Jędrzejewska, Agnieszka Skotnicka, N. Arul Murugan, Filip
Patalas, Wojciech Bartkowiak, Denis Jacquemin, Borys Ośmiałowski
PII:
S0143-7208(19)30399-7
https://doi.org/10.1016/j.dyepig.2019.04.026
DYPI 7481
DOI:
Reference:
To appear in: Dyes and Pigments
Received Date: 19 February 2019
Revised Date: 10 April 2019
Accepted Date: 10 April 2019
Please cite this article as: Grabarz A, Jędrzejewska B, Skotnicka A, Murugan NA, Patalas F, Bartkowiak
W, Jacquemin D, Ośmiałowski B, The impact of the heteroatom in a five-membered ring on the
photophysical properties of difluoroborates, Dyes and Pigments (2019), doi: https://doi.org/10.1016/
j.dyepig.2019.04.026.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. 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.

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT
1
2
The Impact of the Heteroatom in a Five-Membered Ring on the
Photophysical Properties of Difluoroborates
a,
b
b
c
3
4
Anna Grabarz , Beata Jędrzejewska , Agnieszka Skotnicka , N. Arul Murugan , Filip
d
a
e,∗
d,∗
Patalas , Wojciech Bartkowiak , Denis Jacquemin , Borys Ośmiałowski
5
6
7
8
9
0
1
2
3
aDepartment of Physical and Quantum Chemistry, Faculty of Chemistry, Wrocław University of
Technology, Wyb. Wyspiańskiego 27, PL–50370 Wrocław, Poland
Faculty of Chemical Technology and Engineering, UTP University of Science and Technology,
b
Seminaryjna 3, PL-85326 Bydgoszcz, Poland
c
Division of Theoretical Chemistry and Biology, School of Biotechnology, Royal Institute of
1
1
1
1
Technology, SE–10691 Stockholm, Sweden
d
Faculty of Chemistry, Nicolaus Copernicus University, Gagarina 7, PL-87100 Toruń, Poland
e
Laboratoire CEISAM, UMR CNRS 6230, Université de Nantes, 2 Rue de la Houssinière, BP92208,
44322 Cedex 3 Nantes, France
1
4
Abstract
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
A series of novel BF2 complexes, bearing a five-membered heterocyclic ring (with X=NMe,
O, and S), were synthesized and characterized with a focus on the influence of atom ex-
change on the photophysical properties of both unsubstituted and dimethylamino deriva-
tives. The experimental results show that the optical spectra substantially differ in both
sets of dyes. In particular, the dimethylamino series are more strongly affected by het-
eroatom substitution, i.e., the insertion of X=O or X=S in lieu of X=NMe causes sub-
stantial bathochromic shifts of the absorption and emission bands, as well as marked
changes in their topologies. In contrast, the optical spectra of the unsubstituted com-
pounds undergo only relatively small redshifts, and no variation of band shapes is ob-
served. Moreover, the measured absorption spectra of the unsubstituted compounds
bearing X=NMe and X=O are almost identical. Interestingly, the fluorescence yields
of the dimethylamino derivatives are much larger (up to one order of magnitude) than
those of the corresponding unsubstituted compounds. The experimental analyses are
supported by state-of-the-art quantum chemistry calculations, which satisfactorily re-
produced the experimental trends and provided further insights into the observed optical
signatures.
1

ACCEPTED MANUSCRIPT
3
1
Keywords: TD-DFT, difluoroborates, functionalization, charge transfer
∗
Corresponding author
Email addresses: Denis.Jacquemin@univ-nantes.fr (Denis Jacquemin),
borys.osmialowski@umk.pl (Borys Ośmiałowski)
Preprint submitted to Dyes & Pigments
April 12, 2019

ACCEPTED MANUSCRIPT
3
2
1. Introduction
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
Structure-optical property relationships are now established for many classes of dyes,
and one can often safely predict the evolution of absorption spectra with chemical sub-
stitution. This is more difficult to achieve for emission, which is a key challenge in view
of its use in many biomedical imaging applications. Indeed, there is a need to design new
fluorescent probes, effectively exhibiting sensitivity to environmental changes, physiolog-
ical conditions and/or to specific interactions ongoing in biomacromolecules. To achieve
these goals, countless syntheses have been carried out to functionalize fluorophores and
tune their emissive properties, which can be achieved through standard variation of
the substituent(s) [1–6], benzannulation [7–12], elongation of the π-conjugated spacer
separating the donor and acceptor moieties in push-pull systems [13–17], or introduc-
tion of additional chromophores into their structure [18–20]. However, the addition of
several substituents may lead to undesired effects such as the decrease of the fluores-
cence quantum yield (Φfl) inherent to flexible side group that enhance the efficiency of
non-radiative pathways. This explains why the most potent fluorescent molecules re-
main “simple and rigid” derivatives, e.g., xanthenes [21–25], cyanines [26–30], BODIPYs
[26; 31–35], and coumarins [36–40]. Among these classes, BODIPY derivatives occupy
a privileged position [41], because they can be easily functionalized and are very pho-
tostable, explaining the undisputed popularity in this family [42–46]. However, one of
the main disadvantages of regular BODIPY fluorophores is the impossibility to directly
doubly-benzannulate their five-membered rings. Indeed, while single benzannulation has
been achieved [47–49] and has shown to be an efficient strategy for red-shifting their
absorption and fluorescence maxima, double benzannulation has only been achieved at
other positions [50–53]. It is also well recognized that atom exchange, leading to iso-
electronic compounds, may significantly impact the photophysical signatures of dyes. In
fluoroborate dyes, substituent, benzannulation, and π-extension have been extensively
applied to tune the photophysical properties, but atom exchange has been relatively less
studied, though fluorescent techniques are also well suited to assess the impact of the
3

ACCEPTED MANUSCRIPT
6
6
6
6
6
6
6
0
1
2
3
4
5
6
changes induced by atom-by-atom substitutions. In the classes of dyes encompassing
atom-substituted BODIPYs, one should, of course mention: i) aza-BODIPY dyes – with
a nitrogen atom at the meso position – that shows red-shifted optical bands compared to
standard BODIPYs [54–58]; ii) other difluoroborates showing one or both nitrogen atoms
in their N-BF2-N core replaced by oxygen atoms, e.g., boranils [59–62], hydroxyphenyl-
benzoxazole (HBO) complexes [63–65], difluoroboron β-diketonate complexes [66–68],
and others [42; 69–71] (see Figure 1).
Figure 1: The common derivatives of BODIPY dyes, from left to right: aza-BODIPY,
boranils, HBO complexes and difluoroboron β-diketonate complexes.
6
6
6
7
7
7
7
7
7
7
7
7
7
8
8
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
Notably, in contrast to standard BODIPY compounds, some of these derivatives, e.g.,
HBO complexes and boranils, are characterized by improved Stokes shifts (∆ss), usually
larger than 50 nm [72–74], whereas the ∆ss typically do not exceed 20 nm in standard
BODIPYs [31; 41; 75]. Additionally the emission range of those compounds can be
further red-shifted by extending the π-conjugated structure [76; 77]. This suggests that
additional efforts to synthesize and understand the impact of atom substitution in this
class of dyes are of interest. In a recent work, we have studied the impact of atom
exchange directly bonded to the boron atom (OBF2N versus SBF2N) [78]. While this
provided the first example of fluorophores of the BF2 family with a S-B bond, these
molecules were composed of six-membered rings only. In this context, our goal here is to
synthesize NMe2-bearing fluoroborate derivatives containing one five-membered ring and
to compare their photophysical properties with their unsubstituted analogues (Scheme 1).
One can expect that the impact of atom-exchange is stronger in five-membered ring
than in their six-membered counterparts, because of the higher strain in the former
compounds. We underline that atom exchange does not influence only the electronic
4

ACCEPTED MANUSCRIPT
8
8
8
8
8
8
2
3
4
5
6
7
properties but also impacts the vibrations in the five-membered ring and, in turn, can
modify the vibronic couplings and hence the absorption and fluorescence band shapes [79;
80]. In more detail, we have three goals here: i) studying the effect of atom exchange on
fluorophores incorporating or not a strong donor group; ii) probing the electron accepting
properties of three different five-membered heterocycles; and iii) gaining insights into the
electronic structure of the newly synthesized compounds using first principles calculations
Scheme 1: The general structure of BF2-complexes of various heterocycles studied herein.
8
8
8
9
2. Experimental and theoretical methods
2.1. Experimental methods
9
9
9
9
9
9
9
9
0
1
2
3
4
5
6
7
The 2-phenacylheterocycles [49; 81] were first obtained and next converted to their
BF2 complexes following the procedure established for similar derivatives [50; 82]. For
1
11
13
15
19
all presented structures NMR data ( H, B, C, N and F) have been obtained (see
Experimental Section at the end of this work). The fluorescence yields were determined
in chloroform as follows. The fluorescence spectrum of dilute (A ≈ 0.1) solutions of
the compounds were recorded by excitation at the absorption band maximum of the
reference (see Table S3 in the SI). The quantum yield of the tested compounds (Φdye)
was calculated using the following equation [83]:
2
IdyeAref ndye
Φdye = Φref
·
Iref Adye n2
ref
9
9
8
9
where Φref is the fluorescence yield of the reference, Adye and Aref are the absorbance
of the dye and reference samples at the excitation wavelengths, Idye and Iref are the
1
00 integrated emission intensity for the dyes and references samples, ndye and nref are the
5

ACCEPTED MANUSCRIPT
1
1
1
01 refractive indices of the solvents used for the dyes and reference, respectively. Knowing
02 the (average) lifetime of the S1 excited state (τav) and the fluorescence quantum yield
03 (Φfl), the radiative (kr) and nonradiative (knr) rate constants were calculated using:
Φfl
kr =
τav
1
− Φfl
τav
knr =
1
04 2.2. Computational details
The geometries of the six studied compounds were optimized in their electronic ground
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
05
06 state (GS) at the B3LYP/6-311++G(d,p) level of theory. The effects of the solvent
07 (CHCl3 solution is used consistently with experiment, except when noted) were included
08 by means of the linear-response variant of polarizable continuum model (LR-PCM) [84].
09 The minimal nature of the optimal GS geometries was confirmed by evaluation of the
10 Hessian. These minimum-energy geometries were used in all subsequent molecular dy-
11 namics simulations described below. In order to determine the vertical electronic spectra
12 we have applied the state-of-the-art polarizable-embedded resolution-of-identity second-
13 order coupled-cluster (PERI-CC2) method [85]. In more detail, we firstly carried out
14 molecular dynamics (MD) calculations for BF2-complexes in chloroform solvent and the
15 trajectories were subjected to subsequent PERI-CC2/MM calculations. During the MD
16 simulations, the interactions between the solute and solvent subsystems were described
17 considering both electrostatic and van der Waals interactions, and this requires charges
18 and force field parameters for each of the molecules. The electrostatic potential fitted
19 charges for the difluoroborates were obtained using the B3LYP/6-311++G(d,p) level of
20 theory and CHEPLG procedure [86] as implemented in Gaussian 09 software. To de-
21 scribe the van der Waals interactions, we employed General AMBER Force Field (GAFF)
22 [87] and CHCL3.frcmod for BF2-carrying and chloroform solvent molecules, respectively.
23 Selected chloroform force field includes both bond length and bond angle parameters
24 making the solvent flexible. Even though GAFF provides a flexible molecular model
6

ACCEPTED MANUSCRIPT
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
25 for solutes, we treated the difluoroborates molecules with rigid body approximation.
26 The simulations for BF2-carrying compounds in chloroform encompassed a single solute
27 molecule and a few thousands of solvent molecules. The calculations were carried out
28 at room temperature and 1 atmospheric pressure using the Amber16 software [88]. The
29 calculations were performed in an isothermal-isobaric ensemble, so that the density of
30 the solute-solvent system evolves to the right value independent of the initial density.
31 The maintenance of the system at a specific temperature and pressure has been achieved
32 by connecting the system to Langevin’s thermostat and Barendsen’s barostat. The time
33 step for integration of equation of motion was set to 2 fs and the total time scale for the
34 production run was 10 ns. The convergence of key properties (energies and density) in
35 each simulation was checked. Fifty configurations from the last 5 ns simulations selected
36 at equal intervals have been used for computing the spectra using PERI-CC2 method
37 which describes the interaction between the solute-solvent subsystems using electrostatic
38 embedding. The same force-fields as in the MD simulations were used to describe the
39 solvents during the PERI-CC2 calculations. The PERI-CC2 calculations were carried
40 out by using the Turbomole package [89].
41
Since excited-state (ES) properties, e.g., electronic density changes upon electronic
42 excitation, are not available yet at the PERI-CC2 level, we have employed the Kohn-Sham
43 formulation of time-dependent density functional theory (TD-DFT) for their determina-
44 tion. The DFT and TD-DFT calculations were performed with the Gaussian 09 package
45 [90], employing 11 different exchange-correlation functionals (XCF), i.e., BLYP [91; 92],
46 B3LYP [93], PBE0 [94; 95], M06 [96], BHandHLYP [97], CAM-B3LYP [98], M06-2X
47 [96], ωB97X [99], ωB97X-D [100], LC-BLYP [101], and M06-HF [102; 103]. Five excited
48 states were considered. For all DFT/TD-DFT calculations, we used so-called ultrafine
49 integration grid. The density difference plots were simulated with a 0.002 au contour
50 threshold, with density increase (decrease) indicated by red (blue) zones. Subsequent
51 charge-transfer parameters, namely the distance on which charge is transferred (dCT), the
52 amount of charge (qCT) and the dipole moment change upon transition (µCT = µES−µGS)
7

ACCEPTED MANUSCRIPT
1
1
1
1
53 were calculated using Le Bahers’ procedure described in details elsewhere [104; 105]. In
54 this procedure, the CT parameters are determined using the barycenters of density in-
55 crease and decrease upon photon absorption. All calculations of transition energies were
56 performed with the aug-cc-pVDZ basis set [106–108].
1
57 3. Results and discussion
1
1
1
1
1
1
1
1
1
1
1
58
The measured absorption and emission spectra of all six compounds obtained in
59 chloroform are given in Figure 2 and the associated data are listed in Tables 1 and 2. Let
60 us first discuss the results obtained for the absorption spectra. Obviously, the absorption
61 bands of the unsubstituted compounds (R=H) are blue-shifted compared to the R=NMe2
62 dyes, a statement holding irrespective of the X heteroatom. Indeed, for all unsubstituted
63 compounds (X=NMe, O, and S) the absorption band peaks at 350–375 nm, whereas for
64 R=NMe2 the span is twice as large (400–450 nm). In addition, both the shapes and the
65 positions of the absorption bands of the unsubstituted X=NMe and X=O compounds
66 are highly similar, whereas the corresponding NMe2-substituted derivatives show more
67 pronounced discrepancies in their band topologies. Introducing the donor group also
68 yields a significant evolution of the band shape for the thio derivative.
Table 1: Photophysical properties of examined compounds measured in chloroform.
a
fl
b
Struct.
λabs
λfl
ε
∆ss
Φ
Φ
fl
−
1
−1
−1
X
R
[nm] [nm] [M cm
]
[cm
]
[%] [%]
NMe
O
S
NMe NMe2
O
S
H
H
H
349
351
369
398
411
447
420
424
444
443
457
480
21400
4844
4905
4578
2552
2449
1538
< 1 < 1
< 1 < 1
27000
26300
28000
42500
59800
1
< 1
46
49
98
70
NMe2
NMe2
86
68
a
b
-
quantum yield measured with the use of the references, - quantum yield measured
with the use of integrating sphere.
1
69
As can be seen in Table 1, the measurements of the fluorescence quantum yields for
8

ACCEPTED MANUSCRIPT
Figure 2: Experimental absorption and emission spectra in CHCl3.
1
1
1
1
1
1
1
1
1
1
1
1
1
70 R=NMe2 dyes returned very high values (Φfl of 48.7, 98.2, and 70.0% for X=NMe, X=O
71 and X=S respectively) thereby representing huge increases compared to their unsubsti-
72 tuted counterparts (all <1%). Clearly, the enhancement of Φfl values is also dependent
73 on heteratom type, since in both the R=NMe2 and R=H series, Φfl are improved fol-
74 lowing NMe→O and/or NMe→S substitution. At this stage, we hypothesize that the
75 vibrations and rotations around the single bonds separating the BF2 scaffold and the
76 side phenyl ring (see Scheme 1) that are likely to quench the emission of the dyes in
77 solution, are significantly limited when an amino group is added. Consistently with the
78 measured quantum yields, enhanced fluorescence lifetimes are observed for the NMe2
79 substituted dyes (see Table 2 and the fluorescence decay profiles in Figures S10–S15
80 in the SI). For the unsubstituted compounds, two exponential fluorescence decays were
81 observed (See Table 2) with a major component close to 70–90 ps and another at ca.
82 0.9–1.8 ns. In contrast, in the NMe2 series, the fluorescence decay profiles fit well with
9

ACCEPTED MANUSCRIPT
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
2
83 a single-exponential function, giving fluorescence lifetimes of 0.86, 1.59 and 1.78 ns for
84 X=NMe, O, and S, respectively. In other words, in the amino series, the fluorescence
85 lifetimes increase along with the NMe→O→S pattern. From the Table 2 one also notices
86 that the radiative rate constant (kr) increases when the compounds are substituted by
87 electron releasing substituent NMe2, i.e., the enhancement of Φfl originates in both the
88 expected decrease of knr (vide supra) and the increase of kr. The differences in the kr
89 values for R=H series are not significant but the kr/knr ratio between is two- or three
90 orders lower than unity, indicating that the non-radiative deactivation of S1 is efficient.
91 The situation is opposite in the R=NMe2: the kr/knr ratio is close to unity for X=NMe
92 derivative and becomes even higher in the X=S and X=O derivatives.
93
Our next step was to check how solvatochromic effects affect the properties of R=H
94 and R=NMe2 derivatives, thus we measured absorption and emission properties of the
95 examined dyes in three solvents characterized by different dielectric constant. The elec-
96 tronic absorption and fluorescence spectra recorded in methylcyclohexane (MCH), N,N-
97 dimethylformamide (DMF) and tetrahydrofuran (THF) can be found in Figures S8–S9
98 in the SI whereas the corresponding spectroscopic and photophysical properties are col-
99 lected in Tables S5 and S6. As can be seen in Figure S8, the UV-vis absorption spectra of
00 the unsubstituted difluoroborates (R=H) do not display any significant variation of their
01 absorption maxima position, when changing the solvent dielectric constant. In contrast,
02 the absorption maxima of the dimethylamine-substituted derivatives undergo positive
03 solvatochromism, e.g., a noticeable red-shift from 439 nm in MCH to 450 nm in DMF
04 is found for the X=S R=NMe2 compound (see Figure S9 in the SI). These red-shifts of
05 the absorption maxima position along with the high molar absorption coefficients indi-
∗
06 cate that the transitions of the R=NMe2 derivatives have a π − π nature associated
07 with a significant charge transfer (CT) (see Table S5). Similarly, a considerable redshift
08 driven by a solvent dielectric constant change is also observed in the emission spec-
09 tra of X=O R=NMe2 and X=S R=NMe2 compounds, whereas no significant shift was
10 found in the fluorescence spectra of unsubstituted difluoroborates, indicating that the
1
0

ACCEPTED MANUSCRIPT
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
11 CT character of the first singlet excited state is much more marked in the dimethylamine
12 derivatives [109; 110]. Interestingly, the emission spectra of the X=NMe R=NMe2 dye
13 display significantly smaller solvatochromic shifts, than their benzoxazole and benzoth-
14 iazole analogues. In short, the redshifts of the emission maxima positions, as well as
15 higher sensitivity of fluorescence spectra (with respect to absorption) to the dielectric
16 constant of the environment are characteristics of push-pull compounds undergoing CT
17 transitions [109–112].
18
As can be seen in Table S5, the fluorescence quantum yields of the unsubstituted
19 dyes remain very small (<1%), irrespective of the solvent dielectric constant. Obviously,
20 such trifling values do not allow for an unambiguous interpretation of the impact of the
21 solvent dielectric constant on the Φfl; the trends being however similar as in CHCl3. In
22 contrast, the R=NMe2 dyes fluoresce intensively in all studied solvents. The Φfl of the
23 X=S R=NMe2 compound is systematically decreasing with increasing solvent dielectric
24 constant, which is a usual trend in dyes exhibiting large charge separation in their excited
25 state [113; 114]. Interestingly, the opposite is observed for X=NMe R=NMe2 dye, in
26 which fluorescence quantum yield increases from 14.5% (MCH) up to 56.5% (DMF). It
27 is not easy to interpret this peculiar behavior but we hypothesize that it is related to
28 the motion of the additional methyl group that would be somehow diminished in DMF.
29 Indeed, such improvement of the fluorescence quantum yield with increasing in solvent
30 dielectric constant, also known as negative solvatokinetic effect, might be connected
31 with interactions between the excited solute and solvent molecules, which changes the
32 excited-state configuration and the fluorescence intensity [115–120]. However, the above
33 hypothesis needs more precise studies in a wider range of solvents and is out of the scope
34 of the current work. Eventually, for X=O R=NMe2, the Φfl is similar in THF (68.9%)
35 than in MCH (73.5%), but a lower efficiency is observed in the most polar DMF (52.5%)
36 , as usual. We also highlight that, irrespective of the selected solvent, the unsubstituted
37 compounds exhibit a two-exponential decay emission whereas the R=NMe2 dyes show
38 mono-exponential decays, but for X=NMe R=NMe2 in MCH.
1
1

ACCEPTED MANUSCRIPT
Table 2: Fluorescence lifetimes (τ), component amplitudes (α), correlation coefficients
2
(
χ ) and radiative (kr) and non-radiative (knr) rate constants for examined compounds
in chloroform.
Struct.
τ1
τ2
α1
α2
τav
χ2
kr
knr
kr/knr
[s ] × 10 [s−1] × 108
−
1
8
X
R
[ps] [ps]
[ps]
NMe H
72 1630 99.345 0.654 82.2 1.668
71 1981 98.785 1.215 94.2 1.439
86 1021 99.438 0.562 91.3 1.538
0.29
0.33
1.32
5.66
6.16
3.94
121
106
108
5.96
0.12
1.69
0.0024
0.0031
0.0120
0.95
53.6
2.33
O
S
H
H
NMe NMe2
-
-
-
861
1593
1775
-
-
-
100 861 1.395
100 1593 1.077
100 1775 1.172
O
S
NMe2
NMe2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
39
The fluorescence excitation and emission characteristics of all compounds have been
40 measured in THF at different emission and excitation wavelengths. As can be seen in
41 Figures S4 and S5 in the SI, the position of the fluorescence peaks remains constant
42 independently of both the λexc and concentration in the solution for all compounds.
43 This clearly suggests at a rapid internal conversion from higher excited states to the
44 lowest vibrational energy level of the S1 state as well as the presence of only one emitting
45 species. In addition, the fluorescence excitation spectra were recorded for two emission
46 wavelengths (detection at blue and red edges of the emission spectrum), as well as at
47 the maximum of the emission spectrum for diluted and more concentrated solutions.
48 Figures S6 and S7 display the steady-state fluorescence excitation spectra of the tested
49 compounds, and one can observe, on the one hand, spectra independent of the considered
50 wavelength, and, on the other hand, a non-negligible impact of the concentration of the
−
5
51 solution. The differences between absorption and excitation spectra of the dyes in 10
M
52 solutions show the existence of the excited molecules that differ in spatial conformations
53 and/or solvent relaxations.
54
All these experimental results encouraged us to carry out computer simulations in
55 order to shed light onto the electronic structure of these derivatives, one of our goals be-
56 ing to unravel the origin of the band shape differences: multiple electronic state and/or
57 various vibronic couplings. Our results are presented in Table 3. As can be seen, the
1
2

ACCEPTED MANUSCRIPT
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
58 first electronic excited state is well separated from the second one, with an energy gap
59 between them attaining ca. 1 eV for all studied molecules. In addition, the oscillator
60 strengths, f are much larger for the S0→S1 transition (from 1.07 to 1.59) than for the
61 S0→S2 transition (from 0.02 to 0.06). This clearly indicates that the observed spectra
62 are solely due to the first electronic ES. For that excitation, the NMe2-substituted com-
63 pounds exhibit larger oscillator strengths than their unsubstituted counterparts. We note
64 that the predicted f values are rather consistent with measured molar absorptivities (see
65 ε in Table 1): in both cases, significantly lower values are obtained for R=H derivatives.
66 Nevertheless, based on the measurements, one would have expected significantly lower
67 oscillator strength for the X=NMe R=NMe2 derivative. Figure 3 shows the simulated
68 vertical absorption spectra in the 300–550 nm range, considering the absorption bands of
69 the S0→S1 transition. As discussed above, the PERI-CC2 calculations were performed
70 based on 50 solute-solvent configurations (for each compound) considering rigid dyes and
71 flexible chloroform molecules. Therefore, the widths of the absorption bands shown in
72 Figure 3 originate from the inhomogeneous broadening due to solute-solvent interactions,
73 whereas the vibronic couplings are neglected, so that the band widths are logically under-
74 estimated compared to experiment. Nevertheless, by comparing the R=H and R=NMe2
75 derivatives, one clearly notices that the latter are much more sensitive to the solvent
76 microenvironment, explaining the larger inhomogeneous broadening of derivatives pos-
77 sessing a dimethylamino group (see the σ values in Table 3), a feature that can be directly
78 associated with an increase of the intramolecular charge-transfer character when adding
79 the electron-donating group. As neither the presence of multiple electronic states nor
80 explicit solvent-solute interactions can explain the presence of several maxima/shoulders
81 in the absorption spectra of Figure 2, these band shapes should reflect the changes in the
82 vibrational fine structure induced by chemical modifications. To confirm this hypothesis,
83 complementary calculations were performed using CC2 method, but in gas phase (see
84 Figure S3 in the SI). Although these calculations do not account for the solvent envi-
85 ronment, and hence, logically yield band positions that are blue-shifted compared to the
1
3

ACCEPTED MANUSCRIPT
2
2
2
86 measurements and PERI-CC2 results, one notices that the experimental band shapes are
87 reasonably well reproduced, confirming the vibronic nature of the multi-peak absorption
88 bands.
Table 3: Average vertical excitation energies (∆E), standard deviations of ∆E distribu-
∗
tions (σ(∆E) ) and average oscillator strengths (f) computed at the PERI-CC2/aug-cc-
p
∗
pVDZ level of theory in chloroform. FWHM=2 2 ln σ(∆E).
Struct.
R
∆E [eV] ∆E [nm] σ(∆E) [eV]
f
X
S0→S1
NMe
O
S
NMe NMe2
O
S
H
H
H
3.73
3.70
3.44
3.19
3.07
2.90
333
335
361
389
404
428
0.01
0.01
0.01
0.05
0.05
0.06
1.12
1.12
1.07
1.59
1.59
1.57
NMe2
NMe2
S0→S2
NMe
O
S
NMe NMe2
O
S
H
H
H
4.73
4.66
4.42
4.19
4.20
4.00
262
266
280
296
295
310
0.02
0.04
0.05
0.03
0.02
0.02
0.02
0.02
0.05
0.06
0.06
0.03
NMe2
NMe2
2
2
2
2
2
2
2
2
2
2
89
Let us now assess the accuracy of the PERI-CC2 method in reproducing the aux-
90 ochromic shifts on the absorption spectra. In Table 4 the experimental and simulated
91 spectral shifts of the S0→S1 absorption band maxima are listed. In that Table, we
92 used the X=NMe R=H compound as a reference. The experimentally determined shift
93 upon X=NMe→O replacement is trifling for R=H (+2 nm) but significant for R=NMe2
94 (+14 nm). The respective shifts predicted by the PERI-CC2 method are +3 nm and
95 +15 nm, thus demonstrating the excellent agreement with experimental data. Signifi-
96 cantly larger auxochromic effects are observed experimentally for the X=NMe→S sub-
97 stitution, i.e., +20 nm for R=H and +50 nm for R=NMe2, with again simulated shifts
98 in good agreement (+28 nm and +40 nm). One can therefore conclude that the overall
1
4

ACCEPTED MANUSCRIPT
Figure 3: Electronic vertical absorption spectra simulated at the PERI-CC2/aug-cc-
pVDZ level of theory.
2
3
3
3
3
3
3
3
3
3
3
3
3
3
3
99 agreement between the PERI-CC2 simulation results and the experimental data is, at
00 the very least, satisfactory, thus further consolidating the above conclusions regarding
01 the origin of the specific band shapes.
02
In order to obtain additional insights into the electronic structure of the dyes, we have
03 probed the nature of the excited state properties in the Franck-Condon region. Since
04 ES properties cannot yet be determined with the PERI-CC2 approach, we have turned
05 towards TD-DFT. The performance of TD-DFT in predicting ES properties is known to
06 significantly depend on the selected XCF [121; 122], and no single XCF can be consid-
07 ered as outperforming the others on a systematic basis. Hence, in order to select a XCF
08 providing accurate results for the molecules under investigation, we have benchmarked
09 eleven XCFs and compared the obtained GS and ES dipole moments to CC2 reference
10 values. The dipoles were chosen as they are a measure of the quality of the electronic
11 density and are also directly related to the CT character of the electronic transition.
12 These benchmark calculations were performed for isolated (no solvent) molecules. As
13 can be seen in Figure 4, the errors made by DFT for GS dipole moments (µGS) are sub-
1
5

ACCEPTED MANUSCRIPT
Table 4: Excitation energy changes, w.r.t. X=NMe R=H compound, induced by chemi-
cal modifications: comparisons between theory (PERI-CC2 and PCM-TD-B3LYP) and
experiment.
Struct.
S0→S1
exp. exp. PERI-CC2 PERI-CC2 PCM-TD-B3LYP PCM-TD-B3LYP
X
R
[
eV] [nm]
[eV]
[nm]
[eV]
[nm]
NMe H
0.00
-0.02 +2
-0.19 +20
0
0.00
-0.03
-0.29
-0.54
-0.66
-0.83
0
+3
+28
+56
+71
+96
0.00
0.00
-0.17
-0.37
-0.42
-0.57
0
0
+18
+43
+50
+70
O
S
H
H
NMe NMe2 -0.43 +48
O
S
NMe2 -0.54 +62
NMe2 -0.78 +98
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
14 stantially smaller compared to the TD-DFT errors for the ES (µES), which is consistent
15 with previous benchmarks [123; 124]. As a consequence, the errors in the so-called excess
16 dipole moments (µGS-µES) are mainly originating from the TD-DFT part. For the µGS
17 the smallest errors are reached with the ωB97X XCF (smaller than 0.2 D), although
18 all XCFs provide satisfying results. In contrast, for the ES, ωB97X yields the largest
19 maximum deviations (ca. 3.0 D), whereas B3LYP provides the smallest maximum error
20 for both µES (1.3 D) and µES-µGS (1.9 D). B3LYP also gives one of the smallest average
21 errors for µGS (0.3 D), µES (1.2 D), and µES-µGS (1.3 D). In addition, this XCF accu-
22 rately reproduces the impact of chemical substitution on the ES dipoles, but for a slight
23 underestimation of the dipole increase when adding the dimethylamino group. Therefore
24 we select B3LYP for our analysis of the ES properties in CHCl3 solution. As can be seen
25 in Table 4, PCM-TD-B3LYP also fairly reproduces all auxochromic shifts though less
26 accurately than with PERI-CC2, as expected. We are well aware of the inherent limits
27 of B3LYP for describing long-range CT systems [122], but the molecules considered are
28 probably compact enough so that these limits are not relevant here.
29
Table 5 provides a summary of the TD-DFT ES properties computed in the Franck-
30 Condon region together with electron density difference (EDD) plots showing the elec-
31 tronic reorganization accompanying the transition to the lowest electronic excited state.
1
6

ACCEPTED MANUSCRIPT
Figure 4: Errors in TD-DFT dipole moments: maximum (∆max) and average (∆avg)
absolute errors compared to CC2 (see also corresponding Table S1 in the SI). All results
were obtained in gas phase with the aug-cc-pVDZ basis set.
3
3
3
3
3
3
3
3
3
3
3
3
32 As can be seen, for the R=H compounds, the electronic density changes mainly take
33 place within the BF2-containing core and adjacent five-membered ring. In contrast for
34 molecules bearing a NMe2 group, the density changes involve not only the BF2-carrying
35 ring but also the side NMe2-phenyl moiety, the amino group acting, as expected, as a
36 donor (mostly in blue in the EDDs). The CT character of the transitions in NMe2 deriva-
37 tives is reflected by the significant values of the dipole moment differences (µES - µGS)
38 which exceed 6 D for all three dyes, whereas significantly smaller values are obtained for
39 the R=H derivatives. Logically, the charge-transfer distances dCT are increasing upon
40 NMe2 substitution (up to dCT ≈ 2.7 Å). However as one can see the dCT values of
41 X=NMe R=H (dCT = 2.310 Å) and X=O R=H (dCT = 2.110 Å) derivatives are much
42 closer to their NMe2-substituted counterparts, than to X=S R=H (dCT = 1.539 Å). This
43 can be explained by looking at the EED of R=H derivatives collected in Table 5, as one
1
7

ACCEPTED MANUSCRIPT
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
44 can see both X=O and X=NMe derivatives have red (accepting) lobes on the bond con-
45 necting the fluoroborate core to the adjecent phenyl ring. This indicates that dCT in fact
46 should be larger for those two derivatives. Additionally, the R=NMe2 compounds are
47 also characterized by a larger transferred charge (qCT of ca. 0.50 e) than in their R=H
48 counterparts (qCT in the 0.41–0.43 e range). In the R=H series the CT character of the
49 electronic transition is clearly decreasing following X=NMe→O→S, which is confirmed
50 by a systematic decrease of both the dCT (NMe→O=-0.209 Å, O→S=-0.562 Å) and µES-
51 µGS (NMe→O=-0.395 D, O→S=-1.332 D) parameters. An interesting observation can
52 be made by comparing NMe2-substituted compounds, i.e., dCT and µES-µGS values are
53 lowest for X=O (dCT=2.713 Å, µES-µGS=6.537 D), while opposite to R=H, the highest
54 values are obtained for the thio derivatives (dCT=2.751 Å, µES-µGS=6.612 D). In turn
55 X=NMe dCT values are close to X=O, whereas µES-µGS results are similar to those of
56 X=S. Comparing the parameters of Table 5 to the fluorescence quantum yields, one no-
57 tices that the decrease of the dCT and µES-µGS promotes the radiative transition for all
58 three dyes of each series (R=H and R=NMe2), e.g., in the R=H series the quantum yield
59 decreases with increasing CT character. In contrast, Φfl is much higher in the NMe2
60 series, although the CT character is enhanced in the latter. We hypothesize that this
61 result comes from the fact that in the amino dyes, the CT involves the electron lone pair
62 of the amino nitrogen, which allows the formation of a quinoidal structure, and hence
63 lead to the rigidification of the excited state. That conclusion is also supported by the
64 single-exponential fit of the fluorescence decay (see Figures S10–S15 in the SI).
65
To further ascertain the CT character of the amino-substituted dyes, we calculated
66 the excited-state properties in Franck-Condon region of the benzothiazole derivatives
67 in three additional solvents: MCH, DMF and THF. The results of those calculations
68 together with the corresponding EDD plots are presented in Table S3 in the SI. As can
69 be seen, the above-described trends (obtained in chloroform) pertain, e.g., the S0→S1
70 transition is associated with a much smaller µES-µGS in the R=H (not exceeding 3 D)
71 than in the R=NMe2 dye (larger than 5.8 D). The EDD plots also tend to confirm the
1
8

ACCEPTED MANUSCRIPT
Table 5: Electron density difference plots and related electron transition parameters. All
results were obtained at B3LYP/aug-cc-pVDZ level of theory for chloroform solution.
Blue (red) color marks electron density depletion (gain) upon photon absorption. The
contour value was set to 0.002 au.
Struct.
R
dCT
qCT
µES − µGS
EDD plots
X
[Å]
[e]
[D]
NMe
O
H
H
H
2.310 0.427
2.101 0.430
1.539 0.407
2.725 0.504
2.713 0.502
2.751 0.500
4.734
4.339
3.007
6.595
6.537
6.612
S
NMe NMe2
O
S
NMe2
NMe2
3
3
3
3
3
3
3
3
3
72 CT nature of S0→S1 transition of the R=NMe2 derivatives: with density variations in
73 the R=NMe2 derivative showing a clear CT from the p-NMe2-phenyl moiety (donor) to
74 the BF2-bearing moiety (acceptor).
75
To further characterize the first excited state beyond the Franck-Condon region, we
76 have optimized the excited-state geometries of all compounds using TD-DFT. Let us first
77 discuss the changes in bond lengths between the two electronic states. The complete set
78 of data can be found in Table S2 in the SI. We focus our analysis on the geometrical
EG
bond
ES
GS
79 changes, as given by ∆
= bond −bond . A graphical representation of the general
80 variations of geometries can be found in Scheme 2. First, based on the comparison of
1
9

ACCEPTED MANUSCRIPT
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
81 the sum of absolute changes of bond lengths in the series, it appears that the S0→S1
82 excitation induces stronger effects for unsubstituted derivatives (R=H) than that for
EG
83 their substituted analogues (R=NMe2). Indeed, the total variations, Σ∆bond, are equal
84 to 1.249 Å and 1.157 Å, for the R=H and R=NMe2 series, respectively. This holds for the
85 bonds belonging to the heterocyclic part of the structures, confirming that the excitation
86 causes more dramatic changes in geometry for R=H, which is detrimental for emission
87 yields. In contrast, except for the C11-C14 distance, the opposite trends are observed for
88 1,4-phenylene ring moieties. This can be easily explained by the presence of NMe2 donor
89 which strongly affects the electron density distribution thus also the geometry of the
90 R=NMe2 dyes (See Figure S2 in the SI). Second, most of the computed changes in bond
91 distances follow a N→O→S ranking, e.g., for the C10-C11 bond, the excitation-induced
92 changes are maximal for X=S derivatives and minimal for their X=NMe counterparts.
93 In contrast, for the C9-N1 bond distance the above trend is reversed in both dye series
94 (See Figure S1 in the SI). Third, for the side phenyl rings, the expected p-quinoid form
95 (=(C(-CH=CH-)2C)=) is favored upon excitation. This can be deduced based on the
96 fact that two, out of six, ring bonds are shorter, while four of them are elongated when
97 passing from the GS to the ES geometry (see Scheme 2). Interestingly, somewhat similar
98 trends are observed for five-membered ring, i.e., N1-C2 and C2-X3 bonds become longer,
99 while X3-C4 and C9-N1 distances become shorter upon excitation.
Scheme 2: Schematic representation of geometry changes upon excitation in R=H and
R=NMe2 dyes.
2
0

ACCEPTED MANUSCRIPT
4
00 4. Conclusions
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
01
We have investigated the influence of heteroatom (X=O, S, and NMe) substitution
02 in five-membered heterocyclic ring on the photophysical properties of novel BF2 com-
03 plexes containing or not a dimethylamino group (R=NMe2). The experimental results
04 showed that the photophysical properties of the molecules bearing an electron-donating
05 group are substantially different from these observed for unsubstituted analogues, which
06 follows previous findings of our groups [2; 50; 82]. For instance, the NMe→O heteroatom
07 replacement has a negligible impact on the absorption and emission properties of R=H
08 derivatives, but induces significant red-shifts accompanied by changes of the relative in-
09 tensities of absorption/emission shoulders for R=NMe2 compounds. More pronounced
10 changes stem from the NMe→S substitution, i.e., both R=H and R=NMe2 optical bands
11 are substantially redshifted. It is noteworthy that all R=NMe2 derivatives exhibit fluo-
12 rescence quantum yields significantly higher (up to one order of magnitude) than their
13 unsubstituted counterparts. The experimental data were supported by the quantum
14 chemistry calculations, which not only reproduced the experimental trends but also con-
15 tributed to the understanding of electronic structure and excited-state properties: i)
16 the different photophysical properties of the R=NMe2 fluorophores likely result from the
17 more pronounced CT character of their S0→S1 transition, as evidenced by both the large
18 inhomogeneous broadening of the absorption spectra obtained with the PERI-CC2 sim-
19 ulations and the dipole moment differences obtained at various levels of theories; ii) the
20 difference between absorption band topologies of R=H and R=NMe2 have their roots in
21 the CT nature of the latter derivatives, that washes out the impact of vibronic couplings,
22 as PERI-CC2 calculations show that only one electronic state contributes to the overall
23 visible absorption band; and iii) the enhancement of fluorescence quantum yield observed
24 for compounds with R=NMe2 can be linked to stabilization of a more planar structure,
25 both in the ground and the excited electronic states.
2
1

ACCEPTED MANUSCRIPT
4
26 5. Experimental Section
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
27
In general, 2-phenacylbenzoxazoles can be obtained in the condensation of 2-amino-
28 phenol and 3,3-dialkoxy-3,3-dimercapto-1-phenylprop-2-en-1-ones [125; 126] or alkyl ben-
29 zoylacetates [127]. These compounds were also conveniently prepared by subtraction of
30 one of the methyl protons in 2-methylbenzoxazole and treatment of the formed carban-
31 ion with acylating agent such as benzoic acid ester or benzoyl chloride [81; 128–137].
32 All compounds discussed in the present paper were obtained in the two-step synthesis
33 starting from 2-methylbenzoxazole and substituted benzoyl chlorides as before [133]. It
34 is worth underscoring the tautomeric equilibrium exists in these compounds.
◦
◦
35
2-Phenacylbenzoxazole - White solid, yield 59%, Mp 94–96 C; 94–96 C [81], 88–
◦
◦
◦
◦
◦
36 88.5 C [126], 88–88.5 C [128], 88 C [129], 90–91 C [130], 97–98 C [125], and 93.5–
◦
37 94.5 C [131]. In CDCl3 solution 2-phenacylbenzoxazole (ketimine tautomeric form) is in
1
38 equilibrium with (Z)-2-(1,3-benzoxazol-2-yl)-1-phenylethenol (enolimine form). H NMR
39 (CDCl3 from TMS) δ: 12.50 (br s, 1H), 8.06 (m, 1H), 7.89 (m, 3H), 7.73 (m, 1H), 7.62
40 (m, 2H), 7.50 (m, 3H), 7.46 (m, 4H), 7.33 (m, 2H), 7.29 (m, 1H), 6.21 (s, 1H), 4.64
1
3
41 (s, 1H). C NMR δ: 192.4, 166.3, 165.74, 160.4, 151.3, 148.8, 141.3, 139.9, 135.8, 134.1,
42 133.9, 130.6, 128.9, 128.6, 128.6, 125.9, 125.0, 124.6, 124.3, 124.1, 120.0, 117.9, 110.6,
1
5
43 110.2, 83.7, 39.6. N NMR (CDCl3 from MeNO2) δ: -134.4, -164.1. C15H11NO2 Calcd.
44 C 75.94, H 4.67, N 5.90; Found C 75.87, H 4.75, N 5.83.
45
2-(1,3-benzoxazol-2-yl)-1-[4-(dimethylamino)phenyl]ethanone - White solid, yield: 73%,
◦
◦
46 Mp 199–200.4 C; 197–198 C [81]. In CDCl3 solution 2-(1,3-benzoxazol-2-yl)-1-[4-(dimethyl-
47 amino)phenyl]ethanone (ketimine tautomeric form) is in equilibrium with (Z)-2-(1,3-
2
2

ACCEPTED MANUSCRIPT
1
4
4
4
4
4
4
4
48 benzoxazol-2-yl)-1-[4-(dimethylamino)phenyl]ethenol (enolimine form). H NMR (CDCl3
49 from TMS) δ: 7.96 (m, 2H), 7.72 (m, 1H), 7.50 (m, 1H), 7.29 (m, 2H), 6.68 (m, 2H),
50 6.05 (s, 1H, signal of low intensity coming from low amount of the enolimine form), 4,53
1
3
51 (s, 1H), 3.05 (s, 6H). C NMR δ: 190.0, 166.7, 161.6, 153.9, 151.3, 141.5, 131.0, 127.3,
1
5
52 124.8, 124.4, 124.2, 123.7, 123.4, 119.9, 111.5, 110.8, 110.6, 80.4, 40.0, 39.1, N NMR
53 (CDCl3 from MeNO2) δ: -135.6, -322.9. C17H16N2O2, Calcd. C 72.84, H 5.75, N 9.99.
54 Found C 72.66, H 5.88, N 9.86.
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
55
Some of 1-methyl-2-phenacylbenzimidazoles were synthesized according to reaction
56 of benzoyl substituted mercaptals with N-methyl-o-phenylenediamine [138; 139]. Afore-
57 mentioned synthesis, as well as other earlier described in the literature [125; 140], are
58 characterized by low yield, therefore 1-methyl-2-phenacylbenzimidazole were prepared by
59 treating 1,2-dimethylbenzimidazole with benzoyl chloride in the presence of triethylamine
60 followed by thermal decomposition of formed 2-(1-methyl-1H-benzo[d]imidazol-2-yl)-1-
61 phenylvinyl benzoate. Thus, benzoyl chloride (8 mmol) was added at once to the mix-
62 ture of 1,2-dimethylbenzimidazole (0.29 g, 2 mmol), triethylamine (0.81 g, 8 mmol) and
63 diglyme (10 mL) and the resulted mixture was heated (steam bath) and stirred overnight.
64 Addition of water (6 mL) precipitated the byproduct. Solution of morpholine (0.52 g,
65 6 mmol) in methanol (3 mL) was added to crude 2-(1-methyl-1H-benzo[d]imidazol-2-yl)-
66 1-phenylvinyl benzoate and the resulted mixture was refluxed for 5 min. Reaction was
67 quenched by addition of water (3 mL) and the precipitated solid product was recrystal-
68 lized from ethanol.
◦
69
1-Methyl-2-phenacylbenzimidazole - Yellow solid, yield 40%, Mp 134–135.5 C; 119–
2
3

ACCEPTED MANUSCRIPT
◦
◦
◦
◦
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
70 121 C (plates) and 149–152 C (needles) [49], 150–151 C and 134–134.5 C [139], 150–
◦
◦
71 152 C [125; 140], and 150–151.5 C [128; 141]. In CDCl3 solution 1-methyl-2-phenacylbenz-
72 imidazole (ketimine tautomeric form) is in equilibrium with (Z)-2-(1-methyl-1H-benzi-
1
73 midazol-2-yl)-1-phenylethenol (enolimine form). H NMR (CDCl3 from TMS) δ: 14.04
3
74 (br s, 1H), 8.13 (d, 2H, JH,H=7.32 Hz), 7.93 (m, 2H), 7.74 (m, 1H), 7.61 (m, 1H), 7.50
75 (m, 2H), 7.43 (m, 4H), 7.33 (m, 1H), 7.28 (m, 1H), 7.23 (m, 3H), 5.91 (s, 1H), 4.70
1
3
76 (s, 2H), 3.76 (s, 3H), 3.67 (s, 3H). C NMR δ: 194.0, 174.8, 154.4, 148.6, 138.4, 134.9,
77 134.8, 134.0, 133.14, 130.11, 129.9, 128.93, 128.87, 128.83, 128.3, 127.3, 126.2, 124.3,
78 122.9, 122.8, 122.7, 122.3, 122.2, 119.3, 113.9, 110.6, 109.4, 108.3, 77.0, 39.1, 30.5, 29.1.
1
5
79
N NMR (CDCl3 from MeNO2) δ: -137.4, -210.4, -233.3, -253.5. C16H14N2O, Calcd.
80 C 76.78, H 5.64, N 11,19. Found C 76.85, H 5.74, N 11.03.
A slightly modified procedure [139] was used to prepare 1-[4-(dimethylamino)phenyl]-
81
82 2-(1-methyl-1H-benzimidazol-2-yl)ethanone. A mixture of 1,2-dimethylbenzimidazole
83 (0.29 g, 2 mmol), ethyl 4-(dimethylamino)benzoate (0.39 g, 2 mmol), dry benzene (10 mL),
84 dimethylformamide (0.5 mL) and sodium hydride (0.15 g, 60% suspension in mineral oil)
85 was refluxed and stirred magnetically overnight. The reaction was quenched by addition
86 of water (15 mL) and the resulted mixture was extracted with ethyl acetate (20 mL). The
87 organic layer was subsequently washed with water (2 x 10 mL) and brine (2 x 10 mL) and
88 dried with sodium sulfate. Solid impurities were filtered off, the filtrate was concentrated
89 on the rotary evaporator. The obtained solid was recrystallized from ethanol.
90
1-[4-(dimethylamino)phenyl]-2-(1-methyl-1H-benzimidazol-2-yl)ethanone - Yellow solid,
◦
◦
91 yield 36%, Mp 228–230 C, and 228–230 C [49]. In CDCl3 solution 1-[4-(dimethylamino)-
92 phenyl]-2-(1-methyl-1H-benzimidazol-2-yl)ethanone (ketimine tautomeric form) is in equi-
93 librium with (Z)-1-[4-(dimethylamino)phenyl]-2-(1-methyl-1H-benzimidazol-2-yl)ethenol
1
3
94 (enolimine form). H NMR (CDCl3 from TMS) δ: 8.02 (m, 2H), 7.68 (d, 1H, JH,H=8.80 Hz),
95 7.74 (m, 2H), 7.30 (m, 1H), 7.24 (m, 1H), 7.18 (m, 2H), 6.73 (m, 4H), 6.66 (m, 2H), 5.81
1
3
96 (s, 1H), 4.61 (s, 2H), 3.78 (s, 6H), 3.03 (s, 6H), 3.01 (s, 6H). C NMR δ: 191.7, 170.8,
97 153.8, 153.5, 149.8, 142.2, 136.1, 131.9, 131.2, 127.5, 123.7, 122.4, 122.3, 122.0, 119.2,
2
4

ACCEPTED MANUSCRIPT
1
5
4
4
5
98 116.8, 111.4, 110.73, 110.67, 110.4, 109.3, 74.9, 40.1, 40.0, 38.8, 30.5. N NMR (CDCl3
99 from MeNO2) δ: -136.8, -233.8, -317.3. C18H19N3O, Calcd. C 73.69, H 6.53, N 14.32.
00 Found C 73.78, H 6.63, N 14.23.
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
01
2-Methylbenzothiazole (2.5 g, 16.7 mmol) was added to a suspension of NaH (2 g,
02 50 mmol, 60% suspension in oil) in dry THF (15 mL) stirred at room temperature for 2 h.
03 After that time methyl benzoate (3.27 g, 21.8 mmol) or ethyl 4-(dimethylamino)benzoate
04 (4.21 g, 21.8 mmol) was added, and the resulting mixture was heated at reflux and stirred
05 magnetically overnight. After cooling down to room temperature, 2 M HCl aq. was slowly
06 added, and the yellow precipitate was filtered and washed with water. Resulting solid was
07 purified by column chromatography (SiO2, DMC) [142–144]. Other methods of synthesis
08 of 2-phenacylbenzothiazoles are reported in the literature [125; 128; 134; 136; 137; 145–
09 147].
◦
◦
10
2-phenacylbenzothiazole - yellow solid, yield 69%, Mp 110.5–112.5 C, 113.5–114.5 C
◦
◦
◦
◦
◦
11 [134], 110 C [129], 114 C [125], 113–114 C [126], 110–111 C [128], and 116–117 C [142].
12 In CDCl3 solution 2-phenacylbenzothiazole (ketimine tautomeric form) is in equilibrium
1
13 with (Z)-2-(1,3-benzothiazol-2-yl)-1-phenylethenol (enolimine form). H NMR (CDCl3
3
14 from TMS) δ: 13.81 (br s, 1H), 8.09 (d, 2H, JH,H=7.36 Hz), 8.03 (m, 1H), 7.88 (m, 2H),
15 7.80 (m, 2H), 7.59 (m, 4H), 7.52 (m, 4H), 7.39 (m, 1H), 7.30 (m, 1H), 6.38 (s, 1H), 4.85
1
3
16 (s, 1H). C NMR δ: 194.0, 168.1, 165.6, 163.8, 152.3, 150.4, 135.8, 135.2, 134.8, 133.9,
17 131.4, 130.4, 130.0, 129.1, 128.9, 128.7, 128.5, 127.5, 126.5, 126.2, 125.9, 125.2, 124.2,
1
5
18 122.8, 122.4, 121.6, 121.4, 120.0, 90.9, 43.7. N NMR (CDCl3 from MeNO2) δ: -120.0.
19 C15H11NOS, Calcd. C 71.12, H 4.38, N 5.53. Found C 71.20, H 4.51, N 5.42.
2
5

ACCEPTED MANUSCRIPT
5
5
5
5
5
5
20
2-(1,3-benzothiazol-2-yl)-1-[4-(dimethylamino)phenyl]ethanone - yellow solid, yield
◦
1
21 77%, Mp 207–208 C. H NMR (CDCl3 from TMS) δ: 7.99 (m, 3H), 7.86 (d, 1H,
3
3
22
JH,H=8.0 Hz), 7.45 (m, 1H), 7.36 (m, 1H), 6.67 (d, 2H, JH,H=8.0 Hz ), 4.75 (s, 2H),
1
3
23 3.07 (s, 6H). C NMR δ: 191.8, 165.1, 153.8, 152.6, 136.0, 131.1, 127.5, 126.3, 125.8,
1
5
24 124.9, 123.6, 123.6, 122.7, 121.5, 121.3, 110.8, 88.4, 43.3, 40.0. N NMR (CDCl3 from
25 MeNO2) δ: -67.8, -317.1. C17H16N2OS, Calcd. C 68.89, H 5.44, N 9.45. Found C 68.95,
H 5.57, N 9.29. The complexes were prepared based on a literature protocol [82]. For
5
5
26
27 their purification DCM was used as eluent in column chromatography (SiO2).
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
28
1,1-difluoro-5-methyl-3-phenyl-1,5-dihydrobenzo[4,5]imidazo[1,2-c][1,3,2]oxazaborinin-
◦
1
29 10-ium-1-uide (X=NMe, R=H) - White solid, yield 36%, Mp 255.6–256.7 C. H NMR
30 (DMSO-d6 from TMS) δ: 8.09 (m, 2H), 7.77 (m, 1H), 7.68 (m, 1H) 7.56 (m, 3H), 7.44
1
1
31 (m, 2H), 7.05 (s, 1H), 3.98 (s, 3H). B NMR (DMSO-d6 from BF3·Et2O) δ: 1.66 (t).
1
3
32
C NMR δ: 164.8, 150.4, 134.4, 133.8, 132.0, 131.7, 129.2, 126.9, 125.1, 124.8, 114.3,
1
5
19
33 111.8, 81.8, 30.5.
N NMR (DMSO-d6 from MeNO2) δ: -218.76, -242.04.
F NMR
34 (DMSO-d6 from CFCl3) δ: -134.14. C16H13BF2N2O, Calcd. C 64.47, H 4.40, N 9.40;
35 Found C 64.53, H 4.46, N 9.48.
36
3-(4-(dimethylamino)phenyl)-1,1-difluoro-5-methyl-1,5-dihydrobenzo[4,5]imidazo[1,2-
37 c][1,3,2]oxazaborinin-10-ium-1-uide (X=NMe, R=NMe2) - Yellow solid, yield 39%,
◦
1
3
38 Mp 282–283.4 C. H NMR (DMSO-d6 from TMS) δ: 7.97 (d, 2H, JH,H=9.04 Hz), 7.74
3
39 (m, 1H), 7.65 (m, 1H) 7.46 (m, 2H), 6.85 (d, 2H, JH,H=9.12 Hz), 6.79 (s, 1H), 3.96
1
1
13
40 (s, 3H), 3.09 (s, 6H). B NMR (DMSO-d6 from BF3·Et2O) δ: 1.59 (t).
C NMR δ:
41 166.1, 152.7 151.2, 133.8, 131.9, 128.5, 124.6, 124.2, 120.8, 113.8, 111.7, 111.3, 77.9,
1
5
42 ca. 40.64-39.36 (overlapped with solvent), 30.23. N NMR (DMSO-d6 from MeNO2) δ:
2
6

ACCEPTED MANUSCRIPT
1
9
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
43 -211.89, -245.63, -319.77. F NMR (DMSO-d6 from CFCl3) δ: -134.64. C18H18BF2N3O,
44 Calcd. C 63.37, H 5.32, N 12.32; Found C 63.30, H 5.44, N 12.30.
45
1,1-difluoro-3-phenyl-1H-benzo[4,5]oxazolo[3,2-c][1,3,2]oxazaborinin-10-ium-1-uide
◦
1
46 (X=O, R=H) - Yellowish-green solid, yield 38%, Mp 239.2–240.8 C. H NMR (DMSO-
47 d6 from TMS) δ: 8.12 (m, 2H), 7.94 (m, 1H), 7.72 (m, 1H) 7.65 (m, 1H), 7.59 (m, 4H),
3
11
48 7.55 (d, 1H, JH,H=1.64 Hz), 7.22 (s, 1H). B NMR (DMSO-d6 from BF3·Et2O) δ: 1.85,
1
3
49 (t). C NMR δ: 170.8, 165.1, 148.5, 133.1, 130.1, 129.5, 127.5, 126.9, 126.4, 114.9, 112.5,
1
5
19
50 82.1. N NMR (DMSO-d6 from MeNO2) δ: -213.54. F NMR (DMSO-d6 from CFCl3)
51 δ: -135.07. C15H10BF2NO2, Calcd. C 63.20, H 3.54, N 4.91; Found C 63.25, H 3.49,
52 N 4.83.
53
3-(4-(dimethylamino)phenyl)-1,1-difluoro-1H-benzo[4,5]oxazolo[3,2-c][1,3,2]oxazaborinin-
◦
1
54 10-ium-1-uide (X=O, R=NMe2) - Orange solid, yield 41%, Mp 298–300 C. H NMR
3
3
55 (DMSO-d6 from TMS) δ: 7.96 (d, 2H, JH,H=9.12 Hz), 7.85 (d, 1H, JH,H=7.36 Hz), 7.61
3
3
56 (d, 1H, JH,H=7.64 Hz), 7.50 (m, 2H), 6.89 (s, 1H), 6.81 (d, 2H, JH,H=9.16 Hz), 3.07
1
1
13
57 (s, 6H). B NMR (DMSO-d6 from BF3·Et2O) δ: 1.79, (t). C NMR δ: 172.1 153.7,
58 152.4, 148.2, 130.6, 129.6, 127.0, 125.9, 119.1, 114.2, 112.1, 111.8, 77.7, ca. 40.61-39.35
1
5
19
59 (overlapped with solvent).
N NMR (DMSO-d6 from MeNO2) δ: -313.97.
F NMR
60 (DMSO-d6 from CFCl3) δ: -133.19. C17H15BF2N2O2, Calcd. C 62.23, H 4.61, N 8.54;
61 Found C 62.14, H 4.64, N 8.39.
62
1,1-difluoro-3-phenyl-1H-benzo[4,5]thiazolo[3,2-c][1,3,2]oxazaborinin-10-ium-1-uide
◦
1
63 (X=S, R=H) - Yellow solid, yield 43% Mp 244.5–245.7 C. H NMR (DMSO-d6 from
3
3
64 TMS) δ: 8.23 (d, 1H, JH,H=8.0 Hz), 8.13 (m, 2H), 8.00 (d, 1H, JH,H=8.16 Hz) 7.69
1
1
65 (m, 1H), 7.59 (m, 4H), 7.51 (s, 1H). B NMR (DMSO-d6 from BF3·Et2O) δ: 1.64 (t).
1
3
66
C NMR δ: 169.3, 164.9, 142.1, 133.1, 132.8, 129.5, 129.4, 128.9, 127.3, 126.6, 124.1,
1
5
19
67 117.8, 90.8. N NMR (DMSO-d6 from MeNO2) δ: -186.16. F NMR (DMSO-d6 from
68 CFCl3) δ: -132.16. C15H10BF2NOS, Calcd. C 59.83, H 3.35, N 4.65. Found C 59.75,
69 H 3.43, N 4.54.
70
3-(4-(dimethylamino)phenyl)-1,1-difluoro-1H-benzo[4,5]thiazolo[3,2-c][1,3,2]oxazabor-
2
7

ACCEPTED MANUSCRIPT
◦
1
5
5
5
5
5
5
5
71 inin-10-ium-1-uide (X=S, R=NMe2) - Yellow solid, yield 44%, Mp 280–283 C. H NMR
3
72 (DMSO-d6 from TMS) δ: 8.17 (d, 1H, JH,H=8.0 Hz), 7.89 (m, 2H), 7.85 (s, 1H) 7.61
1
1
73 (m, 1H), 7.47 (m, 1H), 7.23 (s, 1H), 6.82 (d, 2H), 3.06 (s, 1H). B NMR (DMSO-d6 from
1
3
74 BF3·Et2O) δ: 1.54 (t). C NMR δ: 168.5, 166.5, 153.5, 142.3, 129.4, 128.5, 125.7, 123.8,
1
5
75 118.8, 117.1, 111.9, 87.6, ca. 40.60-39.35 (overlapped with solvent). N NMR (DMSO-
1
9
76 d6 from MeNO2) δ: -194.22, -314.90.
F NMR (DMSO-d6 from CFCl3) δ: -133.09.
77 C17H15BF2N2OS, Calcd. C 59.32, H 4.39, N 8.14. Found C 59.43, H 4.45, N 8.06.
78 Acknowledgments
5
5
5
5
5
5
5
5
5
5
5
79
B.O. and A.M.G. thank the Polish National Science Centre (Grant No.
80 2017/26/M/ST5/00327). W.B. acknowledges the statutory activity subsidy from the
81 Polish Ministry of Science and Higher Education for the Faculty of Chemistry of Wrocław
82 University of Technology. The authors want to address special and warm gratitude
83 to Dr Robert Zaleśny from Faculty of Chemistry, Wrocław University of Science and
84 Technology for countless valuable discussions which helped to improve manuscript and
85 his kind and precious help with CC2 calculations. Authors are also very graceful to
86 Damian Plazuk from University of Łódź for recording fluorescence quantum yield with
87 the use of the integrating sphere. The calculations were performed at the Wroclaw Center
88 for Networking and Supercomputing.
5
89 Supporting Informations
5
5
5
90
The Supporting Information is available on the Elsevier website. It contains fluo-
91 rescence decay curves, NMR spectra, Cartesian coordinates, computational results and
92 bond-lenghts analysis.
5
93 References
5
5
94
95
[1] I. Kaya, M. Yildirim, A. Avci, Synthesis and characterization of fluorescent polyphenol species de-
rived from methyl substituted aminopyridine based schiff bases: The effect of substituent position
2
8