New Boron(III) Blue Emitters for All-Solution Processed OLEDs: Molecular Design Assisted by Theoretical Modeling

Article abstract of DOI:10.1002/ejic.201900265

Luminescent boron(III) complexes have recently been employed as emitters in organic light-emitting diodes (OLEDs) with reasonable success. They are easy to prepare and sufficiently stable to be used in such devices, being of great interest as a simple mol

Full text of DOI:10.1002/ejic.201900265

A Journal of
Accepted Article
Title: New Boron(III) blue emitters for all-solution processed OLEDs:
molecular design assisted by theoretical modeling
Authors: Cristian A.M. Salla, Jéssica Teixeira dos Santos, Giliandro
Farias, Adailton J. Bortoluzi, Sergio F. Curcio, Thiago Cazati,
Róbert Izsák, Frank Neese, Bernardo de Souza, and Ivan
Bechtold
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Eur. J. Inorg. Chem. 10.1002/ejic.201900265
Link to VoR: http://dx.doi.org/10.1002/ejic.201900265

10.1002/ejic.201900265
European Journal of Inorganic Chemistry
New boron(III) blue emitters for all-solution processed OLEDs: molecular
design assisted by theoretical modeling
Cristian A. M. Salla[a], Jéssica Teixeira dos Santos[b], Giliandro Farias[b], Adailton J.
Bortoluzi[b], Sergio F. Curcio[c], Thiago Cazati[c], Róbert Izsák[d], Frank Neese[d], Bernardo de
Souza[b]*, Ivan H. Bechtold[a]*
[a]Physics Department, Universidade Federal de Santa Catarina, 88040-900 Florianópolis, SC,
Brazil
[b]Chemistry Department, Universidade Federal de Santa Catarina, 88040-900 Florianópolis,
SC, Brazil
[c]Physics Department, Universidade Federal de Ouro Preto, 35400-000 Ouro Preto, MG,
Brazil
[d]Max-Planck-Institut für Kohlenforschung, Kaiser-Wilhelm-Platz 1, 45470 Mülheim an der
Ruhr, Germany
*Email: ivan.bechtold@ufsc.br; tel: +55 48 37212304
bernardo.souza@ufsc.br; tel: +55 48 37213608
Graphical Abstract
We investigate the emission and photophysics from known and new boron(III) complexes, by
both experiments and theory. Within our theoretical framework, we show to be possible to
determine the actual color of emission and predict the fluorescence rates. Solution processed
OLEDs were also built from those, with good efficiency and potential for application, also
displaying the predicted colors.
This article is protected by copyright. All rights reserved.

10.1002/ejic.201900265
European Journal of Inorganic Chemistry
Abstract
Luminescent boron (III) complexes have recently been employed as emitters in organic light
emitting diodes (OLEDs) with reasonable success. They are easy to prepare and sufficiently
stable to be used in such devices, being of great interest as a simple molecular emissive layer.
Although emitters for this class with all colors have already been reported, highly efficient
and stable blue emitters for applications in solution processed devices still pose a challenge.
Here, we report the design, synthesis and characterization of new boron complexes based on
the 2-(benzothiazol-2-yl)phenol ligand (HBT), with different donor and acceptor groups
responsible for modulating the emission properties, from blue to red. The molecular design
was assisted by calculations using our newly developed formalism, where we demonstrate
that the absorption and fluorescence spectra can be successfully predicted, being a powerful
technique to evaluate molecular photophysical properties prior to synthesis. In addition,
density functional theory (DFT) allows to understand the molecular and electronic structure of
the molecules in greater detail. The molecules studied here presented fluorescence efficiency
as high as Φ = 0.88 and all-solution processed OLEDs were prepared and characterized under
ambient atmosphere, after dispersion in the emitting layer. Surprisingly, even considering
these rather simple experimental conditions, the blue emitters displayed superior properties
compared to present literature, in particular with respect to the current efficiency stability.
Keywords: boron(III) complexes; theoretical modeling; blue emitters; solution processed
OLEDs; DFT.
1. Introduction
There has been significant progress in recent years in the design of new materials for
application in OLEDs, particularly for those emitting in the blue region. Despite the
improvements with phosphorescence and thermally activated delayed fluorescence (TADF),
fluorescent blue materials have proven to be the best option so far for the construction of
efficient, stable and long-lived devices in displays and lighting systems.[1] Most current
research involving blue emitters deal with classical scaffolds with known good fluorescence
yields such as anthracene, pyrene, and fluorene. However, the possibility of excimer
formation through packing usually leads to reduced electroluminescent efficiency and
degradation of color purity in such devices.[2] In that context, the color tuning of four-
coordinate organoboron compounds achieved by modifying ligands provides an alternative
for the development of new blue emitters.[3][4] Even so, there are only a few reports on both
This article is protected by copyright. All rights reserved.

10.1002/ejic.201900265
European Journal of Inorganic Chemistry
efficient and stable blue fluorescent complexes and very little on the theoretical modeling of
the emission dynamics.
Boron (III) complexes have been extensively studied in recent years by different
research groups. These compounds are widely employed in organic synthesis,[5] chemical
biology[6] and material sciences,[7] mainly due to versatility, ease of synthesis,
physicochemical stability and low cost.[8] In particular, luminescent organoboron complexes
have received considerable attention in the development of efficient OLEDs[9] because of
their high emission quantum yield, good thermal stability and their charge carrier
properties.[10] Devices based on these materials with emissions from blue to red,[11] white[12]
and near infrared[13] have already been reported. In addition, these materials can be processed
in solution to make thin films or dispersed on host materials, which simplifies the
manufacturing process and reduces the production costs.[10,14]
The type of ligand and the nature of substituent groups on either the ligands or the
boron center itself influences the photophysical properties of these compounds, which are
affected by the boron–chromophore interaction. In particular, ligands with flat and rigid π-
conjugated skeletons are useful for obtaining luminescent complexes for OLED applications,
since they minimize the decay through non-radiative pathways.[15,16] Boron dypyrromethene
complexes (BODIPYs) are among the most well-known boron complexes with an intense
fluorescence and tunable emission wavelengths of this class. Nevertheless, in general,
BODIPYs and their derivatives exhibit self-quenching due to the small Stokes shifts and tight
packing, reducing their emission efficiency in solid state, which has limited their potential
application in OLEDs.[17] Trying to overcome these limitations, boron (III) complexes with
the widely used ligand 2-(benzothiazol-2-yl)phenol (HBT)[18] were recently synthesized. They
displayed high Stokes shifts and high fluorescence intensities at the same time, and their
emission color was shown to be tunable through chemical modifications on the aromatic
moiety.[19]
Herein, we report the design, synthesis and characterization of a series of emissive
neutral boron (III) complexes (Figure 1) based on the HBT ligand containing different donor
and acceptor groups, where the luminescent properties could be modulated, leading from blue
to red emitters. Calculations using DFT assisted molecular and electronic structure
interpretation, as well as some excited state dynamics. We show that our recently developed
formalism[19] in conjunction with time-dependent density functional theory (TD-DFT), allows
the prediction of emissive rates and the fluorescence spectra with impressive accuracy.
Finally, the complexes were dispersed on a double-host emissive layer and all-solution
This article is protected by copyright. All rights reserved.

10.1002/ejic.201900265
European Journal of Inorganic Chemistry
processed OLEDs were prepared. The device constructed from the blue emitter with the
highest fluorescence quantum yield (Φ = 0.88), prepared under normal ambient conditions,
revealed itself to be one of the most efficient all-solution processed devices based on boron
complexes known to date.
Figure 1. Proposed structures of synthesized Boron(III) complexes.
2. Results and Discussion
2.1 Synthesis of the ligands and complexes
The HBT derivatives can be synthesized in a one-pot procedure, starting from commercially
available 2-aminobenzenethiol and substituted phenols in ethanol, using hydrogen peroxide as
oxidant. Then, the difluoroboron complexes B1-B4 can be prepared from the respective
ligand, by treatment with boron trifluoride diethyl ether in the presence of DIPEA. All of the
complexes were obtained from direct crystallization in dichloromethane and for the nitro
substituted compound (B4), single crystals were obtained. A synthetic route for the ligands
and the boron complexes is shown in Scheme 1 and Schemes 1S – 3S.
2.2 Crystal structure and geometries from DFT
The molecular structure of complex B4 was determined by single crystal X-ray diffraction
studies. As can be seen in Figure 2, the boron(III) center adopts a typical tetrahedral
geometry, binding with N1 and O1 of HBT and two fluorine atoms. Bond distances and
angles are reported at the Supporting Information and are similar to those of other related
organoboron compounds previously reported.[10, 20] The molecule crystallizes in a P21/c space
group with four molecules per lattice and between these units, a π-stacking interaction is
observed with a shorter distance between centroids of 3.615 Å and coplanarity angle of 1.78º.
This article is protected by copyright. All rights reserved.

10.1002/ejic.201900265
European Journal of Inorganic Chemistry
Figure 2. ORTEP representation of the molecular structure of complex B4 (left) and packing structure (right).
Unfortunately, despite their high purity, we could not obtain single crystals for the
other complexes. In the absence of accurate experimental data for B1, B2 and B3, we decided
to rely on quantum chemistry and optimized the geometries using DFT at the PBE0/def2-
TVZP(-f) level. Figure 3 shows that the calculated structures around the boron center are quite
similar for all complexes. The vibrational frequencies confirm that the geometries correspond
to a minimum on the potential energy surface, since no negative frequencies were found. For
the complex B2 with a methoxy group, two isomers were considered by rotating the C-O
bond, with an energy increase of 0.693 kcal mol-1 from B2a to B2b. Table 3S of Supporting
Information compares the crystalline structure parameters of B4 with the optimized
theoretical calculations for B1-B4. As can be seen, the theoretical geometry for B4 is in good
agreement with the experiment, where the average error for bond lengths was 0.83% and
1.23% for bond angles, which supports further use of PBE0 in subsequent investigations.
Figure 3. Optimized geometry for complexes B1-B4 obtained from DFT using PBE0/def2-TZVP(-f).
This article is protected by copyright. All rights reserved.

10.1002/ejic.201900265
European Journal of Inorganic Chemistry
2.3 Photophysical Properties
The UV-Vis absorption and emission spectra of complexes B1-B4 obtained in CH2Cl2 are
presented in Figure 4 and summarized in Table 1. All complexes exhibited an intense
absorption band around 300 nm, and lower energy absorption bands with maxima between
360 and 460 nm, which are commonly attributed to π→π* transitions of the HBT ligand.
These were strongly affected by the substituent groups: the bands of B2 and B3, containing
electron donor groups, were red shifted when compared with B1 (without substituent), while
those for B4, which has an electron withdrawing group, were blue shifted.
The emission peaks of the complexes B1, B2 and B4 after excitation at the lower
energy peak were located in the blue and green region, with maxima centered at 424, 481 and
502 nm. B3, with the dimethylamine group, exhibited the lowest energy emission band (595
nm), already suggesting a smaller HOMO-LUMO gap. This indicates possibility of
modulating the emission maxima with a simple modification of the molecular structure. The
results in the solid state, with complexes dispersed in a PMMA matrix, are similar to those
obtained in solution and can be seen in Figure 7S.
Figure 4. Optical absorption (dotted line) and emission (solid line) spectra of the boron (III) complexes in
CH2Cl2 solution 2.010-5 mol L-1. Emission spectra were obtained by excitation at the low energy region of the
first absorption.
Excited state decay curves of the boron complexes were collected at the wavelength of
maximum emission in solution, indicating lifetimes typical of fluorescence, in the range of
This article is protected by copyright. All rights reserved.

10.1002/ejic.201900265
European Journal of Inorganic Chemistry
nanoseconds (Figure 9S). The fluorescent decay curves for B1 and B3 in solution could be
fitted with a mono-exponential function, resulting in a lifetime of 1.4 ns and 15.8 ns,
respectively, as expected for monomeric emission. The fluorescent decay curves of B2 in
solution was best fitted with a bi-exponential function displaying two distinct lifetimes in
solution, 9.7 ns and 5.7 ns, with 51.8% and 48.2% contribution to the total emission. These
agree with the fact that there are two conformers for B2. In order to confirm the individual
emission of both isomers, the lifetime was collected at 445 nm and 505 nm, that is, 30 nm
bellow and above the maximum emission wavelength, where a variation of each contribution
is expected due to distinct emission bands. In both cases, the same two lifetimes (9.7 ns and
5.7 ns) were observed, with contributions of 27.8% and 72.2% for 455 nm, and 68.6% and
31.4% for 505 nm, thus confirming the existence of the isomers absorbing at different
wavelengths. The B4 excited state decay curve in solution could only be fitted satisfactorily
with a triexponential function, resulting in lifetimes of 8.4 ns (7.1%), 1.6 ns (16.5%) and 0.2
ns (76.4%), possibly due to emission from aggregates even at very low concentrations.
When the absolute emission quantum yield was measured, a quite high
photoluminescence quantum yield (PLQY, 훷F) was obtained for complex B2 (훷F = 0.88 in
CH2Cl2), followed by complex B3 (훷F = 0.45 in CH2Cl2), both of which contain donor
groups. The lowest PLQY was observed for complex B4 (훷F = 0.02 in CH2Cl2). This
demonstrates the significant role of the substituents on the luminescent properties of HBT-FB
complexes when compared to complex B1 (훷F = 0.26 in CH2Cl2), without any substituent.
The solid-state luminescence properties followed the same pattern, with only a reduction on
the B2 and B3 PLQY (see Table 1), which can be due to the existence of strong
interchromophore π-π interactions in the solid state, leading to larger nonradiative rates.[10]
The emission of B4 was too low even to allow proper quantification.
Table 1. Photophysical data for the boron (III) complexes.
Solution[a]
Solid State
λabs (nm)/
ε (M-1 cm-1)
362 /26484
[b]
[b]
λem (nm)
ΦF
τS (ns)[c]
λabs (nm)
λem (nm)
ΦF
421
475
595
496
0.26
0.88
0.45
0.02
1.4
8.3[d]
5.4
363
397
454
355
424
481
576
502
0.26
0.53
0.37
-
B1
B2
B3
B4
396/15984
449/485
356/22416
15.8[d]
[a] In CH2Cl2 at 2.010-5 mol L-1. [b] The uncertainty in the ΦF measure was ±0.01. [c] ± 0.1 ns. [d] τ = Σ τi2Ai /ΣτiAi,
due to multi-exponential profile.
This article is protected by copyright. All rights reserved.

10.1002/ejic.201900265
European Journal of Inorganic Chemistry
2.4 Theoretical Modeling of the Electronic Structure and Fluorescence
To have a better understanding of the electronic and luminescent properties of B1-B4, we first
carried out calculations using DFT on all complexes. The calculated HOMO (-1), HOMO,
LUMO and LUMO (+1) are shown in Figure 5, with their energies on the vertical axis. As
can be seen, the HOMO increases in energy from B1 to B3, as the donating power of the
substituent increases, and decreases in B4, while the LUMO energy does not change
significantly. For the complexes B1-B3, the HOMO has major contributions from the phenol
part of the chelate ligands, and has a small contribution from the benzothiazole moiety,
particularly for B1 and B2, whereas the LUMO for these complexes is mostly a π* orbital
spread throughout the ligand. In addition, the HOMO (-1) and LUMO (+1) are quite similar
for these species, being a delocalized π orbital and a π* centered on the heterocycle.
In contrast, the electronic structure of B4 is different. For this complex, the HOMO is
more evenly spread throughout the ligand and the LUMO is centered on both the π* orbitals
of the C-N bond of the thiadiazol ring and the nitro group. The HOMO (-1) for this complex
is centered on the benzothiazole and the LUMO (+1) on the phenol moiety, demonstrating
that the substituents can have a significant impact on energy levels and charge distribution,
which ultimately influences the radiative rates and color, as shown later on.
Figure 5. Frontier molecular orbitals of complexes B1-B4 calculated using PBE0/def2-TZVP(-f).
This article is protected by copyright. All rights reserved.

10.1002/ejic.201900265
European Journal of Inorganic Chemistry
The electronic excitations were modeled by TD-DFT, using linear response CPCM to
account for the solvent effect on the energy, and the calculated absorption spectra as well as
convoluted gaussians reproducing the experimental spectra are shown for each molecule in
Figure 6, along with the experimental results. After specifying the number of Gaussians to be
used in the fitting by selecting prominent peaks and shoulders on the experimental spectra, the
width and intensity of the curves were allowed to vary to achieve the best fit. Allowing the
width to vary may account for broadening due to vibrational progressions, which is why some
features of the spectra were not selected for fitting. While owing to the nature of this process
and the possible failings of the TD-DFT method some electronic transitions may be missed or
misidentified, a tentative assignment is still possible between the Gaussians and the TD-DFT
transition state labels. Table 4S displays some of the most intense calculated transitions, with
the corresponding largest excitation amplitudes. There are three intense transitions predicted
for compounds B1-B3 and the lowest energy transition of all these complexes can be assigned
as HOMO→LUMO. Furthermore, the higher-energy transitions in Table 4S and Figure 6 are
similar for both B2 isomers, but the low energy ones have slightly different energies,
indicating that the experimental spectrum might have contributions from both species. This
also suggests distinct emission bands, as was observed by time-resolved fluorescence
measurements. On the other hand, the complex B4, as expected from the molecular orbital
analysis, has a different profile for the electronic transitions. Its state of lowest energy, which
without the CPCM correction is the fifth singlet state, results from a mixed HOMO→LUMO
(70%) and HOMO→LUMO (+1) (21%) configuration with a transition dipole moment much
smaller than that found for the first excitation of B1-B3. The other high energy transitions are
similar to those of the previous complexes. Comparisons of the theoretical and experimental
oscillator strengths and energy values are shown in Table 2, and the band assignment from
these are shown in Figure 6.
This article is protected by copyright. All rights reserved.

10.1002/ejic.201900265
European Journal of Inorganic Chemistry
Figure 6. Experimental absorption spectra of B1-B4 in CH2Cl2 (left, with tentative assignments) and the
theoretical absorption spectra (right) calculated using PBE0/def2-TZVP(-f) and convoluted with Gaussians of
0.2 eV width.
Since fluorescence generally occurs from the lowest excited state to the ground state,
and its rate depends on the transition matrix element of the dipole moment operator, it is
important to discuss its variation as a function of the molecular structure. The transition
dipole moment can be rationalized as a function of the charge redistribution during the
This article is protected by copyright. All rights reserved.

10.1002/ejic.201900265
European Journal of Inorganic Chemistry
electronic transition which is directly affected by the substituents. From the frontier natural
transition orbitals (NTOs) in Figure 8S, for B1-B3, the excitation occurs mainly from the
phenol ring to the π* orbitals of the C-N bond, and can be classified as a charge transfer-like
transition. The presence of donating groups contributes to delocalizing the HOMO around the
phenol moiety and ends up decreasing the charge transfer character of the excitation (since
there are phenol components also in the LUMO), thus lowering the transition dipole moment
associated with it. For B4, the much smaller transition dipole moment is clearly due to a
larger delocalization of both the HOMO and LUMO, particularly centering the LUMO (+1)
around the phenol ring (see B1-B3 in Figure 5), thus giving rise to a more π-π*-like transition.
Table 2. Comparison between experimental and calculated vertical energy differences and oscillator strengths
using PBE0/def2-TZVP(-f).
Experiment
Energy / eV
Theory
Energy (Sn) / eV
f [a]
f
B1
3.42
3.94
4.07
4.18
0.3422
0.0199
0.2191
0.2829
3.68 (S1)
4.18 (S2)
4.23 (S3)
0.3862
0.1675
0.2615
B2 (a / b)
3.04
3.25
0.0908
0.1660
3.09 / 3.23 (S1)
0.2092 / 0.2885
3.80
3.97
4.18
0.0297
0.1047
0.2720
4.12 / 4.14 (S2)
4.15 / 4.20 (S3)
0.3514 / 0.2602
0.2692 / 0.2711
B3
2.76
3.64
0.0844
0.1024
2.38 (S1)
3.79 (S2)
3.84 (S3)
0.1124
0.0058
0.2311
4.12 (S5)
4.24 (S6)
4.45 (S7)
4.65 (S8)
0.3825
0.1082
0.0051
0.0968
4.52
1.4597
B4
3.23
3.49
4.14
0.0311
0.1789
0.1194
3.46 (S1)
3.65 (S2)
3.86 (S3)
0.0101
0.3726
0.2262
This article is protected by copyright. All rights reserved.

10.1002/ejic.201900265
European Journal of Inorganic Chemistry
3.86 (S4)
4.49 (S7)
0.0172
0.5391
4.20
0.6277
[a] f = 4.32×10-9ʃɛdσ.
The radiative rates (kr) and fluorescence spectra were also calculated for all complexes
using our newly developed approach. [19] The PBE0 and B2PLYP functionals were chosen to
calculate adiabatic electronic energy differences on top of the PBE0 geometries and Hessians.
The calculated rates are displayed in Table 3 and showing that experimental and theoretical
values are in very good agreement. Even for B2, the kr is correctly predicted to be the average
of both isomers. As the radiative rate diminishes along with the transition dipole moment for
B2 and B3, the increase of the emission yield seems to occur due to a smaller non-radiative
rate. Nevertheless, we could predict that the quantum yield would be at least of the same order
from the kr, regardless of the change in emission color.
Table 3. Experimental and calculated rate constants in CH2Cl2 using the def2-TZVP(-f) basis set.
kr(exp)  108 (s-1)[a]
kr(theo)  108 (s-1)
PBE0[b]
kr(theo)  108 (s-1)
B2PLYP[b]
2.51
knr(exp)  108 (s-1)[c]
1.85
1.06
2.23
5.28
0.14
B1
B2
0.99 (B2a)
1.38 (B2b)
0.01
0.85 (B2a)
1.08 (B2b)
0.98
0.83
0.01
1.01
0.62
B3
B4
0.04
0.04
[c]
[a] kr = Φ / τ. [b]using def2-TZVP(-f). knr = (1 - Φ) / τ.
The emission spectra, presented in Figure 7, are also coherent with the experimental
results. The main vibrational progression is due to the C=C and C=N stretching modes, from
the energy separation of these peaks of about 0.15 eV (1200 cm-1). We also calculated the
theoretical observed color and, as one can verify from the areas below the lines, we can
predict them with good agreement to the experiment. The larger error was found in B4, where
we could not reproduce the green color of emission due to the 0-0 energy difference error, a
known issue in DFT.[21] That particular color deviation comes from the fact that the energy
gap was on the border of green and blue and the small shift induced a larger error on the color
perceived. In general, the B2PLYP functional that has a fraction of MP2 correlation
performed better, predicting better energy differences as expected from previous results. [19]
This article is protected by copyright. All rights reserved.

10.1002/ejic.201900265
European Journal of Inorganic Chemistry
Figure 7. Experimental and predicted emission spectra using PBE0 and the B2PLYP functionals using the def2-
TZVP(-f) basis. The colors of the areas represent the actual emission colors obtained from the spectra and the
color matching functions.
2.5 Thermal and Electrochemical Properties
A thermogravimetric analysis (TGA) of B1–B4 under nitrogen atmosphere revealed that the
initial decomposition temperatures are 171, 254, 298 and 104 °C (see Figure 10S). Such
thermal stability indicates that all the complexes can undergo thermal treatments until at least
100 °C and that both B2 and B3 could even be sublimated in evaporated OLEDs. Cyclic
voltammetry in CH2Cl2 solutions were performed to estimate the HOMO and LUMO energy
levels from the oxidation and reduction potentials. Figure 11S shows the cyclic
voltammograms and Table 4 summarizes the electrochemical data. Complexes B1 and B4
exhibited irreversible oxidation, B2 showed a quasi-reversible oxidation and B3 reversible
oxidation peaks. It was found that the substituent effect agrees with the shifts in the UV-Vis
spectra. The complex B4, with the electron withdrawing group, presented the lowest HOMO
and LUMO energy, while B3, with the strongest electron donating group, had the highest
HOMO and LUMO. The HOMO and LUMO energies obtained from voltammetry agree with
the tendency from DFT, also confirming the substituent effect on these orbital energies.
This article is protected by copyright. All rights reserved.

10.1002/ejic.201900265
European Journal of Inorganic Chemistry
Table 4. Electrochemical properties of complexes B1-B4.
Eox (V)
Ered (V)
EHOMO (eV)[a]
-6.22
ELUMO (eV)[a]
-3.11
Bandgap (eV)[b]
-
-1.69
3.11
2.82
2.27
3.20
B1
B2
B3
B4
1.16
0.54
-
-
-
-5.96
-3.14
-5.34
-3.07
-1.16
-6.82
-3.64
[a] Determined from peak potentials, where EHOMO = - (4.8 + Eox) and ELUMO = - (4.8 – Ered).[22]
[b] Calculated from the onset of the absorption spectra in solution.
2.6 Electroluminescent Properties
The electroluminescence (EL) properties of the complexes were then evaluated in all-solution
processed devices with a simple structure: ITO/Pedot:PSS(25 nm)/TcTA:OXD-7:BX(25
nm)/PFNBr:TEAB(35 nm)/CsF(1.3 nm)/Al(100 nm), with BX=B1, B2 or B3. The thickness
of the organic layers was determined with atomic force microscopy, which also presented
mean surface roughness lower than 0.70 nm, see Fig. 12S in supporting information. B4 was
not used in the electroluminescent layer due to the low PLQY. A co-host system containing
TcTA:OXD-7 (7:3 in Wt) as both hole and electron transport layer was conveniently chosen
according to the singlet energy levels to favor energy transfer to the guest, adequate charge
carrier transport and balance inside the emitting layer (EML), as well as good thin film
formation.[23] The PFNBr:TEAB was used as an electron transporting layer (ETL). The total
energy diagrams and electrical properties of the OLEDs build from B1-B3 are presented in
Figure 8 and the results are summarized in Table 5.
Devices with B1-B3 have a turn-on potential around 4 V mainly due to the high
charge injection barriers. The one containing the B2-emitter performed best overall, which
can be mainly attributed to the high PLQY, with maximum values of luminance (L), current
efficiency (CE) and external quantum efficiency (EQE), 357.10 cd m-2, 0.39 cd A-1 and 0.20
% respectively. The latter performance is superior to and more stable than some of the
recently reported evaporated devices[8, 14] and it is among the best known for all-solution
processed devices using boron complexes up to date, if not the best (see Table 5S). Usually,
the TcTA:OXD-7 co-host system is ideal for phosphorescent guests, especially green and
red.[24] However, here it was found promising also for blue fluorescent guests, with good
resulting efficiency. It is important to stress that the devices were processed and characterized
under ambient conditions, which intrinsically reduced their performance. Atomic force
microscopy images of the EML (see Figure 12S) suggest that B3 forms agglomerates,
possibly due to dissolution problems and this may have impaired its performance.
This article is protected by copyright. All rights reserved.

10.1002/ejic.201900265
European Journal of Inorganic Chemistry
Figure 8. (a) Energy diagram and B1-B3 OLED characteristics. (b) current density and luminance vs. voltage.
(c) current efficiency and external quantum efficiency vs. luminance. The arrows indicate the y-labels
corresponding to the curves.
Table 5. Electrical characterization and electroluminescence parameters for the OLEDs B1-B3.
Von (V)[a]
3.80
Lmax (cd m-2)
282.30
CEmax (cd A-1)
EQEmax (%)
0.10
0.17
0.39
0.11
B1
B2
B3
4.39
357.10
0.20
4.27
191.70
0.07
[a] Defined as the voltage at which a luminance of 1 cd m-2 was achieved.
This article is protected by copyright. All rights reserved.

10.1002/ejic.201900265
European Journal of Inorganic Chemistry
The electroluminescence spectra and color coordinates of the devices at maximum
luminance are presented in Figure 9. Here, the emission has bands centered at 462, 478 and
587 nm for devices with B1, B2 and B3, respectively. The shoulder at 430 nm observed for
the B3-device corresponds to the regular PFNBr emission.[25] A lack of an energy barrier to
block holes sometimes allows recombination in ETL and favors emission at these
wavelengths. The EL peaks associated with B2 and B3 agree with solid-state emission while
for B1 a red shift of about 30 nm was observed. The CIE (Commission Internationale de
l’Eclairage, CIEx,y) for the devices using B1-B3 are (0.21;0.28), (0.21;0.29) and (0.51;0.41),
indicating that the compounds can be of interest for constructing white light emitting devices
for lighting systems.[26]
Figure 9. Normalized EL spectra and CIEx,y of OLEDs B1-B3.
3. Conclusion
The design, synthesis and characterization of a new series of boron complexes with the HBT
ligand was presented. Their emission properties could be modulated by inserting donor and
acceptor groups, leading to emission in the blue, green and red regions. Calculations in
conjunction with excited state lifetime measurements confirmed that the complex with a
methoxy donor group (B2) adopts two isomeric conformations, being a strong blue emitter
with very high PLQY (0.88). These compounds exhibited good thermal stability as well as
HOMO and LUMO energy levels appropriate for application in OLEDs. Surprisingly, when
casted as an emissive layer in all-solution processed OLEDs prepared and characterized under
ambient conditions, they demonstrated a superior performance compared to others already
reported, in particular concerning the current efficiency stability. We also wish to highlight
This article is protected by copyright. All rights reserved.

10.1002/ejic.201900265
European Journal of Inorganic Chemistry
the potential of the theoretical formalism used to predict the rates, spectra and actual
fluorescence color of such complexes. That demonstrates the capability of the technique to
assist molecular design by predicting photophysical properties prior to the actual synthesis.
4. Experimental Section
4.1 General Procedures
All starting materials were purchased from Aldrich, Acros, Merck or Vetec and were used
without further purification.
The measurements of carbon, hydrogen and nitrogen percentages for ligands and the
synthesized complexes were performed on a C, H, N - Perkin-Elmer 2400 elemental analyzer.
The spectra in the infrared region were obtained in a Perkin Elmer Spectrum 100
spectrophotometer, in the region of 4000 to 450 cm-1. The method of attenuated total
1
reflectance (ATR) was used to analyze all samples. The H NMR spectra of the ligands were
obtained with a Bruker FT-NMR 200 MHz spectrophotometer and a Bruker FT-NMR Ascend
400 MHz spectrophotometer. Hydrogen chemical shifts were recorded in ppm units using
tetramethylsilane (TMS, δ = 0.00 ppm) as internal reference and deuterated DMSO, acetone
or chloroform as the solvent.
4.2 Synthesis and Characterization
All ligands used were synthesized and purified according to literature methods.[10, 20,27, 28] The
synthesis of the complexes B1, B2, B3 and B4 is described in Scheme 1. In a round bottom
flask the corresponding hydroxybenzothiazole (1.0 mmol) and N,N'-diisopropylethylamine
(DIPEA, dry) (0.25 mL) were dissolved in dry dichloromethane (10.0 mL) under an inert
atmosphere. Then trifluoride diethyl ether (BF3.Et2O) (1.0 mL) was added dropwise at room
temperature. Precipitation began immediately. After one hour, the resulting solid was filtered
and washed with cold dichloromethane. The crude product was purified with crystallization in
1
dichloromethane. All the complexes were characterized by infrared, H NMR, 13C NMR,
melting point and elemental analysis, B1 and B3 are in agreement with the previously
reports[10,20] and the data for B2 and B4 is shown in Figures 1S-6S.
This article is protected by copyright. All rights reserved.

10.1002/ejic.201900265
European Journal of Inorganic Chemistry
Scheme 1. Synthesis of B1-B4 complexes.
B2. Green solid, 72% yield.
Mp: 290.9 - 271 °C. IR (ATR, cm-1): ʋ (C-Har) 3080, ʋ (C-H) 2844, ʋ (C=C) 1572-1485, ʋ
1
(C-N) 1324, ʋ (C-O-C) 1266-1035, ʋ (B-F) 1060, δ (C-Har) 757; H NMR (200 MHz,
acetone-d6) δ/ppm: 8.31 (d, 2H); 7.80 (t, 1H); 7.71 (t, 1H); 7.35 (t, 1H), 7.32 (t, 1H); 7.11 (t,
1H); 3.90 (s, 3H); 13C NMR (100 MHz, DMSO-d6) δ/ppm: 164.9, 152.8, 151.9, 150.9, 135.2,
126.9, 125.5, 122.6, 122.4, 120.2, 119.1, 118.5, 111.6, 56.1; elemental analysis calcd. (%) for
C14H10BF2NO2S: C 55.11, H 3.30, N 4.59; found: C 55.00, H 3.25, N 4.49.
B4. Pale green solid, 74 % yield.
Mp: 291.2 – 292.8 °C. IR (ATR, cm-1): ʋ (C-Har) 3107, ʋ (C=C) 1579-1484, ʋ (N=O) 1508-
1327, ʋ (B-F) 1055, ʋ (C-N) 841, δ (C-Har) 733; 1H NMR (400 MHz, DMSO-d6) δ/ppm: 9.17
(d, 1H); 8.28 (dd, 1H); 8.16 (dd, 2H); 7.58 (td, 1H); 7.49 (td, 1H); 7.27 (d, 1H); 13C NMR (100
MHz, DMSO-d6) δ/ppm: 161.7, 151.7, 140.5, 135.8, 128.5, 127.6, 127.1, 125.9, 124.8, 123.1, 122.6,
120.1, 118.1; elemental analysis calcd. (%) for C13H7BF2N2O3S: C 48.78, H 2.20, N 8,75;
found: C 48,63, H 2.28, N 8.46.
4.3 X-Ray Data Collection and Structure Determination
The X-ray diffraction analysis was performed on a Bruker APEX II DUO diffractometer
using radiation generated by a molybdenum tube (MoKα λ = 0.71073 Å) and a graphite
monochromator. During the analysis, the sample was sustained at 200K. Crystal structure was
solved by direct methods and was refined by the least squares method with complete matrix,
using the programs SIR97 and SHELXL-97 respectively.[29, 30] The graphical representations
of the molecular structures were generated using the program PLATON.[31] The Supporting
Information containing selected crystallographic data is available free of charge on the ACS
Publications website. Full crystallographic tables (including structural factors) have been
This article is protected by copyright. All rights reserved.

10.1002/ejic.201900265
European Journal of Inorganic Chemistry
deposited at the Cambridge Structural Database, as supplementary publication number CCDC
1872766. Copies of these data can be obtained free of charge at www.ccdc.cam.ac.uk.
4.4 Thermogravimetric Analysis and Cyclic Voltammetry Studies
The thermogravimetric analysis was performed with approximately 2 mg of sample in a
platinum crucible, with a nitrogen flow of 100 mL min-1 and a heating ramp of 10 °C min-1 in
a temperature range of 30-900 °C. The equipment used is of the brand Shimadzu, model TGA-
50.
The redox behavior of the complexes was investigated by square wave voltammetry
in a potentiostat-galvanostat BAS (Bioanalytical Systems, Inc.) model Epsilon. A
concentration of 2.510-4 mol L-1 in a CH2Cl2 solution was used under nitrogen atmosphere.
Tetrabutylammonium hexafluorophosphate (0.1 mol L-1) was used as support electrolyte and
the electrochemical cell was composed of three electrodes: work - vitreous carbon; auxiliary -
platinum wire; reference - Ag/AgCl. For correction of the reference electrode, ferrocene (E1/2
vs NHE = 398 mV) was used.
4.5 Spectroscopic Measurements
The electronic spectra in the ultraviolet, visible and near infrared regions were obtained for all
the complexes in the range of 200−800 nm in a Perkin-Elmer spectrophotometer model
Lambda-750. The values of ε are given in L mol-1 cm-1. The emission spectra were obtained
using a Varian Cary Eclipse Fluorescence spectrophotometer. The spectroscopy
measurements were performed in CH2Cl2 solution and solid state using PMMA matrix
containing 10% B1-B4. The analyses were performed using spectroscopic grade solvents,
substrates and quartz cuvettes with a capacity of 1 mL and 1.00 cm of optical path at 25 ºC.
Absolute quantum yields (Φ) were made using a Hamamatsu Photonics Absolute Quantum
Yield Measurement System model c9920-02G which is based on the integrating sphere
method.
Singlet-excited state photoluminescence decay curves were recorded using the time-
correlated individual photon counting method of a FluoTime 200 (PicoQuant) spectrometer.
The excitation was promoted using a 401 nm pulsed diode laser with repeating rates ranging
from 5.0 to 20 MHz. Photoluminescence was collected perpendicular to the excitation and
passed through a polarizer adjusted at the magic angle. The detection system consists of a
monochromator and a multi-channel base photomultiplier (Hamamatsu R3809U-50).
Lifetimes were obtained by fitting a convolution of the instrument's response function and a
This article is protected by copyright. All rights reserved.

10.1002/ejic.201900265
European Journal of Inorganic Chemistry
sum of exponentials to the fluorescence data using the FluoFit® software. The residual plots
and the chi-square (χ2) were used to accurately determine the quality of the accessories during
the analysis procedure.
4.6 Theoretical methods
Geometry optimizations of complexes B1-B4 were carried out in vacuum, using the Orca
4.0.1 software package[32] at the DFT level using the PBE0[33][34] functional and the def2-
TZVP(-f) basis set for all atoms. [35-37] Dispersion effects were included using Grimme’s D3
correction with Becke-Johnson (BJ)[38,39] damping. The RIJCOSX algorithm[40,41] was
employed to accelerate the evaluation of the functionals, using the resolution of identity
approximation for the Coulomb part (RIJ), and the chain of spheres approach for the Fock
exchange (COSX). RIJCOSX requires the specification of an auxiliary basis set for the
Coulomb part and that of a numerical integration grid for the exchange part discussed
elsewhere.[40,41] The vibrational frequencies computed on the optimized geometries of the
complexes included no imaginary ones. TD-DFT was employed to obtain the first 10 singlet
excited states, using the same calculation protocol to optimize the geometry and calculate the
Hessian of the first excited state. Single point calculations using the B2PYLP functional were
also performed[42] for the excitation energies, since it is supposed to give better results,[39]
which was actually corroborated here as well. Images of the complex geometries were
obtained using the Chemcraft program.[43]
In order to calculate the rates, the path integral approach recently developed and
implemented by us in the ORCA_ESD module[19] was employed, using PBE0 Hessians and
geometries for the ground and excited states in all cases. The line width was set to 300 cm-1
and the default Lorentzian line shape was used to achieve a better fit to experimental data.
Also, all frequencies below 300 cm-1 were removed. The default options were used for all the
other parameters. In order to include solvent effects in the excited state energies, the linear
response conductor-like polarizable continuum model (LR-CPCM) was used while the regular
CPCM was employed for the ground state. Since in the case of fluorescence, we already start
at a properly solvated excited state, the actual dielectric constant of the medium was used in
the energy correction expression instead of the square of the refractive index, as an
approximation to the slow relaxation term. The colors were obtained from the spectra using
the SPECTOCOLOR software, also developed by one of us locally. It uses analytic
formulations of the CIE color matching functions [44] to convert from spectral data to several
color spaces and can be obtained free of charge from https://github.com/bersouza/spectocolor.
This article is protected by copyright. All rights reserved.

10.1002/ejic.201900265
European Journal of Inorganic Chemistry
4.7 OLED Device Fabrication
The OLEDs were prepared in ambient conditions on glass substrates coated with ITO (Delta
technologies Co, 4-8 Ω) previously etched and cleaned in a detergent solution, deionized
water, acetone and isopropanol, sequentially in an ultrasonic bath. The cleaned ITO glass was
treated with UV/ozone for 5 min to adjust the work function and improve wetting. A film
with 25 nm of poly(3,4-ethylenedioxythiophene)/poly(4-styrenesulfonate) (PEDOT:PSS) was
spin-casted and then annealed at 160 °C for 30 min as hole-injection layer. Then the 25 nm-
thick EML composed of Tris(4-carbazoyl-9-ylphenyl)amine (TcTA) and 1,3-bis[2-(4-tert-
butylphenyl)-1,3,4-oxadiazo-5-yl]benzene (OXD-7) TcTA:OXD-7 (7:3) co-host were
dissolved in trichloroethylene with 4 wt% of emitter B1-B3 complexes previously optimized,
and was spin-coated and annealed at 60 °C for 10 min. Afterward, a 35 nm-thick ETL
containing
poly[(9,9-bis(3'-((N,N-dimethyl)-N-ethylammonium)-propyl)-2,7-fluorene)-alt-
2,7-(9,9-dioctylfluoren)] dibromide (PFNBr) and tetraethylammonium bromide (TEAB)
PFNBr:TEAB (10:3) dissolved in methanol:water (8:2) was spin-coated and annealed at 60
°C for 10 min.[45] Finally, 1.3 nm of cesium fluoride (CsF) as inject electron layer (IEL) and
100nm of Al as cathode was deposited in a high-vacuum thermal deposition system at a rate
of 0.1 and 1.0 A s-1 at a base pressure of 1×10−6 Torr, respectively. The active area for all
devices was about 4 mm2 and the final device structure was ITO/Pedot:PSS (25
nm)/TcTA:OXD-7:BX (25 nm)/PFNBr:TEAB(35 nm)/CsF (1.3 nm)/Al (100 nm) with
BX=B1, B2 or B3.
The full characterization of the devices was realized using a Keithley Model 2400
source and a Hamamatsu Photonics Absolute Quantum Yield Measurement System model
c9920-02G at ambient conditions without encapsulation. The thickness and morphological
aspects of the thin films were determined with a Nanosurf Flex AFM, operating in tapping
mode under ambient conditions with a scanning rate of 1.0 Hz and 512×512 pixels.[46]
Acknowledgements
The authors are grateful to CNPq (Proj. Univ. 401028/2016-0), FINEP, CAPES, FAPESC,
INCT-INEO, H2020-MSCA-RISE-2017 (OCTA, #778158) and the Max Planck Gesellschaft
for financial support. We are also grateful to Prof. Fernando R. Xavier for the 13C NMR
measurements.
This article is protected by copyright. All rights reserved.

10.1002/ejic.201900265
European Journal of Inorganic Chemistry
References
[1] H. Kuma, C. Hosokawa, Sci. Technol. Adv. Mater. 2014, 15, 034201.
[2] S. Lee, B. Kim, H. Jung, H. Shin, H. Lee, J. Lee, J. Park, Dyes Pigm. 2017, 136, 255–
261.
[3] T. M. H. Vuong, J. Weimmerskirch-Aubatin, J. Lohier, N. Bar, S. Boudin, C. Labbé, F.
Gourbilleau, H. Nguyen, T. T. Dang, D. Villemin, New J. Chem. 2016, 40, 6070–6076.
[4] Z. Zhang, Z. Zhang, H. Zhang, Y. Wang, Dalton Trans. 2017, 47, 127–134.
[5] V. S. Sadu, H.-R. Bin, D.-M. Lee, K.-I. Lee, Sci.Rep. 2017, 7, 242.
[6] D. B. Diaz, A. K. Yudin, Nat. Chem. 2017, 9, 731–742.
[7] D. Frath, J. Massue, G. Ulrich, R. Ziessel, Angew. Chem. Int. Ed. Engl. 2014, 53,
2290–2310.
[8] D. Suresh, B. Ferreira, P. S. Lopes, C. S. B. Gomes, P. Krishnamoorthy, A. Charas, D.
Vila-Viçosa, J. Morgado, M. J. Calhorda, A. L. Maçanita, P. T. Gomes, Dalton Trans.
2016, 45, 15603–15620.
[9] Y.-J. Shiu, Y.-T. Chen, W.-K. Lee, C.-C. Wu, T.-C. Lin, S.-H. Liu, P.-T. Chou, C.-W.
Lu, I.-C. Cheng, Y.-J. Lien, Y.Chi, J. Mater. Chem. C 2017, 5, 1452–1462.
[10] M. Santra, H. Moon, M.-H. Park, T.-W. Lee, Y. K. Kim, K. H. Ahn, Chem. Eur. J.
2012, 18, 9886–9893.
[11] Q. D. Liu, M. S. Mudadu, R. Thummel, Y. Tao, S. Wang, Adv. Funct. Mater. 2005, 15,
143–154.
[12] Y. Liu, J. Guo, H. Zhang, Y. Wang, Angew. Chem. Int. Ed. 2002, 41, 182–184.
[13] M. Sassi, N. Buccheri, M. Rooney, C. Botta, F. Bruni, U. Giovanella, S. Brovelli, L.
Beverina, Sci. Rep. 2016, 6, 34096.
[14] D. Suresh, C. S. B. Gomes, P. S. Lopes, C. A. Figueira, B. Ferreira, P. T. Gomes, R. E.
Di Paolo, A. L. Maçanita, M. T. Duarte, A. Charas, J. Morgado, D. Villa-Viçosa, M. J.
Calhorda, Chem. Eur. J. 2015, 21, 9133–9149.
[15] D.-H. Kim, A. D’Aléo, X. K. Chen, A. D. S. Sandanayaka, D. Yao, L. Zhao, T.
Komino, E. Zaborova, G. Canard, Y. Tsuchiya, E. Choi, J. W. Wu, F. Fages, J. L.
Brédas, J. C. Ribierre, C. Adachi, Nat. Photonics. 2018, 12, 98–104.
[16] D. Suresh; P. S. Lopes, B. Ferreira, C. A. Figueira, C. S. B. Gomes, P. T. Gomes, R. E.
Di Paolo, A. L. Maçanita, M. T. Duarte, A. Charas, J. Morgado, M. J. Calhorda, Chem.
Eur. J. 2014, 20, 4126–4140.
[17] Y. Zhou, J. W. Kim, M. J. Kim, W.-J. Son, S. J. Han, H. N. Kim, S. Han, Y. Kim, C.
Lee, S.-J. Kim, D. H. Kim, J.-J. kim, J. Yoon, Org. Lett. 2010, 12, 1272–1275.
This article is protected by copyright. All rights reserved.

10.1002/ejic.201900265
European Journal of Inorganic Chemistry
[18] J. S. García, F. Maschietto, M. Campetella, I. Ciofini, J. Phys. Chem. A 2018, 122,
375–382.
[19] B. de Souza, F. Neese, R. Izsák, J. Chem. Phys. 2018, 148, 034104.
[20] X. Li, Y.-A. Son, Dyes Pigm. 2014, 107, 182–187.
[21] C. A. Guido, S. Knecht, J. Kongsted, B. Mennucci, J. Chem. Theory Comput. 2013, 9,
2209–2220.
[22] I. Seguy, P. Jolinat, P. Destruel, J. Farenc, R. Mamy, H. Bock, J. Ip, T. P. Nguyen, J.
Appl. Phys. 2001, 89, 5442–5448.
[23] J.-H. Jou, S. Kumar, A. Agrawal, T.-H. Li, S. Sahoo, J. Mater. Chem. C 2015, 3, 2974–
3002.
[24] K. S. Yook, J. Y. Lee, Adv. Mater. 2014, 26, 4218–4233.
[25] S. Ohisa, T. Kato, T. Takahashi, M. Suzuki, Y. Hayashi, T. Koganezawa, C. R.
McNeill, T. Chiba, Y.-J. Pu, J. Kido, ACS Appl. Mater. Inter. 2018, 10, 17318–17326.
[26] S. Reineke, M. Thomschke, B. Lüssem, K. Leo, Rev. Mod. Phys. 2013, 85, 1245–
1293.
[27] J. Massue, D. Frath, G. Ulrich, P. Retailleau, R. Ziessel, Org. Lett. 2012, 14, 230–233.
[28] A. K. Chakraborti, S. Rudrawar, K. B. Jadhav, G.nKaur, S. V. Chankeshwara, Green
Chem. 2007, 9, 1335–1340.
[29] A. Altomare, M. C. Burla, M. Camalli, G. L. Cascarano, C. Giacovazzo, A. Guagliardi,
A. G. G. Moliterni, G. Polidori, R. Spagna, J. Appl. Cryst. 1999, 32, 115-119.
[30] G. M. Sheldrick, Acta Cryst. C Struct. Chem. 2015, 71, 3–8.
[31] A. L. Spek, Acta Cryst. D 2009, 65, 148–155.
[32] F. Neese, Comput. Mol. Sci. 2012, 2, 73-78.
[33] J. P. Perdew, K. Burke, M. Ernzerhof, Phys. Rev. Lett. 1996, 77, 3865–3868.
[34] J. P. Perdew, K. Burke, M. Ernzerhof, Phys. Rev. Lett. 1997, 78, 1396–1396.
[35] A. Schäfer, H. Horn, R. Ahlrichs, J. Chem. Phys. 1992, 97, 2571–2577.
[36] A. Schäfer, C. Huber, R. Ahlrichs, J. Che.m Phys. 1994, 100, 5829–5835.
[37] F. Weigend, R. Ahlrichs, Phys. Chem. Chem. Phys. 2005, 7, 3297–3305.
[38] S. Grimme, J. Antony, S. Ehrlich, H. Krieg, J. Chem. Phys. 2010, 132, 154104.
[39] S. Grimme, S. Ehrlich, L. Goerigk, J. Comput. Chem. 2011, 32, 1456–1465.
[40] R. Izsák, F. Neese, W. Klopper, J. Chem. Phys. 2013, 139, 094111.
[41] R. Izsák, F. Neese, J. Chem. Phys. 2011, 135, 144105.
[42] S. Grimme, F. Neese, J. Chem. Phys. 2007, 127, 154116.
[43] Www.Chemcraftprog.Com.
This article is protected by copyright. All rights reserved.

10.1002/ejic.201900265
European Journal of Inorganic Chemistry
[44] C. Wyman, P. Sloan, P. Shirley, J. of Comp. Graph. Tech. (JCGT) 2013, 2, 1-11.
[45] Y. Tian, X. Xu, J. Wang, C. Yao, L. Li, ACS Appl. Mater. Inter. 2014, 6, 8631–8638.
[46] G. E. Nunes, A. L. Sehnem, I. H. Bechtold, Liq. Cryst. 2012, 39, 205–210.
This article is protected by copyright. All rights reserved.