Tuning the quinoid versus biradicaloid character of thiophene-based heteroquaterphenoquinones by means of functional groups

Article abstract of DOI:10.1021/ja3072385

A series of quinoidal bithiophenes (QBTs) with controlled variations in steric hindrance and electron activity of the substituents has been synthesized. Evidence of their quinoidal versus biradicaloid ground-state electronic character has been experimentally detected and coherently identified as fingerprints by spectroscopic methods such as NMR, UV-vis, multiwavelength Raman. From this analysis, alkoxy groups have been shown to strongly affect the electronic structure and the ground-state energy and stability of QBTs. Quantum-chemical calculations correctly predict the experimental spectroscopic response, even while changing the alkyl on phenone from a tertiary carbon atom to secondary to primary toward an unsubstituted phenone, further confirming the validity of the approach proposed. A control of the electronic structure accompanied by negligible variations of the optical gap of the molecules has thus been demonstrated, extending the potential use of quinoidal species in fields ranging from photon harvesting to magnetic applications.

Full text of DOI:10.1021/ja3072385

Article
pubs.acs.org/JACS
Tuning the Quinoid versus Biradicaloid Character of Thiophene-
Based Heteroquaterphenoquinones by Means of Functional Groups
Eleonora V. Canesi,*,† Daniele Fazzi,† Letizia Colella,†,‡ Chiara Bertarelli,‡ and Chiara Castiglioni‡
†Center for Nano Science and Technology @PoliMi, Istituto Italiano di Tecnologia, Via Pascoli 70/3, 20133 Milano, Italy
‡Dipartimento di Chimica, Materiali e Ing. Chimica “G. Natta”, Politecnico di Milano, Piazza L. Da Vinci 32, 20133 Milano, Italy
S
* Supporting Information
ABSTRACT: A series of quinoidal bithiophenes (QBTs) with controlled
variations in steric hindrance and electron activity of the substituents has
been synthesized. Evidence of their quinoidal versus biradicaloid ground-
state electronic character has been experimentally detected and coherently
identified as fingerprints by spectroscopic methods such as NMR, UV−vis,
multiwavelength Raman. From this analysis, alkoxy groups have been shown
to strongly affect the electronic structure and the ground-state energy and
stability of QBTs. Quantum-chemical calculations correctly predict the
experimental spectroscopic response, even while changing the alkyl on
phenone from a tertiary carbon atom to secondary to primary toward an
unsubstituted phenone, further confirming the validity of the approach
proposed. A control of the electronic structure accompanied by negligible
variations of the optical gap of the molecules has thus been demonstrated,
extending the potential use of quinoidal species in fields ranging from photon harvesting to magnetic applications.
Parallel to the intriguing ground-state properties of quinoidal
compounds, excited states have been investigated as well, both
theoretically (time-dependent DFT, TDDFT, and CASSCF//
CASPT2 calculations) and experimentally (UV−vis and PL).
The presence of a low-lying double exciton state responsible for
the weak features in the high wavelength region of their
absorption spectra was observed, thus highlighting a possible
use of these materials for up- or down-conversion processes
such as singlet fission and triplet−triplet annihilation.14−17
Accordingly, we expect that the ground/excited-state
electronic characters could be tuned not only by increasing
the molecular length but also by the presence and nature of
functional groups bound to the conjugated chain. Tuning the
electronic character without affecting to a large extent the
degree of π electrons conjugation is of interest: indeed, a
separate control of the two characteristics (quinoid vs
biradicaloid structure and excited-state energies) allows the
independent choice of the optical bandgap and the electronic
structure of the ground state. Specifically, longer molecules are
usually characterized by intense absorption in the red and near-
infrared spectral region6 and are interesting for photodetection
in telecommunication18 as well as low bandgap species for
photovoltaics.19 However, biradicaloid structures enable
isomerization,20 which, in principle, could be responsible for
instability and lower charge mobility;21,22 in addition, the
accessibility to dark states is expected to influence the
molecular optical response, in terms both of lifetime and of
INTRODUCTION
■
Thiophene-based quinoidal molecules1−6 have found a renewed
interest in the literature,7−9 especially due to their amphoteric
redox character and high electron affinity,3 which confer on
them air stability and an ambipolar or n-type electronic
behavior.4,10,11
Recently, an unexpected biradicaloid character was shown to
occur in thiophene quinoidal derivatives, which is favored by
increasing the length of the conjugated cores; by combining
theoretical and experimental approaches, Raman spectroscopy
has been lately demonstrated to be a suitable technique to
probe the electronic structure of quinoidal species.7−9,12 The
monitoring, in terms of frequency shifts and intensity
variations, of those Raman active normal modes that are
sensitive to π-electron delocalization (in particular the so-called
Я-modes, corresponding to collective vibrations that maximize
the electron−phonon coupling interaction)13 has been revealed
as a powerful tool to discriminate different highly π-conjugated
compounds on the basis of the character of their ground
electronic state (i.e., quinoid vs biradicaloid). In the case of
thiophene-based heterophenoquinones, it has been found that
by increasing the molecular chain length, a strong red shift
(80−120 cm−1) of the Я-mode frequency is observed until a
turning point, where the Я-mode starts to blue-shift.9 This
behavior has been rationalized in terms of modulation of the
equilibrium molecular structure, while going from a quinoidal
structure (as obtained by density functional theory (DFT)
closed-shell (CS) calculations) for short-length molecules to a
more aromatic-like (e.g., biradicaloid) structure (predicted by
DFT broken symmetry (BS) calculations) for longer systems.8,9
Received: July 24, 2012
Published: October 22, 2012
© 2012 American Chemical Society
19070
dx.doi.org/10.1021/ja3072385 | J. Am. Chem. Soc. 2012, 134, 19070−19083

Journal of the American Chemical Society
Article
The Netherlands), run by a PC with the NOVA 1.8 software. The
working electrode was a 0.071 cm2 glassy carbon GC disk embedded
in Teflon (Amel, IT), and the counter electrode was platinum wire.
Aqueous saturated calomel (SCE) was used as operating reference
electrode, in a double bridge containing a CH2Cl2 + 0.1 M
tetrabutylammonium perchlorate (TBAP) solution, to avoid water
and KCl pollution of the working solution. (Details on cyclic
voltammetry measurements are reported in the Supporting Informa-
tion.)
FT infrared spectra were collected on a Nicolet NEXUS FTIR
spectrometer. Raman spectra were recorded with (i) a Nicolet FT-
Raman Nexus NXR 9650, which employs a Nd/YAG laser with
excitation at λ = 1064 nm, and (ii) a Horiba Jobin Ivon Labram
HR800 equipped with a microscope, using 514.5 and 632 nm
excitation lines. Thermal evolution of the Raman spectra was analyzed
by using the Horiba Jobin Ivon Labram HR800 (excitation 514.5 nm,
nominal power 100 uW) on solid samples, collecting and averaging
two scans with 1 cm−1 spectral resolution.
Synthesis and Electrochemical Characterization. Unless
otherwise specified, all reagents, catalysts, spectroscopic grade, and
reagent grade solvents were commercial (Sigma Aldrich). All reactions
of air- and water-sensitive reagents and intermediates were performed
in dried glassware under argon. The solvents were previously dried by
conventional methods and stored under argon. Air- and water-sensitive
solutions were transferred with hypodermic syringes or double-ended
needles.
intensity of absorptive and emissive transitions. On the other
hand, open-shell species are studied for third-order nonlinear
optical properties23 and for magnetic applications,24 thanks to
their nonvanishing electron spin resonance signals: accordingly,
the presence of such peculiar properties also on shorter
oligomers can be of interest as well. A reliable characterization
of the electronic structure is thus of great relevance, because it
can draw the guidelines for the design of new quinoidal
derivatives.
In this work, we present a series of novel symmetric and
asymmetric thiophene-based heteroquaterphenoquinones,
which differ in the alkyl in the 3,5-positions of the
phenoquinones and bearing electron-active groups on the
bithiophene core. With respect to benzoquinoidal species,
thienoquinoidal oligomers can be easily synthesized, due to
their higher stability, and the versatility of the chemistry of
thiophene gives access to its functionalization.20 5,5′-Bis(3,5-di-
tert-butyl-4-oxo-2,5-cyclohexadiene-1-ylidene)-5,5′-dihydro-
2,2′-bithiophene (QBT-tbu-HH in Figure 1), whose biradica-
General Procedure for the Preparation of Tris[3,5-di(alkyl)-4-
(trimethylsilyloxy)phenyl]boroxine. Butyllithium (2.5 M in hexane,
1.1 equiv) was added dropwise to a stirred suspension of 4-bromo-2,6-
dialkylphenol trimethylsilyl ether25 (1 equiv) in dry TMEDA/THF
(v/v 1:8) at −78 °C. After 1 h at −78 °C, B(O-iPr)3 (3 equiv) was
added, and the resulting solution was stirred at −78 °C for 45 min.
The mixture was stirred overnight, while being allowed to warm to
room temperature. Saturated aqueous NH4Cl solution was added, and
the reaction mixture was extracted with dichloromethane (DCM). The
organic layer was dried on Na2SO4, and the solvent was removed
under reduced pressure. The product was purified using flash
chromatography (silica gel). (a) Tris[3,5-di-tert-butyl-4-
(trimethylsilyloxy)phenyl]boroxine (1): 4-Bromo-2,6-di-tert-butyl-
phenol trimethylsilyl ether (2 g, 5.6 mmol), THF (20 mL),
TMEDA (4 mL), BuLi 2.5 M (8.4 mmol), B(O-iPr)3 (5.2 mL, 22.4
mmol). Flash chromatography using petroleum benzine/ethyl acetate
Figure 1. Chemical structures of the molecules under investigation,
listed by increasing number of alkoxy chains (row) and decreasing
steric hindrance of end units (column).
1
2:1 afforded the desired compound as a yellow solid in 66% yield. H
NMR (400 MHz; CDCl3, 25 °C, TMS): δ 8.194 (s, 6H; H2,H6
phenyl), 1.516 (s, 54H; tert-butyl), 0.471 ppm (s, 27H; −OSiMe3).
(b) Tris[3,5-di-iso-propyl-4-(trimethylsilyloxy)phenyl]boroxine
(2): 4-Bromo-2,6-di-iso-propylphenol trimethylsilyl ether (4.118 g,
12,5 mmol), THF (60 mL), TMEDA (8 mL), BuLi 2.5 M (12.5
mmol), B(O-iPr)3 (37.5 mmol). Flash chromatography using
petroleum benzine/ethyl acetate 2:1 afforded the desired product as
loid nature has been recently investigated,8,14 was considered as
the reference species. In the design of new derivatives, we
considered functional groups affecting the chemical-physical
properties that are commonly addressed for the application
requirements, such as the solubility, the supramolecular
ordering at the solid state, and the energy levels of the frontier
molecular orbitals. Different experimental techniques (NMR,
UV−vis, IR, Raman with different excitation wavelengths),
supported by quantum chemical calculations (DFT, TDDFT,
CASSCF), were used to characterize the new species, aiming
(i) at determining the influence of the substituents on the
stabilization of a predominant quinoidal or a biradicaloid
ground state structure and (ii) at studying the different
interplay between excited states. For each spectroscopic
technique, experimental markers related to the electronic
structure of the investigated species are highlighted.
1
a yellowish solid in 55% yield. H NMR (400 MHz; CDCl3, 25 °C,
3
TMS): δ 7.973 (s, 6H; H2,H6 phenyl), 3.296 (hept, J(H,H) = 6.840
Hz, 6H; −CH iso-propyl), 1.301 (d, 3J(H,H) = 6.840 Hz, 36H; −CH3
iso-propyl), 0.314 ppm (s, 27H; −OSiMe3). The material typical
1
contains ∼20% of the corresponding boronic acid. H NMR (400
MHz; CDCl3, 25 °C, TMS): δ 7.432 (s, 2H; H2,H5 phenyl), 3.296 (m,
2H; −CH iso-propyl), 1.221 (d, 3J(H,H) = 6.840 Hz, 12H; −CH3 iso-
propyl), 0.277 ppm (s, 9H; −OSiMe3).
2-Bromo-3-methoxythiophene (3). A solution of N-bromosucci-
nimide (3.11 g, 17.5 mmol) in dry DMF (12 mL) has been added
dropwise in a solution of 3-methoxythiophene (2 g, 17.5 mmol) in dry
DMF (8 mL). After 1 h, the reaction mixture was extracted with DCM,
the organic layer was dried on Na2SO4, and the solvent was removed
under reduced pressure. Flash chromatography using petroleum
benzine/DCM 9:1 afforded the desired product as a transparent
MATERIALS AND METHODS
■
Apparatus. 1H NMR and 13C NMR spectra were collected
through a Bruker ARX 400. Mass spectroscopy has been carried out
with a Bruker Esquire 3000 plus. UV−vis absorption spectra were
recorded on a Cary 5000 spectrophotometer (Varian). Electro-
chemical measurements were carried out by cyclic voltammetry (CV)
by using an Autolab PGSTAT 30 potentiostat of Eco-Chemie (Utrech,
1
liquid in 88% yield. H NMR (400 MHz; CDCl3, 25 °C, TMS): δ
7.200 (d, J = 5.9 Hz, 1H; H5 thiophene), 6.760 (d, J = 5.2 Hz, 1H; H4
thiophene), 3.889 ppm (s, 3H; −O−CH3).
3-Methoxy-2,2′-bithiophene (4). A solution of 2-bromothiophene
(1.4 g, 8.5 mmol) in dry diethyl ether (5 mL) was added dropwise to a
19071
dx.doi.org/10.1021/ja3072385 | J. Am. Chem. Soc. 2012, 134, 19070−19083

Journal of the American Chemical Society
Article
stirred suspension of magnesium turnings (0.3 g, 12.3 mmol) in dry
diethyl ether (10 mL). After 2 h at reflux, the mixture was cooled to
room temperature, dropped into a stirred solution of 2-bromo-3-
methoxythiophene (1.5 g, 7.7 mmol) and NiCl2(dppp) (0.424 g, 0.77
mmol) in dry diethyl ether (30 mL), and refluxed for 10 h. After an
overnight stirring at room temperature, the reaction mixture was
quenched with ice and an aqueous solution of HCl (2M) and extracted
with diethyl ether/H2O. The organic layer was dried over Na2SO4, and
the solvent was removed under reduced pressure. The product was
purified using silica gel flash chromatography with petroleum benzine/
DCM 9:1 to give the desired product as a yellow liquid in 88% yield.
1H NMR (400 MHz; CDCl3, 25 °C, TMS): δ 7.217 (dd, 3J(H3′,H4′) =
3.6 Hz, 3J(H3′,H5′) = 1 Hz, 1H; H3′ thiophene), 7.182 (dd, 3J(H5′,H4′)
= 5.0 Hz, 3J(H5′,H3′) = 1 Hz, 1H; H5′ thiophene), 7.045 (d, 3J(H,H) =
5.6 Hz, 1H; H5 thiophene), 6.992 (dd, 3J(H43′,H3′) = 3.6 Hz,
3J(H4′,H5′) = 5.3 Hz, 1H; H4′ thiophene), 6.85 (d, J(H,H) = 5.6 Hz,
5,5′-Bis-(3,5-di-tert-butyl-4-hydroxy)phenyl-3,3′-dipentoxy-2,2′-bi-
thiophene (13): 5,5′-Dibromo-3,3′-dipentoxy-2,2′-bithiophene (0.200
g, 0.403 mmol), Pd(PPh3)4 (0.046 g, 0.04 mmol), tris[3,5-di-tert-butyl-
4-(trimethylsilyloxy)phenyl]boroxine (0.348 g, 0.403 mmol),
NaHCO3 solution (5 mL, 5 mmol). Flash chromatography using
petroleum benzine/ethyl acetate 95:5 afforded 13 as a light green solid
in 33% yield. 1H NMR (400 MHz; DMSO, 25 °C, TMS): δ 7.354 (s,
4H; −H phenyl), 7.267 (s, 2H; −OH), 7.106 (s, 2H; H4,H4′
thiophene), 4.202 (t, J(H,H) = 6.32 Hz, 4H; −OCH2), 1.822 (m,
4H; −CH2−), 1.604 (m, 4H; −CH2−), 1.429 (s, 36H; -tert-butyl),
3
1.398 (m, 4H; −CH2−), 0.929 ppm (t, J(H,H) = 7.411 Hz; −CH3).
(d) 5,5′-Bis-(3,5-di-iso-propyl-4-hydroxy)phenyl-3,3′-dipentoxy-
2,2′-bithiophene (14): 5,5′-Dibromo-3,3′-dipentoxy-2,2′-bithiophene
(0.318 g, 0.64 mmol), Pd(PPh3)4 (0.500 g, 0.64 mmol), tris[3,5-di-iso-
propyl)-4-(trimethylsilyloxy)phenyl]boroxine (0.750 g, 0.06 mmol),
NaHCO3 solution (6 mL, 5.76 mmol). Flash chromatography using
petroleum benzine/ethyl acetate 9:1 afforded 14 as a light green solid
in 43% yield. 1H NMR (400 MHz; DMSO, 25 °C, TMS): δ 8.282 (s,
2H; −OH), 7.325 (s, 2H; H4,H4′ thiophene), 7.247 (s, 4H; −H
1H; H4 thiophene), 3.993 ppm (s, 3H; −OCH3).
General Procedure for the Preparation of 5,5′-Dibromo-3,3′-
R1,R2-2,2′-bithiophene. To a solution of 3,3′-R1,R2-2,2′-bithiophene
(1 equiv) in dry DMF was added dropwise N-bromosuccinimide (2
equiv) in dry DMF, in the absence of light. After 1 h, the reaction
mixture was extracted with DCM and washed with water. The organic
layer was dried over Na2SO4, and the solvent was removed under
reduced pressure. The product was purified using flash chromatog-
raphy (silica gel). (a) 5,5′-Dibromo-3-methoxy-2,2′-bithiophene
(6): 3-Methoxy-2,2′-bithiophene (1 g, 5 mmol) in DMF (10 mL),
N-bromosuccinimide (1.78 g, 10 mmol) in DMF (10 mL). Flash
chromatography in petroleum benzine/DCM 9:1 afforded the desired
3
phenyl), 4.212 (t, J(H,H) = 5.883 Hz, 4H; −OCH2), 3.327 (m, 4H;
−CH iso-propyl), 1.827 (m, 4H; −CH2−), 1.627 (m, 4H; −CH2−),
3
1.447 (m, 4H; −CH2−), 1.210 (d, J(H,H) = 6.986 Hz, 24H; −CH3
iso-propyl), 0.950 ppm (t, 3J(H,H) = 7.354 Hz, 6H; −CH3). (e) 5,5′-
Bis-(3,5-dimethyl-4-hydroxy)phenyl-3,3′-dipentoxy-2,2′-bithio-
phene (15): 5,5′-Dibromo-3,3′-dipentoxy-2,2′-bithiophene (0.320 g,
0.645 mmol), Pd(PPh3)4 (0.750 g, 0.065 mmol), 4-hydroxy-3,5-
dimethylphenyl boronic acid pinacol ester (0.400 g, 1.613 mmol),
NaHCO3 solution (6 mL, 5.8 mmol). Flash chromatography using
petroleum benzine/ethyl acetate 2:1 afforded 15 as a light green solid
in 60% yield. 1H NMR (400 MHz; DMSO, 25 °C, TMS): δ 7.254 (s,
2H; H4,H4′ thiophene), 7.199 (s, 4H; −H phenyl), 4.176 (t, 3J(H,H) =
5.883 Hz, 4H; −OCH2), 2.195 (s, 12H; −CH3 phenyl), 1.812 (m, 4H;
−CH2−), 1.573 (m, 4H; −CH2−), 1.433 (m, 4H; −CH2−), 0.959
1
product as a light green solid in 74% yield. H NMR (400 MHz;
CDCl3, 25 °C, TMS): δ 6.915 (d, 3J(H,H) = 3.6 Hz, 1H; H3′
3
thiophene), 6.85 (s, 1H; H4 thiophene), 6.81 (d, J(H,H) = 3.6 Hz,
1H; H4′ thiophene), 3.993 ppm (s, 3H; −OCH3). (b) 5,5′-Dibromo-
3,3′-dipentoxy-2,2′-bithiophene (7): 3,3′-Dipentoxy-2,2′-bithio-
phene26 (1.5 g, 4.44 mmol) in DMF (20 mL), N-bromosuccinimide
(1.74 g, 9.76 mmol) in DMF (10 mL). Flash chromatography using
petroleum benzine/DCM 3:1 afforded the desired product as a yellow
solid in 72% yield. 1H NMR (400 MHz; CDCl3, 25 °C, TMS): δ 6.811
(s, 2H; H4,H4′ thiophene), 4.039 (t, 3J(H,H) = 6.539 Hz, 4H;
−OCH2−), 1.836 (m, 4H; −CH2−), 1.481 (m, 4H; −CH2−), 1.409
3
ppm (t, J(H,H) = 7.383 Hz, 6H; −CH3 pentoxy). (f) 5,5′-Bis-(4-
hydroxy)phenyl-3,3′-dipentoxy-2,2′-bithiophene (16): 5,5′-Dibro-
mo-3,3′-dipentoxy-2,2′-bithiophene (1 g, 2.016 mmol), Pd(PPh3)4
(0.233 g, 0.2 mmol), 4-hydroxyphenylboronic acid (0.8346 g, 6.048
mmol), NaHCO3 solution (18 mL, 18.145 mmol). Flash chromatog-
raphy using petroleum benzine/ethyl acetate 2:1 afforded 16 as a
3
1
(m, 4H; −CH2−), 0.948 ppm (t, J(H,H) = 7.193 Hz, 6H; −CH3).
green powder in 66% yield. H NMR (400 MHz; DMSO, 25 °C,
TMS): δ 7.447 (d, 3J(H,H) = 8.719 Hz, 4H; H2,H6 phenyl), 7.278 (s,
2H, H4,H4′ thiophene), 6.8 (d, 3J(H,H) = 8.719 Hz, 4H; H3,H5
General Procedure for the Preparation of 5,5′-Bis(3,5-dialkyl-4-
hydroxy)phenyl-3,3′-R1,R2-2,2′-bithiophene. To a solution of 5,5′-
dibromo-3,3′-R1,R2-2,2′-bithiophene (1 equiv) and Pd(PPh3)4 (0.1
equiv) in previously degassed DME (2:1 with respect to water) were
rapidly added boronic acid (2.5 equiv) or boroxine (1 equiv) and
aqueous base solution (9 eq, 1 M). The resulting mixture was refluxed
overnight. Next, it was allowed to cool to room temperature, poured
into water, and extracted with diethyl ether or DCM. The organic layer
was dried over Na2SO4, and, after solvent removal, the raw product
was purified using flash chromatography (silica gel). (a) 5,5′-Bis-(3,5-
di-tert-butyl-4hydroxy)phenyl-2,2′-bithiophene (11): 5,5′-Dibromo-
2,2′-bithiophene (0.413 g, 1.276 mmol), Pd(PPh3)4 (0.147 g, 0.1276
mmol), tris[3,5-di-tert-butyl-4-(trimethylsilyloxy)phenyl]boroxine
(1.165 g, 1.276 mmol), Na2CO3 solution (11.5 mL, 11.484 mmol).
Flash chromatography with petroleum benzine/DCM 9:1 afforded the
3
phenyl), 4.173 (t, J(H,H) = 6.321 Hz, 4H; −OCH2), 1.812 (m, 4H;
−CH2−), 1.533 (m, 4H; −CH2−), 1.412 (m, 4H; −CH2−), 0.930
3
ppm (t, J(H,H) = 7.411 Hz, 6H; −CH3).
General Procedure for the Preparation of 5,5′-Bis-(3,5-R3,R3-4-
oxo-2,5-cyclohexadiene-1-ylidene)-3,3′-R1,R2-2,2′-dihydroxy Bithio-
phene. To a solution of 5,5′-bis-(3,5-R3,R3-4-hydroxy)phenyl-3,3′-
R1,R2-2,2′-bithiophene (1 equiv) in DCM were added K3Fe(CN)6 (10
equiv) and an aqueous solution of KOH (100 equiv, 0.3 M). The
resulting mixture was stirred until the starting material completely
reacted (∼3 h). The reaction mixture was extracted with DCM and
washed with water. Solvent removal afforded the desired product in
quantitative yield. (a) 5,5′-Bis-(3,5-di-tert-butyl-4-oxo-2,5-cyclohex-
adiene-1-ylidene)-2,2′-dihydroxy Bithiophene (QBT-tbu-HH):
5,5′-Bis-(3,5-di-tert-butyl-4-hydroxy)phenyl-2,2′-bithiophene (0.641 g,
1.115 mmol, 75 mL), K3Fe(CN)6 (3.67 g, 11.15 mmol), aqueous
KOH solution (6.25 g, 111.5 mmol, 150 mL). The product was
isolated as a green solid. 1H NMR (400 MHz; CDCl3, 25 °C, TMS): δ
1
desired product as a yellow solid in 87% yield. H NMR (400 MHz;
CDCl3, 25 °C, TMS): δ 7.414 (s, 4H; H2,H6 phenyl), 7.123 (d,
3J(H,H) = 3.8 Hz, 2H; H3,H3′ thiophene), 7.086 (d, 3J(H,H) = 3.8 Hz,
2H; H4,H4′ thiophene), 5.278 (s, 2H; −OH), 1.491 ppm (s, 36H; tert-
butyl). (b) 5,5′-Bis-(3,5-di-tert-butyl-4-hydroxy)phenyl-3-methoxy-
2,2′-bithiophene (12): 5,5′-Dibromo-3-methoxy-2,2′-bithiophene
(0.530 g, 1.5 mmol), Pd(PPh3)4 (0.170 g, 0.15 mmol), tris[3,5-di-
tert-butyl-4-(trimethylsilyloxy)phenyl]boroxine (0.9 g, 1.0 mmol),
Na2CO3 solution (13.5 mL, 13.5 mmol). Flash chromatography
using petroleum benzine/DCM 9:1 afforded the desired product as a
yellow solid in 75% yield. 1H NMR (400 MHz; CDCl3, 25 °C, TMS):
δ 7.418 (s, 2H; −H phenyl), 7.384 (s, 2H; −H phenyl), 7.175 (d,
3
7.530 (s, 2H; H thiophene), 7.415 (d, J(H,H) = 5 Hz, 4H; H2,H6
phenyl), 7.246 (s, 2H; H thiophene), 1.369 ppm (s, 36 H; -tert-butyl).
HRMS (ESI): m/z calcd for C36H44O2S2+H+, 573 [M + H+]; found,
573.3 (M+). FT-IR (KBr, cm−1): 1575, 1484, 1452, 1359, 1323, 1026,
989. (b) 5,5′-Bis-(3,5-di-tert-butyl-4-oxo-2,5-cyclohexadiene-1-yli-
dene)-3-methoxy-2,2′-dihydroxy Bithiophene (QBT-tbu-HO): 5,5′-
Bis-(3,5-di-tert-butyl-4-hydroxy)phenyl-3-methoxy-2,2′-bithiophene
(0.250 g, 0.44 mmol), DCM (75 mL), K3Fe(CN)6 (1.7 g, 5 mmol),
aqueous KOH solution (2.8 g, 50 mmol, 150 mL). The product is
isolated as an iridescent green/red solid. 1H NMR (400 MHz; CDCl3,
25 °C, TMS): δ 7.407 (d, 2H; H phenyl), 7.329 (s, 2H; H phenyl),
3
3J(H,H) = 3.8 Hz, 1H; H3′ thiophene), 7.074 (d, J(H,H) = 3.8 Hz,
1H; H4′ thiophene), δ 6.984 (s, 1H; H4 thiophene), 5.227 (s, 2H;
−OH), 4.016 (s, 3H; −OCH3), 1.429 ppm (s, 36H; tert-butyl). (c)
19072
dx.doi.org/10.1021/ja3072385 | J. Am. Chem. Soc. 2012, 134, 19070−19083

Journal of the American Chemical Society
Article
Scheme 1. Synthetic Route for QBTs
7.229 (s, 2H; H3′,H4′ thiophene), 6.768 (s, 1H; H4 thiophene), 4.1 (s,
3H; −OCH3), 1.367 ppm (s, 36H; -tert-butyl). HRMS (ESI): m/z
calcd for C37H46O3S2 + H+, 603 [M + H+]; found, 603.3 (M+). FT-IR
(KBr, cm−1): 1582, 1529, 1487, 1451, 1358, 1329, 1257, 1237, 1088,
1025, 988. (c) 5,5′-Bis-(3,5-di-tert-butyl-4-oxo-2,5-cyclohexadiene-
1-ylidene)-3,3′-dipentoxy-2,2′-dihydroxy Bithiophene (QBT-tbu-
OO): 5,5′-Bis-(3,5-di-tert-butyl-4-hydroxy)phenyl-3,3′-dipentoxy-2,2′-
bithiophene (0.050 g, 0.067 mmol) in DCM (23 mL), K3Fe(CN)6
(0.22 g, 0.67 mmol), aqueous KOH solution (0.375 g, 6.7 mmol, 17
(CH3), 14.003 (CH3). HRMS (ESI): m/z calcd for C42H56O4S2 + H+,
689 [M + H+]; found, 689.4 (M+). FT-IR (KBr, cm−1): 1565, 1531,
1486, 1459, 1407, 1350, 1250, 1104, 1077, 984. (e) 5,5′-Bis-(3,5-
dimethyl-4-oxo-2,5-cyclohexadiene-1-ylidene)-3,3′-dipentoxy-2,2′-
dihydroxy bithiophene (QBT-me-OO): 5,5′-Bis-(3,5-dimethyl-4-
hydroxy)phenyl-3,3′-dipentoxy-2,2′-bithiophene (0.225 g, 0.389
mmol) in DCM (132 mL), K3Fe(CN)6 (1.279 g, 3.89 mmol),
aqueous KOH solution (2.178 g, 38.9 mmol, 97 mL). The product was
1
isolated as a dark blue solid. H NMR (400 MHz; CDCl3, 25 °C,
1
mL). The product was isolated as a purple/blue solid. H NMR (400
TMS): δ 7.339 (d, 4H; −H phenyl), 6.844 (s, 2H; H4,H4′ thiophene),
MHz; CDCl3, 25 °C, TMS): δ 7.323 (d, 4J(H,H) = 2.409 Hz, 2H; −H
4.308 (t, J(H,H) = 5.955 Hz, 4H; −OCH2), 2.131 (s, 12H; −CH3
3
4
phenyl), 7.307 (d, J(H,H) = 2.409 Hz, 2H; −H phenyl), 6.775 (s,
phenyl), 2.021 (m, 4H; −CH2−), 1.672 (m, 4H; −CH2−), 1.433 (m,
4H; −CH2−), 1.047 ppm (t, 3J(H,H) = 7.443 Hz, 6H; −CH3). HRMS
(ESI): m/z calcd for C34H40O4S2 + H+, 577 [M + H+]; found, 577.2
(M+). FT-IR (KBr, cm−1): 1574, 1534, 1486, 1469, 1409, 1371, 1336,
1267, 1196, 1171, 1067, 1048, 1036, 964. (f) 5,5′-Bis-(4-oxo-2,5-
cyclohexadiene-1-ylidene)-3,3′-dipentoxy-2,2′-dihydroxy Bithio-
phene (QBT-H-OO): 5,5′-Bis-(4-hydroxy)phenyl-3,3′-dipentoxy-
2,2′-bithiophene (0.400 g, 0.763 mmol) in DCM (270 mL),
K3Fe(CN)6 (2.510 g, 7.63 mmol), aqueous KOH solution (4.273 g,
76.3 mmol, 190 mL). The product was isolated as a dark blue solid. 1H
NMR (400 MHz; CDCl3, 25 °C, TMS): δ 7.530 (dd, J(H,H) = 9.997
Hz, 4H; H2,H6 phenyl), 6.853 (s, 2H; H4,H4′ thiophene), 6.511 (dd,
J(H,H) = 9.997 Hz, 4H; H3,H5 phenyl), 4.320 (t, 3J(H,H) = 6.051 Hz,
4H; −OCH2), 2.022 (m, 4H; −CH2−), 1.616 (m, 4H; −CH2−),
3
2H; H4,H4′ thiophene), 4.279 (t, J(H,H) = 5.889 Hz, 4H; −OCH2),
1.99 (m, 4H; −CH2−), 1.640 (m, 4H; −CH2−), 1.465 (m, 4H;
−CH2−), 1.362 (s, 18H; -tert-butyl), 1.353 (s, 18H; -tert-butyl), 0.981
3
ppm (t, J(H,H) = 7.228 Hz, 6H; −CH3). APT NMR (300 MHz;
CDCl3): δ 185.261 (C), 163.934 (C), 152.105 (C), 125.949 (C),
122.135 (C), 106.845 (CH), 73.083 (CH2), 35.626 (C), 35.545 (C),
29.731 (CH3), 28.769 (CH2), 23.3 (CH2), 22.322 (CH2), 13.967
(CH3). HRMS (ESI): m/z calcd for C46H64O4S2 + H+, 745 [M + H+];
found, 745.5 (M+). FT-IR (KBr, cm−1): 1578, 1531, 1491, 1452, 1357,
1331, 1232, 1088, 1026, 988. (d) 5,5′-Bis-(3,5-di-iso-propyl-4-oxo-
2,5-cycloesadiene-1-ylidene)-3,3′-dipentoxy-2,2′-dihydroxy Bithio-
phene (QBT-ipr-OO): 5,5′-Di (3,5-di-iso-propil-4-hydroxy)phenyl-
3,3′-dipentoxy-2,2′-bithiophene (0.060 g, 0.087 mmol) in DCM (30
mL), K3Fe(CN)6 (0.286 g, 0.87 mmol), aqueous KOH solution (0.487
g, 8.7 mmol, 22 mL). The product was isolated as a dark blue solid. 1H
3
1.480 (m, 4H; −CH2−), 1.009 ppm (t, J(H,H) = 7.23 Hz; −CH3).
HRMS (ESI): m/z calcd for C30H32O4S2 + H+, 521 [M + H+]; found,
521.2 (M+). FT-IR (KBr, cm−1): 1614, 1599, 1577, 1533, 1481, 1410,
1354, 1259, 1164, 1025, 966.
4
NMR (400 MHz; CDCl3, 25 °C, TMS): δ 7.273 (d, J(H,H) = 2H;
4
−H phenyl), 7.247 (d, J(H,H) = 2H; −H phenyl), 6.853 (s, 2H;
H4,H4′ thiophene), 4.333 (t, 3J(H,H) = 5.951 Hz, 4H; −OCH2), 3.288
(m, 4H; −CH iso-propyl), 2.033 (m, 4H; −CH2−), 1.880 (m, 4H;
Computational Methods. Density functional theory (DFT)
quantum chemical calculations have been carried out for the whole
series of molecules sketched in Figure 1. Different exchange-
correlation (XC) DFT functionals has been considered using the
hybrid Becke’s three-parameters B3LYP27 and range-separated CAM-
B3LYP28 coupled to a double split Pope basis set, 6-31G**. All of the
optimized molecular structures are in a stable minimum of the ground
state potential, and no imaginary frequencies have been computed.
3
−CH2−), 1.680 (m, 4H; −CH2−), 1.204 (d, J(H,H) = 4.883 Hz,
3
12H; −CH3 iso-propyl), 1.187 (d, J(H,H) = 4.883 Hz, 12H; −CH3
3
iso-propyl), 1.014 ppm (t, J(H,H) = 7.324 Hz, 6H; −CH3). APT
NMR (300 MHz; CDCl3): δ 183.867 (C), 164.136 (C), 152.409 (C),
126.273 (C), 122.688 (C), 107.1 (CH), 73.236 (CH2), 35.626 (C),
28.749 (CH2), 28.337 (CH2), 27.082 (CH), 22.387 (CH2), 22.234
19073
dx.doi.org/10.1021/ja3072385 | J. Am. Chem. Soc. 2012, 134, 19070−19083

Journal of the American Chemical Society
Article
Table 1. HOMO−LUMO Energies, HOMO−LUMO Energy Gap Derived from the Electrochemical Data (EHOMO−LUMO(exp)),
a
and Energy Gap Theoretically Predicted (EGAP(theo)
)
compound
HOMOexp [eV]
LUMOexp [eV]
EHOMO−LUMO
[eV]
EGAP(theo) [eV]
(exp)
QBT-tbu-HH
QBT-tbu-HO
QBT-tbu-OO
−5.086
−5.047
−4.905
−3.954
−3.922
−3.836
1.132
1.125
1.069
1.682 (BS)
1.643 (BS)
1.615
a
Computed values are at the (U)B3LYP/6-31G** (S0-CS) and (S0-BS) level; see Materials and Methods.
As was already reported in other studies,3,8,9 the stability of the
As was recently reported,8 Suzuki coupling between a 5,5′-
dibromo-bithiophene and a para-phenol boronic acid or
anhydride is the most effective reaction to yield aromatic
phenol end-capped α,α′-thienylenes. Regarding phenols
bearing bulky tert-butyl or iso-propyl groups, boronic
derivatives are not commercially available, and tris-phenyl
boroxine precursors (1and 2), which are characterized by high
stability, solubility, and easy purification run, were prepared.
Symmetric 3,3′-dipentoxy-2,2′-bithiophene has been synthe-
sized according to the published procedure,26 and further
brominated with NBS in DMF. For asymmetric 3-methoxy-
2,2′-bithiophene, a Kumada coupling between 2-bromo-3-
methoxy thiophene, obtained from bromination of 3-methoxy
thiophene with NBS, and the freshly prepared Grignard reagent
of 2-bromothiophene was carried out. The Suzuki coupling
between the functionalized 5,5′-dibromo-2,2′-bithiophenes and
the benzene boronic derivatives gave the aromatic precursors of
QBTs, which were subsequently oxidized with potassium
ferricyanide to yield the target molecules.
All of the synthesized species are soluble enough to be
characterized by solution techniques. However, as expected, we
observed a reduction of solubility by moving along the vertical
arrow of Figure 1, that is, by going from tert-butyl to less steric
demanding groups (ipr→me), to unsubstituted phenoquinone.
In particular, QBT-H-OO shows a solubility of just a few mg/
ml in chlorinated solvents. These findings are quite surprising if
we consider that relatively long alkoxy chains are present that
were supposed to contribute to the solubility of the material.
However, it is worth noting that alkoxy chains at the 3,3′-
positions of bithiophene do not cause distortion of the
molecular structure due to a possible interaction between the
lone pairs of sulfur and oxygen,20 and/or to the partial double
bond character of the central CC bond. This may favor a strong
intermolecular packing at the solid state.
ground state DFT wave-function has been tested by the keyword
stable=opt. For those cases where the ground-state closed-shell (S0-CS)
DFT wave-function result to be unstable, a full geometry
reoptimization at the unrestricted UDFT level has been carried out,
computing the equilibrium structure for the corresponding ground-
state broken symmetry (S0-BS) DFT wave-function. For both B3LYP
and CAM-B3LYP functionals, the S0-CS and S0-BS cases have been
considered for each molecule featuring a biradicaloid character. For the
case of QBT-H-OO, a complete-active space SCF calculation CASSCF
featuring 12 active electrons in 12 molecular orbitals (CASSCF-
(12,12)) has been carried out with the 6-31G* basis set. Excited-state
calculations have been carried out at the time-dependent TD-DFT
level by using both B3LYP and CAM-B3LYP functional. Vertical
excitation energies and oscillator strengths have been computed for
each molecule. For the case of QBT-H-OO and QBT-H−OH, the
molecular structure of the singlet dipole allowed excited state (S2) has
been fully optimized. Excited-state force field for S2 state has been
computed. Franck−Condon factors have been evaluated (B3LYP and
CAM-B3LYP), as widely reported in literature (notice that due to the
computational cost, tbu groups have been replaced by H in these
calculations).29−37 Medium effects have been considered by
computing, for both ground and excited states, the equilibrium
geometries and force fields using the IEFPCM method38,39 (solvent =
chloroform). Negligible effects have been observed. All calculations
have been carried out by using the Gaussian 09 package.40 The
ground-state CASSCF(12,12)/6-31G* calculation for QBT-H-OO has
been performed with NWChem code.41
RESULTS AND DISCUSSION
■
Synthesis. All of the molecules synthesized in this work
(Figure 1) are obtained by chemical oxidation of phenol end-
capped bithiophene skeletons, where the quinoid character is
induced by the terminal carbonyl groups. Without side groups,
the molecule is almost insoluble in common organic solvents.
For this reason, introduction of lateral substituents can play the
double role of increasing solubility and tuning the electronic
properties. Two different positions of the conjugated skeleton
have been considered, and substituents have been systemati-
cally varied to yield homologous series of compounds, as
follows: (i) alkyl chains on the 3,5-positions of phenone, tert-
butyl (tbu), iso-propyl (ipr), methyl (me), and, as reference,
hydrogen (H); and (ii) alkoxy on the 3,3′ positions of the
bithienylene core. Molecules are labeled as QBT-(x)-Y where
(x) indicates the alkyl groups on phenone (tbu, ipr, me, H) and
Y refers to the presence of alkoxy groups; specifically Y = OO
means the substitution on both the 3,3′ positions of
bithiophene, whereas Y = HO is used when the molecule
bears just one alkoxy group. Molecules synthesized are
summarized in Figure 1, where the systematic variation of the
molecular structure is highlighted.
On the other hand, the length of the alkoxy chain was
expected not to significantly influence the electronic features of
the molecules (this hypothesis would be further confirmed by
calculations, see below); for this reason, in QBT-tbu-HO,
methoxy has not been converted to pentoxy, with tert-butyls
guaranteeing good solubility.
Details on the synthetic procedures and on characterization
methods are reported in the Materials and Methods.
The effect of the presence of donor alkoxy groups on the
frontier energy levels was first investigated through cyclic
voltammetry. Table 1 summarizes the energies of the frontier
molecular orbitals of QBT-tbu-HH, QBT-tbu-HO, and QBT-
tbu-OO, derived using the Fc+/Fc couple as a standard
according to the following equations:
Because functionalities are linked to two different parts of the
molecule, the synthetic scheme proceeded by preparing the
homologous phenol end-capping groups (1, 2, 8, and 9,
Scheme 1a) and the thienylenes (6, 7, and 10, Scheme 1b),
which were then coupled following a combinatorial approach
(Scheme 1c).
E
HOMO (eV) = −e·[Ep,a(V vs Fc+|Fc) +
4.8(V Fc+|Fc vs zero)]
(1)
19074
dx.doi.org/10.1021/ja3072385 | J. Am. Chem. Soc. 2012, 134, 19070−19083

Journal of the American Chemical Society
Article
temperature. A significant shift of the signals of protons on
thiophenes, which move from 7.2−7.5 ppm down to 6.8 ppm
going from QBT-tbu-HH to QBT-tbu-OO, is also observed, as
a result of the deshielding caused by the vicinal oxygen atoms.
Conversely, the alkyl substitution onto the phenoquinone
produces a negligible effect on the magnetic spectroscopic
response of the aromatic nuclei.
Well-defined NMR spectra are analogously shown by the
other QBT-x-OO molecules. This evidence leads to the
hypothesis that in QBT-x-OO species a sole equilibrium
ground-state isomer exists. The QBT-x-OO quinoidal character,
which is expected to disable isomerization in the ground state,
steric hindrance of the alkoxy group, as well as the oxygen−
sulfur interactions can be invoked to rationalize this hypothesis.
13C NMR gives consistent observations, with QBT-x-OO
molecules showing well-defined spectra, whereas just a few very
weak signals were detected for biradicaloid QBTs, in analogy to
what was previously observed for other quinoidal oligothio-
phenes.6
E
LUMO (eV) = −e·[Ep,c(V vs Fc+|Fc) +
+ 4.8(V Fc+|Fc vs zero)]
(2)
where e corresponds to the number of exchanged electrons,
while E is the peak potential (anodic or cathodic, taken with the
maxima criterion).
As expected, the presence of the alkoxy groups raises both
the energy of the HOMO level and that of the LUMO.
However, the LUMO energy is less affected, thus resulting in a
progressive reduction of the HOMO−LUMO energy gap going
from QBT-tbu-HH to QBT-tbu-OO. Computed DFT values
nicely fit the trend of the experimental energy gap.
NMR spectroscopy provided interesting features about the
structure of quinoidal species: molecule QBT-tbu-HH is known
to be characterized by broad H NMR signals (Figure 2).
1,42
1
Ground-State Stability and Equilibrium Molecular
Structures. For each molecule in the first row of Figure 1,
we performed B3LYP/6-31G** closed-shell (CS) geometry
optimization of the ground state (S0); on the S0-CS optimized
geometry, we further checked the wave-function stability at the
(U)B3LYP broken symmetry (BS) level.4,8,9 If a wave-function
instability is found, we fully optimized the molecular structure
at the (U)B3LYP(BS) level. As above cited, for QBT-tbu-HH a
biradicaloid S0-BS ground state has been calculated as more
stable than the quinoidal S0-CS one, with a stabilization energy
ΔE = (S0(BS) − S0(CS)) = −0.94 kcal/mol and a spin
contamination of the resulting wave-function ⟨S2⟩ = 0.67.
Carrying out the same analysis on QBT-tbu-HO, we found that
the introduction of the methoxy group on the bithienylene
affects the stability of the biradicaloid structure with respect to
the quinoidal one. In particular, the electron-donor ability of
the methoxy group destabilizes the biradicaloid form, thus
reducing ΔE with respect to S0-CS (ΔE = −0.18 kcal/mol with
⟨S2⟩ = 0.35).43 The addition of a second electron-donor group
on the bithiophene core (i.e., QBT-tbu-OO) completely
suppresses the biradicaloid character, and no wave-function
instability (BS) is found. As a consequence, QBT-tbu-OO
results to be fully quinoidal (⟨S2⟩ = 0).44 An interesting issue is
the analysis of the optimized equilibrium geometries of QBT-
tbu-Y derivatives. In Figure 3a, we plot the DFT optimized
bond lengths of the carbon−carbon bonds along the π-electron
conjugation path (excluding the capping CO bonds for
which small variations have been calculated). As has been
already discussed,8 QBT-tbu-HH is characterized by a
remarkable bond length equalization in the central part of the
bithiophene unit by moving from the quinoidal S0-CS to the
more stable biradicaloid S0-BS structure. On the contrary, for
QBT-tbu-HO, which shows a biradicaloid character less marked
than that QBT-tbu-HH, the bond length alternation (BLA) for
the stable S0-BS structure is maintained even in the central part
of the molecule, but, in analogy with QBT-tbu-HH, it appears
to be reduced. In Figure 3b, the bond length difference between
S0-BS and S0-CS optimized structures for QBT-tbu-HO (RS0‑BS
− RS0‑CS) is reported. We can observe that, as it happens for
QBT-tbu-HH,8 the structural relaxations mainly involve the
bithiophene unit and the phenone−thiophene inter-ring bonds,
thus localizing the biradicaloid character in the central part of
the molecule. For the case of QBT-tbu-OO, where a quinoidal
(S0-CS) instead of a biradicaloid structure has been predicted,
1
Figure 2. H NMR spectra of QBT-tbu-HH (top) and QBT-tbu-OO
(bottom) in CDCl3.
Takahashi rationalized these spectroscopic features (and also of
analogous heteroquaterphenoquinones; see ref 1) by dynamic
1H NMR spectroscopy, clarifying that QBT exists predom-
inantly in an S-trans conformation in CDCl3 solution with a
relative ratio between the S-trans and the S-cis of 2.23:1 at −40
°C. Coalescence of the NMR signals of the two conformers has
been indeed observed starting from 10 °C. Markedly different
evidence has been detected for QBT-tbu-OO (Figure 2), which
is characterized by well-defined 1H NMR peaks at room
19075
dx.doi.org/10.1021/ja3072385 | J. Am. Chem. Soc. 2012, 134, 19070−19083

Journal of the American Chemical Society
Article
Figure 3. (a) (U)B3LYP/6-31G** bond lengths (Å) for QBT-tbu-
HH S0-CS (red line) and S0-BS (green line), for QBT-tbu-HO S0-CS
(blue line) and S0-BS (purple line), and for QBT-tbu-OO S0-CS
(black line). Bond numbers run along the molecular backbone (see
molecular sketch). (b) Bond length difference (RS0‑BS − RS0‑CS) for
QBT-tbu-HO.
Figure 4. (a) UV−vis absorption spectra in chloroform of QBT-tbu-
HH (black line), QBT-tbu-HO (red line), QBT-tbu-OO (light blue
line), QBT-ipr-OO (blue line), QBT-me-OO (light green line), and
QBT-H-OO (green line); and UV−vis absorption spectra (b) of QBT-
tbu-HH and (c) of QBT-tbu-OO in different solvents, n-hexane (blue
line), THF (red line), chloroform (orange line), and ethanol (green
line).
CC bond lengths show a strong alternation pattern, with a
quasi double CC bond (bond 9 in Figure 3a) connecting the
two thiophene rings.
We can finally state that the introduction of electron-donor
groups on the thienylenic core results in a destabilization of the
S0-BS biradicaloid structure, thus favoring the quinoidal S0-CS
character.
UV−Visible Absorption Spectroscopy. Experimental
UV−vis spectra have been recorded from dilute chloroform
solutions (<10−5 M), which is a good solvent for all of the
molecules under investigation, thus allowing a direct compar-
ison (Figure 4).
(iii) a marked vibronic structure at energy higher than the
main absorption band.
While the main peak shows a modest red shift as a
consequence of the addition of two donor groups, the general
shape of the absorption feature shows a clear evolution, mainly
due to the intensity increase of the lower energy transition lead
by the addition of the alkoxy groups. On the other hand, the
effect of the alkyl groups is less significant: QBT-(x)-OO
species only differ in the relative intensity of the low energy
band, while the frequency position of the absorption maxima is
almost constant.
Polarity of the solvent has a non negligible effect on the
spectral pattern: as shown in Figure 4b and 4c for two
representative cases, QBT-tbu-HH and QBT-tbu-OO, an
increasing solvent polarity determines a batochromic shift of
the whole absorption pattern and an intensity increase of the
lower energy shoulder. However, while the latter effect is weak
for QBT-tbu-HH, it is strongly emphasized in the presence of
the alkoxy substituents.
All QBT species show similar absorption features dominated
by:
(i) a strong absorption peak in the region of 650−700 nm,
which is assigned (see discussion below) to the strongest
dipole allowed transition, described in terms of one-
electron excitations and characterized by a large
HOMO−LUMO contribution.
(ii) a lower energy feature in the range of 750−800 nm
assigned to a transition to a state with a dominant double
excitation character, which is formally forbidden in dipole
approximation as far as it is weakly coupled with single
excitations.
19076
dx.doi.org/10.1021/ja3072385 | J. Am. Chem. Soc. 2012, 134, 19070−19083

Journal of the American Chemical Society
Article
Experimental UV−vis spectra of dilute solutions (<10−5 M)
in an apolar solvent, n-hexane, have been also considered, to
better simulate a condition of isolated molecules (see Figure 5a,
existence of a low-lying dark double exciton state (named S1)
vibronically coupled to the ground biradicaloid state, as
calculated with high level CASPT2 method.14
The recognition of double exciton states in the whole series
of QBT molecules here considered is relevant, because it is
expected to play a fundamental role in ruling the photophysics
of organic compounds being determinant in processes such as
singlet fission, triplet−triplet annihilation, and conical inter-
sections, as was widely discussed in refs 45−48.
For QBT-tbu-HO, the main absorption band assigned to the
S0 → S2 transition slightly changes its energy with respect to
QBT-tbu-HH (red-shift ≈ 6 nm). The vibronic progression
related to the FC factors of these QBT derivatives is quite
similar, with small changes in the relative intensities of the
vibronic peaks. This result remarks the likeness of the electronic
structure of these two molecules for both the ground and the
excited (dipole allowed S2) states. The absorption features at
low energy, which are assigned to a double exciton (S1) state in
analogy with the previous study,14 characterize both species,
with a stronger intensity for QBT-tbu-HO.49
For the fully quinoidal QBT-(x)-OO (with (x) = tbu, ipr, me,
H), the absorption spectra present (i) a red-shift (∼25 nm) of
the main absorption band with respect to the previous cases,
and (ii) a blue-shift and an intensity increase of the low energy
band, which can be assigned to a transition to a double exciton
state, on the basis of the remarkable analogies with another fully
quinoidal compound, referred as 2P-1T as reported in ref 14.
2P-1T is characterized by the same structure as QBT-tbu-HH,
while featuring only one thiophene ring. Indeed, the UV−vis
spectrum of 2P-1T presents an intense absorption band (S0→
S2) characterized by a vibronic structure and a close lower
energy strong band assigned to an excited state of a different
nature, a double excitation state showing a non-negligible
coupling with single excitations.
Figure 5. (a) UV−vis absorption spectra in n-hexane of QBT-tbu-HH
(black line), QBT-tbu-HO (red line), QBT-tbu-OO (light blue line),
and QBT-ipr-OO (blue line); (b) TDB3LYP/6-31G** calculated
absorption spectra for QBT-H-HO and QBT-H-OO (light blue line);
for the first species, both S0-CS (red line) and S0-BS (dark red line)
geometries have been considered in the calculation of the FC
activities; and (c) diagram of the relative positions of the higher
wavelength absorption bands for QBT-tbu-HH, QBT-tbu-HO, and
QBT-tbu-OO.
These findings definitely prove the similarity between
polyenic molecules50,51 and QBT species, which are charac-
terized by the presence of a low-lying double exciton state
independently from the character (quinoidal or biradicaloid) of
their ground state.
To get further insights into the excited-state energies (both
single and double exciton states), we carefully measured the
wavelengths corresponding to all of the absorption lines by
means of the analysis of the negative second derivative of the
absorption spectra, which makes evident the minor transitions,
often appearing as shoulders52 (the detailed analysis is reported
in the Supporting Information). For each QBT-tbu-Y derivative
(row 1 in Figure 1), the energy difference between the
vibrational levels associated with the main electronic transition
(e.g., S0→S2 main absorption band and high energy shoulders)
sets on a constant value of about 0.18 eV, consistent with a
vibronic progression. Conversely, the lower energy band (S0→
S1) shows a gradual decrease of its distance from the main
absorption band (S0→S2) while increasing the number of
alkoxy substituents on the thienyl skeleton. The results are
summarized in Figure 4c, where the experimental energy of the
main absorption band (S0→S2 0−0 transition) and the energy
difference between the lowest band (S0→S1) and the S0→S2 0−
0 transition are reported. In conclusion, the data collected for
the newly synthesized species, besides being perfectly
consistent with the experimental finding previously reported,
show that the effect of the electro-donor groups is to lower the
energy of S2 and increase that of S1.
where spectra of the molecules QBT-me-OO and QBT-H-OO
are not reported due to their poor solubility in n-hexane).
Spectra in hexane allow a more convenient comparison with
TDDFT simulations (Figure 5b) in vacuo and will be analyzed
in deeper detail making explicit reference to the theoretical
results. Also, in this case, we detected strong transitions in the
red spectral region (650−700 nm) with a vibronic structure at
energy higher than that of the absorption maxima, and a
transition at lower energy assigned to a state with double
excitation character, while no absorption features are recorded
after 1000 nm, down to 3000 nm.
The progression of vibronic peaks is very clear in these
spectra and becomes more intense by increasing the number of
lateral alkoxy chains. As was previously reported,14 for QBT-
tbu-HH the vibronic progression related to the strong, sharp
absorption band at 658 nm has been well reproduced and fully
rationalized by calculating the TDDFT Franck−Condon (FC)
activities for the S0→S2 transition (S2 is the dipole allowed
excited state, mainly described as HOMO→LUMO transition);
weaker bands at higher wavelength have been ascribed to the
19077
dx.doi.org/10.1021/ja3072385 | J. Am. Chem. Soc. 2012, 134, 19070−19083

Journal of the American Chemical Society
Article
TDDFT calculations (both B3LYP and CAM-B3LYP
functionals) have been carried out to compute the vertical
excitations and to optimize the molecular geometry in the
dipole allowed excited state, as well as to calculate its force field
(see also the Supporting Information).53
TDDFT FC activities have been computed by projecting the
excited-state geometry on both S0-CS and S0-BS ground-state
structures obtained for the biradicaloid QBT-H-HO species,
whereas, for the quinoidal QBT-H-OO, only the S0-CS ground-
state structure has been considered (see the Supporting
Information for details).54 The calculated absorption spectra
(normalized on the main absorption peak) are reported in
Figure 5b. We can observe that for the case of QBT-tbu-HO
the intensities of the vibronic progression, calculated
considering the S0-BS (biradicaloid) DFT geometry, are
underestimated with respect to the experimental ones; on the
contrary, the FC activity is overestimated when the ground S0-
CS geometry is used in the calculation. These findings are
parallel to what we have already reported for QBT-tbu-HH14
and can be related to the difficulty in giving a quantitatively
reliable description of the biradicaloid ground-state geometry,
which is probably a weighted average between the predicted
DFT S0-CS and S0-BS structures. Indeed, for the case of the
fully quinoidal QBT-tbu-OO, the absorption spectrum related
to the strongly dipole allowed transition is well reproduced by
TDDFT calculations in terms of FC progression. We found
some minor differences by evaluating both ground and excited
states using either TDB3LYP or TDCAM-B3LYP methods: in
particular, the latter predicts a more pronounced bond length
alternation (both in the ground and in the excited states),
increasing the relative intensity of the FC progression (further
details in the Supporting Information).55,56
Figure 6. FT-Raman spectra (powder samples) of QBT-tbu-HH
(top), QBT-tbu-HO (middle), and QBT-tbu-OO (bottom).
for QBT-tbu-HO and at about 1060 cm−1 for QBT-tbu-HH.
The Raman spectrum of QBT-tbu-OO is markedly different:
indeed, the 1800−900 cm−1 spectral region is more selective
than in the previous cases and shows the strongest band,
assigned to the Я-mode, at 1430 cm−1, up-shifted by more than
100 cm−1 with respect to the Я-mode transition of the
biradicaloid species. The occurrence of a Я-mode at higher
energy confirms that QBT-tbu-OO has a quinoidal ground
state, as further confirmed by DFT-BS calculations.
In Figure 7, we report the calculated DFT Raman spectra for
the newly synthesized compounds QBT-tbu-HO (Figure 7a)
and QBT-tbu-OO (Figure 7b). In particular, for QBT-tbu-HO,
the experimental FT-Raman spectrum is compared to DFT
predictions obtained according to the following procedures: (i)
standard method, by using the S0-CS (B3LYP) equilibrium
structure and wave-function, (ii) S0-BS (UB3LYP) equilibrium
structure and CS wave-function, (iii) both structures and wave-
functions at S0-BS level, and, finally, (iv) adopting a fictitious
geometry, obtained after displacing the nuclei along the Я-
mode coordinate,8 in conjunction with CS wave-function.
Method (i) does not reproduce the Raman spectra, because
it predicts the Я-mode (stronger line) frequency about 100
cm−1 higher than the experimental one. This result was
expected, because the biradicaloid S0-BS structure is more
stable (ΔE = −0.18 kcal/mol) than the S0-CS quinoidal one.
We thus adopted approaches (ii) and (iii). Approach (ii) takes
into account only the nuclear effect of the molecular geometry
(S0-BS structure instead of S0-CS (see Figure 3)), whereas
approach (iii) explicitly considers the electronic effects
provided by the electron charge distribution typical of the
biradicaloid structure. In case (ii), the calculated Raman
spectrum is similar to that obtained with the fully S0-CS
approach, showing a few strong band with a high energy Я-
mode at 1410 cm−1. Conversely, case (iii) gives a very different
result: the computed Raman spectrum presents many strong
Raman transitions (especially in the region 1600−1400 cm−1),
and the Я-mode is red-shifted at 1305 cm−1, in good agreement
with the experimental frequency value (1300 cm−1). Moreover,
at 1600 cm−1, the calculated Raman spectrum shows, as the
experimental one, a doublet of peaks, due to the splitting of the
quasi-degenerate CO stretching modes of the two carbonyl
groups. The occurrence of this doublet in the Raman spectrum
further supports the biradicaloid character; in fact, it is also
present in the Raman spectrum of QBT-tbu-HH, but it is
In the frame of the TDDFT method, it is impossible to
describe the lower energy double exciton state: while a possible
computational strategy to gain insights in the nature of this
state should be to use the CASPT2 method, due to the high
computational cost, it is unfeasible for the current study, and
the S0→S1 transition is assigned on the basis of empirical
correlations.
Raman Spectroscopy. In this study, we applied the same
approach developed for QBT-tbu-HH8 to investigate the new
series of QBT species, with the double aim (i) to provide
further insights into their electronic nature and (ii) to confirm
the validity of the theoretical and computational methods we
have used so far.8 The analysis is first focused on the most
intense Raman feature, assigned to the so-called Я-mode,
describing the oscillation of the BLA parameter.8,13,57
Subsequently, we discuss also weaker spectral features, to
better support and to accomplish the vibrational analysis.
QBT-tbu-HH and QBT-tbu-HO present similar Raman
spectra with the most intense peak in the 1800−900 cm−1
region located at ∼1300 cm−1 (Figure 6). As demonstrated
before for QBT-tbu-HH,8 the anomalously low frequency
(∼1300 cm−1) of the Я-mode can be justified only taking into
account a biradicaloid ground-state structure. The similarity in
the main Raman features between QBT-tbu-HH and QBT-tbu-
HO spectra further validates the fact that also QBT-tbu-HO has
a ground state with a biradicaloid nature, as previously stated
on the basis of the predicted stabilization energies and DFT-BS
calculations.
Interestingly, the Raman spectra of these two molecules
present several bands of comparable intensity: in particular,
remarkable features are shown in the 1500−1350 cm−1 range
19078
dx.doi.org/10.1021/ja3072385 | J. Am. Chem. Soc. 2012, 134, 19070−19083

Journal of the American Chemical Society
Article
achieved, the Я-mode being correctly predicted in terms of
both frequency and relative intensity; moreover, also the
doublet at 1600 cm−1 is nicely reproduced. It is worthwhile
noting that the remarkable modulation of the Raman spectrum
obtained by the above procedure provides an independent
proof of the existence of a large electron−phonon coupling,
occurring along the Я direction.
Figure 7b compares the calculated B3LYP/6-31G** Raman
spectrum of QBT-tbu-OO and its FT-Raman experimental
counterpart. In this case, the prediction by standard procedure
is in good agreement with the experimental spectrum, with the
Я-mode at 1420 cm−1 resulting as the most intense Raman
peak in the 1800−900 cm−1 region.
Concerning the minor Raman features, QBT-tbu-OO shows
an interesting pattern in the 1370−1300 cm−1 region. These
peaks can be related to the different dynamical coupling
between the molecular backbone vibrations and those of the
alkyl groups attached onto the phenone. Indeed, by reducing
their hindrance (tbu→ipr→me) toward the unsubstituted
species QBT-H-OO, we observed a variation of the spectral
pattern and relative intensities in the region 1200−1400 cm−1
(as reported in the experimental spectra of Figure 8a), while the
1440 cm−1 band, which is related to the Я-mode, remains
almost unaffected.
The spectral patterns in the 1500−1350 cm−1 region are
nicely reproduced by the calculated Raman spectra, as reported
in Figure 8a. In Figure 8b, the normal modes for the Raman
active bands (red arrows) around 1300 cm−1 are sketched.
Specifically, CC and CS bond stretching of the molecular
backbone are coupled to the CH bending of phenone and
thiophene rings and also to the CH2 waggings of the alkoxy
chains. We observe that by changing the capping phenone, a
different dynamical coupling takes place between a Я-like
oscillation, involving the CC bonds in the molecular backbone
and displacements localized on the end groups, resulting in a
modulation of frequencies and relative Raman intensities.
Finally, following the recent works by Navarrete et al.9,24
dealing with other quinoidal thiophene-based series, the
thermal evolution of the Raman spectra can be analyzed to
discuss the possible presence of low energy species, as for
instance low-lying triplet states, which can show non-negligible
population already at room temperature.
Figure 7. (a) Calculated B3LYP/6-31G** Raman spectra for QBT-
tbu-HO evaluated with (i) S0-CS geometry and wave-function, (ii) S0-
BS (UB3LYP) geometry and CS wave-function, (iii) S0-BS geometry
and wave-function, and (iv) a displaced geometry (along the Я-
coordinate) with CS wave-function. The experimental FT-Raman
spectrum is reported as comparison. Calculated frequencies have been
scaled by a factor 0.96814. (b) B3LYP/6-31G** S0-CS calculated
Raman spectra for QBT-tbu-OO (below) and the experimental one
(above). Calculated frequencies have been scaled by a factor of
0.96814.
In the Supporting Information, Raman spectra of QBT-tbu-
HH and QBT-tbu-OO recorded at different temperatures
(−190 to 80 °C) are reported. In the case of QBT-tbu-OO, the
observed spectra evolution with increasing temperature can be
described in terms of a general broadening and frequency shift
of the Raman bands (which is usually ascribed to the increased
molecular mobility), without clear indication of the presence of
different molecular species characterized by a thermally
modulated relative population. The spectra evolution shown
by QBT-tbu-HH is much more intriguing because it is possible
to detect (Figure 9) a peculiar evolution of the structured band
at about 1060 cm−1, which shows an asymmetric shape already
at −190 °C, resulting in a shoulder on the lower frequency side
at increasing temperatures. At T = 0 °C, the small but clear shift
of the maximum frequency (Δν ≅ 1 cm−1) can be interpreted
as the result of the convolution of the two components, which
contributes to the Raman bands with comparable weights, while
the further peak shift at higher temperature, accompanied by
remarkable band broadening, may be ascribed to effects related
to thermal expansion. These experimental findings suggest that
in the QBT-tbu-HH sample, a molecular species different from
absent in that of QBT-tbu-OO (see Figure 6). However, the
prediction of the Raman spectrum according to case (iii) is
necessarily affected by the intrinsic weakness of the DFT theory
in the description of multireference systems, giving spin-
contaminated wave-functions: we may ascribe to this flaw the
inaccuracy of our calculation in the prediction of the intensity
pattern, giving an underestimation of the intensity of the Я-
mode and a general overestimation of the Raman activity in the
1450−1550 cm−1 region (see Figure 7a).
A strategy we have already adopted to correctly reproduce
the Raman spectra of biradicaloid systems8 is the study of the
dependence of the Raman spectrum on the nuclear geometry: a
modulation of the molecular structure starting from the (mainly
quinoidal) S0-CS equilibrium geometry toward a more aromatic
(biradicaloid) one7−9 is obtained by displacing nuclei along the
Я-coordinate. The Raman spectrum (computed with the S0-CS
wave-function) is subsequently calculated at different frozen
nuclei geometries, corresponding to increasingly large displace-
ments. In Figure 7a, we report the spectrum of QBT-tbu-HO,
obtained after a nuclear displacement of ΔЯ = 0.15 A/amu2; a
very good agreement with the experimental spectrum is
19079
dx.doi.org/10.1021/ja3072385 | J. Am. Chem. Soc. 2012, 134, 19070−19083

Journal of the American Chemical Society
Article
frequency component of the 1060 cm−1 band to the occurrence
of molecules in a triplet state. The computed Raman spectra of
triplet states species (see the Supporting Information) do not
support this finding, and further experimental and theoretical
investigations will be necessary for a reliable assignment.
Multiwavelength Raman Scattering. The spectroscopic
characterization has been extended to the study of the Raman
spectral pattern evolution while changing the laser excitation
line (λexc). The experimental FT-Raman spectra presented in
the previous section were recorded by exciting at λexc= 1064
nm, to avoid resonance phenomena with the strong dipole
allowed electronic transitions of QBTs (λabs ≈ 700 nm). λexc at
higher energy approaches the absorption band of QBTs,
possibly matching the resonance condition. Two different
radiations were chosen with photon energy nearly in
correspondence (λexc = 632 nm) or above (λexc = 514 nm)
the main electronic transition of QBTs.
In Figure 10, the spectra of QBT-tbu-YY, obtained by
exciting at different λexc, are reported: as first evidence, the most
intense Raman band, assigned to the Я-mode, still dominates
the spectra, thus confirming that this peculiar vibration has an
intrinsically very strong Raman activity.
Nevertheless, looking at the intensity pattern, some non-
negligible changes are detected especially for the two molecules
featuring a biradicaloid character, while QBT-tbu-OO spectrum
is quite insensitive to the excitation energy. The intensity
pattern variations have opposite trends in different spectral
regions: in the region 1000−1200 cm−1, where normal modes
are more localized onto the thiophene rings, band intensities
raise by increasing the excitation energy; on the opposite,
transitions associated with normal modes at higher wave-
numbers (e.g., 1600 cm−1) mainly involving the phenones
weaken.
Figure 8. (a) Experimental (gray lines) and theoretical (black lines)
Raman spectra for QBT-(x)-OO species (from the bottom, (x) = H,
me, ipr, tbu). Black and gray circles indicate the Я-mode; arrows
indicate the correspondences between theoretically predicted and
experimental bands. Red arrows indicate one of the most intense
Raman peaks around 1300 cm−1. (b) Calculated normal modes for
QBT-(x)-OO relative to the Raman peaks indicated by the red arrows
in panel (a).
This experimental evidence can find a qualitative explanation
in the frame of the Albrecht theory of the Raman scattering.58,59
When dealing with polyconjugated molecules characterized by
a low energy strongly dipole allowed excited state, named “e”
state, both nonresonant and resonant Raman spectra are
dominated by the contribution of this relevant excited state (e).
If state “e” is characterized by a remarkable relaxation of the
equilibrium geometry, the general intensity pattern can be
predicted simply considering the so-called A term,60 ruled by
the change of the equilibrium molecular structure (ΔXe,g) while
passing from the ground (g) to the relevant excited state (e). As
a consequence, those normal modes (Qi) describing a nuclear
trajectory with a non-negligible projection (ΔQie,g) on ΔXe,g
show a remarkably high Raman activity, and the resulting
Raman spectrum is strong and very selective, independently on
the matching of the resonance condition with state “e”. In other
words, resonant and non resonant Raman patterns are similar
each other, as it has been described since long time in the case
of polyenes.13 In this framework, the origin of the strong
intensity of Я-mode can be justified as follows: the relaxation of
the nuclear geometry in the relevant excited state “e” usually
implies a decrease of the BLA, giving a structural reorganization
parallel to the Я-mode displacements, thus enhancing the A
term. Conversely, when the geometry displacement ΔXe,g is
small, the A term is not further dominant, and additional
contributions to the Raman activity must be taken into
consideration; in addition, contributions from states different
from “e” could become of comparable importance. For this
reason, a weaker and less selective Raman spectrum is expected,
Figure 9. Raman spectra of QBT-tbu-HH with temperature: −190 °C
(blue line), −110 °C (orange line), −30 °C (green line), 0 °C (cyan
line), and 80 °C (magenta line).
the lower energy one occurs. However, on the basis of this
evidence, it is difficult to assign the observed rise of the lower
19080
dx.doi.org/10.1021/ja3072385 | J. Am. Chem. Soc. 2012, 134, 19070−19083

Journal of the American Chemical Society
Article
Figure 10. Top: Raman spectra of (a) QBT-tbu-HH; (b) QBT-tbu-HO; and (c) QBT-tbu-OO in chloroform solution, recorded by exciting at 1064
nm (top), 632 nm (middle), and 514 nm (bottom). Bottom: Bond length differences between the TDDFT optimized geometries of the dipole
allowed excited state S2 and the stable ground state, of (d) QBT-tbu-HH; (e) QBT-tbu-HO; and (f) QBT-tbu-OO.
possibly showing a complex evolution while varying the
excitation energy.
CONCLUSIONS
■
New thiophene-based heteroquaterphenoquinones have been
synthesized with bulky to less steric demanding alkyls on the
3,5-positions of the phenoquinone and/or alkoxy groups on the
bithienylene core.
To verify the reliability of the qualitative interpretation
suggested above, we consider (see Figure 10, bottom) the
differences of the equilibrium bond lengths along the
conjugated sequence of CC bonds, as obtained considering
the TDDFT optimized geometry of the S2 state and that of the
more stable ground state (RS2 − RS0), for the three molecules
under investigation (QBT-H-OO (S0-CS), -HO (S0-BS), and
-HH (S0-BS)).
Spectroscopic characterization highlighted a strong contri-
bution of the donor substituents in stabilizing the quinoidal
ground-state structure of these species and provided the
identification of markers of their electronic structure.
Specifically, quinoidal electronic structures, with respect to
biradicaloid ones, show (i) a displacement toward higher
energy and an intensity increase of the absorption band
associated to the double exciton state, at lower energy with
respect to the strongly dipole allowed excited state, mainly
described by a single (HOMO−LUMO) excitation; (ii) a
sharpening of the 1H NMR signals in the region 6−8 ppm; and
(iii) a remarkable shift of 100 cm−1 observed for the strongest
Raman band assigned to the Я-mode. Raman spectra have been
successfully predicted, also when dealing with the subtle
changes of the Raman pattern associated with the presence of
differently hindered phenones, which indeed require a reliable
description of the vibrational couplings.
From the characterization of this homologous series of
molecules, it followed that, while maintaining the same
oligomer length, a perturbation of the π-electron distribution
of the conjugated main molecular axis by introducing side
donor groups turns into a strong stabilization of the quinoidal
electronic structure, with a negligible decrease of the energy
gap. This represents a new insight into the general class of
thiophene-based quinoidal molecules, where only the depend-
ence of the biradicaloid structure on the effective conjugation
length (e.g., Egap) has been so far deeply characterized.
QBT-H-OO shows the largest bond length differences
between the excited and the ground states (ΔXe,g), whereas
for QBT-H-HO and QBT-H-HH, ΔXe,g decreases by increasing
the biradicaloid ground-state character. According to the
qualitative description given above (i.e., large ΔXe,g with high
projection ΔQie,g over Я coordinates), in the case of QBT-H-
OO the general Raman pattern is well reproduced through the
calculation of the Franck−Condon factors for all of the Raman
active normal vibrations (see the Supporting Information).
On the opposite side, QBTs showing a biradicaloid electronic
ground state are characterized by an equalized sequence of CC
bonds along the conjugation path (small BLA) also in the
ground state, thus featuring lower ΔXe,g with respect to QBT-
H-OO. As a result, the Raman spectrum cannot be simply
interpreted only on the basis of term A; other contributions
have to be taken into account. The consequence is that the Я-
mode is not the only mode dominating the Raman spectrum,
but other modes are activated, thus featuring spectra of
biradicaloid QBTs less selective than for the quinoidal species
and more affected by dispersion while changing the excitation
energy. This effect is usually accompanied by a general
weakening of the Raman activity, which no longer benefits
the (usually large) contribution from term A. Experiments
confirm this behavior because Raman spectra recorded at the
same laser power are weaker and noisier in the case of
molecules characterized by a biradicaloid ground state.
19081
dx.doi.org/10.1021/ja3072385 | J. Am. Chem. Soc. 2012, 134, 19070−19083

Journal of the American Chemical Society
Article
(18) Agostinelli, T.; Caironi, M.; Natali, D.; Sampietro, M.; Dassa,
G.; Canesi, E. V.; Bertarelli, C.; Zerbi, G.; Cabanillas-Gonzalez, J.; De
Silvestri, S.; Lanzani, G. J. Appl. Phys. 2008, 104, 114508.
ASSOCIATED CONTENT
* Supporting Information
■
S
NMR spectra, analysis of absorption spectra (second
derivatives), cyclic voltammetry measurements, DFT,
TDDFT, CASSCF(12,12) optimized geometries and absolute
energies for QBT-x-YY, IEFPCM optimized geometries and
absolute energies, TDDFT computed Franck−Condon factors,
FT-Raman spectra of solution and solid-state samples,
temperature-dependent Raman spectra, computed Raman
spectrum of the triplet state of QBT-tbu-HH, and computed
TDDFT resonant Raman spectrum of QBT-H-OO. This
material is available free of charge via the Internet at http://
pubs.acs.org.
(19) Gonzal
J. Am. Chem. Soc. 2011, 133, 16350.
́ ́
ez, S. R.; Ie, Y.; Aso, Y.; Lopez Navarrete, J. T.; Casado, J.
(20) Suzuki, Y.; Miyazaki, E.; Takimiya, K. J. Am. Chem. Soc. 2010,
132, 10453.
(21) Handa, S.; Miyazaki, E.; Takimiya, K.; Kunugi, Y. J. Am. Chem.
Soc. 2007, 129, 11684.
(22) Wu, Q.; Li, R.; Hong, W.; Li, H.; Gao, X.; Zhu, D. Chem. Mater.
2011, 23, 3138.
(23) Nakano, M.; Kishi, R.; Yoneda, K.; Inoue, Y.; Inui, T.; Shigeta,
Y.; Kubo, T.; Champagne, B. t. J. Phys. Chem. A 2011, 115, 8767.
(24) Ponce Ortiz, R.; Casado, J.; Hernan
́ ́
dez, V.; Lopez Navarrete, J.
T.; Ortí, E.; Viruela, P. M.; Milian, B.; Hotta, S.; Zotti, G.; Zecchin, S.;
́
Vercelli, B. Adv. Funct. Mater. 2006, 16, 531.
(25) Takahashi, K.; Suzuki, T.; Akiyama, K.; Ikegami, Y.; Fukazawa,
Y. J. Am. Chem. Soc. 1991, 113, 4576.
(26) Zotti, G.; Gallazzi, M. C.; Zerbi, G.; Meille, S. V. Synth. Met.
1995, 73, 217.
(27) Becke, A. D. J. Chem. Phys. 1993, 98, 5648.
(28) Yanai, T.; Tew, D. P.; Handy, N. C. Chem. Phys. Lett. 2004, 393,
51.
(29) Di Donato, E.; Vanzo, D.; Semeraro, M.; Credi, A.; Negri, F. J.
Phys. Chem. A 2009, 113, 6504.
(30) Negri, F.; Orlandi, G. In Computational Photochemistry; Olivucci,
M., Ed.; Elsevier: New York, 2005; Vol. 16, p 129.
(31) Negri, F.; Zgierski, M. Z. J. Chem. Phys. 1994, 100, 1387.
(32) Negri, F.; Zgierski, M. Z. J. Chem. Phys. 1994, 100, 2571.
(33) Negri, F.; Orlandi, G.; Zerbetto, F. Chem. Phys. Lett. 1988, 144,
31.
(34) Negri, F.; Orlandi, G. J. Phys. B: At. Mol. Opt. Phys. 1996, 29,
5049.
(35) Santoro, F.; Improta, R.; Lami, A.; Bloino, J.; Barone, V. J. Chem.
Phys. 2007, 126, 084509.
AUTHOR INFORMATION
Corresponding Author
eleonora.canesi@iit.it
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work has been partially supported by Fondazione Cariplo,
grant no. 2011-0368, and by the Italian Scientific and
Technological Research Ministry through PRIN project no.
2008JKBBK4_003.
REFERENCES
■
(1) Takahashi, K.; Gunji, A.; Yanagi, K.; Miki, M. J. Org. Chem. 1996,
61, 4784.
(2) Abbotto, A.; Bradamante, S.; Facchetti, A.; Pagani, G. A. J. Org.
Chem. 1997, 62, 5755.
(3) Casado, J.; Miller, L. L.; Mann, R. K.; Pappenfus, T. M.; Higuchi,
H.; Orti, E.; Milian, B.; Pou-Amerigo, R.; Hernandez, V.; Navarrete, L.
T. J. J. Am. Chem. Soc. 2002, 124, 12380.
(4) Pappenfus, T. M.; Chesterfield, R. J.; Frisbie, D. C.; Mann, R. K.;
Casado, J.; Raff, D. J.; Miller, L. L. J. Am. Chem. Soc. 2002, 124, 4184.
(5) Berlin, A.; Grimoldi, S.; Zotti, G.; Osuna, M. R.; Delgado, R. C.
M.; Ortiz, P. R.; Casado, J.; Hernandenz, V.; Navarrete, L. T. J. J. Phys.
Chem. B 2005, 109, 22308.
(6) Takahashi, T.; Matsuoka, K. I.; Takimiya, K.; Otsubo, T.; Aso, Y.
J. Am. Chem. Soc. 2005, 127, 8928.
(7) Casado, J.; Ortiz, R. P.; Navarrete, L.T. J. Chem. Soc. Rev. 2012,
41, 5672.
(8) Fazzi, D.; Canesi, E. V.; Negri, F.; Bertarelli, C.; Castiglioni, C.
ChemPhysChem 2010, 11, 3685.
(9) Ponce Ortiz, R.; Casado, J.; Rodriguez Gonzales, S.; Hernandez,
V.; Navarrete, L. T.; Viruela, P. M.; Ortì, E.; Takimiya, K.; Otsubo, T.
Chem.-Eur. J. 2010, 16, 470.
(36) Santoro, F.; Lami, A.; Improta, R.; Barone, V. J. Chem. Phys.
2007, 126, 184102.
(37) Barone, V.; Bloino, J.; Biczysko, M.; Santoro, F. J. Chem. Theory
Comput. 2009, 5, 540.
(38) Tomasi, J.; Mennucci, B.; Cammi, R. Chem. Rev. 2005, 105,
2999.
(39) Mennucci, B.; Cammi, R. Continuum Solvation Models in
Chemical Physics: From Theory to Applications; Wiley: Chirchester,
U.K., 2007.
(40) Frisch, M. J.; et al. Gaussian 09, revision A.1; Gaussian, Inc.:
Wallingford, CT, 2009.
(41) Valiev, M.; Bylaska, E. J.; Govind, N.; Kowalski, K.; Straatsma,
T. P.; Van Dam, H. J. J.; Wang, D.; Nieplocha, J.; Apra, E.; Windus, T.
L.; de Jong, W. A. Comput. Phys. Commun. 2010, 181, 1477.
(42) Lanata, M.; Bertarelli, C.; Gallazzi, M. C.; Bianco, A.; Del
Zoppo, M.; Zerbi, G. Synth. Met. 2003, 138, 357.
(43) We calculated S0-CS and S0-BS optimized structures also for a
(10) Ribierre, J.-C.; Watanabe, S.; Matsumoto, M.; Muto, T.; Nakao,
A.; Aoyama, T. Adv. Mater. 2010, 22, 4044.
t
model molecule QBT-H-HO, carrying hydrogen atoms instead of Bu
groups on the phenyl moieties, and we found that ΔE is just slightly
increased (ΔE = −0.21 kcal/mol with ⟨S2⟩ = −0.38).
(44) The same result is obtained from calculation on QBT-x-OO,
with x = ipr, me, H.
(45) Zimmerman, P. M.; Bell, F.; Casanova, D.; Head-Gordon, M. J.
Am. Chem. Soc. 2011, 133, 19944.
(11) Suzuki, Y.; Shimawaki, M.; Miyazaki, E.; Osaka, I.; Takimiya, K.
Chem. Mater. 2010, 23, 795.
(12) Ortiz, R. P.; Casado, J.; Hernandez, V.; Navarrete, L. T. J.;
Viruela, M. P.; Orti, E.; Takimiya, K.; Otsubo, T. Angew. Chem., Int. Ed.
2007, 46, 9057.
(13) Castiglioni, C.; Tommasini, M.; Zerbi, G. Philos. Trans. R. Soc.
(46) Barbatti, M. WIREs: Comp. Mol. Sci. 2011, 1, 620.
(47) Barbatti, M.; Aquino, A. J. A.; Szymczak, J. J.; Nachtigallova, D.;
Lischka, H. Phys. Chem. Chem. Phys. 2011, 13, 6145.
(48) Plasser, F.; Barbatti, M.; Aquino, A. J. A.; Lischka, H. Theor.
Chem. Acc. 2012, 131, 1073.
(49) The transition to S1, described as a double exciton state, is
expected to vanish for molecules featuring an inversion symmetry
centre (S1 is a dark state). Asymmetries induced by the environment,
deviation from the molecular planarity, as well as the introduction of
London, Ser. A 2004, 362, 2425.
(14) Di Motta, S.; Negri, F.; Fazzi, D.; Castiglioni, C.; Canesi, E. V. J.
Phys. Chem. Lett. 2010, 1, 3334.
(15) Smith, M. B.; Michl, J. Chem. Rev. 2010, 110, 6891.
́
(16) Paci, I.; Johnson, J. C.; Chen, X.; Rana, G.; Popovic, D.; David,
D. E.; Nozik, A. J.; Ratner, M. A.; Michl, J. J. Am. Chem. Soc. 2006, 128,
16546.
(17) Quarti, C.; Fazzi, D.; Del Zoppo, M. Phys. Chem. Chem. Phys.
2011, 13, 18615.
19082
dx.doi.org/10.1021/ja3072385 | J. Am. Chem. Soc. 2012, 134, 19070−19083

Journal of the American Chemical Society
Article
asymmetric chemical substituents can concur to the activation of the
S0−S1 transition.
(50) Orlandi, G.; Zerbetto, F.; Zgierski, M. Z. Chem. Rev. 1991, 91,
867.
(51) Kohler, B. E.; Spangler, C.; Westerfield, C. J. Chem. Phys. 1988,
89, 5422.
(52) Hagler, T. W.; Pakbaz, K.; Voss, K. F.; Heeger, A. J. Phys. Rev. B
1991, 44, 8652.
(53) Despite the fact that biradicaloid species, as QBT-tbu-HH and
QBT-tbu-HO, need multireference methods for a reliable description
of both ground and excited states, we used here TDDFT calculations
as the method is likely accurate for describing dipole allowed excited
state characterized by single excitations and shows good performances
in evaluating energy gradients and Hessians.
(54) For TDDFT FC calculations, we considered only hydrogen
t
capped QBT species instead of Bu due to the computationally
demanding evaluation of the force fields in the dipole allowed excited
states.
(55) Jacquemin, D.; Planchat, A.; Adamo, C.; Mennucci, B. J. Chem.
Theory Comput. 2012, 8, 2359.
(56) Jacquemin, D.; Mennucci, B.; Adamo, C. Phys. Chem. Chem.
Phys. 2011, 13, 16987.
(57) Fazzi, D.; Canesi, E. V.; Bertarelli, C.; Castiglioni, C.; Negri, F.;
Zerbi, G. J. Raman Spectrosc. 2010, 41, 406.
(58) Albrecht, A. C. J. Chem. Phys. 1960, 33, 156.
(59) Albrecht, A. C. J. Chem. Phys. 1961, 34, 1476.
(60) Andruniow
3096.
́
, T.; Olivucci, M. J. Chem. Theory Comput. 2009, 5,
19083
dx.doi.org/10.1021/ja3072385 | J. Am. Chem. Soc. 2012, 134, 19070−19083