Addressing Charge-Transfer and Locally-Excited States in a Twisted Biphenyl Push-Pull Chromophore

Article abstract of DOI:10.1002/cphc.201900703

We present the synthesis and spectroscopic characterization of a twisted push–pull biphenyl molecule undergoing photoinduced electron transfer. Steady-state and transient absorption spectra suggest, in this rigid molecular structure, a subtle interplay be

Full text of DOI:10.1002/cphc.201900703

Accepted Article
Title: Addressing charge-transfer and locally-excited states in a twisted
biphenyl push-pull chromophore
Authors: Elisa Campioli, Somananda Sanyal, Agnese Marcelli,
Mariangela Di Donato, Mireille Blanchard-Desce, Olivier
Mongin, Anna Painelli, and Francesca Terenziani
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: ChemPhysChem 10.1002/cphc.201900703
Link to VoR: http://dx.doi.org/10.1002/cphc.201900703
A Journal of
www.chemphyschem.org

10.1002/cphc.201900703
ChemPhysChem
FULL PAPER
Addressing charge-transfer and locally-excited states in a twisted
biphenyl push-pull chromophore
Elisa Campioli,[a] Somananda Sanyal,[a] Agnese Marcelli,[b] Mariangela Di Donato,[b] Mireille Blanchard-
Desce,[c] Olivier Mongin,[d] Anna Painelli,[a] and Francesca Terenziani*[a]
Abstract: We present the synthesis and spectroscopic
characterization of a twisted push-pull biphenyl molecule undergoing
photoinduced electron transfer. Steady-state and transient absorption
spectra suggest, in this rigid molecular structure, a subtle interplay
between locally-excited and charge-transfer states, whose equilibrium
and dynamics is only driven by solvation. A theoretical model is
presented for the solvation dynamics and, with the support of quantum
chemical calculations, we demonstrate the existence of two sets of
states, having either local or charge-transfer character, that only
“communicate” thanks to solvation, which is the sole driving force for
the charge-separation process.
the vibrational and solvation effects in photoinduced
intramolecular electron transfer. All these molecules are
characterized by at least one conformational degree of freedom
strongly coupled to the charge-transfer process: the rotation of the
amino group in aminobenzonitriles, the angle between the planes
of the two anthryl or phenyl rings in bianthryl and biphenyl
derivatives. Different substitutions and derivatizations have been
performed on those systems,[12–14,17,22–27] in order to facilitate or
inhibit the motion along the coupled nuclear coordinate and
investigate its effects on the charge-transfer process. Controlling
the coupling of the charge-transfer process with nuclear degrees
of freedom helps to single out the effects of solvation. The
interaction with the solvent plays a strong role in electron-transfer
processes since charge-transfer states are characterized by large
permanent dipole moments and hence strongly interact with polar
solvents. Optical spectra of model systems for electron transfer
are often characterized by a subtle interplay between locally-
excited (LE) and charge-transfer (CT) states, strongly affected by
the solvent. Many of these systems display dual fluorescence in
polar solvents[12–15,28,29] and show transient optical behavior
strongly dependent on the solvent itself.[17–19,23,26]
In this paper, we investigate a push-pull substituted biphenyl
compound (hereafter TBP) where the electron-donor and the
electron-acceptor groups are connected to the biphenyl central
core via extended π-conjugated bridges favoring long-distance
electron transfer. Various “push-push”, “pull-pull” and “push-pull”
systems have been reported in the literature based on
functionalized biphenyl cores,[26,27] with applications in two-photon
absorption,[30–32] sensing,[33] light-emitting devices,[34] solar energy
conversion.[29,34,35] The extent of π overlap in the two benzene
rings affects the extent of electronic delocalization over the entire
π system. This varies with the torsional angle between the planes
of the two phenyl rings. Specifically, a large conjugation is
expected for a biphenyl system with coplanar phenyl rings. On the
other hand, delocalization is disrupted in compounds where the
two rings are mutually orthogonal. Thermal motion and the
presence of substituents in different positions affect the torsional
angle, with strong effects on the molecular photophysics.[26,27,29]
In the investigated compound, the 2,2’ and 6,6’ positions of the
biphenyl are substituted by methyl groups, whose steric
hindrance imposes a large twist of the two phenyl rings, resulting
in a very poor conjugation between the two halves of the molecule
and hindering the rotation around the central C-C bond.
Accordingly, conformational motion can be disregarded, leaving
solvation as the only coordinate coupled to the electron transfer
process. Upon photoexcitation, a long-distance and long-lived
charge-transfer state is observed, whose photophysics is
governed by polar solvation. In particular, we analyze the interplay
Introduction
Electron transfer is a widespread and widely exploited process in
many different fields, from life-sustaining reactions and
photosynthesis,[1] to molecular electronics,[2,3] light-emitting
devices[4] and solar cells.[5,6] The understanding of charge-transfer
processes has enormously deepened in the last years thanks to
new experimental techniques (among which time-resolved
spectroscopy) and sophisticated computational tools.[7,8] Electron
transfer is affected and controlled by many variables, including
temperature, donor-acceptor distance, coupling to nuclear
degrees of freedom and solvation.[7–10] In this context, systems
where the electron-donor and the electron-acceptor group are
covalently linked through a rigid spacer are particularly useful
molecular models,[11] where the process is typically referred to as
an intramolecular electron (or charge) transfer event.
Substituted and/or rigidified dialkylaminobenzonitriles,[10,12–15]
bianthryl molecules,[16–21] substituted biphenyl derivatives[22–26]
have been extensively investigated as model systems to unravel
[a]
Dr. Elisa Campioli, Dr. Somananda Sanyal, Prof. Dr. Anna Painelli,
Prof. Dr. Francesca Terenziani
Dipartimento di Scienze Chimiche, della Vita e della Sostenibilità
Ambientale, Università di Parma
Parco Area delle Scienze 17/A, 43124 Parma, Italy
E-mail: francesca.terenziani@unipr.it
[b]
[c]
Dr. Agnese Marcelli, Dr. Mariangela Di Donato
LENS, via N. Carrara 11, 50019 Sesto Fiorentino (FI), Italy
Dr. Mireille Blanchard-Desce
Univ. Bordeaux, Institut des Sciences Moléculaires (CNRS UMR
5255), 33405 Talence, France
[d]
Dr. Olivier Mongin
Univ Rennes, CNRS, ISCR (Institut des Sciences Chimiques de
Rennes), UMR 6226, 35000 Rennes, France
Supporting information for this article is given via a link at the end of
the document.
This article is protected by copyright. All rights reserved.

10.1002/cphc.201900703
ChemPhysChem
FULL PAPER
between the locally-excited (LE) and charge transfer (CT) states
in solution by means of steady-state and time-resolved optical
spectroscopy, as well as by making resort to essential-state
models and quantum chemical calculations. We do not limit the
investigation to the first LE and CT states, but recognize and
identify two distinct manifolds of LE and CT states that are not
optically coupled and can interconvert only thanks to solvation
after photoexcitation.
TBP was synthesized in a five-step sequence (Scheme 1) from
commercially available 3,5-dimethylnitrobenzene (1). Reduction
of 1 in the presence of zinc powder and sodium hydroxide led to
hydrazine 2,[36,37] the Zinin rearrangement of which, performed in
refluxing HCl (10%), afforded 2,2’,6,6’-tetramethylbenzidine
(3).[36,37] Diazotization of both amino groups of 3 and treatment of
the resulting diazonium salt with potassium iodide and iodine gave
the diiodo derivative 4.[37,38] Finally, TBP was obtained from the
latter compound by two successive Sonogashira couplings with
alkynes 5[39] and 7[40] (deprotected in situ with TBAF).
Results and Discussion
Scheme 1. Scheme of the synthesis of TBP.
Table 1. Main spectral properties of TBP in solvents of different polarity and in the solid and nanoaggregate state.
[a]
[b]
[c]
[f]
[g]
λabs
εmax
λem
τ [d]
[ns]
φ [e]
kr
knr
[nm]
327
329
326
[M-1cm-1]
[nm]
367
409
484
[108 s-1]
[108 s-1]
Cyclohexane
Toluene
72600
0.70
4.08
11.7
0.650
0.470
0.150
9.28
5.00
-
-
1.15
1.30
Diethyl ether
0.13
0.72
377
592
-
THF
328
328
331
64100
0.024
0.031
0.005
0.04
-
1.63
-
6.00
370
600
< 0.5 [h]
2.06 (88%); 9.90 (12%)
Triacetin
DCM
-
-
377
625
-
0.01
3.68
2.06 (95%); 9.39 (5%)
DMSO
Powders
ONPs
331
-
-
-
-
-
-
-
-
-
-
-
-
443
504
4.64 (9%); 25.5 (32%); 0.8 (68%)
3.64 (2%); 23.6 (27%); 66.5 (71%)
-
332
49200
0.06
[a] Wavelength of maximum absorption ( 1 nm). [b] Molar extinction coefficient at the wavelength of maximum absorption ( 5%). [c] Wavelengths of maximum
emission ( 2 nm). [d] Fluorescence lifetime(s) extracted from the fit of the fluorescence decay measured at the wavelength reported in the same line of column
c (excitation wavelength 340 nm). [e] Fluorescence quantum yield ( 10%). [f] Radiative decay rate obtained from quantum yield and (average) lifetime. [g] Non-
radiative decay rate. [h] Shorter than the instrumental response.
This article is protected by copyright. All rights reserved.

10.1002/cphc.201900703
ChemPhysChem
FULL PAPER
43]
Steady-state spectroscopy. TBP is a push-pull D-π-A molecule
where a dialkylamino donor group (D) and a cyano acceptor group
(A) are connected to a biphenyl moiety through extended π
bridges (Scheme 1). The 2,2’ and 6,6’ positions of the biphenyl
bear methyl groups, whose steric hindrance forces the two phenyl
rings to be almost perpendicular to one another in the ground
state, as suggested by the crystallographic structure of similar
compounds,[27] and confirmed by quantum-chemical calculations
(see computational part). The twisted biphenyl structure forces
very small conjugation between the donor and the acceptor sides
of the molecule, so that the charge-transfer (CT) transition is
expected to have very low intensity.
The absorption and emission spectra of TBP were measured in
solvents of different polarity, namely cyclohexane, toluene, diethyl
ether, tetrahydrofuran (THF), triacetin, dichloromethane (DCM),
dimethyl sulfoxide (DMSO), as well as in the solid and
nanoaggregate state. Spectra are reported in Figure 1, and main
spectral properties are summarized in Table 1.
tolane emission typically stems from a relaxed trans-bent
geometry. The strongly solvatochromic emission observed at
longer wavelengths has a clear CT origin. Since fluorescence
emission typically stems from the lowest-energy excited state, this
behavior suggests that the lowest-energy excited state has a LE
nature in nonpolar solvents, and a CT character in polar solvents,
in agreement with a stabilization of the CT state in polar solvents.
The residual emission observed from the LE state in polar
solvents is explained in terms of a very small population residing
on the LE state and is justified by the very different radiative
emission rates associated to the LE and CT states (Table 1,
compare kr in cyclohexane with values in polar solvents).
Moreover, the intensity of the residual emission band from the LE
state depends on the solvent and can be related to the typical
solvation time: in triacetin, a polar solvent characterized by a very
slow relaxation time, the residual emission from the LE state is
even more intense than the CT emission.
The absorption spectrum is characterized by a band with a
pronounced vibronic structure, with a maximum centered at ~325
nm. When increasing the solvent polarity, no appreciable shift of
the band is observed, while a broadening occurs when going from
the nonpolar solvent cyclohexane to more polar solvents. The
oscillator strength associated to the absorption band amounts to
~4.5, corresponding to a transition dipole moment of ~10 D. Such
an intense band cannot be assigned to the charge transfer
transition from the donor to the acceptor moiety since the twisted
structure suggests very low intensities for CT bands. We then
assign the absorption band to a transition toward a locally-excited
(LE) state. LE states in TBP can be related to excitations involving
separately the two halves of the molecule, each half
corresponding to a substituted diphenylacetylene (also called
tolane).
The solvent dependence of the fluorescence spectrum is fairly
complex and very interesting. In cyclohexane a strong emission
is detected (quantum yield of 65%, lifetime 0.7 ns), partly
overlapping with the absorption band. For increasing solvent
polarity, the emission band progressively shifts to the red and
loses intensity, the quantum yield reducing to less than 1% in
DCM (in this solvent the radiative decay rate is two orders of
magnitude smaller than the non-radiative decay rate, see Table
1). The fluorescence signal is vanishingly small in more polar
solvents (acetonitrile and DMSO). In medium/high polarity
solvents (diethyl ether, THF, triacetin and DCM) a dual emission
is observed, with the main band located at long wavelengths in a
position depending on the solvent polarity, and a second band
located in the same position as the emission band recorded in
cyclohexane.
The emission in cyclohexane is easily explained as stemming
from the same LE state responsible for the lowest-energy
absorption band. The residual emission observed in higher-
polarity solvents in the same spectral position is quite naturally
ascribed to the same LE state (the associated lifetime, similar to
the one measured in cyclohexane, confirms this hypothesis).
Since LE states are localized on each single half of the molecule,
emission from the lowest-energy LE state of TBP is related to the
emission of a substituted tolane. As discussed in the literature,[41–
Figure 1. Top panel: Absorption and emission spectra of TBP in solvents of
different polarity (cyclohexane = cHex; Toluene = Tol; diethyl ether = Et2O;
tetrahydrofuran = THF; triacetin = tAc; dichloromethane = DCM; dimethyl
sulfoxide = DMSO). Bottom panel: Absorption and emission spectra of TBP as
ONPs in water suspension (black lines) and in the solid state (only emission,
red line).
This article is protected by copyright. All rights reserved.

10.1002/cphc.201900703
ChemPhysChem
FULL PAPER
The transition dipole moment associated with fluorescence can
be estimated from the radiative decay rate according to the
Fermi’s golden rule for spontaneous emission:[44] we obtain
values of ~12 D in cyclohexane, ~5 D in toluene, decreasing down
to ~1 D in DCM. These values are in agreement with the
interpretation of the emission stemming from a strongly allowed
LE state in the nonpolar solvent (in fact the transition dipole
moments associated to absorption and emission are very similar
in cyclohexane), and related to an almost forbidden CT transition
in polar solvents. Toluene seems to be an intermediate case,
where the excited-state population could be distributed on both
types of states, suggesting that the solvent-equilibrated LE and
CT states are almost degenerate in that solvent.
According to our interpretation, in the nonpolar solvent absorption
and emission involve the same state with a strong LE character,
while in polar solvents the vertical excitation still populates a state
with strong LE nature, but emission stems from the relaxed
excited state, which has a dominant CT character.
The LE/CT excited-state scenario extracted from steady-state
and lifetime data is summarized in Scheme 2, where the role of
polar solvation in driving the relaxation from the LE to the CT state
is underlined, and the timescales of the competitive decay
processes are reported.
charge-separated state could be of interest, for example, in solar
energy conversion applications. In order to test if this property is
preserved in the solid state, we characterized the spectroscopic
properties of TBP powders. The powders are significantly
fluorescent (quantum yield not estimated), with a spectrum
(Figure 1, bottom panel) that resembles the emission spectrum in
a low-polarity solvent (something in between toluene and diethyl
ether). Much as in solution, we recognize a dual fluorescence
behavior: a weak band at shorter wavelength (ascribed to a
residual emission from the LE state) and a more intense band at
~440 nm (ascribed to the emission from the CT state). The 3-
exponential decay of the fluorescence is characterized by a > 20
ns component with a ~30% weight, confirming the long-living
character of the CT state in the solid state.
Organic nanoparticles (ONPs) in water suspension were also
prepared, to evaluate the effect of nanoaggregation. Absorption
and fluorescence spectra are reported in Figure 1, bottom panel.
The absorption spectrum is very similar to the absorption spectra
measured in solution and the emission band basically recovers
the fluorescence measured in diethyl ether solution (with similar
quantum yield, 6% vs 15%), suggesting that the environment
experienced by the molecules in the ONPs has a polarity very
close to that solvent. The main component of the 3-exponential
decay of the fluorescence intensity is the longest one, amounting
to ~66 ns, confirming again the long-living nature of the CT state.
In none of the aggregated states we observed excitonic effects:
the main effect of intermolecular interactions in the
nanoaggregates and in the microcrystals is that of a dielectric
continuum.
Theoretical modeling of steady-state spectra. To model a D-
π-A push-pull molecule with a CT/LE interplay through an
essential-state description, three basis (diabatic) states have to
be introduced: the neutral form |D-π-A>, the charge-separated
form |D+-π-A-> (a pure CT state), and a locally excited basis state
that we denote with |D-π*-A> (a pure LE state). The electronic
Hamiltonian is defined by the energies associated with these
three basis states, respectively 0, 2η and 2ζ, and by the mixing
matrix elements: <D-π-A|H|D+-π-A> =  휏, and <D-π*-A|H|D+-π-
A> =  β.[45] We define the dipole moment operator in terms of µ0,
the dipole moment of the zwitterionic |D+-π-A> state, which is
very large if compared with <D-π-A|µ|D-π-A> and <D+-π-A|µ|D-
π-A>, so that the last two can be approximated to 0. To account
for the intensity of the LE transition, we introduce an additional
matrix element of the dipole-moment operator: <D-π-A|µ|D-π*-A>
=µ∗.[45]
Polar solvation is described introducing the electric field F
generated by the orientation of the polar solvent molecules
around the solute. F stabilizes the zwitterionic |D+-π-A> basis
state, so that its energy becomes 2η  μ0F. If the solute is
described as a continuum elastic medium, an additional term,
(μ0F)2/(4εor), enters the Hamiltonian to account for the elastic
energy needed to create the field, where εor measures the solvent
reorganization energy. F describes a slow motion and is dealt with
in the adiabatic approximation.[45] The diagonalization of the F-
dependent electronic problem then leads to F-dependent
eigenvalues, i.e. the adiabatic potential energy surfaces (PES) as
Scheme 2. Energy diagram of the lowest-energy electronic singlet states and
possible radiative and non-radiative transitions of TBP in polar solvents (after
fast solvent relaxation). In blue: decay from the LE state to the ground state,
with associated radiative and non-radiative lifetimes and transition dipole
moment. In red: the decay from the CT state to the ground state, with associated
radiative and non-radiative lifetimes and transition dipole moment. In grey: the
non-radiative, spontaneous decay from the LE to the CT state, driven by the
slow component of solvent relaxation, with a timescale depending on the
specific solvent.
An interesting characteristic of the CT state is the associated
lifetime, amounting to ~10 ns in polar solvents: such a long-lived
This article is protected by copyright. All rights reserved.

10.1002/cphc.201900703
ChemPhysChem
FULL PAPER
reported in Figure 2. As a reference, the diabatic states are also
shown as colored lines. In the specific case, the twisted geometry
implies a small mixing, and the ground state strongly resembles
the neutral |D-π-A> form. The two excited states result from the
mixing of the pure CT and pure LE basis states, but the mixing is
significant only in the region where the LE and CT basis states
are almost degenerate (see the zoom in the Figure). In particular,
focusing on the lowest-energy excited state (the one relevant to
fluorescence), we can recognize an almost pure LE character for
negative values of the solvation coordinate, and an almost pure
CT character for positive values of the solvation coordinate. In
other words, in polar solvents the vertical excitation leads to a
state having a mixed LE/CT nature, with the transition having
mainly a LE character (because of the very small transition dipole
moment associated to the CT transition). After photoexcitation we
expect that the system relaxes along the solvation coordinate,
reaching the minimum of the PES, characterized by a strong
(almost pure) CT nature.
The adiabatic approximation safely applies to the solvation
coordinate, but has to be carefully evaluated when dealing with
vibrational motion. In fact, the adiabatic approximation is doomed
to fail whenever the energy difference between the electronic
states becomes comparable to the vibrational quanta, as it is the
case for the LE/CT three-level model. We therefore rely on the
direct diagonalization of the nonadiabatic Hamiltonian matrix.[45]
Briefly, for fixed F values, the (F-dependent) Hamiltonian matrix
is written on the basis obtained as the direct product of the
electronic diabatic states and the eigenstates of the harmonic
oscillators associated with each vibrational coordinate. The
nominally infinite basis associated with each oscillator is
truncated to the M lowest states. The Hamiltonian matrix is then
diagonalized to get numerically exact vibronic eigenstates,
provided M is large enough to get convergence (M = 10 is used
here). Transition and permanent dipole moments are calculated
from the eigenvectors. Steady-state optical spectra are finally
obtained as sum of spectra obtained for different F-values,
weighting each spectrum for the relevant Boltzmann distribution.
Specifically, the ground-state Boltzmann distribution applies to
absorption spectra while, for fluorescence spectra, the Boltzmann
distribution must be calculated based on the energy of the
emitting state.[46]
The nonadiabatic approach is conceptually simple and applies
smoothly even to difficult cases where large anharmonicities are
present and/or several electronic states have similar energies. Its
main drawback is that it hinders the classification of excited states
as “vibrational” or “electronic”: all nonadiabatic eigenstates are
indeed “vibronic”. This is not an issue in the calculation of
absorption spectra but, to calculate fluorescence spectra, the
emissive state must be identified as the lowest excited state with
electronic character. This is possible through a detailed analysis
of the eigenvectors and/or a comparison of transition dipole
moments from the ground state (electronic transitions are typically
characterized by much larger transition dipole moments than
vibrational ones) or permanent dipole moments (CT states are
characterized by large permanent dipole moments).[45,47]
In the essential-state approach, the model parameters are fixed
as to best reproduce the available experimental data, namely
absorption and emission spectra. Best results for TBP are
obtained with the model parameters reported in Table 2.
Calculated absorption and emission spectra (Figure 3) compare
well with the experimental data: in particular, the solvent polarity
Figure 2. Potential energy surfaces relevant to the diabatic basis states (colored
lines) and to the adiabatic eigenstates (black full lines) as a function of the
solvation coordinate. The diabatic curve relevant to the |D-π-A> state is
superimposed to the ground-state curve. The inset shows a magnification of the
region where the diabatic LE (blue line) and CT (red line) states cross.
mainly affects emission spectra, with
a strong positive
solvatochromic effect, while absorption spectra are barely
affected. The intensity of emission, even if the spectra have been
normalized in Figure 3, is also well reproduced, decreasing of a
factor of ~200 from the less polar to the most polar solvent.
To complete the model, we introduce an effective harmonic
vibrational coordinate, 푄∗, with frequency 휔∗. The diabatic |D-π*-
A> state has displaced minimum with respect to |D-π-A> along 푄∗,
so that the energy gaps acquire a linear dependence on 푄∗: 2휉 −
Table 2. Molecular model parameters entering the calculation of the optical
spectra of TBP, as discussed in the text. Γ = 0.06 eV is the effective intrinsic
width assigned to each transition.
휔
∗
2휀 푄 , where 휀 is the vibrational relaxation energy.[45] We do
√
∗ ∗ ∗
휇0
휇∗
휔∗
휀∗
휂
휏
ζ
β
not insert any vibrational coordinate coupling the pure CT state
with the ground state because, in this specific molecule, the
twisted structure hinders any significant contribution of nuclear
motion to the charge-transfer process.
[eV]
[eV]
[eV]
[eV]
[D]
[D]
[eV]
[eV]
1.81
0.02
1.85
0.05
20
8
0.12
0.16
This article is protected by copyright. All rights reserved.

10.1002/cphc.201900703
ChemPhysChem
FULL PAPER
was centered at 330 nm, while the broadband probe pulse
covered the 400-650 nm spectral region.
The positive band dominating the differential absorption spectra
is due to excited-state absorption (ESA) from the state(s)
populated by the pump pulse, toward higher-energy excited
state(s). In cyclohexane (Figure 4a), this band does not show any
appreciable temporal evolution, apart from
a progressive
decrease in intensity due to the depopulation of the excited state
itself, on a timescale somewhat shorter than 1 ns, in agreement
with the measured fluorescence lifetime (Table 1).
Much more interesting is the dynamics of pump-probe spectra
measured in polar solvents. In all polar solvents (diethyl ether,
triacetin, DMSO), two ESA bands are observed in the transient
spectra, located at about 500 and 450 nm. The relative intensity
of the two bands changes with the time delay: for short pump-
probe delays, the long-wavelength feature dominates, while at
longer time delays the short-wavelength band appears,
progressively acquiring intensity at the expense of the long-
wavelength band. The kinetics of this evolution is clearly governed
by solvation. In fact, while the evolution of the spectral shape is
qualitatively similar in all polar solvents, the associated times are
clearly solvent-dependent, being faster in diethyl ether than in
DMSO, and much slower in triacetin, as shown by the kinetic
traces in Figure 5, relevant to the decay of the long-wavelength
band.
Figure 3. Calculated absorption and emission spectra using the model
parameters reported in Table 2. The spectra of different color simulate different
solvents (εor values in the legend).
Transient spectroscopy. To obtain more information on the
nature of the excited states and on the evolution from a state
having LE character to a state having CT character, we performed
transient absorption measurements in solvents of different
polarity and different relaxation times (Figure 4). The pump pulse
0.05
1 ps
1.5 ps
10 ps
15 ps
50 ps
300 ps
500 ps
0.10
1 ps
a
b
1.5 ps
3 ps
0.04
0.08
5 ps
10 ps
15 ps
0.03
0.06
50 ps
100 ps
0.02
0.01
1 ns
0.04
0.02
0.00
0.00
-0.01
-0.02
-0.02
400
450
500
550
600
650
400
450
500
550
600
650
wavelength(nm)
wavelength(nm)
0.14
1 ps
1 ps
0.10
0.08
0.06
0.04
0.02
0.00
1.5 ps
10 ps
15 ps
d
1.5 ps
3 ps
5 ps
10 ps
15 ps
50 ps
100 ps
c
0.12
0.10
0.08
0.06
0.04
0.02
0.00
50 ps
100 ps
300 ps
500 ps
400
450
500
550
600
650
400
450
500
550
600
650
wavelength(nm)
wavelength(nm)
Figure 4. Transient pump-probe spectra measured in (a) cyclohexane, (b) diethyl ether, (c) DMSO and (d) triacetin.
This article is protected by copyright. All rights reserved.

10.1002/cphc.201900703
ChemPhysChem
FULL PAPER
timescale. The short-wavelength signal corresponds to ESA from
the CT state, which is stabilized by polar solvation. In other terms,
after the excitation the population flows on a potential energy
surface (along the solvation coordinate) having a strong LE
character at zero time and a strong CT character when solvation
is accomplished. In transient spectra the feature related to ESA
from the LE state gradually disappears in favour of the ESA
feature from the CT state, with a nice temporary isosbestic point
(for times significantly shorter than the excited-state lifetime)
confirming an equilibrium between two states.
cHex
triacetin
DMSO
Et2O
To get quantitative information on the kinetics of excited state
relaxation involving the conversion from the LE to the CT state,
the transient absorption data recorded in different solvents were
analyzed using singular value decomposition[48,49] and global
analysis,[50] which consists in simultaneously fitting the kinetics at
all the recorded wavelengths with a combination of exponential
functions. Global analysis was performed by imposing a simple
sequential kinetic decay scheme. Except for DMSO, where four
kinetic constants were used, three kinetic constants were
sufficient for a satisfactory fit of the data. Besides kinetic
constants, global analysis also retrieves the associated spectral
components, called Evolution Associated Difference Spectra
(EADS), which are shown in Figure 6 for all the investigated
solvents.
-5
0
5
10
15
20
Time (ps)
Figure 5. Decay of the long-wavelength band in different solvents: cyclohexane
(cHex, black), diethyl ether (Et2O, red), DMSO (blue), triacetin (green).
The two features observed in the transient absorption spectra can
be ascribed to ESA processes occurring from the LE and CT
states. Specifically, the long wavelength feature, losing intensity
with time delay, is ascribed to ESA from the LE state, which is
directly populated by the pump pulse, and progressively
depopulates towards the CT state on the typical solvation
0.05
2.5 ps
a
0.12
1.8 ps
80.3 ps
4.3 ns
b
0.10
0.08
0.06
0.04
0.02
0.00
-0.02
-0.04
252.5 ps
998 ps
0.04
0.03
0.02
0.01
0.00
-0.01
-0.02
400
450
500
550
600
650
400
450
500
550
600
650
wavelength(nm)
wavelength(nm)
0.14
0.12
0.10
0.08
0.06
0.04
0.02
0.00
0.12
0.10
0.08
0.06
0.04
0.02
0.00
2.7 ps
171 ps
3.7 ns
0.6 ps
3.1 ps
46.6 ps
688 ps
d
c
400
450
500
550
600
650
400
450
500
550
600
650
wavelength(nm)
wavelength(nm)
Figure 6. EADS obtained from global analysis of transient absorption data of TBP recorded in: a) cyclohexane; b) diethyl ether; c) DMSO; d) triacetin. The lifetimes
associated to each spectral component are reported on each panel.
This article is protected by copyright. All rights reserved.

10.1002/cphc.201900703
ChemPhysChem
FULL PAPER
Inspection of the EADS highlights that the kinetics of the
interconversion between the LE and CT states depends on the
solvent properties. In cyclohexane (Figure 6a) the spectral
evolution is very limited, suggesting that no significant dynamics
occurs other than the slow relaxation of the LE state to the ground
state (on a timescale of about 1 ns, in agreement with the
measured fluorescence lifetime). In fact, cyclohexane is a non-
polar solvent, so that no significant solvation is expected. In all the
we describe the dynamics of an excited system, obtained
populating, via a vertical transition, the state responsible for
steady-state emission (whose energy, as a function of the
reaction field, is reported in Figure 7 as a black thick line).[47,54]
This defines an effective zero time where vibrational cooling
already occurred, while the solvent distribution is still frozen in the
same configuration as in the ground state.[54] We are thus
assuming that, after photoexcitation, vibrational coordinates relax
other solvents, a conversion from the LE to the CT state is evident, on a faster timescale with respect to solvation, so that it is possible
showing a kinetic behavior highly dependent on the solvent. In
diethyl ether (Figure 6b), the positive band in the first EADS
(broad feature peaking at about 520 nm, black line) can be
interpreted as the ESA from the LE state. In 1.8 ps this spectral
component evolves towards the second one (red line), where the
positive band peaks at 475 nm and is assigned to the ESA from
the CT state. The intensity of this band slightly decreases in about
80 ps and then finally recovers on a long 4 ns timescale. When
going to the more polar solvent DMSO (Figure 6c), the relaxation
from the LE to the CT state appears to occur bi-exponentially on
a 0.6 ps and 3.1 ps timescales (evolution from black to red and
red to blue EADS). The lifetime of the excited state in this solvent
is reduced compared to diethyl ether: most of the transient signal
in fact recovers on a 46 ps timescale, and only a small residual
signal decays on the ns timescale, in agreement with the poor
fluorescence quantum yield in this solvent. Finally, in triacetin
(Figure 6d) the conversion between the LE and CT states occurs
on a significantly longer timescale and is complete in about 170
ps. The excited-state lifetime is quite long in this solvent, being on
the order of 3.7 ns. Our analysis clearly shows that the evolution
of the excited-state population from the LE to the CT state in
medium/high polarity solvents is dictated by the solvent properties.
The associated times in fact compare well with the typical
solvation times reported in the literature for the three solvents: 1.7
ps for diethyl ether,[51] 3.1 ps for DMSO,[52] 80 ps for triacetin.[53]
A puzzling feature of transient spectra is that neither the band
assigned to excited-state absorption from LE state, nor the band
assigned to excited-state absorption from the CT state show any
significant solvatochromism nor any significant shift with the time
delay (the shifts are on the order of 1000 cm-1, i.e. not larger than
a typical vibrational frequency). This is particularly surprising for
the feature ascribed to the CT state, since our data demonstrate
that the energy of the lowest-energy CT state is very sensitive to
solvation (see e.g. the large emission solvatochromism). This
result can be rationalized assuming that the state reached upon
ESA from the low-lying CT state has a very similar solvatochromic
behavior, i.e. it has a similar CT nature as the low-lying CT state,
so that their energy difference stays roughly constant.
Analogously, the excited-state reached upon absorption from the
LE state should have a LE character too.
to separate the vibrational and solvation motions, assuming that
solvation dynamics starts after vibrational cooling.[55] The zero-
time distribution of the reaction-field coordinate (the solvation
coordinate) is shown in Figure 7 as a grey full line (labeled ‘‘0’’),
while the thermally-equilibrated distribution relevant to the excited
state is reported as a grey dashed line (labeled fluo). This relaxed
distribution is calculated as the Boltzmann distribution relevant to
the emissive state.
The evolution of the probability distribution from the effective zero
time to the fully relaxed limit is calculated using the Smoluchowski
diffusion equation:[56]
2
2
휕푤(퐹,푡)
휕푡
1
푉(퐹)
푤(퐹,푡)
=
[푤(퐹, 푡) 휕
+
휕푉(퐹) 휕푤(퐹,푡) + 푘푇 휕
]
(1)
2
2
휏
푠
휕퐹
휕퐹
휕퐹
휕퐹
where w is the F- and t-dependent probability distribution, V(F) is
the PES governing the evolution, 휏푠 is the average solvation time
(typical of each solvent), k is the Boltzmann constant and T the
temperature (set to 298 K). Consistently with the separation
between nuclear and solvation dynamics, vibrational and
electronic degrees of freedom are assumed to be always in
equilibrium with the instantaneous configuration of the solvent.[54]
Figure 7. The energy of the emissive excited state (thick black line) as a function
of the solvation coordinate, calculated for TBP in diethyl ether (model
parameters in Table 2; εor = 0.58 eV, 휏푠 = 1.7 ps). The probability distribution of
the solvation coordinate is reported for different pump-probe delay times (lines
of different colors, each labeled by the corresponding delay time, in ps,
measured with respect to the effective zero time of solvation). The probability
distributions of the solvation coordinate relevant to the effective zero-time of
solvation (labeled “0”) and to the relaxed excited state (labeled fluo) are reported
as grey lines.
Modeling solvation dynamics. In the pump-probe experiment,
a UV or visible pulse creates an out-of-equilibrium population in
the resonant excited state. A delayed broad-band UV-vis pulse
then addresses the spectral properties of the system while the
population evolves towards the excited-state equilibrium and
finally relaxes back to the ground state. To mimic the experiment,
This article is protected by copyright. All rights reserved.

10.1002/cphc.201900703
ChemPhysChem
FULL PAPER
Results about the evolution of the F-distribution with time,
reported in Figure 7, are calculated setting 휏푠 = 1.7 ps, as relevant
to diethyl ether.[51] Calculated distributions are shown as colored
lines for selected delay times (specified by the labels). The
distribution evolves along a PES characterized by a double-
minimum shape with no barrier in between the two minima. The
zero-time distribution is centered in the region of the metastable
LE-like minimum, and the evolution (solvent relaxation) drives the
system towards the thermodynamically stable minimum having
CT character. Having no barrier to cross, the dynamics is
completely governed by the solvation time.
In our three-state model, excited-state absorption cannot be
evaluated, due to the lack of excited states beyond the two
explicitly accounted for. In order to be able to calculate the
excited-state absorption contribution, we thus introduce two ad-
hoc higher-energy excited states: one, having a transition dipole
moment from the LE state and not being solvatochromic, such as
the LE state; the other, having a transition dipole moment from
the CT state and having the same solvatochromic behavior as the
CT state itself. The calculated transient absorption spectra
(including also the contributions from bleaching and stimulated
emission) are reported in Figure 8. The agreement with the
experimental behavior is impressive, demonstrating that solvation
governs the dynamics of the transient spectra, being faster in
diethyl ether, slightly slower in DMSO and very much slower in
triacetin, according to the respective solvation times (1.7 ps for
diethyl ether,[51] 3.1 ps for DMSO,[52] 80 ps for triacetin[53]). In
particular, solvation drives the population from the initially
populated LE state to the CT state; from the LE state, excited-
state absorption leads to a state having LE character (non-
solvatochromic), while, from the CT state, excited-state
absorption leads to a state having the very same solvatochromic
behavior (i.e. another state having CT character). The solvation-
driven evolution of the population from the LE to the CT state
induces an intensity exchange between the two excited-state
absorption bands, which cross in an isosbestic point, proving that
transitions between excited states having different nature is not
allowed.
Based on transient distributions, differential absorption spectra
can be obtained as the difference in the UV-vis absorbance (as
measured by the probe beam) of the optically pumped system and
the system at rest. The absorbance variation (A) has positive
contributions from ESA and negative contributions from ground-
state bleaching and stimulated emission.
To support our interpretation, we performed quantum chemical
calculations, retrieving the energies of a large number of excited
states in order to study their nature.
Quantum-chemical computations. To model the TBP molecule
for theoretical calculations, we have substituted the long chain
alkyl groups in the amino moiety with methyl groups, in the
assumption that this has marginal spectroscopic effects. The
optimized structure of the molecule in gas phase is reported in
Figure 9, showing that the biphenyl moiety is twisted along the
central C-C bond by ~90°, in agreement with the crystallographic
structure of other substituted-biphenyl molecules.[27] This twist
hinders the conjugation of the π cloud across the twisted bridge.
The ground state geometry was also optimized in different
solvents (cyclohexane, toluene and dichloromethane) obtaining
similar results.
Figure 9. Optimized structure of TBP in gas phase.
Figure 8. Calculated pump-probe (differential absorption) spectra for TBP in
different solvents (model parameters in Table 2; εor = 0.58, 0.88, 0.84 eV and
To look into the transition energies of TBP in different solvents
and in gas phase, we carried out TD-DFT calculations: Table 3
summarizes the first few low-energy transitions. Several flavors of
PCM models are available and we adopted the state-specific
휏푠
= 1.7, 3.1, 80 ps for diethyl ether, dimethyl sulfoxide and triacetin,
respectively) for variable delay time (expressed in ps, see legends).
This article is protected by copyright. All rights reserved.

10.1002/cphc.201900703
ChemPhysChem
FULL PAPER
approach,[57] which takes into account the variations of the
polarization of the solvent following the electronic density
rearrangements of the solute.[58]
molecule, geometry optimization in the excited state is
computationally very expensive and we just succeeded to
optimize the geometry in the gas phase, but we could not
calculate vibrational frequencies. Accordingly, we cannot confirm
that the CT state becomes the fluorescent states in polar solvents.
However, the permanent dipole moment calculated for the vertical
lowest CT state is very large (61 and 72 D in gas phase and DCM,
respectively, against a ground-state dipole moment < 10 D). This
highly polar excited state is expected to be strongly stabilized in
polar solvents due to solvent relaxation. While we cannot obtain
the relaxed excited state, a solvent relaxation energy of ~0.5 eV,
corresponding to a mildly polar solvent, is enough to stabilize the
CT state below the lowest-energy LE state, as also confirmed by
the essential-state modeling.
Table 3. TDDFT Results for TBP in Gas phase and different solvents.
Transition Wavelength /
Oscillator Nature of Transitions
Energy [nm / eV] Strength (MO Contributions)
S0→S1
S0→S2
308 / 4.02 (LE)
293 / 4.23
3.03
0.01
H-1→L (38%)
H→L+1 (52%)
Gas
phase
H-1→L (53%)
H→L+1 (37%)
S0→S3
S0→S1
273 / 4.53 (CT)
315 / 3.93 (LE)
0.00
3.26
H→L (87%)
H-1→L (35%)
H→L+1 (55%)
cHex
S0→S2
302 / 4.10
0.03
H-1→L (56%)
H→L+1 (35%)
S0→S3
S0→S1
274 / 4.52 (CT)
316 / 3.92 (LE)
0.00
3.28
H→L (88%)
Table 4. ZINDO Results for TBP showing the first few low energy transitions.
H-1→L (35%)
H→L+1 (56%)
Transition
Wavelength / Energy
[nm / eV]
Oscillator
Strength
Nature of Transitions
(MO Contributions)
S0→S2
303 / 4.09
0.04
H-1→L (57%)
H→L+1 (34%)
Toluene
S0→S1
357 / 3.47 (LE)
1.91
H-1→L (33%)
S0→S3
S0→S4
S0→S1
274 / 4.52
0.06
0.00
3.24
H→L+7 (70%)
H→L (88%)
H→L+1 (43%)
274 / 4.52 (CT)
322 / 3.34 (LE)
S0→S2
S0→S3
S0→S4
352 / 3.53
344 / 3.60
342 / 3.62
0.00
0.00
0.01
H-8→L (56%)
H-1→L (25%)
H→L+1 (65%)
H-5→L+1 (59%)
S0→S2
312 / 3.97
0.17
H→L+6 (66%)
H→L+1 (25%)
H-1→L (39%)
H→L+1 (33%)
DCM
S0→S3
S0→S4
278 / 4.46
0.09
0.00
H→L+6 (67%)
H→L (89%)
S0→S5
S0→S6
S0→S7
S0→S8
307 / 4.04
302 / 4.11
299 / 4.14
291 / 4.26
0.05
0.01
0.00
0.00
H→L+6 (69%)
H→L+5 (37%)
H-1→L+4 (45%)
273 / 4.55 (CT)
Results in Table 3 are consistent with experimental results,
showing a non-solvatochromic absorption. Irrespective of the
polarity of the solvent, the first transition (S0→S1) is a LE
transition with an almost equal weight of two localized excitations
on the donor fragment (H→L+1) and on the acceptor fragment (H-
1→L), see Figures 10 and S1, blue arrows.
H-9→L (34%)
H-1→L+2 (41%)
S0→S9
285 / 4.34 (CT)
280 / 4.43
0.00
0.00
H→L+14 (37%)
H→L+17 (23%)
S0→S10
H→ L+11 (57%)
Figure 10. Frontier Molecular Orbitals of TBP in gas phase (TDDFT level) and
the LE/CT transitions. Blue arrow indicates LE transition and red arrow indicates
CT transition.
Figure 11. Frontier Molecular Orbitals of TBP obtained with ZINDO calculations
showing the LE/CT transitions. Blue arrow indicates LE transition and red arrow
indicates CT transition.
The first CT state is found at higher energy, corresponding to S3
in gas phase and in cyclohexane and to S4 in toluene and DCM.
This CT state corresponds to the H→L transition, displacing
charge from the D to the A site (Figures 10 and 11, red arrows).
The vertical energy of the CT state is marginally affected by the
solvent polarity, as expected (indeed the negligible polarity of the
molecule in the ground state implies a negligible value of the
corresponding equilibrium reaction field). For such a large
To discuss ESA spectra we must address higher-energy excited
states, implying a very expensive calculation within TD-DFT,
which could not be undertaken. We therefore made resort to
semiempirical ZINDO/S calculations in gas phase, as
implemented in the Gaussian09 package.[59] In this way, we were
able to reach states with energies up to 7 eV (180 nm) above the
ground state, comparable to the energy of the excited states
addressed in transient spectroscopy (neither the computational
This article is protected by copyright. All rights reserved.

10.1002/cphc.201900703
ChemPhysChem
FULL PAPER
results nor the experimental ones suggest any ionization effect at
these energies). Computational results for highly excited states
should be taken with care, and cannot be directly compared with
experiment. Indeed, we just exploit them to validate and
generalize the hypothesis, suggested by experimental data, of the
existence of two sets of optically independent excited states.
In Table 4, we show ZINDO/S results for a few low-lying
transitions of TBP in gas phase. In comparison with experimental
data, the energy of the first allowed transition is underestimated
in gas phase by ∼0.3 eV. The lowest-energy LE and CT states
(as labelled in Table 4) have been identified by analyzing the
nature of the MOs most contributing to each transition (see Figure
11). ZINDO results are in line with TD-DFT results, ascribing a LE
character to the lowest-energy transition (S0→S1, calculated at
3.47 eV, i.e. 357 nm). The first transition with CT character is
calculated in ZINDO/S at 4.34 eV (285 nm) as the S0→S9
transition. Having identified the first LE and CT states as S1 and
S9, respectively, we calculated the transition dipole moments
(and more precisely their Cartesian components µti, i=x,y,z, see
Supporting Information, Table S1 and S2) from those two states
towards higher-energy states. We verified that the only significant
component of µt is the one aligned along the molecular axis
(assigned to the x direction). In Table 5, we have tabulated the
transition energies from the first LE state (S1) and from the first
CT state (S9) to higher states, selecting the ones with sizeable µtx
values. The MO’s taking part in these transitions have also been
checked to verify the nature of each transition.
Data in Table 5 demonstrate that the excited states that can be
reached from S1 (the lowest-energy LE state) are different from
those that can be reached from S9 (the lowest-energy CT state).
In particular, the states reachable upon photoexcitation from S1
have a predominant LE nature and those reachable upon
photoexcitation from S9 have a predominant CT nature. This
result confirms and generalizes the optical independence of CT
and LE states suggested by transient spectroscopy.
Conclusions
We have presented a thorough spectroscopic characterization of
a biphenyl-based push-pull chromophore, by means of UV-vis
absorption and emission spectroscopy, as well as pump-probe
transient spectroscopy. The substitution of the 2,2’ and 6,6’
positions of the biphenyl moiety imposes the orthogonality of the
two phenyl rings, decoupling the electron transfer from the donor
to the acceptor site from torsional/vibrational degrees of freedom.
In this way, solvation remains as the only relevant degree of
freedom coupled to the charge transfer. Experimental data,
supported by theoretical modeling through an essential-state
description, suggest a subtle interplay between a LE state and a
CT state. The CT state is sensitive to solvent polarity and is
characterized by a fairly long lifetime (~10 ns). Polar solvation
governs the relative energy of the two relaxed excited states, and
therefore the dynamics following photoexcitation.
Interestingly, transient absorption data suggest that transitions
are allowed only between states having the same (LE or CT)
nature. This interpretation is supported by quantum chemical
calculations addressing excited states up to very high energies:
two different manifolds of states can be recognized, having LE or
CT nature, not interconverted via optical excitation. However, a
spontaneous evolution from the lowest-energy LE state to the
lowest-energy CT state is allowed in polar solvents and is driven
by the solvent relaxation.
While other donor-acceptor systems are known to behave in a
similar way, with a LE/CT interplay revealed by dual fluorescence
and by transient-absorption data, here we investigated for the first
time an elongated donor-acceptor molecule where π-conjugated
bridges act as linkers and assure long-distance electron transfer.
The long-distance charge-separated state, moreover, is
surprisingly long-lived (both in polar solvents and in the solid
state), most probably as a result of the blocked conformation of
the central biphenyl core, which is substituted by sterically
hindering methyl groups.
Table 5. Transition dipole moments and energies from the lowest-energy LE
state (S1) to higher-energy LE states and from the lowest-energy CT state (S9)
to higher-energy CT states.
Lowest LE State
Higher LE States
µtx [a.u.]
0.3062
1.0317
0.8502
-3.0683
1.8033
3.0691
∆E / λ [eV / nm]
S4
0.27 / 4592
S12
S14
S22
S24
S26
1.23 / 1008
1.99 / 623
S32
S33
-0.9658
-0.6430
S1
2.30 / 539
2.64 / 470
S40
S41
S57
0.8957
-0.6838
0.3845
S60
S66
-0.3162
0.2124
3.02 / 410
S69
0.3696
µtx [a.u.]
-1.1938
0.2932
-0.5346
0.7922
-0.3681
0.2252
-0.3640
0.7169
0.1052
To the best of our knowledge, here we offer for the first time a
convincing demonstration of the presence of two sets of optically-
independent excited states, having either LE or CT character. The
two sets can only “communicate” thanks to solvation, which is the
only driving force for the charge-separation process.
Lowest CT State
Higher CT States
∆E / λ [eV / nm]
S13
S30
S31
S37
S52
S53
S55
S67
S71
0.39 / 3180
1.30 / 954
1.54 / 805
S9
Experimental Section
1.97 / 630
2.33 / 532
Synthesis
1,2-Bis(3,5-dimethylphenyl)hydrazine (2). Zinc powder (25.04 g, 0.393
mol) was added to a suspension of 3,5-dimethylnitrobenzene (1) (10.01 g,
66.2 mmol) in EtOH (40 mL) under argon. The mixture was refluxed until
This article is protected by copyright. All rights reserved.

10.1002/cphc.201900703
ChemPhysChem
FULL PAPER
a solution was obtained. Heating was removed and a solution of NaOH
(15.06 g, 0.376 mol) in water (50 mL) was added dropwise, in order to
maintain the reflux. After addition of about one third of NaOH, heating was
needed again, and after addition was complete, the mixture was refluxed
for 2 h. Zinc powder was then added again (10.08 g, 0.153 mol), and the
mixture was refluxed for another 2 h. The hot suspension was filtered
through Celite and washed with hot EtOH (60 mL) into a mixture of AcOH
(45 mL), water (105 mL) and sodium bisulfite (1.0 g). The solution was
cooled with a cold water bath, and the precipitate was filtered off and
recrystallized from heptane to give 3.501 g (44%) of 2.[36,37] 1H NMR (300
MHz, CDCl3): δ = 6.51 (s, 6 H), 5.48 (br s, 2 H), 2.26 (s, 12H).
CDCl3): δ = 7.66 (d, 2H, J = 8.7 Hz), 7.61 (d, 2H, J = 8.7 Hz), 7.38 (d, 2H,
J = 9.0 Hz), 7.36 (s, 2H), 7.32 (s, 2H), 6.59 (d, 2H, J = 9.0 Hz), 3.29 (t, 4
H, J = 7.7 Hz), 1.93 (s, 6H), 1.90 (s, 6H), 1.60 (m, 4 H), 1.34 (m, 12 H),
0.92 (t, 6H, J = 6.7 Hz); 13C NMR (75 MHz, CDCl3): δ = 148.0, 141.3, 138.7,
136.3, 135.5, 133.0, 132.2, 132.1, 131.0, 130.6, 128.6, 123.1, 120.7, 118.7,
111.3, 108.9, 94.4, 90.6, 87.4, 87.1, 51.1, 31.9, 27.3, 27.0, 22.8, 19.8, 19.7,
14.2; HRMS (ESI, CH2Cl2/CH3OH 90:10): m/z calcd for C45H51N2:
619.4052 ([M+H]+); found: 619.4047.
Preparation of Organic Nanoparticle Suspensions. 200 μL of a 1mM
THF solution of the compound was added to 9.800 mL of bidistilled water
under vigorous stirring, to achieve a final nominal concentration of 2×10-5
M. The suspension was kept under vigorous stirring for about 15 minutes.
2,2',6,6'-Tetramethylbenzidine
(3).
A
solution
of
1,2-bis(3,5-
dimethylphenyl)hydrazine (2) (3.502 g, 14.57 mmol) in 165 mL of HCl
(10%) was bubbled with argon for 20 min, and refluxed for 2 h. The mixture
was cooled to rt, and a solution of NaOH (30 g in 120 mL of water) was
added dropwise until pH ~ 10-11. The product was extracted with ether
(3x). The combined organic layers were washed with brine until neutral pH,
dried with Na2SO4, filtered and evaporated under reduced pressure. The
crude product was purified by column chromatography on silica gel
(EtOAc/heptane 1:2), to give 2.316 g (66%) of 3.[36,37] 1H NMR (300 MHz,
CDCl3): δ = 6.49 (s, 4 H), 3.53 (br s, 4 H), 1.83 (s, 12 H).
Absorption and Fluorescence Spectroscopy. Absorption and emission
measurements were performed at room temperature on freshly prepared
solutions. For fluorescence measurements, the optical density was kept ≤
0.1 (roughly corresponding to a 10-6 M) concentration, in order to minimize
inner-filter effects. All solvents were spectroscopic or HPLC grade and
were used as received. Absorption measurements were performed with a
Perkin-Elmer Lambda 650 double-beam UV/Vis spectrophotometer. To
estimate the molar extinction coefficients in chloroform, the Lambert-Beer
law was verified in the concentration range 10-4 - 10-6 M. Fully corrected
steady-state emission and excitation spectra were obtained with a
Fluoromax-3 Horiba Jobin-Yvon fluorometer (detection at 90 degrees)
equipped with a Xenon lamp as the excitation source. A diluted solution (c
~ 10-6 M) of fluorescein in NaOH(aq) 0.1 M (φ = 0.9, λexc = 460 nm) was
used as reference standard for the determination of the fluorescence
quantum yields. In order to integrate the emission spectra for evaluating
the quantum yield, even in the cases where the spectra were not “complete”
because we had to stop measuring at 670 nm, the spectra were fitted (on
the wavenumber scale) with Gaussian lines. The emission spectrum of the
powders was collected placing the sample on a front-face vertical sample
holder, with an inclination of the plane of 30 degrees with respect to the
propagation direction of the incoming light; an excitation and an emission
polarisers were placed at the magic angle, in order to decrease the
scattered light on the detector. Fluorescence decays were measured with
the TCSPC (Time-Correlated Single-Photon Counting) technique, using
the Fluoromax-3 instrument equipped with a Horiba Fluoro-Hub analyzer.
A pulsed NanoLed at λ = 340 nm was used as the excitation source. The
prompt signal was acquired by measuring the scattering of an aqueous
suspension obtained by adding two drops of Ludox HS-40 (Sigma-Aldrich)
to approximately 4 mL of water. Lifetimes (τ) values were extracted by
reconvolution analysis of the decay profiles and the goodness of the fit was
evaluated with the chi-square test (results were retained if χ2 < 1.2).
4,4'-Diiodo-2,2',6,6'-tetramethylbiphenyl (4). A solution of NaNO2 (0.875 g,
12.68 mmol) in water (3 mL) was added to a suspension of 1,2-bis(3,5-
dimethylphenyl)hydrazine (3) (1.386 g, 5.767 mmol) in water (22 mL) and
sulfuric acid (97%, 12 mL) cooled in an ice bath. The mixture was stirred
at ~ 5 °C for 30 min, and a solution of I2 (4.142 g, 16.32 mmol) and KI
(4.579 g, 27.58 mmol) in water (7.5 mL) was added. The mixture was
stirred for another 20 min at ~ 5 °C. Water (30 mL) and CH2Cl2 (75 mL)
were added, and the solution was stirred at rt overnight. Sodium thiosulfate
(4.89 g) was added, and the mixture was stirred for 10 min. The precipitate
was filtered off and washed with CH2Cl2. The two layers were separated
and the aqueous layer was extracted with CH2Cl2 (2 x 45 mL). The
combined organic layers were then washed again (3x) with sodium
thiosulfate and brine, and finally dried with MgSO4. The solvent was
removed under reduced pressure, and the crude product was purified by
column chromatography on silica gel (heptane), to give 1.805 g (68%) of
4.[37,38] 1H NMR (300 MHz, CDCl3): δ = 7.49 (s, 4 H), 1.84 (s, 12 H).
4-[(4'-Iodo-2,2’,6,6’-tetramethyl[1,1'-biphenyl]-4-yl)ethynyl]-N,N-dihexyl-
benzenamine (6). A solution of 4,4'-diiodo-2,2',6,6'-tetramethylbiphenyl (4)
(0.8074 g, 1.749 mmol) and N,N-dihexyl-4-[2-(trimethylsilyl)ethynyl]-
benzenamine[39] (5) (0.1795 g, 0.503 mmol) in toluene (5 mL) and Et3N
(1.5 ml) was bubbled with argon for 20 min. CuI (4.8 mg, 0.0175 mmol),
Pd(PPh3)2Cl2 (13.7 mg, 0.0175 mmol) and TBAF (1 M in THF, 1 mL) were
successively added, and the mixture was stirred at 40 °C overnight.
Solvents were evaporated and the residue was purified by column
chromatography on silica gel, eluting with CH2Cl2/heptane (gradient from
5:95 to 20:80), to give 0.1976 g (64%) of 6. 1H NMR (300 MHz, CDCl3): δ
= 7.50 (s, 2H), 7.37 (d, 2H, J = 8.9 Hz), 6.58 (d, 2H, J = 8.9 Hz), 7.29 (s,
2H), 3.28 (t, 4 H, J = 7.6 Hz), 1.88 (s, 6H), 1.86 (s, 6H), 1.53 (m, 4 H), 1.33
(m, 12 H), 0.91 (t, 6H, J = 6.8 Hz); 13C NMR (75 MHz, CDCl3): δ = 147.8,
139.2, 138.4, 138.0, 136.3, 135.4, 132.8, 130.4, 122.9, 111.2, 108.8, 92.6,
90.4, 87.0, 51.0, 31.7, 27.2, 26.8, 22.7, 19.6, 19.3, 14.0; HRMS (ESI,
CH2Cl2/CH3OH 90:10): m/z calcd for C36H46NI: 619.2675 (M+.); found:
619.2671.
Transient Absorption Spectroscopy. The apparatus used for the
transient absorption spectroscopy measurements is based on
a
Ti:sapphire regenerative amplifier (BMI Alpha 1000) system pumped by a
Ti:sapphire oscillator (Spectra Physics Tsunami). The system produces
100 fs pulses at 785 nm, 1 kHz repetition rate and average power of 450
mW. Excitation pulses at 330 nm have been obtained by pumping an
optical parametric amplifier (TOPAS, light conversion) by a portion of the
fundamental 785 nm. The output of the TOPAS has been first frequency
doubled and then mixed with residual 800 nm radiation to generate the
desired pump wavelength. The probe pulse was generated by focusing a
small portion of the fundamental laser output radiation on a 2 mm thick
sapphire window. The pump beam polarization has been set to magic
angle with respect to the probe beam by rotating a /2 plate. Excitation
powers were on the order of 100 nJ. Pump-probe delays were introduced
by sending the probe beam through a motorized stage. Multichannel
detection was achieved by sending the white light continuum after passing
through the sample to a flat field monochromator coupled to a home-made
CCD detector. Measurements were carried out in a 2 mm thick quartz cell.
To avoid sample photodegradation and multiple photon excitation, the
solution was refreshed using magnetic stirring. Data were analysed
4-[[4’-[[4-(Dihexylamino)phenyl]ethynyl]-2,2’,6,6’-tetramethyl[1,1'-
biphenyl]-4-yl]ethynyl]benzonitrile (TBP). A solution of iodo compound 6
(87.3 mg, 0.141 mmol) and 4-[2-(trimethylsilyl)ethynyl]benzonitrile (7)[40]
(35.3 mg, 0.177 mmol) in toluene (2 mL) and Et3N (0.5 mL) was bubbled
with argon for 20 min. CuI (2 mg, 0.01 mmol), Pd(PPh3)2Cl2 (7.4 mg, 0.01
mmol) and TBAF (1 M in THF, 0.5 mL) were successively added, and the
mixture was stirred at 40 °C overnight. Solvents were evaporated and the
residue was purified by column chromatography on silica gel, eluting with
CH2Cl2/heptane (1:4), to give 0.0627 g (72%) of TBP. 1H NMR (300 MHz,
This article is protected by copyright. All rights reserved.

10.1002/cphc.201900703
ChemPhysChem
FULL PAPER
through singular value decomposition and global analysis performed using
the software Glotaran.[60]
[13] K. A. Zachariasse, M. Grobys, T. von der Haar, A. Hebecker, Y. V. Il’ichev,
Y.-B. Jiang, O. Morawski, W. Kühnle, J. Photochem. Photobiol. A Chem.
1996, 102, 59–70.
[14] K. A. Zachariasse, S. I. Druzhinin, W. Bosch, R. Machinek, J. Am. Chem.
Soc. 2004, 126, 1705–1715.
Computational details. All the quantum chemical calculations for TBP in
gas phase and solvents have been carried out using Density Functional
Theory (DFT) and Time Dependent DFT (TDDFT) as implemented in the
Gaussian 09 package.[59] The 6-31+g(d) basis set was adopted for all
calculations and a long-range Coulomb attenuated functional CAM-
B3LYP[61] was used both for geometry optimization and for the calculation
of vertical excited states. Calculation were run in gas phase and in
dichloromethane (DCM), cyclohexane and toluene, in the polarizable
continuum model (PCM)[62] approach (non-equilibrium, state-specific) as
implemented in Gaussian 09 package. The optimized ground state (S0)
geometries have been confirmed to be in local minimum energy structures
by analyzing the vibrational frequencies in the ground state. To look into
the excited state geometry in gas phase for TBP, optimization of the first
charge transfer (CT) state and the first local excited state (LE) were carried
out using the above-mentioned level of theory. We tried to look into the
optimized geometry of the excited state even in solvent, but due to
computational limitations and heavy nature of the calculations, it could not
be achieved. Since we are interested to the calculation of high-energy
excited states, a task that becomes too expensive in TDDFT, we have
resorted to the semi-empirical ZINDO calculations, as implemented in G09,
to look into the nature of the transitions.
[15] S. Murali, V. Kharlanov, W. Rettig, A. I. Tolmachev, A. V. Kropachev, J.
Phys. Chem. A 2005, 109, 6420–6429.
[16] N. Nakashima, M. Murakawa, N. Mataga, Bull. Chem. Soc. Jpn. 1976, 49,
854–858.
[17] S. A. Kovalenko, J. L. Pérez Lustres, N. P. Ernsting, W. Rettig, J. Phys.
Chem. A 2003, 107, 10228–10232.
[18] M. Jurczok, P. Plaza, M. M. Martin, Y. H. Meyer, W. Rettig, Chem. Phys.
2000, 253, 339–349.
[19] M. Jurczok, P. Plaza, W. Rettig, M. M. Martin, Chem. Phys. 2000, 256,
137–148.
[20] N. Mataga, H. Yao, T. Okada, W. Rettig, J. Phys. Chem. 1989, 93, 3383–
3386.
[21] S. Hashimoto, A. Yabushita, T. Kobayashi, K. Okamura, I. Iwakura, Chem.
Phys. 2018, 512, 128–134.
[22] W. Rettig, M. Maus, R. Lapouyade, Berichte der Bunsengesellschaft für
Phys. Chemie 1996, 100, 2091–2096.
[23] M. Maus, W. Rettig, G. Jonusauskas, R. Lapouyade, C. Rullière, J. Phys.
Chem. A 1998, 102, 7393–7405.
[24] M. Maus, W. Rettig, D. Bonafoux, R. Lapouyade, J. Phys. Chem. A 1999,
103, 3388–3401.
[25] M. Maus, W. Rettig, J. Phys. Chem. A 2002, 106, 2104–2111.
[26] J. Dobkowski, M. Kijak, I. V. Sazanovich, J. Waluk, J. Phys. Chem. B 2015,
119, 7294–7307.
Acknowledgements
[27] D. Vonlanthen, J. Rotzler, M. Neuburger, M. Mayor, Eur. J. Org. Chem.
2010, 2010, 120–133.
The authors gratefully acknowledge financial support from the
Italian Ministero dell’Istruzione, dell’Università e della Ricerca
(MIUR): grant RBFR10Y5VW (FIRB Futuro in Ricerca 2010) and
grant “Dipartimenti di Eccellenza” (DM 11/05/2017 n. 262). This
project has received funding from the European Union's Horizon
2020 research and innovation programme under grant agreement
No 812872 (TADFlife). This research benefits from the HPC (High
Performance Computing) facility of the University of Parma, Italy.
We thank Cédric Rouxel and Cynthia Vandermeulen for
assistance in the synthesis of TBP.
[28] W. Weigel, W. Rettig, M. Dekhtyar, C. Modrakowski, M. Beinhoff, A. D.
Schlüter, J. Phys. Chem. A 2003, 107, 5941–5947.
[29] S. Sasaki, G. P. C. Drummen, G. Konishi, J. Mater. Chem. C 2016, 4,
2731–2743.
[30] O. Mongin, L. Porrès, M. Charlot, C. Katan, M. Blanchard-Desce, Chem.
Eur. J. 2007, 13, 1481–1498.
[31] F. Bureš, RSC Adv. 2014, 4, 58826–58851.
[32] L. Donato, A. Mourot, C. M. Davenport, C. Herbivo, D. Warther, J. Léonard,
F. Bolze, J.-F. Nicoud, R. H. Kramer, M. Goeldner, et al., Angew. Chem.
Int. Ed. 2012, 51, 1840–1843.
[33] J. Cody, C. J. Fahrni, Tetrahedron 2004, 60, 11099–11107.
[34] M. Gantenbein, M. Hellstern, L. Le Pleux, M. Neuburger, M. Mayor, Chem.
Mater. 2015, 27, 1772–1779.
Keywords: electron transfer • dual fluorescence • solvation
dynamics • time-resolved spectroscopy • theoretical modelling
[35] O. V. Mikhnenko, R. Ruiter, P. W. M. Blom, M. A. Loi, Phys. Rev. Lett. 2012,
108, 137401.
[36] A. Helms, D. Heiler, G. McLendon, J. Am. Chem. Soc. 1992, 114, 6227–
6238.
[1] M. Mamedov, Govindjee, V. Nadtochenko, A. Semenov, Photosynth. Res.
2015, 125, 51–63.
[37] D. Vonlanthen, J. Rotzler, M. Neuburger, M. Mayor, Eur. J. Org. Chem.
2010, 2010, 120–133.
[2] H. Klauk, Organic Electronicsꢀ: Materials, Manufacturing and Applications,
Wiley-VCH, 2006.
[38] S. Suzuki, R. Sugimura, M. Kozaki, K. Keyaki, K. Nozaki, N. Ikeda, K.
Akiyama, K. Okada, J. Am. Chem. Soc. 2009, 131, 10374–10375.
[39] C. Rouxel, M. Charlot, O. Mongin, T. R. Krishna, A.-M. Caminade, J.-P.
Majoral, M. Blanchard-Desce, Chem. Eur. J. 2012, 18, 16450–16462.
[40] S. P. McIlroy, E. Cló, L. Nikolajsen, P. K. Frederiksen, C. B. Nielsen, K. V.
Mikkelsen, K. V. Gothelf, P. R. Ogilby, J. Org. Chem. 2005, 70, 1134–1146.
[41] J. Saltiel, V. K. R. Kumar, J. Phys. Chem. A 2012, 116, 10548–10558.
[42] M. Krämer, U. H. F. Bunz, A. Dreuw, J. Phys. Chem. A 2017, 121, 946–
953.
[3] L. Venkataraman, J. E. Klare, C. Nuckolls, M. S. Hybertsen, M. L.
Steigerwald, Nature 2006, 442, 904–907.
[4] A. P. Kulkarni, C. J. Tonzola, A. Babel, S. A. Jenekhe, Chem. Mater. 2004,
16, 4556–4573.
[5] S. Günes, H. Neugebauer, N. S. Sariciftci, Chem. Rev. 2007, 107, 1324–
1338.
[6] B. Kippelen, J.-L. Brédas, Energy Environ. Sci. 2009, 2, 251.
[7] V. May, O. Kꢀhn, Charge and Energy Transfer Dynamics in Molecular
Systems, Wiley-VCH, 2011.
[43] M. Flock, L. Bosse, D. Kaiser, B. Engels, I. Fischer, Phys. Chem. Chem.
Phys. 2019, 21, 13157–13164.
[8] T. Kumpulainen, B. Lang, A. Rosspeintner, E. Vauthey, Chem. Rev. 2017,
117, 10826–10939.
[44]P. Meystre, M. Sargent, in Elem. Quantum Opt., Springer Berlin Heidelberg,
Berlin, Heidelberg, 2007, pp. 51–91.
[9] C. A. Rumble, M. Maroncelli, J. Chem. Phys. 2018, 148, 193801.
[10] K. A. Zachariasse, S. I. Druzhinin, O. Morawski, B. Kozankiewicz, J. Phys.
Chem. A 2018, 122, 6985–6996.
[45] C. Sissa, V. Calabrese, M. Cavazzini, L. Grisanti, F. Terenziani, S. Quici,
A. Painelli, Chem. Eur. J. 2013, 19, 924–935.
[11] K. K. Mentel, A. Serra, P. E. Abreu, L. G. Arnaut, Nat. Commun. 2018, 9,
2903.
[46] B. Boldrini, E. Cavalli, A. Painelli, F. Terenziani, J. Phys. Chem. A 2002,
106, 6286–6294.
[12] W. Schuddeboom, S. A. Jonker, J. M. Warman, U. Leinhos, W. Kuehnle, K.
A. Zachariasse, J. Phys. Chem. 1992, 96, 10809–10819.
This article is protected by copyright. All rights reserved.

10.1002/cphc.201900703
ChemPhysChem
FULL PAPER
[47] F. Terenziani, A. Painelli, Phys. Chem. Chem. Phys. 2015, 17, 13074–
13081.
[48] E. R. Henry, Biophys. J. 1997, 72, 652–673.
[49] E. R. Henry, J. Hofrichter, 1992, pp. 129–192.
[50] I. H. M. van Stokkum, D. S. Larsen, R. van Grondelle, Biochim. Biophys.
Acta - Bioenerg. 2004, 1657, 82–104.
[51] E. R. Middelhoek, H. Zhang, J. W. Verhoeven, M. Glasbeek, Chem. Phys.
1996, 211, 489–497.
[52] M. A. Kahlow, T. J. Kang, P. F. Barbara, J. Phys. Chem. 1987, 91, 6452–
6455.
[53] G. C. Walker, E. Aakesson, A. E. Johnson, N. E. Levinger, P. F. Barbara,
J. Phys. Chem. 1992, 96, 3728–3736.
[54] F. Terenziani, A. Painelli, Chem. Phys. 2003, 295, 35–46.
[55] R. S. Fee, M. Maroncelli, Chem. Phys. 1994, 183, 235–247.
[56] S. Mukamel, Principles of Nonlinear Optical Spectroscopy, Oxford
University Press, New York, 1995.
[57] R. Improta, V. Barone, F. Santoro, Angew. Chem. Int. Ed. 2007, 46, 405–
408.
[58] D. Jacquemin, A. Planchat, C. Adamo, B. Mennucci, J. Chem. Theory
Comput. 2012, 8, 2359–2372.
[59] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J.
R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G. A. Petersson, et
al., Gaussian 09, Revis. D.01, Gaussian, Inc., Wallingford CT 2009.
[60] J. J. Snellenburg, S. P. Laptenok, R. Seger, K. M. Mullen, I. H. M. van
Stokkum, J. Stat. Softw. 2012, 49, DOI 10.18637/jss.v049.i03.
[61] T. Yanai, D. P. Tew, N. C. Handy, Chem. Phys. Lett. 2004, 393, 51–57.
[62] J. Tomasi, B. Mennucci, R. Cammi, Chem. Rev. 2005, 105, 2999–3094.
This article is protected by copyright. All rights reserved.

10.1002/cphc.201900703
ChemPhysChem
FULL PAPER
Entry for the Table of Contents
Layout 1:
FULL PAPER
The photoinduced, solvent-driven
intramolecular electron transfer is
followed in real time, demonstrating
the existence of two mutually-
orthogonal sets of locally-excited and
charge-transfer states.
Elisa Campioli, Somananda Sanyal,
Agnese Marcelli, Mariangela Di Donato,
Mireille Blanchard-Desce, Olivier
Mongin, Anna Painelli, Francesca
Terenziani*
Page No. – Page No.
LE
Addressing charge-transfer and
locally-excited states in a twisted
biphenyl push-pull chromophore
CT
This article is protected by copyright. All rights reserved.