Thiophenylazobenzene: An Alternative Photoisomerization Controlled by Lone-Pair???π Interaction

Article abstract of DOI:10.1002/anie.201909739

Azoheteroarene photoswitches have attracted attention due to their unique properties. We present the stationary photochromism and ultrafast photoisomerization mechanism of thiophenylazobenzene (TphAB). It demonstrates impressive fatigue resistance and photoisomerization efficiency, and shows favorably separated (E)- and (Z)-isomer absorption bands, allowing for highly selective photoconversion. The (Z)-isomer of TphAB adopts an unusual orthogonal geometry where the thiophenyl group is perfectly perpendicular to the phenyl group. This geometry is stabilized by a rare lone-pair???π interaction between the S atom and the phenyl group. The photoisomerization of TphAB occurs on the sub-ps to ps timescale and is governed by this interaction. Therefore, the adoption and disruption of the orthogonal geometry requires significant movement along the inversion reaction coordinates (CNN and NNC angles). Our results establish TphAB as an excellent photoswitch with versatile properties that expand the application possibilities of AB derivatives.

Full text of DOI:10.1002/anie.201909739

A Journal of the Gesellschaft Deutscher Chemiker
Angewandte
Chemie
International Edition
www.angewandte.org
Accepted Article
Title: Thiophenylazobenzene – an alternative photoisomerization
controlled by lone pair∙∙∙π interaction
Authors: Chavdar Slavov, Chong Yang, Andreas H. Heindl, Hermann
A. Wegner, Andreas Dreuw, and Josef Wachtveitl
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: Angew. Chem. Int. Ed. 10.1002/anie.201909739
Angew. Chem. 10.1002/ange.201909739
Link to VoR: http://dx.doi.org/10.1002/anie.201909739
http://dx.doi.org/10.1002/ange.201909739

10.1002/anie.201909739
Angewandte Chemie International Edition
RESEARCH ARTICLE
Thiophenylazobenzene – an alternative photoisomerization con-
trolled by lone pair∙∙∙π interaction
Chavdar Slavov, #[a] Chong Yang, #[b,d] Andreas H. Heindl,[c] Hermann A. Wegner,[c] Andreas Dreuw,*[b] and Josef
Wachtveitl*[a]
Abstract: Azoheteroarene photoswitches have attracted attention
due to their unique properties. We present here the stationary
photochromism and the ultrafast photoisomerization mechanism of
in substantial geometrical and spectral changes.[4] Spectrally,
ABs are characterized by a strong ππ* transition band (S2),
typically located in the 300-350 nm range for the (E)-isomer and
the 250-290 nm range for the (Z)-isomer. They also show a
weakly allowed nπ* transition band (S1) in the 400-450 nm range
for both isomers. The photoisomerization quantum yield (QY) of
thiophenylazobenzene (TphAB)
– an alternative sulfur-based
azoheteroarene design. TphAB demonstrates impressive fatigue
resistance and photoisomerization efficiency, and shows favorably
separated (E)- and (Z)-isomer absorption bands, allowing for highly
selective photoconversion. The (Z)-isomer of TphAB adopts an
unusual orthogonal geometry where the plane of the thiophene is
perfectly perpendicular to that of the benzene. This geometry is
stabilized by a rare lone pair∙∙∙π interaction between the S-atom and
the benzene. The photoisomerization of TphAB occurs on the sub-ps
to ps timescale and is governed by the lone pair∙∙∙π interaction.
Therefore, the adoption and disruption of the orthogonal geometry
requires significant movement along the inversion reaction
coordinates (CNN and NNC angles). Our results establish TphAB as
an excellent photoswitch with versatile properties that expand the
application possibilities of AB derivatives.
ABs is excitation wavelength dependent: QYE(ππ*) Z ~13%,
→
QYE(nπ*) Z ~29%, QYZ(ππ*) E ~24-36%, QYZ(nπ*) E ~ 47-51%.[4-5]
→
→
→
The E → Z photoisomerization of ABs proceeds on the sub-ps to
ps timescale after both nπ* and ππ* excitation. After S2
excitation, ultrafast (~100 fs) relaxation to S1 state is observed.[6]
The decay of the S1 state is described by two lifetimes (~400 fs,
~2 ps), originally attributed to direct and diffusive motion from
the Franck–Condon region to the conical intersection with the
ground state.[6b, 6c, 7] Relaxation lifetimes after S1 excitation are
slightly different, indicating that the relaxation pathways are
differently populated.[6b, 6d] Recently, only the longer (~2 ps) S1
decay component was associated with E → Z isomerization,
while the shorter one (~400 fs) was assigned to a nonreactive
relaxation pathway from a region on the S1 potential energy
surface (PES) that is accessible only after S2 excitation.[6f] While
this result designates an important tendency, it does not yet
explain the presence of a similar (~400 fs) lifetime component
after S1 excitation. Also, additional pathways may be involved
making the sharp distinction between the timescales of the
relaxation pathways in the ultrafast data hard.[8] In fact, lifetime
distribution analysis has shown that the two S1 relaxation
lifetimes discussed above belong to a broad ~700 fs lifetime
Introduction
The operation of molecular photoswitches is based on reversible
transformation of the switching molecule between states of
different physicochemical properties (e.g. geometrical structure,
dipole moment, absorption spectrum, redox potential, etc).[1] The
utilization of light as a trigger allows easy manipulation which,
combined with the instantaneous property change of the
molecule, makes photoswitches extremely attractive for
application in chemical science and technology.[2]
Azobenzenes (AB) are a prominent group of widely utilized
photoswitches. They are highly fatigue resistant and relatively
easy to synthesize.[3] The photochromism of ABs is based on the
E ⇄ Z photoisomerization of the central N═N bond, which results
distribution.[6e] This indicates
a significant overlap of the
timescales of the reactive and the nonreactive pathways.[8] The
Z → E photoisomerization is ultrafast with a main lifetime of
~150 fs and a minor contribution of ~1 ps lifetime.[6b, 6d, 7a] In both
isomerization directions, relaxation to the ground state is
followed by vibrational cooling (10–20 ps lifetime).[6b, 6e, 6f, 7a, 9]
The photoisomerization mechanism of ABs has been analyzed
extensively via quantum chemical calculations.[4] In latest works,
a
consensus arises that inversion-assisted rotation is the
[a]
[b]
[b]
Dr. C. Slavov, Prof. J. Wachtveitl
dominant photoisomerization mechanism particularly in
condensed phase.[6f, 8, 10] The lower isomerization QY after ππ*
(S2) excitation was attributed to
Institute of Physical and Theoretical Chemistry
Goethe University, Frankfurt am Main, Germany
wveitl@theochem.uni-frankfurt.de
a nonreactive internal
Dr. C. Yang, Prof. A. Dreuw
conversion channel from the S1 to the ground state accessible
only after S2 → S1 relaxation[6f, 10f] or to a nonreactive channel
Interdisciplinary Center for Scientific Computing (IWR)
University of Heidelberg, Heidelberg, Germany
andreas.dreuw@iwr.uni-heidelberg.de
A. H. Heindl, Prof. H. A. Wegner
due to crossing of the S2 and S3 potential energy surfaces[11]
.
In contrast to the conventional, well-studied ABs, the
photochemical properties of azoheteroarenes remain largely
unexplored. Recent reports reveal that azoheteroarenes, like
arylazopyrazoles and arylazopyrroles, can adopt different
ground state conformations stabilizing or destabilizing the (Z)-
isomer, which allows for remarkable tunability of the thermal
relaxation rate – from seconds to ~1000 days.[12] In other
Institute of Organic Chemistry
Center for Materials Research (LaMa)
Justus Liebig University, Giessen, Germany
# These authors contributed equally.
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/anie.201909739
Angewandte Chemie International Edition
RESEARCH ARTICLE
azoheteroarenes impressively high thermal relaxation rates in
the microsecond[13] and even in the nanosecond[14] timescale
have been achieved. Furthermore, due to distortions because of
steric effects, the spectral properties of azoheteroarenes are
altered. This often results in favorable separation between the
(E)- and the (Z)-isomer bands, permitting >98% conversion in
12b, 12d]
each direction.[12a,
Evidently, choice, position and
orientation of the heteroaryl and its substituents represent a new
tuning dimension for ABs. Moreover, the presence of
heteroatoms permits new functional designs previously
unavailable in conventional ABs. Therefore, azoheteroarenes
offer an untapped potential for further optimization and
expansion of the capabilities of AB photoswitches.
In this work, we explore a different azoheteroarene design,
where one of the benzenes of conventional AB is substituted by
a thiophene. We present its synthesis, along with the theoretical
and experimental investigation of the photoisomerization of this
thiophenylazobenzene (TphAB) photoswitch.
Figure 1. (A) Photostationary state PSS365 – (E)-isomer absorbance difference
spectra; (B) Absorption spectra of the (E)-isomer and the PSS365 (see Fig. S5
for the pure (Z)-isomer spectrum); (C) TphAB isomerization; (D) Eyring plot for
the thermal Z → E relaxation (see also Table S1 for the used parameters).
Results and Discussion
Synthesis
The investigated TphAB 2 was synthesized based on the
coupling of aryldiazonium salts with aryllithium compounds[15]
(Scheme 1). Herein, thiophene (1) was readily lithiated at the 2-
The E → Z conversion under 365 nm irradiation is extremely
efficient, resulting in less than 3% (E)-isomer in the
photostationary state PSS365 (Fig. S5). This unusually high
photoconversion level is due to the favorable separation of the
(E)-isomer ππ* band from the (Z)-isomer absorption band in
TphAB. The (Z)-isomer present in the PSS285 is also very low
(~13%). The QY determination (see SI and Ref.[16]) for the two
photoisomerization reactions of TphAB after ππ* excitation
reveals some of the highest ever reported QYs for an AB system,
position,
and was
subsequently
added to
a
4-
methylphenyldiazonium tetrafluoroborate suspension at low
temperature to obtain TphAB 2 in 67% yield (see Supporting
Information (SI)).
S
S
N2BF4
THF
-78 °C to rt
67%
Li
n-BuLi
N
S
N
THF
0 °C
1a
1
2
with impressive QYE
of ~44% and QYZ
of ~65%.
E
Z
(ππ*) →
(nπ*) →
These QYs are even significantly higher than the ones typically
reported for nπ* excitation of the conventional ABs[4]. TphAB
shows very high fatigue resistance to repeated photoswitching.
After 50 photocycles equivalent to ~9 h of high intensity light
exposure at most ~3% degradation is detected (Fig. S1).
The Z → E thermal relaxation of TphAB is considerably shorter
than that of AB with half-life of ~120 min (20 °C). The
temperature dependence of the thermal relaxation rate was
determined at four different temperatures between 10 °C and
40 °C to obtain the thermodynamic parameters for the
corresponding transition state (Fig. 1D).
Scheme 1. Synthesis of thiophenylazobenzene 2.
Photochromic properties
All spectroscopic experiments with TphAB were performed in
acetonitrile. The absorption spectrum of the thermodynamically
stable (E)-isomer of TphAB shows a dominant 365 nm ππ*
absorption band and a weaker ~450 nm nπ* band (Fig. 1 A and
B). Irradiation of the (E)-isomer with 365 nm leads to E → Z
photoisomerization of the central N═N bond (Fig. 1C). The (Z)-
isomer is characterized by two main absorption bands: i) a
~285 nm ππ* absorption band, which has about half of the
intensity of the (E)-isomer ππ* band (365 nm); and ii) a ~450 nm
nπ* band, which is only about as intense as the one of the (E)-
isomer. Due to the identical spectral location and intensity of the
nπ* bands of the (E)- and the (Z)-isomers, the Z ⇄ E
photoconversion of TphAB is achieved by irradiation in the
corresponding ππ* absorption bands.
Quantum chemical calculations
Theoretical calculations using the second-order algebraic
diagrammatic construction scheme for excitation energies
(ADC(2))[17] and linear-response time-dependent density
functional theory (TDDFT)[18] (see SI) were performed to gain
insight into the molecular properties of the studied TphAB.
Ground state geometry optimization of TphAB yielded two stable
geometries for both the (E)- and the (Z)-isomer in dependence
of the rotational orientation of the thiophene ring with respect to
the azophenyl group – TphAB-1 and TphAB-2 (see Fig. 2 left).
This article is protected by copyright. All rights reserved.

10.1002/anie.201909739
Angewandte Chemie International Edition
RESEARCH ARTICLE
Similar to AB, the (E)-isomers of TphAB are planar. However,
the (Z)-isomers of TphAB adopt a rather unusual geometry,
where the thiophene ring lies in plane with the CNNC moiety,
while the benzene ring is either perfectly orthogonal to this plane
(TphAB-1) or slightly twisted away from it (TphAB-2) (Fig. 2, left).
This is in stark contrast to AB,[4] and the related azothiophene,[19]
where both rings are twisted away from the CNNC plane. Similar
unusual geometries of the (Z)-isomer have been reported for
some nitrogen-based azoheteroarenes,[12a, 12b, 12d] but not for the
conventional AB. Interestingly, it was found that in these
compounds the orthogonal geometry is disfavored when a bulky
substituent is present in ortho position to the CNNC group.[12d]
Alternatively, the orthogonal geometry is adopted when an H-
atom is present in ortho position due to favorable C–H∙∙∙π
interaction.[12d] Remarkably, TphAB adopts an orthogonal
geometry only when the S-atom in ortho position of the
thiophene faces the benzene ring (TphAB-1, Fig. 2 left), while
the twisted geometry is realized when the S-atom is facing away
from the benzene (TphAB-2¸ Fig. 2 left). The differences in the
adopted conformations by TphAB and by nitrogen-based
azoheteroarenes[12d] can be explained by differences in the
attractive and repulsive interactions. In TphAB-1, the orthogonal
structure of the (Z)-isomer is stabilized by a favorable interaction
between the lone pair of the S-atom and the π-system of the
benzene ring (lone pair∙∙∙π interaction[20]). Based on the results
from azoheteroarenes,[12d] one would expect the adoption of an
orthogonal structure also in TphAB-2 (Z)-isomer, where the
ortho H-atom on the thiophene ring is involved in C–H∙∙∙π
interaction. However, in this configuration (TphAB-2), with the S-
atom facing away from the benzene ring, the lone pair of the S-
atom comes into a close contact with the lone pair of the N-atom
from the azo group, which results in a repulsive interaction and
the ensuing twisting of the thiophene ring.
configuration is dominant (~97.2%). The excitation energies for
all the isomers obtained from the theoretical calculations
(Fig. S3 & S4, at the ADC(2) and the BHLYP level of theory) and
the Boltzmann distributions were used to simulate the theoretical
absorption spectra of the isomer mixtures (Fig. 2, right, Fig. S5).
These calculated spectra reproduce the experimental spectra of
the (E)- and (Z)-isomers of TphAB in acetonitrile (Fig. 1B) very
well given an energy blue shift of ~0.61 eV. For the (E)-isomer, a
strong transition is present at ~320 nm, which has ππ* character
(see Fig. S2 with the attachment/detachment densities). The nπ*
transition is only weakly allowed in the (E)-isomer and therefore
the intensity in the >400 nm range is small. The calculated
spectrum of the (Z)-isomer of TphAB shows
a strong
contribution in the 250-300 nm region due to two ππ* transitions
(Fig. 2, right). The attachment/detachment densities for the
dominant TphAB-1 configuration (Fig. S2), show
a very
interesting character for these ππ* transitions. The lower energy
one is located entirely on the azothiophenyl group, while the
higher energy one shows charge transfer character from the
benzene to the thiophene. Concerning the nπ* transition of the
(Z)-isomer, it appears that in the orthogonal geometry the
transition is very weak and has negligible contribution to the
absorption spectrum (Fig. 2, right) in contrast to AB. This result
is in agreement with the experiential absorption spectrum
(Fig. 1B) where the intensity in the >400 nm region is very
similar for both the (E)- and the (Z)-isomer. Noteworthy, in the
twisted geometry of the (Z)-isomer, the nπ* transition is stronger
(Fig. S3 and S4) due to the smaller angle between the plane of
the nonbonding orbitals of the azo group and the azothiophenyl
plane. Nevertheless, the contribution of the twisted geometry to
the experimental spectrum is negligible as this configuration is
essentially not present at room temperature (see above).
Ultrafast E → Z photoisomerization
The E → Z photoisomerization of TphAB after 355 nm excitation
in the ππ* absorption band of the (E)-isomer was studied by
ultrafast transient absorption spectroscopy (see SI and ref.[16] for
a description of the pump-probe set-up). The ultrafast dynamics
(Fig. 3A and Fig. S6A) resembles that of the conventional AB[6b,
6e]. However, the main excited state absorption (ESA) bands
appear shifted to the red by ~30 nm (to 510 nm and 420 nm). On
the sub-250 fs timescale, the ESA located at ~510 nm decays
concomitantly with the rise of the ESA at 420 nm (Fig. S6). In
turn, the decay of the 420 nm ESA proceeds on the early ps
timescale but is not associated with complete recovery of the
ground state bleach (GSB) band at 365 nm. Instead, a hot
ground state band contribution is observed on the lower energy
side of the GSB, which decays on the 10 ps time scale. No
product band formation is visible in the transient absorption data
in the detected wavelength range. This is due to the dominant
absorption of the (E)-isomer in this spectral window, which
obscures the formation of the (Z)-isomer. Nevertheless,
formation of the (Z)-isomer can be indirectly deduced by the
strong nondecaying GSB signal on the nanosecond timescale,
which is indicative for the depopulation of the initial (E)-isomer.
Figure 2. Geometry optimized conformations adopted by TphAB (left);
absorption spectra calculated at the ADC(2) level of theory (see Table S1 and
Figs. S3 and S4) and taking into account the distribution of the conformations
(right).
Although the geometry optimization of TphAB yields two stable
geometries, the twisted and the orthogonal conformation of the
(Z)-isomer, their Boltzmann distribution indicates that at room
temperature the orthogonal structure (TphAB-1) represents
~99.5% of the population. Interestingly, the Boltzmann
distribution of the (E)-isomers shows that also the TphAB-1
This article is protected by copyright. All rights reserved.

10.1002/anie.201909739
Angewandte Chemie International Edition
RESEARCH ARTICLE
Ultrafast Z → E photoisomerization
The ultrafast Z → E photoisomerization of TphAB was
investigated after 455 nm excitation in the nπ* (S1) absorption
band of the (Z)-isomer. On the sub-250 fs timescale, the
transient absorption data (Fig. 4A, Fig. S6B) shows a broad ESA
signal over the complete detection range. This ESA is
interrupted only by the GSB at ~450 nm (in the range of the nπ*
absorption of the (Z)-isomer). This early ESA undergoes an
ultrafast blue shift, which results in complete ESA/GSB signal
compensation and an ESA rise in the 365 nm range. Therefore,
after ~200 fs only positive transient absorption is present in the
detected spectral range. The ESA then decays on the sub-ps to
the early ps timescale. This decay is most obvious above
430 nm, while below 430 nm the ESA transforms into the
product absorption signature (330-420 nm). The product
absorption continues to grow strongly on the ps timescale and
results in an intense, long-lived product band, associated with
the efficient Z → E isomerization. Initially the product band
appears broader on the red (~400-450 nm) spectral side.
However, this red side undergoes a blue shift (from ~420 nm to
365 nm) on the 10 ps timescale due to the cooling of the hot
ground state product (Fig. 4A).
Figure 3. (A) Transient absorption data of the (E)-TphAB measured in
acetonitrile after 355 nm excitation (ππ* band). (GSB, negative absorption
difference signal, light to dark blue; excited-state and hot ground-state
absorption, positive absorption difference signal, yellow through red to black).
(B) Corresponding LDMs obtained from the lifetime distribution analysis (see
SI) of the time-resolved data in (A). LDMs are read as decay-associated
spectra from global lifetime analysis: (i) positive (red) amplitudes account for
decay of absorption or rise of GSB; (ii) negative (blue) amplitudes account for
rise of absorption or decay of GSB.
The experimental data were analysed by lifetime distribution
analysis (see SI and ref.[21]) and the corresponding lifetime
distribution map is presented in Fig. 3B. The decay of the
510 nm ESA is described by a ~100-200 fs lifetime distribution
(positive amplitude) and can be assigned to the decay of the
initially excited ππ* state (S2) of the (E)-isomer into the nπ* state
(S1) (negative amplitudes). The lifetime distributions for this
process are not fully resolved due to the limited time resolution
(~100 fs) of the experiments. Based on similar studies of the
6f]
conventional AB with higher time resolution[6d,
it can be
expected that the lifetime of this relaxation is even shorter.
Therefore, in our experiments this lifetime possibly overlaps with
the lifetime describing the relaxation on the S1 surface. The
decay of the 420 nm ESA, ascribed to the nπ* state (S1), is
characterized by a relatively broad lifetime distribution centered
at 950 fs (positive amplitude). Similar, albeit slightly shorter,
lifetime distribution was found in AB.[6e] Therefore, the decay of
TphAB from the S1 excited state to the ground state appears to
Figure 4. (A) Transient absorption data of the (Z)-TphAB measured in
acetonitrile after 455 nm excitation (nπ* band). (B) Corresponding LDMs
obtained from the lifetime distribution analysis of the time-resolved data in (A).
be slightly slower. Typically, this decay dynamics is fitted by two
The ultrafast dynamics of the (Z)-isomer is exceptionally
nonexponential as illustrated by all of the lifetime distributions
below 1-2 ps (Fig. 4B). Above 425 nm, a pair of a positive and
negative, tilted and elongated lifetime distributions stretches
from <100 fs to ~1 ps. They account for the ultrafast,
nonexponential, excited state relaxation dynamics that proceeds
from the S1 Franck-Condon region towards the S1 minimum and
through the conical intersection with the ground state. Similarly,
in the 350-430 nm range, a strong, elongated and tilted (from
short to long wavelengths) negative lifetime distribution is
present that also stretches from <100 fs to 1-2 ps. The shorter,
blue side of this distribution is associated with the early ESA
shift dynamics due to the relaxation on the S1 PES, while the
longer red side is associated with S1 to S0 transition and the
formation of the hot ground state product band. Again, due to
the strong nonexponentiality of the dynamics, the relaxation on
the S1 and to the ground state cannot be observed as separate
lifetime distributions. The cooling dynamics of the hot ground
6f]
lifetime components[6b,
via conventional global lifetime
analysis[21]. However, lifetime distribution analysis indicates that
those are artificially assigned to describe the rather broad
distribution of relaxation processes. Nevertheless, it was
proposed that the slower S1 relaxation pathways are reactive,
while the fast ones are nonreactive.[6f, 8] In this respect, the shift
of the corresponding lifetime distribution in TphAB towards
longer lifetimes (dominance of the reactive pathways) may
potentially explain the much higher isomerization QY (44%, see
above) as compared to AB (~10%)[4-5]
.
After the relaxation of the S1 state, a pair of a negative (365 nm)
and a positive (405 nm), 10-20 ps, elongated and tilted lifetime
distributions are observed that describe the nonexponential
cooling dynamics in the ground state.
This article is protected by copyright. All rights reserved.

10.1002/anie.201909739
Angewandte Chemie International Edition
RESEARCH ARTICLE
state photoproduct is described by a pair of a positive and a
negative distribution at ~8-20 ps around 350-430 nm (Fig. 4B).
attempts to adopt a twisted conformation, which apparently is
not accessible, and thus it moves towards the orthogonal
conformation, which is stabilized by the lone pair∙∙∙π interaction.
Overall, the E → Z photoisomerization from S1 is dominated by
the CNNC torsional reaction coordinate. However, this torsional
motion appears to be assisted by significant changes along the
inversion reaction coordinate (>∆15° for ∠CNN and ∠NNC).
These changes essentially adjust the relative position of the
benzene and the thiophene rings, and result in the adoption of
the orthogonal configuration by the (Z)-isomer. Furthermore, this
additional degree of freedom is most likely the reason for the
observed (Fig. 3B) slightly longer S1 lifetime of (E)-TPhAB than
of (E)-AB[6e] (lifetime distributions centered at 950 fs and 700 fs,
correspondingly).
Photoisomerization mechanism
It is generally accepted that torsion of the central CNNC moiety
8, 10]
plays a dominant role in the isomerization of ABs.[6f,
Therefore, we performed relaxed PES scans along the torsion
coordinate of TphAB in the S1 (nπ*) state (Fig. S7). The resulting
PESs for the orthogonal (TphAB-1) and the twisted (TphAB-2)
geometries resemble the shape of those in AB with a conical
intersection with the ground state at CNNC angle of ~90°.
However, given the unconventional geometry adopted by the
(Z)-isomer (Fig. 2), we decided to further investigate the
contribution of other reactive coordinates to the isomerization
mechanism operating in TphAB. In principle, nuclear dynamics
simulations would be required to obtain a dynamical picture of
the reaction mechanism, which are not feasible at present.
Instead, we performed unconstrained geometry optimization for
both the (E)- and (Z)-TphAB-1 (orthogonal geometry) starting
from the corresponding Franck-Condon geometries in S1 (see SI
for details). These calculations mimic the structural relaxation
processes and identify the relevant reaction coordinates. The
shape of the potential energy surfaces obtained along theses
optimizations allows for conclusions with respect to the dynamic
mechanism. The optimizations bring the molecule from the
Franck-Condon region to the conical intersection with the ground
state. At this point, we switched the unconstrained optimization
to the S0 PES to obtain the complete isomerization pathway. The
optimizations of (E)- and (Z)-TphAB-1 unveil contributions of
further coordinates beyond the typical CNNC torsion. For clarity,
we name the angle on the thiophene side of the molecule CNN,
while NNC corresponds to the angle on the benzene side, and
examine the changes along these two coordinates (Figs. 5, 6,
S8 and S9, Tables S2 and S3).
Isomerization mechanism of (E)-TphAB-1. The S1
unconstrained optimization of (E)-TphAB-1 shows a relatively
large increase (~15°) in both the CNN and the NNC angles,
while the CNNC angle remains at ~180° (Fig. 5 and S8). This is
followed by a large reduction of the CNNC angle to 116° and
simultaneously a minor reduction in the CNN and the NNC
angles to ~125° (Fig. 5, red circle 3, and Fig. S8). At this point,
the unconstrained optimization converges to a local minimum
due to the presence of a small barrier on the S1 PES on the way
to the conical intersection with the ground state at CNNC of 90°.
From this point, we performed constrained geometry
optimizations with fixed CNNC angles between 116° and 90° to
reach the conical intersection. During this torsion of the CNNC
angle, the NNC angle increases back to ~132°, while the CNN
angle reduces further to ~120°. At CNNC of 90°. We continued
the unconstrained optimization on the S0 PES to obtain the
complete isomerization pathway. The optimization shows a
gradual decrease of the CNNC angle to 0°, and of the NNC
angle to 122° (Fig. 5 and S8A). In contrast, the NNC angle first
increases from ~120° to ~130° to finally equilibrate at 127°
(Fig. 5). The changes in the CNN and the NNC angles in the
ground state optimization indicate that the molecule initially
Figure 5. Example geometries from the unconstrained/partially constrained
(∠CNNC 116° to 90°) optimization of (E)-TphAB-1 illustrating the isomerization
mechanism on the S1 PES (top). PES obtained from the
unconstrained/partially constrained optimization of (E)-TphAB-1 started from
the Franck-Condon region on S1. Note: CNN is the angle on the thiophene
side of the molecule, while NNC is the angle on the benzene side (bottom).
Isomerization mechanism of (Z)-TphAB-1. The unconstrained
relaxation of (Z)-TphAB-1 from the Franck-Condon region in S1
is dominated initially by a large opening motion of the NNC
angle, which increases by ~35° to ~158° (Fig. 6 and S9). During
the last 5° of this opening, the torsion of the CNNC moiety is
activated and the CNNC angle quickly reaches 30°. The initial
This article is protected by copyright. All rights reserved.

10.1002/anie.201909739
Angewandte Chemie International Edition
RESEARCH ARTICLE
changes of the CNN angles are minor (<5°). The following
geometric changes towards the conical intersection with the
ground state are governed by the torsion of the CNNC moiety,
which brings the CNNC angle from 30° to 90°. This torsional
motion is accompanied by a decrease in the NNC angle to 130°.
From the conical intersection the unconstrained optimization
proceed on the ground state towards the (E)-isomer. This
relaxation is associated with a decrease of both the CNN and
the NNC angles to 115° and a concomitant increase of the
CNNC angle to 180°. Similar to the E → Z photoisomerization
direction, the Z → E direction is also dominated by the torsional
motion about the CNNC moiety. However, here the inversion
reaction coordinate plays an even more important role, as we
detected much larger changes in the CNN and the NNC angles
(∆25° and ∆42°, correspondingly) than during the transformation
of the (E)- to the (Z)-isomer.
effectively free the torsional reaction coordinate. The disruption
of the lone pair∙∙∙π interaction is achieved via the opening of the
NNC angle, which pulls the S-atom of the thiophene ring away
from the plane of the benzene ring. Therefore, the initial motion
on the S1 PES is led by the NNC opening. Apparently, above
150° of the NNC angle, the strength of the lone pair∙∙∙π
interaction is sufficiently reduced and the CNNC torsion is
activated. From this point, the mechanism is governed by the
torsional reaction coordinate, while the CNN and the NNC
moieties work towards planarization of the TphAB molecule. The
complex changes that (Z)-TphAB-1 undergoes along the S1 PES
also explain the strongly nonexponential dynamics observed in
the transient absorption data (sub-250 fs timescale) discussed
above. Such nonexponential dynamics is not present in the
relaxation of (Z)-AB.[6e]
Conclusions
The uncommon properties of azoheteroarenes have drawn
significant attention as an alternative design to the popular
AB.[22] However, those studies have been focused mostly on the
investigation of nitrogen-based azoheteroarenes. Here we
present a detailed study on the photochromism of a different
thiophenylazobenzene (TphAB) photoswitch. We show that the
TphAB photoswitch has outstandingly high photoisomerization
QYs (QYE
= 44%, QYZ
= 65%.), photoconversion
E
Z
(ππ*) →
(nπ*) →
levels (PSS365 contains only ~3% (E)-isomer, while PSS285
contains only ~13% (Z)-isomer), and fatigue resistance. Our
theoretical calculations demonstrate that the (Z)-isomer of
TphAB adopts a geometry where the thiophene ring is perfectly
orthogonal to the benzene ring, with the S-atom facing the
benzene ring. This orthogonal geometry is stabilized by a rare
lone pair∙∙∙π interaction between the S-atom and the benzene
ring. We reveal that while the ultrafast photoisomerization of
TphAB occurs on a timescale similar to that of AB, the
corresponding dynamics is remarkably rich. The torsional motion
about the CNNC moiety is the dominant reaction coordinate.
However, the formation and disruption of the unusual orthogonal
geometry of the (Z)-isomer requires significantly larger changes
along the inversion coordinate (∠CNN and ∠NNC) than those
discussed in AB[10h]. Therefore, the presence of the lone pair∙∙∙π
interaction and the ensuing orthogonal geometry add additional
degree of freedom, as compared to AB, which evidently alters
the photochemistry of TPhAB.
Our work delivers important insight into the molecular basis of
the photoisomerization mechanism operating in TphAB, which is
relevant not only to azoheteroarenes but also to conventional
ABs. We also establish TphAB as an excellent photoswitch with
versatile properties, in many ways better than the once of the
widely utilized conventional AB and many of its derivatives.
Figure 6. Example geometries from the unconstrained optimization of (Z)-
TphAB-1 illustrating the isomerization mechanism on the S1 PES (top). PES
obtained from the unconstrained optimization of (Z)-TphAB-1 started from the
Franck-Condon region on S1. Note: CNN is the angle on the thiophene side of
the molecule, while NNC is the angle on the benzene side (bottom).
This photoisomerization mechanism of (Z)-TphAB-1 can be
straightforwardly explained considering the orthogonal geometry
of the isomer and the lone pair∙∙∙π interaction (see above).
Essentially, to initiate the isomerization first a disruption of the
lone pair∙∙∙π interaction is required. Such a disruption would then
Acknowledgements
C.S. and J.W. acknowledge funding by the DFG (WA 1850/4-2).
H.A.W. acknowledge funding by the DFG (WE 5601/6-1). We
This article is protected by copyright. All rights reserved.

10.1002/anie.201909739
Angewandte Chemie International Edition
RESEARCH ARTICLE
[12] a) T. Wendler, C. Schütt, C. Näther, R. Herges, J. Org. Chem. 2012, 77,
3284-3287; b) C. E. Weston, R. D. Richardson, P. R. Haycock, A. J. P.
White, M. J. Fuchter, J. Am. Chem. Soc. 2014, 136, 11878-11881; c)
Y.-T. Wang, X.-Y. Liu, G. Cui, W.-H. Fang, W. Thiel, Angew. Chem. Int.
Ed. 2016, 55, 14009-14013; d) J. Calbo, C. E. Weston, A. J. P. White,
H. S. Rzepa, J. Contreras-García, M. J. Fuchter, J. Am. Chem. Soc.
2017, 139, 1261-1274.
thank Daniel Bauer and Adeel Ahmad for their help in data
acquisition.
Keywords: photochromism • photoswitch •
thiophenylazobenzene • time-resolved spectroscopy •
isomerization mechanism
[13] a) J. Garcia-Amorós, M. Díaz-Lobo, S. Nonell, D. Velasco, Angew.
Chem. Int. Ed. 2012, 51, 12820-12823; b) J. Garcia-Amorós, M. C. R.
Castro, P. Coelho, M. M. M. Raposo, D. Velasco, Chem. Commun.
2013, 49, 11427-11429.
[1]
[2]
B. L. Feringa, W. R. Browne, Wiley-VCH, 2001.
a) M.-M. Russew, S. Hecht, Adv. Mater. 2010, 22, 3348-3360; b) Y.
Wang, Q. Li, Adv. Mater. 2012, 24, 1926-1945; c) W. Szymański, J. M.
Beierle, H. A. V. Kistemaker, W. A. Velema, B. L. Feringa, Chem. Rev.
2013, 113, 6114-6178; d) L. Schweighauser, H. A. Wegner,
Chembiochem 2015, 16, 1709-1711; e) N. Kameta, M. Masuda, T.
Shimizu, Chem. Eur. J. 2015, 21, 8832-8839; f) O. Parlak, S. Beyazit,
M. J. Jafari, B. T. S. Bui, K. Haupt, A. Tiwari, A. P. F. Turner, Adv.
Mater. Interfaces 2016, 3, 1500353; g) J. Thevarpadam, et al., Angew.
Chem. Int. Ed. 2016, 55, 2738-2742; h) S. Neri, S. Garcia Martin, C.
Pezzato, L. J. Prins, J. Am. Chem. Soc. 2017, 139, 1794-1797; i) T.
Podewin, et al., Chem. Sci. 2017, 8, 4644-4653; j) M. D. Pescher, L. J.
G. W. van Wilderen, S. Grützner, C. Slavov, J. Wachtveitl, S. Hecht, J.
Bredenbeck, Angew. Chem. Int. Ed. 2017, 56, 12092-12096; k) L. Dong,
Y. Feng, L. Wang, W. Feng, Chem. Soc. Rev. 2018, 47, 7339-7368; l)
X. Liu, J. Zhang, M. Fadeev, Z. Li, V. Wulf, H. Tian, I. Willner, Chem.
Sci. 2019, 10, 1008-1016.
[14] J. Garcia-Amoros, A. Bucinskas, M. Reig, S. Nonell, D. Velasco, J.
Mater. Chem. C 2014, 2, 474-480.
[15] M. J. Hansen, M. M. Lerch, W. Szymanski, B. L. Feringa, Angew. Chem.
Int. Ed. 2016, 55, 13514-13518.
[16] C. Slavov, et al., Phys. Chem. Chem. Phys. 2015, 17, 14045-14053.
[17] A. Dreuw, M. Wormit, WIREs Comput. Mol. Sci. 2015, 5, 82-95.
[18] A. Dreuw, M. Head-Gordon, Chem. Rev. 2005, 105, 4009-4037.
[19] P. R. Huddleston, V. V. Volkov, C. C. Perry, Phys. Chem. Chem. Phys.
2019, 21, 1344-1353.
[20] M. Egli, S. Sarkhel, Acc. Chem. Res. 2007, 40, 197-205.
[21] C. Slavov, H. Hartmann, J. Wachtveitl, Anal. Chem. 2015, 87, 2328-
2336.
[22] S. Crespi, N. A. Simeth, B. König, Nat. Rev. Chem. 2019, 3, 133-146.
[3]
[4]
E. Merino, Chem. Soc. Rev. 2011, 40, 3835-3853.
H. M. D. Bandara, S. C. Burdette, Chem. Soc. Rev. 2012, 41, 1809-
1825.
[5]
[6]
V. Ladányi, P. Dvořák, J. Al Anshori, Ľ. Vetráková, J. Wirz, D. Heger,
Photochem. Photobiol. Sci. 2017, 16, 1757-1761.
a) T. Fujino, S. Y. Arzhantsev, T. Tahara, J. Phys. Chem. A 2001, 105,
8123-8129; b) H. Satzger, C. Root, M. Braun, J. Phys. Chem. A 2004,
108, 6265-6271; c) T. Pancur, F. Renth, F. Temps, B. Harbaum, A.
Kruger, R. Herges, C. Nather, Phys. Chem. Chem. Phys. 2005, 7,
1985-1989; d) M. Quick, et al., J. Phys. Chem. B 2014, 118, 8756-
8771; e) C. Slavov, C. Yang, L. Schweighauser, C. Boumrifak, A.
Dreuw, H. A. Wegner, J. Wachtveitl, Phys. Chem. Chem. Phys. 2016,
18, 14795-14804; f) A. Nenov, et al., J. Phys. Chem. Lett. 2018, 1534-
1541.
[7]
a) T. Nägele, R. Hoche, W. Zinth, J. Wachtveitl, Chem. Phys. Lett. 1997,
272, 489-495; b) Y. C. Lu, C. W. Chang, E. W. G. Diau, J. Chin. Chem.
Soc. 2002, 49, 693-701; c) H. Satzger, S. Spörlein, C. Root, J.
Wachtveitl, W. Zinth, P. Gilch, Chem. Phys. Lett. 2003, 372, 216-223.
C. Xu, L. Yu, F. L. Gu, C. Zhu, Phys. Chem. Chem. Phys. 2018, 20,
23885-23897.
[8]
[9]
a) P. Hamm, S. M. Ohline, W. Zinth, J. Chem. Phys. 1997, 106, 519-
529; b) T. Fujino, S. Y. Arzhantsev, T. Tahara, Bull. Chem. Soc. Jpn.
2002, 75, 1031-1040.
[10] a) M. Böckmann, N. L. Doltsinis, D. Marx, Phys. Rev. E 2008, 78,
036101; b) Y. Ootani, K. Satoh, A. Nakayama, T. Noro, T. Taketsugu, J.
Chem. Phys. 2009, 131, 194306; c) M. Böckmann, N. L. Doltsinis, D.
Marx, J. Phys. Chem. A 2010, 114, 745-754; d) G. Tiberio, L. Muccioli,
R. Berardi, C. Zannoni, ChemPhysChem 2010, 11, 1018-1028; e) O.
Weingart, Z. Lan, A. Koslowski, W. Thiel, J. Phys. Chem. Lett. 2011, 2,
1506-1509; f) V. Cantatore, G. Granucci, M. Persico, Comp. Theor.
Chem 2014, 1040–1041, 126-135; g) E. M. M. Tan, S. Amirjalayer, S.
Smolarek, A. Vdovin, F. Zerbetto, W. J. Buma, Nat. Commun. 2015, 6,
5860; h) P. Tavadze, G. Avendaño Franco, P. Ren, X. Wen, Y. Li, J. P.
Lewis, J. Am. Chem. Soc. 2018, 140, 285-290.
[11] a) L. Gagliardi, G. Orlandi, F. Bernardi, A. Cembran, M. Garavelli,
Theor. Chem. Acc. 2004, 111, 363-372; b) S. Yuan, Y. Dou, W. Wu, Y.
Hu, J. Zhao, J. Phys. Chem. A 2008, 112, 13326-13334; c) J. Casellas,
M. J. Bearpark, M. Reguero, ChemPhysChem 2016, 17, 3068-3079.
This article is protected by copyright. All rights reserved.

10.1002/anie.201909739
Angewandte Chemie International Edition
RESEARCH ARTICLE
Entry for the Table of Contents
ARTICLE
Thiophenylazobenzene is a sulfur-
based azoheteroarene photoswitch
with impressive fatigue resistance,
photoisomerization efficiency, and
highly selective photoconversion. Its
(Z)-isomer adopts unusual orthogonal
geometry, stabilized by rare,
Chavdar Slavov, Chong Yang, Andreas
H. Heindl, Hermann A. Wegner, Andreas
Dreuw, and Josef Wachtveitl
Page No. – Page No.
Thiophenylazobenzene – an
alternative photoisomerization con-
trolled by lone pair∙∙∙π interaction
intramolecular lone pair∙∙∙π
interaction. The interaction governs
the ultrafast photoisomerization, which
requires a mechanism involving
significant movement along the
inversion reaction coordinate.
This article is protected by copyright. All rights reserved.