Hole Catalysis as a General Mechanism for Efficient and Wavelength-Independent Z → E Azobenzene Isomerization

Article abstract of DOI:10.1016/j.chempr.2018.06.002

Whereas the reversible reduction of azobenzenes has been known for decades, their oxidation is destructive and as a result has been notoriously overlooked. Here, we show that a chain reaction leading to quantitative Z → E isomerization can be initiated before reaching the destructive anodic peak potential. This hole-catalyzed pathway is accessible to all azobenzenes, without exception, and offers tremendous advantages over the recently reported reductive, radical-anionic pathway because it allows for convenient chemical initiation without the need for electrochemical setups and in the presence of air. In addition, catalytic amounts of metal-free sensitizers, such as methylene blue, can be used as excited-state electron acceptors, enabling a shift of the excitation wavelength to the far red of the azobenzene absorption (up to 660 nm) and providing quantum yields exceeding unity (up to 200%). Our approach will boost the efficiency and sensitivity of optically dense liquid-crystalline and solid photoswitchable materials. Video Abstract: [Figure presented] Molecular switches are a key ingredient in stimulus-responsive and adaptive materials and devices. Light is among the most attractive stimuli, yet photoswitches often require intense irradiation with high-energy UV light and suffer from inefficient switching as well as fatigue. Thus, the design of robust and efficient photoswitches constitutes an important challenge to boost the sensitivity and energy efficiency of the respective materials and devices. Here, we describe that the isomerization of azobenzene switches from their less stable Z isomer back to the more stable E isomer can be triggered by tiny, i.e., catalytic, amounts of holes caused by chemical, electrochemical, or photochemical oxidation. Our method is generally applicable to the entire family of azobenzene switches, does not require expensive equipment, and allows the reliable and efficient operation of these photoswitches by using red light with quantum efficiencies up to 200%. An efficient and generally applicable method is developed for operating azobenzene molecular switches by using catalytic amounts of holes (via an oxidant) or photons (via a photosensitizer). The pathway allows for indirect Z → E photoisomerization using lower-energy light than required for direct azobenzene excitation and with high quantum yields exceeding unity. The method should help to enhance the sensitivity of photoresponsive materials and devices with high optical density.

Full text of DOI:10.1016/j.chempr.2018.06.002

Article
Hole Catalysis as a General Mechanism for
Efficient and Wavelength-Independent Z / E
Azobenzene Isomerization
Alexis Goulet-Hanssens,
Clemens Rietze, Evgenii Titov,
Leonora Abdullahu, Lutz
Grubert, Peter Saalfrank, Stefan
Hecht
sh@chemie.hu-berlin.de
HIGHLIGHTS
Efficient and quantitative Z / E
isomerization enabled by hole-
catalyzed pathway
Initiation of radical chain reaction
by either chemical oxidation or
photo-oxidation
Azobenzene isomerization using
red light with quantum yields
exceeding unity
Generally applicable method to
enhance sensitivity of azobenzene
switches
An efficient and generally applicable method is developed for operating
azobenzene molecular switches by using catalytic amounts of holes (via an oxidant)
or photons (via a photosensitizer). The pathway allows for indirect Z / E
photoisomerization using lower-energy light than required for direct azobenzene
excitation and with high quantum yields exceeding unity. The method should help
to enhance the sensitivity of photoresponsive materials and devices with high
optical density.
Goulet-Hanssens et al., Chem 4, 1–16
July 12, 2018 ª 2018 Elsevier Inc.
https://doi.org/10.1016/j.chempr.2018.06.002

Please cite this article in press as: Goulet-Hanssens et al., Hole Catalysis as a General Mechanism for Efficient and Wavelength-Independent Z /
E Azobenzene Isomerization, Chem (2018), https://doi.org/10.1016/j.chempr.2018.06.002
Article
Hole Catalysis as a General Mechanism
for Efficient and Wavelength-Independent
Z / E Azobenzene Isomerization
Alexis Goulet-Hanssens,1 Clemens Rietze,2 Evgenii Titov,2 Leonora Abdullahu,1 Lutz Grubert,1
Peter Saalfrank,2 and Stefan Hecht1,3,
*
The Bigger Picture
SUMMARY
Molecular switches are a key
ingredient in stimulus-responsive
and adaptive materials and
devices. Light is among the most
attractive stimuli, yet
Whereas the reversible reduction of azobenzenes has been known for decades,
their oxidation is destructive and as a result has been notoriously overlooked.
Here, we show that a chain reaction leading to quantitative Z / E isomerization
can be initiated before reaching the destructive anodic peak potential. This
hole-catalyzed pathway is accessible to all azobenzenes, without exception,
and offers tremendous advantages over the recently reported reductive,
radical-anionic pathway because it allows for convenient chemical initiation
without the need for electrochemical setups and in the presence of air. In addi-
tion, catalytic amounts of metal-free sensitizers, such as methylene blue, can be
used as excited-state electron acceptors, enabling a shift of the excitation wave-
length to the far red of the azobenzene absorption (up to 660 nm) and providing
quantum yields exceeding unity (up to 200%). Our approach will boost the
efficiency and sensitivity of optically dense liquid-crystalline and solid photo-
switchable materials.
photoswitches often require
intense irradiation with high-
energy UV light and suffer from
inefficient switching as well as
fatigue. Thus, the design of robust
and efficient photoswitches
constitutes an important
challenge to boost the sensitivity
and energy efficiency of the
respective materials and devices.
Here, we describe that the
isomerization of azobenzene
switches from their less stable Z
isomer back to the more stable E
isomer can be triggered by tiny,
i.e., catalytic, amounts of holes
caused by chemical,
INTRODUCTION
Achieving precise spatiotemporal control over light-responsive systems is one of the
main goals of modern photochromism research1–3 and a key requirement for exploit-
ing molecular photoswitches to externally regulate the functions of materials and de-
vices.4 Azobenzene5,6 is one of the most frequently used photoswitches and is easily
converted from its thermodynamically stable E isomer to its metastable Z isomer
with light. Previously, Z / E isomerization has been catalyzed via ionic mechanisms
with various Brønsted/Lewis acids and bases,7 including iodine,8,9 tri-coordinate
phosphorus species,10 and organometallic complexes.11 In contrast to these
methods, which rely on bimolecular complexes to catalyze isomerization, we
recently reported that the electrochemical reduction of azobenzene can quantita-
tively isomerize a solution of Z-azobenzene to its E isomer,12 and thus catalytic elec-
trons13 provide an orthogonal handle to light to control this versatile molecular
switch. Although the reductive pathway is powerful and widely applicable to all azo-
benzenes, severe limitations include the need for an oxygen-free environment, the
need for an electrochemical setup, and the lack of easy-to-handle reagents to initiate
the radical chain. After our initial study,12 we were motivated by the work of Ge-
scheidt and coworkers, who studied the oxidatively catalyzed Z / E isomerization
of 1,10-azonorbornane.14 Some of the most recently reported and interesting azo-
benzene structures for biological applications tend to be rather electron rich,
bearing ortho-alkoxy groups15,16 or strong donors in para position to the N=N azo
electrochemical, or
photochemical oxidation. Our
method is generally applicable to
the entire family of azobenzene
switches, does not require
expensive equipment, and allows
the reliable and efficient
operation of these photoswitches
by using red light with quantum
efficiencies up to 200%.
Chem 4, 1–16, July 12, 2018 ª 2018 Elsevier Inc.
1

Please cite this article in press as: Goulet-Hanssens et al., Hole Catalysis as a General Mechanism for Efficient and Wavelength-Independent Z /
E Azobenzene Isomerization, Chem (2018), https://doi.org/10.1016/j.chempr.2018.06.002
Figure 1. Scope of This Work
The light-triggered isomerization of azobenzene to its metastable Z isomer can be reversed
thermally or with another wavelength of light (center). Alternatively, redox chemistry can be used to
trigger a catalytic chain to enable quantitative isomerization back to the thermodynamically more
stable E isomer. Whereas the reductive pathway (left) is known, this work is focusing on the hitherto
unknown oxidative pathway (right) and demonstrates its practicality.
bond,17 thus making them more amenable to oxidation (lower oxidation potential)
than to reduction (higher reduction potential). However, unlike the one-electron
reduction of azobenzene, which is reversible, with few exceptions, the one-electron
oxidation is not reversible and as a result has not been extensively studied.18 There is
some evidence in the literature that the oxidative pathway via radical cations is
available to azobenzene given that it was suggested in an early and inspiring
report of electron acceptors decreasing the thermal half-life of Z-azobenzene by
Schulte-Frohlinde7 and as a mechanism to explain the accelerated thermal Z / E
isomerization of azobenzene on gold nanoparticles.19,20 Although mechanistically
the reductive and oxidative pathways can be considered analogs to acid- and
base-catalyzed reactions,13,21 to the best of our knowledge the scope of the hole-
catalyzed oxidative pathway and how it relates to its reductive counterpart have
not been reported thus far.
In this work, we show that the oxidative hole-catalyzed pathway (as illustrated in
Figure 1) is a fully functional counterpart to the reductive electron-catalyzed pathway
and beyond it offers several key advantages. Although the formal oxidation event is
destructive when conducted electrochemically, the rate of electron transfer
following rapid initial Z / E isomerization is fast enough that decomposition does
not occur. Because single-electron radical oxidants can be prepared as bench-stable
salts, azobenzene electrocatalytic isomerization can now be triggered anywhere on
the bench, away from electrochemical setups.
1Department of Chemistry & IRIS Adlershof,
RESULTS AND DISCUSSION
Humboldt-Universitat zu Berlin,
¨
Brook-Taylor-Straße 2, 12489 Berlin, Germany
Attaining Hole-Catalyzed Isomerization
Early on in our studies, we wanted to identify the structural features that govern the
nature of the azobenzene oxidation, being either reversible or irreversible.18 We
sought to first study the reversible oxidation, aided by the few examples in the liter-
ature that have a recurring motif: azobenzenes bearing either bis-para-hydroxy22,23
or (bis-)para-dialkylamino functionalities,24 enabling the formation of a stable bis-
quinoidal species upon oxidation (Figure 2B), as supported by density functional
2Theoretische Chemie, Institut fur Chemie,
¨
Universitat Potsdam, Karl-Liebknecht-Straße
¨
24–25, 14476 Potsdam-Golm, Germany
3Lead Contact
*Correspondence: sh@chemie.hu-berlin.de
https://doi.org/10.1016/j.chempr.2018.06.002
2
Chem 4, 1–16, July 12, 2018

Please cite this article in press as: Goulet-Hanssens et al., Hole Catalysis as a General Mechanism for Efficient and Wavelength-Independent Z /
E Azobenzene Isomerization, Chem (2018), https://doi.org/10.1016/j.chempr.2018.06.002
Figure 2. Oxidation of Azobenzenes
(A) Cyclovoltammograms of azobenzenes 1, 1b, 1c, and 1i (1.0 3 10ꢀ3 M in CH3CN containing 0.1 M
Bu4NPF6) depicting the oxidation waves.
(B) The fully reversible two-electron oxidation in the case of azobenzene 1 is attributed to the stable
bis-quinoid structure of the dication.
theory (DFT) calculations (see Section 7.2.4 in the Supplemental Experimental Pro-
cedures). To avoid complicating secondary proton-transfer reactions as in the case
of hydroxy derivatives, we initially investigated azobenzene 1, which carries two
para-dimethylamino groups. The cyclic voltammogram of 1 shows a fully reversible
oxidation wave (Figure 2A). Spectroelectrochemical measurements (see Figures S55
and S56) and titrations with ceric ammonium nitrate (CAN; see Figure S71) demon-
strate that azobenzene 1 undergoes a two-electron oxidation at near-identical
potentials and not a one-electron oxidation as previously reported.24 Moreover,
moving the electron-rich dimethylamino groups to the meta position relative to
the central azo bond leads to an irreversible oxidation (see Figure S42). Indeed,
both para-dimethylamino groups are needed for a fully reversible oxidation, as
illustrated by the fact that the oxidation wave of 4-dimethylamino-40-methoxyazo-
benzene 1a, carrying a weaker donor (Hammett constants: sp(OCH3) = ꢀ0.27 versus
sp(N(CH3)2) = ꢀ0.83),25 is no longer fully reversible (see Figure S31). Indeed,
all nine other 4-dimethylaminoazobenzenes in the series (see Figure S1) demon-
strate semi-reversible oxidation waves regardless of the 40-R functional group (see
Figures S31–S39).
By plotting the peak potentials for this series (for data collection see Table S1) as a
function of the Hammett sp values, one can see that the cathodic (reduction) peak
potential is far more sensitive to the nature of the substituent R than the anodic
(oxidation) peak potential (see Figures S27 and S28). A possible explanation involves
Chem 4, 1–16, July 12, 2018
3

Please cite this article in press as: Goulet-Hanssens et al., Hole Catalysis as a General Mechanism for Efficient and Wavelength-Independent Z /
E Azobenzene Isomerization, Chem (2018), https://doi.org/10.1016/j.chempr.2018.06.002
oxidation at the amino group, leading to rearrangement into a semi-quinone struc-
ture in which the conjugation with the R group is broken, and thus the latter does not
significantly influence the stability of the radical cation. In contrast, reduction popu-
lates the p*-orbital, which is delocalized over the entire molecule and is thus more
sensitive to the appended chemical functionality. Moreover, in both the oxidation
and reduction of 1, the potential deviates significantly from the trend set by the other
compounds in the series. Indeed, the absolute shift of the anodic peak potential in
the oxidation of the doubly dimethylamino-substituted azobenzene 1 (sp = ꢀ0.83)25
in relation to the monofunctionalized azobenzene 1c (336 mV) is much larger than for
the corresponding derivative 1i (78 mV), carrying a nitro group that is of similar ab-
solute strength (sp = 0.78).25 This asymmetry is the result of a subsequent reorgani-
zation of the molecule to the bis-quinoid structure (Figure 2B). The latter is able to
fully protect azobenzene 1 upon oxidation, whereas the oxidation waves of all other
azobenzenes in the series are semi-reversible. Furthermore, in the absence of any
dialkylamino group, oxidation is fully irreversible.
In our previous work,12 we used substoichiometric reduction to initiate a chain
reaction, during which the Z-configured radical anion rapidly isomerizes to the
more stable E-configured radical anion. The latter is able to reduce neutral Z-azo-
benzene, yielding neutral E isomer as well as the initial Z-radical anion and thereby
propagating the chain to enable rapid and quantitative isomerization of the entire
Z-azobenzene solution (see Figure 1). In analogy to this mechanism, we speculated
about the existence of a corresponding catalytic cycle based on radical cations
instead of radical anions. Despite the short thermal half-life of Z-1, which amounts
to just about 4 min in acetonitrile at room temperature, we were able to investigate
the radical cation isomerization in solution with the aid of chemical oxidants, in this
case CAN, as a one-electron oxidant. A fresh solution of predominantly 1,+ was pre-
pared by the slow dropwise addition of one equivalent of CAN to a solution of 1 un-
der rapid stirring. Note that because of the practically identical anodic potentials of
the first and second oxidation in 1 (see Figure S30), some dication 12+ is always
generated as a side product, and thus conversion to the single radical species 1,+
is not quantitative, i.e., the concentration of 1,+ is always lower than the mol %
oxidant added. More importantly, addition of 0.3 mol % CAN, and thus formation
of less than 0.3 mol % of the radical cation 1,+, caused immediate and quantitative
isomerization to the E isomer, clearly demonstrating a catalytic isomerization via the
radical cation species (see Figure S73).
We were curious whether this catalytic isomerization behavior is limited to azoben-
zenes capable of undergoing a (semi-)reversible oxidation because this would
largely limit the scope and impact of this technique. To our delight, treatment
of a solution of Z-2 with a solution containing 0.5 mol % of CAN (Figure 3A) led
to an immediate and complete Z / E isomerization, showing that the electron-
transfer reaction responsible for propagation is faster than the irreversible oxida-
tion event, which would lead to destruction of the azobenzene. Because the anodic
peak potential of CAN (0.68 V; see Figure S29) is close to that of 2 (1.00 V; see
Figure S41), rapid agitation was necessary for achieving quantitative isomerization,
given that the addition of the oxidant without stirring led to incomplete isomeriza-
tion in all cases as a result of undesired side reactions of the radical cation instead
of chain propagation necessary for catalytic isomerization. Treatment of 2 with
superstoichiometric amounts, i.e., 2 equiv, of CAN shows a slow bleaching of
the azobenzene absorption band (see Figure S74). This bleaching is accelerated
upon irradiation at 365 nm, suggesting that the photochemically generated Z
isomer is more prone to degradation by the oxidant.
4
Chem 4, 1–16, July 12, 2018

Please cite this article in press as: Goulet-Hanssens et al., Hole Catalysis as a General Mechanism for Efficient and Wavelength-Independent Z /
E Azobenzene Isomerization, Chem (2018), https://doi.org/10.1016/j.chempr.2018.06.002
Figure 3. Spectral Changes upon Isomerization
(A) Absorption spectra during quantitative isomerization of Z-enriched azobenzene 2 (97% Z isomer
after 365 nm irradiation) after the addition of 0.5 mol% CAN oxidant solution demonstrate
chemocatalytic isomerization under aerobic conditions.
(B) Corresponding electrocatalytic isomerization occurs at a lower potential (inset: red curve) than
the irreversible anodic peak potential of 2 (inset: black curve), allowing for complete isomerization
of the sample without oxidative destruction of the azobenzene.
We performed DFT calculations including solvent effects (for full details on the theoret-
ical methods and results, see Section 7 of the Supplemental Experimental Procedures).
The calculations indicate that in the case of species 1, the oxidation of the Z isomer
should be facilitated given that the adiabatic ionization Gibbs free energy in acetonitrile
is about 0.2 eV lower than that of the corresponding E isomer (see Table S7). This
Chem 4, 1–16, July 12, 2018
5

Please cite this article in press as: Goulet-Hanssens et al., Hole Catalysis as a General Mechanism for Efficient and Wavelength-Independent Z /
E Azobenzene Isomerization, Chem (2018), https://doi.org/10.1016/j.chempr.2018.06.002
prediction was confirmed through spectroelectrochemical experiments showing that
the isomerization indeed occurs before the anodic peak potential of E-2 (Figure 3B).
These results suggest that a strong oxidant is not required to trigger the catalytic
isomerization and that compounds with lower potential might be beneficial in prevent-
ing the unwanted destruction of the E-isomer product. According to spectroelectro-
chemistry the isomerization of Z-azobenzene 2 occurs more than 200 mV before the
anodic peak potential (Figure 3B), suggesting that this process can be addressed
non-destructively. Indeed, we have repeated the electrochemical switching five times
without a measurable loss of the azobenzene (see Figure S66).
Oxidative Pathway Permits Chemically Initiated Catalysis
When designing photoswitchable molecules, much attention is given to developing
switches with an absorption spectrum that stretches as far into the visible as
possible, ideally reaching red wavelengths.26 When employing electrochemical
means to trigger catalytic isomerization, one could think of a low electrical potential
as akin to using red light, in that it mitigates the chances of unselective and poten-
tially damaging redox reactions. Since isomerization of azobenzene 2 is taking place
at 0.5 V versus Fc/Fc+, it implies that chemical oxidants weaker than CAN should be
able to trigger catalytic isomerization, and hence we were curious as to how low their
oxidizing power and, thus, reduction potential could be.
The role of the chemical oxidant is to generate a population of azobenzene radical
cations (Figure 4A) while assuming that the two isomers exhibit potentials compara-
ble enough to allow for electron transfer and propagation of the chain. One can
estimate the relative amount of azobenzene radical generated in solution from the
potential difference of the azobenzene and the chemical oxidant by using a simpli-
fied equilibrium constant (Equation 1; for derivation and explanation, see Section 4
of the Supplemental Experimental Procedures). This shows that the potential gap
(DE) between the radical cation (or anion) source and the azobenzene is exponen-
tially proportional to the amount of oxidized (or reduced) azobenzene and thus
should be predictive of the ability of that source to initiate catalytic isomerization.
K = e38:94DE
(Equation 1)
Triarylamines are very well-suited compounds for tuning the amount of oxidized azo-
benzene species because they form stable radical cations with anodic peak potentials
that can be widely tunedon thebasis of thesubstitutionpatternsof the aryl rings27 (cyclic
voltammograms of some symmetrically substituted triarylamines measured under our
experimental conditions, and thus comparable to the azobenzenes presented here,
are shown in Figure S50). The triarylamine radical cation salts are readily available, either
commercially as in the case of T2 or by treating the triarylamines with antimony penta-
chloride to afford the corresponding radical cations as their SbCl6 salts.28,29 These are
bench-stable solids, which can be dosed in with ease, thus allowing us to carry out oxi-
dations at a well-defined ‘‘potential’’ yet in the absence of an electrochemical setup.
If we consider three possible triarylamine radical cations, T1,+, T2,+, and T3,+, which
have three different anodic peak potentials of 380, 700, and 910 mV versus Fc/Fc+
(Figures 4B and 4C), respectively, their oxidative strength relative to that of azoben-
zenes 2, 3, and 4 (see Figure S1) will be different. Ignoring solvent effects or reorga-
nization energies allows us to estimate the fraction of azobenzene radical cation
in solution. This radical population should be directly proportional to the
efficiency of the isomerization. In the case of azobenzene 2, potential differences
of DE = ꢀ620 mV for T1,+, ꢀ300 mV for T2,+, and ꢀ90 mV for T3,+ should lead to
ratios of roughly 3.3 3 1011:1 for T1,+:T2,+, 8.4 3 106:1 for T2,+:T2,+, and 300:1
6
Chem 4, 1–16, July 12, 2018

Please cite this article in press as: Goulet-Hanssens et al., Hole Catalysis as a General Mechanism for Efficient and Wavelength-Independent Z /
E Azobenzene Isomerization, Chem (2018), https://doi.org/10.1016/j.chempr.2018.06.002
Figure 4. Chemical Oxidants
(A) Chemical oxidants generate a population of azobenzene radical cations to initiate oxidative
electrocatalysis.
(B) Triarylamines T1–T3 , which were used as oxidants after oxidation to their corresponding radical
cations.
(C) Different potential gaps associated with azobenzenes 2–4 in combination with T1–T3, leading
to different initiation efficiencies.
for T3,+:T2,+, respectively. This example illustrates that because of the logarithmic
relationship of Equation 1, small changes in the DE value lead to a wide range of
efficiencies for radical cation formation.
Below a certain concentration threshold, we found our isomerization experiments
difficult to reproduce and attribute this threshold effect to the non-linear stability
Chem 4, 1–16, July 12, 2018
7

Please cite this article in press as: Goulet-Hanssens et al., Hole Catalysis as a General Mechanism for Efficient and Wavelength-Independent Z /
E Azobenzene Isomerization, Chem (2018), https://doi.org/10.1016/j.chempr.2018.06.002
Figure 5. Oxidatively Induced Catalytic Isomerization
Isomerization of Z-2 induced by addition of catalytic amounts of T2 +, indicating that on average
d
each formed radical cation isomerizes approximately 155 azobenzene molecules before being
quenched.
of the radical cation upon dilution (see Figures S60 and S61). Because the
radical source T1,+ is so potent in catalyzing the isomerization of 2, less than
1 mol % is required. At spectroscopic concentrations of 10ꢀ5 M, the amount of
radical cation added is therefore below 10ꢀ7 M, from which a fraction is quenched
upon dilution and thus precludes reliable dosing of the initiator. After
this threshold is surpassed, the additions are reliable and indeed correspond to
the added quantities. To exclude this effect, we employed concentrations of
10ꢀ4 M and monitored the less intense n-p* transition. In these experiments,
we found that both T1,+ and T2,+ were able to catalyze the isomerization of
2. However, whereas T2,+ (DE = 620 mV) could be used to isomerize 3, T1,+
(DE = 1,220 mV) was not able to induce isomerization. Because azobenzenes
4 and 5 (see Figure S1) could not be isomerized with T3,+ (DE R 950 mV),
a
gap of around 620 mV (T2,+ to 3), again as in the case of T3,+:2,+
,
appeared to be near the limit required for successfully initiating the radical chain
reaction.
Interestingly, when an oxidant with a smaller gap to the azobenzene favoring the
formation of the azobenzene radical cation is used, it leads to a faster equilibration
of the system but also a more rapid consumption of the reactive radical species
through side reactions. This means that the azobenzene can effectively be ‘‘titrated’’
from its metastable Z to its thermodynamically stable E isomer. Under the conditions
and stirring rate employed in our experiment, we could therefore determine a
turnover number TON = 155 for the isomerization of azobenzene 2 induced by
oxidant T2,+ (Figure 5). In the case of T1,+, which has a longer equilibration time
(see Figure S62), this results in a lower TON = 116. These few examples demonstrate
the wide parameter space for using chemical oxidants to trigger Z / E isomerization
and thus to reset azobenzene photoswitches.
8
Chem 4, 1–16, July 12, 2018

Please cite this article in press as: Goulet-Hanssens et al., Hole Catalysis as a General Mechanism for Efficient and Wavelength-Independent Z /
E Azobenzene Isomerization, Chem (2018), https://doi.org/10.1016/j.chempr.2018.06.002
Figure 6. Efficiencies of Hole and Electron Catalysis
Amount of electrons (in mol %) that needs to be ejected from (oxidative pathway, red) or injected
into (reductive pathway, black) samples of azobenzenes 2–5 at the given potentials to trigger
quantitative Z / E isomerization (for corresponding plots of each azobenzene derivative, see
Figures S67–S70).
Oxidative versus Reductive Catalytic Isomerization: Both Are Possible
Having established the availability of the radical cation pathway to azobenzenes
where the radical anion pathway is known, we were interested in comparing the ef-
ficiencies of both opposite pathways (judged by the minimum amount of ejected or
injected electron measured in coulometry) and their sensitivity on the redox poten-
tial. Over the course of many experiments, we found that in the case of the relatively
electron-rich azobenzene 2, the oxidative pathway is more efficient than the reduc-
tive pathway, and additional coulometry on azobenzenes 3–5 (see Figures S67–S70)
showed that both pathways are available across the electrochemical spectrum of
azobenzenes (Figure 6). Whereas 3 is the parent azobenzene, 4 has the longest re-
ported half-life (2 years) of any azobenzene to date,30 and 5 is one of the most
Chem 4, 1–16, July 12, 2018
9

Please cite this article in press as: Goulet-Hanssens et al., Hole Catalysis as a General Mechanism for Efficient and Wavelength-Independent Z /
E Azobenzene Isomerization, Chem (2018), https://doi.org/10.1016/j.chempr.2018.06.002
electron-poor azobenzenes that we have measured.12 For the unsubstituted azo-
benzene 3, both pathways are equivalent in efficiency, whereas electron-poor azo-
benzenes 4 and 5 are more efficiently isomerized via the reductive pathway. It
must be noted that the oxidative potential of +2.2 V versus Fc/Fc+ that we had to
use for azobenzene 5 represents the technical limit in acetonitrile because the
solvent and electrolyte start to be oxidized at this potential (see Figure S47). Please
note that these efficiencies are coarse estimates given that they are also dependent
on the stirring rate during coulometry (see Section 3 of the Supplemental Experi-
mental Procedures). Although the efficiency can vary slightly in either direction, it
is important to note that the oxidative pathway is accessible for all azobenzenes
we have studied and appears to be general.
Initially, we were surprised that the efficiency of the oxidative pathway is overall com-
parable with that of the reductive pathway, given that the latter results in a stable
azobenzene radical anion. However, in both cases the radical ions formed are not
trapped on the azobenzene molecules and represent only a small population of
the sample during the catalytic reaction. In the case of the oxidative pathway, if
the lifetime on the azobenzene radical cation is not sufficiently long for its decompo-
sition to occur, then this process becomes irrelevant and thus does not impede
catalytic isomerization. Using Eyring transition state theory based on DFT, we calcu-
lated the rates for Z / E isomerization of azobenzenes 1–5 in acetonitrile solution
(and in the gas phase) at room temperature. It turns out that the radical (anion or
cation) isomerization rates are between 14 and 17 orders of magnitude faster than
their neutral species (see Table S8). This tremendous difference is in part due to a
change in the isomerization geometry (see Figure S90). When compared with the
charge-neutral inversion pathway, the radical anions follow a rotational pathway in
all cases, whereas the radical cations undergo rotation for 1 and 2 but inversion
for 3 and 4 and an inversion or rotation pathway for 5 (see Table S8). Note that in
all cases, the E isomer is more stable than the Z isomer, providing the necessary ther-
modynamic driving force for thermal isomerization regardless of occurring via the
radical cation, radical anion, or charge-neutral species (see Table S8).
In almost all cases, the isomerization was calculated to proceed somewhat faster
through the radical cation versus the radical anion pathway given that the correspond-
ing computed rate constants are about two orders of magnitude higher (see Table S8).
In view of the very fast isomerization rate constants for both pathways, the overall reac-
tion rate is, however, limited by the diffusion and the (intermolecular) electron-transfer
steps.31 To gain better insight into the latter, we estimated their thermodynamics by
calculating reaction energies for the charge-transfer reactions E,+/ꢀ + Z / E + Z,+/ꢀ
that enable propagation (see Section 7 of the Supplemental Experimental Procedures).
In the oxidative case, an efficient chain reaction is expected for all azobenzenes 1–5 as a
result of the negative free energy for this propagation step (DG between ꢀ1.4 and
ꢀ7 kcal,molꢀ1 in acetonitrile; see Table S8). In contrast, for the reductive pathway the
propagation reaction is almost thermoneutral, sometimes slightly positive or slightly
negative depending on the azobenzene derivative. In addition, the computed ionization
potentials (on various levels of theory) support the trend suggested by Figure 6 that
higher potentials are needed to oxidize and isomerize the molecules in the order
from 1 to 5 (see Table S7 and Figure S89).
Catalytic Isomerization Using Photoinitiated Catalysis: Quantum Yields above
Unity
Orthogonal and catalytic processes are attractive and underdeveloped tools to
achieve precise and improved control over switchable systems. We were intrigued
10
Chem 4, 1–16, July 12, 2018

Please cite this article in press as: Goulet-Hanssens et al., Hole Catalysis as a General Mechanism for Efficient and Wavelength-Independent Z /
E Azobenzene Isomerization, Chem (2018), https://doi.org/10.1016/j.chempr.2018.06.002
about the possibility of using photoinduced electron transfer to trigger switching in
an oxidative context, as it would drastically increase the availability and application
of photoelectron-transfer agents.31–34 In our previous work, we matched the LUMO
(lowest unoccupied molecular orbital) level of our azobenzene to allow for a photo-
induced reduction event leading to catalytic isomerization, yet selective excitation
was not possible because of overlapping absorption of the photoreducing agent
and the azobenzene.12 To photogenerate a radical cation, we sought a photoelec-
tron-transfer agent, which would have an excitation wavelength red from where azo-
benzenes absorb. For these reasons, we chose to employ methylene blue (MB),
which features an excitation wavelength centered at 655 nm and is reported to
exhibit an excited-state oxidation potential of around 1.56 V for the S1 state and
1.60 V for the T1 state (values versus standard calomel electrode [SCE], measured
in methanol).31,35 These values match the anodic peak potential of azobenzene 2
(measured 1.00 V versus Fc/Fc+, corresponding to 1.41 V versus SCE) well and
thus should allow triggering of Z / E isomerization of 2 with red light at 660 nm
via an indirect photoelectron-transfer catalytic pathway, akin to photoredox catal-
ysis,31–34 but because the catalyst in our case initiates a chain reaction, we refer to
it as photoinitiator. Indeed, a solution of Z-2 can be quantitatively and efficiently iso-
merized to E-2 with 660 nm red light in the presence of MB (Figure 7). Because of the
indirect excitation mechanism, the amount of time to fully recover the E isomer de-
pends on the concentration of MB. Full recovery of the E isomer, however, cannot be
achieved by direct excitation because the best possible photostationary state (PSS),
obtained by irradiation at 436 nm (in the absence of MB), contains only 69% E isomer
as a result of the poor spectral separation of the n-p* bands. Although this system
functions well in the presence of oxygen, it is far more efficient in degassed solution
as the excited triplet state of the MB is not quenched by oxygen, allowing for more
Z-2,+ to be formed in a given time period.
There is a fortuitous coincidence in using MB as the photoelectron-transfer agent in
that the compound has a very low absorbance around 365 nm with a molar extinction
coefficient that is about 30 times lower than that of E-2 (see Figure S77). Hence,
no difference in the composition of the Z-enriched PSS of azobenzene 2 was
observed in the absence of MB when compared with up to 10 mol % MB upon
irradiation at 365 nm in aerated acetonitrile (see Figure S78), allowing for true in
situ switching in both the E / Z and Z / E directions with 365 and 660 nm light,
respectively. Such a system shows no fatigue of the azobenzene after five cycles of
365-nm-induced E / Z isomerization followed by 660-nm-induced Z / E isomeri-
zation. Furthermore, an equimolar mixture of 2 and MB showed no sign of azoben-
zene photobleaching after 2 hr of irradiation. Indeed, the system does not seem to
be limited by the survival of the azobenzene but rather the survival of the MB photo-
electron-transfer agent under constant excitation (as a result of the generation
of singlet-oxygen causing decomposition). Therefore, at very low loadings
(<0.1 mol %), the amount of MB present might not be sufficient to survive excitation
until quantitative Z / E isomerization has been achieved.
The use of triplet energy-transfer agents in conjunction with azobenzene has been
previously studied, which included the study of MB.36,37 In view of these studies
and reports that the triplet energy level of azobenzene does not seem to vary widely
with functionalization,38–40 we were considering triplet energy transfer from MB to
Z-2 as a mechanistic alternative to the electron-transfer pathway. To confirm that
indeed the latter is operating in our system, we measured quantum yields for the
entire Z / E isomerization process. In these experiments, MB was irradiated at
579 nm, when actinometry can be performed, and the isomerization rate of
Chem 4, 1–16, July 12, 2018
11

Please cite this article in press as: Goulet-Hanssens et al., Hole Catalysis as a General Mechanism for Efficient and Wavelength-Independent Z /
E Azobenzene Isomerization, Chem (2018), https://doi.org/10.1016/j.chempr.2018.06.002
Figure 7. Isomerization Induced by Red Light
(A) Efficient Z / E isomerization of azobenzene 2 in the presence of MB with red light (>610 nm).
With 1 mol % MB, the degassed acetonitrile solution of 2 (3.5 3 10ꢀ5 M) is quantitatively isomerized
within 14 s. The expanded region from 550 to 750 nm shows the corresponding bleaching of the MB.
(B) Z / E isomerization kinetics as a function of MB loading with a 660 nm LED in combination with a
610 nm cutoff filter in aerated acetonitrile.
azobenzene Z / E isomerization was monitored as a function of irradiation time. The
thus obtained quantum yield for the process encompasses the photoexcitation of
MB as well as the subsequent electron-transfer reaction and is highly concentration
and solvent dependent. Because the photostability of MB is strongly dependent on
the solvent, giving rise to fast bleaching in acetone and DMSO, we chose to conduct
our experiments in acetonitrile, which is more desirable to work in than
12
Chem 4, 1–16, July 12, 2018

Please cite this article in press as: Goulet-Hanssens et al., Hole Catalysis as a General Mechanism for Efficient and Wavelength-Independent Z /
E Azobenzene Isomerization, Chem (2018), https://doi.org/10.1016/j.chempr.2018.06.002
Table 1. Quantum Yields and PSS Composition for Indirect Isomerization of Selected
Azobenzenes in Acetonitrile with 579 nm in the Presence of 20 mol % MB,PF6
Azobenzene
Faerated
0.013
Fdegassed
0.99
PSS
2
3
4
5
100% E
97% E
92% E
89% E
0.0004
0.0018
0.0005
0.0018
0.023
0.28
dichloromethane, which performed equally well. We measured quantum yields for
this system between F = 0.63 and F = 1.99 depending on the solvent, the
counter-ion of the MB, and concentration and ratio of azobenzene 2 to MB (see
Tables S4–S6). A quantum yield F > 1 is clearly indicative of a photoinduced
electron-transfer mechanism as only a radical chain reaction—and not a triplet
energy-transfer pathway—can account for more than one switching event per
photon absorption. If we, however, consider the reported intersystem crossing
quantum yield of MB as its chloride salt (FISC = 0.52)41 and our measured value of
FISC = 0.3 for the PF6 salt (see Figure S88), all of the isomerization quantum yields
are higher than the quantum yield of the intersystem crossing, clearly implying
that in all cases photoinduced electron transfer (subsequently initiating hole
catalysis) is taking place and not triplet energy transfer (sensitization).
After exploring the energetically ‘‘matched’’ case of 2 in combination with MB, we
were eager to test azobenzenes 3–5, which do not possess oxidation potentials
that match the excited MB. To our surprise, MB was still able to trigger indirect
Z / E isomerization in all three azobenzenes. Measuring quantum yields for these
reactions, however, showed much smaller values and a high dependence on
whether oxygen was present in solution or not (Table 1). The obtained values are
similar to the ones reported for using MB, eosin Y, and fluorescein to sensitize
azobenzene isomerization,37 suggesting a triplet energy-transfer pathway. More-
over, the PSS of the azobenzenes 3–5 did not reach 100% E, which we interpret as
additional proof for triplet sensitization because the radical pathway is unidirec-
tional. Note that the quantum yield for the sensitized Z / E isomerization of 5 is
rather high (F = 0.28), indicating that the triplet energy level of both isomers must
be rather low (<33 kcal,mol )
ꢀ1 42 to allow for triplet energy transfer to both isomers
resulting in a PSS composed of 89% E-5.
By uncovering both complementary pathways, either via radical anions7 or cations,
we have made a wide variety of azobenzene derivatives amenable to a range of
photoelectron-transfer agents with matching HOMO (highest occupied molecular
orbital) or LUMO energies, thereby allowing for the selection of suitable excitation
wavelengths that do not interfere with the direct (conventional) operation of the
photoswitch.
Conclusions
We began our journey into the redox chemistry of azobenzenes by exploring the
reversibility of their oxidation. Whereas strongly electron-donating substituents in
the para position, such as dimethylamino groups in azobenzene 1, stabilize the
oxidation products via a semi-quinone structure, fewer electron-donating residues
lead to semi-reversible and irreversible oxidation of azobenzenes 2–5. Despite
this daunting instability, the transiently formed radical cations are able to success-
fully initiate a radical chain reaction leading to quantitative Z / E isomerization
Chem 4, 1–16, July 12, 2018
13

Please cite this article in press as: Goulet-Hanssens et al., Hole Catalysis as a General Mechanism for Efficient and Wavelength-Independent Z /
E Azobenzene Isomerization, Chem (2018), https://doi.org/10.1016/j.chempr.2018.06.002
of all azobenzene derivatives, regardless of the irreversible nature of their
oxidation. From a more practical standpoint, the oxidative, radical-cation-catalyzed
isomerization can be initiated with shelf-stable single electron oxidants, which can
conveniently be used in the presence of oxygen and thus without the need for so-
phisticated electrochemical setups. Both pathways—involving either stable radical
anions or unstable radical cations—rely on dramatic acceleration of the Z / E isom-
erization in the radical ions and are comparable in their efficiency. The availability of
the radical-cation-catalyzed pathway allows for the use of photoelectron-transfer
agents with strongly red-shifted absorption. In the presence of catalytic amounts
of MB and 660 nm red light, azobenzene 2 underwent quantitative Z / E isomeriza-
tion with quantum yields as high as F = 1.99. Such systems allow for both efficient
direct E / Z photoisomerization at 365 nm as a result of negligible absorption of
the MB and quantitative indirect Z / E isomerization at 660 nm as a result of the
operating photoinitiated hole catalysis.
Thus far, we have been studying our systems in solution, often at spectroscopic con-
centrations. We envision a high impact in optically dense systems that are notori-
ously difficult to isomerize quantitatively, such as liquid crystals42 or the solid
state.43,44 The applicability of the oxidative, radical-cation-catalyzed pathway in
the presence of oxygen and water should furthermore provide an additional tool
for the biology and photopharmacology community.45,46 Future work in our
laboratories will focus on using this mechanism to selectively operate molecular
switches in complex systems with orthogonal stimuli and to exploit the underlying
charge-transfer process.
EXPERIMENTAL PROCEDURES
Full experimental procedures are provided in the Supplemental Information.
SUPPLEMENTAL INFORMATION
Supplemental Information includes Supplemental Experimental Procedures, 90 fig-
ures, and 10 tables and can be found with this article online at https://doi.org/10.
1016/j.chempr.2018.06.002.
A
video abstract is available at https://doi.org/10.1016/j.chempr.2018.06.
002#mmc2.
ACKNOWLEDGMENTS
We thank Sergey Kovalenko for time-resolved measurements as well as Lukas
Gallandi for valuable discussions and assistance with the GW method. A.G.-H. is
indebted to the Alexander von Humboldt-Foundation and the Fonds de Recherche
du Que´ bec – Nature et Technologies (FQRNT) for providing postdoctoral fellow-
ships. Generous support from the Daimler and Benz Foundation (via 32-02/14),
the German Research Foundation (via SFB 658), and the European Research Council
(via ERC-2012-StG_308117 ‘‘Light4Function’’) is gratefully acknowledged.
AUTHOR CONTRIBUTIONS
A.G.-H. and L.A. synthesized all azobenzene derivatives; A.G.-H. carried out all
spectroscopic investigations, actinometry, and chemical catalysis, as well as
sensitization experiments; L.G. carried out the cyclovoltammetry, coulometry, and
spectroelectrochemistry; C.R., E.T., and P.S. performed all computational work;
A.G.-H. and S.H. designed the experiments and wrote the paper.
14
Chem 4, 1–16, July 12, 2018

Please cite this article in press as: Goulet-Hanssens et al., Hole Catalysis as a General Mechanism for Efficient and Wavelength-Independent Z /
E Azobenzene Isomerization, Chem (2018), https://doi.org/10.1016/j.chempr.2018.06.002
DECLARATION OF INTERESTS
The authors declare no competing interests.
Received: January 30, 2018
Revised: March 23, 2018
Accepted: June 4, 2018
Published: June 28, 2018
REFERENCES AND NOTES
1. Feringa, B.L., and Browne, W.R., eds. (2011).
Molecular Switches, Second Edition (Wiley-
VCH Verlag GmbH).
and near infrared photoswitches based on
azonium ions. Chem. Commun. (Camb.) 51,
12981–12984.
resonance and field parameters. Chem. Rev.
91, 165–195.
26. Ble´ ger, D., and Hecht, S. (2015). Visible-light-
activated molecular switches. Angew. Chem.
Int. Ed. 54, 11338–11349.
2. Tian, H., and Zhang, J., eds. (2016).
Photochromic Materials (Wiley-VCH Verlag
GmbH).
16. Dong, M., Babalhavaeji, A., Samanta, S.,
Beharry, A.A., and Woolley, G.A. (2015). Red-
shifting azobenzene photoswitches for in vivo
use. Acc. Chem. Res. 48, 2662–2670.
27. Dapperheld, S., Steckhan, E., Grosse
Brinkhaus, K., and Esch, T. (1991). Substituted
triarylamine cation-radical redox systems -
synthesis, electrochemical and spectroscopic
properties, Hammet behavior, and suitability
as redox catalysts. Chem. Ber. 124, 2557–
2567.
3. Yokoyama, Y., and Nakatani, K., eds. (2017).
Photon-Working Switches (Springer).
17. Trads, J.B., Burgstaller, J., Laprell, L., Konrad,
D.B., de la Osa de la Rosa, L., Weaver, C.D.,
Baier, H., Trauner, D., and Barber, D.M.
(2017). Optical control of GIRK channels
using visible light. Org. Biomol. Chem. 15,
76–81.
4. Russew, M.-M., and Hecht, S. (2010).
Photoswitches: from molecules to materials.
Adv. Mater. 22, 3348–3360.
5. Hartley, G.S. (1937). The cis-form of
azobenzene. Nature 140, 281.
28. Connelly, N.G., and Geiger, W.E. (1996).
Chemical redox agents for organometallic
chemistry. Chem. Rev. 96, 877–910.
18. Stradins, J.P., and Glezer, V.T. (1979). Azo,
azoxy and diazo compounds. In Encyclopedia
of Electrochemistry of the Elements, volume
XIII, A.J. Bard and H. Lund, eds. (Dekker),
pp. 163–208.
6. Bandara, H.M.D., and Burdette, S.C. (2012).
Photoisomerization in different classes of
azobenzene. Chem. Soc. Rev. 41, 1809–
1825.
29. Bell, F.A., Ledwith, A., and Sherrington, D.C.
(1969). Cation-radicals: tris-(p-bromophenyl)
ammonium perchlorate and
hexachloroantimonate. J. Chem. Soc. C Org.
20, 2719–2720.
¨
19. Hallett-Tapley, G.L., D’Alfonso, C., Pacioni,
N.L., McTiernan, C.D., Gonza´lez-Be´ jar, M.,
Lanzalunga, O., Alarcon, E.I., and Scaiano, J.C.
(2013). Gold nanoparticle catalysis of the cis-
trans isomerization of azobenzene. Chem.
Commun. (Camb.) 49, 10073–10075.
7. Schulte-Frohlinde, D. (1958). Uber die
thermische und katalytische cis/trans-
Umlagerung substituierter Azobenzole. Justus
Liebigs Ann. Chem. 612, 131–138.
30. Ble´ ger, D., Schwarz, J., Brouwer, A.M., and
Hecht, S. (2012). o-Fluoroazobenzenes as
readily synthesized photoswitches offering
nearly quantitative two-way isomerization with
visible light. J. Am. Chem. Soc. 134, 20597–
20600.
8. Yamashita, S. (1961). Iodine-catalyzed thermal
cis-trans isomerization of azobenzene. Bull.
Chem. Soc. Jpn. 34, 842–845.
20. Titov, E., Lysyakova, L., Lomadze, N., Kabashin,
A.V., Saalfrank, P., and Santer, S. (2015).
Thermal cis-to-trans isomerization of
azobenzene-containing molecules enhanced
by gold nanoparticles: an experimental and
theoretical study. J. Phys. Chem. C 119, 17369–
17377.
9. Arnaud, R., and Lemaire, J. (1974).
Isomerisation cis-trans de I’azobenzene
catalyse´ e par I’iode. III. Can. J. Chem. 52, 1868–
1871.
31. Romero, N.A., and Nicewicz, D.A. (2016).
Organic photoredox catalysis. Chem. Rev. 116,
10075–10166.
32. Majek, M., and Jacobi von Wangelin, A.
(2016). Mechanistic perspectives on organic
photoredox catalysis for aromatic
substitutions. Acc. Chem. Res. 49, 2316–
2327.
10. Hall, C.D., and Beer, P.D. (1991). Trico-
ordinate phosphorus compounds as catalysts
for the isomerization of (Z)- to (E)-
azobenzene. J. Chem. Soc. Perkin Trans. 2,
1947–1950.
21. Luca, O.R., Gustafson, J.L., Maddox, S.M.,
Fenwick, A.Q., and Smith, D.C. (2015). Catalysis
by electrons and holes: formal potential scales
and preparative organic electrochemistry. Org.
Chem. Front. 2, 823–848.
33. Xuan, J., and Xiao, W. (2012). Visible-light
photoredox catalysis. Angew. Chem. Int. Ed.
51, 6828–6838.
11. Nakamura, A., Doi, K., Tatsumi, K., and Otsuka,
S. (1975). Catalysis by low-valent diamagnetic
transition-metal complexes of the cis-trans
isomerization of azobenzenes. J. Mol. Catal. 1,
417–429.
22. Eriksson, A., and Nyholm, L. (2001).
Coulometric and spectroscopic investigations
of the oxidation and reduction of some
azosalicylic acids at glassy carbon electrodes.
Electrochim. Acta 46, 1113–1129.
34. Prier, C.K., Rankic, D.A., and MacMillan,
D.W.C. (2013). Visible light photoredox
catalysis with transition metal complexes:
applications in organic synthesis. Chem. Rev.
113, 5322–5363.
12. Goulet-Hanssens, A., Utecht, M., Mutruc, D.,
Titov, E., Schwarz, J., Grubert, L., Ble´ ger, D.,
Saalfrank, P., and Hecht, S. (2017).
Electrocatalytic Z / E isomerization of
azobenzenes. J. Am. Chem. Soc. 139,
335–341.
23. Eriksson, A., and Nyholm, L. (1999).
Comparison of the electrochemical properties
of some azosalicylic acids at glassy carbon
electrodes by cyclic and hydrodynamic
voltammetry. Electrochim. Acta 44, 4029–
4040.
35. Shen, T., Zhao, Z.-G., Yu, Q., and Xu, H.-J.
(1989). Photosensitized reduction of benzyl by
heteraoatom-containing anthracene dyes.
J. Photochem. Photobiol. A Chem. 47,
203–212.
13. Studer, A., and Curran, D.P. (2014). The
electron is a catalyst. Nat. Chem. 6, 765–773.
24. Kaplan, M.L., Haddon, R.C., Raghavachari, K.,
Menezes, S., Schilling, F.C., Hauser, J.J., and
Marshall, J.H. (1982). Electrical-conductivity in
Tcnq salts of bis(4-dimethylaminophenylimino)
sulfur and its structural analogs. Mol. Cryst. Liq.
Cryst. 80, 51–66.
36. Ronayette, J., Arnaud, R., Lebourgeois, P., and
Lemaire, J. (1974). Isome´ risation
14. Gescheidt, G., Lamprecht, A., Heinze, J.,
Schuler, B., Schmittel, M., Kiau, S., and
photochimique de l’azobenze` ne en solution. I.
Can. J. Chem. 52, 1848–1857.
Ruchardt, C. (1992). Electron-transfer-catalyzed
¨
cis-trans isomerization of 1,10-azonorbornane.
Helv. Chim. Acta 75, 1607–1612.
37. Ronayette, J., Arnaud, R., and Lemaire, J.
(1974). Isome´ risation photosensibilise´ e par
des colorants et photore´ duction de
15. Dong, M., Babalhavaeji, A., Hansen, M.J.,
Ka´lma´n, L., and Woolley, A. (2015). Red, far-red,
25. Hansch, C., Leo, A., and Taft, R.W. (1991). A
survey of Hammett substituent constants and
Chem 4, 1–16, July 12, 2018
15

Please cite this article in press as: Goulet-Hanssens et al., Hole Catalysis as a General Mechanism for Efficient and Wavelength-Independent Z /
E Azobenzene Isomerization, Chem (2018), https://doi.org/10.1016/j.chempr.2018.06.002
l’azobenze` ne en solution. II. Can. J. Chem. 52,
1858–1867.
azobenzenes: an energy transfer investigation.
J. Photochem. 23, 249–256.
44. Castellanos, S., Goulet-Hanssens, A., Zhao, F.,
Dikhtiarenko, A., Pustovarenko, A., Hecht, S.,
Gascon, J., Kapteijn, F., and Ble´ ger, D. (2016).
Structural effects in visible-light-responsive
metal-organic frameworks incorporating
ortho-fluoroazobenzenes. Chem. Eur. J. 22,
746–752.
38. Pietro Bortolus, S.M. (1979). Cis-trans
photoisomerization of azobenzene. Solvent
and triplet donor effects. J. Phys. Chem. 83,
648–652.
41. Montalti, M., Credi, A., Prodi, L., and Gandolfi,
M.T. (2006). Handbook of Photochemistry,
Third Edition (CRC Press).
42. Ichimura, K. (2000). Photoalignment of
liquid-crystal systems. Chem. Rev. 100,
1847–1874.
45. Broichhagen, J., Frank, J.A., and Trauner, D.
(2015). A roadmap to success in
photopharmacology. Acc. Chem. Res. 48,
1947–1960.
39. Gagliardi, L., Orlandi, G., Bernardi, F.,
Cembran, A., and Garavelli, M. (2004). A
theoretical study of the lowest electronic states
of azobenzene: the role of torsion coordinate in
the cis-trans photoisomerization. Theor. Chem.
Acc. 111, 363–372.
43. Bushuyev, O.S., Tomberg, A., Friscic, T., and
Barrett, C.J. (2013). Shaping crystals with light:
crystal-to-crystal isomerization and
photomechanical effect in fluorinated
azobenzenes. J. Am. Chem. Soc. 135, 12556–
12559.
46. Lerch, M.M., Hansen, M.J., van Dam, G.M.,
Szymanski, W., and Feringa, B.L. (2016).
Emerging targets in photopharmacology.
Angew. Chem. Int. Ed. 55, 10978–10999.
40. Monti, S., Dellonte, S., and Bortolus, P. (1983).
The lowest triplet state of substituted
16
Chem 4, 1–16, July 12, 2018