The Photodynamic Covalent Bond: Sensitized Alkoxyamines as a Tool to Shift Reaction Networks Out-of-Equilibrium Using Light Energy

Article abstract of DOI:10.1021/jacs.8b03633

We implement sensitized alkoxyamines as "photodynamic covalent bonds" - bonds that, while being stable in the dark at ambient temperatures, upon photoexcitation efficiently dissociate and recombine to the bound state in a fast thermal reaction. This type of bond allows for the photochemically induced exchange of molecular building blocks and resulting constitutional variation within dynamic reaction networks. To this end, alkoxyamines are coupled to a xanthone unit as triplet sensitizer enabling their reversible photodissociation into two radical species. By studying the photochemical properties of three generations of sensitized alkoxyamines it became clear that the nature and efficiency of triplet energy transfer from the sensitizer to the alkoxyamine bond as well as the reversibility of photodissociation crucially depends on the structure of the nitroxide terminus. By employing the thus designed photodynamic covalent bonding motif, we demonstrate how to use light energy to shift a dynamic covalent reaction network away from its thermodynamic minimum into a photostationary state. The network could be repeatedly switched between its minimum and kinetically trapped out-of-equilibrium state by thermal scrambling and selective photoactivation of sensitized alkoxyamines, respectively.

Full text of DOI:10.1021/jacs.8b03633

Article
pubs.acs.org/JACS
Cite This: J. Am. Chem. Soc. XXXX, XXX, XXX−XXX
The Photodynamic Covalent Bond: Sensitized Alkoxyamines as a
Tool To Shift Reaction Networks Out-of-Equilibrium Using Light
Energy
Martin Herder and Jean-Marie Lehn*
́ ́ ́ ́
Institut de Science et d’Ingenierie Supramoleculaires, Universite de Strasbourg, 8 allee Gaspard Monge, 67000 Strasbourg, France
S
* Supporting Information
ABSTRACT: We implement sensitized alkoxyamines as “photo-
dynamic covalent bonds”bonds that, while being stable in the
dark at ambient temperatures, upon photoexcitation efficiently
dissociate and recombine to the bound state in a fast thermal
reaction. This type of bond allows for the photochemically
induced exchange of molecular building blocks and resulting
constitutional variation within dynamic reaction networks. To
this end, alkoxyamines are coupled to a xanthone unit as triplet
sensitizer enabling their reversible photodissociation into two
radical species. By studying the photochemical properties of three
generations of sensitized alkoxyamines it became clear that the nature and efficiency of triplet energy transfer from the sensitizer
to the alkoxyamine bond as well as the reversibility of photodissociation crucially depends on the structure of the nitroxide
terminus. By employing the thus designed photodynamic covalent bonding motif, we demonstrate how to use light energy to
shift a dynamic covalent reaction network away from its thermodynamic minimum into a photostationary state. The network
could be repeatedly switched between its minimum and kinetically trapped out-of-equilibrium state by thermal scrambling and
selective photoactivation of sensitized alkoxyamines, respectively.
states after photoexcitation. Consequently, it is a straightfor-
ward strategy to couple molecular switches to dynamic self-
assembly and self-organization processes and take advantage of
their ability by populating their photostationary state to shift
the whole system out-of-equilibrium using light energy.5 In a
number of beautiful examples this has been achieved by
employing changes of the molecular geometry of the switching
unit to modulate the outcome of self-organization,6 using the
switch as an auxiliary group modulating electronic properties
and thus the kinetics of formation and exchange of a binding
motif,7 or by using the switch as a catalyst allowing for the
spatiotemporal control over the self-organization process.8
Here we present an unprecedented concept to photoinduce
out-of-equilibrium states within a dynamic covalent reaction
network utilizing “photodynamic covalent bonds” instead of
photoswitches. After excitation, these bonds directly undergo
dissociation into reactive building blocks, which in a fast
thermal back reaction recombine forming (new) molecular
entities (Scheme 1a). While there exist many examples for the
photocleavage of covalent bonds, only a few systems are
available that undergo reversible light-induced dissociation, i.e.,
that are able to recombine to the bound state. This principle
applies for light induced bond formation and scission based on
[2 + 2] or [4 + 4] cycloadditions/reversions between coumarin,
thymine, cinnamate, and anthracene derivatives, which have
INTRODUCTION
■
Molecular self-assembly and self-organization processes medi-
ated by supramolecular or dynamic covalent interactions
generally approach a thermodynamic equilibrium with a stable
composition of (supra)molecular entities. Thereby, under given
surrounding conditions (e.g., temperature, pH, or the presence
of a binder) the “fittest” among all possible species is selectively
amplified. Due to its inherent constitutional dynamicity the
system is able to adapt in response to changes of the
surrounding conditions by the expression of other species
and population of new energetic minima.1 However, in living
organisms, many self-organized systems reside in an out-of-
equilibrium state, i.e., they are accessed and held up by
consuming energy, which gives them unique properties that are
inaccessible within the thermodynamic minimum. Thus, for the
synthetic chemist, designing self-organized systems that can be
driven into out-of-equilibrium states is an attractive challenge in
order to increase the complexity of attainable structures and
functions.2 While this can be achieved using chemical fuels3
as is usually the case in biological processesa particularly
advantageous energy source is light due to its noninvasive
character and its high spatial, temporal, and energetic
resolution.
Indeed, a single photoexcited molecule is inherently in a
dissipative out-of-equilibrium state and shows properties very
different from that of the ground state, e.g., regarding its acidity
or redox potentials. Furthermore, photochromic switches and
motors4 allow for the controlled access of metastable ground
Received: April 4, 2018
© XXXX American Chemical Society
A
DOI: 10.1021/jacs.8b03633
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
equilibrium state using light energy (Scheme 2): When mixing
an AOA possessing sensitizers on both the alkyl and nitroxide
Scheme 1. (a) Concept of a Photodynamic Covalent Bond;
(b) Heat Induced Reversible Dissociation of AOAs; (c)
Literature Examples of Sensitized AOAs Utilized for the
Photoinitiation of NMP19f,k,l
Scheme 2. Switchable Dynamic Covalent Reaction Network
Consisting of a Sensitized (SS) and a Nonsensitized (RR)
a
AOA
a
While heat induces formation of the mixed derivatives (RS and SR)
and thus equilibration of the network towards the thermodynamic
minimum, selective activation of sensitized AOAs SS, RS, and SR with
light and the resulting enrichment of the nonsensitized analogue RR
drives it back to the initial out-of-equilibrium state.
termini (SS) with a nonsensitized analogue (RR), thermal
activation and exchange of the AOAs will lead to an equilibrium
situation with the homosubstituted (RR, SS) and mixed (RS,
SR) derivatives being present in equal amounts. However, upon
irradiation of the equilibrium mixture with light only AOAs
carrying a sensitizer on either terminus (SS, RS, SR) are
activated by intramolecular energy transfer and thus undergo
exchange, while the nonsensitized derivative RR stays inactive.
Thus, RR and at the same time due to stoichiometric reasons
its agonist SS accumulate simultaneously in the mixture driving
the overall reaction back to the initial out-of-equilibrium state.
In the present work we develop photodynamic AOAs using
xanthone as the sensitizing unit. Thereby, three generations of
AOAs (Scheme 3) differing in the molecular architecture of the
nitroxide terminus are synthesized and their exchange behavior
is investigated in depth before finally realizing the photo-
dynamic reaction network that can be reversibly switched
between an equilibrium and a kinetically trapped out-of-
equilibrium state by means of thermal and photochemical
activation, respectively.
been utilized in the context of dynamic covalent chemistry.9
Nevertheless, in these systems, the unbound state is in general
thermodynamically more stable and unselective deep-UV light
is needed to induce photodissociation due to the decreased π-
conjugation of the cycloaddition products. Photocleavable
disulfides and diselenides10 as well as hexaarylbiimidazoles11
represent true photodynamic covalent bonds. However, while
the former efficiently respond only to deep-UV irradiation, the
recombination of the latter is rather slow.12 Besides, the leuco
form of triarylmethane dyes13 and specific metal−ligand
interactions14 are able to photodissociate reversibly.
We presumed that photocleavable alkoxyamines (AOAs)
could be a highly interesting alternative: AOAs possess a C-ON
group that is stable at room temperature while it dissociates
into a transient alkyl radical and a persistent nitroxide radical at
elevated temperatures (Scheme 1b).15 Due to the Persistent
Radical Effect16 the homolytic dissociation is reversible in the
absence of oxygen and therefore AOAs could successfully be
applied in numerous thermally operated dynamic covalent
systems.17 In addition, in the context of photoinitiation of
Nitroxide Mediated Polymerizations (NMP)18 it has recently
been shown that by inter- or intramolecular sensitization by a
chromophore, which after excitation undergoes intersystem
crossing to the triplet state, cleavage of the AOA bond may also
be induced photochemically at room temperature19 (see for
example NMP photoinitiators in Scheme 1c). Thereby, the
possibility to use near-UV light for the homolytic cleavage and
the rapid recombination reaction of the radicals are highly
advantageous.
RESULTS AND DISCUSSION
■
Choosing the Sensitizer. In order to photoinduce AOA
dissociation certain requirements have to be fulfilled by the
sensitizer: (i) its selective excitation should be possible in the
presence of AOAs or nitroxides to avoid direct photoexcitation
of these molecules and potential resulting side reactions, (ii) it
should undergo efficient intersystem crossing and its triplet
lifetime should be long enough, and (iii) its triplet energy has
to be sufficiently high to allow for triplet energy transfer to the
AOA bond. Inspection of the UV spectra of a model AOA
PhEt-TEMPOPh as well as the corresponding free nitroxide
TEMPOPh (Figure 1a) reveals that there is a window between
300−400 nm where selective excitation of a sensitizer is
possible. Though the molar absorptivity of nitroxides at
wavelengths >400 nm is very low, still their photoexcitation
should be avoided as they tend to easily undergo proton
abstraction even from very inert molecules in the excited
state.20 Thus, in this work a lamp-filter setup was used to ensure
When utilized in a dynamic covalent reaction network,
photosensitization of AOA dissociation enables the selective
addressing and activation of individual constituents of the
network in dependence of the presence of the sensitizer within
the molecule. Conceptually, this allows for the construction of
reaction networks that can be repeatedly driven into an out-of-
B
DOI: 10.1021/jacs.8b03633
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Scheme 3. Structures of all AOAs Synthesized and Investigated in This Study
irradiation of the samples only with UV light between 310−400
nm.
exchange of the nitroxide moiety of the model AOA PhEt-
TEMPO with another nitroxide TEMPOPh.
In an intermolecular sensitization experiment, it has been
demonstrated19m that the sensitizer’s triplet energy must exceed
a value of around 260 kJ/mol in order to efficiently be
quenched by an AOA and xanthone (λmax = 336 nm, ΦISC = 1,
ET = 310 kJ mol−1)21 was found to be suited. To reproduce this
finding under the experimental conditions employed in the
present work and to test other potentially suited sensitizers, a
simple screening experiment consisting of intermolecular
sensitization of the model AOA PhEt-TEMPO in the presence
of air was set up. When air is present during AOA dissociation
the process becomes irreversible22 as the transient benzyl
radical reacts with molecular oxygen to form the corresponding
ketone, benzyl alcohol, and benzylhydroperoxide while the
nitroxide radical is liberated (Figure 1b). The rate of
decomposition of the parent AOA and the emergence of
First Generation AOAs. On the basis of a one-step
synthesis of AOAs starting from benzyl bromide and tert-
butylnitrone,24 a very fast access to a bis-xanthone substituted
AOA XEt-TXENO was given (Figure 2a). The brominated
xanthone precursor XEtBr was obtained from 2-ethylxanthone
using a mild visible light mediated bromination protocol25
avoiding otherwise predominant elimination and overbromina-
tion reactions (see the SI for all synthetic details). The
architecture of the resulting nitroxide building block TXENO
(tert-butyl-(1-xanthon-2-ylethyl)nitroxide) is closely related to
the well-known TIPNO motif (tert-butyl-(1-isopropyl-1-
phenylmethyl)nitroxide), which is heavily used for NMP26
and thus promises outstanding properties. Furthermore, in
XEt-TXENO the distance between the central C-ON bond and
both xanthone units is minimized in order to ensure optimal
intramolecular triplet energy transfer.19h Note that due to its
two stereogenic centers, XEt-TXENO was obtained as a 1:1
mixture of diastereomers and in the following was used as such.
To realize the reaction network that can be light-driven into
the out-of-equilibrium state (Scheme 2) XEt-TXENO was
combined with an equimolar amount of its nonsensitized,
phenyl substituted analogue PhEt-TPENO and subjected to
heating to 110 °C in a closed tube under oxygen free conditions
to induce thermal scrambling of the AOAs, i.e., to go to the
thermodynamic minimum. The reaction was followed by
UHPLC as due to the presence of diastereomers and rotational
isomers for all involved species and generally rather broad
peaks, 1H NMR spectra could not be interpreted. The
chromatograms (Figure 2b) show that after 24 h of heating
as expected ca. 50% of exchange products with mixed xanthone
and phenyl substitution were formed. Unfortunately, no precise
quantification could be performed due to poor peak separation
even under most optimized UHPLC conditions. Upon
subsequent irradiation of the mixture with UV light the relative
amount of the exchange products compared to XEt-TXENO
clearly decreases. However, in an absolute sense there is only a
slight increase of the peak areas corresponding to XEt-TXENO
observed before all peak areas slowly decrease with prolonged
1
oxidation products was followed using H NMR spectroscopy.
This and all following experiments were conducted in benzene
as preliminary tests on AOA sensitization showed that in other
solvents such as dioxane, methanol, methylene chloride, or
acetonitrile small amounts of undesired byproducts due to
reactions with the solvent are formed. Indeed, during the
screening it turned out that xanthone is by far the most efficient
sensitizer among the molecules tested giving 45% AOA
decomposition after 4 h of irradiation. Notably, benzophenone,
which was utilized before to photoinitiate NMP,19 gave only 7%
decomposition whereas 3-acetylcoumarin performed slightly
better with 10% decomposition.23
Consequently, in the following xanthone is used as the
sensitizing unit. The incorporation into different AOA and
nitroxide structures only marginally altered its absorption
properties (see UV spectra in Figures S1 and S2). In order to
prove its ability to sensitize not only AOA decomposition but
also exchange reactions an intermolecular sensitization experi-
ment under oxygen free conditions was followed using UHPLC
(Figure S3). Thereby it could be demonstrated that 2-
ethylxanthone as well as methyl xanthone-2-carboxylate, both
mimicking those xanthone motifs used later on within AOA
structures (compare Scheme 3), are able to sensitize the
C
DOI: 10.1021/jacs.8b03633
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 2. (a) One-step synthesis of first generation AOAs. (b)
UHPLC traces (diode array detector at 336 nm) of samples taken
from an equimolar mixture of XEt-TXENO and PhEt-TPENO in
benzene (10 mM, nitrogen atmosphere) during heating (110 °C) and
subsequent irradiation with UV light (rt). While the two diastereomers
ds1 and ds2 of XEt-TXENO show up as two well separated peaks,
diastereomers of the exchange products, being indistinguishable due to
their identical m/z ratio and UV spectra, show up as two nonbaseline
resolved peaks and a third peak fully overlapping with that of the
starting material. Note the doubled molar absorptivity of XEt-TXENO
at the detection wavelength as compared to the other xanthone
containing compounds.
Figure 1. (a) UV−vis spectra of prototype AOA PhEt-TEMPOPh,
nitroxide TEMPOPh, and xanthone in acetonitrile (25 °C). The
dashed line represents a 100-fold magnification of the spectrum of
TEMPOPh. The dotted line shows the transmission spectrum of the
filter combination used together with a 125 W high-pressure mercury
lamp for irradiation experiments. (b) Intermolecular sensitization of
photochemical AOA decomposition using different sensitizers:
Normalized peak areas corresponding to the AOA’s benzylic CH
proton (filled symbols) and the CH3 group of the formed ketone
bly, free TXENO radical was detected only as a minimum trace
in all chromatograms recorded during the decomposition
reaction. This stands in stark contrast to other photo-
decomposition experiments on AOAs (vide infra) where
generally the liberated nitroxide radical is stable and can be
detected by UHPLC. As TIPNO as well as its methyl
substituted analogue TPENO are expected to be thermally
stable at room temperature27 this finding hints to a pronounced
photochemical decomposition of the TXENO moiety that
prevents the realization of the desired network. Consequently, a
redesign of the system is necessary, which also should avoid the
formation of diastereomers in order to simplify the analysis of
the reaction networks.
Second Generation AOAs. To ensure maximum stability
of the nitroxide building blocks an AOA design based on the
well-known TEMPO (2,2,6,6-tetramethylpiperidinenitroxide)
moiety was chosen. The synthesis of second generation AOAs
is accomplished by generating benzylic radicals in the presence
of a free TEMPO nitroxide.28 Thereby, 4-hydroxy-TEMPO
serves as a building block that is easily prefunctionalized with
the xanthone sensitizer by esterification. Note that despite the
relatively large spatial separation between the AOA bond and
1
(open symbols) as obtained from H NMR spectra of an equimolar
mixture of PhEt-TEMPO and a sensitizer in aerated C6D6 (c = 20
mM) during irradiation with UV light. Peak areas are normalized to
the initial value of the benzylic CH proton and to the number of
corresponding protons. The blank sample was irradiated without
adding any sensitizer.
irradiation time. This is due to the formation of a number of by
products containing the xanthone chromophore upon UV
irradiation, whose peaks arise at tR < 1 min. Thus, overall
photodegradation of the system prevents reestablishing the
initial out-of-equilibrium state by light irradiation.
In order to gain more insight into the reason for
photodegradation XEt-TXENO was also subjected to photo-
chemical decomposition in the presence of air (Figure S4).
After 30 min of irradiation the starting material was completely
decomposed and the corresponding ketone (2-acetylxanthone)
was identified as the main reaction product besides smaller
amounts of the benzyl alcohol (1-xanthon-2-ylethanol) as well
as the amine N-tert-butyl-N-(1-xanthon-2-yl)ethylamine. Nota-
D
DOI: 10.1021/jacs.8b03633
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
the sensitizer attached to the TEMPO building block for very
similar AOA-sensitizer architectures reported before (Scheme
1c), the emergence of an intramolecular triplet energy transfer
after photoexcitation was postulated.19d,h,k,l
To test for the ability of second generation AOAs to undergo
the desired thermal and photochemical exchange reactions, the
bis-functionalized derivative XEt-TEMPOX was mixed with an
excess (3 equiv) of its nonsensitized analogue PhEt-TEMPO
and heated to 100 °C in the absence of oxygen (Figure 3). After
synthesized allowing for a detailed investigation of the
photochemical behavior of second generation AOAs in
dependence of the position of the sensitizer. Both compounds
were subjected to photodecomposition to the ketone and
nitroxide in the presence of air (Figure S6). It becomes
immediately evident that the photodecomposition of XEt-
TEMPO is much faster than the reaction of PhEt-TEMPOX.
For the former an effective quantum yield29 of 0.68 was
determined, demonstrating the high efficiency of the photo-
cleavage process when the sensitizer is situated on the alkyl
terminus.19a At a constant photon flux the time needed to reach
complete photodecomposition of XEt-TEMPO shortens when
reducing the concentration while for PhEt-TEMPOX there is
no significant change. The latter indicates that the rate of the
decomposition reaction after excitation of PhEt-TEMPOX
decreases when lowering its concentrations. Another indication
for a fundamentally different photochemical behavior of the
two compounds is obtained when performing the photo-
decomposition in the presence of excess naphthalene, an
efficient quencher of the triplet state of xanthone.30 While for
XEt-TEMPO there is no significant influence on the
decomposition rate, in the case of PhEt-TEMPOX the reaction
is almost completely inhibited. Taken together these findings
indicate that while the former undergoes the expected efficient
intramolecular triplet energy transfer after excitation of the
sensitizer, in the latter case dissociation of the AOA bond is
only achieved via a bimolecular process, i.e., intermolecular
energy transfer.
To confirm this assumption, initial reaction rates were
measured for the photodecomposition of both compounds at
varying concentrations. This was done by monitoring the
disappearance of the AOA’s characteristic CH−ON signal
(Figure 4) and the emergence of a signal for the resulting
ketone’s CH3 group (Figure S13) using 1H NMR spectroscopy
under precisely controlled irradiation conditions (for photo-
decomposition NMR spectra and further details of the
measurement see Figures S9−S12 in the SI). It becomes
evident that for XEt-TEMPO the initial rate of the reaction is
Figure 3. Attempted switching of the reaction network consisting of
second generation sensitized AOAs: UHPLC traces (diode array
detector at 336 nm) of samples taken from a mixture of XEt-
TEMPOX (10 mM) and PhEt-TEMPO (30 mM) in benzene after
heating (100 °C) and subsequent irradiation with UV light (rt) under
oxygen free conditions.
26 h of heating exchange to the mixed derivatives XEt-TEMPO
and PhEt-TEMPOX was observed by UHPLC with a
conversion of 61%. Importantly, no other byproducts were
detected. At this point heating was stopped and the mixture was
subjected to UV irradiation with the expectation to reverse the
direction of the exchange reaction (vide supra). Surprisingly,
the reaction went on into the forward direction, i.e., the mixed
exchange products XEt-TEMPO and PhEt-TEMPOX con-
tinued to form, until the reaction leveled off at a conversion of
approximately 85% after 130 min of irradiation. Notably, when
performing the same irradiation experiment on the mixture of
XEt-TEMPOX and PhEt-TEMPO without the initial heating
step, again the clean formation of the exchange products is
observed with a conversion of 87% after 120 min of irradiation
(Figure S5). This finding strongly suggests that there is an
indirect pathway to photocleave the nonsensitized AOA PhEt-
TEMPO, which does not absorb the employed UV light, as
otherwise a solely photochemically induced exchange could not
proceed. As a consequence, it is not possible to reverse the
direction of the exchange reaction as PhEt-TEMPO cannot be
accumulated.
Figure 4. Initial reaction rates for the photodecomposition of model
AOAs of the second and third generation as a function of AOA
1
concentration in aerated C6D6 as determined by H NMR spectros-
copy. Monitoring signals (XEt-TEMPO: CHarom at δ = 7.46 ppm,
PhEt-TEMPOX: PhEt-CH3 at δ = 1.46 ppm, PhEt-TMIOX: TMIOX-
CH3 at δ = 1.40 ppm) were selected to avoid overlap with other signals
of the AOA or photoproducts. Initial rates for PhEt-TEMPOX were
multiplied with a factor of 60 for the sake of clarity.
To explain this finding the two mixed AOAs XEt-TEMPO
and PhEt-TEMPOX, each possessing only one xanthone group
at the alkyl or the nitroxide terminus, respectively, were
E
DOI: 10.1021/jacs.8b03633
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
independent of the concentration and thus it follows a
unimolecular reaction scheme including intramolecular triplet
energy transfer from the sensitizer to the AOA bond (see
section 9 of the SI for a detailed analysis of rate equations). In
contrast, for PhEt-TEMPOX a linear dependence of the initial
reaction rate with the concentration is observed, which proves
that the energy transfer between the excited sensitizer and the
AOA bond is a bimolecular process. Thus, it can be concluded
that the distance between the sensitizer and the AOA bond is
too large to allow for an intramolecular energy transfer.
As a consequence, upon mixing PhEt-TEMPOX with a
nonsensitized AOA Cum-TEMPOPh under oxygen free
conditions and applying UV irradiation, an exchange reaction
occurs (Figure S7b), as due to its intermolecular nature, there is
a similar probability that after excitation of the sensitizer the
energy is transferred to another molecule PhEt-TEMPOX or
Cum-TEMPOPh. In contrast, when using XEt-TEMPO
instead of PhEt-TEMPOX only small traces of the exchange
product are observed (Figure S7a) as the intramolecular energy
transfer within XEt-TEMPO is much faster than any
intermolecular process. A similar behavior is expected for the
bis-sensitizer substituted compound XEt-TEMPOX (Scheme
4) used in the reaction network. Due to the fact that the
possessing the xanthone moiety at the alkyl radical terminus
readily cleave at room temperature under photoirradiation.
Thus, by combining these different types of AOAs and by
choosing the appropriate conditions the order in which the
individual AOAs undergo exchange can be precisely controlled
and thus a dynamic reaction network can adapt multiple
different states on the way toward the global thermodynamic
minimum. To demonstrate this, a mixture of Cum-TEMPO
and XEt-TEMPO with an excess (3 equiv) of another free
nitroxide TEMPOPh(OMe) was first heated to 110 °C and
afterward subjected to UV irradiation (Figure 5, left). Upon
Scheme 4. Energy Transfer Pathways in Second Generation
Sensitized AOAs
TEMPOX unit transfers its triplet energy after excitation in an
intermolecular fashion also activating the fully nonsensitized
constituent of the reaction network its accumulation and thus
the light-driven population of the out-of-equilibrium state is
prevented.
Interestingly, the finding that the sensitizer substituted to the
TEMPO moiety undergoes exclusively intermolecular energy
transfer to the AOA bond stands in contrast to assumptions
made in the literature. In particular, Goto et al.19k reported
unimolecular photodissociation kinetics of the related com-
pound PhEt-TEMPOQ (see Scheme 1c) sensitized by a
quinoline unit in acetonitrile. To verify our results, we also
subjected PhEt-TEMPOQ to the photodecomposition reaction
and monitored its concentration dependence (Figure S8). It is
clear that with the conditions used in this work (benzene as the
solvent; for the influence of acetonitrile vide infra) the
photodissociation behavior of PhEt-TEMPOQ is very similar
to that of PhEt-TEMPOX undergoing intermolecular energy
transfer with bimolecular reaction kinetics.
Though second generation AOAs do not allow for the
reversible light-induced population of an out-of-equilibrium
state in a dynamic reaction network, they still possess the
interesting property that their photochemical and thermal
dissociation reactions are highly orthogonal:31 ethylbenzene
derived AOAs like PhEt-TEMPO and XEt-TEMPO undergo
thermal exchange reactions only at high temperatures (>100
°C) whereas cumene derived AOAs like Cum-TEMPO already
cleave at moderate heating (>60 °C) and finally AOAs
Figure 5. Orthogonal exchange of second generation AOAs as
followed by 1H NMR spectroscopy: Consecutive heating (60 °C, 9 h)
and UV irradiation (rt, 30 min) steps of a mixture of XEt-TEMPO and
Cum-TEMPO in C6D6 (5 mM, oxygen-free conditions) in the
presence of three equivalents of nitroxide TEMPOPh(OMe). Peak
areas are normalized to their initial values. Lines are drawn as guide for
the eye.
thermal treatment only the cumene derived AOA undergoes
exchange of its nitroxide moiety forming Cum-TEMPOPh-
(OMe) until reaching the expected equilibrium situation.
Thereby the other exchange product XEt-TEMPOPh(OMe)
was not observed at all. During subsequent UV irradiation now
exclusively XEt-TEMPO undergoes exchange, while the
amount of Cum-TEMPO and Cum-TEMPOPh(OMe)
remains unchanged. The order of the activation stimuli may
be reversed (Figure 5, right) with the same selectivity observed.
The initial photoactivation seems to be slightly less selective as
a small amount of the nondesired exchange product Cum-
TEMPOPh(OMe) is formed (7%). This hints to a partial
involvement of intermolecular sensitization processes either as
F
DOI: 10.1021/jacs.8b03633
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
side reaction of XEt-TEMPO or by traces of xanthone
containing decomposition products formed (vide infra).
a sealed NMR tube under oxygen free conditions. By careful
deconvolution of the NMR spectra all species involved in the
reaction could be identified and their relative quantities
Third Generation AOAs. As it became clear that the
distance between the C-ON bond and the sensitizer on the
nitroxide terminus has to be reduced to allow for an
intramolecular energy transfer, a new nitroxide motif, as stable
as the TEMPO moiety, was needed. The nitroxide TMIO32
(1,1,3,3-tetramethylisoindolineoxide) is reported to possess
similar electronic and steric properties compared to TEMPO,
shows excellent stability,33 and contains an aromatic benzene
ring that could be extended to the desired xanthone moiety in
close proximity to the N−O bond. Thus, to access the desired
structure, TMIO was iodinated twice34 in order to apply a Pd-
catalyzed xanthone formation protocol, originally reported on
1,2-dibromobenzene as substrate,35 forming the xanthone
containing nitroxide TMIOX (Scheme 5). Generation of
1
determined (Figure 6b and 6d, for full H NMR spectra see
Scheme 5. Synthesis of the Xanthone Derived Nitroxide
TMIOX and Third Generation Sensitized AOA XEt-TMIOX
benzylic radicals in the presence of TMIOX gave the bis-
xanthone functionalized third generation AOA XEt-TMIOX as
well as the single-substituted compound PhEt-TMIOX.
The ability of the TMIOX motif to intramolecularly transfer
the triplet energy after excitation was tested as before by
investigating the kinetics of photodecomposition of PhEt-
TMIOX in the presence of air. Following the reaction by
1
UHPLC as well as H NMR spectroscopy (Figures S14 and
S15) gives a very similar picture to the behavior of XEt-
TEMPO. In particular, initial reaction rates are independent of
the concentration (Figure 4) clearly speaking for a predom-
inantly intramolecular triplet energy transfer from the TMIOX
moiety to the AOA bond. Yet, it can be deduced that the
reaction is considerably slower by a factor of 2 as compared to
XEt-TEMPO. Interestingly, and in contrast to XEt-TEMPO,
upon addition of naphthalene the photodissociation is
quenched to a certain extent (Figure S14). This indicates that
due to the slightly larger distance between sensitizer and AOA
bond within the TMIOX moiety the rate of triplet energy
transfer is somewhat smaller compared to the sensitizer located
on the alkyl terminus. This is also reflected in the lower
effective quantum yield for the photocleavage of PhEt-TMIOX
of 0.37. Nevertheless, the compound can still be selectively
photoaddressed in the presence of other AOAs as demon-
strated by irradiating a mixture with Cum-TEMPOPh in the
absence of oxygen (Figure S16). Only small amounts (5%) of
the exchange product are formed.
Figure 6. Realization of the photodynamic reaction network using
third generation sensitized AOAs: Sealed NMR tubes containing
degassed solutions of equimolar amounts of XEt-TMIOX and PhEt-
TMIO in C6D6 (5 mM) were heated (110 °C, 44 h) or irradiated with
UV light (rt, 330 min), as indicated. (a) Reaction scheme. (b)
Evolution of the network during consecutive long-time heating and
irradiation. Relative peak areas of the starting material and reaction
product obtained by deconvolution of 1H NMR spectra as well as the
normalized overall integral of the overlapping signals for the benzylic
CH protons are shown. (c) Evolution of the network upon irradiation
only. (d) Selected spectral regions utilized for the quantification by
deconvolution (for full NMR spectra see SI).
Figure S17, for a plot of absolute quantities see Figure S18). As
expected, upon heating smooth exchange to XEt-TMIO and
PhEt-TMIOX occurs until reaching the thermodynamic
minimum at a 1:1 mixture of starting materials and exchange
products. Thereby, a longer reaction time of 44 h was necessary
due to the slightly higher thermal barrier for the dissociation of
Taking advantage of these promising properties of the
TMIOX building block, the desired reaction network could
eventually be realized: equimolar amounts of XEt-TMIOX and
nonsensitized PhEt-TMIO were mixed and heated to 110 °C in
G
DOI: 10.1021/jacs.8b03633
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
TMIO AOAs in comparison to previous experiments using
TEMPO.33 Markedly, upon subsequent UV irradiation fast
regeneration of XEt-TMIOX and PhEt-TMIO at the expense
of the mixed AOAs is observed. Upon extending the UV
irradiation up to 330 min a photostationary state consisting of a
91:9 mixture of starting materials and mixed compounds
evolves. From that point on no further shift of the equilibrium
is observed and photodecomposition of the involved species
(vide infra) becomes the major process. Notably, when starting
with the same system and subjecting it to UV irradiation only, a
similar photostationary state of 93:7 starting materials to
exchange products is formed (Figure 6c). Obviously, a
quantitative back reaction to the initial state cannot be obtained
due to partial intermolecular activation of the nonsensitized
AOA within the reaction network. This may be caused by a
partial intermolecular energy transfer of the TMIOX motif
(vide supra) and the presence of small amounts of
decomposition products possessing a xanthone unit.
order to increase the dynamic range of the switching process,
the amount of nonsensitized AOA was increased to three
equivalents (Figure S20). As expected from the stoichiometry,
thermal scrambling of the mixture gives a ratio of XEt-TMIOX:
XEt-TMIO of 28:72 while the UV induced back reaction levels
off at a 79:21 mixture.
Limits of the Reaction Network. To get further insight
we investigated the emergence of byproducts under thermal
and photochemical activation depending on the position of the
sensitizer and the substitution pattern using model AOAs. The
principal findings are summarized in Scheme 6 (see Section 8
Scheme 6. Thermal and Photochemical Side Reactions of
Sensitized AOAs
Indeed, monitoring of the overall signal for the benzylic CH
protons reveals a slow decrease of the total amount of AOAs
present in the sample during switching the network between
the two states. As major side products two styrenes and two
ketones, both stemming from the PhEt or XEt building blocks,
respectively, were identified in the NMR spectra (Figure S19)
as well as in UHPLC analysis of the mixture at the end of the
reaction. UHPLC also proved the presence of free TMIOX
nitroxide as well as the corresponding reduced hydroxylamine.
Notably, both styrenes evolve during the thermal activation
step, while the ketones start to form upon UV irradiation of the
sample. During the UV step the amount of styrene1, stemming
from the PhEt moiety, does not change anymore, while
styrene2, stemming from the XEt moiety continues to form.
Despite the slow decomposition of the sample, several cycles
of switching of the reaction network between the equilibrium
and out-of-equilibrium state could be performed (Figure 7).
For this experiment heating and irradiation times were
restricted to 24 h and 60 min, respectively, to minimize
degradation. After four switching cycles, during which the
relative amounts of starting material and exchange products
were fully reproduced in the two states of the reaction network,
still 81% of the initial total amount of AOAs was present. In
of the SI for detailed data). Notably, upon thermal activation of
PhEt-TEMPO and PhEt-TMIO styrene and the corresponding
hydroxylamines were detected as sole and quantitatively formed
decomposition products (Scheme 6a, Figure S21). Thereby the
disproportionation reaction of PhEt-TMIO is significantly
slower than that of PhEt-TEMPO. It is known for TEMPO
based AOAs that substitution of the ethylbenzene fragment at
the alkyl chain can diminish disproportionation.36 Thus,
PhEt(OBz)-TEMPO was synthesized and tested. Though
during its thermal activation the corresponding styrene (as both
E and Z isomer) was still observed, the disproportionation
reaction was much slower than for PhEt-TEMPO (Figure
S21). To check if this strategy could increase the fatigue
resistance of the reaction network during thermal activation,
the benzoyloxy substituted third generation AOA XEt(OBz)-
TMIOX was synthesized. Indeed, upon thermal equilibration of
its mixture with nonsensitized PhEt(OBz)-TEMPO the
according exchange products formed quantitatively and no
styrene was detected at all (Figure S22). Unfortunately, upon
Figure 7. Repeated switching of the reaction network shown in Figure
6 (C6D6, 5 mM, oxygen free conditions) using shorter heating (Δ, 24
h) and irradiation (UV, 60 min) steps. Relative peak areas as obtained
by deconvolution of specific spectral regions and the overall benzylic
CH signal area are shown. Lines are drawn as guide for the eye.
H
DOI: 10.1021/jacs.8b03633
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
subsequent photoactivation of the reaction network as well as
irradiation of XEt(OBz)-TMIOX alone fast decomposition of
the AOAs occurred. Investigation of decompositions products
(Figures S23−S24) hints to a photoinduced beta cleavage
facilitated by the OBz group as the major decomposition
pathway (Scheme S5).
We demonstrated that, using photochemical and thermal
stimuli, such photodynamic covalent reaction network can be
reversibly switched (multiple times) between its initial out-of-
equilibrium state and the thermally equilibrated thermodynamic
state.
In the present state, the driving force for the network to react
toward the thermodynamic minimum consists mainly in the
entropy of mixing, i.e., the depths of the potential wells on the
left and the right side of the equilibrium are approximately the
same (Figure 8). This results from the fact that all constituents
Gratifyingly XEt-TEMPO and PhEt-TMIOX are much more
resistant toward photodegradation. However, their stability
under continuous irradiation in the absence of oxygen, i.e.,
while constantly undergoing reversible photocleavage, is highly
dependent on the position of the sensitizing group within the
AOA architecture. For XEt-TEMPO photodegradation is
limited to slow radical disproportionation (Scheme 6b)
resulting in a loss of 10% of the starting material after 4 h of
irradiation (Figure S25a). In contrast, for PhEt-TMIOX
photodegradation is significantly faster with a loss of 30%
after the same irradiation time (Figure S25b). Thereby,
acetophenone was formed quantitatively and a secondary
amine stemming from reduction of the TMIOX moiety was
detected by UHPLC analysis of the reaction mixture. This
strongly indicates that during photoactivation of PhEt-TMIOX
the desired cleavage of the C-ON bond competes with the
cleavage of the CO-N bond leading to stoichiometric formation
of the ketone and amine (Scheme 6c). The latter reaction is a
typical side-reaction of AOAs,37 in particular with photo-
sensitizers placed on the nitroxide terminus.38 In case of PhEt-
TMIOX we estimated relative rates of 100:1 for the two
reaction pathways by comparing its photodecomposition in
presence and absence of oxygen.
Photoirradiation of model AOAs was also carried out using
acetonitrile39 as solvent. Thereby, XEt-TEMPO and PhEt-
TMIOX did not show any significant difference to pure
benzene regarding their photostability and photodegradation in
absence and presence of oxygen, respectively (Figure S26). In
contrast, PhEt-TEMPOX decomposed much faster in acetoni-
trile than in benzene and the rate was independent of the
presence of oxygen (Figure S27a). The formation of N-(1-
phenylethyl)acetamide as reaction product (Figure S27b)
indicates a change of the photosensitization mechanism from
triplet energy transfer to photoelectron transfer when going
from benzene to acetonitrile.20a This leads to the release of a
benzyl cation,40 which reacts with acetonitrile in a Ritter type
reaction41 to yield the acetamide product (Scheme S6). The
differing sensitization mechanisms in benzene and acetonitrile
presumably also explain the contrary photokinetics of PhEt-
TEMPOQ observed by Goto et al.19k and us.
Figure 8. Thermodynamic picture of the photodynamic reaction
network showing the generation of an out-of-equilibrium state under
irradiation (bottom). Black arrows indicate the probability of thermal
dissociation and recombination reactions while red arrows indicate the
probability of photoexcitation from either side of the equilibrium. As
photoexcitation is selective for constituents possessing a sensitizer
(depicted in red) the amount of nonsensitized building blocks
reaching the intermediate high energy region depletes over time
leading to a higher probability for the formation of homosubstituted
constituents and thus a net flow of material from the right to the left
side of the equilibrium under irradiation.
of the reaction network possess the same AOA linkage with
similar steric and electronic properties. Upon heating the
relatively large barrier between the potential wells can be
overcome, i.e., radical species are formed as high energy
intermediates, and the reaction toward exchange products
proceeds until the net-flow of material from both potential
wells is equal, representing the thermodynamic minimum state.
This state is frozen at room temperature. However, photo-
chemical activation of sensitized AOAs allows population of the
high-energy intermediate region via the triplet excited state of
the molecules. While the radical intermediates recombine
statistically to form the reaction products of either side of the
equilibrium, due to the fact that only AOAs possessing a
sensitizer can undergo photoexcitation the amount of non-
sensitized building blocks that reach the intermediate region
depletes over time. This results in a net flow of material from
the right to the left side of the equilibrium until a
photostationary state has been reached, i.e., the (re)generation
of the kinetically trapped out-of-equilibrium state. The
selectivity for specific constituents of the network, i.e., the
dependency of photoactivation on the “information” if a
sensitizer is present in the molecule, allows for its shifting out-
CONCLUSION
■
By utilizing photosensitized reversible dissociation of AOAs
into their alkyl radical and nitroxide building blocks we were
able to develop a “photodynamic covalent bond” that is stable
in the absence of light and can only be activated at elevated
temperatures, while undergoing efficient and controlled
dissociation upon UV irradiation at room temperature, thus
inducing the exchange of building blocks and the formation of
new constituents within a dynamic covalent reaction network.
When AOAs are sensitized in a solely intramolecular fashion,
which has been achieved in this work by the structural
optimization of the nitroxide terminus and detailed analysis of
photodissociation kinetics, specific constituents of the network
can be selectively addressed by irradiation, while nonsensitized
analogues accumulate. This results in a light-driven population
of states different from the overall thermodynamic minimum.
I
DOI: 10.1021/jacs.8b03633
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
of-equilibrium without an intermediate change of the relative
ground state energies of the two states. This is reminiscent to
Leigh’s molecular information ratchet.42 Thereby, the detour
over the excited state and subsequent relaxation is necessary to
allow for the reduction of the entropy of the system during this
process.
REFERENCES
■
(1) (a) Lehn, J.-M. Angew. Chem., Int. Ed. 2015, 54, 3276−3289.
(b) Lehn, J.-M. Angew. Chem., Int. Ed. 2013, 52, 2836−2850. (c) Lehn,
J.-M. Chem. Soc. Rev. 2007, 36, 151−160.
(2) (a) Sorrenti, A.; Leira-Iglesias, J.; Markvoort, A. J.; de Greef, T. F.
A.; Hermans, T. M. Chem. Soc. Rev. 2017, 46, 5476−5490. (b) van
Rossum, S. A. P.; Tena-Solsona, M.; van Esch, J. H.; Eelkema, R.;
Boekhoven, J. Chem. Soc. Rev. 2017, 46, 5519−5535. (c) Mattia, E.;
Otto, S. Nat. Nanotechnol. 2015, 10, 111−119.
(3) Boekhoven, J.; Hendriksen, W. E.; Koper, G. J. M.; Eelkema, R.;
van Esch, J. H. Science 2015, 349, 1075−1079.
(4) Feringa, B. L. Angew. Chem., Int. Ed. 2017, 56, 11060−11078.
(5) Kathan, M.; Hecht, S. Chem. Soc. Rev. 2017, 46, 5536−5550.
(6) (a) Vantomme, G.; Hafezi, N.; Lehn, J.-M. Chem. Sci. 2014, 5,
With the presented network it is not possible to selectively
access the right side of the equilibrium as both states are equally
populated in the thermodynamic minimum. However, a larger
driving force for the formation of exchange products can be
implemented by increasing the difference in free Gibbs energy
between the two states, e.g., by reducing the steric congestion
of the exchange products or by incorporating the homosub-
stituted AOA linkages within the main chain of oligomeric or
polymeric structures and thus increasing the entropic cost for
the population of the left side of the equilibrium. In particular
the latter case is highly interesting from the point of view of
dynamic materials: one could design a network consisting of
monomeric species in the thermodynamic minimum while their
oligo- and polymerization would be driven by light on going to
the out-of-equilibrium state. Moreover, in photodynamic
covalent bonds the photoexcitation and energy transfer steps
are directly linked to bond cleavage and exchange. Thus,
photodynamic covalent libraries could be constructed and
selection of specific constituents could be installed not on the
basis of binding to a target molecule, as it is classically done in
dynamic combinatorial chemistry, but based on the photo-
chemical and photophysical properties of the individual
constituents, such as the efficiency of energy transfer.
Furthermore, whereas the present AOA system allows for
access to a kinetically trapped out-of-equilibrium state with a
rather high thermal barrier toward the energetic minimum, it
could in principle be turned into a dissipative system by
implementing AOAs, which undergo thermal exchange at or
slightly above room temperature. Research in these directions is
currently ongoing in our laboratory.
1475−1483. (b) Barrell, M. J.; Campana, A. G.; von Delius, M.;
̃
Geertsema, E. M.; Leigh, D. A. Angew. Chem., Int. Ed. 2011, 50, 285−
290. (c) Ingerman, L. A.; Waters, M. L. J. Org. Chem. 2009, 74, 111−
117. (d) Klajn, R.; Wesson, P. J.; Bishop, K. J. M.; Grzybowski, B. A.
Angew. Chem., Int. Ed. 2009, 48, 7035−7039. (e) Eelkema, R.; Pollard,
M. M.; Vicario, J.; Katsonis, N.; Ramon, B. S.; Bastiaansen, C. W. M.;
Broer, D. J.; Feringa, B. L. Nature 2006, 440, 163.
(7) (a) Kathan, M.; Kovarí
̌ ̌
cek, P.; Jurissek, C.; Senf, A.; Dallmann,
A.; Thunemann, A. F.; Hecht, S. Angew. Chem., Int. Ed. 2016, 55,
̈
13882−13886. (b) Fuhrmann, A.; Gostl, R.; Wendt, R.; Kotteritzsch,
̈
̈
J.; Hager, M. D.; Schubert, U. S.; Brademann-Jock, K.; Thunemann, A.
̈
F.; Nochel, U.; Behl, M.; Hecht, S. Nat. Commun. 2016, 7, 13623.
̈
(8) (a) Maity, C.; Hendriksen, W. E.; van Esch, J. H.; Eelkema, R.
Angew. Chem., Int. Ed. 2015, 54, 998−1001. (b) Kundu, P. K.;
Samanta, D.; Leizrowice, R.; Margulis, B.; Zhao, H.; Borner, M.;
Udayabhaskararao, T.; Manna, D.; Klajn, R. Nat. Chem. 2015, 7, 646−
̈
652.
(9) (a) Frisch, H.; Marschner, D. E.; Goldmann, A. S.; Barner-
Kowollik, C. Angew. Chem., Int. Ed. 2018, 57, 2036−2045. (b) Claus,
T. K.; Telitel, S.; Welle, A.; Bastmeyer, M.; Vogt, A. P.; Delaittre, G.;
Barner-Kowollik, C. Chem. Commun. 2017, 53, 1599−1602. (c) He,
H.; Feng, M.; Chen, Q.; Zhang, X.; Zhan, H. Angew. Chem., Int. Ed.
2016, 55, 936−940. (d) Tron, A.; Thornton, P. J.; Lincheneau, C.;
Desvergne, J.-P.; Spencer, N.; Tucker, J. H. R.; McClenaghan, N. D. J.
Org. Chem. 2015, 80, 988−996. (e) Kaur, G.; Johnston, P.; Saito, K.
Polym. Chem. 2014, 5, 2171−2186. (f) Zhang, Q.; Qu, D.-H.; Wu, J.;
Ma, X.; Wang, Q.; Tian, H. Langmuir 2013, 29, 5345−5350.
(10) (a) Klepel, F.; Ravoo, B. J. Org. Biomol. Chem. 2017, 15, 3840−
3842. (b) Li, L.; Feng, W.; Welle, A.; Levkin, P. A. Angew. Chem., Int.
Ed. 2016, 55, 13765−13769. (c) Shaobo, J.; Wei, C.; Ying, Y.;
Huaping, X. Angew. Chem., Int. Ed. 2014, 53, 6781−6785. (d) Li, J.;
Carnall, J. M. A.; Stuart, M. C. A.; Otto, S. Angew. Chem., Int. Ed. 2011,
50, 8384−8386. (e) Otsuka, H.; Nagano, S.; Kobashi, Y.; Maeda, T.;
Takahara, A. Chem. Commun. 2010, 46, 1150−1152.
(11) (a) Ahn, D.; Zavada, S. R.; Scott, T. F. Chem. Mater. 2017, 29,
7023−7031. (b) Kimoto, A.; Niitsu, S.; Iwahori, F.; Abe, J. New J.
Chem. 2009, 33, 1339−1342. (c) Kikuchi, A.; Iyoda, T.; Abe, J. Chem.
Commun. 2002, 1484−1485.
(12) Sathe, S. S.; Ahn, D.; Scott, T. F. Ind. Eng. Chem. Res. 2015, 54,
4203−4212.
(13) (a) Huajie, L.; Yun, X.; Fengyu, L.; Yang, Y.; Wenxing, W.;
Yanlin, S.; Dongsheng, L. Angew. Chem., Int. Ed. 2007, 46, 2515−2517.
(b) Irie, M. J. Am. Chem. Soc. 1983, 105, 2078−2079.
ASSOCIATED CONTENT
■
S
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/jacs.8b03633.
Details on instrumental methods, synthesis and charac-
terization of all compounds, UV−vis spectra, and further
photochemical studies (PDF)
AUTHOR INFORMATION
■
Corresponding Author
*lehn@unistra.fr
ORCID
Jean-Marie Lehn: 0000-0001-8981-4593
(14) (a) Bahreman, A.; Limburg, B.; Siegler, M. A.; Koning, R.;
Koster, A. J.; Bonnet, S. Chem. - Eur. J. 2012, 18, 10271−10280.
(b) Burnworth, M.; Tang, L.; Kumpfer, J. R.; Duncan, A. J.; Beyer, F.
L.; Fiore, G. L.; Rowan, S. J.; Weder, C. Nature 2011, 472, 334.
(c) Yamashita, K.-i.; Kawano, M.; Fujita, M. J. Am. Chem. Soc. 2007,
129, 1850−1851. (d) Mobian, P.; Kern, J.-M.; Sauvage, J.-P. Angew.
Chem., Int. Ed. 2004, 43, 2392−2395.
(15) (a) Audran, G.; Bremond, P.; Marque, S. R. A. Chem. Commun.
2014, 50, 7921−7928. (b) Studer, A. Chem. Soc. Rev. 2004, 033, 267−
273.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
M.H. thanks the Alexander von Humboldt Foundation for
granting a postdoctoral fellowship. We thank the ERC
(Advanced Research Grant SUPRADAPT 290585) for financial
support.
(16) Fischer, H. Chem. Rev. 2001, 101, 3581−3610.
J
DOI: 10.1021/jacs.8b03633
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(17) (a) Wessely, I.; Mugnaini, V.; Bihlmeier, A.; Jeschke, G.; Brase,
S.; Tsotsalas, M. RSC Adv. 2016, 6, 55715−55719. (b) Zhang, Z. P.;
Lu, Y.; Rong, M. Z.; Zhang, M. Q. RSC Adv. 2016, 6, 6350−6357.
(c) Otsuka, H. Polym. J. 2013, 45, 879−891. (d) Zhang, Z. P.; Rong,
M. Z.; Zhang, M. Q.; Yuan, C. Polym. Chem. 2013, 4, 4648−4654.
(e) Blinco, J. P.; Fairfull-Smith, K. E.; Micallef, A. S.; Bottle, S. E.
Polym. Chem. 2010, 1, 1009−1012. (f) Schulte, B.; Tsotsalas, M.;
Becker, M.; Studer, A.; De Cola, L. Angew. Chem., Int. Ed. 2010, 49,
6881−6884. (g) Otsuka, H.; Aotani, K.; Higaki, Y.; Takahara, A. Chem.
Commun. 2002, 2002, 2838−2839.
(29) Irradiation wavelength: 340 nm. Note that due to the multistep
process of photocleavage and subsequent trapping of the alkyl radical
by molecular oxygen, the quantum yield is dependent on irradiation
conditions and oxygen concentration.
(30) Satzger, H.; Schmidt, B.; Root, C.; Zinth, W.; Fierz, B.; Krieger,
F.; Kiefhaber, T.; Gilch, P. J. Phys. Chem. A 2004, 108, 10072−10079.
(31) For a review on the utility of orthogonality on dynamic covalent
chemistry see: (a) Wilson, A.; Gasparini, G.; Matile, S. Chem. Soc. Rev.
2014, 43, 1948−1962 For related orthogonal photochemical or
́
mechanochemical activation of disulfides see:. (b) Gaston, O. A.;
Agustina, L. V.; Escalante, E. A.; Furlan, F. R. L. Chem. - Eur. J. 2018,
24, 3141−3146. (c) Fritze, U. F.; von Delius, M. Chem. Commun.
2016, 52, 6363−6366. (d) Belenguer, A. M.; Friscic, T.; Day, G. M.;
Sanders, J. K. M. Chem. Sci. 2011, 2, 696−700.
(18) Chen, M.; Zhong, M.; Johnson, J. A. Chem. Rev. 2016, 116,
10167−10211.
́ ́
(19) (a) Baron, M.; Morris, J. C.; Telitel, S.; Clement, J.-L.; Lalevee,
J.; Morlet-Savary, F.; Spangenberg, A.; Malval, J.-P.; Soppera, O.;
Gigmes, D.; Guillaneuf, Y. J. Am. Chem. Soc. 2018, 140, 3339−3344.
(b) Bottle, S. E.; Clement, J.-L.; Fleige, M.; Simpson, E. M.;
Guillaneuf, Y.; Fairfull-Smith, K. E.; Gigmes, D.; Blinco, J. P. RSC
Adv. 2016, 6, 80328−80333. (c) Su, J.; Liu, X.; Li, M.; Zhang, T.; Cui,
Y. Int. J. Polym. Sci. 2016, 2016, 8. (d) Su, J.; Liu, X.; Hu, J.; You, Q.;
Cui, Y.; Chen, Y. Polym. Int. 2015, 64, 867−874. (e) Telitel, S.; Telitel,
(32) Griffiths, P.; Moad, G.; Rizzardo, E.; Solomon, D. Aust. J. Chem.
1983, 36, 397−401.
(33) Marque, S.; Le Mercier, C.; Tordo, P.; Fischer, H. Macro-
molecules 2000, 33, 4403−4410.
(34) Fairfull-Smith, K. E.; Debele, E. A.; Allen, J. P.; Pfrunder, M. C.;
McMurtrie, J. C. Eur. J. Org. Chem. 2013, 2013, 4829−4835.
(35) Wang, S.; Xie, K.; Tan, Z.; An, X.; Zhou, X.; Guo, C.-C.; Peng,
Z. Chem. Commun. 2009, 6469−6471.
(36) (a) Skene, W. G.; Scaiano, J. C.; Yap, G. P. A. Macromolecules
2000, 33, 3536−3542. (b) Ohno, K.; Tsujii, Y.; Fukuda, T.
Macromolecules 1997, 30, 2503−2506.
́ ́
S.; Bosson, J.; Lalevee, J.; Clement, J.-L.; Godfroy, M.; Fillaut, J.-L.;
Akdas-Kilig, H.; Guillaneuf, Y.; Gigmes, D.; Soppera, O. Langmuir
2015, 31, 10026−10036. (f) Morris, J.; Telitel, S.; Fairfull-Smith, K. E.;
Bottle, S. E.; Lalevee, J.; Clement, J.-L.; Guillaneuf, Y.; Gigmes, D.
Polym. Chem. 2015, 6, 754−763. (g) Telitel, S.; Telitel, S.; Bosson, J.;
(37) (a) Marshall, D. L.; Gryn’ova, G.; Coote, M. L.; Barker, P. J.;
Blanksby, S. J. Int. J. Mass Spectrom. 2015, 378, 38−47. (b) Hodgson, J.
L.; Roskop, L. B.; Gordon, M. S.; Lin, C. Y.; Coote, M. L. J. Phys.
Chem. A 2010, 114, 10458−10466.
́ ́
Spangenberg, A.; Lalevee, J.; Morlet-Savary, F.; Clement, J.-L.;
Guillaneuf, Y.; Gigmes, D.; Soppera, O. Adv. Mater. Interfaces 2014,
1, 1400067. (h) Versace, D.-L.; Guillaneuf, Y.; Bertin, D.; Fouassier, J.
P.; Lalevee, J.; Gigmes, D. Org. Biomol. Chem. 2011, 9, 2892−2898.
́
(38) Guillaneuf, Y.; Bertin, D.; Gigmes, D.; Versace, D.-L.; Lalevee,
(i) Guillaneuf, Y.; Versace, D.-L.; Bertin, D.; Lalevee
Fouassier, J.-P. Macromol. Rapid Commun. 2010, 31, 1909−1913.
(j) Versace, D.-L.; Lalevee, J.; Fouassier, J.-P.; Guillaneuf, Y.; Bertin,
́
, J.; Gigmes, D.;
J.; Fouassier, J.-P. Macromolecules 2010, 43, 2204−2212.
(39) Due to solubility issues a 4:1 acetonitrile/benzene mixture had
to be used.
́
D.; Gigmes, D. Macromol. Rapid Commun. 2010, 31, 1383−1388.
(k) Goto, A.; Scaiano, J. C.; Maretti, L. Photochem. Photobiol. Sci. 2007,
6, 833−835. (l) Hu, S.; Malpert, J. H.; Yang, X.; Neckers, D. C.
Polymer 2000, 41, 445−452. (m) Scaiano, J. C.; Connolly, T. J.;
Mohtat, N.; Pliva, C. N. Can. J. Chem. 1997, 75, 92−97.
(20) (a) Morris, J. C. PhD Thesis, Queensland University of
Technology, 2017; DOI: 10.5204/thesis.eprints.110527. (b) Morris, J.
C.; Walsh, L. A.; Gomes, B. A.; Gigmes, D.; Fairfull-Smith, K. E.;
Bottle, S. E.; Blinco, J. P. RSC Adv. 2015, 5, 95598−95603.
(21) Montalti, M.; Credi, A.; Prodi, L.; Gandolfi, M. T. In Handbook
of Photochemistry, 3rd ed.; CRC Press: Boca Raton, 2006; pp 83−351.
(22) Li, L.; Hamer, G. K.; Georges, M. K. Macromolecules 2006, 39,
9201−9207.
(40) (a) Zhang, L.; Laborda, E.; Darwish, N.; Noble, B. B.; Tyrell, J.
H.; Pluczyk, S.; Le Brun, A. P.; Wallace, G. G.; Gonzalez, J.; Coote, M.
L.; Ciampi, S. J. Am. Chem. Soc. 2018, 140, 766−774. (b) Zhu, Q.;
Gentry, E. C.; Knowles, R. R. Angew. Chem., Int. Ed. 2016, 55, 9969−
9973.
́
(41) Guerinot, A.; Reymond, S.; Cossy, J. Eur. J. Org. Chem. 2012,
2012, 19−28.
(42) Serreli, V.; Lee, C.-F.; Kay, E. R.; Leigh, D. A. Nature 2007, 445,
523.
(23) This experiment has only qualitative character as the amount of
AOA decomposition is highly dependent on concentrations, light
intensity, the molar absorptivity of the sensitizer, its triplet lifetime,
and extent of triplet quenching by the released nitroxide. However, it
resembles the conditions used later for realizing photodynamic
reaction networks.
(24) Grubbs, R. B.; Wegrzyn, J. K.; Xia, Q. Chem. Commun. 2005,
80−82.
(25) (a) Shibatomi, K.; Zhang, Y.; Yamamoto, H. Chem. - Asian J.
2008, 3, 1581−1584. (b) Zhang, Y.; Shibatomi, K.; Yamamoto, H.
Synlett 2005, 2005, 2837−2842.
(26) Benoit, D.; Chaplinski, V.; Braslau, R.; Hawker, C. J. J. Am.
Chem. Soc. 1999, 121, 3904−3920.
(27) TPENO was reported to slowly decompose “over several days at
0 °C”, see ref 24. TIPNO decomposes only at elevated temperatures
by disproportionation, see: Nilsen, A.; Braslau, R. J. Polym. Sci., Part A:
Polym. Chem. 2006, 44, 697−717.
(28) (a) Willenbacher, J.; Wuest, K. N. R.; Mueller, J. O.; Kaupp, M.;
Wagenknecht, H.-A.; Barner-Kowollik, C. ACS Macro Lett. 2014, 3,
574−579. (b) Matyjaszewski, K.; Woodworth, B. E.; Zhang, X.;
Gaynor, S. G.; Metzner, Z. Macromolecules 1998, 31, 5955−5957.
(c) Connolly, T. J.; Baldoví, M. V.; Mohtat, N.; Scaiano, J. C.
Tetrahedron Lett. 1996, 37, 4919−4922.
K
DOI: 10.1021/jacs.8b03633
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX