Improving the fatigue resistance of diarylethene switches

Article abstract of DOI:10.1021/ja513027s

When applying photochromic switches as functional units in light-responsive materials or devices, an often disregarded yet crucial property is their resistance to fatigue during photoisomerization. In the large family of diarylethene photoswitches, format

Full text of DOI:10.1021/ja513027s

Article
pubs.acs.org/JACS
Improving the Fatigue Resistance of Diarylethene Switches
Martin Herder, Bernd M. Schmidt, Lutz Grubert, Michael Patzel, Jutta Schwarz, and Stefan Hecht*
̈
Department of Chemistry, Humboldt-Universitat zu Berlin, Brook-Taylor-Straße 2, 12489 Berlin, Germany
̈
S
* Supporting Information
ABSTRACT: When applying photochromic switches as
functional units in light-responsive materials or devices, an
often disregarded yet crucial property is their resistance to
fatigue during photoisomerization. In the large family of
diarylethene photoswitches, formation of an annulated isomer
as a byproduct of the photochromic reaction turns out to
prevent the desired high reversibility for many different
derivatives. To overcome this general problem, we have
synthesized and thoroughly investigated the fatigue behavior of
a series of diarylethenes, varying the nature of the hetaryl
moieties, the bridging units, and the substituents. By analysis of
photokinetic data, a quantification of the tendency for
byproduct formation in terms of quantum yields could be achieved, and a strong dependency on the electronic properties of
the substituents was observed. In particular, substitution with 3,5-bis(trifluoromethyl)phenyl or 3,5-bis(pentafluorosulfanyl)-
phenyl groups strongly suppresses the byproduct formation and opens up a general strategy to construct highly fatigue-resistant
diarylethene photochromic systems with a large structural flexibility.
Scheme 1. Isomerization Behavior and Byproduct Formation
of Diarylethenes
INTRODUCTION
■
Photochromic molecules, which interconvert reversibly between
two states by irradiation with light, increasingly receive attention
as photoswitches that allow control over complex functions on
the molecular scale, e.g., reactivity and catalysis,1 single-molecule
fluorescence,2 or supramolecular binding,3 as well as functions of
large molecular ensembles, e.g., photoactuation and optome-
chanics4 or charge transport in organic electronic devices.5
Nevertheless, in order to eventually render such functional
systems technologically relevant, it is mandatory to use switches,
which in addition to undergoing a highly efficient photo-
isomerization reaction in both directions show no or at least
extremely low fatigue over a large number of switching cycles.
Even the occurrence of a slight side reaction that is hardly
observed when the photoisomerization is performed only once
will lead to a significant loss of photochromic material after tens
or hundreds of switching cycles. In this regard diarylethene
switches, interconverting between the ring-open isomer A and
the ring-closed isomer B by irradiation with UV light and visible
light (Scheme 1), are generally regarded to possess outstanding
properties. These are manifested in the fact that they typically
show fast and nearly quantitative interconversion between
thermally stable isomers, with some derivatives being reported
to last more than 104 switching cycles without any sign of
degradation.6 In particular for structural motifs possessing either
benzothiophene7 or β-methyl-substituted thiophenes8 as hetaryl
moieties in conjunction with a central hexafluorocyclopentene
bridge (Scheme 2), no byproduct formation has been observed.
However, it has been noted by several groups that distinct
diarylethene derivatives show different types of photochemical
side reactions, e.g., oxidation or elimination reactions of the ring-
closed isomer.9 In particular, the formation of an annulated ring
system C (Scheme 1) has been reported for perfluoro-
cyclopentene and perhydrocyclopentene derivatives.8,10 The
structure of the byproduct has been proven by X-ray
crystallography.8,10g It is formed upon excitation of the ring-
closed isomer with UV light by a formal 1,2-dyotropic
rearrangement, which may proceed via a concerted mechanism
as shown in Scheme 1 or via radical or ionic intermediates.10g,11
A
Received: December 22, 2014
© XXXX American Chemical Society
A
DOI: 10.1021/ja513027s
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Scheme 2. Rare Examples of Highly Fatigue-Resistant
Diarylperfluorocyclopentenes
Scheme 3. Diarylethene Derivatives Investigated in This
Work
recent theoretical work identified the cleavage of the C−S bond
of one of the thiophene rings as the excited-state reaction
coordinate leading to the byproduct.12 Although there are no
systematic investigations on the influence of diarylethene
structure on byproduct formation, it has been noted in earlier
work that dithienylperhydrocyclopentenes suffer more pro-
nounced fatigue as compared to the respective perfluoro-
cyclopentene derivatives.13 For the latter, the presence of an
additional methyl group in the 4-position (β-position) of the
thiophene ring or the use of benzothiophenes as hetaryl moieties
is essential to fully shut off the reaction pathway to the annulated
ring system.8,14 Additionally, oxidation of one thiophene or
benzothiophene unit to the S,S-dioxide or substitution with
acetyl groups may lead to an improved photochemical stability.15
Note that, for the photoconversion of diarylethenes in the single-
crystalline state, the formation of the annulated byproduct has
not been observed.8
Recently, large structural variations of the parent diarylethene
motif have been made in order to implement distinct properties
or functions, e.g., functionalization of the bridging ethylene unit
with reactive or catalytically active moieties,16 incorporation of
the central double bond into heteroaromatic rings in terarylenes
giving rise to a highly efficient isomerization behavior,17 or use of
double-bond-containing structures different from the parent
thiophene or benzothiophene units.18 In this context, it is highly
important to find a general concept of implementing fatigue
resistance to diarylethenes while maintaining a maximum degree
of structural and synthetic flexibility and hence not being
restricted to the parent motifs shown in Scheme 2.
In our ongoing work on diarylethene switches, we noticed a
ubiquitous appearance of the condensed ring system C for quite
different structures. Nevertheless, we noticed as well that there is
a strong dependency of the amount of byproduct formed on the
substitution pattern of the adjacent phenyl rings. In order to gain
a systematic insight into this structure-switch relationship, we
prepared a series of diarylethene derivatives and investigated
their fatigue behavior in detail. Starting with the parent
dithienylperhydrocyclopentene structure 1 (Scheme 3), sub-
stituents on the phenyl ring were varied from strongly donating
(4-Me2N, 1a) to strongly accepting (3,5-(CF3)2, 1h) moieties.
To study the effect of the perfluorocyclopentene bridge,
derivatives 2c, 2h, and 2i were prepared. In addition, thiazole
analogues 3 and 4 as well as two non-symmetrically substituted
dithienylcyclopentenes 5 were investigated. Last but not least,
dithienylcyclopentene 6c was synthesized to determine the effect
of the additional β-methyl group in the case of a bridge other than
perfluorocyclopentene.
thien-4-yl)cyclopentene with the corresponding aryl halides.13
Perhydrocyclopentene derivatives bearing thiazole units (3d,h)
were synthesized by cross-coupling of substituted thiazolylbor-
onic acids or stannanes and 1,2-dibromocyclopentene.20
Analogous compounds bearing the perfluorocyclopentene
bridge were obtained by borylation and subsequent Suzuki
cross-coupling of 1,2-(2-chloro-5-methylthien-4-yl)perfluoro-
cyclopentene13 or by cross-coupling of corresponding thienyl
and thiazolyl boronic esters with 1,2-dichlorohexafluoro-
cyclopentene.21 Compounds 1b,13 1c,22 1d,23 1e,13 2c,24 2h,25
and 6c26 have been reported earlier. Details on the synthesis and
characterization are given in the Supporting Information (SI).
Fatigue Behavior. An example of the typical spectral
response of diarylethene switches upon irradiation with UV
light is shown in Figure 1 for the methyl-substituted compound
1c. While the ring-open isomer possesses a strong absorbance in
the UV range, upon irradiation with 313 nm light, a new broad
band centered at 522 nm builds up that is characteristic for the
ring-closed isomer with its extended π-conjugation (Figure 1a).
The fast growth of the visible band is completed after 60 s when a
plateau is reached; i.e., the absorbance does not increase upon
further UV illumination. This is the expected behavior for a
photochemical equilibrium between the ring-open and ring-
closed isomers, leading to a photostationary state (PSS).
Importantly, when the sample after reaching the PSS is further
subjected to UV illumination, a slow decrease of the visible
absorbance can be observed, accompanied by a small
hypsochromic shift of the band maximum (Figure 1b). Analysis
of the irradiation solution by ultraperformance liquid chroma-
tography (UPLC) reveals the emergence of a second photo-
product upon longer irradiation times (Figure 1d). In contrast to
the formation of the ring-closed isomer, this process is not
reversible; i.e., upon visible light irradiation to induce the ring-
opening reaction, a weak absorbance in the visible range is
RESULTS AND DISCUSSION
■
Synthesis. The investigated diarylethenes were synthesized
following established procedures.19 Perhydrocyclopentene de-
rivatives 1a−h and 5a,b were easily obtained by borylation and
subsequent Suzuki cross-coupling of 1,2-(2-chloro-5-methyl-
B
DOI: 10.1021/ja513027s
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 1. UV/vis spectroscopy of an acetonitrile solution of 1c (2.38 × 10−5 M, 25 °C). (a) Irradiation with UV light (310 nm) until reaching maximum
absorbance in the visible range. (b) Prolonged UV irradiation (310 nm) after reaching maximum absorbance. (c) Evolution of absorbance in the visible
range during repetitive switching cycles consisting of alternating UV (310 nm, 90 s) and visible irradiation (>500 nm, 600 s). (d) Evolution of UPLC
traces (absorbance monitoring by diode array detector integrated between 250 and 800 nm) upon continuous UV irradiation (310 nm).
retained and corresponds to the formed byproduct. Although the
observed byproduct formation seems to be very slow, it leads to a
significant loss of photochromic material already after a few
switching cycles, which consist of short UV irradiation until
reaching the (pseudo)PSS and subsequent visible light
irradiation until no further spectral change can be observed
(Figure 1c).
derivatives, for diarylethenes bearing thiazoles as aryl moieties,
the formation of a byproduct has not been reported yet. In
contrast, they were originally characterized as highly fatigue
resistant.27
To prove that the photochemical byproduct formed from
dithiazolylethene 3d is structurally analogous to the thiophene
counterpart, it has been isolated by irradiation and subsequent
chromatography on a preparative scale. Again, 1H NMR
spectroscopy shows signals for two magnetically inequivalent
methyl groups (Figure S31), while 13C NMR shows two signals
for quaternary carbon atoms at high field. Long-range couplings
observed in 2D NMR spectra are similar to those observed for
the byproduct of 1c (Figures S34 and S35). Note that for all
compounds reported in this study, UPLC/MS analysis of
irradiated samples verifies that the emerging byproduct is an
additional isomer of the parent, possessing an m/z ratio identical
to that of the respective ring-open and ring-closed forms. Given
the close similarities between all compounds in the observed
UV/vis spectra and corresponding UPLC/MS traces, we can
assume that, in any case, the fatigue of the photochrome is based
on the formation of the same condensed ring system as a
byproduct. Only in the case of electron-rich dithienylcyclopen-
tenes 1a−f and 5a,b some additional unspecific byproduct
formation, in the form of oxidation reactions, can be detected by
UPLC/MS (peaks corresponding to (M+16)+ and (M+32)+).
However, in argon-saturated solutions, the emergence of the
rearranged isomer is much faster than photooxidation.
The identity of the formed byproduct with the condensed ring
system C shown in Scheme 1, which was reported earlier by the
group of Irie,8 was proven by its preparative isolation and
1
subsequent NMR spectroscopic examination. While H NMR
spectra of the ring-open and ring-closed isomers of 1c show only
one signal for the methyl groups attached to the reactive carbons,
two separate singlets, each integrating for three protons, appear
in the spectrum of the isolated byproduct (see Figure S30). 2D
NMR spectroscopy proves the structure by unambiguous
assignment of the observed signals and analysis of the coupling
paths (see Figures S32 and S33). Notably, the downfield-shifted
part of the 1H NMR spectrum of the byproduct is nearly identical
to that of the ring-closed isomer, which is an indication for their
similar π-electronic structure.
From inspection of the irradiation spectra of all other
diarylethene derivatives discussed here (collected in section 2.1
of the SI), it can be deduced that the fatigue exemplified in
Figure 1 for compound 1c is a general feature of diarylethenes,
independent of the nature of their aryl moieties or their bridging
unit. While, as discussed above, the formation of the annulated
ring system has been noted earlier for some dithienylethene
C
DOI: 10.1021/ja513027s
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Table 1. Photochemical Properties of Investigated Diarylethenes in Acetonitrile
λmax/nm (ε/104 M−1 cm−1
)
yield/% (310 nm)
ring-open
isomer
ring-closed
isomer
a
b
c
c
c
d
compd
byproduct PSS byproduct
ΦAB (313 nm)
ΦBA (313 nm)
ΦBC (313 nm)
ΦBA (546 nm)
1a
1b
1c
1d
1e
1f
331 (4.33)
283 (3.19)
279 (3.45)
278 (3.25)
284 (3.71)
273 (2.83)
334 (3.58)
284 (3.16)
289 (3.87)
300 (3.45)
302 (3.43)
316 (2.13)
332 (1.80)
307 (3.11)
326 (3.37)
282 (3.04)
271 (2.75)
537 (2.66)
520 (1.90)
522 (1.97)
520 (1.78)
532 (2.16)
512 (1.18)
556 (2.32)
548 (1.96)
588 (1.79)
590 (1.54)
590 (1.58)
500 (1.55)
522 (1.34)
538 (1.13)
570 (2.54)
546 (1.82)
483 (1.29)
530 (1.44)
509 (0.84)
508 (1.13)
510 (0.94)
514 (1.17)
476 (0.51)
532 (1.18)
524 (0.95)
550 (0.85)
549 (0.72)
545 (0.76)
521 (0.80)
540 (0.77)
538 (0.64)
539 (1.22)
515 (0.82)
459 (0.35)
99
89
94
98
96
96
98
97
94
96
96
94
95
89
97
99
91
71
86
79
73
66
37
25
6
0.74 0.09
0.47 0.05
0.43 0.05
0.47 0.05
0.42 0.05
0.41 0.04
0.65 0.07
0.44 0.04
0.62 0.08
0.58 0.06
0.50 0.05
0.53 0.05
0.56 0.06
0.45 0.05
0.03 0.06
0.03 0.03
0.01 0.03
0.02 0.05
0.02 0.03
0.01 0.05
0.06 0.05
0.02 0.06
0.03 0.03
0.05 0.05
0.04 0.05
0.03 0.05
0.03 0.05
0.06 0.04
0.008 0.003
0.007 0.002
0.009 0.003
0.013 0.003
0.006 0.002
0.006 0.001
0.0029 0.0002
0.0072 0.0006
0.008 0.0006
0.0093 0.0007
0.0055 0.0004
0.020 0.002
0.0021 0.0002
0.0046 0.0004
0.015 0.001
0.017 0.001
0.018 0.001
0.021 0.002
0.015 0.001
0.039 0.003
0.0022 0.0002
0.0058 0.0005
0.043 0.003
1g
1h
2c
2h
2i
0.0013 0.0003
0.0004 0.0001
0.0006 0.0002
0.0004 0.0001
0.00019 0.00004
0.007 0.001
13
8
3
3d
3h
4h
5a
5b
6c
65
22
13
6
0.002 0.0005
0.0008 0.0002
0.0009 0.0002
0.0013 0.0002
0.051 0.005
0.56 0.06
0.47 0.05
0.005 0.008
0.02 0.04
0.04 0.03
19
26
0.004 0.001
a
b
Amount of ring-closed isomer in the (pseudo)PSS reached after UV irradiation (310 nm), obtained by UPLC. Amount of byproduct after 30 min
c
of UV irradiation (310 nm), obtained by UPLC. Quantum yields obtained by nonlinear regression of photokinetic data under UV irradiation
(313 nm). Ring-opening quantum yields under visible light irradiation (546 nm) obtained from the initial slope of photoconversion.
d
Interestingly, close inspection of the UV/vis spectra (section
2.1 of the SI) and spectroscopic data given in Table 1 reveals
subtle differences in the spectroscopic characteristics of the
byproduct when comparing dithienylethenes 1c, 1h, and 2h with
their thiazole analogues 3d, 3h, and 4h. For all compounds, both
the ring-closed isomer and the byproduct possess very similar
absorbances in the visible region, due to analogous π-conjugation
pathways throughout their molecular backbone. Nevertheless,
the molar absorptivity of the byproduct is significantly lower,
leading to the typical decrease of the absorption band upon
prolonged UV irradiation. For thiophene derivatives, the
decrease is accompanied by a hypsochromic shift, which is
more pronounced for electron-deficient derivatives (Δλmax = 14
nm for 1d, Δλmax = 41 nm for 2h). However, thiazole derivatives
show a bathochromic shift of 21 and 18 nm for 3d and 3h,
respectively; in the case of 4h, the λmax values of the ring-closed
form and the byproduct are identical.
To gain further insight into the spectroscopic characteristics,
some calculations were performed on the prototype thiophene-
and thiazole-containing structures 1c and 3d. Following
procedures outlined by Jacquemin et al.,28 ground-state geo-
metries were optimized using density functional theory (DFT) at
the B3LYP/6-311G(d,p) level of theory, and vertical transition
energies were computed by time-dependent DFT using the
CAM-B3LYP as well as PBE0 functionals together with the 6-
311+G(2d,p) basis set (for details, see section 6 of the SI). It has
been shown that CAM-B3LYP and PBE0 give a good agreement
between experimental spectra and calculated vertical transition
energies of diarylethenes.29 In the cases of 1c and 3d, CAM-
B3LYP reproduces the experimentally observed λmax for all three
isomers very accurately (Figures S55 and S57), while the PBE0
functional significantly underestimates the transition energies
(Figures S56 and S58). Nevertheless, both methods confirm a
hypsochromic shift of λmax for 1c (11 nm with CAM-B3LYP/15
nm with PBE0) and a bathochromic shift of λmax for 3d (24 nm
with CAM-B3LYP/21 nm with PBE0) when going from the ring-
closed isomer to the byproduct. The reduced oscillator strength
of the lowest energy transition of the byproducts is demonstrated
as well. Inspecting the optimized geometries, one finds that the
major difference between isomers of 1c and 3d lies in the twisting
of the adjacent phenyl groups, while deviations in the central
hexatriene/cyclohexadiene core are small (see Table S4 for
structural parameters). In general, the dihedral angle θ between
thiophenes and phenyl rings in 1c is larger for all isomers than
that between thiazoles and phenyl rings in 3d. Importantly, θ
increases on going from the ring-closed isomer to the byproduct
due to increased steric repulsion by expanding the former
thiophene/thiazole 5-membered ring to a 6-membered ring.
While for 3d the increase is moderate (θ = 4° for the ring-closed
isomer/θ = 20° for the byproduct), it is larger for 1c (θ = 13° for
the ring-closed isomer/θ = 40° for the byproduct). Thus, a
smaller increase of the dihedral angle of 3d due to less steric
repulsion by the nitrogen atom may lead to a better
delocalization of π-electron density, resulting in the exper-
imentally observed bathochromic shift, while a stronger twisting
of the phenyl group shifts the absorbance maximum of the
byproduct of 1c to shorter wavelengths (Figure S59).
Quantification of Byproduct Formation. Along the whole
series of diarylethene derivatives possessing electronically
different substitution patterns, significant differences in the rate
of byproduct formation have been observed. At first glance, a
good indication for the performance of a switch is the amount of
byproduct formed after a certain time of UV irradiation under
comparable conditions, which can readily be determined by
UPLC. Although this can only serve as a crude estimate since the
exact amount of byproduct depends on the sample concen-
tration, molar absorptivities at the irradiation wavelength, and
quantum yields for all photochemical reactions, the obtained
conversions clearly show a strong correlation of the fatigue
behavior with the substitution pattern (Figure 2, Table 1). After
30 min of high-intensity UV irradiation using a 1000 W Xe arc
lamp together with a 310 nm interference filter, the most
electron-rich dithienylethenes 1a−d, bearing perhydrocyclopen-
tene bridges and electron-donating or electron-neutral sub-
D
DOI: 10.1021/ja513027s
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
were obtained (Table 1). Note that assuming a different model
with the byproduct being formed parallel to the ring-closed
isomer by excitation of the ring-open isomer does not lead to a
satisfying fit of the experimental data, in accordance to previous
findings.10g A detailed discussion of the photokinetic model, the
fitting procedure, and experimental errors can be found in
section 1.4 of the SI.
Quantum Yields for Ring Closure and Ring Opening.
The data in Table 1 show that the majority of the investigated
compounds possess the expected photochemical behavior of
diarylethenes substituted with methyl groups and phenyl rings at
the inner and outer α-positions of the hetaryl groups,
respectively.13,32 Quantum yields for ring closure are in the
range of 0.4−0.6, reflecting the thermal equilibrium between the
parallel and antiparallel conformers of the ring-open isomer, with
only the latter being photoactive.6b Quantum yields for ring
opening are more than 1 order of magnitude smaller and strongly
wavelength-dependent; i.e., the ring-opening reaction performed
upon irradiation with visible light proceeds much more slowly
than the reaction with UV light.33 Due to the ratio of ΦAB:ΦBA,
the amount of ring-closed isomer in the PSS generally exceeds
90%, although both isomers absorb UV light. Only compounds
1a and 5a deviate from the described behavior, with the former
possessing a slightly larger quantum yield for ring closure of 0.74,
and the latter showing a significantly reduced photoreactivity,
with ΦAB being only 0.05. Whereas a differing population of the
antiparallel and parallel conformers in the electronic ground state
may serve as explanation for these deviations, for compound 5a
an electronic reason, i.e., the pronounced donor−acceptor
character of the π-electronic system, seems plausible as well. This
is manifested in the fact that quantum yields for both ring closure
and ring opening are reduced by approximately 1 order of
magnitude. In addition, compound 5a shows a strong bath-
ochromic shift of the absorption bands of all three isomers,
pointing to a charge-transfer character of the excitation.
Quantum Yields for Byproduct Formation. The
measured quantum yields for byproduct formation confirm the
qualitatively observed trends. The largest values for ΦBC are
found for donor-substituted and unsubstituted dithienylethenes
possessing the perhydrocyclopentene bridge (1a−d), between
0.007 and 0.013. Thus, for these structures, the rate of byproduct
formation is in the same order of magnitude as the rate of the
ring-opening reaction, leading to the observed fast depletion of
the photochromic material under continuous UV irradiation and
high irreversibility during full switching cycles (Figure 1c,
FigureS25). A comparison between structures 1d and 3d shows
that employing less electron-rich thiazoles instead of thiophenes
increases the fatigue resistance only marginally. In contrast, the
perfluorination of the cyclopentene bridge has a much larger
effect. Compared to perhydrocyclopentene 1c, ΦBC is reduced by
a factor of 15 to a value of 0.0006 for perfluorocyclopentene 2c.
Importantly, in the perhydrocyclopentene series, byproduct
formation is suppressed to the same degree by introduction of
acceptor groups on the adjacent phenyl rings. While a small effect
is observed for substitution with p-bromophenyl groups (1e) and
pentafluorophenyl groups (1f), substitution with an ester group
in the para position of the phenyl ring (1g), exerting a −M effect,
reduces ΦBC to a value of 0.001. An exceptional improvement of
the fatigue behavior is found when chemically inert trifluoro-
methyl groups, exerting a strong −I effect, are attached at the
meta positions of the phenyl rings. With ΦBC = 0.0004 for
compound 1h, byproduct formation is suppressed by a factor of
20−30 when compared to the unsubstituted or donor-
Figure 2. Amount of byproduct formed after 30 min of UV irradiation
(1000 W Xe arc lamp, 310 nm interference filter, I0 = 5.67 × 10−9 E s−1
cm−3 determined by ferrioxalate actinometry) of acetonitrile solutions (c
= 2.5 × 10−5 M) of diarylethenes 1a−6c.
stituents on the adjacent phenyl rings, are almost quantitatively
converted to the byproduct. For analogous compounds
possessing substituents with increasing acceptor strength, e.g.,
bromine (1e), fluorine (1f), ester (1g), or trifluoromethyl (1h)
groups, the yield of byproduct is more and more suppressed,
down to a minimum amount of ca. 6% for compound 1h.
Notably, structurally analogous diarylethenes of the thiophene
and thiazole series possess a comparable fatigue resistance.
Introduction of the perfluorocyclopentene bridge strongly
reduces byproduct formation for donor- or electron-neutral-
substituted compounds, while no further suppression is observed
for trifluoromethyl-substituted compound 2h when compared to
the respective perhydrocyclopentene 1h.
In order to obtain more precise information on the
dependency of fatigue behavior on the substitution pattern, a
precise quantification of the rate of byproduct formation is
needed. Therefore, quantum yields for all photochemical steps
involved in the isomerization and fatigue of diarylethenes have
been determined by means of measurement of absorbance−time
profiles under continuous UV irradiation and their evaluation by
numerical nonlinear regression.30 For this purpose, the molar
absorptivities of the ring-closed isomer and byproduct have to be
known. However, as their UV/vis spectra overlap with that of the
ring-open isomer, a calculation from pure photokinetic data
without simplifying assumptions or the use of special irradiation
conditions is impossible.31 A preparative isolation of the isomers
is laborious and in some cases very difficult due to their similar
structures and susceptibility to oxidation. Thus, a careful analysis
of irradiated samples at different stages of the irradiation was
conducted by UPLC, and spectra of the pure isomers were
calculated by determination of conversions to the ring-closed
isomer and byproduct (for details, see section 1.2 of the SI). A
comparison of the spectra obtained by this method with spectra
of isolated isomers of compound 3d shows an excellent
agreement (Figure S3).
As a kinetic model for the nonlinear regression, a photo-
chromic ABC(3Φ) system according to the reaction scheme
shown in Scheme 1 was assumed, consisting of a photochemical
equilibrium between species A and B and a subsequent
irreversible photochemical reaction from species B to C. Since
the corresponding rate equations cannot be integrated in a closed
form, they were treated numerically. By fitting the theoretical
model to the experimental data, quantum yields ΦAB for ring
closure, ΦBA for ring opening, and ΦBC for byproduct formation
E
DOI: 10.1021/ja513027s
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
substituted analogues 1a−d. Importantly, the CF3 substitution
does not alter the rates of the desired photochemical isomer-
ization reactions between the ring-open and ring-closed isomers
and has only a minor influence on the electronic spectra
(Figure 3). The positive effect of the 3,5-bis(trifluoromethyl)-
substituents. Besides modulation of the rates of ring closure and
ring-opening in the case of compound 5a (vide supra), it was
found that for both structures byproduct formation is strongly
suppressed compared to the symmetrically donor-substituted
compounds. This strategy of using a non-symmetrical sub-
stitution pattern to implement fatigue resistance via the 3,5-
bis(trifluoromethyl)phenyl group on one side and a functional
moiety on the other side of the diarylethene recently allowed us
to remote-control the reactivity of a furan in a Diels−Alder
reaction.18a Importantly, in this particular case, a perfluorinated
bridge could not be used to provide improved fatigue resistance
for electronic reasons, as it would significantly reduce the Diels−
Alder reactivity of the attached furan. This nicely demonstrates
the power of imparting fatigue resistance via the diarylethene
termini and not the bridge.
The second previously known strategy to impart fatigue
resistance to diarylethenes making use of β-methyl-substituted
thiophenes was reported only for compounds possessing the
perfluorocyclopentene bridge.8 To test if the β-methyl
substituent has an effect on the fatigue behavior when other
bridges are used, the photochemistry of compound 6c,
possessing β-methyl-substituted thiophenes and the perhydro-
cyclopentene bridge, was investigated. In fact, 6c shows
significant byproduct formation with a quantum yield only
slightly lower than that of compound 1c without β-methyl
substituents. This finding supports theoretical investigations,14
which attribute the lack of fatigue of β-methyl-substituted
dithienylperfluorocyclopentenes to steric hindrance between the
β-methyl group and the fluorine atoms, which are not present in
compound 6c.
Figure 3. UV/vis absorbance spectra of an acetonitrile solution of 1h
(2.09 × 10−5 M, 25 °C) during repetitive switching cycles consisting of
alternating UV (310 nm, 60 s) and visible irradiation (>500 nm, 600 s).
Inset: Evolution of absorbance in the visible range.
phenyl groups can be observed in the dithiazolylethene series as
well. However, compared to the parent structure 3d, the
quantum yield of byproduct formation of compound 3h is
reduced by only a factor of 3. The combination of the 3,5-
bis(trifluoromethyl)phenyl substituents with the perfluoro-
cyclopentene bridge in diarylethene structures 2h and 4h gives
a further reduction of ΦBC only in the case of the thiazole-
containing compound. In the thiophene series, the performances
of the perhydro- and perfluorocyclopentene switches 1h and 2h
are very similar. The large improvement of fatigue resistance for
structures 1h, 2h, 3h, and 4h is manifested as well in the excellent
reversibility during full switching cycles (Figure 3, Figure S25).
In an attempt to further suppress byproduct formation by
substitution with strong acceptor groups, compound 2i was
synthesized, possessing pentafluorosulfanyl groups. Compared
to CF3, the SF5 group is an even stronger electron acceptor while
being structurally related and chemically inert.34 In fact,
comparing structures 2h and 2i upon introduction of SF5
groups, ΦBC is further reduced by a factor of 2. Thus far,
compound 2i shows the highest fatigue resistance of all
diarylethene structures in our hands. Note that, similar to CF3
groups, SF5 substitution only marginally alters the photochromic
properties of the parent unsubstituted compound.
Solvent and Wavelength Dependence of ΦBC. The
photochemistry of dithiazolylethene 3d was investigated in
different solvents and under different irradiation conditions
(Table 2, Figures S23 and S24). Using solvents with lower
Table 2. Solvent and Wavelength Dependence of Quantum
Yields of Compound 3d
PSS
λirr/
nm
yield/
a
b
b
b
solvent
%
ΦAB
ΦBA
ΦBC
MeCN
CH2Cl2
C6H12
313
313
313
297
334
94
94
96
88
96
0.53 0.05
0.59 0.06
0.73 0.07
0.51 0.05
0.60 0.06
0.03 0.05
0.03 0.04
0.02 0.06
0.04 0.03
0.04 0.08
0.007 0.001
0.007 0.001
0.008 0.002
0.007 0.001
0.008 0.001
MeCN
MeCN
a
b
Amount of ring-closed isomer in the (pseudo)PSS. Quantum yields
obtained by nonlinear regression of photokinetic data under UV
irradiation (313 nm).
Given the outstanding performance of CF3- and SF5-
substituted diarylethenes in terms of fatigue resistance, we
speculated that it could be sufficient to functionalize only one
hetaryl terminus of the structure with this motif while still
suppressing byproduct formation. This would enable a flexible
choice of the structure of the second terminus, thus allowing for
implementing any desired chemical functionality to be
modulated by the photochromic reaction. To answer this
question, non-symmetrically substituted dithienylethenes 5a
and 5b were synthesized, possessing the 3,5-bis(trifluoro-
methyl)phenyl group on one thiophene ring and a 4-N,N-
dimethylamino- or 4-methoxy-substituted phenyl group on the
other. In terms of fatigue, this substitution pattern combines one
of the best-performing acceptors with poor-performing donor
polarity than acetonitrile causes a small increase of the quantum
yield for ring closure, while quantum yields for ring opening as
well as byproduct formation are not altered. Interestingly,
quantum yields for all three isomerization reactions are not
affected within the error of measurement by the choice of the
irradiation wavelength in the UV range. For photochemical ring
closure of diarylethenes, the insensitivity to the irradiation
wavelength was expected.33a The previously observed wave-
length dependence of the quantum yield for the ring-opening
process33 is manifested in the much lower values for excitation
with visible light (Table 1), yet probably could not be resolved in
our experiments within the UV range due to the high statistical
error of this parameter.
F
DOI: 10.1021/ja513027s
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
In accordance with previous reports,8,10g we find that
byproduct formation is not induced by irradiation of the ring-
closed isomer using visible light (see section 1.6 of the SI). This
may be ascribed to a thermal barrier in the S1 excited state,
making excitation to a higher excited state by UV light necessary
to deliver sufficient kinetic energy upon internal conversion to
the S1 state to overcome the barrier.12 Thus, one would also
expect a dependence of the quantum yield for byproduct
formation on the wavelength within the UV range. However, we
find that ΦBC is essentially wavelength independent between 297
and 334 nm.
strong donor substitution (1a) and >2.00 V in the case when the
perfluorocyclopentene bridge is used (2h, 2i, and 4h). The same
trend is observed for ring-closed isomers, with their first
a1
oxidation potential Ep between −0.37 V (1a) and 1.35 V
(4h). In particular, oxidation potentials of ring-closed isomers
may be related to the observed trends in the fatigue behavior of
the photochromes. In fact, starting from the parent dithienylcy-
lopentene 1c, the largest effect on oxidation potentials has the
a1
perfluorination of the cyclopentene bridge shifting Ep of the
ring-closed isomer by 520 mV (1c vs 2c) to more positive values,
accompanied by a massive improvement of the fatigue resistance.
Introduction of CF3 groups in 1h, exerting the same effect on the
fatigue, gives a potential shift of 260 mV. Comparing 1h and 2h,
one finds again a shift of 520 mV arising from the perfluorination
of the bridge; nevertheless, there is no difference in the rate of
byproduct formation between these two structures. The
replacement of CF3 with SF5 groups, giving rise to the structure
with the lowest rate of byproduct formation, is accompanied by a
further increase of the oxidation potential by 110 mV.
In contrast to this general trend, the thiazole-containing
structure 3d becomes oxidized at slightly higher potentials than
1h, despite showing a much higher rate of byproduct formation.
Again, introduction of CF3 groups in 3h shifts the potential by
210 mV.
Another example showing that there is no straightforward
correlation between electrochemical oxidation potentials and
photochemical fatigue is the pentafluorophenyl-substituted
compound 1f. It was synthesized with the expectation that the
pentafluorophenyl group would have an effect similar to that of
the 3,5-bis(trifluoromethyl)phenyl group in 1h. Indeed, both
structures become oxidized at very similar potentials in the ring-
open and ring-closed states. However, for 1f the quantum yield
for byproduct formation is larger by more than 1 order of
magnitude. Possibly the +M effect of fluorine atoms directly
attached to a π system stabilizes the transition state or
intermediate during formation of the annulated isomer.
Electronic Effects of the Substituents. To estimate the
effects of the substituents on the diarylethenes’ electronic
structure, cyclic voltammetry measurements were performed
(see section 5 of the SI). Perhydrocyclopentene-substituted
dithienylethenes 1a−e, 5a, 5b, and 6c show the expected
electrochemical behavior, consisting of an irreversible oxidation
of the ring-open isomer followed by the appearance of cathodic
waves at potentials similar to those of the two reversible
oxidations of the respective ring-closed isomers, which indicates
a thermal cyclization reaction in the dicationic state.35
Interestingly, in the case of the non-symmetrically substituted
derivative 5a, two separated oxidation waves can be observed for
the ring-open isomer, the first being fully reversible and only the
second leading to the cyclization reaction. For the acceptor-
substituted analogues 1f, 1g, and 1h, the oxidative cyclization
cannot be observed. The same holds for thiazole-substituted
compounds 3d and 3h. For the perfluorocyclopentene
derivatives, an oxidative cyclization takes place only for the
most electron-rich derivative 2c. Reductions, which are generally
observed within the accessible potential range for all ring-closed
isomers and only for acceptor-substituted ring-open isomers, do
not induce isomerization reactions.
Peak potentials (against Fc/Fc+) reported in Table 3 show
that, by changing the substitution pattern, the electron density
within the π-electronic system is varied over a broad range. Thus,
ring-open isomers become oxidized between 0.13 V in the case of
DFT calculations at the B3LYP/6-31G(d) level of theory were
performed to estimate the influence of the substituents on the
geometry and relative ground-state stability of the three isomers
(see section 6.2 of the SI). Therefore, the hypothetical derivatives
1i (R = 3,5-(SF5)2), 2d (R = H), and 3c (R = Me) (see
Scheme S2) were also included. Remarkably, in all cases the
geometry optimization led to very similar structures of the
diarylethene cores, independent of the substituents on the
phenyl rings, with only small, non-systematic variations in
geometric parameters such as dihedral angles, the distance
between the ring-closing carbons, or the length of the C−S bonds
(Table S6). Computed NICS values36 of the thiophene or
thiazole rings in the ring-open isomers also show only minimal
variation upon exchange of substituents (Table S6). However,
ground-state energies of the ring-closed isomers and byproducts,
relative to the respective ring-open isomers, slightly rise within a
series of similar structures on going from donor to acceptor
substitution (Figure 4, Table S6). While the byproducts are
generally much less stable than the ring-closed isomers, their
further destabilization by acceptor substitution may lead to an
increased fatigue resistance. Nevertheless, comparing the
different series (cyclopentene vs perfluorocyclopentene and
thiophene vs thiazole switches), it is obvious that, for compounds
experimentally possessing similar fatigue behavior (e.g., 1h and
2h), the stabilities of the byproducts are very different. Thus,
more advanced quantum mechanical calculations concerning
Table 3. Anodic and Cathodic Peak Potentials Determined by
Cyclic Voltammetry in Acetonitrile
a
open isomer
closed isomer
compd
Epa1/V
Epc1/V
Epa1/V
Epa2/V
Epc1/V
1a
1b
1c
1d
1e
1f
0.13
0.62
0.79
0.79
0.81
1.08
0.87
1.08
1.08
>2.00
>2.00
0.84
1.06
>2.00
0.22
0.74
0.77
<−2.80
<−2.80
<−2.80
<−2.80
−2.86
−2.79
−2.48
<−2.80
−2.53
−2.56
−2.63
−2.85
−2.38
−2.11
−2.71
<−2.80
<−2.80
−0.37
−0.07
−0.02
0.03
−2.40
−2.30
−2.25
−2.17
−2.18
−1.90
−1.87
−1.84
−1.56
−1.27
−1.20
−1.99
−1.61
−1.08
−2.02
−2.02
−2.45
0.06
0.20
0.27
0.31
0.55
0.39
0.53
0.07
0.29
1g
1h
2c
2h
2i
0.13
0.24
0.50
0.76
0.87
1.02
0.61
0.88
3d
3h
4h
5a
5b
6c
0.40
0.61
1.35
b
−0.13
0.05
0.09
0.28
0.22
−0.02
a
All values are reported against the ferrocene/ferrocenium redox
b
a2
couple as external standard. Ep = 0.64 V.
G
DOI: 10.1021/ja513027s
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
at 405 nm with the initial slope method gave a surprisingly high
value of ΦAB,eff = 0.40. Of course, this value is highly dependent
on concentrations, light intensity, and triplet lifetimes due to the
bimolecular triplet energy-transfer process. Nevertheless, the fact
that the effective quantum yield of the sensitized ring closure in
this experiment is as high as the quantum yield for ring closure by
direct excitation with 313 nm light (ΦAB = 0.43, see Table 1)
shows that the cyclization reaction of 1c within the triplet state is
highly efficient. Kinetic traces recorded during irradiation of 1c
alone with 313 nm light, and of the mixture of 1c and BA with
405 nm light, both show the emergence of a (pseudo)PSS with a
subsequent decrease of the absorbance in the visible range.
However, UPLC analysis of the irradiated samples reveals that
the byproduct is exclusively formed upon excitation with 313 nm
light (Figure S28b). Note that the decrease in absorbance during
the sensitized cyclization can be attributed to residual traces of
oxygen in the degassed solution, being transformed into singlet
oxygen by the sensitizer and thus unspecifically degrading the
diarylethene.
Figure 4. Energies (including zero-point corrections) of ring-closed
isomers and byproducts, relative to the corresponding ring-open
isomers, as obtained from DFT calculations at the B3LYP/6-31G(d)
level of theory. Calculations were also performed on the hypothetical
derivatives 1i, 2d, and 3c.
In a second experiment, methylene blue (MB, ΦISC = 0.5239)
was used to sensitize the formation of the triplet selectively for
the ring-closed isomer of 1c, whose triplet energy was calculated
to lie far below that of MB (Table S3). Importantly, MB can be
selectively excited at longer wavelengths than 1c_closed, thus
preventing (singlet) resonance energy transfer. As expected,
during irradiation of a mixture of 1c_open and MB or a mixture
of 1c_closed and MB with 654 nm light, no changes in the
absorbance spectra could be observed (Figure S29). While triplet
energy transfer from MB to 1c_open is not possible, and thus no
reaction can take place, the latter result points to the inability of
the triplet excited state of the ring-closed isomer either to
undergo ring opening or to form the byproduct. For the ring-
opening reaction, this has been attributed earlier to the presence
of a barrier on the triplet reaction pathway.37b From the
photochemistry of the singlet excited state of the ring-closed
isomer, it can be deduced that the barrier for byproduct
formation is even higher than the barrier for ring opening,12 as
excess energy in terms of excitation with UV light is needed. In
analogy to ring opening, byproduct formation in the triplet
excited state following the same pathway would not be expected.
The singlet and triplet pathways for the isomerization and
byproduct formation of diarylethenes are summarized in
Scheme 4.
stabilities in the excited state and reaction barriers will be needed
to rationalize the experimental trends.
Byproduct Formation via Triplet Excited States? From
the theoretical standpoint,12 it seems possible that byproduct
formation is induced by intersystem crossing of the excited state,
facilitated by the presence of sulfur atoms, and subsequent
relaxation of the triplet species to a different conical intersection
than that leading to ring opening. To address this question, the
fatigue behavior of compound 1c in the presence of triplet
sensitizers was investigated. It is well known in the literature that
diarylethenes covalently bound to diverse transition metal
complexes undergo efficient ring closure via metal-to-ligand
charge-transfer (MLCT) excitation, subsequent intersystem
crossing, and transfer of the triplet energy to the photochromic
core.37 In an intermolecular fashion, cyclization of a diarylethene
via its triplet state has been achieved using xanthone as an organic
sensitizer.38 In our work, we chose butane-2,3-dione (biacetyl,
BA) as an intermolecular sensitizer undergoing quantitative
intersystem crossing (ΦISC = 139). It can be selectively excited at
405 nm in the presence of the ring-open or ring-closed isomers of
1c. Although its absorption is red-shifted relative to that of the
ring-open isomer of 1c, the triplet energy of BA is within the
range of the triplet energy of 1c_open, which was estimated by
TD-DFT calculation (Table S3). Thus, effective triplet energy
transfer from BA to the ring-open isomer of 1c can be expected.
In fact, irradiation of a thoroughly degassed mixture of
1c_open and excess BA in acetonitrile with 405 nm light results
in efficient formation of the ring-closed isomer (Figure S26).
Importantly, when irradiating 1c alone or the mixture of 1c and
BA in the presence of oxygen, no changes in the absorption
spectra can be observed. A significant difference in the rate of
conversion can be observed comparing a sample that was
thoroughly degassed by freeze−pump−thaw cycles and one that
was only treated with a stream of argon for 10 min (Figure S28a).
The observed impact of traces of oxygen and the fact that, in the
presence of the ring-open isomer of 1c, the phosphorescence of
BA is quenched (Figure S27) prove the involvement of triplet
species and the efficient energy transfer from the BA moiety to
the diarylethene. Due to the low absorbance of BA, the absolute
rate of conversion to the ring-closed isomer via the triplet
pathway is relatively low compared to that of the isomerization of
1c alone with 313 nm light (Figure S28a). However,
determination of an effective quantum yield for the cyclization
Scheme 4. Singlet and Triplet Pathways for the Isomerization
and Byproduct Formation of Diarylethenes (S = Triplet
Sensitizer)
H
DOI: 10.1021/ja513027s
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
These results show that the sensitized cyclization of a
diarylethene via its triplet state, in combination with low-energy
excitation of the ring-closed isomer for the reverse ring-opening
reaction, may represent an alternative way to guarantee high
fatigue resistance while efficiently operating the switch, given that
strict oxygen-free conditions are applied.
ASSOCIATED CONTENT
* Supporting Information
Details on synthesis and characterization; photochemical and
electrochemical procedures; NMR spectra, UV/vis spectra, and
cyclic voltammograms of all compounds; and details on quantum
mechanical calculations. This material is available free of charge
via the Internet at http://pubs.acs.org.
■
S
CONCLUSION
■
AUTHOR INFORMATION
Corresponding Author
*sh@chemie.hu-berlin.de
Notes
■
A detailed investigation of the photochemistry of a series of
diarylethenes, systematically varying the nature of the hetaryl
moieties, the substituents on the adjacent phenyl rings, and the
bridging moiety, allowed for the identification of the previously
known annulated byproduct as the major source of fatigue
appearing ubiquitously over the broad range of structures.
Photokinetic measurements and evaluation of quantum yields for
the elementary photochemical processes revealed a strong
dependency of the rate of byproduct formation on the electronic
properties of the substituents. It appears that when more and
stronger electron-accepting units are introduced to the parent
structure, less fatigue due to formation of the annulated isomer is
observed. While the stabilizing effect of the perfluoro-
cyclopentene compared to perhydrocyclopentene bridge was
recognized earlier, we found that, in the latter structures, an
equally high fatigue resistance can be imparted by substitution
with 3,5-bis(trifluoromethyl)phenyl groups. By exchanging CF3
with SF5 groups, a further improvement can be observed, making
compound 2i the diarylethene structure with the highest fatigue
resistance in our hands thus far. Importantly, CF3 or SF5 groups
are chemically inert, and their introduction into diarylethenes
does not alter their efficient isomerization between the ring-open
and ring-closed isomers.
While our results show a clear correlation between the
electronic nature of the substituents and the fatigue behavior of
diarylethene photochromes, they are only phenomenological at
the current point. From a mechanistic standpoint, substitution of
the diarylethene core with donor or acceptor groups may
alternate the relative stability of intermediates or transition-state
structures on pathways leading to either ring-opening or
byproduct formation after excitation of the ring-closed isomer.12
However, for obtaining a deeper understanding, the observation
of byproduct formation using time-resolved spectroscopic
techniques together with a theoretical rationalization supported
by high-level quantum mechanical calculations of the excited-
state potential energy surface will be needed.
On the basis of our experimental results, we propose a strategy
for imparting excellent fatigue resistance to diarylethenes while
allowing a high flexibility in the choice of the components out of
which the photochromic system may be assembled. Non-
symmetrically substituted derivatives 5a and 5b show that it is
sufficient to substitute one hetaryl ring with a 3,5-bis(trifluoro-
methyl)phenyl group, leaving the other aryl ring and the bridging
moiety of the diarylethene structure open to modifications.
Notably, the reported compounds demonstrate that the
performance of dithienylperhydrocyclopentenes, which typically
show a low reversibility in their switching behavior, can be
massively improved. Experiments using an organic sensitizer to
induce the cyclization reaction of the ring-open isomer via its
triplet state revealed that the byproduct is not formed at all under
these conditions, rendering triplet-sensitized isomerization an
attractive alternative for the operation of diarylethenes.
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The authors are grateful to Jana Hildebrandt for synthetic
support and Dr. Beatrice Braun for obtaining the single-crystal X-
ray structure of 3d. The authors thank Prof. Klaus Meerholz
(University of Cologne) for stimulating discussions about the
potential involvement of triplet states in byproduct formation.
Generous support by the German Research Foundation (DFG
via SFB 658, project B8) and the European Research Council
(ERC via ERC-2012-STG_308117 “Light4Function”) is grate-
fully acknowledged. BASF AG, Bayer Industry Services, and
Sasol Germany are thanked for generous donations of chemicals.
REFERENCES
■
(1) (a) Gostl, R.; Senf, A.; Hecht, S. Chem. Soc. Rev. 2014, 43, 1982−
̈
1996. (b) Neilson, B. M.; Bielawski, C. W. ACS Catal. 2013, 3, 1874−
1885. (c) Stoll, R. S.; Hecht, S. Angew. Chem., Int. Ed. 2010, 49, 5054−
5075.
(2) (a) Raymo, F. M. Phys. Chem. Chem. Phys. 2013, 15, 14840−14850.
(b) Fukaminato, T. J. Photochem. Photobiol., C 2011, 12, 177−208.
(c) Cusido, J.; Deniz, E.; Raymo, F. M. Eur. J. Org. Chem. 2009, 2009,
2031−2045. (d) Heilemann, M.; Dedecker, P.; Hofkens, J.; Sauer, M.
Laser Photon. Rev. 2009, 3, 180−202.
́
(3) (a) Szymanski, W.; Beierle, J. M.; Kistemaker, H. A. V.; Velema, W.
A.; Feringa, B. L. Chem. Rev. 2013, 113, 6114−6178. (b) Natali, M.;
Giordani, S. Chem. Soc. Rev. 2012, 41, 4010−4029. (c) Yagai, S.;
Kitamura, A. Chem. Soc. Rev. 2008, 37, 1520−1529.
(4) (a) Kim, T.; Zhu, L.; Al-Kaysi, R. O.; Bardeen, C. J. ChemPhysChem
2014, 15, 400−414. (b) Bleger, D.; Yu, Z.; Hecht, S. Chem. Commun.
2011, 47, 12260−12266.
(5) (a) Orgiu, E.; Samorì, P. Adv. Mater. 2014, 26, 1827−1845.
(b) Shallcross, R. C.; Zacharias, P.; Kohnen, A.; Korner, P. O.; Maibach,
̈
̈
E.; Meerholz, K. Adv. Mater. 2013, 25, 469−476. (c) Tsujioka, T.; Irie,
M. J. Photochem. Photobiol., C 2010, 11, 1−14.
(6) (a) Irie, M.; Fukaminato, T.; Matsuda, K.; Kobatake, S. Chem. Rev.
2014, 114, 12174−12277. (b) Irie, M. Chem. Rev. 2000, 100, 1685−
1716.
(7) Hanazawa, M.; Sumiya, R.; Horikawa, Y.; Irie, M. J. Chem. Soc.,
Chem. Commun. 1992, 206−207.
(8) Irie, M.; Lifka, T.; Uchida, K.; Kobatake, S.; Shindo, Y. Chem.
Commun. 1999, 747−750.
(9) (a) Shoji, H.; Kobatake, S. Chem. Commun. 2013, 49, 2362−2364.
(b) Liu, H.-h.; Chen, Y. New J. Chem. 2012, 36, 2223−2227.
(c) Higashiguchi, K.; Matsuda, K.; Yamada, T.; Kawai, T.; Irie, M.
Chem. Lett. 2000, 29, 1358−1359.
(10) (a) Hirose, T.; Inoue, Y.; Hasegawa, J.-y.; Higashiguchi, K.;
Matsuda, K. J. Phys. Chem. A 2014, 118, 1084−1093. (b) Harvey, E. C.;
Areephong, J.; Cafolla, A. A.; Long, C.; Browne, W. R.; Feringa, B. L.;
Pryce, M. T. Organometallics 2014, 33, 447−456. (c) Sakano, T.;
Imaizumi, Y.; Hirose, T.; Matsuda, K. Chem. Lett. 2013, 42, 1537−1539.
(d) Ikeda, H.; Sakai, A.; Kawabe, A.; Namai, H.; Mizuno, K. Tetrahedron
Lett. 2008, 49, 4972−4976. (e) Sevez, G.; Gan, J.; Pan, J.; Sallenave, X.;
Colin, A.; Saadoui, H.; Saleh, A.; Vogtle, F.; Pozzo, J.-L. J. Phys. Org.
̈
I
DOI: 10.1021/ja513027s
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Chem. 2007, 20, 888−893. (f) Peters, A.; Branda, N. R. Adv. Mater. Opt.
Electron. 2000, 10, 245−249. (g) Higashiguchi, K.; Matsuda, K.;
Kobatake, S.; Yamada, T.; Kawai, T.; Irie, M. Bull. Chem. Soc. Jpn. 2000,
73, 2389−2394.
(35) (a) Staykov, A.; Areephong, J.; R. Browne, W.; L. Feringa, B.;
Yoshizawa, K. ACS Nano 2011, 5, 1165−1178. (b) Browne, W. R.; de
Jong, J. J. D.; Kudernac, T.; Walko, M.; Lucas, L. N.; Uchida, K.; van
Esch, J. H.; Feringa, B. L. Chem.Eur. J. 2005, 11, 6414−6429.
(c) Guirado, G.; Coudret, C.; Hliwa, M.; Launay, J.-P. J. Phys. Chem. B
2005, 109, 17445−17459.
(36) (a) Schleyer, P. v. R.; Manoharan, M.; Wang, Z.-X.; Kiran, B.; Jiao,
H.; Puchta, R.; van Eikema Hommes, N. J. R. Org. Lett. 2001, 3, 2465−
2468. (b) Schleyer, P. v. R.; Maerker, C.; Dransfeld, A.; Jiao, H.;
Hommes, N. J. R. v. E. J. Am. Chem. Soc. 1996, 118, 6317−6318.
(37) (a) Monaco, S.; Semeraro, M.; Tan, W.; Tian, H.; Ceroni, P.;
Credi, A. Chem. Commun. 2012, 48, 8652−8654. (b) Indelli, M. T.;
Carli, S.; Ghirotti, M.; Chiorboli, C.; Ravaglia, M.; Garavelli, M.;
Scandola, F. J. Am. Chem. Soc. 2008, 130, 7286−7299. (c) Ko, C.-C.;
Kwok, W.-M.; Yam, V. W.-W.; Phillips, D. L. Chem.Eur. J. 2006, 12,
5840−5848. (d) Jukes, R. T. F.; Adamo, V.; Hartl, F.; Belser, P.; De Cola,
L. Inorg. Chem. 2004, 43, 2779−2792.
(11) Ikeda, H.; Kawabe, A.; Sakai, A.; Namai, H.; Mizuno, K. Res. Chem.
Intermed. 2009, 35, 893−908.
(12) Mendive-Tapia, D.; Perrier, A.; Bearpark, M. J.; Robb, M. A.;
Lasorne, B.; Jacquemin, D. Phys. Chem. Chem. Phys. 2014, 16, 18463−
18471.
(13) de Jong, J. J. D.; Lucas, L. N.; Hania, R.; Pugzlys, A.; Kellogg, R.
M.; Feringa, B. L.; Duppen, K.; van Esch, J. H. Eur. J. Org. Chem. 2003,
2003, 1887−1893.
(14) Patel, P. D.; Mikhailov, I. A.; Belfield, K. D.; Masunov, A. E. Int. J.
Quantum Chem. 2009, 109, 3711−3722.
(15) (a) Jeong, Y.-C.; Park, D. G.; Kim, E.; Ahn, K.-H.; Yang, S. I.
Chem. Commun. 2006, 1881−1883. (b) Jeong, Y.-C.; Yang, S. I.; Kim, E.;
Ahn, K.-H. Tetrahedron 2006, 62, 5855−5861.
(38) Murata, R.; Yago, T.; Wakasa, M. Bull. Chem. Soc. Jpn. 2011, 84,
1336−1338.
(39) Montalti, M.; Credi, A.; Prodi, L.; Gandolfi, M. T. Handbook of
Photochemistry, 3rd ed.; CRC Press: Boca Raton, FL, 2006; pp 83−351.
(16) (a) Asadirad, A. M.; Boutault, S.; Erno, Z.; Branda, N. R. J. Am.
Chem. Soc. 2014, 136, 3024−3027. (b) Neilson, B. M.; Bielawski, C. W.
Chem. Commun. 2013, 49, 5453−5455. (c) Neilson, B. M.; Bielawski, C.
W. J. Am. Chem. Soc. 2012, 134, 12693−12699. (d) Lemieux, V.;
Spantulescu, M.; Baldridge, K.; Branda, N. Angew. Chem., Int. Ed. 2008,
47, 5034−5037.
(17) (a) Nakashima, T.; Kajiki, Y.; Fukumoto, S.; Taguchi, M.; Nagao,
S.; Hirota, S.; Kawai, T. J. Am. Chem. Soc. 2012, 134, 19877−19883.
(b) Fukumoto, S.; Nakashima, T.; Kawai, T. Angew. Chem., Int. Ed. 2011,
50, 1565−1568.
(18) (a) Gostl, R.; Hecht, S. Angew. Chem., Int. Ed. 2014, 53, 8784−
̈
8787. (b) Cahova,
3190.
(19) Szalok
́
H.; Jaschke, A. Angew. Chem., Int. Ed. 2013, 52, 3186−
̈
́
i, G.; Pozzo, J.-L. Chem.Eur. J. 2013, 19, 11124−11132.
(20) Herder, M.; Utecht, M.; Manicke, N.; Grubert, L.; Patzel, M.;
̈
Saalfrank, P.; Hecht, S. Chem. Sci. 2013, 4, 1028−1040.
(21) Hiroto, S.; Suzuki, K.; Kamiya, H.; Shinokubo, H. Chem. Commun.
2011, 47, 7149−7151.
(22) Orgiu, E.; Crivillers, N.; Herder, M.; Grubert, L.; Patzel, M.;
̈
Frisch, J.; Pavlica, E.; Duong, D. T.; Bratina, G.; Salleo, A.; Koch, N.;
Hecht, S.; Samorì, P. Nat. Chem. 2012, 4, 675−679.
(23) Lucas, L. N.; de Jong, J. J. D.; van Esch, J. H.; Kellogg, R. M.;
Feringa, B. L. Eur. J. Org. Chem. 2003, 2003, 155−166.
(24) Irie, M.; Lifka, T.; Kobatake, S.; Kato, N. J. Am. Chem. Soc. 2000,
122, 4871−4876.
(25) Chung, J. W.; Yoon, S.-J.; Lim, S.-J.; An, B.-K.; Park, S. Y. Angew.
Chem., Int. Ed. 2009, 48, 7030−7034.
(26) Gostl, R.; Kobin, B.; Grubert, L.; Patzel, M.; Hecht, S. Chem.
̈
̈
Eur. J. 2012, 18, 14282−14285.
(27) Uchida, K.; Ishikawa, T.; Takeshita, M.; Irie, M. Tetrahedron 1998,
54, 6627−6638.
(28) (a) Maurel, F.; Perrier, A.; Jacquemin, D. Int. J. Quantum Chem.
2012, 112, 1122−1133. (b) Jacquemin, D.; Perpet
Perrier, A. l. J. Phys. Chem. C 2010, 114, 9489−9497. (c) Jacquemin, D.;
Perpete, E. A.; Ciofini, I.; Adamo, C. Acc. Chem. Res. 2008, 42, 326−334.
(29) Jacquemin, D.; Perpete, E. A.; Scuseria, G. E.; Ciofini, I.; Adamo,
̀
e, E. A.; Maurel, F. o.;
̀
̀
C. J. Chem. Theory Comput. 2007, 4, 123−135.
(30) (a) Delbaere, S.; Vermeersch, G.; Micheau, J.-C. J. Photochem.
Photobiol., C 2011, 12, 74−105. (b) Higashiguchi, K.; Matsuda, K.; Irie,
M. Angew. Chem., Int. Ed. 2003, 42, 3537−3540. (c) Deniel, M. H.;
Lavabre, D.; Micheau, J. C. In Organic Photochromic and Thermochromic
Compounds; Crano, J., Guglielmetti, R., Eds.; Kluwer Academic: New
York, 1999; Vol. 2, pp 167−209.
(31) Maafi, M.; Brown, R. G. Photochem. Photobiol. Sci. 2008, 7, 1360−
1372.
(32) Irie, M.; Sakemura, K.; Okinaka, M.; Uchida, K. J. Org. Chem.
1995, 60, 8305−8309.
(33) (a) Santos, A. R.; Ballardini, R.; Belser, P.; Gandolfi, M. T.; Iyer, V.
M.; Moggi, L. Photochem. Photobiol. Sci. 2009, 8, 1734−1742. (b) Sumi,
T.; Takagi, Y.; Yagi, A.; Morimoto, M.; Irie, M. Chem. Commun. 2014,
50, 3928−3930.
(34) Sheppard, W. A. J. Am. Chem. Soc. 1962, 84, 3072−3076.
J
DOI: 10.1021/ja513027s
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX