Improved solid-state photomechanical materials by fluorine substitution of 9-anthracene carboxylic acid

Article abstract of DOI:10.1021/cm502866e

Four fluorinated derivatives of 9-anthracene carboxylic acid (9AC), a molecule that shows a reversible photomechanical response in its crystal form, have been synthesized and characterized. The spectroscopic properties and crystal structures of 4-fluoro-9-anthracene carboxylic acid (4F-9AC), 2-fluoro-9-anthracene carboxylic acid (2F-9AC), 10-fluoro-9-anthracene carboxylic acid (10F-9AC), and 2,6-difluoro-9-anthracene carboxylic acid (2,6DF-9AC) are all very similar to those of 9AC. However, their photomechanical properties vary widely. 405 nm light was used to induce [4 + 4] photodimerization and a mechanical response in crystalline microneedles and ribbons. Both the photodimer dissociation rate and the mechanical recovery varied by more than an order of magnitude, with 4F-9AC exhibiting the most rapid recovery time, on the order of 30 s. Nanoindentation measurements show that this crystal has a slightly reduced elastic modulus and a significantly reduced hardness, making it less brittle than the 9AC crystal. Large 4F-9AC crystals remain intact after irradiation, without fragmenting, while microneedles can undergo more than 100 mechanical bending cycles. Given the similarity of the crystal packing in all five molecules, the improved photomechanical properties must arise from subtle changes in intermolecular interactions or possibly differences in disorder. These results demonstrate that it is possible to significantly improve the properties of photomechanical materials through small modifications of the molecular structure.

Full text of DOI:10.1021/cm502866e

Article
pubs.acs.org/cm
Improved Solid-State Photomechanical Materials by Fluorine
Substitution of 9‑Anthracene Carboxylic Acid
Lingyan Zhu,† Fei Tong,† Christopher Salinas,‡ Muhanna K. Al-Muhanna,§ Fook S. Tham,†
David Kisailus,‡,∥ Rabih O. Al-Kaysi,*,⊥ and Christopher J. Bardeen*,†
†Department of Chemistry, University of California, Riverside, 501 Big Springs Road, Riverside, California 92521, United States
‡Department of Chemical and Environmental Engineering, University of California, Riverside, 900 University Avenue, Riverside,
California 92521, United States
§King Abdulaziz City for Science and Technology, P.O Box 6086, Riyadh 11442, Kingdom of Saudi Arabia
∥Materials Science and Engineering Program University of California, Riverside, 900 University Avenue, Riverside, California 92521,
United States
⊥Department of Basic Science, College of Science and Health Professions-3124, King Saud bin Abdulaziz University for Health
Sciences, and King Abdullah International Medical Research Center, Ministry of National Guard Health Affairs, Riyadh 11426,
Kingdom of Saudi Arabia
S
* Supporting Information
ABSTRACT: Four fluorinated derivatives of 9-anthracene carboxylic
acid (9AC), a molecule that shows a reversible photomechanical
response in its crystal form, have been synthesized and characterized.
The spectroscopic properties and crystal structures of 4-fluoro-9-
anthracene carboxylic acid (4F-9AC), 2-fluoro-9-anthracene carboxylic
acid (2F-9AC), 10-fluoro-9-anthracene carboxylic acid (10F-9AC), and
2,6-difluoro-9-anthracene carboxylic acid (2,6DF-9AC) are all very
similar to those of 9AC. However, their photomechanical properties vary
widely. 405 nm light was used to induce [4 + 4] photodimerization and a
mechanical response in crystalline microneedles and ribbons. Both the
photodimer dissociation rate and the mechanical recovery varied by more
than an order of magnitude, with 4F-9AC exhibiting the most rapid
recovery time, on the order of 30 s. Nanoindentation measurements
show that this crystal has a slightly reduced elastic modulus and a significantly reduced hardness, making it less brittle than the
9AC crystal. Large 4F-9AC crystals remain intact after irradiation, without fragmenting, while microneedles can undergo more
than 100 mechanical bending cycles. Given the similarity of the crystal packing in all five molecules, the improved
photomechanical properties must arise from subtle changes in intermolecular interactions or possibly differences in disorder.
These results demonstrate that it is possible to significantly improve the properties of photomechanical materials through small
modifications of the molecular structure.
without a host, can also function as photomechanical materials
and have several potential advantages.5−11 From a mechanistic
standpoint, the structure of molecular crystals can be precisely
characterized using X-ray diffraction, making it possible to
study in detail how reaction dynamics lead to crystal
deformation. Molecular crystals tend to have higher elastic
moduli than polymers.12 Irie and co-workers have shown that
the photomechanical response of diarylethene crystals can be
very fast and persist even at very low temperatures.13−15
In order to use a molecular crystal as a reversible
photoactuator, the underlying photochemistry must be
reversible. For example, the ring closing reaction of diary-
INTRODUCTION
■
Photons can be transformed into mechanical motion by
photochemical reactions in which molecules change shape or
bonding configuration. If photochemically reactive molecules
are organized, for example in a molecular crystal or a liquid
crystalline polymer, then each photochemical reaction can
“push” in the same direction. The sum of many such
displacements, on the angstrom scale, can result in a large
total displacement on the scale of micrometers or more. This is
the origin of photomechanical effects in molecular materials.1
Such materials have potential applications as photocontrolled
actuators and as elements of photon-fueled machines. Photo-
mechanical materials where the photoreactive molecule is
embedded in a polymer host have been developed and have
demonstrated interesting photomechanical properties.2−4 Neat
molecular crystals, composed solely of photoreactive molecules
Received: August 4, 2014
Revised: September 10, 2014
© XXXX American Chemical Society
A
dx.doi.org/10.1021/cm502866e | Chem. Mater. XXXX, XXX, XXX−XXX

Chemistry of Materials
Article
lethene photochromes can be initiated by ultraviolet light, while
the reverse reaction requires illumination at visible wavelengths
to return the ring to its open form.16 The use of a single light
source to induce repeated deformations simplifies the operation
of such actuators but requires a photoreaction that can be
reversed by ambient thermal fluctuations. One such reaction is
the [4 + 4] photodimerization of crystalline 9-anthracene
carboxylic acid (9AC), which generates a metastable head-to-
head dimer that spontaneously dissociates to the monomer pair
at room temperature.17 This reaction formed the basis of
molecular crystal nanorods and microribbons that could
reversibly bend and twist in response to a pulse of
monochromatic light.18−20 As a practical photoactuator ma-
terial, however, 9AC has several drawbacks. First, this molecule
is not particularly photostable, perhaps due to endoperox-
idation reactions with ambient O2, and typically a crystal can
only undergo 10 cycles before a significant amount of response
is lost.18 Second, the slow dimer dissociation time (∼400 s)
means that the duty cycle of an actuator based on this molecule
will be low. Third, 9AC crystals are relatively brittle, and
crystals with thicknesses greater than about 10 μm tend to
crack or shatter under illumination. We and others have shown
that the use of nanoscale crystals provides a general solution to
the problem of fragmentation,21−23 but it is desirable to find a
material which has a useful photomechanical response over a
larger size range. To circumvent these problems, we recently
explored a family of 9AC derivatives with substituents at the
10-position, opposite the COOH group.24 While none of those
derivatives improved upon the performance of 9AC, 10-fluoro-
9-anthracene carboxylic acid (10F-9AC) did exhibit a reversible
photomechanical response, albeit much slower than 9AC.
In this paper, we continue our pursuit of improved
derivatives of 9AC. We have concentrated our efforts on
fluoro-substituted derivatives, since this atom is close in size to
hydrogen and should minimize size-induced disruptions of the
crystal lattice. At the same time, the presence of electron-
withdrawing groups should lower the HOMO and LUMO
levels, increasing the stability of the molecule with respect to
oxidative side reactions.25 The goal of the research is to identify
molecules that have (a) a faster dissociation and mechanical
recovery rate and (b) greater photostability. With this in mind,
we undertook a comparative study of the 9AC derivatives
shown in Figure 1. We have previously reported results on 9AC
and 10F-9AC but the other three molecules, 4-fluoro-9-
anthracene carboxylic acid (4F-9AC), 2-fluoro-9-anthracene
carboxylic acid (2F-9AC), and 2,6-difluoro-9-anthracene
carboxylic acid (2,6dF-9AC), are newly synthesized derivatives.
For completeness, in this paper we consider all five molecules
as a group. We find that the spectroscopic properties of the
molecules in both solution and solid-state are very similar. The
behaviors of these molecules diverge strongly when their
photomechanical properties are considered, however. The most
promising molecule, 4F-9AC, has a back-reaction rate that is
almost 1 order of magnitude faster than that of 9AC, as well as
higher photostability with the ability to undergo >100
photomechanical cycles without loss of function. Using
nanoindentation measurements, we find that crystals composed
of 4F-9AC have a similar elastic modulus to 9AC, but are a
factor of 2 less hard. This decreased hardness allows larger 4F-
9AC crystals to remain intact while they undergo photo-
mechanical deformations. 4F-9AC retains the properties of
9AC that make it a photomechanical material (head-to-head
crystal packing mediated by hydrogen-bonding interactions)
Figure 1. Molecular structures of anthracene-9-carboxylic acid (9AC)
and its fluorinated derivatives: (a) anthracene-9-carboxylic acid (9AC);
(b) 4-fluoro-anthracene-9-carboxylic acid (4F-9AC); (c) 2-fluoro-
anthracene-9-carboxylic acid (2F-9AC); (d) 10-fluoro-anthracene-9-
carboxylic acid (10F-9AC); and (e) 2,6-difluoro-anthracene-9-
carboxylic acid (2,6DF-9AC).
but is significantly superior to 9AC in terms of properties that
would be desired for a practical photoactuator. The results of
this paper show that it is possible to systematically vary the
molecular structure to improve the solid-state photomechanical
properties.
EXPERIMENTAL SECTION
■
Synthesis of Fluoro-9-anthracene Carboxylic Acid Deriva-
tives. All starting materials were purchased from the Tokyo Chemical
Industry Company (TCI) and used without further purification. Dry
diethyl ether was distilled over lithium aluminum hydride and stored
over activated molecular sieves (4 Å). All other organic solvents were
distilled over anhydrous CaCl2 prior to use. Glassware was kept dry in
an oven at 60 °C prior to use. Melting points (uncorrected) were
measured on a 1101D Mel-Temp digital melting point apparatus. IR
measurements were performed using IR Affinity-1 FTIR from
Shimadzu. GC-MS data were collected using Shimadzhu GCMS-
QP2010 using a 30 m long general-purpose capillary GC column.
4-Fluoro-9-anthracene Carboxylic Acid (4F-9AC). The detailed
synthesis of 4F-9AC is depicted in Scheme 1. The syntheses of the
other fluoro-9-anthracene carboxylic acid derivatives are similar and
are described in the Supporting Information.
1-Anthraquinone Diazonium Tetrafluoroborate 1c.26 In a
250 mL Erlenmeyer flask, 1-aminoanthraquinone 1a (3 g, 0.013 mol)
was dissolved in 98% H2SO4 (50 mL) at room temperature to give a
dark brown solution. NaNO2 (2.5 g, 0.036 mol), previously dissolved
in ice-cold 98% H2SO4 (25 mL), was added slowly to 1a in H2SO4 at 0
°C. The temperature was kept at <5 °C throughout the addition
process, which took 10−15 min. The reaction mixture was stirred at
<5 °C for 1 h before pouring the contents over 400 mL of crushed ice.
1-Anthraquinone diazonium hydrogen sulfate 1b precipitated out as a
sparingly soluble pink solid. To the aqueous suspension of 1b, NaBF4
(3 g, 0.028 mol) was added and stirred at 0 °C for 2 h to precipitate
out the insoluble tetrafluoroborate salt 1c. The pinkish solid was
filtered over crushed ice and then washed with ice-cold water (50 mL)
followed by cold ethanol (10 mL). The powder was rinsed with
hexane (20 mL) before leaving it to air-dry at room temperature for
several hours. We obtained 1c as a white pink powder (3.6 g, 84%
yield).
1-Fluoro-anthraquinone 1d. Compound 1c (3.6 g, 0.011 mol)
was mixed with six times its weight in sand (20 g), placed on a 25 cm
× 25 cm peace of aluminum foil, and folded into a flat puck-shaped
disk. The disk was placed flat inside a 500 mL beaker and decomposed
at 300 °C for 24 h. The beaker was covered with aluminum foil to
collect the sublimed 1d as large red needle-like crystals (2.0 g, 78%
yield) with a mp of 230−231 °C (literature mp: 230−232 °C27). The
B
dx.doi.org/10.1021/cm502866e | Chem. Mater. XXXX, XXX, XXX−XXX

Chemistry of Materials
Article
vacuum (200 mbar) at 90 °C for several hours to remove any trace of
water in preparation for the lithiation step.
Scheme 1. Synthetic Scheme for Compound 4-Fluoro-9-
anthracene Carboxylic Acid (4F-9AC)
4-Fluoro-9-anthracene Carboxylic Acid (4F-9AC). In a 50 mL
oven-dried round-bottom flask, dry diethyl ether (10 mL) was added
to 1-bromopropane (0.54 g, 0.4 mL, 0.0044 mol). Freshly cut lithium
wire (0.18 g, 0.03 mol) was flattened and then stirred with the 1-
bromopropane in diethyl ether under an argon atmosphere to prepare
n-propyllithium. Unreacted lithium was fished out and reweighed to
confirm conversion of most of the 1-bromopropane to its lithium salt.
The n-propyllithium/diethyl ether solution was cooled to −30 °C
using Lab Armor dry bath beads before adding 1f (0.2 g, 7.3 × 10−4
moles) and stirring for 20 min to produce 1g. While maintaining the
temperature at −30 °C, CO2 gas was bubbled through the mixture for
20 min using two 1 L balloons inflated with dry CO2. The reaction was
quenched by adding it to crushed ice (50 mL) followed by extracting
the insoluble organic residue with ethyl acetate (50 mL) followed by
hexanes (2 × 50 mL). The acid was precipitated out of the aqueous
solution by adding concentrated HCl (3 mL). The yellow precipitate
was suction filtered and recrystallized from boiling ethanol (10 mL)
and water (10 mL). We obtained a yellow powder of 1h (0.13 g, 77%
1
yield) with a mp of 257−258 °C. H NMR (CDCl3) δ (ppm): 7.37
(1H, dd), 7.61 (3H, m), 7.84 (1H, d), 8.03 (1H, d), 8.26 (1H, d), 8.86
(1H, s). νmax (KBr disc) cm−1: 3500−2800 (broad, COOH), 1682
(CO), 1639 (CC), 1559, 1457, 1278, 1249, 720.
X-ray Structure Determination. Crystals of the other fluoro-9-
anthracene carboxylic acids were prepared by subliming a sample
under normal atmospheric pressure at 180 °C. They were mounted on
a cryo-loop glass fiber with paratone oil. Single-crystal X-ray diffraction
data were collected on a Bruker APEX2 platform-CCD X-ray
diffractometer system (fine focus Mo radiation, λ = 0.71073 Å,
50KV/35 mA power) at 100 K. The frames were collected for a sphere
of reflections and integrated using the Bruker SAINT software package
and using a narrow-frame integration algorithm. Absorption
corrections were applied to the raw intensity data using the SADABS
program. The Bruker SHELXTL software package was used for phase
determination and structure refinement. The 4F-9AC, 10F-9AC, and
2,6DF-9AC crystals were all found to be twinned structures. The 2F-
9AC crystal included about 20% of a 2-fluoro-anthracene byproduct,
likely the result of decomposition during the sublimation process. The
10F-9AC crystal had about 5% of a residual 10-bromo-9-anthracene
carboxylic byproduct left over from the synthesis. Both 4F-9AC and
2,6DF-9AC had significant disorder in the fluorine positions due to
different orientations of the anthracene in the stack. All these
complications were taken into account in the diffraction analysis, and
details of the analysis can be found in the Supporting Information.
Absorption and Fluorescence Spectroscopy. Ultraviolet−
visible absorption spectra of solution samples are collected using a
Varian Cary 50 UV/vis spectrophotometer with background
correction. Tetrahydrofuran (THF, ACS, 99%+) was used as the
solvent. Steady-state fluorescence spectra of solutions are collected by
Spex Fluorolog Tau-3 fluorescence spectrophotometer. Peak absor-
bances are kept below 0.2 in order to avoid self-absorption artifacts.
The samples were excited at 350 nm, and the spectra were collected
from 400 to 650 nm.
crystals can be easily separated from the decomposed byproducts that
formed a black solid lump with the sand inside the aluminum foil. GC-
MS revealed that the product was 95% 1d with M+ = 226
corresponding to the molecular ion peak and fragments at 198 (loss
of one CO) and 170 (loss of two CO).
1-Fluoroanthracene 1e. Compound 1d (2.0 g, 0.0088 mol) was
suspended with 27% ammonium hydroxide solution (50 mL), water
(50 mL), and CuSO4 (0.05 g) in a 500 mL round-bottom flask. Zn
powder (4 g, 0.0625 g), prewashed with acid, was added to the mixture
and stirred vigorously at 80 °C. The round-bottom flask was fitted
with a six segment Snyder column that allows excess ammonia gas to
escape while stopping air from flowing back inside the flask. The
reaction changes from deep red to a turbid white after 24 to 36 h when
all of the 1d is reduced to the corresponding anthracene 1e. The
suspension of 1e and Zn(OH)2 was filtered and stirred with HCl (200
mL, 2 M) to dissolve the Zn(OH)2. The undissolved 1e was suction
filtered and washed with water before recrystallizing from boiling 1-
propanol (30 mL) and water (30 mL). We obtained pale green
platelike crystals of 1e (0.81 g, 47% yield) with a mp of 108−109 °C
(literature mp: 108 °C28,29). 1H NMR (CDCl3) δ (ppm): 7.32 (1H, t),
7.35 (1H, m), 7.40 (1H, d), 7.48 (1H, d), 7.51 (1H, m), 7.77−8.00
(2H, dd), 8.4 (1H, s), 8.6 (1H, s). GC-MS demonstrated 98% 1e with
a major peak corresponding to M+ = 196 and minor fragment at 170.
9-Bromo-4-fluoroanthracene 1f. Compound 1e (0.80 g, 0.0041
mol), previously dried under vacuum, was dissolved in warm CCl4 (30
mL). Bromine liquid (0.72 g, 0.23 mL, 1.1 equiv) was added slowly
and allowed to react at room temperature overnight. The mixture was
filtered through silica gel (10 g, flash chromatography grade) then
eluted with 200 mL of hexane. The organic solvent was removed
under reduced pressure to yield yellow crystals of 1f (0.82 g, 72%
Time-Resolved Fluorescence Measurements. A few drops of
concentrated ethanol solution of 4F-9AC were placed on a glass
coverslip. After evaporation of the ethanol, a thin polycrystalline film
of 4F-9AC was left on the coverslip. Other samples were prepared
using the same procedure. Time-resolved fluorescence lifetime data
were collected using a Hamamastsu C4334 Streakscope. The solid
samples were kept under a vacuum in a Janis ST100 cryostat and
excited using front-face geometry. The excitation source was a 200 fs
pulse centered at 387 nm, derived from frequency doubling the 775
nm output of a Spectra Physics MaiTai Ti:sapphire oscillator. The
laser repetition rate was decreased to 1 MHz using a pulse picker to
avoid thermal effects.
1
yield) with a mp of 99−101 °C. H NMR (CDCl3) δ (ppm): 7.16
(1H, t), 7.49 (2H, m), 7.55 (1H, dd), 8.04 (1H, d), 8.30 (1H, d), 8.48
(1H, d), 8.71 (1H, s). GC-MS showed a very pure product with two
equal intensity peaks at M+ = 276 and 274, corresponding to a
monobrominated product and fragment peaks at 196 and 194
corresponding to loss of one bromine. The powder was dried under
Preparation of Crystalline Microribbons and Microneedles.
All the microcrystals were prepared using the floating drop method.30
A total of 2.1 mg of 4F-9AC was first dissolved in 1.0 mL of filtered
ethyl acetate (Sigma, 99.5+%), and then this solution was slowly added
C
dx.doi.org/10.1021/cm502866e | Chem. Mater. XXXX, XXX, XXX−XXX

Chemistry of Materials
Article
on top of Milli-Q Millipore purified H2O in a Petri dish (VWR, 60 ×
15 mm). The bilayer solution was covered and left undisturbed in the
dark for 48 h. As the ethyl acetate evaporated, 4F-9AC slowly
crystallized out as ribbons and needles floating on top of the water
surface. Other microcrystals were prepared using the same procedure.
The fluoro-9-anthracene carboxylic acids formed a majority of
crystalline needles, in contrast to 9AC.
Optical Microscopy Measurements. To prepare samples, a few
drops of 50% H3PO4 were added to water containing suspended
microcrystals in order to slow down H2O evaporation during the
measurements. A drop of the suspension was transferred to a
microscope slide and covered with a microscope coverslip. To monitor
the photomechanical motion, videos were taken using a digital camera
AMScope MU900 (resolution 9 M pixels). The crystals were irradiated
by 405 nm light for a few seconds and then allowed to recover in the
dark.
Fluorescence Recovery Measurements. A few drops of
concentrated ethanol solution of the compound were placed on a
glass coverslip. After evaporation of the ethanol, a thin polycrystalline
film was left on the coverslip. To measure the photorecovery process,
the polycrystalline film was irradiated by a 325 nm He−Cd laser for
8−18 s to photo bleach the samples around 50%. The laser power was
then attenuated by a factor of 1000 or more using neutral density
filters that were applied to prevent further photochemistry. The 3.1
ND filters were used for 9AC, 4F-9AC, and 2F-9AC, while the 3.6 ND
filters were used for 10F-9AC and 2,6DF-9AC. This still allowed the
fluorescence signal of the surviving monomer molecules to be detected
to monitor the back reaction. The signals were read out by a lock-in
amplifier and collected using a LabView computer program.
Nanoindentation Measurements. Nanomechanical testing was
performed using a TI-950 Nanoindenter (Hysitron Minneapolis, MN),
equipped with a conical indenter tip with a 1 μm radius. Two partial
unload−displacement controlled indents were performed per sample
at 20 different depths ranging from 20 to 450 nm. From these
indentation events values of hardness, H, were determined using the
relation31
Figure 2. Top and side views of crystal structures of fluorinated
anthracene-9-carboxylic acid derivatives: (a) 9AC; (b) 4F-9AC; (c)
2F-9AC; (d) 10F-9AC; and (e) 2,6DF-9AC.
P
A
all aligned to one side of the stack, which facilitates hydrogen
bonding between neighboring stacks. It is this interaction that
forces the 9AC derivatives to adopt the head-to-head stack
motif, in contrast to the head-to-tail motif seen in crystals of
most 9-substituted anthracene derivatives. The steric interfer-
ence of the two COOH groups after [4 + 4] photodimer
formation is believed to be the dominant factor in destabilizing
the dimer and leads to the reversible photochemistry of the
9AC crystal.17 All four crystals share this packing motif with
9AC, and thus we expect all four compounds to exhibit
reversible photochemistry in the crystal form, like 9AC.
Given the similarity of the crystal structures, it is not
surprising that the photophysical properties of the different
crystals are similar. The addition of a fluorine atom to an
anthracene generally has only a small effect on the anthracene’s
spectroscopic properties.32,33 In dilute solution, the steady-state
absorption spectra of all five compounds exhibit similar vibronic
peak structures in their absorption spectra, resembling
unsubstituted anthracene (Supporting Information, Figure
S1). They also show the broad emission spectrum characteristic
of anthracene carbonyl compounds, which results from a partial
charge transfer from the ring to the oxygen atoms on the
COOH group.34 The only noticeable change is a slight redshift
in the fluorinated compounds that arises from an inductive
effect from the extra electrons of the substituent atoms. As
expected, this redshift is most pronounced for 2,6DF-9AC. The
fluorescence lifetimes of the molecules in solution (Table 2) are
also similar, averaging 16 ns. The exception is 10F-9AC, whose
lifetime of 5.7 ns is shorter than the other fluorinated
derivatives. In solution, it does not appear that fluorine
H =
(1)
where P is the maximum applied load and A is the area of contact.
Values of the reduced elastic modulus, Er, were obtained using31
π
2
S
A
Er =
(2)
where S is the measured stiffness of the material upon unloading of the
indenter.
RESULTS AND DISCUSSION
The synthesis of 4F-9AC, 2F-9AC, and 2,6-DF-9AC has not
been previously reported. The structures of these compounds
were confirmed by H NMR, IR, and single crystal X-ray. All
compounds were obtained in decent yields with high (>98%)
purities as determined by H NMR. The growth of crystals
suitable for XRD measurements was somewhat challenging.
Successful deposition of XRD worthy 4F-9AC and 2,6DF-9AC
crystals was achieved via slow sublimation at 180 °C under
normal pressure. Growing crystals of 2F-9AC was challenging
due to decarboxylation at elevated temperatures, leading to
crystals of 2F-9AC doped 2-fluoro-anthracene, unsuitable for
XRD. Suitable crystals of 2F-9AC were slowly grown at 150 °C
under reduced pressure over a period of 5 days. We found that
the crystal structures of the molecules in Figure 1 are
remarkably similar. In Figure 2 we show representative views
of the crystal packing for the five compounds that are the
subject of this paper. The crystal parameters are summarized in
Table 1. From Figure 2, we can see that all four compounds
crystallize in a π-stacked arrangement. The COOH groups are
■
1
1
D
dx.doi.org/10.1021/cm502866e | Chem. Mater. XXXX, XXX, XXX−XXX

Chemistry of Materials
Article
Table 1. Crystal Parameters and X-ray Structure Refinement Parameters for 2-Fluoro-9-anthracene Carboxylic Acid (2F-9AC),
4-Fluoro-9-anthracene Carboxylic Acid (4F-9AC), and 2,6-Difluoro-9-anthracene Carboxylic Acid (2,6DF-9AC), the New
Molecules Synthesized in This Work
2F-9AC
4F-9AC
2,6DF-9AC
empirical formula C15H9FO2
C15H9.21FO2
240.44
C15H8F2O2
258.21
formula weight
temperature
wavelength
240.22
100(2) K
100(2) K
100(2) K
0.71073 Å
monoclinic
P2(1)/n (No. 14)
0.71073 Å
monoclinic
P2(1)/n (No. 14)
0.71073 Å
monoclinic
P2(1)/n (No. 14)
crystal system
space group
unit cell
dimensions
a = 3.7695(7) Å, b = 9.4431(18) Å,
c = 29.486(6) Å, α = 90°, β = 90.018(3)°,
γ = 90°
1049.6(3) Å3
a = 3.7914(4) Å, b = 9.5647(10) Å,
c = 28.905(3) Å, α = 90°, β = 90.472(2)°,
γ = 90°
1048.15(19) Å3
a = 3.7404(2) Å, b = 9.9250(7) Å,
c = 29.0381(19) Å, α = 90°, β = 90.211(1)°,
γ = 90°
1077.99(12) Å3
volume
Z
4
4
4
density
1.520 Mg/m3
1.524 Mg/m3
1.591 Mg/m3
(calculated)
absorption
0.112 mm−1
0.112 mm−1
0.128 mm−1
coefficient
F(000)
496
497
528
crystal size
0.47 × 0.30 × 0.02 mm3
0.54 × 0.05 × 0.04 mm3
0.33 × 0.06 × 0.02 mm3
theta range for
data collection
2.07 to 30.03°
1.41 to 27.10°
2.05 to 29.57°
index ranges
−5 ≤ h ≤ 5, −13 ≤ k ≤ 13, −41 ≤ l ≤ 41
−4 ≤ h ≤ 4, −12 ≤ k ≤12, −37 ≤ l ≤ 37
−5 ≤ h ≤ 5, −13 ≤ k ≤ 13, −40 ≤ l ≤ 40
reflections
22215
22651
24006
collected
independent
reflections
3076 [R(int) = 0.0928]
99.90%
2300 [R(int) = 0.0742]
100.00%
3016 [R(int) = 0.0403]
99.90%
completeness to
θ = 30.03°
absorption
correction
semiempirical from equivalents
0.9981 and 0.9494
semiempirical from equivalents
0.9960 and 0.9413
semiempirical from equivalents
0.9974 and 0.9593
max and min
transmission
refinement
method
full-matrix least-squares on F2
3076/0/177
full-matrix least-squares on F2
2300/588/236
full-matrix least-squares on F2
3016/0/196
data/restraints/
parameters
goodness-of-fit
on F2
1.094
1.047
1.067
final R indices
[I > 2σ(I)]
R1 = 0.0653, wR2 = 0.1405
R1 = 0.0755, wR2 = 0.1445
0.337 and −0.285 e·Å−3
R1 = 0.0542, wR2 = 0.1399
R1 = 0.0693, wR2 = 0.1494
0.372 and −0.302 e·Å−3
R1 = 0.0488, wR2 = 0.1240
R1 = 0.0621, wR2 = 0.1324
0.327 and −0.248 e·Å−3
R indices (all
data)
largest diff. peak
and hole
Table 2. Fluorescence Lifetimes in the Solid-State and in
Solution for 9AC, 2F-9AC, 4F-9AC, 10F-9AC, and 2,6DF-
9AC
solid-state fluorescence decays are biexponential in general,
with a short component on the order of 3 ns and a longer-lived
component on the order of 20−30 ns (Table 2). The exception
once again is 10F-9AC, whose long-time decay component is
significantly shorter at 10.7 ns. The nonexponential decays of
the 9AC derivatives indicates that the structural disorder
inferred from the XRD data leads to a distribution of relaxation
times and possibly reaction rates.
Given the structural and spectroscopic similarities of the five
compounds, we had little reason to expect dramatically different
photomechanical properties. We began by measuring the
lifetime of the metastable photodimer in different crystals. To
accomplish this, we utilized a fluorescence recovery experiment.
In this experiment, a brief pulse of 325 nm laser light reacts a
fraction of the monomer, typically about 50%. The laser is then
attenuated to a level where the photochemistry is negligible but
the fluorescence signal from the surviving monomer molecules
can still be detected. The dimer photoproduct has negligible
absorption at 325 nm and does not contribute to the
fluorescence signal. As the photodimer dissociates back into
a
solution τS
(ns)
τ1 (ns)
τ2 (ns)
A1
A2
9AC
6.0 0.4
3.1 0.2
1.8 0.1
3.2 0.1
3.8. 0.2
33.0 0.5
24.0 0.3
21.7 0.1
10.7 0.1
18.1 0.6
0.31
0.42
0.15
0.50
0.66
0.69
0.58
0.85
0.50
0.34
15.3 0.1
18.0 0.1
17.2 0.2
5.7 0.5
2F-9AC
4F-9AC
10F-9AC
2,6DF-9AC
15.6 0.2
a
The solid state fluorescence lifetime decay was fit to a biexponential
decay of the form A1 exp[t/τ1] + A2 exp[t/τ2]. The solution lifetime
decay was fit to a single exponential decay of the form A exp[t/τS].
substitution has a dramatic effect on the molecular electronic
state structure and dynamics. The solid-state fluorescence and
excitation spectra are shown in Figure 3. Again, the different
compounds show similar broadened excitation and emission
spectra. The solid-state fluorescence line shape is characteristic
of excimer formation, as expected for the π-stacked crystal. The
E
dx.doi.org/10.1021/cm502866e | Chem. Mater. XXXX, XXX, XXX−XXX

Chemistry of Materials
Article
model the kinetics, we simply define τ1/2 as the time required
for the fluorescence to attain 50% of its asymptotic value. Table
3 gives the average τ1/2 values for the different compounds
Table 3. Fluorescence Recovery Half Time (τ1/2) of 9AC and
Its Fluorinated Derivatives
τ1/2 (s)
9AC
200 20
2F-9AC
4F-9AC
10F-9AC
2,6DF-9AC
55
25
5
3
400 40
60
6
shown in Figure 1. 4F-9AC undergoes a fluorescence recovery
that is almost 10 times more rapid than that of 9AC, while 10F-
9AC is significantly slower than 9AC. 2F-9AC and 2,6DF-9AC
were considerably more rapid than 9AC but still a factor of 2
slower than 4F-9AC. The large variation in recovery times was
unexpected given the similarity of crystal structures and
spectroscopic properties of these compounds.
An important question is whether the mechanical recovery
mirrors the fluorescence recovery. We recently analyzed the
connection between reaction progress and mechanical
deformation for the solid-state [4 + 4] photodimerization of
9-methlyanthracene, showing that the mechanical response
mirrors the reaction progress but with a slight time lag.36 We
have also shown that 9AC nanorod bending closely follows the
photochemical reaction progress.19 Even though 4F-9AC is a
different molecule, we expect a similar dynamical picture to
hold for it as well, with the mechanical response time of 4F-
9AC being ∼10× faster than that of 9AC. In Figure 5 we
Figure 3. Solid-state excitation and emission spectra of anthracene-9-
carboxylic acid (9AC) and its fluorinated derivatives: (a) 9AC; (b) 2F-
9AC; (c) 4F-9AC; (d) 10F-9AC, and (e) 2,6DF-9AC. The solid line
plots are the excitation spectra, and the dotted line plots are the
emission spectra. For the emission spectra, samples were excited at
350 nm. Solution spectra are given in the Supporting Information
(Figure S1).
monomers, the fluorescence signal recovers to an asymptotic
value. Ideally, this value is identical to the fluorescence signal
before the light pulse, but in most cases it is slightly (10−20%)
below that value due to crystal movement or photobleaching.
Typical data traces are shown in Figure 4 for 4F-9AC, 2F-9AC,
Figure 4. Fluorescence recovery curves of 4F-9AC (black), 2F-9AC
(red), and 9AC (blue). Prior to the probing of the photorecovery, all
three crystalline films were photoreacted by ∼50%. For comparison,
the residual signal after photoreaction was subtracted from the data,
and the recovery curves were normalized so they recovered to an
asymptotic value of 1. In the experiment, 9AC and 2F-9AC crystals
recovered to 70−80% of their original signals, while 4F-9AC crystals
recovered to 94% of its original signal. The fluorescence recovery data
of 2,6DF-9AC and 10F-9AC crystals are given in the Supporting
Information, Figure S3.
Figure 5. Series of optical microscopy images of (a) 9AC and (b) 4F-
9AC. The 4F-9AC crystal completely untwists in 25 s, while the 9AC
crystal requires 420 s to unbend. Both scale bars are 50 μm.
compare a series of images for microribbons composed of 9AC
and 4F-9AC after irradiation causes them to twist. The 4F-9AC
ribbon completely untwists within 30 s, while the 9AC ribbon
requires ∼400 s. Thus, the mechanical recovery rate closely
tracks the fluorescence recovery rate, as expected. We find that,
for all the compounds, given comparable crystals, the
mechanical recovery times parallel the fluorescence recovery
times. Quantitative comparison of the mechanical recovery
times is difficult due to the fact that different crystals have
and 9AC. It is clear that the recovery rates are very different in
these three crystals. It is also clear that the recovery is
nonexponential, from the time-lag in the growth at early times.
This nonexponential behavior is a symptom of nucleation
kinetics, which has been observed in the solid-state
dimerization of 9AC.35 Rather than try to quantitatively
F
dx.doi.org/10.1021/cm502866e | Chem. Mater. XXXX, XXX, XXX−XXX

Chemistry of Materials
Article
different shapes and degrees of adhesion to the substrate,
leading to variations in the observed recovery times. Typically,
4F-9AC shows the most rapid mechanical recovery, with 2F-
9AC slower by a factor of ∼2, 9AC slower by a factor of ∼10,
and 10F-9AC slower by a factor of ∼25. We found it difficult to
obtain consistent results for the mechanical recovery of 2,6DF-
9AC, since it often did not recover its original shape, despite
the fact that the fluorescence recovered. In order to compare
absolute mechanical recovery times between different molecular
crystals, we would need uniform samples where all the crystals
are the same size and shape. In this paper only the qualitative
correlations as shown in Figure 5 are presented.
The more brittle nature of 9AC compared to 4F-9AC can be
attributed to the greater hardness of 9AC, and thus we expect
this crystal to crack more readily in response to the internal
strain generated by the photodimerization. This is consistent
with our observations, suggesting that mechanical measure-
ments of the reactant crystal may provide useful parameters for
evaluating photomechanical molecular crystals.
In addition to being less prone to fragmentation, 4F-9AC
crystals also show improved mechanical cycling behavior. It was
previously found that 9AC nanorods and microribbons could
be repeatedly bent or twisted, but only for a limited number of
cycles, typically less than 1018. In contrast, ribbons and needles
composed of 4F-9AC could undergo 100 or more photo-
mechanical cycles without noticeable loss of response. An
example is shown in Figure 7, where a 4F-9AC microneedle
bends under the repeated application of 405 nm light. The
crystal shows complete motion and full recovery even after 115
cycles. Additional images of repeated microribbon twisting can
be found in the Supporting Information (Figure S5). This
In addition to their faster recovery time, 4F-9AC crystals
were more resistant to cracking and fracture than 9AC crystals.
In Figure 6, two similarly sized, large (>20 μm) crystals of 9AC
Figure 6. Optical microscopy images of 9AC and 4F-9AC crystals. (a)
9AC crystal before and after irradiation. The 9AC crystal shatters into
pieces after irradiation. (b) 4F-9AC before and after irradiation. After
irradiation, the 4F-9AC crystal stays intact. Both scale bars are 50 μm.
and 4F-9AC are shown before and after irradiation at 405 nm
for 5 s. These crystals are too large to exhibit the bending and
twisting seen in smaller needles and ribbons.37 The 9AC crystal
loses many surface fragments during exposure and obviously
never recovers its original morphology, while the 4F-9AC
crystal stays intact and does not fragment. This appears to be a
general property of the 4F-9AC crystals. Even relatively large
crystals could survive irradiation conditions that would cause
similar 9AC crystals to disintegrate. This improved robustness
was previously observed for 10F-9AC crystals as well.24 In
order to investigate the physical origin of this effect, the
mechanical properties of the monomer crystals were measured
using a nanoindentation method.31,38 The measured reduced
elastic modulus and hardness for 9AC were determined to be
9.58 and 0.46 GPa, respectively, while the reduced elastic
modulus and hardness of 4F-9AC were 6.98 and 0.27 GPa,
respectively (Supporting Information, Figure S4). Post-
indentation optical analysis of 9AC revealed that the material
was fractured. The Er values are typical for organic molecular
crystals. Both the reduced elastic modulus and hardness of 4F-
9AC are less than 9AC, but the hardness decreases more than
the elastic modulus. The tendency of a material to develop
cracks is proportional to the ratio of hardness to elastic
modulus,39,40 which is 0.0486 for 9AC and 0.0392 for 4F-9AC.
Figure 7. Snapshots of repeated cycles of 4F-9AC crystalline needle’s
photoinduced bending and recovery: (a) 1st cycle; (b) 50th cycle; and
(c) 115th cycle. The irradiation source was a 100 W Hg lamp. Scale
bar is 20 μm.
G
dx.doi.org/10.1021/cm502866e | Chem. Mater. XXXX, XXX, XXX−XXX

Chemistry of Materials
Article
enhanced robustness puts this group of molecules in the same
class as the fluorinated diarylethenes.13,14 The intermolecular
anthracene [4 + 4] dimerization reaction, however, is much
more disruptive than the intramolecular diarylethene ring
opening−closing reaction in the sense that it involves larger
nuclear displacements by more atoms. Given this, the fact that
the 4F-9AC crystals remain intact and responsive after multiple
reaction cycles is remarkable. The fluorinated anthracene is
expected to be more stable with respect to photochemical
oxidation side reactions.41 What is perhaps more surprising is
that crystal defect formation and fracture do not seem to be
major degradation pathways in this crystal. For the latter effect,
the reduced hardness of the 4F-9AC crystal may be an
important factor that affects degradation.
The results described above demonstrate that the photo-
mechanical behavior of the 4F-9AC crystal represents a
significant improvement over that of 9AC, but the physical
origin of this improved behavior is not clear. Fluorine
substitution has been used as a strategy for changing crystal
packing to enhance photochemical reactivity.42,43 But as
discussed above, all five compounds have very similar crystal
packing geometries and photophysical properties. It should be
noted, however, that the crystals have varying amounts of
disorder and some variations in twinning behavior and impurity
levels. One concern is that the different mechanical behaviors
could be due to differences in crystal quality between the
different molecules. But defects and impurities would not be
expected to change the dimer dissociation rates, which result
from very localized molecular interactions. The fact that the
different mechanical response rates mirror the fluorescence
recovery rates suggests that the changes in the mechanical
properties are due to changes in the intrinsic crystal structure,
as opposed to differences in crystal quality.
If we assume that the changes in behavior are due to intrinsic
crystal properties, then the fluorine atoms must be the key.
Organic fluorine substitution can lead to multiple different
hydrogen and halogen bonding interactions which have the
potential to modify material properties.44 We suspect that the
faster dimer dissociation and increased plasticity of the 4F-9AC
crystals both stem from changing noncovalent interactions in
the crystal. There are several possible culprits. Fluorine
substitution can affect π−π interactions due to the electron-
withdrawing effects of the F atoms, leading to enhanced
photodimerization. However, the small change in the
interplanar distances (Table 4) and the similarity in crystal
packing suggests that this effect is not large. Noncovalent
bonding interactions, for example H···F interactions, can
stabilize crystal structures. If we assume that these bonds are
broken when the photodimer is formed, then the stabilization
energy of the H···F interactions would provide an extra driving
force for the dissociation reaction and lead to more rapid
recovery. The photodimer dissociation would be driven partly
by the energetic gain derived from reestablishing the hydrogen
bond network. Determining which molecule has the strongest
hydrogen bonding interactions from the crystal structures in
Figure 2 is not easy, especially given the disorder in the crystal
structures. In order to assess how intermolecular forces affect
the energetics of the monomer−dimer reaction, we are
collaborating with a theoretical group. Ideally, we would be
able to correlate the relative stability of the photodimers with
the observed recovery rate. Although the microscopic origins of
the different behaviors are currently unclear, the observation
that large changes in the photomechanical performance of the
crystals can be achieved by small changes in the constituent
molecular structure represents the most important conclusion
of this paper.
CONCLUSIONS
■
We have synthesized and characterized four fluorinated
derivatives of 9AC, a molecule that shows a reversible
photomechanical response in its crystal form. These new
molecules have similar molecular electronic structure and
dynamics, as well as similar crystal packing motifs, but different
photomechanical properties. One of these derivatives, 4F-9AC,
exhibits an order of magnitude improvement in both its
recovery time and its cycling ability relative to 9AC. It appears
that subtle interactions arising from changes in the π−π
interactions or hydrogen bonding between molecules are
responsible for the changes in the crystal photomechanical
properties. An improved understanding of these interactions
will require the help of computational chemistry to quantify
energetic differences between the different crystals. The results
in this paper demonstrate that it is possible to significantly
improve the properties of photomechanical materials based on
the anthracene [4 + 4] photodimerization reaction through
molecular engineering. It is likely that advances in both organic
synthesis and crystal engineering will lead to dynamic molecular
crystals with improved performance.
ASSOCIATED CONTENT
* Supporting Information
■
S
1H-NMR of 4F-9AC and synthesis of fluoro−anthracene
carboxylic acids; spectroscopic properties of fluoro−anthracene
carboxylic acids; photomechanical properties of 9AC and 4F-
9AC; and crystallographic information for 2F-9AC, 4F-9AC,
2,6DF-9AC, and 10F-9AC. This material is available free of
charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Authors
■
*E-mail: christopher.bardeen@ucr.edu (C.J.B).
*E-mail: kaysir@ksau-hs.edu.sa; rabihalkaysi@gmail.com
(R.O.K.).
Table 4. Distances between Two Adjacent π−π Stacked
Molecules of 9AC, 2F-9AC, 4F-9AC, 10F-9AC, and 2,6DF-
9AC
Notes
The authors declare no competing financial interest.
Dπ−π (Å)
a
9AC
3.473
3.438
3.446
3.436
3.412
ACKNOWLEDGMENTS
■
2F-9AC
4F-9AC
10F-9AC
2,6DF-9AC
This research was supported by the National Science
Foundation Grant DMR-1207063 (C.J.B.), by the Air Force
Office of Scientific Research Grant FA9550-12-1-0245 (D.K.),
and by the King Abdulaziz City for Science and Technology
(KACST) Grant AT-30-435 (R.O.K.).
b
a
b
Data taken from ref 45. Data taken from ref 24.
H
dx.doi.org/10.1021/cm502866e | Chem. Mater. XXXX, XXX, XXX−XXX

Chemistry of Materials
Article
(37) Kim, T.; Zhu, L.; Mueller, L. J.; Bardeen, C. J. CrystEngComm
2012, 14, 7792−7799.
REFERENCES
■
(1) Kim, T.; Zhu, L.; Al-Kaysi, R. O.; Bardeen, C. J. ChemPhysChem
2014, 15, 400−414.
(38) Varughese, S.; Kiran, M. S. R. N.; Ramamurty, U.; Desiraju, G.
R. Angew. Chem., Int. Ed. 2013, 52, 2701−2712.
(39) Lawn, B. R.; Evans, A. G.; Marshall, D. B. J. Am. Ceram. Soc.
1980, 63, 574−581.
(2) Finkelmann, H.; Nishikawa, E.; Pereira, G. G.; Warner, M. Phys.
Rev. Lett. 2001, 87, 015501/1−015501/4.
(3) Ikeda, T.; Mamiya, J.; Yu, Y. Angew. Chem., Int. Ed. 2007, 46,
506−528.
(40) Lawn, B. R.; Marshall, D. B. J. Am. Ceram. Soc. 1979, 62, 347−
350.
(4) Barrett, C. J.; Mamiya, J.; Yager, K. G.; Ikeda, T. Soft Matter 2007,
3, 1249−1261.
(41) Matsubara, Y.; Kimura, A.; Yamaguchi, Y.; Yoshida, Z. Org. Lett.
2008, 10, 5541−5544.
(5) Lange, C. W.; Foldeaki, M.; Nevodchikov, V. I.; Cherkasov, V. K.;
Abakumov, G. A.; Pierpont, C. G. J. Am. Chem. Soc. 1992, 114, 4220−
4222.
(42) Kumar, V. A.; Begum, N. S.; Venkatesan, K. J. Chem. Soc., Perkin
Trans. 2 1993, 463−467.
(43) Sonoda, Y. Molecules 2011, 16, 119−148.
(44) Berger, R.; Resnati, G.; Metrangolo, P.; Weber, E.; Hulliger, J.
Chem. Soc. Rev. 2011, 40, 3496−3508.
(6) Koshima, H.; Nakaya, H.; Uchimoto, H.; Ojima, N. Chem. Lett.
2012, 41, 107−109.
(7) Koshima, H.; Takechi, K.; Uchimoto, H.; Shiro, M.; Hashizume,
D. Chem. Commun. 2011, 47, 11423−11425.
(8) Naumov, P.; Sahoo, S. C.; Zakharov, B. Z.; Boldyreva, E. V.
Angew. Chem., Int. Ed. 2013, 52, 9990−9995.
(9) Bushuyev, O. S.; Tomberg, A.; Friscic, T.; Barrett, C. J. J. Am.
Chem. Soc. 2013, 135, 12556−12559.
(45) More, R.; Scholz, M.; Busse, G.; Busse, L.; Paulmann, C.;
Tolkiehn, M.; Techert, S. Phys. Chem. Chem. Phys. 2012, 14, 10187−
10195.
(10) Nath, N. N.; Pejov, L.; Nichols, S. M.; Hu, C.; Saleh, N.; Kahr,
B.; Naumov, P. J. Am. Chem. Soc. 2014, 136, 2757−2766.
(11) Irie, M.; Yokoyama, Y.; Seki, T. New Frontiers in Photochromism;
Springer: Japan, 2013.
(12) Roberts, R. J.; Rowe, R. C.; York, P. Powder Technol. 1991, 65,
139−146.
(13) Kobatake, S.; Takami, S.; Muto, H.; Ishikawa, T.; Irie, M. Nature
2007, 446, 778−781.
(14) Morimoto, M.; Irie, M. J. Am. Chem. Soc. 2010, 132, 14172−
14178.
(15) Terao, F.; Morimoto, M.; Irie, M. Angew. Chem., Int. Ed. 2012,
51, 901−904.
(16) Irie, M. Chem. Rev. 2000, 100, 1685−1716.
(17) Ito, Y.; Fujita, H. J. Org. Chem. 1996, 61, 5677−5680.
(18) Al-Kaysi, R. O.; Bardeen, C. J. Adv. Mater. 2007, 19, 1276−
1280.
(19) Good, J. T.; Burdett, J. J.; Bardeen, C. J. Small 2009, 5, 2902−
2909.
(20) Zhu, L.; Al-Kaysi, R. O.; Bardeen, C. J. J. Am. Chem. Soc. 2011,
133, 12569−12575.
(21) Takahashi, S.; Miura, H.; Kasai, H.; Okada, S.; Oikawa, H.;
Nakanishi, H. J. Am. Chem. Soc. 2002, 124, 10944−10945.
(22) Al-Kaysi, R. O.; Muller, A. M.; Bardeen, C. J. J. Am. Chem. Soc.
2006, 128, 15938−15939.
(23) Bucar, D. K.; MacGillivray, L. R. J. Am. Chem. Soc. 2007, 129,
32−33.
(24) Zhu, L.; Al-Kaysi, R. O.; Dillon, R. J.; Tham, F. S.; Bardeen, C. J.
Cryst. Growth Des. 2011, 11, 4975−4983.
(25) Kaur, I.; Jia, W.; Kopreski, R. P.; Selvarasah, S.; Dokmeci, M. R.;
Pramanik, C.; McGruer, N. E.; Miller, G. P. J. Am. Chem. Soc. 2008,
130, 16274−16286.
(26) Rochlin, E.; Rappaport, Z. J. Org. Chem. 2003, 68, 216−226.
(27) Yakamoto, G.; Suzuki, M.; Oki, M. Bull. Chem. Soc. Jpn. 1983,
56, 306−313.
(28) Bergmann, E. D.; Berkovic, S. J. Org. Chem. 1961, 26, 919−923.
(29) Dewar, M. J. S.; Michl, J. Tetrahedron 1970, 26, 375−384.
(30) Campione, M.; Ruggerone, R.; Tavazzi, S.; Moret, M. J. Mater.
Chem. 2005, 15, 2437−2443.
(31) Oliver, W. C.; Pharr, G. M. J. Mater. Res. 2004, 19, 3−20.
(32) Medina, B. M.; Beljonne, D.; Egelhaaf, H. J.; Gierschner, J. J.
Chem. Phys. 2007, 126, 111101/1−111101/6.
(33) Tannaci, J. F.; Noji, M.; McBee, J.; Tilley, T. D. J. Org. Chem.
2007, 72, 5567−5573.
(34) Dey, J.; Haynes, J. L.; Warner, I. M.; Chandra, A. K. J. Phys.
Chem. A 1997, 101, 2271−2278.
(35) More, R.; Busse, G.; Hallmann, J.; Paulmann, C.; Scholz, M.;
Techert, S. J. Phys. Chem. C 2010, 114, 4142−4148.
(36) Kim, T.; Zhu, L.; Mueller, L. J.; Bardeen, C. J. J. Am. Chem. Soc.
2014, 136, 6617−6625.
I
dx.doi.org/10.1021/cm502866e | Chem. Mater. XXXX, XXX, XXX−XXX