Photochromic molecular gyroscope with solid state rotational states determined by an azobenzene bridge

Article abstract of DOI:10.1021/jo402516n

We describe the synthesis, characterization, photochemical isomerization, and rotational dynamics of a crystalline molecular gyroscope containing an azobenzene bridge (trans-2) that spans from one end of the stator to other, with the intention of explorin

Full text of DOI:10.1021/jo402516n

Article
pubs.acs.org/joc
Photochromic Molecular Gyroscope with Solid State Rotational
States Determined by an Azobenzene Bridge
Patrick Commins and Miguel A. Garcia-Garibay*
Contribution from the Department of Chemistry and Biochemistry, University of California, Los Angeles, California 90095-1569,
United States
S
* Supporting Information
ABSTRACT: We describe the synthesis, characterization, photo-
chemical isomerization, and rotational dynamics of a crystalline
molecular gyroscope containing an azobenzene bridge (trans-2) that
spans from one end of the stator to other, with the intention of
exploring its function as a molecular brake. While single crystal X-ray
diffraction analysis of a photochemically inactive dichloromethane
solvate was used to confirm the molecular and packing structures of
trans-2, a nanocrystalline pseudopolymorph was shown to be
photoactive, and it was analyzed by powder X-ray diffraction
(PXRD), scanning electron microscopy, and variable temperature
2
solid state H NMR before and after photoisomerization. It was shown that the nanocrystalline suspension irradiated with λ =
340 nm reaches a photostationary state with 34% of cis-isomer, as compared to that observed in solution where the
2
corresponding value is 74%. Line shape analysis of solid state H NMR spectra of a phenylene-d4 isotopologue, obtained as a
function of temperature, indicated that rotation in crystals of the trans-2 isomer, with a mean activation energy of 4.6 0.6 kcal/
mol and a pre-exponential factor exp(29.4 1.7), is ten times faster than that of samples containing the cis-2 isomer, which has a
higher mean activation energy of 5.1 0.6 kcal/mol and a lower pre-exponential factor of exp(27.9 1.3).
one of the bridges to collapse toward the central phenylene,
ultimately hindering the desired rotation.
INTRODUCTION
■
Interest in artificial molecular machines has driven chemists to
explore the design and synthesis of molecules that can be
switched from one state to another1 with suitable external
stimuli in order to display functions that take advantage of
thermal oscillations. A few years ago we suggested that a
promising platform for the design of artificial molecular
machines might take the form of “amphidynamic crystals”2
built with rigid components that sustain their lattice
architecture and mobile components that explore fast, large
amplitude motions, designed to perform specific functions. To
realize a combination of fast motion and crystalline order, we
decided to emulate and analyze the structure of macroscopic
gyroscopes, which may form an ordered lattice while allowing
for the fast motion of their internal rotator (Figure 1). To
ensure rapid rotational dynamics, we selected a 1,4-substituted
phenylene as the rotator and two alkynyl linkages as a nearly
“barrierless” axle.3 One approach toward accessing molecular
frameworks with controllable molecular dynamics has been
utilizing shielded or “bridged” molecular gyroscopes.4 In our
initial work, the rotator was shielded from intermolecular close
contacts by a pair of bulky trityl groups acting as the stator. The
installation of groups that bridge the structure from one trityl
stator to the other in order to prevent interdigitation and
provide better shielding to the rotator was met with partial
success. While three benzophenone bridges decreased the
extent of molecular interdigitation,5 there was a tendency for
The tendency for a bridge to collapse toward the rotator
inspired us to explore the use a photoactivatable bridging group
that might serve as a simple molecular brake, which may help
modify the rotational dynamics of the molecule (Figure 1B).
Based on its similarity with the previously studied benzophe-
none bridges, and the previous work in the field using the
azobenzene chromophore,6 we selected a 4,4′-dimethyl-
azobenze as a photoresponsive molecular bridge. We report
here the synthesis, characterization, photochromic behavior,
and rotational dynamics of the singly azobenzene bridged
molecular gyroscope 2. Although we encountered challenges
associated with the desolvation and lack of reactivity of single
crystals grown from dichloromethane and diethyl ether, we
showed that nanocrystals obtained by precipitation of a THF
solution of trans-2 in water led to photoactive crystals, which
were photochemically investigated in the form of nanocrystal-
line suspensions. The polymorphic nature of the nanocrystals
was determined by PXRD, their habits analyzed by scanning
electron microcopy (SEM), and the remarkable thermal
stability of the photogenerated cis-azobenzene isomer estab-
2
lished by UV−vis absorption methods. Using solid state H
NMR of filtered nanocrystals, we were able to confirm that
rotational exchange dynamics in the original trans-2 structure
Received: November 13, 2013
Published: January 15, 2014
© 2014 American Chemical Society
1611
dx.doi.org/10.1021/jo402516n | J. Org. Chem. 2014, 79, 1611−1619

The Journal of Organic Chemistry
Article
Figure 1. (A) A molecular gyroscope with three benzophenone bridges. (B) An idealized model of an azobenzene bridged molecular gyroscope
where a change from trans- to cis-azobenzene configurations may change the rotational motion of the central phenylene.
Scheme 1
Figure 2. (A) X-ray-determined molecular structure of trans-2 as a dichloromethane solvate. (B) Packing of trans-2 in sheets made with pairs of
molecules having adjacent azobenzene bridges on one side and creating a cavity for two molecules of dichloromethane on the other.
has a lower energy barrier than that of the metastable cis-2
form.
The azobenzene bridged molecular gyroscope trans-2 was
prepared by slow addition of 5 and diphenol molecular rotor 6
into a flask with potassium carbonate in DMF to provide high
dilution conditions (1.0 mM based upon 6) to favor
macrocyclization. The bridged molecular gyroscope trans-2
was obtained in a 47% yield, or 69% yield per bond formed,
which is in good agreement with previous macrocyclization
reactions.5 A deuterated isotopologue was also synthesized to
study the rotational dynamics of trans-2 and cis-2 using
RESULTS AND DISCUSSION
■
Synthesis and Characterization. The synthesis of trans-2
was achieved using the convergent synthesis shown in Scheme
1. Samples of 1,2-bis(4-(bromomethyl)phenyl)diazene 5 were
obtained from p-toluidine 3 using CuBr(I) and dimethylami-
nopyridine (DMAP) in air to afford 4,4′-dimethylazobenzene
4.7 Compound 4 was then doubly brominated by refluxing N-
bromosuccimide (NBS) and benzoyl peroxide in carbon
tetrachloride for 6 h to afford 5 in 60% yield.8
2
quadrupolar spin−echo H NMR spectroscopy. The isotopo-
logue trans-2-d4 was synthesized in the same manner using 1,4-
dibromobenzene-d4 instead of 1,4-diiodobenzene in the
1612
dx.doi.org/10.1021/jo402516n | J. Org. Chem. 2014, 79, 1611−1619

The Journal of Organic Chemistry
Article
Figure 3. (Left) SEM image of nanocrystals of trans-2 formed by the reprecipitation method. The particles are well faceted and have an average size
of ca. 96 nm. (Right) PXRD spectrum of filtered nanocrystals of trans-2 (red line) and the experimental PXRD spectrum of desolvated crystals trans-
2 crystallized from dichloromethane and diethyl ether (blue line).
Sonogashira coupling.5 The resulting trans-2 (and trans-2-d4)
were purified by column chromatography and characterized by
13C NMR, 1H NMR, attenuated total reflectance (ATR) FTIR,
and high-resolution mass spectrometry. By 1H NMR,
compound trans-2 showed the expected AA′BB′ spin system
associated with the 1,4-disubstituted azobenzene bridge at 7.84
and 7.38 ppm, a 1,3-disubstituted aryl splitting pattern from the
trityl stator, and the only nonaromatic signal is a singlet located
at 5.20 ppm from the benzylic methylene protons of the
azobenzene bridge. A diagnostic singlet at 6.84 ppm was
attributed to the four phenylene protons of the motionally
averaged rotator. Upon irradiation with 340 nm UV light to
generate the cis-2 isomer, we noticed that the AA′BB′ spin
system of the azobenzene bridge showed an upfield shift to the
heavily populated aromatic region between 7.40 and 6.91 ppm.
The methylene singlet also experienced an upfield shift to 4.91
ppm, which was a useful diagnostic to determine the ratio
that is not accounted for is presumed to have evaporated before
or while the structure was acquired. There is an intermolecular
close contact between a proton in the phenylene rotator and
two of the azobenzene carbons of 2.6 Å and 2.7 Å. There is an
additional intermolecular close contact of 3.1 Å present
between one of the methylene protons and a carbon of a
neighboring trityl stator.
Nanocrystalline Suspensions. To study the photo-
isomerization of trans-2 in the solid state, we decided to
explore the formation of nanocrystalline suspensions in water.
It has been shown that suspended nanocrystals constitute an
ideal medium to analyze solid state photochemical reactions
with conventional transmission spectroscopy methods. When
their size is smaller than the wavelength of light, suspended
crystals may be compared with large proteins or protein
complexes, and many of the challenges associated with bulk
solids, which are related to their high optical density,
birefringence, and light scattering, are strongly diminished.11
Nanocrystalline suspensions of trans-2 were obtained from
THF and water by the reprecipitation method (also known as
solvent-shift method), which consists of the addition of a
concentrated solution of the desired compound in a good
water-miscible solvent into vigorously stirring water.12 The
nanocrystalline suspensions of trans-2 were analyzed by
dynamic light scattering (DLS) and SEM. Nanocrystalline
specimens were shown to have an average size of ca. 96 nm by
DLS (Figure S20), which was in good agreement with the size
of crystals observed in the SEM image shown in Figure 3. The
image also revealed that nanocrystals are well faceted and have
a rectangular habit. To determine the nature of nanocrystals of
trans-2 obtained from THF−water, we measured their powder
X-ray diffraction (PXRD). Their diffractogram is shown in
Figure 3 along with the one obtained with powdered crystals
grown from CH2Cl2. Not surprisingly, the two diffractograms
do not match and the one obtained from the filtered
nanocrystals reveals significantly broader peaks, consistent
with the very small size of the nanocrystalline specimens,13
and a very broad baseline indicating that at least part of the
filtrate is amorphous. It is important to note that both the
nanocrystalline suspension and the bulk powder of trans-2-d4
contain >97% of trans isomer, so the difference in the
diffraction data is not attributed to different contributions of
the cis isomer. While numerous crystallizations were attempted
to obtain diffraction quality single crystals of trans-2 by
evaporation of tetrahydrofuran and water, only aggregates
1
between the two isomers. The H NMR spectra for both the
trans-2-d4 and cis-2-d4 are nearly identical to the natural
abundance samples, except that the phenylene rotator singlets
are no longer present.
Crystallization Studies. Diffraction quality single crystals
of trans-2 were obtained by slow evaporation of a 1:1 solution
of dichloromethane and diethyl ether. The diffraction data was
acquired at 100 K and solved in the monoclinic system, space
group P21/c with four molecules of trans-2 and ca. 3.3
molecules of dichloromethane per unit cell (Figure 2A). The
position of the phenylene rotator is disordered over two sites
with occupancies of 73% and 27% that differ by an angle of 37°.
The molecular structure is characterized by a conformation
where the three C−Ar bonds on the two trityl groups adopt a
staggered orientation and adopt angles that result in the same
axial chirality. The two alkynes are bent such that the central
phenylene rotator is bowed toward the azobenzene nitrogens.
The two aryl rings that compose the azobenzene moiety are
twisted 39° away from planarity. The molecular gyroscopes
pack in layers in a manner that is similar to that observed for
several previously reported bridged structures.5,9 Each molecule
has one of its two trityl groups engaged in a sixfold phenyl
embrace.10 The other trityl group is slightly offset from its
neighboring trityl group, preventing a sixfold phenyl embrace,
but forming two edge-to-face interactions. There is one
molecule of dichloromethane disordered over two sites with
occupancies of 46% and 37% located in the space between the
pairs of molecular gyroscopes (Figure 2b). The 17% of solvent
1613
dx.doi.org/10.1021/jo402516n | J. Org. Chem. 2014, 79, 1611−1619

The Journal of Organic Chemistry
Article
could be obtained under those conditions. Crystallizations
attempted by layering a solution of trans-2 in THF on top of
water yielded either conglomerates or very fine needles.
between pairs of azobenzene moieties in the crystal, or it could
also be due to the crowded environment surrounding
azobenzene, which has several close contacts in its crystal
environment. To our surprise, when analogous experiments
were carried out with nanocrystalline suspensions, we
discovered that the trans−cis photoisomerization takes place
as it does in solution (Figure 5), but under otherwise similar
Photoisomerization. The photoisomerization of the
azobenzene bridge molecular gyroscope trans-2 was first
studied by UV−vis spectroscopy in a dilute benzene solution
and then as a nanocrystalline suspension in water. The UV−vis
of trans-2 shows the typical π→π* and n→π* transition of
azobenzene at ca. 340 and 440 nm, respectively.14 There is also
a very strong absorbance located at 290 nm (not shown)
attributed to the benzene chromophores in the two trityl
groups. Irradiation of trans-2 in solution at 340 nm for various
time increments resulted in the diagnostic decrease in the
absorbance of the 340 nm band, which was accompanied by a
slight increase of the absorption band at 440 nm, as expected
for the isomerization of trans-2 to the cis-2 (Figure 4 top
Figure 5. (Top) Photoisomerization of nanocrystalline trans-2 to cis-2
upon irradiation at 340 nm until it reaches a photostationary state, and
(Bottom) preferential photoisomerization of cis-2 to trans-2 by
irradiation at 440 nm.
conditions, it appeared to be significantly slower. A photosta-
tionary state was reached after exposing the sample to UV light
for approximately 165 min, and the reduction in the maximum
absorbance at 340 nm was only 46% as compared to the change
observed in solution. This change corresponds to a trans:cis
1
Figure 4. (Top) The photoisomerization of trans-2 to cis-2 in benzene
upon irradiation at 340 nm until it reaches a photostationary state, and
(Bottom) preferential photoisomerization of cis-2 to trans-2 by
irradiation at 440 nm and return to the trans-2 form.
photostationary state of 66:34, which was confirmed by H
NMR analysis (Figure S21). Conversion of cis-2 to trans-2 by
irradiation at 440 nm was shown as a relatively fast increase in
the absorption band at 340 nm, such that continued irradiation
returns the sample to the trans-2 isomer within 34−40 min
(Figure 5 bottom frame).
frame). Analysis of a sample that had reached its photosta-
tionary state as judged by change in its UV−vis spectrum was
Thermal Reversibility. The thermal equilibration of cis-2
back to trans-2 was analyzed by measuring the recovery of the
intense absorption of trans-2 at 334 nm. An aqueous
nanocrystalline suspension of trans-2 was irradiated at 365
nm until it reached its photostationary state after 120 min. The
suspension was monitored by UV−vis absorption at 334 nm
after intervals of 10 to 20 h for a total of 340 h at room
temperature, when ca. 60% of the original absorption value had
been recovered. At that time, the sample was exposed to broad
visible light and full recovery was observed after a few minutes,
indicating that it had the potential of going back to the original
state but was undergoing a very slow thermal isomerization
process. While the data is insufficient for a rigorous kinetic
analysis, it was clear that a fit of the thermal isomerization data
1
evaluated by H NMR, which established a trans:cis ratio of
26:74. Irradiation of a sample that had reach the photosta-
tionary state at λ = 440 nm, where cis-2 can be preferentially
excited, showed a rapid recovery of the absorption band at 340
nm such that the sample returned to its initial state after only
ca. 36 min (Figure 4, bottom frame).
In contrast to the changes observed in solution, exposure of
the dry dichloromethane-containing crystals of trans-2 to UV
light for up to 24 h revealed no changes in the UV−vis
spectrum. Since there are many examples of photoactive
azobenzene chromophores in the solid state,6,14 it seems
reasonable that the photostable nature of these crystals may be
due to self-quenching arising from tight face-to-face interaction
1614
dx.doi.org/10.1021/jo402516n | J. Org. Chem. 2014, 79, 1611−1619

The Journal of Organic Chemistry
Article
2
would require a model with at least two components recovering
with very different rates, with the fast component being
relatively minor (Figure S22). Assuming that full thermal
recovery would occur after very long times, the data could be fit
reasonably well by a short component of 19.9 h and a long
component of 330 h (13.75 days) with weighted contributions
of 0.7% and 99.3%, respectively (Figure S23). This result
indicated that the long lifetime of the metastable cis-2 form
should make it possible to study the rotational dynamics of the
central phenylene-d4 for both the stable (100% trans-2) and
metastable (34% cis-2) forms using the slow quadrupolar echo
2H NMR technique, which can take tens of hours. For
simplicity, for the remainder of this paper, we will refer to
metastable crystals containing ca. 34% cis-2 just as “cis-2”.
case. The H NMR spectrum is dominated by the interaction
between the nuclear spin and the electric quadrupole moment
at the nucleus. It has an orientation dependence that makes it
sensitive to dynamics in the range of ca. 103 to 108 s−1. Each
deuteron provides a doublet with a splitting frequency Δν that
depends on the orientation angle β that the C−2H bonds make
with respect to the external field:
Δv = 3/4(e2qzzQ /h)(3 cos2 β − 1)
= 3/4QCC(3 cos2 β − 1)
The magnitude of Δν for deuterons with an axially symmetric
electric field gradient tensor depends on the electric quadrupole
moment of the deuteron, Q, and on the magnitude of the
principal component of electric field gradient tensor, qzz, which
lies along the C−2H bond. Combined with the electric charge
and Plank constant, e and h, the first term in the equation
represents the quadupolar coupling constant QCC, for which
we have used a value 180 kHz for aromatic deuterons. Static
powder samples have all the β angles represented and consist of
a combination of doublets with populations and intensities that
reflect a statistical distribution with respect to the external field,
known as a Pake pattern (Figure 8a). The corresponding broad
symmetric spectrum characterized by two sharp maxima and
two feet separated by ca. 125 kHz and 250 kHz, respectively, as
shown in Figure 8a. By contrast, powder samples with
phenylene groups undergoing a fast site exchange described
as 180° rotations are subject to dynamic averaging and result in
a narrower symmetric spectrum with the two maxima separated
by distance of only ca. 32 kHz and two sets of shoulders with
separations of ca. 115 and 150 kHz (Figure 8d). Spectral
changes caused by 180° rotation in the intermediate exchange
regime denote an evolution between the two limits as
illustrated with the examples in Figure 8b and c. Importantly,
several excellent programs are available to simulate these
spectra with models that include sites with variable flip angles,
populations, and energy barriers.16 In addition, there are also
other groups modeling the theoretical dynamics of molecular
gyroscopes using computational chemistry.17
Figure 6. Thermal isomerization of nanocrystalline cis-2 into trans-2
determined by measuring the recovery of the UV−vis absorption the
as a function of time. The two red dots indicate the absorbance before
UV irradiation at λ = 365 nm.
The photochemical stability of the azobenzene bridge was
evaluated by multiple isomerization cycles both in benzene (not
shown) and as a nanocrystalline suspension in water (Figure 7).
Measurements of nanocrystalline samples of trans-2 and cis-2
were carried out at 46.07 MHz with a 90° pulse width of 2.5 μs,
echo and refocusing delays of 50 and 42 μs, respectively, using a
15 s delay for relaxation between pulses. Samples of
nanocrystals of trans-2 were prepared using the reprecipitation
method,12 centrifuged, decanted, and the resulting pellet was
dried in a desiccator at 298 K for 24 h. Samples of cis-2 were
prepared in the same manner, except that the nanocrystalline
suspension was exposed to light at λ = 365 nm for 4 h prior to
centrifugation.
Samples of trans-2 were measured between 170 and 296 K,
but only those acquired below 220 K are shown in the left
column in Figure 9. Spectra acquired above 220 K were
characteristic of rotational motion in the fast exchange regime
with some additional narrowing mechanism (please see below).
Similarly, nanocrystals of cis-2 were analyzed in the range of
200−260 K as illustrated on the right column of Figure 9. As in
the case of trans-2, spectra acquired at higher temperatures
displayed further narrowing that deviates from a twofold
flipping process and were analyzed further. One can see that the
spectra measured at any given temperature are different for the
two samples, indicating that the motion of pure trans-2 is
indeed affected by the partial isomerization into the cis-isomer.
This is evident in the spectra measured at 220 K where one can
Figure 7. Changes in absorbance of suspended nanocrystals of trans-2
alternatively irradiated with 365 nm and white light. A similar behavior
with a greater change on optical densities was observed in solution.
Azobenzene is a robust chromophore, capable of undergoing
hundreds of isomerization cycles without significant photo-
degradation.6 When exposed to 365 nm light for 15 min and
allowed to re-equilibrate under ambient light for 12 h,
molecular gyroscope 2 showed good cyclability over 12 cycles
both in solution and as a nanocrystalline suspension (Figure 7).
2
Variable Temperature H Spin−Echo NMR Analysis.
Deuterium NMR is a powerful technique for determining
rotational dynamics in the solid state.15 It relies on the use of
samples labeled with deuterium atoms in positions where
motion will be probed, the central phenylene rotator in our
1615
dx.doi.org/10.1021/jo402516n | J. Org. Chem. 2014, 79, 1611−1619

The Journal of Organic Chemistry
Article
Figure 8. Examples of simulated 2H NMR spectra corresponding to phenylene deuterons undergoing two-site 180° flips with exchange frequencies
of (a) kex ≤ 103 s−1, (b) kex = 105 s−1, (c) kex = 106 s−1, and (d) kex ≥ 108 s−1 with a quadrupolar coupling constant of 180 kHz and an asymmetry
parameter η = 0.
Figure 9. Spin−echo 2H NMR spectra for nanocrystals of the trans-2 (left) and cis-2 (right) isomers. The heavy blue line represents the experimental
data and the thin orange line indicates the spectra simulation, obtained as described in the text. The mean rotational rate used for each fitting is listed
next to the corresponding spectrum. A line broadening of 5 kHz was used during processing.
see that rotation in crystals of trans-2 is already in the fast
exchange regime (kex ≥ 108 s−1) while rotation in crystals
containing cis-2 remains at intermediate values, clearly showing
that solid state isomerization leads to more hindered rotation.
1616
dx.doi.org/10.1021/jo402516n | J. Org. Chem. 2014, 79, 1611−1619

The Journal of Organic Chemistry
Article
The spectra shown in Figure 9 are generally consistent with
the expected twofold flipping process, but there are features
indicating the coexistence of sites where rotation occurs with
faster and slower exchange rates. The co-occurrence of fast and
slow sites in the sample can be deduced from spectra that
contain peaks characteristic of the two dynamic ranges. For
example, the spectrum of trans-2 measured at 170 K has two
sets of maxima separated by ca. 125 and 32 kHz, which may be
approximated as a sum of spectra such as those shown in Figure
8a and d. Spectra measured at higher temperatures display an
increasing contribution from the faster components as seen by
the greater intensity of the two inner peaks. Recognizing the
solid has an amorphous component as indicated by PXRD, the
heterogeneous nature of the spectra is unsurprising, and we
explored a model that involves two discrete sites with different
dynamics and a model that describes a system with a
distribution of sites and activation energies leading to a
0−10°. However, the data obtained at 296 K is much narrower
on the top than on the bottom, and does not resemble any of
the simulated librations. We interpret the spectral changes in
Figure 11A as a function of increasing temperature as resulting
from a heterogeneous distribution of large librational
amplitudes. This interpretation is in agreement with the sample
heterogeneity displayed in the PXRD and by the need for a log-
Gaussian model to simulate the spectra obtained in the
intermediate regime.
The spectra of nanocrystals of cis-2 were acquired at 200−
296 K (Figure 9). The spectra were also simulated using a
weighted average of a log-Gaussian distribution of jump rates,
but a distribution width of σ = 2.0 was necessary to obtain a
good fit. Since the irradiated cis-2 nanocrystals are a 31:69%
solid solution of cis-2 and trans-2 isomers, a more
heterogeneous mixture of activation energies would be
expected and this assumption is reflected in the line-shape
analysis simulation, which required a σ = 2 value. The exchange
frequencies were simulated to be 3.2, 8.9, 24, and 66 MHz for
data acquired at 200, 220, 240, and 260 K, respectively. It is
important to note that the thermal cis to trans isomerization
during the acquisition of the spectra was considered to be small
because the lifetime of the thermal isomerization was estimated
to be almost fourteen days (330 h) at 296 K, and the
acquisition of each spectrum was approximately 12 h at
temperatures that are 30−96 K lower, where the lifetime should
2
distribution of exchange rates. The H NMR simulations were
performed using the program Express 1.0.16b with a
quadrupolar coupling constant of 180 kHz and an asymmetry
parameter of η = 0. Attempts to use a simple two-site model
generally gave unsatisfactory results. The most reasonable
simulations of the trans-2 spectra shown in Figure 9 required a
model based on log-Gaussian distribution of jump rates with a
width of σ = 1.5. The simulations gave a reasonable match to
the experimental spectra with exchange frequencies of 7.3, 40,
and 170 MHz for data acquired at 170, 200, and 220 K,
respectively. Arrhenius analysis revealed an average activation
energy Ea = 4.6 0.6 kcal/mol and a pre-exponential factor of
exp(29.42 1.71) s−1 (Figure 10, top).
2
be much longer. The H NMR calculated mean activation
energy for rotation of cis-2 is 5.1 0.6 kcal/mol, which is 0.5
kcal/mol higher than the barrier for the trans isomer and the
pre-exponential factor is exp(27.86 1.32) s−1. Noticing that
there is an overlap between the margin of error and calculated
difference between the two isomers, further statistical anaylysis
was performed and the data was determined to have a z-score
of 0.98, indicating that there is a confidence level of 84% that
the data is significant. In addition to the relatively small
differences in activation barriers, an almost 10-fold difference in
exchange frequencies for trans-2 and cis-2 in the range of 200−
220 K can be traced to differences in the two mean pre-
exponential factors, which differ by a factor of 4.6. One of the
factors previously shown to lower the pre-exponential factor of
a molecular rotor is the amorphicity of the sample,19 which in
this case is consistent with the high heterogeneity displayed by
the cis-2 isomer, which required a kinetic model with a wider
barrier distribution width. Unfortunately, a single crystal X-ray
structure of the phase corresponding to the nanocrystalline
solid could not be obtained, so the structural details that may
be related to different rotational rates cannot be determined.
Figure 10. Arrhenius plots derived from the variable temperature
2
spin−echo H NMR data and simulations of trans-2 (blue diamonds)
and cis-2 (red squares). The activation energy and pre-exponential
factors for trans-2 are 4.6 0.6 kcal/mol and exp(29.42 1.71) s−1,
and those for cis-2 are 5.1 0.6 kcal/mol, and exp(27.86 1.32) s−1,
respectively.
CONCLUSION
■
While good quality dichloromethane-containing single crystals
of trans-2 could be used to confirm its molecular and packing
structure by single crystal X-ray diffraction determination, they
were shown to be photostable. By contrast, nanocrystalline
specimens obtained by solvent shifting from saturated THF to
water resulted in a photochemically active solvent-free
pseudopolymorph. Nanocrystalline samples of trans-2 displayed
promising photochemical reversibility, and the thermal stability
of the photogenerated cis-2 isomer was shown to be
considerable. The rotational dynamics measurements by
The spectra of trans-2 acquired above 230 K had an
appearance that indicate an exchange process beyond the fast
exchange regime for a 180° rotation with noticeable narrowing
as the temperature increased up to 296 K (Figure 11A). We
considered that the observed narrowing may be explained by
increasingly larger oscillations of the phenylene rotator by an
angle Δ°, in addition to the discrete 180° flips. The effects of
increasing angular displacements (Δ°) on the appearance of the
spectrum were explored with simulations of twofold flipping
motions in the fast exchange regime (>108 Hz) with added
oscillations between Δ = 0° and Δ = 60°, as shown in Figure
11B.18 One can see that the experimental spectrum acquired at
230 K resembles simulations with oscillations in the range of
2
spin−echo H NMR with filtered nanocrystals showed that
spectra obtained as a function of temperature could only be
simulated with a model that considers a Gaussian distribution
of activation energies centered at 4.6
0.6 kcal/mol with a
1617
dx.doi.org/10.1021/jo402516n | J. Org. Chem. 2014, 79, 1611−1619

The Journal of Organic Chemistry
Article
Figure 11. (A) Quadupolar spin echo 2H NMR of trans-2 measured between 230 and 296 K showing an increasing narrowing of the spectrum. (B)
Simulations of a model based on a twofold flipping in the fast exchange regime with increasing varying degrees of librational motion with angular
displacement ranging from 0° to 60°.
carbonate (82 mg, 0.6 mmol) and dry DMF (50 mL) and put under
Ar: dissolved 5 (55 mg, 0.15 mmol and 6-d4 (96 mg, 0.15 mmol) in
dry DMF (4 mL) in separate round bottoms under Ar; added 5, and 6-
d4 to the reaction at a rate of 0.5 mL per hour for 8 h. The reaction was
stirred at 40 °C for 24 h. The organic phase was diluted with DCM,
and washed with saturated aqueous LiCl (3 × 50 mL), water (50 mL),
and brine (50 mL). The combined organic fractions were dried over
MgSO4 and concentrated on a rotary evaporator. The residue was then
purified by flash chromatography on silica gel (1:1 hexanes:DCM) to
afford 54 mg (42%) of yellow powder. Mp = 295−296 °C. HRMS
ESI-TOF(+) calcd for [C62H41 D4N2O2]+: 853.3732; Found: 853.3729
(M+H).
width σ = 1.5, while cis-2 has a higher activation energy
centered at 5.1 0.6 kcal/mol with a broader distribution
indicated by width of σ = 2.0. While the exact underlying
mechanism responsible for the different rotational states cannot
be ascertained without detailed structural knowledge, this study
shows that carefully designed functionalization of molecular
gyroscopes may lead to molecules with rotational dynamics that
are amenable to control by external stimuli, a feature of high
promise in the field of crystalline molecular machines.
EXPERIMENTAL SECTION
■
trans-2-d4: FTIR (solid, HATR, cm−1). 3059, 2925, 1728, 1595, 1488,
All commercially available compounds were used without further
purification. Samples of 4,4′-dimethylazobenzene 4 and 1,2-bis(4-
(bromomethyl)phenyl)diazene 5 were prepared as described in the
literature.7,8
1
1447, 1260, 1230, 1033, 906, 729, 696. H NMR trans (400 MHz,
CDCl3): δ 7.84 (d, J = 8.2 Hz, 4H), 7.38 (d, J = 8.2 Hz, 4H), 7.35−
7.27 (m, 20H), 7.11 (apparent triplet, J = 8.2, 7.8 Hz, 2H), 6.99 (ddd, J
= 8.2, 2.5, 0.9 Hz, 2H), 6.84 (s, 4H), 6.82 (dd, J = 2.5, 1.8 Hz, 2H),
6.27 (ddd, J = 7.8, 1.8, 0.9 Hz 2H) 5.20 (s, 4H). 13C NMR trans (125
MHz, CDCl3): 158.1, 151.7, 146.8, 145.0, 140.8, 131.7 (t, JCD = 25
Hz), 129.1, 128.7, 128.0, 127.2, 126.8, 123.1, 122.7, 121.7, 117.0,
115.4, 96.3, 84.9, 69.9, 56.1.
Azobenzene Bridged Molecular Gyroscope, 2. A 100 mL
round-bottom flask equipped with a magnetic stir bar was charged with
potassium carbonate (82 mg, 0.6 mmol) and dry DMF (50 mL) and
put under Ar: dissolved 5 (55 mg, 0.15 mmol) and 6-d4 (96 mg, 0.15
mmol) in dry DMF (4 mL) in separate round bottoms under Ar;
added 5, and 6-d4 to the reaction at a rate of 0.5 mL per hour for 8 h.
The reaction was stirred at 40 °C for 24 h. The organic phase was
partitioned with DCM, and washed with saturated aqueous LiCl (3 ×
50 mL), water (50 mL) and brine (50 mL). The combined organic
fractions were dried over MgSO4 and concentrated on a rotary
evaporator. The residue was then purified by flash chromatography on
silica gel (1:1 hexanes:DCM) to afford 60 mg (47%) of yellow powder.
Mp = 295−296 °C. HRMS ESI-TOF(+) calcd. for [C62H45N2O2]+:
849.3481; Found: 849.3492 (M+H).
cis-2-d4: FTIR (solid, HATR, cm−1). 3055, 2925, 1596, 1488,
1
1447,1378, 1248, 1034, 908, 795, 757, 729, 696. H NMR cis (500
MHz, CDCl3): δ 7.64 (dd, J = 2.5, 1.9 Hz, 2H), 7.36 (d, J = 8.5 4H),
7.3−7.20 (m, 20H), 7.17 (apparent triplet, J = 8.0, 8.0 Hz, 2H), 6.90
(ddd J = 8.0, 1.9, 0.7 Hz, 2H), 6.89 (d, J = 8.5 Hz, 4H), 6.41 (ddd, J =
8.0, 2.5, 0.7 Hz, 2H), 4.91 (s, 4H). 13C NMR cis (125 MHz, CDCl3):
158.8, 153.0, 146.9, 144.9, 135.9, 131.1 (t, J = 25, 1C) 129.1, 128.7,
128.6, 128.0, 127.0, 126.98, 123.0, 121.9, 120.8, 115.6, 113.1, 97.4,
84.9, 69.9, 56.3.
trans-2: FTIR (solid, HATR, cm−1). 3060, 2923, 2852, 2247, 1714,
1,4-Bis(3-(3-hydroxyphenyl)-3,3-diphenylpropynyl)benzene-d4, 6-
d4. An oven-dried 250 mL round-bottom flask equipped with a
magnetic stir bar was charged with 7 (2.21 g, 3.27 mmol) and dry
DCM (30 mL) under an argon atmosphere. The flask was cooled to 0
°C in an ice bath, and 1 M BBr3 in DCM (9.8 mL, 9.8 mmol) was
added. The reaction was stirred for 3 h at 0 °C, quenched by addition
of ice, and diluted with hexanes. The precipitate is filtered and washed
with 0 °C DCM. Recovered 1.61 g (76%) of yellow powder.
Compound was used in next step without further purification. FTIR
(solid, HATR, cm−1): 3531, 3356 (broad), 3058, 1591, 1488, 1446,
1
1595, 1488, 1259, 1230, 1033, 906, 729, 696. H NMR trans (300
MHz, CDCl3): δ 7.84 (d, J = 8.7 Hz, 4H), 7.38 (d, J = 8.2 Hz, 4H),
7.35−7.27 (m, 20H), 7.11 (apparent triplet, J = 8.2, 7.8 Hz, 2H), 6.99
(ddd, J = 8.2, 2.5, 0.9 Hz, 2H), 6.84 (s, 4H), 6.82 (dd, J = 2.5, 1.8 Hz,
2H), 6.25 (ddd, J = 7.8, 1.8, 0.9 Hz 2H) 5.20 (s, 4H). 13C NMR (100
MHz, CDCl3): 158.2, 151.8, 145.8, 145.0, 140.8, 131.2, 129.1, 128.7,
128.0, 127.2, 126.7, 123.1, 122.7, 121.7, 117.5, 115.4, 96.3, 85.0, 69.9,
56.2.
cis-2: FTIR (solid, HATR, cm−1). 3060, 2923, 2853, 1720, 1596,
1
1488, 1447, 1238, 1033, 907, 791, 729, 696. H NMR (300 MHz,
1
1262, 1183, 1031, 832, 781, 754, 697. H NMR (500 MHz, DMSO-
CDCl3): δ 7.64 (dd, J = 2.5, 1.9 Hz, 2H), 7.43 (s, 4H), 7.3−7.20 (m,
24H), 7.17 (apparent triplet, J = 8.0, 8.0 Hz, 2H), 6.90 (ddd J = 8.0,
1.9, 0.7 Hz, 2H) 6.89 (d, J = 8.5 Hz, 4H), 6.41 (ddd, J = 8.0, 2.5, 0.7
Hz, 2H), 4.91 (s, 4H). 13C NMR (125 MHz, CDCl3): 158.8, 153.0,
146.9, 145.0, 144.9, 135.9, 129.1, 128.6, 128.1, 127.0, 126.98, 126.92,
126.9, 121.9, 120.8, 115.6, 113.1, 97.4, 84.9, 69.9, 56.3.
d6): δ 7.25−7.20 (m, 20H), 7.12 (apparent triplet, J = 8.2, 7.8 Hz, 2H),
6.89 (dd, J = 2.5, 1.8 Hz 2H), 6.78 (ddd, J = 7.8, 1.8, 0.9 Hz 2H), 6.60
(ddd, J = 8.2, 2.5, 0.9 Hz, 2H) 4.66 (s, 6H). 13C NMR (125 MHz,
DMSO-d6): δ 157.6 146.3, 145.0, 131.7 (t, JCD = 25 Hz) 129.6, 129.0,
128.7, 127.5, 122.8, 119.9, 116.3, 114.5, 97.7, 84.8, 56.2, 56.0. HRMS
ESI-TOF(-) Calcd for [C49H30D4O4]− 691.2786; Found: 691.2783 (M
+HCOO).
Azobenzene Bridged Molecular Gyroscope-d4, trans-2-d4. A 100
mL equipped with a magnetic stir bar was charged with potassium
1618
dx.doi.org/10.1021/jo402516n | J. Org. Chem. 2014, 79, 1611−1619

The Journal of Organic Chemistry
Article
1,4-Bis(3-(3-methoxyphenyl)-3,3-diphenylprop-1-yn-1-yl)-
benzene-d4, 7-d4. A 250-mL round-bottom flask equipped with a
magnetic stir bar was charged with 3-(3-methoxyphenyl)-3,3-
diphenylpropyne (1.1 g, 3.69 mmol), 1,4-dibromobenzene-d4 (451
mg, 1.85 mmol), diisopropylamine (5 mL), dry THF (10 mL), CuI
(35 mg, 0.18 mmol) and Pd(PPh3)2Cl2 (130 mg, 0.18 mmol) added,
and the reaction was heated to reflux for 48 h. The reaction was
quenched with aqueous saturated ammonium chloride and partitioned
with diethyl ether. The organic phase was washed with water (2 × 40
mL) and brine (40 mL). The combined organic fractions were dried
over MgSO4, filtered, and concentrated on a rotary evaporator. The
residue was purified by flash chromatography on silica gel (2:1
hexanes:DCM) to afford 909 mg (73%) of a yellow solid: mp = 204−
205 °C. FTIR (solid, HATR, cm−1): 3056, 2930, 2831, 1601,1602,
1580, 1489, 1446, 1430, 1315, 1290, 1243, 1132, 1051, 881, 780, 756.
1H NMR (400 MHz, CDCl3): δ 7.36−7.22 (m, 22H), 6.95 (apparent
Garibay, M. A. Top. Curr. Chem. 2006, 262, 179. Vogelsberg, C. S.;
Garcia-Garibay, M. A. Chem. Soc. Rev. 2012, 41, 1892.
(3) Saebo, S.; Almolof, J.; Boggs, J. E.; Stark, J. G. J. Mol. Struct.
(Theochem) 1989, 200, 361.
(4) (a) Han, J.; Deng, C.; Fang, R.; Zhao, D. Y.; Wang, L. Y.;
Gladysz, J. A. Organometallics 2010, 29, 3231. (b) Hess, G. D.;
Harnpel, F.; Gladysz, J. A. Organometallics 2007, 26, 5129. (c) Setaka,
W.; Yamaguchi, K. J. Am. Chem. Soc. 2013, 135, 14560. (d) Setaka, W.;
Yamaguchi, K. J. Am. Chem. Soc. 2012, 134, 12458. (e) Zeits, P. D.;
Rachiero, G. P.; Harnpel, F.; Reibenspies, J. H.; Gladysz, J. A.
Organometallics 2012, 31, 2854.
(5) Commins, P.; Nunez, J. E.; Garcia-Garibay, M. A. J. Org. Chem.
̃
2011, 76, 8355.
(6) (a) Kobatake, S.; Takami, S.; Muto, H.; Ishikawa, T.; Irie, M.
Nature 2007, 446, 778. (b) Koshima, H.; Ojima, N.; Uchimoto, H. J.
Am. Chem. Soc. 2009, 131, 6890. (c) Natanshohn, A.; Rochan, P.
Chem. Rev. 2002, 102, 4139.
(7) Zhang, C.; Jiao, N. Angew. Chem. 2010, 49, 6174.
(8) Ahmad, R. K.; Faure, D.; Goddard, P.; Oda, R.; Bassani, D. M.
Org. Biomol. Chem. 2009, 7, 3173.
triplet, J = 8.2, 7.8 Hz, 2H), 6.90 (dd J = 2.6, 1.5 Hz 2H), 6.86 (ddd, J
= 7.8, 1.8, 0.9, 2H), 6.82 (ddd, J = 8.2, 2.5, 0.9, 2H), 3.75 (s, 6H). 13
C
NMR (125 MHz, CDCl3): δ 159.3, 146.8, 145.1, 131.1 (t, JCD = 25
Hz) 129.1, 128.9, 128.0, 126.9, 121.8, 115.5, 112.0, 97.3, 84.8, 56.2,
55.1. HRMS ESI-TOF(+) calcd for [C50H35D4O2]+ 675.3201; Found:
675.3212 (M+H).
(9) Nunez, J. E.; Nataranjan, A.; Khan, S. I.; Garcia-Garibay, M. A.
̃
Org. Lett. 2007, 9, 3559.
(10) (a) Dance, I. Mol. Cryst. Liq. Cryst. 2005, 440, 265. (b) Stopin,
A.; Garcia-Garibay, M. A. Crys. Growth. Des. 2012, 12, 3792.
(11) (a) Lebedeva, N. V.; Tarasov, V. F.; M. J. E. Resendiz, M. J. E.;
Garcia-Garibay, M. A.; White, R. C.; Forbes, M. D. E. J. Am. Chem. Soc.
2010, 132, 82. (b) Chin, K. K.; Natarajan, A.; Gard, M. N.; Campos, L.
M.; Johansson, E.; Shepherd, H.; Garcia-Garibay, M. A. Chem.
Commun. 2007, 41, 4266. (c) Veerman, M.; Resendiz, M. J. E.; Garcia-
Garibay, M. A. Org. Lett. 2006, 8, 2615.
(12) Kasai, H.; Nalwa, H. S.; Oikawa, H.; Okada, S.; Matsuda, H.;
Minami, N.; Kuakuta, A.; Ono, K.; Mukoh, A.; Nakanishi, H. Jpn. J.
Appl. Phys. 1992, 31, 1132.
(13) West, R. W. Solid State Chemistry and its Applications; John
Wiley & Sons Ltd: Chichester, United Kingdom, 2003.
(14) (a) Kumar, G. S.; Neckers, D. C. Chem. Rev. 1989, 89, 1915.
(b) Bandara, H. M.; Burdette, S. C. Chem. Soc. Rev. 2012, 41, 1809.
(15) Hoatson, G. L.; Vold, R. L. NMR: Basic Princ. Prog. 1994, 32, 1.
(b) Fyfe, C. A. Solid State NMR for Chemists; CFC Press: Guelph,
Ontario, 1983.
(16) The simulation program NMR-WEBLAB from the Spiess lab is
available on the web: Macho, V.; Brombacher, L.; Spiess, H. W. Appl.
Magn. Reson. 2001, 20, 405. (b) The program Express 1.0 from the
Vold lab is also available on the web: Vold, R. L.; Hoatson, G. L. J.
Magn. Reson. 2009, 198, 57.
(17) (a) Marahatta, A. B.; Kanno, M.; Hoki, K.; Setaka, W.; Irle, S.;
Kono, H. J. Phys. Chem. C 2012, 116, 24845. (b) Akimov, A. V.;
Mandal, D.; Chernyak, V. Y.; Sinitsyn, N. A. J. Chem. Phys. 2013, 138,
024109. (c) Akimov, A. V.; Kolomeisky, A. B. J. Phys. Chem. C 2011,
115, 13584.
ASSOCIATED CONTENT
■
S
* Supporting Information
Spectroscopic data (1H, 13C NMR, FTIR) for compounds
trans-2, trans-2-d4, 7, and 8. Crystallographic information file
(cif) for compound trans-2. Differential scanning calorimetry
and thermogravimetric analysis for nanocrystals of trans-2,
dynamic light scattering of nanocrystalline suspensions of trans-
2, and decay kinetics for thermal isomerization of cis-2. This
material is available free of charge via the Internet at http://
pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
*E-mail: mgg@chem.ucla.edu.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported by NSF IGERT: Materials Creation
Training Program (MCTP) − DEG-0654431 and by grant
DMR1101934. This material is based upon work supported by
the National Science Foundation under equipment grant no.
CHE-1048804.
(18) Simulations were performed using NMR-WEBLAB from the
Spiess lab.
(19) O’Brien, Z. J.; Karlen, S. D.; Khan, S.; Garcia-Garibay, M. A. J.
Org. Chem. 2010, 75, 2482.
REFERENCES
■
(1) (a) Koumura, N.; Zijlstra, R.; Delden, R.; Harada, N.; Feringa, B.
Nature 1999, 401, 152. (b) Balzani, V.; Credi, A.; Raymo, F.; Stoddart,
F. Angew. Chem., Int. Ed. 2000, 39, 3348. (c) Kelly, T. R.; De Silva, H.;
Silva, R. A. Nature 1999, 401, 150. (d) Karim, A. R.; Linden, A.;
Baldridge, K. K.; Siegel, J. S. Chemical Science 2010, 1, 102.
(e) Muraoka1, T.; Kinbara1, K.; Aida, T. Nature 2006, 440, 512.
(f) Coskun, A.; Banaszak, M.; Astumian, D. R.; Stoddart, J. F.;
Grzybowski, B. A. Chem. Soc. Rev. 2012, 41, 19. (g) Browne, W. R.;
Feringa, B. L. Nat. Nanotechnol. 2006, 1, 25. (h) Michl, J.; Sykes, E. C.
H. ACS Nano 2009, 3, 1042. (i) Horie, M.; Suzaki, Y.; Hashizume, D.;
Abe, T.; Wu, T. D.; Sassa, T.; Hosokai, T.; Osakada, K. J. Am. Chem.
Soc. 2012, 134, 17932. (j) Lemouchi, C.; Iliopoulos, K.; Zorina, L.;
Simonov, S.; Wzietek, P.; Cauchy, T.; Rodríguez-Fortea, A.; Canadell,
E.; Kaleta, J.; Michl, J.; Gindre, D.; Chrysos, M.; Batail, P. J. Am. Chem.
Soc. 2013, 135, 9366.
(2) (a) Khuong, T.-A. V.; Nunez, J. E.; Godinez, C. E.; Garcia-
̃
Garibay, M. A. Acc. Chem. Res. 2006, 39, 413. (b) Karlen, S. D.; Garcia-
1619
dx.doi.org/10.1021/jo402516n | J. Org. Chem. 2014, 79, 1611−1619