Quantum amplified isomerization: A new concept for polymeric optical materials

Article abstract of DOI:10.1021/ma050348k

The preparation and evaluation of a new class of photoresponsive polymers are described on the basis of a process called quantum amplified isomerization (QAI). The QAI process utilizes photoinitiated, cation radical isomerization chemistry in a polymeric medium. Two classes of materials are described: one where the QAI reactant is molecularly doped in the polymer matrix and another where the reactant is part of a functionalized polymer. Quantum yield experiments demonstrate that the isomerization reaction can proceed by a chain process with modest efficiencies. Photochemical conversion experiments show that high extents of conversion of the QAI reactants are possible. The rate and extent of conversion are strongly correlated to the glass transition temperature of the polymer. For molecularly doped polymers, hypotheses to explain the high conversions based on diffusion or phase separation of the reactants were tested and excluded. Models are discussed to rationalize experimental factors that affect the quantum yields and the photochemical conversions.

Full text of DOI:10.1021/ma050348k

7684
Macromolecules 2005, 38, 7684-7694
Quantum Amplified Isomerization: A New Concept for Polymeric
Optical Materials1
⊥
,†
‡
†
†
†
J. G. Gillmore, J. D. Neiser, K. A. McManus, Y. Roh, G. W. Dombrowski,
,
‡
,†
,§
,§
T. G. Brown,* J. P. Dinnocenzo,* S. Farid,* and D. R. Robello*
Department of Chemistry, Center for Photoinduced Charge Transfer, University of Rochester,
Rochester, New York 14627-0216; The Institute of Optics, University of Rochester, Rochester, New York
4627-0186; and Research Laboratories, Eastman Kodak Company, Rochester, New York 14650-2116
1
Received February 17, 2005; Revised Manuscript Received July 1, 2005
ABSTRACT: The preparation and evaluation of a new class of photoresponsive polymers are described
on the basis of a process called quantum amplified isomerization (QAI). The QAI process utilizes
photoinitiated, cation radical isomerization chemistry in a polymeric medium. Two classes of materials
are described: one where the QAI reactant is molecularly doped in the polymer matrix and another where
the reactant is part of a functionalized polymer. Quantum yield experiments demonstrate that the
isomerization reaction can proceed by a chain process with modest efficiencies. Photochemical conversion
experiments show that high extents of conversion of the QAI reactants are possible. The rate and extent
of conversion are strongly correlated to the glass transition temperature of the polymer. For molecularly
doped polymers, hypotheses to explain the high conversions based on diffusion or phase separation of
the reactants were tested and excluded. Models are discussed to rationalize experimental factors that
affect the quantum yields and the photochemical conversions.
Introduction
Scheme 1
Photoinduced electron transfer (PET) reactions in
rigid matrices form the basis of several technologically
important processes. Of these, the most widely applied
are silver halide photography and electrophotography,
where electron transfer or charge transport occurs at
an interface or in a heterogeneous medium.2 PET
processes between molecularly dissolved reactants in
polymeric media are also well-known. Examples of the
latter include the PET generation of free radicals or
+•
cation radical (R ) isomerizes to a product cation radical
+•
(
P ) that is capable of undergoing electron-transfer
3
acids that can be used to initiate polymerizations or
+•
reaction with another R to form R . The latter process
will be energetically favorable if the oxidation potential
of P is greater than that of R. It is important to note
that rapid diffusion is necessary for both initiation and
propagation of the chain reaction.
induce other reactions, such as those used in photoli-
4
thography. In most of these processes amplification of
the primary photochemical reaction is an important
component. In silver halide photography and acid-
induced chemical transformations, the photochemical
process leads to the formation of a latent image, an
entity that can be used in a subsequent treatment
We set out to determine the feasibility of photochemi-
cally initiated cation radical chain isomerization reac-
tions in the restrictive environment of rigid polymer
matrices. Scheme 2 illustrates the concept, which we
(
development) as a catalyst to amplify the effect of the
primary step. In photoinitiated polymerization the
conversion of many monomeric units into a polymeric
molecule is also a form of amplification.
6
refer to as quantum amplified isomerization (QAI). It
is important to point out that, unlike ion radical chain
reactions in polar solvents where diffusion and solvation
of ion radicals occur readily, the high viscosity and the
relatively low solvating power of polymers seemed likely
to limit the extent of chain propagation in polymeric
media since, as described above, diffusional processes
are required in fluid media with these types of reactions
for chain initiation and propagation. In addition, since
Herein we describe a new class of photoresponsive
materials based on photoinduced electron-transfer chain
isomerizations in polymeric matrices. Photoinduced
electron-transfer reactions in fluid media have long been
demonstrated to have the potential for amplification by
5
a chain mechanism. These reactions usually take place
-
•
+•
in polar solvents such as acetonitrile, where free (i.e.,
fully separated) ion radicals can be formed. As shown
in Scheme 1, chain reactions can occur when a reactant
diffusional separation of S R ion radical pairs would
also be severely impaired in polymer matrices, it seemed
probable that chain-terminating return electron transfer
to regenerate S and R might be a dominant process.
Despite these potential difficulties, we recognized that
if it was possible to achieve ion radical chain isomer-
izations in polymersseven with modest efficienciessthe
resulting materials could have novel photoresponsive
properties.
†
Department of Chemistry, University of Rochester.
‡
§
⊥
The Institute of Optics, University of Rochester.
Eastman Kodak Company.
Current address: Department of Chemistry, Hope College,
Holland, MI 49424.
*
T.G.B.); jpd@chem.rochester.edu (J.P.D.); samir.farid@kodak.com
S.F.); douglas.robello@kodak.com (D.R.R.).
Corresponding authors. E-mail: brown@optics.rochester.edu
When assessing the potential for achieving ion radical
chain isomerizations in a polymeric matrix, it is instruc-
(
(
1
0.1021/ma050348k CCC: $30.25 © 2005 American Chemical Society
Published on Web 08/16/2005

Macromolecules, Vol. 38, No. 18, 2005
Quantum Amplified Isomerization 7685
Scheme 2
Scheme 3
Table 1. Quantum Yields for Isomerization of 1 and 3 in
Polymeric Media
sensitizera
φisom
b
reactant
[reactant]
1
1
1
1
1
3
3
3
0.1
0.5
0.5
0.5
2.0
0.1
0.5
0.5
DCA
1.6
2.2
3.2
4.2
5.3
0.4
0.6
1.2
DCAt
2-Bu -DCA
t
2,6-diBu -DCA
t
2,6-diBu -DCA
DCA
DCA
NMQ
tive to consider the possible fates of geminate ion radical
pairs produced by PET in low-polarity solvents where
out-of-cage diffusion is not a dominant reaction. As
shown in Scheme 3, there are several processes to
consider: (1) adiabatic isomerization of the reactant
cation radical to a product cation radical, (2) nonradia-
tive return electron transfer, (3) radiative return elec-
tron transfer (i.e., exciplex fluorescence), (4) intersystem
crossing to form the triplet excited state of the sensitizer
a
t
0
.02 M sensitizer. DCA ) 9,10-dicyanoanthracene; 2-Bu -DCA
t
)
2-tert-butyl-9,10-dicyanoanthracene; 2,6-diBu -DCA ) 2,6-di(tert-
butyl)-9,10-dicyanoanthracene; NMQ ) N-methylquinolinium
b
hexafluorophosphate. Isomerization quantum yields were re-
corded at low conversion and were corrected for incomplete
interception of sensitizer excited states.
Because compound 1 is somewhat thermally labile,8
we considered the possibility that its isomerization may
have been triggered by heat dissipation during the
photoreaction. To test this hypothesis, PMMA films of
were prepared containing ferrocene in place of DCA.
The excited state of ferrocene is incapable of oxidizing
, and upon photoexcitation, ferrocene is known to
(
or the reactant), and (5) bond-coupled return electron
transfer, a process that yields the neutral isomerization
product. Of the reactions shown in Scheme 3, only
adiabatic isomerization (reaction 1) has the potential
1
-
•
+•
for chain propagation, where interception of S /P by
1
-
•
+•
a reactant molecule leads to S /R ; all other reactions
are chain-termination steps.
undergo rapid nonradiative decay to its ground state
with high efficiency. Therefore, photolysis of PMMA
9
films containing ferrocene and 1 should convert all of
the absorbed light energy into heat, thus providing an
appropriate control experiment to determine whether
sample heating could be the cause of isomerization of
the 1 with DCA and the other electron-transfer sensitiz-
ers. In practice, 405 nm irradiation of films containing
0.5 M 1 and 0.6 M ferrocene resulted in no detectable
isomerization, even after long irradiation times. This
experiment rules out photochemical heating as an
explanation for the photoinduced isomerization of 1 in
PMMA. Nonetheless, the relatively low thermal stability
of compound 1 prompted us to investigate other cation
radical chain isomerizations in polymers.
Results
Molecularly Doped Polymers. Adiabatic isomer-
ization within the primary ion radical pair (reaction 1
in Scheme 3) is reported to be the dominant process for
reaction of cage compound 1 with the singlet excited
state of 9,10-dicyanoanthracene (DCA) in nonpolar
solvents. Thus, 1/DCA seemed to be an ideal reactant/
sensitizer pair to initially test the potential of initiating
a cation radical chain isomerization in a polymer matrix.
7
Photoinduced electron-transfer isomerization of hex-
amethyl Dewar benzene (3) in polar solvents is known
to lead to hexamethylbenzene (4) via a cation radical
chain mechanism.5a It is also known that isomerization
within the primary ion radical pair can occur in low-
polarity solvents.10 In nonpolar solvents, however, chain-
Poly(methyl methacrylate) (PMMA) films (≈20 µm)
containing dissolved 1 and 0.01 M DCA were prepared
on poly(ethylene terephthalate) sheets. The films were
irradiated with uniformly intense light from a Hg arc
lamp at 405 nm, where DCA selectively absorbs. Fol-
lowing irradiation, known areas of the irradiated films
were dissolved in CH2Cl2 containing an internal stan-
dard. After precipitation of the PMMA, the filtered
solutions were analyzed by HPLC to quantify the
number of molecules of 2 that were formed. Experiments
conducted with films containing 0.1 and 0.5 M 1 resulted
in isomerization quantum yields, corrected for incom-
plete interception of the excited sensitizer, of 1.6 and
+
•
+•
propagating isomerization of 3 to 4 principally
competes with bond-coupled return electron transfer
(reaction 5 in Scheme 3), which is a chain-terminating
reaction.1
0,11
Although the latter reaction would lower
the chain length for isomerization of 3 to 4, the higher
thermal stability of 3 and the ease of synthesizing
functionalized derivatives offered major advantages over
1 to probe photoresponsive polymers based on QAI.
2
.2, respectively. As shown in Table 1, upon further
increases in reactant concentration and changes in the
sensitizer structure, corrected quantum yields in the
range of 4-5 were obtained for the isomerization
reaction, consistent with a chain isomerization process.
In practice, the quantum yield for isomerization of
3/DCA in PMMA was lower than that for the isomer-

7686 Gillmore et al.
Macromolecules, Vol. 38, No. 18, 2005
ization of 1 by a factor of ca. 4 (see Table 1). As found
for 1, the quantum yield for isomerization of 3 increased
slightly with increasing reactant concentration. The
corrected isomerization quantum yield for 3 was also
found to increase when DCA was replaced by the
cationic sensitizer N-methylquinolinium hexafluoro-
phosphate (NMQ). These results will be discussed later.
Importantly, PMMA films containing 1 or 3 and DCA
showed substantial, although incomplete, conversion
upon irradiation, ≈75% and ≈45%, respectively. If the
reactants are assumed to be uniformly distributed in
the polymeric media, it is possible to estimate the charge
-•
+•
+•
separation distances between DCA and 1 or 3 that
would be required to account for the maximal conver-
sions. On the basis of the DCA concentration of 0.01 M,
the average separation distance between sensitizer
molecules is ≈70 Å. To account for ≈75% conversion of
Figure 1. Photochemical conversion experiments with 0.5 M
1
(circles) or 3 (squares) in PMMA containing 0.01 M DCA
1
, a charge separation distance of ≈32 Å would be
using 407 nm laser. Interpolated lines are drawn through the
experimental data points. The dashed line shows predicted
conversion for 3 based on the diffusion model (see text); this
line neglects the small extent of prompt conversion (<2%) due
to static quenching of the sensitizer.
required. The extent of conversion of 3 (≈45%) requires
a charge separation distance of ≈27 Å. Such separation
distances are noteworthy given the low dielectric of the
polymer medium.
We considered two alternative hypotheses that could
explain the relatively high conversions of 1 and 3: (i)
diffusion and (ii) phase separation. Diffusion of the
reactants during irradiation would obviate the need for
long-distance charge separation. Similarly, phase sepa-
ration of the reactants would provide high local concen-
trations of 1 or 3 in the polymer, which would reduce
the extent of charge separation required for high
conversion.
assumes that electron transfer from 3 to DCA* can occur
at distances that are slightly larger than van der Waals
contact. The calculated value for kdiff based on the above
-
3
-1 -1
estimates is ≈1.4 × 10
M
s .
ab ab
kdiff ) 4πr D NA
(7)
The magnitudes of the rate constant for diffusional
separation of the S/R pair (k-diff) and for the DCA-
photosensitized isomerization of DCA/3 in PMMA (krxn)
were needed to determine which step in the kinetic
scheme would be rate limiting. The value of k-diff can
be determined from kdiff and the effective equilibrium
constant for formation of the encounter pair DCA/3 that
leads to quenching of DCA* (K ) kdiff /k-diff). The
equilibrium constant can be directly measured from
fluorescence quenching. Steady-state fluorescence ex-
periments show that ≈2/3 of the DCA excited states are
Estimates of the diffusion coefficients of the reactants
in PMMA were needed to test the diffusion hypothesis.
These were obtained from a forced Rayleigh scattering
(
FRS) experiment performed by holographic exposure.12
A 20 µm thick coating on glass containing 0.5 M 3 and
.010 M DCA in PMMA was irradiated by using two
0
equal-intensity 407 nm laser beams that were crossed
at an angle of 35.55° normal to the film plane. The
resulting periodic spatial modulation of 3 and its
isomerization product, 4, gave rise to a diffraction
grating with a period of 350 nm. The strength of the
grating was measured by using 633 nm light from a
HeNe laser incident at the angle appropriate to satisfy
the Bragg condition. Long-term monitoring of the grat-
ing showed a gradual drop in the diffraction efficiency
of 10% over 20 days. If all of this decrease in diffraction
efficiency is attributed to diffusion, the resulting diffu-
-1
quenched by 0.5 M 3 in PMMA. Thus, K ≈ 4 M , which
-
4
-1
when combined with kdiff gives k-diff ≈ 3.5 × 10
s .
The value of krxn can be determined from the photon
flux, the number of molecules of DCA in the exposed
region of polymer film, and the fraction of light ab-
sorbed. Combining these experimental quantities gives
krxn ) 1.6 photons absorbed per DCA molecule per
second. Because k-diff , krxn, essentially every excited
DCA/3 pair undergoes photoreaction before diffusional
separation can occur. This requires that diffusional
encounter of 3 and DCA be rate limiting for the
hypothetical kinetic scheme shown in eq 6. Since DCA
is not consumed in the photoisomerization reaction, the
concentration of DCA remains constant. Therefore, the
rate constant for diffusion (and isomerization) should
-
19
2
sion coefficient (D) is calculated to be ≈9 × 10
cm /
13
s. On the basis of the similar molecular volumes of 3
252 Å ) and DCA (269 Å ), it is reasonable to assume
3
3 14
(
that they will have similar diffusion coefficients. These
diffusion coefficients can be used to test the diffusion
model in eq 6, where S is the sensitizer (DCA in this
case), R is the reactant (1 or 3), and P is the isomer-
ization product (2 or 4).
obey simple first-order kinetics, where the observed rate
kdiff
hν
-5 -1
constant is equal to kdiff [DCA], or 1.4 × 10
s . This
S + R yk-diffz S/R
8 S/P
(6)
krxn
rate constant was used to predict the conversion vs
irradiation time for photosensitized isomerization of 3
with DCA for a diffusional process (see dashed line in
Figure 1). The data clearly show that the diffusion
model is kinetically incompetent to explain the DCA-
photosensitized conversion of 3 to 4 in PMMA.
The rate constants for diffusion of 3 and DCA in
PMMA (kdiff) were estimated by using the Smoluchowski
equation, eq 7.15 The sum of the diffusion coefficients
(Dab) for 3 and DCA was approximated as twice the
diffusion coefficient measured for 3 in PMMA. The
encounter radius (rab) was estimated to be 10 Å, which
is ca. 2 Å greater than the sum of the average molecular
radii of 3 (3.9 Å) and DCA (4.0 Å). This estimate for rab
It was not possible to determine the diffusion coef-
ficient for 1 in PMMA by FRS because of the slow
thermal isomerization of 1 to 2 over extended time
periods. However, on the basis of the larger molecular

Macromolecules, Vol. 38, No. 18, 2005
Quantum Amplified Isomerization 7687
3
3
volume of 1 (416 Å ) vs 3 (252 Å ), it is expected that
samples prepared with and without added 3 were found
to have similar yields, molar mass distributions, and
glass transition temperatures. Moreover, 3 was recov-
ered quantitatively after the polymerization. These
results indicated that 3 did not interfere with the free
radical polymerization and that similar polymerization
of 5 should be feasible.
The synthesis of Dewar benzene-substituted monomer
5 is shown below. Ethyl ester 6 was prepared by the
method previously reported for the analogous methyl
ester.17 Reduction of 6 with diisobutylaluminum hydride
(DIBAL-H) provided alcohol 7, which was subsequently
esterified with methacryloyl chloride to give 5.
1
6
D(1) , D(3). 1 undergoes isomerization even more
rapidly than 3; therefore, diffusion is clearly kinetically
incompetent to explain the observed DCA-photosensi-
tized conversion of 1 to 2 in PMMA. Finally, in addition
to the quantitative arguments against diffusion for 1
and 3, we note that the diffusion model predicts that
the extent of conversion to product should exhibit an
exponential dependence on irradiation time that leads
to complete isomerization of the reactants, which is not
observed experimentally.
Having excluded diffusion as a hypothesis to explain
the high rates and extents of conversions for 1 and 3 in
polymeric media, we next sought to test whether phase
separation could rationalize the conversion data. To test
for phase separation, PMMA films containing 1 or 3
were analyzed by transmission electron microscopy
(
TEM). TEM images of PMMA samples containing no
additive or containing 1 at concentrations of 0.5 and of
.0 M were indistinguishable, even at 50 000-fold mag-
2
nification. Similarly, no evidence for phase separation
was found for PMMA films containing 3 at 0.5 M. Only
at high concentrations of 3 (1.5-2.0 M) in PMMA was
phase separation evident by films that were visually
cloudy.
Although the TEM experiments unambiguously rule
out phase separation on length scales of >20 nm, they
cannot exclude phase separation on significantly smaller
length scales. Therefore, to more critically test the phase
separation hypothesis, we prepared polymers in which
the QAI reactants were covalently attached to the
polymer backbone. The goal was to determine whether
functionalized polymerssin which both phase separa-
tion and diffusion are prohibitedsbehaved similarly to
the molecularly doped systems described above with
respect to the QAI process.
Functionalized Polymers. Functionalized polymers
were prepared with a Dewar benzene moiety rather
than the cage compound (1) because of the relative ease
of synthesis of Dewar benzene derivatives as well as
their thermal stability. Methacrylic monomer 5 (DBMA)
was chosen as a target monomer for conventional free
radical polymerization.
Monomer 5 was successfully copolymerized with
varying relative amounts of methyl methacrylate, n-
hexyl methacrylate, or cyclohexyl methacrylate. Copoly-
mers containing high concentrations of the Dewar
benzene-derived repeat units (>50 mol %) tended to
yellow and became insoluble upon storage at room
temperature over a period of days to weeks, which is
due to an, as yet, unidentified cross-linking reaction.
However, copolymers more dilute in Dewar benzene
moieties were found to be stable for long periods of time.
We concentrated our studies on copolymers containing
approximately 10 mol % of Dewar benzene repeat units.
These samples remained colorless and soluble through-
out the course of our experiments. In addition, the
concentration of Dewar benzene moieties in the 10%
copolymers is somewhat similar to those employed in
experiments in which 3 was molecularly doped into host
polymers. We presume that the distributions of repeat
units in the copolymers were random because reactivity
ratios of nearly all simple, alkyl methacrylates are close
to unity.18
Photoinduced isomerizations were first studied with
copolymers prepared from Dewar benzene monomer 5
and methyl methacrylate, p(DBMA/MMA), and from 5
and n-hexyl methacrylate, p(DBMA/HMA). Polymer
films containing sensitizer were coated on poly(ethylene
terephthalate) sheets and irradiated at 405 nm, as
described above. After dissolution of the polymer films
in CDCl3, the extent of isomerization of the Dewar
benzene to benzene moieties was assessed by integra-
1
tion of the ester R-methylene H NMR resonances of the
reactant vs product moieties. The ratio of the R-meth-
ylene integrals vs the R-ester proton resonances from
the comonomer was nearly constant (within 10%) as a
function of photochemical conversion, consistent with
a clean isomerization reaction for the functionalized
polymers. Conversion experiments with p(DBMA/MMA)
and p(DBMA/HMA) containing DCA showed significant
differences (see Figure 2). Interestingly, the limiting
conversion for p(DBMA/MMA) is less than one-half that
observed for 3 dissolved in PMMA. These results are
discussed below.
One obvious difference between p(DBMA/MMA) and
p(DBMA/HMA) is their glass transition temperatures,
115° vs 4°, respectively. To test whether varying Tg
could be the cause of the contrasting conversion results
in Figure 2, the photochemical conversion of p(DBMA/
Before embarking on the synthesis of 5, the feasibility
of free radical polymerization in the presence of hex-
amethyl Dewar benzene (3) was performed to test
whether possible chain transfer reactions involving the
allylic hydrogen abstraction from the Dewar benzene
moiety might interfere with the polymerization. A bulk
polymerization of methyl methacrylate (MMA) using the
conventional free radical initiator azobis(isobutyroni-
trile) (AIBN), and in the absence and presence of 3 (17
mol %), was carried out. The isolated polymers for the

7688 Gillmore et al.
Macromolecules, Vol. 38, No. 18, 2005
HMA), ≈90%, but similar to that found for p(DBMA/
MMA) (see Figure 2). These data again point to Tg being
an important factor controlling the extent of conversion.
Overall, the combined conversion experiments show
that molecularly doped and functionalized QAI polymer
systems behave similarly with regard to extent of
conversion, so long as systems of similar Tg are com-
pared. We next sought to compare molecularly doped
vs functionalized polymer systems in terms of their
relative quantum yields for isomerization.
Unlike the conversion experiments, quantum yield
experiments with molecularly doped polymers showed
that Tg did not substantially affect the reaction quantum
yields, presumably because the quantum yields are
measured at low conversion of the reactants. For
example, the isomerization quantum yields for 1 in
PMMA and poly(n-butyl methacrylate) were indistin-
guishable within experimental error. Isomerization
quantum yields for the functionalized polymers were,
nevertheless, measured with plasticized samples to
provide a closer comparison to the molecularly doped
systems. Quantum yields for p(DBMA/MMA) and 0.5
M 3 in PMMA were compared with both DCA and NMQ
as photosensitizers. For DCA, the quantum yields were
Figure 2. Photochemical conversion experiments with p(DB-
MA/HMA) (circles), p(DBMA/cHMA) (squares), and p(DBMA/
MMA) (triangles) containing ≈0.01 M DCA using a Hg arc
lamp. Interpolated lines are drawn through the experimental
data points.
0
.3 and 0.6 for functionalized and molecularly doped
QAI polymers, respectively. With NMQ, the respective
quantum yields were 1.0 and 1.2. Overall, the quantum
yields for functionalized and molecularly doped QAI
polymers are quite similar. The slightly lower quantum
yields observed for the functionalized polymer may be
attributable to the necessity of carrying the reactions
to higher conversions to obtain acceptable signal-to-
1
noise for the analytical method employed ( H NMR vs
GC). The quantum yield differences are rather small,
however, and suggest that the molecularly doped QAI
polymers and the functionalized polymers behave simi-
larly.
Finally, although diffusion in functionalized polymers
was expected to be slow, a forced Rayleigh scattering
experiment was conducted to test this prediction. As
described above, a weak diffraction grating was written
with a 350 nm period in p(DBMA/MMA) containing
DCA and 20 wt % plasticizer. Long-term monitoring of
the grating showed negligible loss (<0.03%) in its
diffraction efficiency over 23 days, demonstrating neg-
ligible diffusion of the polymer chains in the function-
alized polymer, even when heavily plasticized.
Figure 3. Photochemical conversion of p(DBMA/MMA)/DCA
with various weight percentages of plasticizer (di-n-butyl
phthalate). Interpolated lines are drawn through the experi-
mental data points.
MMA) was studied with 0, 5, and 10 wt % of an added
plasticizer, dibutyl phthalate (DBP). The Tgs of these
film compositions were estimated to be 115, 89, and 65
C, respectively.1 As shown in Figure 3, increasing
9
°
amounts of DBP in films of p(DBMA/MMA) lead to
increasing conversion. These results with plasticized
polymers, as well as the results with unplasticized,
functionalized polymers, are consistent with Tg playing
an important role in controlling the limiting extent of
conversion. Interestingly, the estimated Tg of the film
containing 5 wt % DBP (89 °C) is reasonably similar to
that determined for 0.5 M 3 in PMMA (95 °C), as are
their extents of conversion, ≈40% and ≈45%, respec-
tively. These data suggest that the higher limiting
conversion found for 3 dissolved in PMMA vs the
unplasticized, functionalized polymer p(DBMA/MMA)
is due to the plasticizing effect of 3 in the former system,
which decreases the Tg of the molecularly doped polymer
and thereby increases the ultimate conversion.
Discussion
The experiments described above provide the first test
of the QAI concept for polymeric materials based on
electron-transfer-initiated, ion radical chain isomeriza-
tion. The results show that isomerization quantum
yields can exceed unity, although the chain amplifica-
tion is significantly lower than that observed in polar,
fluid solution. Even when the quantum yield is <1, as
for some cases with the isomerization of 3 to 4, chain
amplification is nonetheless evidenced by the fact that
the isomerization quantum yield, corrected for incom-
plete interception of the excited sensitizer, increases
with increasing reactant concentration. That the quan-
tum yields for the QAI reactions are lower in polymeric
media than in polar solvents is not surprising. In polar,
fluid solvents, diffusive separation of the sensitizer
anion radical and the substrate cation radical in the
geminate ion radical pair can readily occur. Ion radical
separation naturally leads to a diminished rate for
Conversions were also compared for p(DBMA/HMA)
and the copolymer of 5 with cyclohexyl methacrylate,
p(DBMA/cHMA). Although these two copolymers are
expected to have reasonably similar polarities, they have
significantly different Tgs (4 vs 104 °C). The maximal
extent of conversion for p(DBMA/cHMA) was found to
be ≈40%, much lower than that observed for p(DBMA/

Macromolecules, Vol. 38, No. 18, 2005
Quantum Amplified Isomerization 7689
chain-terminating return electron transfer and conse-
quently greater isomerization quantum yields. Simi-
larly, chain-propagating electron transfer from the
neutral QAI reactant to the product cation radical is
enhanced in fluid solution by a diffusional process. As
demonstrated by the forced Rayleigh scattering experi-
ments, neither of these chain-propagating processes can
occur rapidly in a rigid polymer medium due to the low
rate constants for diffusion. In addition, the polymer
matrix may inhibit the chain isomerization process in
more subtle ways. For example, the rigid polymer
medium may sterically constrain isomerization of the
reactant cation radicals, thereby allowing chain-termi-
nating return electron transfer to more effectively
compete. In addition, the rigid nature of the polymer
makes it intrinsically nonpolar. This makes chain
propagation more difficult in a polymer because, in the
absence of molecular diffusion, propagation requires
charge separation of the sensitizer anion radical and the
chain-propagating cation radicals. Given the above
obstacles to ion radical chain isomerizations in polymer
hosts, it is remarkable that the reactions proceed at all
beyond the contact ion radical pair stage.
reaction spheres contain a reactant molecule within
their enclosed quenching spheres, the λS term must be
multiplied by PQ(0), the probability that a quenching
sphere contains no quencher. PQ(0) is equal to exp(-
20
λQ). Making this substitution into eq 9, and noting that
λS ) λR - λQ, gives eq 10.
φ
) {λ - λ P (0)}/f
q
(9)
isom
R
S
Q
φ
) {λ - (λ - λ ) exp(-λ )}/f
q
(10)
isom
R
R
Q
Q
The reaction sphere for DCA/1 upon initial excitation
can be calculated using eq 10 and the experimental
values for φisom, λQ, and fq. The calculated values for λR
at 0.5 and 0.1 M 1 are 1.87 and 1.89, respectively. These
values can be used to determine the average chain
propagation volumes and radii via eq 8 by substituting
λR for λQ. Within the framework of this reaction sphere
model, the average chain propagation radii are ≈20 and
≈
11 Å at 0.1 and 0.5 M 1, respectively. These results
are consistent with the qualitative conclusion made
above, namely that a larger fraction of chain propaga-
tion occurs outside the quenching sphere at lower
reactant concentration.
Quantum Yields. The rigid polymer medium clearly
restricts diffusion of QAI reactants and products in the
matrix; therefore, it is of interest to consider how the
isomerization proceeds under these constraints. We
begin with the reaction of bishomocubane 1 with DCA
in polymer. Because diffusion is negligible in the
polymer medium, fluorescence quenching of the singlet
excited state of DCA by 1 must occur by static quench-
ing. Static quenching is commonly interpreted according
A possible explanation for the greater chain propaga-
tion distance at lower reactant concentration can be
inferred from the reaction quantum yield dependence
on the concentration of 1. The fact that the quantum
yields do not increase in direct proportion to [1] suggests
that the ion radical chain initiation efficiency decreases,
or termination increases, or both, as the reactant
concentration increases. The effect of concentration on
the efficiencies of chain initiation/termination can be
plausibly related to the separation distance between the
geminate ion radical pairs formed upon initial photo-
induced electron transfer from 1 to the excited state of
DCA. At high concentration of 1, a greater fraction of
sensitizers will have a reactant molecule in contact with
the sensitizer, which will lead to the formation of
“contact” or “tight” ion radical pairs. In contrast, at low
concentration of 1, a greater fraction of “separated” or
“loose”, geminate ion radical pairs will be formed. We
propose that the changes in the relative proportion of
contact vs separated, geminate ion radical pairs might
result in differential efficiencies of chain initiation and
termination. The analysis begins by first understanding
that the chain propagation step involving electron
2
0
to the Perrin model, eq 8, where φ0 and φ are the
integrated fluorescence intensities in the absence and
presence of quencher. The average number of quenchers
in the quenching sphere (λQ) is equal to VNA[Q], where
V is the volume that contains effective quenchers, NA
is Avogadro’s number, and [Q] is the quencher concen-
tration. For 0.5 M 1 in polymer, φ0/φ was 2.0, from which
3
21
V was determined to be ≈2240 Å . This corresponds
to a quenching sphere with an average radius of ≈8 Å.
The average numbers of molecules within the 8 Å
quenching sphere (λQ) are 0.14 and 0.69 for 0.1 and 0.5
M 1, respectively.
ln(φ /φ) ) λ ) VN [Q]
(8)
0
Q
A
The quantum yields measured at 0.1 and 0.5 M 1 are
-
•
+•
1
.6 and 2.2, respectively. On the basis of photochemical
transfer from 1 to DCA /2 will, in general, require
some charge separation for both contact and separated
geminate ion radical pairs. Because of an inverse
dependence on ion radical pair separation distance, the
Coulombic barrier for charge separation will be larger
for contact geminate ion radicals pairs than for sepa-
rated ones. Consequently, chain-terminating return
electron transfer for contact, geminate ion radical pairs
will more effectively compete with chain propagation
than for separated, geminate pairs. This would explain
why φisom does not increase in direct proportion to [1].
The preceding analysis leads to several important
conclusions. First, it suggests that increasing the reac-
tant concentration to increase φisom will yield diminish-
ing returns because higher substrate concentrations will
necessarily increase the fraction of contact ion radical
pairs, which are less efficient at chain initiation. In
addition, the model suggests that a more effective
strategy for increasing φisom would be to use sterically
encumbered sensitizers that prevent “contact” geminate
measurements in nonpolar fluid media, the geminate
-
•
+•
ion pair DCA /1 undergoes adiabatic isomerization
-
•
+•
to DCA /2 with essentially unit efficiency. Assuming
this phenomenon holds in the polymeric matrix, the
amplification due to chain propagation would be 0.6 and
1
.2, respectively. Thus, as the concentration of 1 in-
creases by a factor of 5, chain propagation increases only
by a factor of 2. This suggests that the amplification is
proportionally more effective at lower concentration,
which is consistent with a larger propagation distance
at low concentration.
The average distance of chain propagation required
by the isomerization quantum yields (φisom) can be
estimated by using eq 9, where λR is the average, initial
number of reactant molecules in the reaction sphere,
λS is the average number of molecules in the shell that
lies outside the quenching sphere but inside the reaction
sphere, and fq is the fraction of excited-state DCA
molecules that are quenched by 1. Because not all

7690 Gillmore et al.
Macromolecules, Vol. 38, No. 18, 2005
ion radical pairs from being formed. This is a potentially
fruitful direction for future research on electron-transfer-
based QAI materials.
between the sensitizer anion radical and the reactant/
product cation radicals. Consistent with this expecta-
tion, the conversion studies for both molecularly doped
and functionalized polymers show that the isomeriza-
tion quantum yields decrease dramatically with in-
creased conversion.
In contrast to the results with DCA and 1, a some-
what different picture emerges for the isomerization of
Dewar benzene 3 with DCA as sensitizer. The values
of λQ from fluorescence quenching experiments are
calculated to be ≈0.3 at 0.1 M and ≈1 at 0.5 M 3. The
experimental isomerization quantum yields are 0.4 and
The glass transition temperature (Tg) of the polymer
medium had a significant effect on the quantum yield
and ultimate extent of conversion. We interpret this
observation in terms of the ability of the medium to
accommodate chain-propagating charge separation. With
functionalized QAI polymers, the Tg was changed either
by altering the polymer structure or by the addition of
plasticizers. An instructive example in the former
category is the comparison of p(DBMA/HMA) and
p(DBMA/cHMA). These two polymers contain similar
alkyl ester groups but have quite different Tgs. The
n-hexyl methacrylate polymer, which has the lower Tg,
shows much higher overall conversion upon prolonged
irradiation (Figure 2), suggesting that the polymer is
better able to accommodate charge separation. This can
be attributed to increased segmental mobility of the
polymer chains, which increases the effective solvating
ability of the polymer.
When the Tg is lowered by the addition of plasticizers,
there is a similar effect on the quantum yield and overall
extent of conversion (Figure 3). In this case, the cause
of the effect is less clear. The added plasticizer could
increase the polarity of the medium by either increasing
the polymer chain mobility or by increasing the average
polarity of the medium. Although our experiments do
not distinguish between these possibilities, studies with
solvatochromic dyes in methacrylate polymers have
shown that the addition of plasticizers principally
increases polarity by increasing polymer chain motion.24
The effect of Tg on reaction efficiency also provides
an explanation for the higher conversion observed when
0.6, respectively. A comparison of the λQ and φisom values
show that the initial-reaction spheres for 3 are, on
average, comparable to or smaller than the fluorescence
quenching spheres. Although the quantum yield does
increase with increasing [3], suggesting that chain
isomerization does occur, the reaction is significantly
less efficient than for 1. The lower efficiency for isomer-
ization of 3 could have several origins. For example, the
isomerization of 3 to 4 might proceed by an intrinsically
less efficient cation radical chain reaction than for 1 to
2
. Quantum yields in fluid solution suggest just the
2
2
opposite is true, however. Therefore, a more likely
explanation is that the cation radical chain isomeriza-
tion of 3 suffers from a chain termination process that
1
does not. Solution studies of 3 and DCA in a nonpolar
solvent have shown that the predominant decay process
-
•
+•
of the DCA /3 geminate ion radical pair is return
1
1
-•
electron transfer to neutral 3. In contrast, the DCA /
+
•
1
exciplex is known to undergo adiabatic isomerization
to DCA /2 with unit efficiency in nonpolar solvent.7
Similar behavior in PMMA would explain the lower
isomerization quantum yields for 3 vs 1.
-
•
+•
Finally, we note that the isomerization quantum
yields for 1 and 3 in polymeric media depend on the
structure of the sensitizer. For 1, the quantum yields
were found to increase along the series DCA, 2-tert-
butyl-DCA (BDCA), and 2,6-di-tert-butyl-DCA (DB-
DCA). This trend can be rationalized in terms of the
expected rate constants for chain-terminating return
electron transfer. The return electron transfers are
highly exothermic for all of these sensitizers and should
lie in the Marcus inverted region. On the basis of the
reduction potentials for DCA (-0.91), BDCA (-0.95),
and DBDCA (-0.99 V vs SCE), return electron transfer
will be least exothermic for DCA and most exothermic
for DBDCA. Consequently, the rate constants for return
electron transfer are expected to decrease from DCA to
BDCA to DBDCA. Thus, the increase in corrected
quantum yields is most likely due to the differences in
reaction energetics for return electron transfer, although
a contribution from modest steric crowding might play
a minor role.
3
is molecularly doped into PMMA, rather than when
the Dewar benzene is incorporated into a functionalized
polymer (compare Figures 1 and 3). When 3 is added
as a dopant, it acts to plasticize the polymer medium
and thereby leads to an increase in the extent of
conversion. When molecularly doped and functionalized
polymers of similar Tg are compared, the extents of
conversions are similar. It is worth noting that the
plasticizers used for the functionalized polymerss
phthalate esterssare significantly more polar than 3.
This provides preliminary evidence that the plasticizers
principally increase the polarity of the polymer medium
by increasing polymer chain mobility rather than by
increasing the average polarity by dilution with a polar
dopant.
The isomerization quantum yield for 3 in polymer was
found to be higher with N-methylquinolinium hexafluo-
rophosphate (NMQ) than with DCA as a sensitizer (1.2
vs 0.6; see Table 1). The reduction potential for NMQ
is less negative than that for DCA (-0.85 vs -0.91 V
vs SCE), which would be expected to increase the rate
constant for return electron transfer for NMQ and lead
Finally, that 1 is capable of achieving a greater overall
conversion than 3, upon prolonged irradiation in poly-
meric media, is consistent with the greater isomeriza-
tion quantum yields for 1. As discussed above, chain-
terminating return electron transfer is less prevalent
for 1; consequently, chain initiation is more efficient for
1
than for 3.
2
3
to lower quantum yield. At present, the cause of the
greater quantum yield for NMQ is not clear; one
possibility is that the partitioning between return
electron transfer and isomerization of the geminate ion
radical pairs differs for the positively charged and
neutral sensitizers.
Experimental Section
Materials. Dichloromethane was distilled from phosphorus
pentoxide under nitrogen and stored over anhydrous sodium
carbonate prior to use. Ether was purified by passing the
2
5
solvent over a bed of activated alumina. Hexamethyl Dewar
benzene (3) was fractionally distilled from lithium aluminum
hydride through a glass helices-packed column at reduced
pressure (bp 47 °C, 17 mmHg). Bis(3-tert-butyl-4-hydroxy-5-
Conversion Studies. As discussed above, the con-
version of significant numbers of reactant molecules
outside the quenching sphere requires charge separation

Macromolecules, Vol. 38, No. 18, 2005
Quantum Amplified Isomerization 7691
methylphenyl) sulfide was a gift from Kodak Research Labo-
ratories and used as received. Methacryloyl chloride was
purified immediately before use by fractional distillation at
atmospheric pressure from a few milligrams of bis(3-tert-butyl-
mounts. The first rail mount held an Oriel 12730 post-mounted
2 in. filter holder that contained the filters used to select a
single mercury line: a Schott GG-395 long-pass cutoff filter
followed by an Oriel 56541 interference filter for 405 nm
irradiations or a Schott WG-335 long-pass cutoff filter followed
by an Oriel 56521 interference filter for 334 nm irradiations.
Solutions used for chemical actinometry were placed in an
Oriel 13950 1 cm cuvette holder, which held the interior front
face of a standard 1 cm cuvette flush with the mounting post.
The front face of the cuvette holder was equipped with a black-
4-hydroxy-5-methylphenyl) sulfide. Triethylamine was distilled
from calcium hydride at atmospheric pressure under nitrogen
before use. Compound 1 was prepared by a literature
7b
procedure. Poly(methyl methacrylate) (PMMA) was obtained
w g
from Polysciences (M ) 25 000, T ) 119 °C) and was used
as received. 2,2′-Azobis(isovaleronitrile) (Vazo-67) was purified
by recrystallization from boiling ligroin. 1,3,5-Trioxane was
sublimed before use. 9,10-Dicyanoanthracene (DCA) was
recrystallized twice from boiling pyridine and once from boiling
acetonitrile. 2-tert-Butyl-9,10-dicyanoanthracene (BDCA) and
2
anodized aluminum light shield containing a 1.45 cm aperture
(0.85 × 1.70 cm) that was centered on the output beam of the
lamp. Film samples were mounted between aluminum plates
(vide infra) in a Thor Labs FP01 sample holder that was
modified to accept an Oriel post so that the film plane was
flush with the mounting post.
2
,6-di-tert-butyl-9,10-dicyanoanthracene (DBDCA) were gifts
from Kodak Research Laboratories and were used as received.
N-Methylquinolinium hexafluorophosphate (NMQ) was pre-
Size exclusion chromatography (SEC) was performed with
uninhibited tetrahydrofuran (THF) (with 0.2% acetone added
to the injection solvent as a flow marker to correct the nominal
1.0 mL/min flow rate) using three Polymer Laboratories 7.5
mm × 300 mm PLgel mixed-C columns, calibrated with narrow
molecular weight distribution polystyrene standards between
molecular weight 580 and 2 300 000, enabling determination
26
pared by a literature procedure. Poly(ethylene terephthalate)
film backing (ESTAR Base) was a gift from Kodak Research
Laboratories and was used as received. The dimensions of the
ESTAR Base were 100 µm thick × 5 in. wide. The ESTAR Base
was coated with a thin (<0.5 µm) layer of latex (15/79/6
acrylonitrile/vinylidene chloride/acrylic acid) to improve coat-
ing adhesion. All other materials were commercially available
and used as received.
of number-average (M
(M ) molecular weights (corrected for axial dispersion assuming
a Gaussian band-broadening function, when appropriate), as
well as polydispersity index (PDI ) M /M ) and intrinsic
n w
), weight-average (M ), and z-average
z
Instrumentation. All laboratory operations with polymer
1
film samples were performed under yellow lights. H NMR
w
n
spectra were recorded on 400 MHz spectrometers equipped
with either a quad nuclear probe or a broadband proton X
probe. UV-vis spectra were recorded on a diode array spec-
trophotometer. High-resolution mass spectrometry was per-
formed on a JEOL HX-110 mass spectrometer operated at
viscosity (dL/g) in THF at 30 °C. The detectors, arranged in
series, were a Waters 2487 Spectroflow spectrophotometric
detector (UV), a Precision Detectors PD2000 light-scattering
(LS) detector operating at 15° and 90°, a Viscotek H502A
differential viscometer (DV), and a Waters 410 differential
refractive index (DRI) detector. The columns, LS, DV, and DRI
detector temperatures were maintained at 30 °C. Polystyrene
(PS) equivalent molecular weights are reported.
10 000 resolution (10% valley definition). Electron impact (EI)
was used wherever feasible with an ionization voltage of 70
eV and perfluorokerosene as a calibrant. Fast atom bombard-
ment (FAB) was performed in a 3-nitrobenzyl alcohol matrix
vs a poly(ethylene glycol) calibrant when EI did not give a
substantial parent ion signal.
The majority of the fluorescence spectra were recorded on
a spectrofluorometer equipped with double-grating excitation
and emission monochromators, a 450 W xenon arc lamp source,
and a thermoelectrically cooled photomultiplier tube detector
and were continuously corrected for fluctuations in excitation
intensity by using a Rhodamine-B reference quantum counter.
Corrections for PMT response were applied to all spectra based
on calibration with a standard lamp according to the manu-
facturer’s instructions.
A TA Instruments model 2920 differential scanning calo-
rimeter (DSC) was used either with a standard heating mode
(linear temperature ramp) or modulated temperature mode
(sinusoidal T-modulation imposed on linear T-ramp) to deter-
g
mine glass transition temperature (T ) values. A nitrogen
purge (50 mL/min) was used unless small molecule dopants
were present. Heating rates were 10 °C/min for standard DSC
or 5 °C/min for modulated DSC (with modulation amplitude
of 0.531 °C and 40 s modulation period, used to enhance weak
transitions, especially at lower temperatures). T
g
values
reported are at the step-transition midpoint. Typically, results
of a second heat are reported, unless an exotherm correspond-
ing to significant isomerization of pendant moieties was
Analytical gas chromatography was performed on a chro-
matograph equipped with a Hewlett-Packard Ultra-2 (5%
phenyl/methyl polysiloxane) capillary column (∼15 m × 0.25
mm × 0.32 µm film thickness).
g g
observed above T or when trying to ascertain the T of a
polymer with dissolved volatile small molecules. The sample
size was typically 10-15 mg, and if dopants were present, a
hermetically sealed aluminum pan was employed to prevent
loss of volatiles during heating.
High-performance liquid chromatography (HPLC) was per-
formed using a PAH Hypersil 150 mm × 3 mm × 5 mm
reverse-phase column, maintained at 15 °C. Elution (2 mL/
min) was according to the following gradient: 25% water/75%
methanol maintained for 5 min and then ramped to 10% water/
Transmission electron microscopy (TEM) was performed on
polymer samples coated on the ESTAR Base. The film samples
were mounted in a microtome, and thin (≈0.1 µm) sections
were prepared. These sections were transferred to a copper
grid with a carbon-coated Formvar film and examined in an
FEI CM 20 transmission electron microscope. The sections
were examined in a cooling holder at liquid nitrogen temper-
atures to minimize beam damage to the samples.
Di-tert-amyl Phthalate. A 500 mL round-bottomed flask
was fitted with a Soxhlet extractor containing a disposable
cellulose thimble filled with 45 g of Linde 4 Å activated
molecular sieves. The flask was charged with 25 mL of 0.2 M
potassium tert-amoxide in tert-amyl alcohol, 200 mL of ben-
zene, and 8.06 g (41.5 mmol) of dimethyl phthalate, and the
reaction mixture was brought to reflux under a nitrogen
atmosphere. Additional 0.2 M potassium tert-amoxide in tert-
amyl alcohol (10 mL) was added after 40 and 90 h. After 110
h, the reaction vessel was cooled to room temperature, and
the reaction mixture was transferred to a separatory funnel
containing 100 mL of water and 50 mL of hexane. The layers
were separated, and the organic layer was washed with water
and brine and dried over anhydrous sodium carbonate. Vola-
9
0% methanol over 25 min. After 0.1 min at the final eluent
concentration, the column was flushed with 100% THF to elute
any polymer residue for 8 min, after which the column was
reequilibrated with 25% water/75% methanol for 8 min prior
to the next injection. Typical injection volumes were 3 µL, and
detection was by diode-array UV-vis spectrum with integra-
tion at individually selected wavelengths for each analyte of
interest.
Optical bench experiments were performed on an apparatus
consisting of an Oriel 68805 200 W universal power supply
powering an Oriel 66002 research lamp housing with integral
collimator assembly equipped with an Oriel 6283 200 W Hg
lamp mounted to an Oriel 11150 high-stability optical rail.
Connected in series to the collimator assembly (kept fully
defocused) of the lamp housing was an Oriel 71430 manual
shutter, an Oriel 6194 liquid filter kept charged with deionized
water, and an Oriel 77660 flanged holder containing a frosted
quartz diffusing element (Edmund Industrial Optics, part no.
F-32406). Also mounted to the rail were two Oriel 11680 rail

7692 Gillmore et al.
Macromolecules, Vol. 38, No. 18, 2005
tiles were removed by rotary evaporation, and the residue was
further concentrated in vacuo (ca. 0.5 mmHg) with gentle
warming to give 11.5 g of a light yellow oil that was distilled
Recrystallization from hexanes yielded 13.0 g (62%) of a white
1
crystalline solid; mp 57-58.5 °C. H NMR (CDCl
3
): δ 4.15 (2
dd, 2H), 1.70 (s, 3.06 H), 1.64 (s, 2.99 H), 1.61 (s, 2.98 H), 1.18
1
3
(
bp 127 °C, 0.7 mmHg) to give 9.6 g (76%) of a viscous, clear,
(s, 3.05 H), 1.14 (s, 2.99 H), 1.10 (t, 0.94 H). C NMR (CDCl
δ 148.2, 146.3, 144.6, 143.5, 57.9, 56.0, 54.8, 11.9, 11.6, 11.3,
10.9, 10.1. Exact mass, C12 18O: calculated 178.1358, found
3
):
1
colorless oil. H NMR (CDCl
3
): δ 7.64 (m, 1.94 H), 7.47 (m,
1
3
1
.96 H), 1.91 (q, 3.99 H), 1.58 (s, 12.21 H), 0.96 (t, 5.90 H).
C
H
NMR (CDCl
3
): δ 166.8, 133.9, 130.4, 128.8, 84.4, 33.7, 25.5,
(M + H): calculated
) -64 °C, T (onset)
m/z 178.1358. While this compound has been previously
2
7
8
.4. Exact mass (FAB-HRMS), C18
H
27
O
4
reported, its characterization has not.
3
)
07.1909, found m/z 307.1891. DSC: T
1 °C, T (onset) ) 23 °C.
MMA Polymerization with and without 3. Two test
g
c
Preparation of 5. To a solution of 5.80 g (32.5 mmol) of
alcohol 7, 6.20 mL (4.50 g, 44.5 mmol) of triethylamine, 0.210
g (1.72 mmol) of 4-dimethylaminopyridine, and 20 mg of bis-
(3-tert-butyl-4-hydroxy-5-methylphenyl) sulfide in 110 mL of
dry dichloromethane stirring under nitrogen was added drop-
wise 3.30 mL (3.50 g, 33.8 mmol) of methacryloyl chloride.
After addition was complete, the reaction mixture was heated
to reflux for 2.5 h. After cooling to room temperature, the
reaction mixture was diluted with 200 mL of ether, succes-
sively washed with water, saturated aqueous sodium bicar-
bonate, 1 M aqueous sodium hydroxide, water, and brine. The
organic phase solution was dried over anhydrous sodium
sulfate, and solvent was removed by rotary evaporation to give
m
tubes were each charged with 0.0138 g of 2,2′-azobis(isobuty-
ronitrile) (AIBN) (0.084 mmol) and 2 mL (1.872 g, 18.7 mmol)
of methyl methacrylate (MMA). To one tube was added 0.120
mL (0.0964 g, 0.594 mmol) of 3. Both tubes were sealed with
septa and the contents vigorously purged with nitrogen for 10
min. The tubes were then immersed in an oil bath maintained
at 65 °C for 24 h. The tubes were broken, and the hard, glassy
polymers in each were dissolved in 15 mL of dichloromethane
with sonication. A portion of each solution was precipitated
into vigorously stirred methanol and dried in a vacuum oven
1
at 80 °C for 2 h. H NMR analysis showed the polymers to be
a clear yellow oil. Vacuum distillation (63 °C, 0.05 mmHg) gave
1
indistinguishable. Molecular weight determinations were made
by size exclusion chromatography. Polymer prepared in the
7.05 g (88%) of a clear colorless oil. H NMR (CDCl
3
): δ 6.15
(s, 0.99 H), 5.60 (s, 0.99 H), 4.66 (br s, 1.99 H), 2.00 (s, 2.93
H), 1.75 (s, 3.04 H), 1.61 (s, 3.00 H), 1.60 (s, 2.98 H), 1.18 (s,
presence of 3 gave M
n
) 145 000, M
w
) 537 000, and PDI )
) 278 000,
w g
) 771 000, and PDI ) 2.77. T s for both polymer samples
1
3
3
.71. Polymer prepared in absence of 3 gave M
n
3.05 H), 1.15 (s, 3.03 H). C NMR (CDCl
144.5, 143.2, 141.8, 136.7, 125.4, 59.8, 56.2, 55.0, 18.6, 12.0,
11.4, 11.2, 10.9, 10.0. Exact mass, C16 calculated
3
): δ 167.5, 150.5,
M
were indistinguishable (127 °C), as were the transition en-
thalpies 0.047 W/g.
22 2
H O :
246.1621, found m/z 246.1624. This compound was routinely
stored over a few milligrams of inhibitor, bis(3-tert-butyl-4-
hydroxy-5-methylphenyl) sulfide. The inhibitor was removed
by passage of the monomer through a short column of basic
alumina prior to use.
The recovery of 3 was determined by repeating the polym-
erization of MMA in the presence of 3 in a fashion identical to
that described above. The resulting hard, glassy polymer was
dissolved in dichloromethane containing an internal standard
(
adamantane). The solution was analyzed by gas chromatog-
Preparation of Functionalized Copolymers. p(DBMA/
MMA). A solution of 4.29 g (17.4 mmol) of 5, 15.7 g (157 mmol)
of methyl methacrylate, and 0.85 g (4.42 mmol) of Vazo-67
[2,2′-azobis(isovaleronitrile)] in 100 mL of chlorobenzene was
deaerated by sparging with nitrogen for 10 min and then
heated at 65-70 °C for 24 h. The resulting polymer was
precipitated into excess methanol, subsequently dissolved in
dichloromethane, and reprecipitated into excess methanol. The
polymer was collected and dried overnight at 80 °C in a
vacuum oven to give 11.6 g of p(DBMA/MMA). The mole ratio
of Dewar benzene to methyl methacrylate monomer units
raphy, and the response factor-corrected percentage of 3
remaining was found to be 98%.
Preparation of 6. A three-necked flask equipped with an
addition funnel, nitrogen inlet, and rubber septum was
charged with 34.2 g (256 mmol) of aluminum chloride, 140 mL
of dry dichloromethane, and a magnetic stir bar. The flask was
immersed in an ice-water bath, and 26.3 g (486 mmol) of
2
-butyne in 70 mL of dry dichloromethane was slowly added
via the addition funnel. The reaction mixture was stirred for
0 min and then transferred via Teflon cannula into a solution
2
1
of 20.1 g (179 mmol) of ethyl 2-butynoate in 70 mL of dry
dichloromethane maintained at 0 °C. The resultant deep red-
brown clear solution was stirred for 15 min at 0 °C, and then
a solution of 70 mL of dimethyl sulfoxide (DMSO) and 170 mL
of dichloromethane was added slowly via an addition funnel.
The reaction mixture was stirred for 45 min and then poured
into 500 mL of ice-water. The resulting mixture was twice
extracted with pentane, and the combined organic extracts
were successively washed with water and brine and then dried
over anhydrous sodium sulfate. The solvent was removed by
rotary evaporation to give 46 g of a clear, light orange oil that
present in the copolymer was determined to be ≈1:9 by H
NMR spectroscopy by integration of the resonances at δ 4.4
ppm (Dewar benzene, allylic ester methylene) and 3.6 ppm
(methyl ester). The molecular weight distribution of the
copolymer was measured by SEC against polystyrene stan-
dards in tetrahydrofuran solvent. M
n
) 27 100; M
w
) 46 900;
) was found
PDI ) 1.73. The glass transition temperature (T
g
to be 115 °C by differential scanning calorimetry performed
in a nitrogen atmosphere at a heating rate of 10 °C/min.
p(DBMA/HMA). An analogous procedure using 2.11 g (8.57
mmol) of 5, 12.90 g (75.77 mmol) of n-hexyl methacrylate, and
0.41 g (2.13 mmol) of Vazo-67 in 75 mL of chlorobenzene gave
was vacuum-distilled (63-67 °C/0.03-0.05 mmHg) to give 28.5
1
g (72%) of a colorless oil. H NMR (CDCl
3
): δ 4.20 (q, 2.01 H),
6.9 g of copolymer. M
n w g
) 19 800; M ) 44 000; PDI ) 2.22; T
2
.05 (s, 3.09 H), 1.65 (s, 3.22 H), 1.60 (s, 2.95 H), 1.30 (t, 2.95
) 4 °C.
13
H), 1.25 (s, 2.95 H), 1.20 (s, 2.82 H). C NMR (CDCl
3
): δ 168.6,
64.5, 145.5, 141.9, 138.4, 59.3, 56.9, 54.7, 14.5, 14.2, 11.3, 11.1,
0.7. Exact mass, C14 : calculated 220.1464, found m/z
p(DBMA/cHMA). An analogous procedure using 0.99 g
(4.02 mmol) of 5, 6.11 g (36.3 mmol) of cyclohexyl methacrylate,
and 0.38 g (1.98 mmol) of Vazo-67 in 75 mL of chlorobenzene
1
1
20 2
H O
2
20.1449. This procedure is analogous to that reported for the
gave 6.2 g of copolymer. M
3.37; T ) 106 °C.
n w
) 12 900; M ) 43 500; PDI )
1
7
methyl ester.
g
Preparation of 7. To 26.0 g (118 mmol) of ester 6 in 500
Preparation of Polymer Films. Dichloromethane solu-
tions of polymers were knife-coated using a 0.005 in. draw-
bar style coating knife on ESTAR Base sheets. Solutions
typically contained ≈1 g of polymer in ≈4 mL of dichlo-
romethane, variable amounts of sensitizer and reactant (for
molecularly doped samples), and plasticizer. All solutions were
filtered through Centrex MF-5 (0.45 µm nylon membrane)
centrifugal filter tubes before coating. For quantum yield
measurements on functionalized polymer films, di-tert-amyl
phthalate was most commonly used as a plasticizer because
it permitted more accurate integration of NMR signals at low
conversion. The ESTAR Base support was mounted on an
mL of dry ether at 0 °C under nitrogen was added 250 mL
(
250 mmol) of a 1.0 M solution of diisobutylaluminum hydride
in hexanes over 40 min. The reaction mixture was stirred 1 h
at 0 °C and 2 h at room temperature. The mixture was cooled
in an ice bath, and 11 mL (8.7 g, 270 mmol) of methanol was
slowly added. A saturated aqueous solution (200 mL) of sodium
sulfate was added, plus an additional 100 mL of ether.
Anhydrous magnesium sulfate was added until a granular
precipitate formed. The mixture was filtered, and the precipi-
tate was washed with dry ether. The combined filtrate was
concentrated to give a pale yellow oil that slowly crystallized.

Macromolecules, Vol. 38, No. 18, 2005
Quantum Amplified Isomerization 7693
aluminum coating block that was maintained at 23-25 °C by
a circulating water bath. The coating block had a hinged cover
that was closed to protect the coating while air-drying. After
drying on the coating block for 10-20 min, the coatings were
cut into individual films using a rotary blade paper trimmer.
These films were mounted between two 50 mm × 60 mm ×
Conversion Experiments. Irradiation of films for conver-
sion experiments were done in a similar manner for quantum
yield measurements except for films containing molecularly
doped 3. Irradiations of these latter films were performed
under a nitrogen atmosphere in a home-built purge cell to
minimize loss of 3/4 that occurred upon prolonged irradiation
in air. The fractional conversion of reactants was determined
as described for the quantum yield measurements.
1/16 in. aluminum plates. The plates had a 23 mm wide × 6
mm deep notch cut in the bottom (50 mm) edge, which fit into
a Thor Labs FP01 sample holder used for photolysis experi-
ments. The plates also contained a 1 in. or a 1.5 in. diameter
aperture that was centered 30 mm above the top of the notch.
Mounted films were dried in vacuo (0.01 mmHg) at either 75-
Control Experiment with 1/Ferrocene/PMMA. Polymer
films containing 0.5 M 1 and 0.6 M ferrocene in PMMA were
prepared as described above. The films were irradiated at 405
1
nm, and the extent of conversion was determined by H NMR
8
0 °C for 1.5 h or room temperature for 12-16 h. Both sets of
using 1,3,5-trioxane as an internal standard, as described
earlier. For a photon radiant energy (i.e., number of photons
absorbed) that resulted in ≈30% conversion of 1 to 2 with DCA
as photosensitizer, films with 1/ferrocene showed <1% isomer-
ization (four repetitions).
drying conditions were demonstrated to leave e1.5 wt %
1
residual dichloromethane in the polymer films by H NMR
analysis vs an internal standard. Under the latter drying
conditions, the thermal isomerization of 1 to 2 in films was
typically <4%. Dried coatings had an ultimate thickness of
Forced Rayleigh Scattering Experiments. Experiments
were conducted with polymer films coated on glass using an
optical table equipped with air suspension legs to provide
vibration isolation. Irradiation was provided by a krypton ion
laser operating at 407 nm. The laser light was spatially filtered
using a UV-grade 0.25 NA microscope objective and a 10 µm
diameter pinhole held by a Optosigma 118-0910 spatial filter
mount. The spatially filtered light was collimated using a 50
mm focal length, 1 cm diameter plano-convex lens. The
collimated beam was split by a 1 × 2 cm Stocker-Yale Lasiris
phase mask (1060.02 ( 0.02 nm grating optimized for first-
order diffraction at 400 nm). The first-order beams were
crossed at the axis of a computer-controlled rotation stage. The
sample was mounted in a Thor Labs FP01 sample holder so
as to hold the film plane at the center of the rotation axis.
The angle of interference (35.55° from normal) and the
wavelength of the laser defined the grating period (350 nm).
Grating formation was monitored with 632.8 nm light from a
helium-neon (HeNe) laser (∼10 mW). The HeNe laser light
was chopped at ≈210 Hz using a Stanford Research Systems
SR540 chopper wheel and frequency controller. The HeNe
probe beam was Bragg matched to the grating (incident at
64.7° from normal), and the diffracted beam was monitored
using a large area (0.43 in. diameter active area) silicon
detector that was interfaced to a personal computer via a lock-
in amplifier set to the chopper frequency.
After writing the diffraction grating, the sample was
transferred to a second computer-controlled sample holder/
rotation stage assembly and mounted for long-term monitor-
ing. The HeNe beam was switched from one stage to the other
with a manual flip mirror. The sample and rotation state were
covered with a light-tight box that contained two small
openings to allow the probe light in and out. During long-term
monitoring the HeNe incident intensity (Iinc) was monitored
as well as the diffracted intensity (Idiff). Idiff/Iinc was measured
to correct for any fluctuations in HeNe source intensity.
Switching from measurement of incident to diffracted light
intensity was computer-automated by using a New Focus 8891
electromechanical flip mirror. A custom-built switch box was
used to switch between output of incident or diffracted beam
detector to the lock-in amplifier. Data were collected and
averaged over 2 min for each detector every 30 min for as long
as desired, typically several weeks. Between data collections,
the motorized flip mirror was positioned to block the sample
from the HeNe beam.
≈
20 µm. After drying, films were examined by UV-vis
spectroscopy in multiple locations at the λmax of the sensitizer
to test uniformity of the film thickness. Only films with OD
variations of <5% were used. Coatings on glass were prepared
in a similar manner.
Fluorometry. The fraction of sensitizer-excited states
intercepted (quenched) by reactant molecules was determined
by fluorometry. Film samples with similar optical density and
sensitizer concentration, with and without quencher, were cut
out by using a 1.5 in. arch punch and mounted between two
1
.5 in. diameter quartz disks in Newport LH-150 mounts. The
mounts were secured in the spectrofluorometer with a home-
built stage. Spectra were recorded in front-face mode, and
samples were excited at the same wavelength used for optical
bench irradiations. The emission spectrum of an identically
mounted film of ESTAR Base was used to background correct
all emission spectra.
Quantum Yield Measurements. Irradiations for quantum
yield determinations were done on the optical bench in an air
atmosphere. The photon flux was determined using phenan-
2
8
threnequinone and 1 mM trans-stilbene in benzene for 405
2
9
nm irradiations and Aberchrome 540 (ca. 5 mM on toluene)
for 334 nm irradiations. Irradiated films were prepared for
analysis by GC or NMR to determine the molecules of product
formed by photolysis as follows. The portion of the film that
had been irradiated through the 1 in. aperture of the alumi-
num plates was marked with a metal scribe using the
aluminum plate as a guide. The film was removed from
between the metal plates, and an area within the scribed
region was cut out with a 15/16 in. arch punch. The polymer
film was cut in half and placed into a scintillation vial. For
polymer films containing molecularly doped cage compound 1
3
samples, 0.7 mL of a CDCl solution containing 1,3,5-trioxane
as internal standard was added, and the vial was sonicated
for 10 min and then allowed to stand 20-30 min with
occasional swirling. The solution was pipetted away from the
Estar film into amber-colored NMR tubes. The numbers of
molecules of 1 and 2 (integrating the four bridgehead methyne
signals for each) were determined vs the known number of
1
molecules of trioxane by H NMR integration. For polymer
films containing molecularly doped 3, 0.4 mL of dichlo-
romethane and a known amount of n-tetradecane as internal
standard were added to the vial. The vial was sonicated for
15 min, and then 1.4 mL cyclohexane was added to precipitate
the polymer. The mixture was filtered and the solution
analyzed by gas chromatography to determine the GC response
factor corrected number of molecules of 3 and 4 present.
Functionalized polymer samples were analyzed as for those
containing 1 except 1,1,2,2-tetrabromoethane was used as an
internal standard. Sufficiently long delay times were used in
the NMR experiments to ensure complete relaxation of nuclei
between rf pulses.
Acknowledgment. Research support was provided
by the National Science Foundation (DMR-0071302)
and, in part, by the Center for Electronic Imaging
Systems (CEIS), a NYSTAR-designated Center for
Advanced Technology. We thank Thomas H. Mourey
and Kim Le for SEC measurements; Michael R. Landry
and Roger E. Moody for DSC measurements; Thomas
C. Jackson for high-resolution MS measurements; Donald
L. Black, Theodore R. Van Dam, and Stephen C. Stoker
for TEM imaging; and Tamara F. Marchinsin and
The reported quantum yields are corrected for the fraction
of incident light absorbed by the samples and for the fraction
of excited states that were quenched, as determined by
fluorometry (vide supra).

7694 Gillmore et al.
Macromolecules, Vol. 38, No. 18, 2005
Timothy C. Schunk for HPLC methods development, all
members of Kodak R&D laboratories. We are grateful
to Ralph H. Young (University of Rochester) for numer-
ous instructive and enlightening discussions. We are
also pleased to acknowledge valuable advice and en-
couragement from Jack C. Chang (Eastman Kodak
Company).
(12) Cf. (a) Hervet, H.; Urbach, W.; Rondelez, F. J. Chem. Phys.
1978, 68, 2725. (b) Lodge, T.; Chapman, B. Trends Polym.
Sci. 1997, 5, 122. (c) Yu, H. Polym. Prepr. 1991, 32, 398.
2
2
(
13) The diffusion coefficient (D) is equal to (Λ /8π t) ln(I0/It),
where Λ is the grating period, I0 is the initial diffraction
efficiency, and I is the diffraction efficiency at time t (ref 12).
t
(14) PC Model, ver. 8.5, Serena Software.
(15) Smoluchowski, M. V. Z. Phys. Chem. 1917, 92, 129.
(
16) Berens, A. R.; Hoppenberg, H. B. J. Membr. Sci. 1982, 10,
283.
References and Notes
(
17) (a) Dopper, J. H.; Greijdanus, B.; Oudman, D.; Wynberg, H.
Tetrahedron Lett. 1975, 48, 4297. (b) van Rantwijk, F.;
Timmermans, G. J.; van Bekkum, H. Pays-Bas 1976, 95, 39.
(
1) Taken in part from the Ph.D. Thesis of Jason G. Gillmore,
University of Rochester, New York, 2003.
(
2) (a) Fyson, J. R.; Twist, P. J.; Gould, I. R. In Electron Transfer
in Chemistry: Molecular-level Electronics, Imaging and
Information, Energy and the Environment; Balzani, V., Ed.;
Wiley-VCH: New York, 2001; Vol. 5, p 285. (b) West, W.;
Gilman, P. B. In The Theory of the Photographic Process, 4th
ed.; James, T. H., Ed.; MacMillan: New York, 1977; p 251.
(18) Greenley, R. Z. In Polymer Handbook, 4th ed.; Brandrup, J.,
Immergut, E. H., Grulke, E. A., Eds.; Wiley: New York, 1999;
pp II/181-308.
(19) Glass transition temperatures were generally experimentally
determined but could be accurately estimated from the weight
fractions of the polymer components. See: Cowie, J. M. G.
Polymers: Chemistry & Physics of Modern Materials; Chap-
man & Hall: New York, 1991; p 327.
(
c) Borsenberger, P. M.; Weiss, D. S. Organic Photoreceptors
for Xerography; Marcel Dekker: New York, 1988.
(
3) (a) Crivello, J. V. NATO Sci. Ser., Ser. E: Appl. Sci. 1999,
(
20) (a) Perrin, F. C. R. Acad. Sci. Paris 1924, 178, 1978. (b)
Frank, I. M.; Vavilov, S. I. Z. Phys. 1931, 69, 100. (c) Perrin,
F. Ann. Chem. Phys. 1932, 17, 283. (d) Lakowicz, J. R.
Principles of Fluorescence Spectroscopy, 2nd ed.; Kluwer:
New York, 1999; Chapter 8.
359, 45. (b) Belfield, K. D., Crivello, J. V., Eds.; Photoinitiated
Polymerization; ACS Symposium Series 847; American Chemi-
cal Society: Washington, DC, 2003. (c) Fouassier, J. P.,
Rabek, J. F., Eds.; Radiation Curing in Polymer Science and
Technology; Elsevier: New York, 1993.
(
21) φ0/φ was found to increase slightly with increasing [DCA].
This observation can reasonably be ascribed to the interven-
tion of F o¨ rster energy transfer at high [DCA]. Excited DCA
molecules that lack a molecule of 1 in their quenching sphere
are more likely to undergo F o¨ rster energy transfer with
neighboring, ground-state DCA molecules as [DCA] increases.
This will lead to a decrease in the fluorescence quantum yield
(
4) (a) Stewart, M. D.; Patterson, K.; Somervell, M. H.; Willson,
C. G. J. Phys. Org. Chem. 2000, 13, 767. (b) Reiser, A.
Photoreactive Polymers; Wiley: New York, 1989.
(
5) See, for example: (a) Evans, T. R.; Wake, R. W.; Sifain, M.
M. Tetrahedron Lett. 1973, 701. (b) Okada, K.; Hisamitsu,
K.; Mukai, T. Tetrahedron Lett. 1981, 1251. (c) Crellin, R.
A.; Lambert, M. C.; Ledwith, A. J. Chem. Soc., Chem.
Commun. 1970, 682. (d) Majima, T.; Pac, C.; Sakurai, H.
Chem. Lett. 1979, 1133. (e) Jones, G., II.; Becker, W. G.;
Chiang, S.-H. J. Photochem. 1982, 19, 245. (f) Borsub, N.;
Kutal, C. J. Am. Chem. Soc. 1984, 106, 4826. (g) Stufflebeme,
G.; Lorenz, K. T.; Bauld, N. L. J. Am. Chem. Soc. 1986, 108,
(
φ) with increasing [DCA] and an increase in φ0/φ. Since
application of eq 8 requires a value of φ0/φ that is uncon-
taminated by any contribution from F o¨ rster energy transfer,
the value of φ0/φ was determined for 0.5 M 1 in PMMA from
plot of φ0/φ vs [DCA] by extrapolation to [DCA] ) 0.
(
(
(
22) Quantum yields for cation radical chain isomerization of 3
are generally greater than those for 1; cf. refs 5a and 5b.
23) These reductions potentials were measured in a polar solvent
and may not reflect the potentials in a polymer medium.
24) (a) Sah, R. E.; Baur, G.; Kelker, H. Appl. Phys. 1980, 23, 369.
4
234. (h) Bauld, N. L. Tetrahedron 1989, 45, 5307. (i) Nelsen,
S. F. Acc. Chem. Res. 1987, 20, 269. (j) Tamai, T.; Ichinose,
N.; Tanaka, T.; Sasuga, T.; Hashida, I.; Mizuno, K. J. Org.
Chem. 1998, 63, 3204.
(
6) Robello, D. R.; Dinnocenzo, J. P.; Farid, S.; Gillmore, J. G.;
Thomas, S. W., III In Chromogenic Phenomena in Polymers;
Jenekhe, S. A., Kiserow, D. J., Eds.; ACS Symposium Series
(b) Eisenbach, C. D.; Sah, R. E.; Baur, G. J. App. Polym. Sci.
1983, 28, 1819.
(
25) (a) Pangborn, A. B.; Giardello, M. A.; Grubbs, R. H.; Rosen,
R. K.; Timmers, F. J. Organometallics 1996, 15, 1518. (b)
Alaimo, P. J.; Peters, D. W.; Arnold, J.; Bergman, R. G. J.
Chem. Educ. 2001, 78, 64.
8
1
88; American Chemical Society: Washington, DC, 2004; p
35.
(
7) (a) Hasegawa, E.; Okada, K.; Mukai, T. J. Am. Chem. Soc.
1
984, 106, 6852. (b) Hasegawa, E.; Okada, K.; Ikeda, H.;
Yamashita, Y.; Mukai, T. J. Org. Chem. 1991, 56, 2170.
8) Mukai, T.; Yamashita, Y. Tetrahedron Lett. 1978, 357.
9) Van Haver, Ph.; Viaene, L.; Van der Auweraer, M.; De
Schryver, F. C. J. Photochem. Photobiol. A: Chem. 1992, 63,
(26) Dockery, K. P.; Dinnocenzo, J. P. Farid, S.; Goodman, J. L.;
Gould, I. R.; Todd, W. P. J. Am. Chem. Soc. 1997, 119, 1877.
(27) Gorycki, P. D.; Macdonald, T. L. Chem. Res. Toxicol. 1994,
7, 745.
(28) (a) Bohning, J. J.; Weiss, K. J. Am. Chem. Soc. 1966, 88, 2893.
(b) Brown-Wensley, K. A.; Mattes, S. L.; Farid, S. J. Am.
Chem. Soc. 1978, 100, 4162.
(
(
265.
(
(
10) (a) Taylor, G. N. Z. Phys. Chem. (Wiesbaden) 1976, 101, 237.
(
b) Jones, G., II.; Chiang, S.-H. J. Am. Chem. Soc. 1979, 101,
7
3
1
421. (c) Jones, G., II.; Chiang, S.-H. Tetrahedron 1981, 37,
397. (c) Jones, G., II.; Becker, W. G. J. Am. Chem. Soc. 1983,
05, 1276.
(
29) (a) Hatchard, C. G.; Parker, C. A. Proc. R. Soc. London, Ser.
A 1956, 235, 518. (b) Heller, H. G.; Langan, J. R. J. Chem.
Soc., Perkin Trans. 2 1981, 341.
11) Kiau, S.; Shukla, D.; Farid, S.; Young, R.; Dinnocenzo, J. P.,
manuscript in preparation.
MA050348K