Nonlinear solvent water effects in the excited-state (formal) intramolecular proton transfer (ESIPT) in m-hydroxy-1,1-diaryl alkenes: Efficient formation of m-quinone methides

Article abstract of DOI:10.1021/ja983557b

The photohydration of hydroxy-substituted 1,1-diaryl alkenes 1-3 has been studied in aqueous CH₃CN solution. Evidence for formation of quinone methide intermediates was provided by product studies and by observation of its absorption spectrum by laser flash photolysis. For the meta isomers, the proposed mechanism of m-quinone methide formation probably involves a solvent-mediated ("proton-relay") excited-state (formal) intramolecular proton transfer (ESIPT) from the phenol hydroxyl group to the β-carbon of the alkene moiety in neutral aqueous CH₃CN solution, either in a concerted manner or via two very fast steps. The m-quinone methides are then trapped by water to form the corresponding diaryl ethanol product with high overall quantum yield. Evidence for the ESIPT pathway was provided by fluorescence and LFP measurements. The addition of small amounts of water (<0.8 M in CH₃CN) decreased the fluorescence emissions of 1 and 2 with a concomitant increase in production of m-quinone methides. Stem-Volmer analyses of fluorescence data revealed a dynamic and a minor static quenching component, both of which involved a water trimer cluster. The degree of charge transfer from the phenol (phenolate) oxygen to the alkene β-carbon, which may be thought of as the driving force for this efficient ESIPT, is pronounced for the meta isomer in the excited state, consistent with Zimmerman's "meta-ortho effect". Alkenes 1 and 2 were shown to be more efficient in m-quinone methide photogeneration than the hydroxy-substituted benzyl alcohol 8, which required conditions of higher water content for similar quantum yields.

Full text of DOI:10.1021/ja983557b

J. Am. Chem. Soc. 1999, 121, 4555-4562
4555
Nonlinear Solvent Water Effects in the Excited-State (Formal)
Intramolecular Proton Transfer (ESIPT) in m-Hydroxy-1,1-diaryl
Alkenes: Efficient Formation of m-Quinone Methides1
Maike Fischer and Peter Wan*
Contribution from the Department of Chemistry, Box 3065, UniVersity of Victoria,
Victoria, British Columbia, Canada V8W 3V6
ReceiVed October 8, 1998
Abstract: The photohydration of hydroxy-substituted 1,1-diaryl alkenes 1-3 has been studied in aqueous
CH3CN solution. Evidence for formation of quinone methide intermediates was provided by product studies
and by observation of its absorption spectrum by laser flash photolysis. For the meta isomers, the proposed
mechanism of m-quinone methide formation probably involves a solvent-mediated (“proton-relay”) excited-
state (formal) intramolecular proton transfer (ESIPT) from the phenol hydroxyl group to the â-carbon of the
alkene moiety in neutral aqueous CH3CN solution, either in a concerted manner or via two very fast steps.
The m-quinone methides are then trapped by water to form the corresponding diaryl ethanol product with high
overall quantum yield. Evidence for the ESIPT pathway was provided by fluorescence and LFP measurements.
The addition of small amounts of water (<0.8 M in CH3CN) decreased the fluorescence emissions of 1 and
2 with a concomitant increase in production of m-quinone methides. Stern-Volmer analyses of fluorescence
data revealed a dynamic and a minor static quenching component, both of which involved a water trimer
cluster. The degree of charge transfer from the phenol (phenolate) oxygen to the alkene â-carbon, which may
be thought of as the driving force for this efficient ESIPT, is pronounced for the meta isomer in the excited
state, consistent with Zimmerman’s “meta-ortho effect”. Alkenes 1 and 2 were shown to be more efficient in
m-quinone methide photogeneration than the hydroxy-substituted benzyl alcohol 8, which required conditions
of higher water content for similar quantum yields.
Introduction
excited-state intramolecular proton transfers (ESIPTs). Direct
evidence for cluster involvement in ESIPT has been obtained
in the gas phase, although these clusters may differ from those
found in the liquid phase.6 Theoretical calculations have also
been used to predict cluster stability and structure.7 Cyclic
systems which allow the hydrogen bonds to approach linearity
for optimum interaction are favored. The size of the cluster has
been shown to depend on the distance between the proton
acceptor and donor as well as on the threshold proton affinity,
which depends on the solvent and the acidity and basicity of
the substrate groups.4-7
ESIPT occurs most readily for those molecules whose acidity
and basicity of both the proton donor and acceptor groups,
respectively, are enhanced upon excitation. ESIPT can readily
lead to the formation of tautomers that are difficult to form via
thermal methods.2,5,8,9 Kalanderopoulos and Yates9 reported the
efficient photohydration of o-hydroxystyrene, which was cited
as the first clear example of an irreversible ESIPT. Due to the
increased acidity of the phenol moiety and the simultaneous
increase in basicity of the alkene moiety upon excitation, direct
Excited-state proton transfers (ESPTs) have been extensively
studied due to their fundamental interest as well as their potential
applications.2 They can act as probes of microscopic solvation
and have revealed the presence of hydrogen-bonded solvent
clusters.3 Fluorescence studies of intermolecular ionizations3,4
and proton-relay proton transfers5 have implied that solvent
clusters can accept or donate protons and that they can mediate
* To whom correspondence should be sent. E-mail: pwan@uvic.ca.
Tel: (250) 721-7150. Fax: (250) 721-7147.
(1) Part of this work has appeared in preliminary form: Fischer, M.;
Wan, P. J. Am. Chem. Soc. 1998, 120, 2680.
(2) (a) Kasha, M.; McMorrow, D.; Parthenopoulos, D. A. J. Phys. Chem.
1991, 95, 2668. (b) Kasha, M. J. Chem. Soc., Faraday Trans. 2 1986, 82,
2379. (c) Kasha, M. Acta Phys. Pol. 1987, 71A, 717. (d) Arnaut, L. G.;
Formosinho, S. J. J. Photochem. Photobiol. A: Chem. 1993, 75, 1. (e)
Formosinho, S. J.; Arnaut, L. G. J. Photochem. Photobiol. A: Chem. 1993,
75, 21.
(3) (a) Moore, R. A.; Lee, J.; Robinson, G. W. J. Phys. Chem. 1985, 89,
3648. (b) Lee, J.; Robinson, G. W.; Webb, S. P.; Philips, L. A.; Clark, J.
H. J. Am. Chem. Soc. 1986, 108, 6538. (c) Robinson, G. W. J. Phys. Chem.
1991, 95, 10386.
(4) (a) Yao, S. H.; Lee, J.; Robinson, G. W. J. Am. Chem. Soc. 1990,
112, 5698. (b) Tolbert, L. M.; Haubrich, J. E. J. Am. Chem. Soc. 1994,
116, 10593. (c) Budac, D.; Wan, P. J. Org. Chem. 1992, 57, 887.
(5) (a) Moog, R. S.; Maroncelli, M. J. Phys. Chem. 1991, 95, 10359.
(b) Chang, C. P.; Hwang, W. C.; Kuo, M. S.; Chou, P. T.; Clements, J. H.
J. Phys. Chem. 1994, 98, 8801. (c) Herbich, J.; Hung, C. Y.; Thummel, R.
P.; Waluk, J. J. Am. Chem. Soc. 1996, 118, 3508. (d) Schwartz, B. J.;
Peteanu, L. A.; Harris, C. B. J. Phys. Chem. 1992, 96, 3591. (e) Mukaihata,
H.; Nakagawa, T.; Kohtani, S.; Itoh, M. J. Am. Chem. Soc. 1994, 116, 10612.
(f) Konijnenberg, J.; Ekelmans, G. B.; Huizer, A. H.; Varma, C. A. G. O.
J. Chem. Soc., Faraday Trans. 2 1989, 85(1), 39. (g) Chou, P. T.; Wei, C.
Y.; Chang, C. P.; Chiu, C. H. J. Am. Chem. Soc. 1995, 117, 7259.
(6) (a) Knochenmuss, R.; Cheshnovsky, O.; Leutwyler, S. Chem. Phys.
Lett. 1988, 144, 317. (b) Nimlos, M. R.; Kelley, D. F.; Berstein, E. R. J.
Phys. Chem. 1989, 93, 643. (c) Nakajima, A.; Hirano, M.; Hasumi, R.;
Kaya, K.; Watanabe, H.; Carter, C. C.; Williamson, J. M.; Miller, T. A. J.
Phys. Chem. A 1997, 101, 392.
(7) (a) Field, M. J.; Hillier, I. H.; Guest, M. F. J. Chem. Soc., Chem.
Commun. 1984, 1310. (b) Fowler, J. E.; Schaefer, H. F. J. Am. Chem. Soc.
1995, 117, 446. (c) Koehler, J. E. H.; Saenger, W.; Lesyng, B. J. Comput.
Chem. 1987, 8, 1090. (d) Gandour, R. D. Tetrahedron Lett. 1974, 295. (e)
Nguyen, M. T.; Raspoet, G.; Vanquickenborne, L. G.; Van Duijnen, P. T.
J. Phys. Chem. A 1997, 101, 7379.
10.1021/ja983557b CCC: $18.00 © 1999 American Chemical Society
Published on Web 04/29/1999

4556 J. Am. Chem. Soc., Vol. 121, No. 19, 1999
Fischer and Wan
ESIPT from the phenol moiety to the â-carbon of the alkene
moiety occurred. The resulting o-quinone methide (o-QM) was
trapped by water to form the aryl ethanol product. We decided
to test the generality of using hydroxy-substituted styrenes as
general QM precursors. Hydroxy-substituted 1,1-diaryl alkenes
1-3 were studied whose acid and base moieties are distal from
each other. We show that photohydration of these substrates
occurs efficiently in aqueous solution, and a solvent-mediated
ESIPT is proposed as being involved in forming the corre-
sponding QM intermediate. Previously, we have shown that
hydroxy-substituted benzyl alcohols and related compounds can
act as general QM photogenerators in aqueous solution.10 The
mechanism involves an initial adiabatic dissociation to the
phenolate, whose formation facilitates hydroxide loss at the
benzylic position in a subsequent step. The relative photo-
solvolytic quantum yields were in the order o > m . p. The
methoxy derivatives, which can only form carbocation inter-
mediates, showed much less or no reaction. In the present study,
we show that 1-3 react significantly more efficiently than the
corresponding hydroxy-substituted benzyl alcohols since the
water-mediated ESIPT mechanism occurs even at low water
concentrations, consistent with a concerted (or very fast)
deprotonation-protonation process.
Figure 1. UV-vis traces for photohydration of 2 in 1:1 H2O-CH3CN.
Each trace, after the first initial absorption, represents 5 min of
photolysis at 254 nm.
nm lamps; ∼15 °C; argon-purged solutions; 2-10 min irradia-
tion). Photolysis of 1-3 yielded the corresponding diaryl ethanol
(e.g., 8 from 1) as the exclusive photoproduct, which is
consistent with a m-QM (i.e., 9 from 1) or a corresponding
diarylmethyl carbocation intermediate being trapped by water.10
Initially, 2 appeared to be photostable in aqueous solution.
However, irradiation in D2O-CD3CN revealed deuterium
incorporation at the 2-position, even at low water concentrations
(eq 1). Photolysis in an NMR tube, which allowed direct analysis
Results and Discussion
During the course of our studies of the photosolvolytic
behavior of a variety of hydroxy-substituted benzyl alcohols,
we found that several of these alcohols readily eliminated to
the aryl alkene during their preparation. This led us to study
the possible photohydration of 1-3, to investigate the effect of
a distal hydroxy substituent on the reaction and whether the
corresponding QMs can be formed as intermediates. The
R-phenyl group ensured a long enough lifetime of the generated
QM (if formed) for detection by nanosecond laser flash
photolysis (LFP). Compounds 4-7 were required for compari-
son purposes in fluorescence, LFP, and product studies since
they lack the hydroxy substituent and thus have only the
potential of leading to the carbocation (or possibly radical
cation). Benzyl alcohol 8 was included to allow for a comparison
between using hydroxy-substituted 1,1-diaryl alkenes vs hy-
droxy-substituted benzyl alcohols as potential QM precursors.
Product Studies. Photolyses were carried out using ∼10-3
M solutions of the substrates dissolved in 1:1 (v/v) aqueous
H2O-CH3CN (Rayonet RPR-100 photochemical reactor; 254
without workup, resulted in only diaryl ethanol formation. This
indicated that the suspected m-QM intermediate 10 (or the
corresponding diarylmethyl carbocation) did not re-form 2 but
rather led entirely to the diaryl ethanol product, which dehy-
drated back to 2 on workup. Photohydration of 1 in H2O-
CH3CN yielded the expected diaryl ethanol 8 cleanly, even at
high conversions (>90%), without dehydration on workup. Also,
photolysis in D2O-CD3CN showed only diaryl ethanol produc-
tion and no deuterium incorporation at the â-position. This lack
of a reverse proton transfer (from the â-carbon) is consistent
with deprotonation of a carbon-hydrogen bond being generally
slow in the ground state and, therefore, unlikely to compete
with nucleophilic attack at the benzylic carbon.
(8) (a) Itoh, M.; Yoshida, N.; Takashima, M. J. Am. Chem. Soc. 1985,
107, 4819. (b) Itoh, M.; Adachi, T. J. Am. Chem. Soc. 1984, 106, 4320. (c)
Choi, J. D. C.; Fugate, R. D.; Song, P.-S. J. Am. Chem. Soc. 1980, 102,
5293. (d) Chou, P. T.; Martinez, M. L.; Cooper, W. C.; McMorrow, D.;
Collins, S. T.; Kasha, M. J. Phys. Chem. 1992, 96, 5203. (e) Chen, Y.;
Gai, F.; Petrich, J. W. J. Am. Chem. Soc. 1993, 115, 10158. (f) Suzuki, T.;
Okuyama, U.; Ichimura, T. J. Phys. Chem. A 1997, 101, 7047.
(9) Kalanderopoulos, P.; Yates, K. J. Am. Chem. Soc. 1986, 108, 6290.
(10) (a) Wan, P.; Barker, B.; Diao, L.; Fischer, M.; Shi, Y.; Yang, C.
Can. J. Chem. 1996, 74, 465. (b) Diao, L.; Yang, C.; Wan, P. J. Am. Chem.
Soc. 1995, 117, 5369. (c) Yang, C. M.Sc. Thesis, University of Victoria,
1994. (d) Diao, L.; Wan, P. Unpublished results.
The UV-vis absorbance of either 1 or 2 decreased smoothly
on photolysis in H2O-CH3CN (Figure 1), the extent of which
increased with increasing H2O concentration (observable in the

Efficient Formation of m-Quinone Methides
J. Am. Chem. Soc., Vol. 121, No. 19, 1999 4557
0-1 M H2O range). Hence, this provides a simple method for
monitoring the extent of photohydration of these compounds.
This absorbance loss was observed to be largest for H2O, less
for D2O, and significantly less for CH3OH (at the same
concentration). This trend is consistent with direct involvement
of H2O (or CH3OH) in the primary photochemical step. A
similar study at high pH (pH > 10) revealed a decrease in
absorbance loss (and hence photohydration yield) of either 1
or 2, which is contrary to that observed for hydroxy-substituted
benzyl alcohols, whose photosolvolysis quantum yields have
been shown to increase at high pH due to direct excitation of
the phenolate.10 This difference suggests that, for 1 and 2,
phenolate formation inhibits diaryl ethanol production, which
is inconsistent with a stepwise mechanism in which the phenol
dissociates adiabatically on excitation, followed by protonation
of the alkene moiety.
Both 1 and 2 gave similar absolute quantum yields for
formation of the corresponding diaryl ethanol (Φp ) 0.22 (
0.02 and 0.24 ( 0.02, respectively in 1:1 H2O-CH3CN). These
values are comparable to that reported for o-hydroxystyrene in
100% water (Φp ) 0.19).9 In contrast, 4 and 5 exhibited much
lower quantum yields (Φp < 0.05) for photohydration. Since
the hydroxy and methoxy substituents have similar electronic
effects, this difference in quantum yields implies that a different
mechanism operates for each. Lacking any sort of electron-
donating substituent, 6 did not photoprotonate under similar
conditions, although the quantum yield for photohydration of
the parent styrene has been reported to be 0.025 in 100% water
at pH 7.11a Only after extensive irradiation did 7 convert to an
observable amount of the corresponding diaryl ethanol product.
Relative quantum yields for conversion of 1 and 3 to their
respective diaryl ethanol product were ∼2.5 times less for the
latter. Thus, 1 apparently generates the corresponding QM more
efficiently, which is consistent with Zimmerman’s “meta-ortho
effect”;12 i.e., the greater electronic transmission effect of the
m-hydroxy group in the excited-state allows more facile
photoprotonation. The above results from product studies
suggest that an ESIPT mechanism leading to m-QM for 1 and
2 is operative, rather than the less efficient direct photoproton-
ation11 of the alkene moiety, leading to the carbocation observed
for the methoxy-substituted and unsubstituted derivatives.
Laser Flash Photolysis (LFP). LFP of 1 or 8 in neat H2O
Figure 2. Absorption spectra of transients generated via LFP of 1 (or
8) in H2O at pH 1 (2), 7 (b), and 11 (9), monitored ∼50 ns after the
laser pulse.
formation of the corresponding p-QM.10 For 2, LFP in 9:1 H2O-
CH3CN yielded a transient whose absorption (λmax ) 415 nm)
was similar to that of 9. However, in this case, a sum of two
first exponential decays was observed (τ ) 4.98 ( 0.07 and
1.11 ( 0.02 µs, with a 36% and 64% contribution, respectively).
The absorption spectra monitored at different time intervals were
the same, showing only the expected variation in intensity. This
is consistent with overlapping carbocation and QM absorptions.
Support for these similar QM and carbocation absorptions was
provided by obtaining similar transient spectra for LFP of 2 in
acidic and basic conditions. At low pH, only carbocation is
expected to form, since the mechanism of acid-catalyzed
photoprotonation of the double bond should dominate in acidic
media, as have been observed for other aryl alkenes.9,11 In turn,
only m-QM should be observable at high pH, since the hydroxy
group would be deprotonated. The transient spectrum observed
at pH 12 was identical to that at neutral pH, and that observed
at pH 1 was blue-shifted by only ∼3 nm. Also, the transient
produced at high and low pHs could be fitted by a single first
exponential decay, indicative of the presence of only one species.
These results suggest that, in neutral conditions, the two
transients with lifetimes τ ) 4.98 ( 0.07 and 1.11 ( 0.02 µs
can be assigned as the m-QM 10 and the corresponding
carbocation 12, respectively (eq 2). We propose that the initially
(pH 7) yielded the same strongly absorbing transient (λmax
)
425 nm) whose decay ended at baseline and was fitted to a
single exponential (τ ) 47 ( 1 ns) (Figure 2). The lifetime and
wavelength maximum of this transient are very similar to those
observed for m-QM 11, which we have previously characterized
by LFP from the corresponding hydroxy-substituted benzyl
alcohol.10 Although this transient absorbs in the same region
as other arylmethyl carbocations, its lifetime is too long. Based
on the results of the product studies and previous LFP data,10
the transient is assigned as 9, whose deprotonated m-hydroxy
group is strongly electron donating (σ ) -0.47), in contrast to
the protonated m-hydroxy group, which is electron withdrawing
(σ ) +0.12) and would thus lead to a shorter lifetime than the
unsubstituted carbocation analogue, the extent of which will
depend on the value of F, which is unknown.10 LFP of the
p-isomer 3 in 9:1 H2O-CH3CN resulted in a very long-lived
transient (λmax ) 350 nm; τ . 1 ms), consistent with the
generated m-QM 10 is sufficiently long-lived for an approach
to equilibrium with the carbocation 12 (requires protonation of
the phenolate anion, which in neutral pH is not expected to be
have a diffusion-controlled rate of protonation from water, the
dominate general acid in solution). However, we do not have a
full understanding of the lifetime difference of the carbocation
at neutral and at low pHs; the carbocation is apparently shorter-
lived (0.26 µs) in the pH 1-3 range. This may reflect
microsolvation effects which depend on the generation pathway.
The potential equilibrium between the initially photogenerated
QM and corresponding carbocation is not observed for 9 (or
11)10b in neutral pH since the lifetime of the initially photo-
(11) (a) Wan, P.; Culshaw, S.; Yates, K. J. Am. Chem. Soc. 1982, 104,
2509. (b) Wan, P.; Shukla, D. Chem. ReV. 1993, 93, 571. (c) Isaks, M.;
Yates, K.; Kalanderopoulos, P. J. Am. Chem. Soc. 1984, 106, 2728. (d)
Wan, P.; Yates, K. ReV. Chem. Int. 1984, 5, 157. (e) Yates, K.; McEwen,
J. J. Am. Chem. Soc. 1987, 109, 5800.
(12) (a) Zimmerman, H. E. J. Am. Chem. Soc. 1995, 117, 8988. (b)
Zimmerman, H. E.; Sandel, V. R. J. Am. Chem. Soc. 1963, 85, 915.

4558 J. Am. Chem. Soc., Vol. 121, No. 19, 1999
Fischer and Wan
Table 1. Photophysical Parameters and Quenching Rate Constants
parameter
1
2
τfa (ns)
5.6 ( 0.3
0.20 ( 0.01
345
3.1 ( 0.5
1.1 ( 0.1
0.08 ( 0.01
0.2 ( 0.2
1.7 ( 0.2
0.047 ( 0.001
425
7.6 ( 0.4
0.23 ( 0.02
353
Φfb
λemc (nm)
kq (H2O)d (108 M-3 s-1
kq (D2O)d (108 M-3 s-1
)
)
1.5 ( 0.3
kq (CH3OH)d (108 M-3 s-1
)
Ks (M-3
)
)
1.3 ( 0.2
1.1 ( 0.2
4.98 ( 0.07h
415
Kd (M-3
τQMe (µs)
λQMf (nm)
g
Φp
0.22 ( 0.02
0.24 ( 0.02
a Fluorescence lifetime in neat CH3CN as measured by single photon
counting. b Fluorescence quantum yield in neat CH3CN. c Fluorescence
emission λmax.
d Dynamic quenching rate constant ((10%) with H2O,
D2O, or CH3OH as quencher. e Lifetime of m-QM in neat H2O for 1
f
and 9:1 H2O-CH3CN for 2. λmax of m-QM. g Photohydration quantum
yield in 1:1 H2O-CH3CN. h An overlapping decay of carbocation 12
(1.11 ( 0.02) µs was also observed.
Figure 3. Decay traces of m-QM 10 generated via LFP of 2 at low
water concentrations (in CH3CN): (a) 0.3, (b) 0.8, (c) 1.7, and (d) 2.8
M.
generated QM (determined primarily by nucleophilic attack by
water) is much shorter, preventing potential equilibration to the
carbocation. LFP of 1 or 8 at pH 1 yielded a blue-shifted (∼12
nm) absorption spectrum (Figure 2), attributable to the carbo-
cation (which is now formed with higher quantum yield via
photoprotonation) which partially decayed within the laser pulse
(therefore, the amount generated at low pH is actually higher
than that shown by its monitored absorbance). LFP of 1 at pH
11 yields essentially the same spectrum as that observed at pH
7, consistent with a m-QM intermediate (9) at both of these
pHs (Figure 2).
Further support for our assignments made above is provided
by LFP of the corresponding methoxy-substituted derivatives
4 and 5, which did not give strongly absorbing transients as
observed for 1 and 2. In 7:3 H2O-CH3CN, only 5 gave a
significant transient (λmax ) 415 nm, τ ) 145 ( 4 ns). This
species has a lifetime that is significantly different from that
assigned to the carbocation from 2. With this in mind, we have
assigned this transient to the radical cation of 5 since methoxy-
substituted styrene derivatives have been shown to give radical
cations (in LFP experiments) with pseudo-first-order decays that
are insensitive to oxygen.13 Both 4 and 5 also gave weak
absorptions at 335 nm with lifetimes of 344 ( 8 and 403 ( 2
ns, respectively, which are unassigned. It should be noted that
all of these transients are significantly weaker than the transients
assigned to QMs for 1-3. The quantum yields for formation
of these transients (from 4 and 5) are probably very low since
no products directly attributable to radical cations or radicals
are observed in the product studies. If simple carbocations are
formed from 4 and 5 (as suggested from the product studies),
their transients were not readily detectable in neutral pH.
Significantly, m-QMs 9 and 10 were photogenerable even at
low water concentrations (in CH3CN) but not in neat CH3CN,
as shown by observation of their characteristic absorption
spectra. The amount produced increased with increasing water
content, and due to nucleophilic quenching by water, a
concomitant decrease in lifetime was observed (Figure 3).
(Lifetime traces were best fitted by a sum of two single
exponentials, both of which showed the same trend in ∆A and
decay with respect to water concentration, suggesting differ-
entially solvated QMs.) Consistent with the similar diaryl ethanol
Figure 4. Representative fluorescence quenching traces of 1 by added
H2O in CH3CN (λex ) 290 nm, λem ) 345 nm).
product quantum yields (Φp) measured for 1 and 2, the relative
amounts of 9 and 10 produced, as measured by ∆As in these
LFP experiments, were similar (assuming 9 and 10 have similar
extinction coefficients). However, 8 did not produce any m-QM
at these low water concentrations, again suggesting that m-QM
formation from 1 and 2 involves a more efficient process (e.g.,
a solvent-mediated ESIPT) rather than the stepwise process
proposed for hydroxy-substituted benzyl alcohols.10
Fluorescence Studies. The strong fluorescence emission
bands observed for 1 and 2 (Table 1) are assigned as the normal
rather than the tautomer emission based on their similarity to
known aryl alkene emissions and the lack of a large Stokes
shift. The emissions (Figure 4) were broad and lacked the fine
structure observed in the excitation spectrum. This suggests that
the molecular geometry in the excited state is less rigid than
the ground state, which is consistent with a charge-transfer
excited state where the alkene bond has been significantly
weakened. Since Φf values are similar for 1 and 2, the charge-
transfer excited state is considered to have a planar structure,
as indene 2 cannot twist significantly. Both 6 and 3 were very
weakly fluorescent.
To gain mechanistic insight into m-QM formation from the
excited singlet state, fluorescence quenching studies by water
were carried out. Efficient quenching of the fluorescence
emission of 1 and 2 by H2O in CH3CN was observed (∼2 M
(13) (a) Johnston, L. J.; Schepp, N. P. J. Am. Chem. Soc. 1993, 115,
6564. (b) Johnston, L. J.; Schepp, N. P. AdV. Electron-Transfer Chem. 1996,
5, 41.

Efficient Formation of m-Quinone Methides
J. Am. Chem. Soc., Vol. 121, No. 19, 1999 4559
reaction to give the same QM as from 1. Also significant was
that water had no effect on the fluorescence emissions of 4-7
at the concentrations used above or at much higher concentra-
tions. Unfortunately, fluorescence quenching of the para isomer
3 could not be studied due to its very weak emission.
Mechanism. The following mechanism for the photohydra-
tion of 1 and 2 is proposed (Scheme 1). The excited state of 1
(or 2) undergoes an ESIPT mediated by a water trimer cluster,
giving rise to the m-QM, which in principle can equilibrate with
the corresponding carbocation by protonation. Our data cannot
distinguish between a concerted proton transfer (via a Grothuss-
type mechanism) or one in which the phenolic proton is ionized
to the water cluster followed by a fast reprotonation at the alkene
â-carbon. Time-resolved studies using picosecond (or faster)
resolution may be required in order to address the mechanism
in more detail. It is clear, however, that the meta grouping of
proton donor and proton acceptor does not hinder formal ESIPT,
which was unexpected. The fast overall rates and high quantum
efficiencies observed are further surprises. The overall ESIPT
is irreversible; the m-QM is completely trapped by water to form
the diaryl ethanol product. The mechanism is supported by the
following observations. The fluorescence intensity of 1 or 2
decreased with increasing water concentration in CH3CN,
accompanied by an increase in the amount of m-QM produced.
These effects were observed at very low water concentrations
in essentially bulk CH3CN. This is consistent with a fluorescence
quenching process leading directly to m-QM formation. The
unsubstituted and methoxy-substituted derivatives and 8 showed
no fluorescence quenching at these low water concentrations,
and all (with the exception of 8) gave only very low diaryl
ethanol product quantum yields. This suggests that the hydroxyl
group is necessary for efficient photohydration and thus rules
out a mechanism involving initial direct photoprotonation of
the alkene moiety. The Stern-Volmer analyses of the fluores-
cence quenching of 1 and 2 revealed a cubed dependence on
water concentration, which suggests the involvement of a water
trimer cluster in the quenching process leading to m-QM
formation. Association of 1 and 2 with a water trimer in the
ground state was also implied. However, the low KS values
suggest that this interaction is very weak but mechanistically
relevant. Direct excitation of this ground-state complex results
in formation of the m-QM directly (probably within a few
picoseconds or less), since the water molecules are already
correctly arranged for ESIPT. The observation of a static
component in the fluorescence quenching further strengthens
our claim that the formal ESIPT is essentially concerted.
Figure 5. Stern-Volmer plots for steady-state (Φfo/Φf, b) and time-
resolved (τo/τ, 9) data in the fluorescence quenching of 2 by H2O (in
CH3CN). The lines drawn are fits to Φfo/Φf ) 1 + (KS + KD)[H2O]3
and τo/τ ) 1 + KD[H2O]3.
H2O resulted in significant quenching) (Figure 4). Methanol
quenched to a much lesser extent, while THF did not quench
at all. A steady-state Stern-Volmer plot was not linear. The
observed curved dependence on water (quencher) concentration
has been previously observed for several other systems.3,4 The
curved Stern-Volmer plot can be fitted to the third power in
H2O concentration (Figure 5). The high-order dependence on
water concentration has been attributed to the involvement of
clusters of size n during excited-state proton transfer.3,4 Both
the steady-state and the lifetime data of 1 and 2 showed a cubed
water dependence. This suggested that water in the form of a
trimer was involved in the deactivation of the singlet excited
state. Significantly, three water molecules would be able to span
the distance between the hydroxy and alkene moieties of 1 and
2 such that the hydrogen bonds approach linearity for optimum
interaction. The extent of quenching as measured by fluores-
cence lifetimes (τ) was significantly less than that observed from
fluorescence intensities (Φf) (Figure 5). This difference implies
that, besides the dynamic ESIPT quenching, static quenching
involving weak complexation with a water trimer was occurring
in the ground state. This weak complexation, however, could
not be observed by UV-vis. Combining the information
provided by the steady-state (eq 3) and the lifetime analyses
(eq 4), static and dynamic Stern-Volmer constants (KS and Kd;
Kd ) kqτ°) were determined (where τo and Φo denote values in
the absence of quencher and kq represents the rate constant of
quenching) (Table 1).
The lack of a normal fluorescence emission has been observed
previously for fast proton transfers upon excitation of ground-
state complexes.5 Any other nonspecifically solvated complexes
are presumed to fluoresce like the uncomplexed form, since they
first need to undergo solvent reorganization before reacting to
form the m-QM. Indeed, the reduced quenching effect observed
for methanol is consistent with involvement of a solvent cluster
as the bulkier and less polar methanol is less able to mediate
the proton transfer.3,4 Further evidence for an ESIPT pathway
was provided by comparison with results for 8, where much
higher water concentrations were required for fluorescence
quenching and for observable m-QM formation. This difference
is indicative of a more efficient process operative for 1 and 2
(such as ESIPT) rather than the two-step process proposed for
the hydroxy-substituted benzyl alcohols.10 Production of m-QM
as well as that of diaryl ethanol product were diminished at
high pH for both 1 and 2, which again is consistent with an
ESIPT mechanism, since if a prior adiabatic dissociation of the
Φfo/Φf ) (1 + KS[Q]3)(1 + KD[Q]3) ≈ 1 + (KS + KD)[Q]3
(3)
τo/τ ) 1 + KD[Q]3
(4)
Dynamic quenching is more rapid for 1 than for 2, which is
consistent with 1 exerting less of a steric effect on the formation
of the cyclic water trimer and substrate complex. For the
dynamic quenching component of 1, a kinetic solvent isotope
effect of 2.8 was observed, which is consistent with those found
for other systems involving solvent clusters in ESPTs.3-6
In contrast, water had a much lesser quenching effect on the
fluorescence of 8 (even 100% H2O did not cause complete
quenching), which is consistent with a different mechanism of

4560 J. Am. Chem. Soc., Vol. 121, No. 19, 1999
Fischer and Wan
Scheme 1
extraction with CH2Cl2, drying with MgSO4) showed an ∼82%
conversion from 8 to 1. Purification was achieved using a silica
columnsan initial wash with hexanes, followed with CH2Cl2 to elute
phenol were involved, more QM and thus more product would
have been expected at high pH, as observed for the hydroxy-
substituted benzyl alcohols.10
1
the oil 1; H NMR and IR data of 1 were consistent with literature
Summary. Hydroxy-substituted 1,1-diaryl alkenes 1-3 photo-
hydrate efficiently via the corresponding QM intermediate. The
results for the meta isomers are consistent with a mechanism
involving water-mediated ESIPT of the phenolic proton to the
â-carbon of the alkene. This ESIPT is not reversible: the QM
is trapped completely by water to form the diaryl ethanol
product. These systems are significant as not only can they be
used as efficient new photogenerators of QM intermediates, but
they can also potentially be used as probes of the dynamics of
water-mediated reactions.
1
data:14 IR (cm-1) 3700-2940, 1610; H NMR (300 MHz, CDCl3) δ
5.04 (s, 1H, OH), 5.44 (d, J ) 1.5 Hz, 1H, dCH2), 5.45 (d, J ) 1.5
Hz, 1H, dCH2), 6.79 (m, 2H, arom), 6.93 (m, 1H, arom), 7.16-7.36
(m, 6H, arom); HRMS (EI), calcd for C14H12O, 196.0889, obsd
196.0887.
7-Hydroxy-3-phenylindene (2). Under nitrogen, a solution of
4-hydroxy-1-indanone (2.0 g, 13 mmol) in THF (250 mL) was added
dropwise to PhMgBr [67 mmol, prepared by adding PhBr (7.1 mL, 67
mmol) in THF (100 mL) to Mg turnings (1.66 g, 68 mmol) in THF
(100 mL) and refluxing for 30 min] in an ice bath. The mixture was
stirred at room temperature for 1 h, refluxed for 3 h, and stirred
overnight. Standard workup gave a dark emulsified solution which was
first filtered before being dried with MgSO4. The sticky crude solid
contained no alcohol but rather showed a ∼95% conversion directly
to 2. A silica column was used for purificationsan initial wash with
hexanes was followed by using CH2Cl2 to elute the white solid 2: mp
123-124 °C; IR (cm-1) 3510-3120, 1620; 1H NMR (300 MHz,
CDCl3) δ 3.45 (d, J ) 2.2 Hz, 2H, CH2), 4.78 (s, 1H, OH), 6.57 (t, J
) 2.2 Hz, 1H, dCH), 6.72 (dd, J ) 6.7, 2.2 Hz, 1H, arom), 7.17-
7.62 (m, 7H, arom); HRMS (EI), calcd for C15H12O, 208.0889, obsd
208.0886.
Experimental Section
General. 1H NMR spectra were recorded on a Bruker AC-300 (300
MHz) instrument, and high-resolution mass spectra (HRMS) were
determined with a Kratos Concept H (EI) instrument. Melting points
were obtained on a Koefler hot-stage microscope and are uncorrected.
Acetonitrile used for the fluorescence studies was dried over CaH2 and
distilled before use. THF and Et2O were also dried and distilled.
Materials. The synthetic precursors 4-hydroxy-1-indanone, 4-meth-
oxy-1-indanone, 1-indanone, 3-hydroxyacetophenone, 3-methoxyaceto-
phenone, and 4-hydroxyacetophenone as well as 6 were purchased from
Aldrich. The initial step in the preparation of compounds 1-5 was the
addition of PhMgBr to the corresponding ketone in THF. Standard
workup of the crude alcohol involved addition of NH4Cl until pH ∼9,
extraction with CH2Cl2, drying with MgSO4, filtration, and evaporation
to dryness in vacuo.
1-m-Hydroxyphenyl-1-phenyl-1-ethanol (8) and m-Hydroxy-r-
phenylstyrene (1). 3-Hydroxyacetophenone (4.0 g, 29 mmol) in THF
(200 mL) was added dropwise to PhMgBr [91 mmol, prepared by
adding PhBr (9.6 mL, 91 mmol) in THF (50 mL) to Mg turnings (2.3
g, 95 mmol) in THF (50 mL) and an I2 crystal] under nitrogen in an
ice bath. More THF (50 mL) had to be added since the mixture became
very thick during reaction. The solution was refluxed for 3 h, and after
standard workup, ∼77% conversion to 8 was obtained. CHCl3 was used
to induce crystallization of the crude oil, and ligroines was used as the
recrystallization solvent to obtain the white solid 8: mp 113-114 °C;
IR (cm-1) 3570-2720; 1H NMR (300 MHz, (CD3)2CO) δ 1.83 (s, 3H,
CH3), 4.55 (s, 1H, OH), 6.61 (dd, J ) 8.3, 1.5 Hz, 1H, arom), 6.80-
7.46 (m, 8H, arom), 8.14 (s, 1H, OH); HRMS calcd for C14H14O2,
214.0994, obsd 214.0994. To obtain 1, H2SO4(aq) (2.7 mL of 2.5 N, 10
mL of H2O added, 6.8 mmol of H+) was added dropwise to crude 8
(∼5.8 g, ∼21 mmol at 77%) in CH3CN (100 mL). After refluxing for
1 h, workup (addition of saturated NaCl and 1 M NaOH until pH ∼5,
p-Hydroxy-r-phenylstyrene (3). A solution of p-hydroxyacetophe-
none (5.0 g, 37 mmol) in THF (250 mL) was added dropwise to
PhMgBr [0.21 mol, prepared by adding PhBr (22 mL, 0.21 mol) in
THF (50 mL) to Mg turnings (5.0 g, 0.21 mol) in THF (50 mL) and an
I2 crystal] in an ice bath. After the solution was stirred for 30 min, the
reaction was complete, and standard workup yielded ∼97% of alcohol.
A few grains of p-toluenesulfonic acid were added to the crude alcohol
(∼0.5 g, ∼2.3 mmol) in toluene (100 mL), followed by a 1 h reflux.
1
Standard workup afforded the oil 3 (98%); H NMR data of 3 were
consistent with literature data:15 IR (cm-1) 3700-2800, 1610; 1H NMR
(300 MHz, CDCl3) δ 4.97 (s, 1H, OH), 5.34 (d, J ) 1.4 Hz, 1H, dCH2),
5.38 (d, J ) 1.4 Hz, 1H, dCH2), 6.78 (d, J ) 8.6 Hz, 2H, arom),
7.19-7.35 (m, 7H, arom); HRMS (EI), calcd for C14H12O, 196.0889,
obsd 196.0885.
m-Methoxy-r-phenylstyrene (4). 3-Methoxyacetophenone (2.0 g,
13 mmol) in THF (100 mL) was added dropwise to PhMgBr (20 mL
of 1 M Aldrich solution, 20 mmol) under nitrogen in an ice bath. The
solution was stirred for 1 h at room temperature, heated at ∼45 °C for
(14) Bruce, J. M.; Chaudry, A.; Dawes, K. J. Chem. Soc., Perkin Trans.
1 1974, 288.
(15) Oude-Alink, B. A. M.; Chan, A. W. K.; Gutsche, C. D. J. Org.
Chem. 1973, 38, 1993.

Efficient Formation of m-Quinone Methides
J. Am. Chem. Soc., Vol. 121, No. 19, 1999 4561
1 h, and stirred overnight, followed by standard workup. The reaction
was not pushed to completion since several side products formed.
H2SO4(aq) (5 mL of 50%, 90 mmol H+) was added to the above crude
mixture in CH3CN (50 mL). Refluxing this solution for 3 h, followed
by workup (extraction with CH2Cl2 and drying with MgSO4), afforded
a relatively clean reaction mixture of ∼43% 4 and ∼57% of starting
ketone. For purification, a silica column was usedsan initial wash with
hexanes, followed by 2:8 CH2Cl2-hexanes to elute the oil 4: 1H NMR
and IR data of 4 were consistent with literature data;16 IR (cm-1) 1600;
1H NMR (300 MHz, CDCl3) δ 3.78 (s, 3H, OCH3), 5.45 (s, 2H, dCH2),
6.83-6.95 (m, 3H, arom); 7.20-7.38 (m, 6H, arom); HRMS (EI), calcd
for C15H14O, 210.1045, obsd 210.1044.
deuterated at position 2. The percent conversion was determined from
the decrease in area of the proton at position 2, due to deuterium
incorporation. The relative intensities of MS (CI) of the mixture [m/z
209 (M+ + 1, 2) and 210 (M+ + 1, deuterated 2)] suggested 76%
conversion. A 25-mg sample dissolved in 1% D2O (0.55 M) in CH3CN
and irradiated for 4 min (16 lamps) yielded a 36% conversion.
Increasing the irradiation time to 15 min (8 lamps) gave a 60%
conversion. The expected diaryl ethanol 4-hydroxy-1-phenyl-1-indanol
(12% and 28%) was produced upon irradiation (25 and 75 min) of a
10-mg sample dissolved in 3 mL of D2O-CD3CN in an 1H NMR
tube: 1H NMR (partial) (300 MHz, CDCl3) δ 2.34 (m, 1H, CH2), 2.76
(m, 1H, CH2), 2.94 (m, 1H, CH2), 6.49 (d, J ) 7.7 Hz, 1H, arom),
6.70 (d, J ) 8.1 Hz, 1H, arom). Upon the addition of acid in excess
D2O to the mixture containing 28% product, complete elimination to
deuterated 2 occurred (25%). In acid, 2 was found to be thermally
labilesa 5.5-mg sample was dissolved in 0.8 mL of CD3CN and 1
drop of concentrated D2SO4, upon which 63% of deuterated 2 was
observed.
Photolysis of 3. A 25-mg sample dissolved in H2O-CH3CN (100
mL) was irradiated (2 min, 8 lamps), yielding 9% 1-p-hydroxyphenyl-
1-phenyl-1-ethanol: 1H NMR (partial) (300 MHz, CDCl3) δ 1.86 (s,
3H, CH3). Increasing the irradiation time to 10 min resulted in 38%.
Photolysis of 4. A 26.3-mg sample dissolved in 1:1 H2O-CH3CN
was irradiated for 10 min (8 lamps), resulting in 8% conversion to
1-m-methoxyphenyl-1-phenyl-1-ethanol: 1H NMR (partial) (300 MHz,
CDCl3) δ 1.93 (s, 3H, CH3), 3.77 (s, 3H, OCH3), 6.77 (dd, J ) 8.8,
2.9 Hz, 1H, arom). Upon increasing the irradiation time to 30 min (16
lamps), a 25-mg sample resulted in 44% conversion.
7-Methoxy-3-phenylindene (5). 4-Methoxy-1-indanone (0.5 g, 3
mmol) dissolved in THF (60 mL) was added dropwise to PhMgBr (10
mL of 1 M Aldrich solution, 10 mmol). The solution was stirred for
30 min and heated at 40-50 °C for 2 h. Standard workup showed an
∼85% conversion directly to 5 with 15% ketone remaining. A silica
column with hexanes as eluant was used to isolate the oil 5: IR (cm-1
)
1
1610; H NMR (300 MHz, CDCl3) δ 3.47 (d, J ) 2.2 Hz, 2H, CH2),
3.93 (s, 3H, OCH3), 6.58 (t, J ) 2.2 Hz, 1H, dCH), 6.81 (d, J ) 8.2
Hz, 1H, arom), 7.21-7.63 (m, 7H, arom); HRMS (EI), calcd for
C16H14O, 222.1045, obsd 222.1042.
3-Phenylindene (7). Following a procedure similar to one reported
previously,17 PhMgBr (98 mL of 1 M Aldrich solution, 98 mmol) in
Et2O (200 mL) was added dropwise to 1-indanone (10.0 g, 75.7 mmol)
in Et2O (100 mL) in an ice bath, resulting in a milky white solution.
After the solution was stirred at room temperature for 3.5 h, saturated
NH4Cl (70 mL) was added slowly, turning the solution a clear yellow
color. The ether layer was decanted, filtered, and dried with Na2SO4,
and the solvent was removed in vacuo, resulting in ∼38% 7 and ∼57%
alcohol with 5% ketone remaining. To a solution of this crude mixture
(2.0 g) in CH3CN (100 mL), H2SO4(aq) (10 mL of 5%, 18 mmol H+)
was added. After heating of the mixture for 1.5 h at 40-50 °C in a
water bath, stirring overnight, and workup (saturated NaCl, extraction
with CH2Cl2, drying with MgSO4), complete conversion of alcohol to
7 occurred. Purification was achieved using a silica column with
hexanes as eluant: the 1H NMR data of 7 were consistent with literature
Photolysis of 7. Only after 1 h of irradiation did a 25-mg sample in
H2O-CH3CN (100 mL) yield 10% 1-phenyl-1-indanol as product: 1H
NMR (partial) (300 MHz, CDCl3) δ 2.49 (m, 1H, CH2), 2.71 (m, 1H,
CH2), 2.95 (m, 1H, CH2), 3.16 (m, 1H, CH2). Increasing the irradiation
time to 2 h resulted in 32% alcohol product.
Quantum Yields. Product quantum yields (Φp) for 1 and 2 were
determined using an Oriel 200-W Hg arc lamp and an Applied Physics
monochromator set for 254 nm. Potassium ferrioxalate19 was used as
the chemical actinometer. UV-vis spectroscopy was used to monitor
for the extent of photohydration of 1 and 2 (1, ꢀ ) (2.20 ( 0.08) ×
103 cm-1 M-1 at 290 nm; 2, ꢀ ) (1.84 ( 0.05) × 103 cm-1 M-1 at 300
nm). Solutions of 1 and 2 (∼10-3 M) in H2O-CH3CN were irradiated
for 5 min in UV-vis cuvettes (conversion ∼10%). Both the standard
and sample solutions were purged with argon 10 min before and during
irradiation in order to eliminate oxygen and ensure a mixed solution.
Only relative quantum yields were determined for 3 and 4. A relative
quantum yield study vs pH was carried out by monitoring the
absorbance change at λmax (loss in absorbance kept below 25%).
Solutions of ∼10-3 M 1 or 2 in 9:1 H2O-CH3CN at a specific pH
were purged for 5-10 min with argon prior to irradiation (2-4 min)
in a Rayonet RPR 100 photochemical reactor with eight lamps at 254
nm using a merry-go-round.
The UV-vis traces for photohydration of 2 in 1:1 H2O-CH3CN
(Figure 1) were obtained in the following manner. A 3.7-mg sample
of 2 in 3 mL of H2O-CH3CN was purged with argon for 5 min, and
then a 0.1-mL aliquot was removed, diluted to 3 mL with H2O-CH3CN,
and its absorption spectrum taken. The remaining concentrated solution
was purged again for 5 min, irradiated for 5 min at 254 nm (8 lamps),
and its absorption spectrum taken (again by removing a 0.1-mL aliquot),
and this process was repeated until a total of 30 min irradiation time.
Laser Flash Photolysis (LFP). All transient spectra and lifetimes
were obtained using a Nd:YAG laser (Spectra Physics Quanta-Ray,
GCR, <30 mJ) with pulse width of ∼10 ns and excitation wavelength
of 266 nm. Flow cells were used, and solutions were purged with
oxygen for approximately 10 min before excitation. Optical densities
at 266 nm were kept below 0.3, ensuring substrate concentrations e10-5
M. No significant differences were observed in LFP runs carried out
in aqueous solution under N2 or O2 purging, indicating that T-T
absorptions are insignificant in our LFP data.
1
data;18 IR (cm-1) 1600; H NMR (300 MHz, CDCl3) δ 3.51 (d, J )
2.2 Hz, 2H, CH2), 6.58 (t, J ) 2.2 Hz, 1H, dCH), 7.22-7.64 (m, 9H,
arom); HRMS (EI), calcd for C15H12, 192.0940, obsd 192.0935.
Product Studies. A solution of ∼10-3 M compound was poured
into a quartz vessel with a cooling finger and purged with argon for
approximately 15 min before and continuously during irradiation in a
Rayonet RPR 100 photochemical reactor using 254-nm lamps (16 lamps
were used unless otherwise stated). The solution was saturated with
NaCl, extracted with CH2Cl2, and dried with MgSO4. The same
procedure was followed for the control, except the argon purge and
1
the irradiation step were omitted. H NMR was used to calculate the
percentage conversion. Isolation of the diaryl ethanol product proved
to be problematic due to its lability on the silica preparative TLC
(conversion back to the diaryl alkene occurred). Therefore, only partial
1H NMR is reported, giving the distinguishable shifts characteristic of
the alcohol product in the mixture with the alkene starting material.
Photolysis of 1. A 25-mg sample was irradiated for 2 min (8 lamps)
in 1:1 H2O-CH3CN (100 mL), yielding 25% of 8. The identity of the
1
product was based on its distinct H NMR peaks within the mixture,
being characteristic of those of 8 synthesized previously. Repeating
the above using an irradiation time of 5 min (16 lamps) resulted in
clean conversion to 94% of 8. The photosolvolysis intermediate of 1
can also be trapped by CH3OHsirradiation for 5 min (16 lamps) in
1:1 H2O-CH3OH gave 76% methyl ether product [1H NMR (partial)
(300 MHz, CDCl3) δ 1.83 (s, 3H, CH3), 3.15 (s, 3H, OCH3)] and 22%
of 8 with 2% of 1 remaining.
Photolysis of 2. A 25-mg sample dissolved in 30 mL of D2O-
CH3CN was irradiated for 30 min, resulting in 82% conversion to 2
(16) Arnold, D. R.; Du, X.; Henseleit, K. M. Can. J. Chem. 1991, 69,
839.
(17) Parham, W. E.; Wright, C. D. J. Org. Chem. 1957, 22, 1473.
(18) (a) Parham, W. E.; Egberg, D. C. J. Org. Chem. 1972, 37, 1545.
(b) Greifenstein, L. G.; Lambert, J. B.; Nienhuis, R. J.; Fried, H. E.; Pagani,
G. A. J. Org. Chem. 1981, 46, 5125.
Steady-State and Time-Resolved Fluorescence Measurements.
All solutions (absorbance ∼0.1 at λex ) 290 nm (except 300 nm for 3
(19) Murov, S. L. Handbook of Photochemistry; Dekker: New York,
1973.

4562 J. Am. Chem. Soc., Vol. 121, No. 19, 1999
Fischer and Wan
and 275 nm for 8) were purged with argon for 10 min before analysis.
Steady-state fluorescence spectra were obtained using a Photon
Technology International A-1010 (PTI) Quanta-Master luminescence
spectrometer. Fluorescence quantum yields (Φf) in CH3CN were
determined by comparing the integrated emission bands of 1 and 2
with that of 2-aminopyridine in 0.05 M H2SO4 and correcting for
differences in the refractive index.20 The fluorescence lifetimes were
measured on a PTI LS-1 instrument using the time-correlated single-
photon-counting technique (5000 or 10 000 counts).
Acknowledgment. Support of this research by NSERC
(Canada) and the University of Victoria is gratefully acknowl-
edged. M.F. thanks NSERC for a post-graduate scholarship.
(20) Eaton, D. E. Pure Appl. Chem. 1988, 60, 1107.
JA983557B