Photoannulation of 4,4-dimethylcyclohex-2-en-1-one to 1,1-diphenylethylene

Article abstract of DOI:10.1021/jp026635x

The photoannulation of 4,4-dimethylcyclohex-2-en-1-one (DMC) to 1,1-diphenylethylene (DPE) results in the [2+2] formation of cyclobutanes exhibiting regio- and stereoselectivity. Additionally, regio- and stereoisomers of substituted octahydrophenanthrenes are produced. Formally, these are derived from alpha-ortho coupling of a 1,4-biradical, followed by rearomatization. From laser flash photolysis experiments, we conclude that the temporal absorbance profile observed following excitation of DMC in the presence of DPE is a composite of multiple triplet state transients that includes triplet state 1,1-diphenylethylene and likely includes a triplet state 1,4-biradical. This analysis of the transient data reveals energy transfer and product forming interactions between ³DMC and ground-state DPE. The present work renders our earlier interpretation [J. Am. Chem. Soc. 1996, 118, 8741],¹ that the transient observed in this system was a ³DMC-DPE exciplex, unnecessary and unlikely.

Full text of DOI:10.1021/jp026635x

J. Phys. Chem. A 2003, 107, 3277-3286
3277
Photoannulation of 4,4-Dimethylcyclohex-2-en-1-one To 1,1-Diphenylethylene
Richard A. Caldwell,* Ralf Constien, and Bryan G. Kriel
Department of Chemistry, The UniVersity of Texas at Dallas, Richardson, Texas 75083
ReceiVed: July 29, 2002; In Final Form: February 26, 2003
The photoannulation of 4,4-dimethylcyclohex-2-en-1-one (DMC) to 1,1-diphenylethylene (DPE) results in
the [2+2] formation of cyclobutanes exhibiting regio- and stereoselectivity. Additionally, regio- and
stereoisomers of substituted octahydrophenanthrenes are produced. Formally, these are derived from alpha-
ortho coupling of a 1,4-biradical, followed by rearomatization. From laser flash photolysis experiments, we
conclude that the temporal absorbance profile observed following excitation of DMC in the presence of DPE
is a composite of multiple triplet state transients that includes triplet state 1,1-diphenylethylene and likely
includes a triplet state 1,4-biradical. This analysis of the transient data reveals energy transfer and product
forming interactions between 3DMC and ground-state DPE. The present work renders our earlier interpretation
[J. Am. Chem. Soc. 1996, 118, 8741],1 that the transient observed in this system was a 3DMC-DPE exciplex,
unnecessary and unlikely.
Introduction
The [2+2] photocycloaddition of R, â-unsaturated carbonyl
compounds to olefins is well established; in fact, it has been
proffered as the most widely used of photosynthetic reactions.2
This annulation has become the preferred route for the construc-
tion of cyclobutane containing compounds and for synthetic
strategies that incorporate such structures. The mechanism of
this reaction has received considerable study. Initially, Corey
proposed a mechanism involving an “oriented π-complex”
(which now would be called an exciplex) to explain regio- and
stereoselectivity in triplet state photocycloadditions of cyclo-
hexenones.3 Although many subsequent investigations thor-
oughly documented the existence and the importance of singlet
exciplexes in singlet state photocycloadditions, by direct detec-
tion, there have been no clear examples of direct detection of
exciplexes in triplet state photocycloadditions.
Indirect evidence, however, abounds. Loutfy and de Mayo
have rationalized their data for the cycloadditions of ketones to
alkenes using an exciplex.4,5 Exciplexes have been claimed in
a number of intramolecular reactions.6-10 Caldwell has found
evidence for exciplexes in the Paterno-Buchi reaction; this was
latter supported by the work of Wagner and Winnik.11-15 Wilson
and Halpern have reported exciplex formation between aliphatic
ketones and benzene derivatives,16,17 and Haselbach has pre-
sented indirect evidence for an exciplex in quenching experi-
ments using benzophenone.18
Figure 1. Major products of DMC and DPE photoannulation reaction.
conclude that exciplexes are not a requirement in explaining
the regiochemical outcome of enone-alkene photoannula-
tions.26-28
Chapman and Rettig have conducted a prior study of the
DMC-DPE photoaddition.29 This study, conducted in benzene
and in tertiary butanol, reports the formation of the trans
cyclobutane, 1, Figure 1. In previous communications, our
laboratory has reported on the DMC-DPE annulation in
cyclohexane1 and reported the simultaneous formation of
cyclobutanes and octahyrophenanthrenes in the p-acetyl styrene-
styrene system.30 The major products of the DMC-DPE
annulation are presented in Figure 1. In our initial DMC-DPE
study, we concluded that the transient spectroscopy indicated a
triplet exciplex intermediate. The study we here describe leads
to a different and we now believe more likely interpretation of
the transient data, although the existence of a triplet exciplex is
not expressly excluded from the reaction. We now believe that
the transient temporal profile results from the composite
Alternatively, Bauslaugh has explained the regio and stereo-
selectivity of ketone-olefin photoannulations on the basis of
the stability of the biradical intermediates.19 Schuster has
questioned the need for exciplexes in cyclohexenone photoan-
nulations and supports an interpretation using a biradical
model.20-23 Peters has interpreted the quenching of ketone
triplets by electron rich alkenes as direct formation of a
biradical.24,25 Biradical-trapping studies have led Weedon to
3
3
absorbance of DMC, DPE, and probably at least one precy-
cloaddition 1,4-biradical. The reversibility of the formation of
the head-to-tail biradical has been demonstrated by irradiation
of the cyclobutanes 1 and 2. The same species are found as in
the photoannulation of DMC and DPE, albeit in different yields.
* To whom correspondence should be addressed. Richard A. Caldwell,
M/S FN32, University of Texas at Dallas, P.O. Box 830688, Richardson
TX 75083-0688. Phone: 972-883-2516 (menu item 7). Fax: 972-883-6371.
E-mail: caldwell@utdallas.edu.
10.1021/jp026635x CCC: $25.00 © 2003 American Chemical Society
Published on Web 04/04/2003

3278 J. Phys. Chem. A, Vol. 107, No. 18, 2003
Caldwell et al.
Experimental Section
ether. The aldehydes 7 and 8 were isolated by preparative TLC
(silica, petroleum ether/10% diethyl ether).
4,4-Dimethylcyclohex-2-en-1-one (DMC) and 1,1-diphenyl-
ethylene (DPE) were obtained from Aldrich Chemical and were
purified by silica gel column chromatography using hexanes
and ethyl acetate, followed by Kugelrohr distillation. UV-vis
spectra were obtained with a HP 8450A spectrophotometer and
the IR spectra with a Nicolet Avatar 360 FT-IR. Gas chroma-
tography was performed on a Shimadzu GC-14 with flame
ionization detection (FID) and on a Finnignan GCQ GC/MS.
The methods of ionization used for the mass spectral analysis
were electron impact (70 eV), methane, and ammonia chemical
ionization. Analytical and semipreparative HPLC separations
were performed on a HP 1090 liquid chromatograph with a
diode array detector (using 270 and 330 nm detection). The
columns used were Lichrosphere (Merck) Si 100, 5 µm (4.0
mm o.d. × 250 mm), and Selectosil (Phenomenex) Si 100, 5
µm (10 mm o.d. × 250 mm). NMR experiments were conducted
using a Varian 500 MHz UnityINOVA spectrometer on CDCl3
solutions containing 0.03% tetramethylsilane. Melting points
were obtained with a Hoover capillary melting point apparatus.
Galbraith Laboratories (Knoxville, TN) performed the elemental
analyses.
Laser flash photolysis experiments were conducted as previ-
ously described with Mode-Locking and Q-Switched Continuum
Nd:YAG lasers, model Y6571C-10.32 An Osram model XBO
150 W/I high-pressure xenon lamp and CVI Digikrom 240
monochromator and photomultiplier provided detection. Data
were collected using a Tektronix DSA 602 digitizing signal
analyzer and transferred to a DTK 486DX personal computer
for analysis. Laser and flash lamp control and the kinetic
analyses were performed with the PC RAD software package
from Kinetic Instruments (Austin, TX). The experiments, unless
otherwise noted, were conducted in cyclohexane at room
temperature in a 5 × 10 mm quartz cuvette. The full-widths-
at-half-maximum height for the mode locked and Q-switched
laser pulses were 1.6 and 5 ns, respectively. The mode-locked
laser was used when the full temporal profile was to be fitted.33
The Q-switch laser was used when fitting only the transient
signal decay and also to collect time-resolved absorption
spectra.34 Figures 4, 6, 8, and 9 exhibit a feature at 50 ns that
is attributed to a cable ring in the laser flash photolysis
equipment. For elevated temperature studies, the cuvette was
placed in a brass block fitted with resistive heating elements.
Dodecane and hexadecane were used as solvents for experiments
conducted at elevated temperatures.
Preparative irradiation of 4, 4-dimethylcyclohex-2-en-1-one
to 1, 1-diphenyethylene was performed with a solution of DMC
(2.45 g, 19.71 mmol) and DPE (7.16 g, 39.72 mmol) in 250
mL of cyclohexane. The solution was irradiated for 24 h using
a standard quartz immersion well photochemical reactor with a
Hanovia 679A 450W medium pressure mercury lamp. A
Corning 3320 uranyl glass filter was added for triplet sensitized
irradiations. The solution was deoxygenated and continuously
mixed by purging with nitrogen for 30 min before and during
the irradiation. After concentrating the reaction mixture to 65
g/L, compound 1 was crystallized at room temperature. The
isolated crystals were further purified by crystallization from
hexanes. The liquor from the initial crystallization was further
concentrated by a factor of 2 and a second crystallization was
performed at -10 °C. These crystals were composed of a
mixture of compounds 3, 4, 5, and 6. These compounds were
isolated by preparative TLC and by semipreparative HPLC.
Compounds 5 and 6 are the result of hydrogen abstraction of
compound 3, resulting in one and two additional unstaturations
(respectively) of the newly formed six-member ring.
The following laser wavelengths, nominal energies, and
3
analytical wavelengths were used for the kinetic studies: DMC
(355 nm, 5 mJ, 300 nm); 3DPE, sensitized by thioxanthone (355
3
nm, 5 mJ, 335 nm); head-to-tail biradical, by decomposition
of 1 and 2 (266 nm, 10 mJ, 335 nm); DMC-DPE temporal
3
profile, resulting from quenching of DMC by DPE (355 nm,
15 mJ, 335 nm). Time-resolved spectroscopy was conducted
on these transients under the same experimental conditions. The
analytical wavelength varied between 300 and 370 nm. Mea-
surement of the quenching rate constant for the energy transfer
of the 3thioxanthone-DPE experiment was performed with
procedures that were previously described.35 The quenching rate
3
constant for the DMC-DPE system was measured similarly;
however; the pseudo-first-order rate constants for the sensitizer
were determined using two kinetic models, vide infra. A DMC
nm
concentration of 79 mmol, OD355
) 1.0 cm -1, and DPE
concentrations between 0 and 0.8 mmol were used in the DMC-
DPE quenching experiments.
Compound 2 could not be isolated in sufficient quantity for
spectral characterization. Its presence was confirmed in the
DMC-DPE reaction mixture by retention time and spectral
matching of HPLC and GC/MS experiments. Preparation of a
standard of cis-5,5-dimethyl-7,7-diphenylbicyclo[4.2.0]octan-
2-one, 2, was performed by adding compound 1 (0.5 mg, 1.64
mmol) to a solution of sodium hydroxide (0.62 g, 15.5 mmol)
in 50 mL of methanol with stirring at room temperature for 12
h. The reaction mixture was neutralized with dilute hydrochloric
acid and extracted 3 times with dichloromethane (100 mL total
volume). Pure 2 was obtained by column chromatography with
petroleum ether/10% diethyl ether.
The transient signal modeling of the laser flash photolysis
studies was performed using two kinetic models. Both models
employ the method of finite differences.36 The mode locked
laser pulse, as detected by the monochronomator, was used to
generate the ERF for iterative reconvolution. Simulations
modeled either the first-order sequential two-reaction scheme
describing the formation of an exciplex or a composite signal
of two first-order decays. The composite model was used to
describe two competing interactions of DMC and DPE. Valida-
tion of the simulators was performed by demonstrating their
ability to fit exact expressions of the respective kinetic model.
The simulators were written in Microsoft Excel workbook files
using the Solver add-in. The Microsoft Excel Solver uses a
generalized reduced gradient (GRG2) nonlinear optimization
code. Linear and integer problems use the simplex method with
bounds on the variables and the branch-and-bound method.37
Fitting was performed by minimization of the sum-of-squares
of the residues of the differences between the simulated and
the experimental data. The sum-of-squares was performed at
one-half nanosecond increments.
Two aldehydes were also observed in the GC/MS and HPLC
analysis but were not isolated. Preparation of standards of 2-(1,1-
dimethyl-allyl)-3,3-diphenyl-trans-cyclobutanecarboxalde-
hyde, 7 and 2-(1,1-dimethyl-allyl)-3,3-diphenyl-cis-cyclobutan-
ecarboxaldehyde, 8, were performed as follows. Solutions of 1
and 2, respectively, (100 mg, 0.32 mmol in 10 mL cyclohexane)
were irradiated in a quartz microscale photochemical reaction
assembly with a mercury lamp at a principal output of 254 nm.31
The reaction mixtures were evaporated and diluted in diethyl

Photoannulation of DMC to DPE
J. Phys. Chem. A, Vol. 107, No. 18, 2003 3279
TABLE 1: X-ray Structure Data of the Cyclobutane 1 and
Octahydrophenanthrene 3
Results
Preparative photoaddition of DMC to DPE resulted in a
mixture of products (Figure 1). The major products were
compounds 1, 2, 3, and 4 (35, 6, 17, and 14 wt %, respectively).
Another 19% of the material balance is accounted for by six
isomers of 3 and 4, none greater than 5 wt %. Compounds 5
and 6, resulting from hydrogen extraction of 3, represent
4% of the material balance, 2% each. Four percent of the bal-
ance was comprised of 7 and 8, approximately two percent
each.
The irradiation of 1 afforded the aldehyde 7 as the major
product with small amounts of DMC and DPE. The irradiation
of 2 resulted primarily in DMC and DPE, with compound 8 as
a minor product. The GC/MS analyses of these reactions indicate
that compounds 1-6 were present in varying amounts for both
reactions.
Spectroscopic Data. trans-5,5-Dimethyl-7,7-diphenylbicyclo-
[4.2.0]octan-2-one, 1: white crystals (1.81 g, 35.3%) mp 167-
168 °C; 1H NMR δ 0.34 (3H, s), 1.25 (3H, s), 1.68-1.74 (1H,
ddd), 1.78-1.88 (1H, dt), 2.08-2.14 (1H, ddt), 2.28-2.34 (1H,
m), 2.34-2.42 (1H, t), 2.90-2.94 (1H, d, J3.39-3.45 ) 13.7 Hz),
2.94-3.0 (1H, dd), 3.39-3.45 (1H, m), 7.14-7.30 (10H, m);
13C NMR δ 18.41, 18.47, 28.68, 34.42, 34.84, 38.10, 42.38,
46.85, 56.84, 60.43, 125.67, 126.40, 128.12, 128.16, 128.41,
141.82, 150.17, 209.67; MS (EI, 70 eV) m/z 304 (M+, 2), 248
(15), 180 (100), 165 (41). Anal.Calcd for C22H24O: C, 86.80;
H, 7.95; O, 5.26. Found: C, 87.25 H, 7.96 O, 4.79. HREI MW
Calcd: 304.1824, Found: 304.1824.
cis-5,5-Dimethyl-7,7-diphenylbicyclo[4.2.0]octan-2-one, 2:
clear crystals (0.34 g, 6.6%) mp 83-84 °C; 1H NMR δ 0.90-
0.96 (1H, ddd), 1.03-1.06 (3H, s), 1.11-1.08 (3H, s), 1.08-
1.34 (1H, m), 2.17-2.23 (1H, m), 2.29-2.37 (1H, m) 2.58-
2.65m (1H, dd), 2.86-2.92 (1H, dt), 3.33-3.38 (1H, dd), 3.51-
3.55 (1H, d, J2.86-2.92 ) 8.3), 7.16-7.30 (10H, m); 13C NMR δ
26.15, 29.47, 31.94, 33.67, 36.79, 37.00, 40.58, 53.59, 55.89,
125.55, 126.14, 126.62, 127.62, 128.11, 129.77, 142.31, 152.19,
215.85; MS (EI, 70 eV) m/z 304 (M+, 3), 248 (15), 180 (100),
165 (44). Anal.Calcd for C22H24O: C, 86.80; H, 7.95; O, 5.26.,
HREI MW Calcd: 304.1824, Found: 304.1835.
1-Keto-4,4-dimethyl-9-cis-phenyl-1,2,3,4,4a,9,10-trans-10a-
octahydrophen-anthrene, 3: white crystals (0.85 g, 17%) mp
164-165 °C; 1H NMR δ 1.00 (3H, s), 1.69 (3H, s), 1.64-1.78
(2H, m), 1.92-2.10 (1H, dt), 2.28-2.46 (3H, m), 2.55-2.58
(1H, dt), 2.85-2.9 (1H, d, J2.55-2.58 ) 12.67 Hz), 4.4-4.8 (1H,
dd, J2.55-2.58 ) 2 Hz, J2.28-2.46 ) 1 Hz), 7.0-7.24 (8H, m),
7.62-7.64 (1H, d); 13C NMR δ 23.66, 29.13, 32.06, 35.06,
35.21, 36.32, 40.19, 42.12, 43.94, 48.71, 125.92, 126.06, 126.09,
128.11, 128.86, 131.02, 138.40, 139.81, 145.27, 213.77; MS
(EI, 70 eV) m/z 304 (M+, 7), 286 (3), 248 (100), 233 (18), 205
(31). Anal. Calcd for C22H24O: C, 86.80; H, 7.95; O, 5.26.
Found: C, 87.33; H, 7.95; O, 4.72; HREI MW Calcd: 304.1824,
Found: 304.1831.
4-Keto-1,1-dimethyl-9-cis-phenyl-1,2,3,4,4a,9,10-cis-10a-oc-
tahydrophenanthrene, 4: white crystals, (0.70 g, 14%) mp 114-
115 °C; 1H NMR δ 0.80 (3H, s), 1.32 (3H, s), 1.62-1.69 (1H,
m), 1.78-1.85 (2H, m), 1.95-2.02 (1H, m), 2.02-2.1 (1H, m),
2.32-2.40 (1H, m), 2.60-2.68 (1H, dd), 3.95-4.05 (1H, d,
J2.02-2.1 ) 1.8 Hz), 4.30-4.38 (1H, dd, J1.78-1.85 ) 1.8 Hz,
J1.95-2.02 <1 Hz), 6.95-7.15 (9H, m); 13C NMR δ 26.95, 27.84,
30.59, 32.52, 35.47, 38.24, 42.67, 43.87, 52.49, 125.88, 127.12,
128.00, 128.69, 129.18, 130.66, 131.27, 134.42, 137.61, 147.33,
210.90; MS (EI, 70 eV) m/z 304 (M+, 97), 286 (18), 271 (22),
245 (30), 233 (88), 204 (100), HREI MW Calcd: 304.1824,
Found: 304.1829.
compound 1
compound 3
formula
formula wt.
crystal system
space group
color
unit cell: a,b,c Å
Z
O1C22H24
304.44
monoclinic
P21/N (no. 14)
clear
11.251,10.884, 14.336 8.684,9.567, 19.776
4
O1C22H24
304.44
monoclinic
P21/C (no. 14)
clear
4
R
RW
0.0453
0.0456
1.168
0.0645
23
0.0388
0.0376
0.745
0.0685
GOF
LinAbs.Coeff.
temp. °C
23
wavelength of rad.(λ, Å) Mo KR (0.71069)
2θ range 4° e 2θe50°
number of data & cutoff 1073 & 3
parameters refined 280
calculated density g/mL 1.157
intensity measure ω-2θ,scan
Mo KR (0.71073)
4° e 2θe50°
1658 & 3
304
1.235
ω-2θ,scan
1-Keto-4,4-dimethyl-9-phenyl-1,2,3,4,4a,-trans-10a-hexahy-
drophenanthrene, 5: white crystals, (0.10 g, 2%) mp 126-127
1
°C; H NMR δ 1.532 (3H, s), 1.537 (3H, s), 1.82-1.84 (1H,
m), 1.84-1.86 (1H, d), 2.35-2.75 (1H, t), 2.45-2.46 (1H, t),
2.63-2.67 (1H, t), 2.67-2.72 (1H, t), 2.95-3.05 (1H, d), 3.11-
3.18 (1H, dd), 6.73-6.75 (1H, d), 7.05-7.38 (8H, m), 7.79-
7.82 (1H, d); 13C NMR δ 19.37, 29.69, 30.30, 32.56, 37.61,
43.73, 47.23, 51.51, 125.76, 126.00, 126.57, 126.71, 127.34,
128.21, 128.71, 137.00, 137.08, 139.66, 140.06, 208.21; MS
(EI, 70 eV) m/z 302 (M+, 41), 285 (6.3), 246 (19), 231 (8), 204
(100), HREI MW Calcd: 302.1668, Found: 302.1676.
1-Keto-4,4-dimethyl-9-phenyl-1,2,3,4a-tetrahydrophenan-
339
Max
355
Max
threne, 6: white crystals, (0.11 g, 2%); UV-vis ꢀ ) ꢀ
)
835 cm-1 mol-1 L (cyclohexane); H NMR δ 1.84 (6H, s),
1
2.08-2.22 (2H, m), 2.78-2.82 (2H, m), 7.4-7.9 (10H, m); 13
C
NMR δ 13.9, 26.7, 34.39, 41.05, 124.09,125.04, 127.4, 128.99,
127.21, 128.27, 129.46, 130.03, 131.03, 131.67, 135.46, 138.62,
139.56, 140.20, 149.35; MS (EI, 70 eV) m/z 300 (M+, 81), 285
(100), 270 (4), 267 (5), 257 (52). HREI MW Calcd: 300.1512,
Found: 300.1516.
2-(1,1-Dimethyl-allyl)-3,3-diphenyl-cyclobutanecarbalde-
1
hyde, 7: oil, (15 mg,); H NMR δ 0.70 (3H, s), 0.79 (3H, s),
2.31-2.43 (1H, m), 2.90-3.09 (1H, m), 3.39 (2H, dd), 4.68-
4.79 (2H, m)5.40 (1H, dd), 9.66 (s,1H), 13C NMR δ 25.2, 33.5,
39.6, 44.6, 53.2, 53.5, 112.0, 125.9, 126.8, 127.0, 127.6, 129.8,
143.8, 145.4, 151.5, 203.1; MS (EI, 70 eV) m/z 304 (M+, 1),
205 (5), 180 (100), 165 (61), 152 (4), 129 (5), 115 (11), 91
(10). IR (KBr): 1766 cm-1 (CdO).
The X-ray structure of trans-5,5-dimethyl-7,7-diphenyl-
bicyclo[4.2.0]-octan-2-one 1 and 4, 4-dimethyl-9-phenyl-3,4,-
4a,9,10,10a-hexahydro-2H-phenanthren-1-one, 3, were deter-
mined, see Table 1 and Figure 2. The preliminary examination
and data collection were accomplished using an Enraf-Nonius
CAD4 diffractometer. Cell constants were obtained by a least-
squares fit of 25 reflections. The space groups of the monoclinic
cells were determined to be P21/n (no 14) for compound 1 and
21/C (no 14) for compound 3. The structures were solved by
direct methods and refined on the basis of 1073 observed
reflections for 1 and 1658 observed reflections for 3. Hydrogen
atoms were calculated at idealized positions and included in
the calculations but not refined. Least-squares refinement was
conducted in the range of 4.00° < 2θ < 50.0° and converged
at R(1) ) 0.0453, R(3) ) 0.0388, and Rw(1) ) 0.0456, Rw(3)
)
0.0376.
The literature values for the lifetimes and the λmax of 3DMC
and 3DPE were reproduced. 3DMC, generated by direct excita-

3280 J. Phys. Chem. A, Vol. 107, No. 18, 2003
Caldwell et al.
Figure 2. Thermal ellipsoid plot of the head-to-tail trans-cyclobutane, compound 1, and the head-to-tail trans-octahydrophenanthrene, compound
3.
transients contribute to yield a composite signal. The lifetime
of the first transient was constrained to 36 ns (the lifetime of
3DPE) and the lifetime of the second transient was constrained
3
to 78 ns (the lifetime of the [biradical]). As seen in Figure 9,
this also resulted in a good fit of the DMC-DPE temporal profile.
Quenching experiments were conducted in order to determine
kq, the quenching rate constant, using the relationship kobs
)
3
kd DMC + kq[Q]. Here, kobs is the observed quenching rate,
and it is dependent on the quencher concentration [Q]. The rate
3
of decay of triplet state DMC is kd DMC, and kq is the quenching
rate constant. Before fitting the data, a signal correction was
3
3
performed for the DMC signal. The DMC correction was
calculated using the rate constant for 3DMC quenching by DPE
(determined iteratively) and the 3DMC signal amplitude-to-laser
power ratio, which was used to scale the amplitude of the
correction. The corrected data was then fitted using the exciplex
model and the composite model, Figure 10. The quenching curve
3
Figure 3. Time-resolved absorption spectrum for the head-to-tail -
[biradical] formed by irradiation of the cyclobutane 1, λmax ) 335 nm.
tion at 355 nm, exhibits an absorption spectrum with a λmax of
295 nm and decays with a lifetime of τ ) 24 ( 1 ns. DPE,
3
3
for the exciplex model yields kd DMC ) 4.91 × 10 7 s-1. The
generated by energy transfer from thioxanthone, has a λmax of
335 nm and a lifetime of τ ) 36 ( 1 ns. The time-resolved
spectrum for the head-to-tail 3biradical generated by photolysis
of 1 is presented in Figure 3 and has λmax ) 335 nm and a
lifetime of τ ) 36 ( 1 ns. A similar spectrum was obtained by
the photolysis of 2. The temporal profile for the head-to-tail
3biradical generated by photolysis of either 1 or 2 resulted in a
78 ns transient. The laser flash photolysis of 1 is presented in
Figure 4. Irradiation of 1 results in 7 as the major product, and
irradiation of 2 results in DPE and DMC as major products.
The photolysis of 2 resulted in a similar temporal profile with
the same first-order lifetime of 78 ( 2 ns. The time-resolved
spectrum for DMC-DPE laser flash photolysis experiment has
a λmax of 337 nm, Figure 5. A first-order fit of the DMC-DPE
temporal profile at 335 nm results in a lifetime of τ ) 49 ns,
Figure 6, which could be used to infer a new species. However,
the measured first-order lifetime of the DMC-DPE temporal
profile is dependent on the analytical wavelength, Figure 7, vide
infra.
3
quenching curve for the composite model yields kd DMC )
7
4.07 × 10 s-1. The independently measured (and literature)
3
7
value of kd DMC is 1/τ ) 4.17 × 10 s-1 is noted in Figure
10. Using the composite model, an Arrhenius plot was generated
within a temperature range of 25 to 150 °C with five temperature
stations. Cyclohexane, dodecane, and hexadecane were used as
solvents. Overlap experiments were performed at the transition
temperatures, and in all cases, these resulted in comparable
results for both solvents. The activation energy for the quenching
3
of DMC by DPE was measured to be 1.1 kcal/mol. The rate
constant is significantly slower than that for a diffusion-limited
event, as is known to be the case for excitation transfer from
nonvertical donors, and the activation energy is lower than that
expected for diffusion.11 There is no requirement that the
activation energy exceed that for diffusion in such a case,
because the quenching act involves both formation and separa-
tion of reactant contact pairs, and diffusive control operates in
both diretions.
The DMC-DPE temporal data of Figure 6 was also fitted
using two kinetic models. In Figure 8, the fit using the exciplex
model (first-order sequential, two reaction kinetics) resulted in
a lifetime of τ ) 48 ns, which is a reasonable agreement with
the first-order fit. Because a small permanent absorbance was
observed, a correction was applied by assuming that it resulted
from decay of the exciplex and that its magnitude was
proportional to the extent of decay. The DMC-DPE temporal
profile was also fitted using a model in which two first-order
Discussion
Compounds 1 and 2 exhibit retro-[2+2] mass spectral
fragmentation, with a low intensity molecular ion (304 amu)
and a base ion at m/z ) 180 (MW of DPE). In the H NMR
spectrum, 1 is best distinguished by the upfield proton shift at
δ ) 0.34 ppm. This shift is due to a phenyl ring affect. The
trans isomer is evidenced by the bridging methine’s axial-axial
coupling constant of 13.7 Hz. Compound 2 exhibits chemical
1

Photoannulation of DMC to DPE
J. Phys. Chem. A, Vol. 107, No. 18, 2003 3281
3
Figure 4. Temporal curve depicting first-order decay of the head-to-tail [biradical] formed by irradiation of the cyclobutane 1, τ ) 78 ( 2 ns,
measured at 335 nm, 27 °C.
Figure 5. Time-resolved absorption spectrum for the DMC-DPE temporal signal, λmax ) 337 nm.
Figure 6. First-order decay of the DMC-DPE temporal profile, τ ) 49 ( 0.5 ns, measured at 335 nm, 27 °C. [DPE] ) 0.7 M.
shifts of the beta hydrogens in the cyclohexane spin system at
δ ) 0.94 and 1.05 ppm because of shielding by the phenyl group
that is trans to the ring junction methines. The bridging methine
hydrogens of 2 possess a relatively large coupling constant of
8.3 Hz due to the small dihedral angle. The mass spectrum of
compound 3 reflects the head-to-tail regiochemistry. The base
ion at m/z ) 248 amu exhibits enolization followed by a retro-
Diels-Alder fragmentation, expected for the head-to-tail isomer
but not the head-to-head isomer.38 The stereochemistry of
compound 3 is trans, as determined by the methine coupling
constant of 12.7 Hz. The stereochemistry of the benzylic
hydrogen, with respect to the methine hydrogen R to the
carbonyl is also trans, this is determined by the coupling
constants of 1-2 Hz with its vicinal hydrogens. These coupling
constants define the hydrogen position as pseudoequatorial and
the phenyl group as axial. The dominant mass spectral ion
fragment of compound 4 is at m/z ) 204 amu; this is the
1-phenylnaphthalene radical cation. In the 1H NMR, the bridging
methine stereochemistry is cis with respect to the methine
hydrogens. This is determined from the methine coupling
constant of 1.8 Hz. The benzylic hydrogen is in an equatorial
position, as seen by the associated coupling constants of 1.8
and less than 1.0 Hz. The phenyl group is pseudoaxial and cis
with respect to the methine hydrogen. Compound 5 exhibits an
additional unsaturation with respect to the 1:1 adducts (M +
)
302 amu) and head-to-tail regiochemistry, again displaying the
retro-Diels-Alder fragmentation observed with compound 3
(m/z ) 246 amu). Reasonable agreement with the calculated
olefinic hydrogen shift δ ) 6.4 ppm with the observed shift of
δ ) 6.6 ppm support the assignment of the additional unsat-
uration. The phenanthrene, 6, exhibits two additional unsatura-
tions with respect to compound 3 (M + ) 300 amu). The large
intensity of the molecular ion and the m-15 ion (m/z ) 285)
indicate development of new aromaticity.

3282 J. Phys. Chem. A, Vol. 107, No. 18, 2003
Caldwell et al.
Figure 7. Dependence of the DMC-DPE transient signal lifetime on the analytical wavelength.
Figure 8. Temporal curve depicting the fit for the first-order sequential two reaction exciplex model for the DMC-DPE temporal profile, τ ) 48
( 2 ns.
Figure 9. Temporal curve depicting the composite signal fit for the DMC-DPE temporal profile, τ1 ) 36, τ2 ) 78 ns.
There are three elementary experiments that may be used to
determine whether laser flash photolysis has produced a single
transient. First, time-resolved spectra should be, except for
amplitude, identical at all times. Second, the natural log of the
decay of the temporal curve should yield a straight line at all
wavelengths. Third, the first-order lifetime of the temporal
profile must be invariant over a wide range of wavelengths.
For a temporal profile composed of multiple transients, the first-

Photoannulation of DMC to DPE
J. Phys. Chem. A, Vol. 107, No. 18, 2003 3283
3
3
Figure 10. Quenching crves generated by the first-order sequential two reaction model and by the composite DPE and head-to-tail biradical
model.
order lifetime is a function of the extinction coefficients and
major product, possesses the characteristics of, and we believe
may well be, the long-lived transient. We expect little variation
of triplet 1,4-biradical lifetime with structure (i.e., head-to-tail
or head-to-head), so we must not rule out isomeric biradicals.
Indeed Weedon39 has shown that biradical yields may be poorly
predicted from product yields.
the concentrations of the species in the system: ODtλ ) ∑in)1
i
ꢀiλcoe-k (t). If the lifetimes and extinction coefficients of the
transients significantly differ, time-resolved spectra may vary
with time, logarithmic plots of decay may be nonlinear, and
the apparent (best-fit) first-order decay rate constant may vary
with wavelength.
To test this biradical hypothesis, the composite model (3DPE,
3
τ1 ) 36 ns and head-to-tail [biradical], τ2 ) 78 ns) was used
We compare the present results to our first DMC-DPE laser
flash photolysis study.1 It is clear from the comparison that such
variations may be small enough that the experimental outcomes
can be misinterpreted. The DMC-DPE time-resolved spectra
in Figure 3 are acceptably uniform, consistent with a single
transient species. Note, however, that multiple transients each
possessing similar spectra may also generate spectra uniform
enough to lead to an unwarranted conclusion of a single
transient. The natural log of the DMC-DPE temporal profile
in Figure 6 appears to be a straight line at normal laser energies
because of poor signal-to-noise late in the decay. Repeating this
experiment at higher high laser energy, i.e., at better signal-to-
noise, we can discern a small curvature. In the third test, at
normal laser energies, the first-order lifetime of the DMC-
DPE temporal profile was a constant 49 ns, over the region
were sufficient absorbance could be measured. This experiment
was also repeated using higher laser energy (21 mJ, focused),
and the dependence of the lifetime on the analytical wavelength
became evident (Figure 7). This is further evidence of the
presence of more than one transient species. The lifetime has a
maximum of 49 ns over a range from 330 to 343 nm and a
minimum value of 38 ns at 365 nm. The 38 ns lifetime is
indicative of the presence of 3DPE, which has a 36 ns lifetime.
We now conclude that a single transient is insufficient to explain
all of the kinetic results.
to fit the DMC-DPE temporal profile. As seen in Figure 9,
the two signals can be combined to fit the DMC-DPE temporal
profile. However, the fit by the exciplex model is comparable;
Figure 8. No conclusive decision can be made based on the fit
alone. From the regression of the quenching data, Figure 10,
3
the predicted rate of decay of DMC by the composite model
is (4.04 ( 0.1) × 107 s-1, a three percent error as compared to
the known rate of decay. The regression for the exciplex model
is 4.91 ( 0.1 × 107 s-1, a seventeen percent error in the
3
prediction of the rate of thermal decay of DMC. Clearly, the
composite model is in better agreement with the known rate of
decay of 3DMC. In summation, the curvature of the natural log
of the temporal profile, the dependence of the first-order lifetime
on wavelength, and the accuracy of the composite model’s
quenching curve are strong arguments that the observed DMC-
DPE temporal profile is (primarily) the result of 3DPE and the
absorbance of a longer-lived species which we expect to be the
head-to-tail 3[biradical]. Other transient species may be present;
however, they are either minor contributors or they closely
mimic one of these two transients.
3
3
The composite DPE and biradical transient argument can
be used to interpret the oxygen and azulene quenching experi-
ments presented in our earlier communication.1 The production
of singlet oxygen by the DMC-DPE system was undertaken
in aerated cyclohexane. S∆ values for the production of singlet
oxygen by DMC (OD355 ) 1.0; ΦT ) 1.0) and DPE (sensitized
by p-methoxyacetophenone, OD355 ) 1.0) were determined to
be 0.55 and 0.19, respectively.40 The Φ∆ for a solution of DMC
(OD355 ) 1.0) and DPE (0.4M) can be predicted from the DMC
and DPE triplet lifetimes, the rate at which the enone triplet is
quenched by DPE (7.5 × 107 L mol-1 s-1), the oxygen
quenching rates (3DMC ) 4.6 × 109 L mol-1 s-1, 3DPE ) 6.3
× 109 L mol-1 s-1) and the S∆ for the oxygen quenching events.
From the new information, we assign, as well as possible,
the transients that are present. First, the shorter lived of the two
3
transients is likely to be DPE, because the 38 ns lifetime is
observed in the spectral region where 3DPE is expected to afford
the dominant absorbance. Second, at least one transient in the
composite has a lifetime significantly greater than 38 ns. Third,
the ratio of the extinction coefficients of the transients must be
relatively constant in the 330-343 nm region. The head-to-tail
3[biradical], cf. Figures 4 and 7, which is the precursor of the

3284 J. Phys. Chem. A, Vol. 107, No. 18, 2003
Caldwell et al.
Figure 11. Reaction scheme for DMC-DPE interactions involving energy transfer, 1,4 annulation, and 1,6 annulation.
If the interaction between the enone triplet and DPE produced
DPE triplets exclusively, the predicted Φ∆ would be 0.092 (
0.004, which is greater than the observation of 0.080 ( 0.004.
The contribution due to quenching of 3DMC is predicted to be
0.069. From this, we conclude that there is modestly more
of these treatments, we suggest that the 3DMC-DPE interactions
partition about one to one between energy transfer and biradical
formation channels.
3
In the azulene quenching experiments, quenching of DMC
by DPE with a rate constant of 7.5 × 107 L mol-1 s-1 leads to
a predicted Φ3AZ*(DPE)/Φ3AZ* of 0.44 (the ratio of the azulene
triplet in the presence and absence of DPE).1 This is to be
compared with the measured Φ3AZ*(DPE)/Φ3AZ* value of 0.47.
The reasonable agreement between the predicted and observed
values indicates that the predominant, and probably the only,
3
singlet oxygen formed than may be explained by DMC-O2
quenching alone. We attribute the excess to 3DPE-O2 quench-
ing, further supporting the presence of 3DPE. That the observed
Φ∆ of 0.080 is between the predicted extremes of complete
energy transfer and no energy transfer is indicative that some
3
3DMC-DPE interactions do not result in DPE formation (as
3
3
3
source of azulene is DMC. DPE is known41 not to transfer
is clear regardless from the formation of the cycloadducts). From
3
3
this analysis, we estimate the yield of 3DPE from the DMC-
energy to azulene, and thus the most reasonable source of -
azulene is 3DMC. 3DMC can generate 3DPE by energy transfer.
The energy transfer is exothermic by 11.5 kcal/mol (DMC -
ET(rel) 63 ( 0.7; DPE - ET(spec) 60.8, ET(rel) 52.1 ( 1.8 kcal
DPE interactions to be 58 ( 18%. A similar result is obtained
from a very different approach. In composite transient modeling,
3
the DPE curve is responsible for approximately 56% of the
mol-1). We have considered that DPE could be generated by
3
composite signal, as determined from the intensities of the
3[biradical] and 3DPE absorbance profiles in Figure 9, assuming
their extinction coefficients are of similar magnitude. As a result
a Schenck-like mechanism. However, this involves an endo-
thermic step (conversion of the Schenck biradical to 3DPE and

Photoannulation of DMC to DPE
J. Phys. Chem. A, Vol. 107, No. 18, 2003 3285
DMC) with an enthalpy of approximately 7.7 kcal/mol, as
estimated by Benson calculations.42,43
ring closure to form stable products. Two types of ring closure
occur. The first is the 1,4-ring closure typical of this reaction;
however, head-to-head cyclobutanes are not formed. The second
type of annulation is an alpha-ortho closure resulting in
unsaturated, substituted octahydrophenanthrenes. No regiose-
lectivity is observed for the octahydrophenanthrenes, from which
we conclude that ring closure is not problematic, as it appears
to be in the 1,4-ring closure. We have concluded that multiple
transient species are involved in this reaction. Among these
It is notable that we observed no head-to-head cyclobutanes.
We conclude that they are formed in at most very low yield.
The head-to-tail biradical has been observed, by photolysis of
1 and 2, with a measured lifetime of 78 ns. The presence of at
least one 78 ns transient with λmax ca. 335 nm is consistent with
our analysis of the DMC-DPE quenching experiments. The
formation of compound 4 may be evidence for the existence of
3
3
transients DPE is a major component. A prima facie case has
a head-to-head biradical, a possibility we are currently inves-
been developed for the presence of the head-to-tail triplet state
biradical. The evidence presented here does not forbid the
existence of an exciplex intermediate. However, there is no basis
in the transient data to invoke a triplet state exciplex.
tigating. The presence of a six-member ring head-to-head
product leads us to believe that it is the failure of ring closure
that prevents the formation of the cyclobutane regio-isomers.
We have at present no convincing explanation for the absence
of head-to-head products.
The hypothesis that the head-to-tail 3biradical is common to
both the DMC-DPE photocycloaddition and the photofrag-
mentation of 1 and 2 places constraints on the products in the
two directions. Our observation that photolysis of either 1 or 2
produces DMC and DPE is consistent with this hypothesis.
Because aldehydes are only observed in the photolysis of either
1 or 2 and not in the photocycloaddition direction, we believe
they must arise by a pathway unavailable in the photocyclo-
addition direction. The aldehydes are Norrish I type products,
requiring R cleavage of ketone 1 or 2, and are clearly
inaccessible in the photocycloaddition direction.
Acknowledgment. This work was supported by the Robert
A. Welch Foundation (Grant AT-0532) and the National Science
Foundation (Grant CHE-9422824). We are indebted to Dr.
Duane Hrncir for the X-ray crystallographic analysis of com-
pounds 1 and 3.
Supporting Information Available: Tables of positional
and thermal parameters and complete bond distances and angles.
Exerts of Excel spreadsheets showing data and cell formulas.
This material is available free of charge via the Internet at
http://pubs.acs.org.
Kelly, Kelly, and McMurry44 have studied the photocyclo-
addition of 3-phenyl-2-cyclohexenone and the 1-phenylpropene
geometric isomers. They have discussed the role of biradicals
in both cycloaddition and the concomitant geometric isomer-
ization of 1-phenylpropene, concluding from triplet excitation
energies that the geometric isomerization probably arises via
the 1,4-biradical and is unlikely to involve excitation transfer.
Our observation of 3DPE demonstrates excitation transfer in the
present case. The triplet excitation energies of 1-phenylpropene
and DPE are similar,45 but DMC has a triplet excitation energy
substantially higher than that of 3-phenyl-2-cyclohexenone and
thus is a better triplet excitation donor.46 The energies in our
case do in fact reasonably permit excitation transfer to DPE, so
there is no anomaly. Kelly et al. also assigned an unidentified
transient as either a triplet exciplex (by analogy to our earlier
report1) or a triplet biradical analogous to those we consider
here. The present results show that our previous study provides
unreliable precedent for a triplet exciplex. Therefore, the
unidentified transient is more likely a triplet biradical.
The presence of octahydrophenanthrenes 3 and 4 requires
hydrogen transfer in each case. The mechanism for octahydro-
phenanthrene formation most reasonably involves triene inter-
mediates, formed by 1,6 coupling of the biradicals, as shown
in Figure 11. The 1,7-hydrogen shift would lead to rearomati-
zation in a strongly exothermic process. Although there is no
cogent direct evidence for the trienes, we note that their expected
absorption in the 300-400 nm spectral region may well be the
source of the permanent absorbance we observe in many of the
temporal profiles.
References and Notes
(1) Caldwell, R. A.; Hrncir, D. C.; Munoz, T.; Unett, D. J. J. Am. Chem.
Soc. 1996, 118, 8741.
(2) Schuster, D. I.; Lem, G.; Kaprindis, N. A. Chem. ReV. 1993, 93, 3.
(3) Corey, E. J.; Bass, J. D.; LeMahieu, R.; Mitra, R. B. J. Am. Chem.
Soc. 1964, 86, 5570.
(4) Loutfy, R. O.; Yip, R. W.; Dogra, S. K. Tetr. Lett. 1977, 2843.
(5) Loutfy, R. O.; de Mayo, P. J. Am. Chem. Soc. 1977, 99, 3559.
(6) Zachariasse, K. A.; Busse, R.; Schrader, U.; Kuhnle, W. Chem.
Phys. Lett. 1982, 89, 303.
(7) Zachariasse, K. A. Trends Photochem. Photobiol. 1994, 3, 211.
(8) Subudhi, P. C.; Lim, E. C. Chem. Phys. Lett. 1978, 58, 62.
(9) Locke, R. J.; Lim, E. C. J. Phys. Chem. 1989, 93, 6017.
(10) Locke, R. J.; Lim, E. C. Chem. Phys. Lett. 1989, 160, 96.
(11) Caldwell, R. A. J. Am. Chem. Soc. 1970, 92, 1439.
(12) Caldwell, R. A. J. Am. Chem. Soc. 1973, 95, 1690.
(13) Caldwell, R. A.; Sovocool, G. W.; Gajewski, R. P. J. Am. Chem.
Soc. 1973, 95, 2549.
(14) Kochevar, I. E.; Wagner, P. J. J. Am. Chem. Soc. 1972, 94, 3859.
(15) Maharaj, U.; Winnink, M. A. J. Am. Chem. Soc. 1981, 103, 2328.
(16) Wilson, T.; Halpern, A. M. J. Am. Chem. Soc. 1980, 102, 7279.
(17) Wilson, T.; Halpern, A. M. J. Am. Chem. Soc. 1981, 103, 2412.
(18) Jacques, P.; Allonas, X.; Von Raumer, M.; Suppan, P.; Haselbach,
E. J. Photochem. A: Chem. 1997, 111, 41.
(19) Bauslaugh, P. G. Synthesis 1970, 287.
(20) Schuster, D. I.; Heibel, G. E.; Brown, P. B.; Turro, N. J.; Kumar,
C. V. J. Am. Chem. Soc. 1988, 110, 8261.
(21) Schuster, D. I. In The Chemistry of Enones; Patai, S., Rappoport,
Z., Eds.; John Wiley and Sons Ltd: New York, 1989; pp 623-756.
(22) Schuster, D. I.; Heibel, G. E.; Woning, J. Angew. Chem., Int. Ed.
Engl. 1991, 30, 1345.
(23) Schuster, D. I.; Lem, G.; Kaprindis, N. A. Chem. ReV. 1993, 93, 3.
(24) Freilich, S. C.; Peters, K. S. J. Am. Chem. Soc. 1981, 103, 6255.
(25) Freilich, S. C.; Peters, K. S. J. Am. Chem. Soc. 1985, 107, 3819.
(26) Andrew, D.; Hastings, D. J.; Weedon, A. C. J. Am. Chem. Soc.
1994, 116, 10870.
(27) Hastings, D. J.; Weedon, A. C. J. Am. Chem. Soc. 1991, 113, 8525.
(28) Maradyn, D. J.; Weedon, A. C. Tetrahedron Lett. 1994, 35, 8107.
(29) Rettig, T. A., Ph.D. Thesis, Iowa State University, Ames, Iowa,
1965.
(30) Our laboratory has observed this type of photoannulation with
p-acetylstyrene and styrene, a tetralin was produced in low yield. See:
Caldwell, R. A.; Diaz, J. F.; Hrncir, D. C.; Unett, D. J. J. Am. Chem. Soc.
1994, 116, 8138.
Conclusion
A reaction mechanism for the photoaddition of DMC to DPE
is presented in Figure 11. There are two initial interactions
between 3DMC and DPE: energy transfer and biradical forma-
3
tion. After the energy transfer, the DPE thermally relaxes to
ground-state DPE without chemical bond formation. The
biradicals may revert to ground-state reactants or may undergo
(31) Penn, J. H.; Orr, R. D. J. Chem. Educ. 1989, 66, 86.

3286 J. Phys. Chem. A, Vol. 107, No. 18, 2003
Caldwell et al.
(32) Caldwell, R. A. In Kinetics and Spectroscopy of carbenes and
Biradicals; Platz, M. S., Ed.; Plenum Publishing Corp.: New York, 1990;
pp 77-116.
(40) S∆ and Φ∆ refer to the fraction of oxygen quenching events leading
to the production of singlet oxygen and the overall singlet oxygen quantum
yield, respectively. The latter is simply the product of the quantum yield
for formation of the triplet state being quenched by oxygen, ΦT, and S∆.
For more information, see: Gorman, A. A.; Rodgers, M. A. J. In Handbook
of Organic Photochemistry; Scaiano, J. C., Ed.; CRC Press: Boca Raton,
FL, 1989; pp 229-250.
(33) A temporal profile is a plot of optical density versus time of transient
species.
(34) A time-resolved spectra are spectra collected at different times in
the decay of a transient.
(35) Ni, T.; Caldwell, R. A.; Melton, L. A. J. Am. Chem. Soc. 1989,
111, 457.
(36) Milne-Thomson, L. M. The Calculus of Finite Differences.;
Macmillan & Co. Ltd.: New York, 1960.
(37) For more information on the internal solution process used by
Solver, contact: Frontline Systems, Inc. PO Box 4288 Incline Village, NV
89450-4288. Phone: (702) 831-0300.
(38) Silverstein, R. M.; Bassler, G. C.; Morril; T. C. Spectrometric
Identification Of Organic Compounds, 4th ed.; John Wiley and Sons: New
York, 1981; pp 15 and 208-233.
(39) Andrew, D.; Weedon, A. C. J. Am. Chem. Soc. 1995, 117, 5647-
5663. Maradyn, D. J.; Weedon, A. C. J. Am. Chem. Soc. 1995, 117, 5359-
5360.
(41) Unpublished result of Unett, D. J. 3DPE generated by thioxanthone
energy transfer is not quenched by the addition of azulene.
(42) Bader, M. J. Chem. Educ. 1980, 57, 703.
(43) Lowry, T. H.; Richardson, K. S. Mechanism And Theory In Organic
Chemistry, 3rd ed.; Harper & Row Publishers: New York, 1987; pp 159-
177.
(44) Kelly, J. F. D.; Kelly, J. M.; McMurry, T. B. H. J. Chem. Soc.,
Perkin Trans. 2 1999, 1933-1941.
(45) Caldwell, R. A.; Ni Tuqiang, L. A.; Melton, J. Am. Chem. Soc.
1989, 111, 457-464.
(46) Caldwell, R. A.; Tang, W.; Schuster, D. I.; Heibel, G. E. Photochem.
Photobiol. 1990, 52, 645-648.