Photochemistry and photophysics of halogen-substituted dibenzothiophene oxides

Article abstract of DOI:10.1021/jo0490346

Dibenzothiophene-5-oxide (DBTO) cleanly produces dibenzothiophene (DBT) on direct photolysis, but with very low quantum yield. A proposed mechanism involves scission of the S-O bond which is coupled to an intersystem crossing step, thus producing the sulfide and O(³P) via a unimolecular pathway. To test this hypothesis, heavy atom substituted DBTOs were prepared and photolyzed. Iodo-, bromo-, and chloro-substituted DBTOs show higher quantum yields for deoxygenation than does the parent molecule, in the order consistent with an intersystem crossing-related heavy atom effect. 2-Iododibenzothiophene also undergoes photochemical deiodination. Phosphorescence data are consistent with heavy-atom assisted intersystem crossing.

Full text of DOI:10.1021/jo0490346

VOLUME 69, NUMBER 24
NOVEMBER 26, 2004
© Copyright 2004 by the American Chemical Society
Articles
Photochemistry and Photophysics of Halogen-Substituted
Dibenzothiophene Oxides1
Mrinmoy Nag and William S. Jenks*
Department of Chemistry, Iowa State University, Ames, Iowa 50011
wsjenks@iastate.edu
Received June 8, 2004
Dibenzothiophene-5-oxide (DBTO) cleanly produces dibenzothiophene (DBT) on direct photolysis,
but with very low quantum yield. A proposed mechanism involves scission of the S-O bond which
is coupled to an intersystem crossing step, thus producing the sulfide and O(3P) via a unimolecular
pathway. To test this hypothesis, heavy atom substituted DBTOs were prepared and photolyzed.
Iodo-, bromo-, and chloro-substituted DBTOs show higher quantum yields for deoxygenation than
does the parent molecule, in the order consistent with an intersystem crossing-related heavy atom
effect. 2-Iododibenzothiophene also undergoes photochemical deiodination. Phosphorescence data
are consistent with heavy-atom assisted intersystem crossing.
Introduction
and some of its derivatives, it is the major sulfur-con-
taining product.3,13-16 From a mechanistic perspective,
many experimental results are consistent with simple
S-O cleavage that yields the sulfide and O(3P).13,14,16 For
example, product studies based on the oxidation of
solvents and other reactive traps in the presence of
DBTO are consistent with expectations for O(3P). Though
no direct evidence has been obtained for the formation
of this reactive intermediate, the evidence pointing
toward a unimolecular mechanism is compelling.10,13,16
One of the fundamental photochemical reactions of
aromatic sulfoxides is deoxygenation to form the corre-
sponding sulfide on direct irradiation.2-12 Though the sul-
fide is usually a minor component of the product mixture,
in the photolyses of dibenzothiophene-5-oxide (DBTO)
(1) Photochemistry and Photophysics of Aromatic Sulfoxides. 21. For
part 20, see: J. Am. Chem. Soc. 2002, 124, 2544-2547.
(2) Shelton, J. R.; Davis, K. E. Int. J. Sulfur Chem. 1973, 8, 217-
228.
(3) Gurria, G. M.; Posner, G. H. J. Org. Chem. 1973, 38, 2419-2420.
(4) Still, I. W. J.; Arora, P. C.; Chauhan, M. S.; Kwan, M.-H.;
Thomas, M. T. Can. J. Chem. 1976, 54, 455-470.
(5) Still, I. W. J. In The Chemistry of Sulfones and Sulfoxides; Patai,
S., Rappaport, Z., Stirling, C. J. M., Eds.; John Wiley & Sons Ltd.:
New York, 1988; pp 873-887.
(6) Still, I. W. J.; Cauhan, M. S.; Thomas, M. T. Tetrahedron Lett.
1973, 1311-1314.
Thus, DBTO derivatives are promising candidates as
photochemical precursors for the study of O(3P) chemistry
in solution.14 Other methods of oxygen atom formation
invariably either require very high energy irradiation or
precursors that are themselves involve oxidants (e.g.,
ozone). To further test and exploit the O(3P) hypothesis,
(7) Still, I. W. J.; Thomas, M. T. Tetrahedron Lett. 1970, 4225-4228.
(8) El Amoudi, M. S. E. F.; Geneste, P.; Olive´, J.-L. J. Org. Chem.
1981, 46, 4258-4262.
(9) Jenks, W. S.; Gregory, D. D.; Guo, Y.; Lee, W.; Tetzlaff, T. In
Organic Photochemistry; Ramamurthy, V., Schanze, K. S., Eds.; Marcel
Dekker: New York, 1997; Vol. 1, pp 1-56.
(10) Kumazoea, K.; Arima, K.; Mataka, S.; Walton, D. J.; Thiemann,
T. J. Chem. Res., Synop. 2003, 60-61.
(11) Thiemann, T.; Kumazoe, K.; Arima, K.; Mataka, S. Kyushu
Daigaku Kino Busshitsu Kagaku Kenkyusho Hokoku 2001, 15, 63-
71.
(13) Lucien, E.; Greer, A. J. Org. Chem. 2001, 66, 4576-4579.
(14) Thomas, K. B.; Greer, A. J. Org. Chem. 2003, 68, 1886-1891.
(15) Wan, Z.; Jenks, W. S. J. Am. Chem. Soc. 1995, 117, 2667-2668.
(16) Gregory, D. D.; Wan, Z.; Jenks, W. S. J. Am. Chem. Soc. 1997,
119, 94-102.
(12) Thiemann, T.; Ohira, D.; Arima, K.; Sawada, T.; Mataka, S.;
Marken, F.; Compton, R. G.; Bull, S.; Davies, S. G. J. Phys. Org. Chem.
2000, 13, 648-653.
10.1021/jo0490346 CCC: $27.50 © 2004 American Chemical Society
Published on Web 10/26/2004
J. Org. Chem. 2004, 69, 8177-8182
8177

Nag and Jenks
SCHEME 1. Relative Energetics for DBTO Excited
SCHEME 2. Compound Numbering and
Preparation
States17 and S-O Dissociation18
however, improved substrates are needed because the
chemical quantum yield for DBTO photolysis is less than
0.01.16
TABLE 1. Luminescence Quantum Yields
b
compd
Φfa
Φp
One rationalization of the low quantum yield is that
the deoxygenation proceeds by a mechanism in which
S-O bond stretching must be accompanied by intersys-
tem crossing at some point before complete bond scission.
Such a mechanism is consistent with the energetics of
the reaction for DBTO, as illustrated in Scheme 1. The
fluorescent singlet energy of DBTO is about 82 kcal
mol-1, and the phosphorescent triplet energy is about 61
DBT
1a
0.09c
0.018
0.045
0.036
0.057
1b
1c
1e
DBTO
2a
0.001
0.065
0.049
0.55
2b
2gd
kcal mol-1 17
. The S-O bond dissociation energy, forming
a Naphthalene in cyclohexane was used as the actinometer, with
excitation at 265 nm.23 Spectra were obtained in cyclohexane at
room temperature. Data were not obtained for 1f. Otherwise,
unlisted compounds did not detectably fluoresce. b Benzophenone
used as an actinometer with excitation at 265 nm.23 Spectra were
obtained at 77 K in EPA frozen organic glass. c Literature value.24
d Measured with 280 nm excitation to avoid a minimum in the
absorption spectrum.
O(3P), has been estimated to be 75-77 kcal mol-1 using
computational methods.18,19 These energetics imply that
DBTO must be on a path toward scission when isc occurs
and not ever reach the phosphorescent triplet, which
lacks sufficient energy for S-O dissociation. At the far
end of the mechanistic continuum suggested by this idea
is the notion that initial scission could result in an ion
pair of DBT•+ and O-, followed by back-electron transfer
to the ground state of DBT and O(3P).
Results
To the extent that one accepts the unimolecular
scission hypothesis and that O(3P) is the reactive inter-
mediate, two goals present themselves for optimization
of DBTO-like sources of oxygen atoms: (1) a red-shifting
of the absorption such that the extinction coefficient is
large at more convenient laser lines, such as 355 nm, and
(2) an increase in the efficiency of the reaction, so as to
generate a higher concentration of O(3P) with a given
pulse of light. In this paper, we address a series of
experiments aimed mainly at the second of these goals.
One reasonable strategy toward increasing the prob-
ability of all isc events is heavy atom substitution. Heavy
atom substitution facilitates spin-orbit coupling and,
hence, increases the probability (and the absolute rate
constants) for both radiative and nonradiative spin-
inverting processes.20,21 Here, we report how adding
halogen and acetyl substituents affects the photochem-
istry and emission of the DBTO nucleus.
Compounds. Substituted DBTs (1a-g, Scheme 2)
were prepared from DBT by lithiation and quenching
with an electrophile, or by electrophilic aromatic substi-
tution, as detailed in the Supporting Information. Oxida-
tion, using m-CPBA at moderately low temperature,
provided the corresponding DBTOs (2a-g) in satisfactory
unoptimized yields.22
Luminescence. Fluorescence spectra of the DBT
derivatives 1a-f at room temperature in cyclohexane
were very similar to that of DBT, save for the intensity.
Quantum yields are reported in Table 1. The acetyl
derivative 1g did not detectably fluoresce; neither did any
of the DBTO derivatives 2a-g.
At 77 K in EPA glass,25 DBTO has a weak phospho-
rescence.17 Similar, but somewhat more intense, spectra
were obtained from 2a and 2b, and approximate quan-
tum yields were measured with respect to benzophe-
(17) Jenks, W. S.; Lee, W.; Shutters, D. J. Phys. Chem. 1994, 98,
2282-2289.
(22) Compound 2b contained a small amount of an uncharacterized
impurity, thought to be an isomer or overhalogenated analogue due
to its NMR spectrum, as shown in the Supporting Information. We do
not believe this substantially affected any of the reported results.
(23) Eaton, D. F. In Handbook of Organic Photochemistry; Scaiano,
J. C., Ed.; CRC Publishers: Boca Raton, FL, 1989; Vol. I, pp 231-
239.
(24) Murov, S. L.; Carmichael, I.; Hug, G. L. Handbook of Photo-
chemistry, 2nd ed.; Marcel Dekker: New York, 1993.
(25) Ethyl ether, pentane, and (ethyl) alcohol, 5:5:2.
(18) Jenks, W. S.; Matsunaga, N.; Gordon, M. J. Org. Chem. 1996,
61, 1275-1283.
(19) This is 10-12 kcal/mol lower than the ordinary S-O BDE of a
sulfoxide because of the aromaticity of dibenzothiophene.
(20) Turro, N. J. Modern Molecular Photochemistry; Benjamin/
Cummings Publishing Co.: Menlo Park, 1978.
(21) Gilbert, A.; Baggott, P. J. Essentials of Molecular Photochem-
istry; CRC Press: Boca Raton, 1991.
8178 J. Org. Chem., Vol. 69, No. 24, 2004

Halogen-Substituted Dibenzothiophene Oxides
TABLE 2. Quantum Yields of Deoxygenation
SCHEME 3. Photochemistry and Quantum Yields
for Photolysis of 2a in Acetonitrile
a
compd
Φdeox
2a
0.0083b ( 0.0001c
0.0053 ( 0.0001
0.0045 ( 0.0006
0.0056 ( 0.0004
0.0093 ( 0.0009
0.0034 ( 0.0003
2b
2c
2d
2e
2f
2g
∼0
DBTO
0.0024d
0.0220
0.10
2b in benzene
2b in cyclohexene
2b in 1:9 cyclohexene/acetonitrile
0.010
a Data were measured by appearance of the sulfide. Solvent is
CH3CN, unless otherwise noted. The actinometer was acetophe-
none formation from valerophenone.28 b This is the apparent one-
photon yield for formation of 1a, but is a lower limit, due to the
efficient photodeiodination of this product. c Quoted error bars are
standard deviations. d Literature value.16
Because the quantum yield of deiodination of 1a is so
much larger than that of any of the other processes, the
appearance of DBT is approximately linear with time
during the early stages of the photolysis, and thus
apparent one-photon quantum yields are obtained for the
two true primary products (1a and DBTO) and the
secondary product DBT. The apparent quantum yields
for the primary products are lower than the actual values,
which could not be obtained. Because of the efficient
conversion of 1a to DBT, it can be concluded that the
actual primary quantum yield for dehalogenation of 2a
is substantially greater than that for deoxygenation.
To ensure the validity of the interpretations here,
photolyses were carried out through higher conversions.
To demonstrate the two-photon pathway to DBT, the
irradiations were done using different irradiation wave-
lengths, using RMR-3000 or RMR-3500 fluorescent tubes,
centered at about 300 or 350 nm, respectively, in a
Rayonet minireactor. The data are shown in Figure 1
with the abscissa normalized to conversion of 2a, rather
than time, for easier comparison of the two. Because of
the extended photolysis times required (25 and 37 h,
respectively), the photolyses were stopped when none of
the starting material remained. Nonetheless, high yields
(84% and 91%) of DBT were obtained. Because of the
relatively high initial concentration of 2a (ca. 4 mM),
some oxidized DBTO was also obtained. This product has
been previously observed from DBTO itself29 and has
been assigned the structure of the corresponding sultine
(sufinic ester).
The extinction coefficients for absorption in the 300-
320 nm range rank in the order 2a > DBTO > 1a >
DBT. Near 350 nm, the order is the same, but the
sulfoxides have much greater extinction coefficients than
the sulfides. Additionally, some discoloration of the
solutions occurs, consistent with the formation of a small
quantity of molecular iodine, but no effort was made to
document the colored material. The fraction of light
absorbed by each compound affects the time course of the
product mixture, as indicated in the figure. The essential
result, however, is the clear induction period associated
with formation of DBT, which indicates that it is a
secondary photoproduct. Also, the comparable initial
yields of DBTO and 1a at low conversion are reflective
none.23 The spectroscopic triplet energies do not vary
widely within the series. These weak spectra are char-
acterized by an onset near 450 nm and are easily
obscured by the much more phosphorescent correspond-
ing sulfides, whose spectra have an onset near 400 nm.26
The acetyl derivative 2g is highly phosphorescent at 77
K. However, its unstructured spectrum is considerably
different, with an onset near 370 nm.27
Photolyses. Photoreactions of sulfoxides 2a-g were
carried out in acetonitrile using initial concentrations
ranging from 1.5 to 4 mM. The absorption spectra have
maxima near 320 nm, so all compounds were excited at
that wavelength ((12 nm) using a Xe-arc lamp filtered
through a monochromator. Product analysis was done
when the reaction had reached no greater than 10%
conversion.
The acetyl derivative 2g was apparently inert to
photolysis, but the other compounds 2a-f provided the
corresponding deoxygenated sulfides. The quantum yields
of sulfide formation in acetonitrile were measured, rela-
tive to the Type II reaction of valerophenone, and are
reported in Table 2.28 The precision of the results, as
reflected in the standard deviations from multiple runs,
demonstrates that these differences, though modest, are
statistically significant.
In addition to the expected sulfide 1a, the iodosulfoxide
2a produced dehalogenated products, i.e., DBTO and
DBT. Such dehalogenations were not observed from any
other sulfoxide. Over the course of extensive irradiation
of 2a, the ratio of the products varied with conversion,
with DBT building up at the expense of the other two,
consistent with its formation being via secondary pho-
tolysis.
Quantum yield measurement is thus more compli-
cated for 2a, because of the two primary photochem-
ical reactions: deoxygenation and dehalogenation. Ad-
ditionally, both secondary reactions lead to DBT. How-
ever, the quantum yields of the secondary reactions could
be measured directly, and the data are presented in
Scheme 3.
(26) See, for example, ref 16.
(27) The phosphorescence spectrum of 1g is considerably more
structured and has a blue edge onset near 400 nm.
(28) Wagner, P. J.; Kelso, P. A.; Kemppainen, A. E.; McGrath, J.
M.; Schott, H. N.; Zepp, R. G. J. Am. Chem. Soc. 1972, 94, 7506-7512.
(29) Jenks, W. S.; Taylor, L. M.; Guo, Y.; Wan, Z. Tetrahedron Lett.
1994, 35, 7155-7158.
J. Org. Chem, Vol. 69, No. 24, 2004 8179

Nag and Jenks
quantum yields were also measured and are reported in
Table 2.
Discussion
The data presented here support the essential hypoth-
eses used in designing the experiments. The existence
of a heavy atom effect is demonstrated by the phospho-
rescence data, in which the iodosulfoxide and bromosul-
foxide have considerably greater quantum yields than
DBTO. In line with expectations, Cl substitution has
little heavy atom effect. Compounds 2c and 2f are not
phosphorescent, and their deoxygenation quantum yields
are only marginally higher than that of DBTO. Thus 2c
and 2f are good control compounds to show that the
larger effects in other compounds do not come from a
more mundane source, such as desymmetrizing of the
chromophore.
Greater phosphorescence yields can result either from
greater efficiency in formation of the triplet, or from a
greater fraction of triplets undergoing emissive decay, or
both. We cannot distinguish among these possibilities.
Nonetheless, there is some evidence (see below) for a
significant triplet yield for 2a. In any case, it is quite
reasonable to assign the greater phosphorescence quan-
tum yield to heavy-atom-assisted mechanisms. It should
also be noted that the fluorescence yields of the substi-
tuted DBTs are lower than that of DBT itselfsconsistent
with a heavy atom effectsthough the value for 1c is lower
than might be expected for a simple trend among the
halogenated compounds.
In a previous work,16 we reported that the use of
oxygen, isoprene, or cyclopentadiene as potential triplet
quenchers did not lower the quantum yield of DBTO
deoxygenation. This showed that deoxygenation does not
proceed through a typically long-lived triplet state of the
sulfoxide. The facts that triplet quenchers do not lower
the efficiency of deoxygenation and the triplet excitation
energy is below the energy required to carry out S-O
scission imply that a higher yield of the spectroscopic
triplet would not be inherently favorable for deoxygen-
ation yields. Thus, the classic increase of the intersystem
crossing rate in arenes by heavy atoms could be coun-
terproductive. Nonetheless, small but reproducible and
significant increases in deoxygenation are obtained with
the iodosulfoxide 2a and the bromosulfoxides 2b, 2d, and
2e, relative to DBTO and the chlorinated sulfoxides.
Especially taking into account that it is likely that the
triplet yields of 2a, 2b, 2d, and 2e are higher than that
of DBTO, we assert that the increased quantum yields
for deoxygenation in this series of compounds are con-
sistent with the idea that S-O scission is coupled with
intersystem crossing by way of an excited-state surface
crossing to a dissociative triplet state. Similar mecha-
nistic assertions have been discussed, for example, in the
extensive published work on the photodissociation of aryl
halides.30 Recent computational papers by Liu and co-
workers derives explicit potential energy curves for
halobenzenes and discuss the surface crossings promoted
by halogen-induced spin-orbit coupling in detail.31,32
FIGURE 1. Products observed on photolysis of 2a in aceto-
nitrile, with initial concentration of 4.0 mM. The abscissas are
normalized to percentage decomposition of 2a. The lines are
smooth fits to help draw the eye and of no physical significance.
Top: excitation from 300 nm fluorescent tubes. Total time was
1500 min. Bottom: excitation from 350 nm fluorescent tubes.
Total time was 2200 min.
of the comparable quantum yields of their formation
reported in Scheme 3.
The photochemistry of 2b was examined in benzene
and cyclohexene to verify that the same type of interme-
diate was formed as in photolysis of DBTO. Product
studies carried out by the methods previously re-
ported15,16 showed the same products of solvent oxidation,
in nearly identical yield, were obtained using either
DBTO or 2b. The major oxidized product from benzene
is phenol, while cyclohexene yields both cyclohexenol and
cyclohexene oxide in very similar amounts.15,16 The
(30) Leading references for this literature are available in refs 31
and 32.
(31) Liu, Y.-J.; Persson, P.; Karlsson, H. O.; Lunnell, S.; Kadi, M.;
Karlsson, D.; Davidsson, J. J. Chem. Phys. 2004, 120, 6502-6509.
8180 J. Org. Chem., Vol. 69, No. 24, 2004

Halogen-Substituted Dibenzothiophene Oxides
SCHEME 4. Model To Explain Solvent Dependence
on Quantum Yields Based on Competition between
Reaction with Solvent and DBT
Another mechanistic extreme that cannot be elimi-
nated is the photodissociation to an ion pair, presumably
DBT•+/O•-, followed by back-electron transfer that leads
to DBT and ground-state O(3P). This back-electron-
transfer pathway would naturally compete with simple
recombination and thus could also result in a low
quantum yield. If the ion pair is born in a singlet state,
then heavy atom substitution on the DBT moiety might
facilitate the back-electron-transfer pathway and raise
the quantum yield of O(3P) formation.
The three bromo-substituted compounds were inves-
tigated to determine empirically whether a higher quan-
tum yield could be obtained by moving the substituent
or adding a second one. From the practical perspective
of trying to use these compounds in independent studies
of oxygen atom chemistry, the differences are not espe-
cially significant, though adding a second bromine does
enhance reactivity.
olefins.16,35 As a result, the observed quantum yields are
higher in cyclohexane than acetonitrile.
The strong luminescence and lack of deoxygenation
from 2g can be rationalized in at least two ways. First,
the data are consistent with a case in which intersystem
crossing to the luminescent triplet is very efficient and
in which the photoreactivity of that state is very low. The
blue-shifted emission spectrum of 2g, relative to the other
emissive sulfoxides, suggests that the triplet energy may
be on par with or just below the S-O bond energy, but
the carbonyl group may perturb the nature of the state
more than the halogens. The lack of structure in the
spectrum suggests it is not a typical localized nπ* state
of aromatic ketones such as benzophenone, but assign-
ment beyond that is speculative. There is not a high
correlation between phosphorescence yields and reactiv-
ity in another prototypical reaction of sulfoxides, photo-
induced stereomutation, but the phosphorescence yields
for compounds whose racemization have also been stud-
ied are all lower than for 2g.36-38 A second rationalization,
also consistent with all the data, is that the introduction
of the acetyl substituent and its attendant electronic
influences directly perturbs the coupling between ISC
and deoxygenation.
An alternative heavy atom strategy, to be reported
separately, is the use of selenoxides instead of sulfoxides
in a DBTO-like molecule. In this case, the perturbation
to the system is inherently greater because of changes
in bond strength, bond lengths, aromaticity of the re-
duced compound, etc. However, the central location of the
heavy atom at the key atomic position involved in the
S-O (or Se-O) cleavage may turn out to be a more
dramatically successful strategy, in terms of producing
a high quantum yield O-atom donor.
The initial attractiveness of the higher quantum yield
for iodinated sulfoxide 2a is quickly quashed by the
complications involved with the dehalogenation reaction.
The fact that only it and its corresponding sulfide 1a
undergo the photodehalogenation reaction suggests that
this is an orthodox homolysis reaction, which is usually
assigned to triplet-state chemistry. (See, for example, refs
33 and 34.) The bond energy for aryl iodides is of the
order of 64 kcal mol-1 33,34 That is just a few kcal mol-1
.
higher than the energy of the spectroscopic triplet of 2a,
probably within reach at room temperature, given the
moderate quantum yield. The triplet energies of the DBT
derivatives is about 70 kcal mol-1 17
which is above the
,
energy required for C-I homolysis and presumably
contributes to the higher quantum yield for deiodination
of 1a.
On the other hand, arene-Br bond energies are near
70 kcal mol-1 33,34
With triplet energies of the order of 60
.
kcal mol-1, it is not surprising that 2b, 2d, and 2e do
not debrominate competitively with deoxygenation. On
the other hand, exhaustive photolysis might have led to
dehalogenation of the corresponding sulfides, given their
higher triplet energies.
The solvent dependence on the quantum yield observed
for 2b is consistent with our previous observations for
DBTO,16 though with values all somewhat larger than
for the parent. We interpret the solvent dependence in
terms of a model suggested in Scheme 4.
On scission, O(3P) is formed in the immediate vicinity
of DBT. As a result, the reactive intermediate may
immediately react with the sulfide to reoxidize it, diffuse
away from the sulfide before reacting with anything, or
react with solvent before escaping from the cage. To the
extent that one solvent reacts with O(3P) faster than
another one, the observed quantum yield will be higher.
Acetonitrile is not a particularly active substrate for
reactions with O(3P).35 More reactive molecules include
halide ions, and those with oxidizable sulfur atoms or
Conclusions
Halo-substituted dibenzothiophenes show a modest
improvement in the quantum yield for deoxygenation,
relative to the parent DBTO. The trend order of iodo >
bromo > chloro ∼ H allows assignment of this to a heavy
atom effect. This is consistent with a proposed mecha-
nism of unimolecular S-O scission in which bond stretch-
ing is coupled to intersystem crossing, presumably into
the T0 substate. From a quantitative point of view, the
effect is smaller than that which would be optimum for
(32) Liu, Y.-J.; Persson, P.; Lunnell, S. J. Phys. Chem. A 2004, 108,
2339-2345.
(33) Grimshaw, J.; de Silva, A. P. Chem. Soc. Rev. 1981, 10, 181-
203.
(34) Schutt, L.; Bunce, N. J. In CRC Handbook of Organic Photo-
chemistry and Photobiology, 2nd ed.; Horspool, W. M., Lenci, F., Eds.;
CRC Press: Boca Raton, 2004; pp 38-1-38-18.
(35) Bu¨cher, G.; Scaiano, J. C. J. Phys. Chem. 1994, 98, 12471-
12473.
(36) Lee, W.; Jenks, W. S. J. Org. Chem. 2001, 66, 474-480.
(37) Tsurutani, Y.; Yamashita, T.; Horie, K. Polymer J. 1998, 30,
11-16.
(38) Tsurutani, Y.; Machida, S.; Horie, K.; Kawashima, Y.; Nakano,
H.; Hirao, K. J. Photochem. Photobiol. A 1999, 122, 161-168.
J. Org. Chem, Vol. 69, No. 24, 2004 8181

Nag and Jenks
demanding mechanistic studies of O(3P) solution-phase
chemistry that require substantial concentrations of
intermediates, such as flash photolysis.
Preliminary photoreactions and some product analysis
studies were done using a fan-cooled Rayonet minireactor at
room temperature using broadly emitting 300 nm fluorescent
tubes. The same reactor was used for the data shown in Figure
1, using both the 300 nm bulbs (RMR-3000) and 350 nm bulbs
(RMR-3500).
Experimental Section
Emission Spectra. Phosphorescence spectra were recorded
at 77 K, as reported previously.17 The samples were contained
in 5 mm Suprasil cylindrical tubes within the Dewar. Fluo-
rescence data were collected at ambient temperature with Ar-
flushed samples that had an optical density of about 0.1 at
the excitation wavelength, usually 265 nm.
Photoreactions. All solvents were “spectro grade” or the
equivalent and deoxygenated by sparging with argon bubbles
for 10 min prior to photolysis. Cyclohexene was refluxed over
Na under an Ar atmosphere and distilled immediately before
use.
The initial concentrations of all photolyses were in the range
of 1.0 to 4.0 mM. Prior to photolysis, the samples were checked
for purity using a HP 5890 II gas chromatograph equipped
with ZB-5 capillary columns and a flame ionization detector.
Dodecane was used as internal standard for all photoreactions.
Valerophenone was used as actinometer for quantum yield
measurements,28 and samples were irradiated in 1 cm square
cells.
Acknowledgment. The financial support of the
National Science Foundation (CHE 0213375) is grate-
fully acknowledged. We thank Dr. Jacob Petrich for the
use of the SPEX Fluoromax instrument for fluorescence
spectra.
Quantum yields were measured using a 75 W Xe arc lamp
fitted to a monochromator set to 320 nm. The full output of
the monochromator was absorbed by the samples, whose
optical density at the excitation wavelength exceeded 2.
Actinometry was repeated frequently to avoid any effect on
quantum yield measurement due to any long-term drift in the
light flux.
Supporting Information Available: Compound prepara-
tion, emission, and NMR spectra. This material is available
free of charge via the Internet at http://pubs.acs.org.
JO0490346
8182 J. Org. Chem., Vol. 69, No. 24, 2004