Triplet energy distribution in photoinitiators containing two dissociable groups

Article abstract of DOI:10.1021/jo0202278

Suitable probe molecules containing the benzophenone chromophore and one dissociable bond [perester 1 or C-X (4, 5)] or two such bonds (2, 3) have been synthesized and studied to examine intramolecular triplet energy dispersion. Triplet energies and phosphorescence quantum yields as well as quantum efficiencies of bond scissions have been studied, and the flow of triplet energy in such molecules is discussed. Upon irradiation at 350 nm in either benzene or methanol, the target peresters undergo dissociation of both cleavable groups, producing a pair of radicals. The presence of a benzylic chloride function has little influence on the efficiency of perester dissociation. However, the presence of a benzylic bromide function was found to decrease the quantum yield of decomposition of the perester function of 3. This can be explained by taking into account the effect of the heavy atom and the rate of cage geminate radical pair recombination. The nature of the heavy atom perturbation, however, was found to be different in 5 as compared with 3.

Full text of DOI:10.1021/jo0202278

Tr ip let En er gy Distr ibu tion in P h otoin itia tor s Con ta in in g Tw o
Dissocia ble Gr ou p s
Bipin K. Shah and Douglas C. Neckers*
Center for Photochemical Sciences, Bowling Green State University, Bowling Green, OH 434031
neckers@photo.bgsu.edu
Received April 2, 2002
Suitable probe molecules containing the benzophenone chromophore and one dissociable bond
[perester 1 or C-X (4, 5)] or two such bonds (2, 3) have been synthesized and studied to examine
intramolecular triplet energy dispersion. Triplet energies and phosphorescence quantum yields as
well as quantum efficiencies of bond scissions have been studied, and the flow of triplet energy in
such molecules is discussed. Upon irradiation at 350 nm in either benzene or methanol, the target
peresters undergo dissociation of both cleavable groups, producing a pair of radicals. The presence
of a benzylic chloride function has little influence on the efficiency of perester dissociation. However,
the presence of a benzylic bromide function was found to decrease the quantum yield of
decomposition of the perester function of 3. This can be explained by taking into account the effect
of the heavy atom and the rate of cage geminate radical pair recombination. The nature of the
heavy atom perturbation, however, was found to be different in 5 as compared with 3.
CHART 1. P h otod issocia ble P er ester s a n d
Su bstitu ted Ben zop h en on es.
In tr od u ction
Benzophenone is a widely used triplet sensitizer. Its n
f π* transition produces the triplet state exclusively due
to the unit quantum yield of the S1 f T1 intersystem
crossing (ISC).2 The triplet energy of benzophenone can
be transferred, though inefficiently, to dialkyl peroxides
to cause homolytic scission of the peroxy bond.3 In
contrast, such triplet energy transfer may be made highly
efficient by linking the perester bond directly to the
benzophenone chromophore. This is clearly reflected by
much higher rates and quantum yields of decomposition
of benzophenone peresters.4
The lifetime of the triplet state of benzophenone in neat
di-tert-butyl peroxide at 25 °C is 60 ns.5 This is about 80
times longer than the lifetime of the triplet state of tert-
butyl p-benzoylperbenzoate.6 In fact, the quantum ef-
ficiency of photodecomposition of the tert-butyl aroylper-
benzoates was found to be at least three times that of
benzophenone-sensitized decomposition of benzoyl per-
oxide, clearly indicating the efficacy of the intramolecular
process in contrast to bimolecular energy transfer.4
Intramolecular triplet energy transfer from the ben-
zophenone chromophore can also enhance dissociation of
the C-Cl and C-Br bonds.7 However, the rates and
quantum yields of photolytic C-X bond dissociation in
appropriately substituted benzophenones have not been
reported. The free radicals generated following intramo-
lecular energy dispersion have been shown to have
application in polymerization7,8 as well as in photo-cross-
linking and photografting.9
Though the observations are relatively old, a detailed,
systematic investigation of the factors affecting this
process has not been provided. We synthesized suitably
designed probe molecules in order to understand the
mechanism and chemical consequences of the intramo-
lecular dispersion of the benzophenone triplet energy,
especially when two cleavable groups are available in the
same molecule. We selected the C-X and perester bonds
which have different homolytic bond dissociation ener-
gies. The compounds studied (Chart 1) are tert-butyl 4-(4′-
* To whom correspondence should be addressed. Telephone: (419)
372-2034. Fax: (419) 372-0366.
(1) Contribution no. 467 from the Center for Photochemical Sciences.
This paper is dedicated to Prof. Dr. J . W. Neckers on the occassion of
his 100th birthday.
(2) Scaiano, J . C. Handbook of Organic Photochemistry; CRC
Press: Boca Raton, FL, 1989; Vol. I, p 378.
(3) Walling, C.; Gibian, M. J . J . Am. Chem. Soc. 1965, 87, 3413.
(4) Thijs, L.; Gupta, S. N.; Neckers, D. C. J . Org. Chem. 1979, 44,
4123.
(5) Scaiano, J . S.; Wubbels, G. G. J . Am. Chem. Soc. 1981, 103, 640.
(6) Morlino, E. A.; Bohorquez, M. D.; Neckers, D. C.; Rodgers, M.
A. J . J . Am. Chem. Soc. 1991, 113, 3599.
(7) (a) Pacifici, J . G.; Wang, R. H. S.; Newland, G. C. U.S. Patent
4,043,887, 1977. (b) Krefeld, H. R.; Krefeld-Bockum, H. R.; Krefeld,
H. U.S. Patent 3,686,084, 1972.
(8) Gupta, S. N.; Gupta, I.; Neckers, D. C. J . Polym. Sci., Polym.
Chem. Ed. 1981, 19, 103.
(9) Gupta, I.; Gupta, S. N.; Neckers, D. C. J . Polym. Sci., Polym.
Chem. Ed. 1982, 20, 147.
10.1021/jo0202278 CCC: $22.00 © 2002 American Chemical Society
Published on Web 07/27/2002
J . Org. Chem. 2002, 67, 6117-6123
6117

Shah and Neckers
TABLE 1. Tr ip let En er gy of P er ester sa
perester
triplet energy (kcal/mol)
1
2
3
77.0
77.2
77.5
a
Solvent ) EPA; temperature at 77 K; excitation at 350 nm.
flexible. Triplet energies determined from the 0-0 band
of the phosphorescence spectra are given in Table 1. The
triplet energy of benzophenone is 68 kcal/mol,10 while
those of the peresters are higher (77 kcal/mol). Thus, the
presence of a tert-butyl perester side chain causes an
increment of about 9 kcal/mol in the triplet energy.
Each perester undergoes homolytic bond dissociation,
forming radical products upon irradiation at 350 nm in
benzene (Table 2). The thermal and primary photolytic
products were similar for 1 and 2 except that the
corresponding acids and phenyl esters were also obtained
during photolysis. The products of the thermolysis of 3
do not include those formed by cleavage of the C-Br
bond. On the other hand, photolysis of this perester yields
mainly those products in which the Br atom is no longer
retained. This is because the flow of the triplet energy
takes place in both directions in the molecule, causing
scission of each dissociable group. Thus, only the perester
bond dissociates during thermolysis, while both the C-Br
and perester bonds dissociate during photolysis.
F IGURE 1. UV-vis absorption spectra of peresters. Concn:
1 ) 6.88 × 10-3 M, 2 ) 6.96 × 10-3 M, and 3 ) 7.18 × 10-3 M.
This is also true for 2, in which dissociation of both
the C-Cl and perester bonds was observed. Products
from cleavage of the C-Cl bond were found from NMR
studies. We observed only one product derived from
dissociation of the C-Cl bond in 2 as a primary photo-
product (during GC/MS analysis) when photolysis was
carried out in benzene. Most of the products observed
after 2 min of irradiation retained the Cl atom. Never-
theless, dissociation of the C-Cl bond was clearly ob-
served when photolysis was carried out in methanol.
Those photoproducts that formed because of coupling of
F IGURE 2. Phosphorescence spectra of peresters. Excitation
wavelength ) 350 nm.
methylbenzoyl)perbenzoate (1), tert-butyl 4-(4′-chloro-
methylbenzoyl)perbenzoate(2),tert-butyl4-(4′-bromomethyl-
benzoyl)perbenzoate (3), 4-(chloromethyl)benzophenone
(4), and 4-(bromomethyl)benzophenone (5). In 2 and 3,
the C-X and perester bonds are at the remote para
positions of the benzophenone chromophore. Similar
compounds with only one dissociable group [either per-
ester 1 or C-X (4, 5)] were studied for comparison. The
rates, quantum efficiencies, and a mechanism of decom-
position of these peresters are reported herein. The
phosphorescence quantum yields are presented, and the
intramolecular flow of the triplet energy is discussed.
•
the CH2-OH radical (formed after H-abstraction from
•
the methanol molecule) with the CH2-Ar radicals (gen-
erated by dissociation of the C-X bond) were observed
from both 2 and 3.
Mech a n ism . The mechanism of photolysis of simple
peresters is well established.4 A similar mechanism is
proposed for the photodecomposition of peresters that
contain two dissociable groups (Scheme 1). The mecha-
nism is based on the analysis of the primary photolysis
products obtained within 1-2 min of irradiation at 350
nm in benzene. The homolytic bond dissociation energies
of benzylic C-Br and C-Cl bonds are 55 and 59 kcal/
mol, respectively, while that of a perester bond is 30 kcal/
mol.10 It is unlikely that dissociation of both the C-X and
perester bonds occurs simultaneously in 2 or 3, given that
the triplet energy of these molecules is about 77 kcal/
mol. The possibility of a one-photon-two-bond-dissocia-
tion mechanism can be discarded on energy grounds.
Resu lts a n d Discu ssion
Sp ectr a a n d P r od u ct An a lysis. UV-vis spectra
(Figure 1) of peresters 1-3 show the characteristic
absorption of the benzophenone chromophore with an
absorption corresponding to the n f π* transition of the
carbonyl group with a maximum at 346-348 nm. The
ꢀ350 values in benzene are found to be 197, 175, and 188
M-1 cm-1 for 1, 2, and 3, respectively. Phosphorescence
spectra of the peresters are relatively structureless
(Figure 2) as well as broader and blue shifted compared
to that of benzophenone. This may be due to the presence
of the perester side chain that renders the molecule more
Dissociation of the perester bond is the only photo-
chemical option in 1. Thus, rapid intramolecular triplet
energy transfer leads to scission of that bond. The formed
(10) Murov, S. L.; Carmichael, I.; Hug, G. L. Handbook of Photo-
chemistry; Marcel Dekker: New York, 1993.
6118 J . Org. Chem., Vol. 67, No. 17, 2002

Triplet Energy Distribution in Photoinitiators
TABLE 2. Th er m a l a n d P r im a r y P h otop r od u cts
perester
thermal productsa
photoproductsa
photoproductsb
1
CH3-Ph-CO-Ph
CH3-Ph-CO-Ph
CH3-Ph-CO-Ph
CH3-Ph-CO-Ph-Ph
CH3-Ph-CO-Ph-Ph
CH3-Ph-CO-Ph-CO-OCH3
CH3-Ph-CO-Ph-COOH
CH3-Ph-CO-Ph-COOH
CH3-Ph-CO-Ph-CO-OPh
CH3-Ph-CO-Ph
2
3
ClCH2-Ph-CO-Ph
ClCH2-Ph-CO-Ph-Ph
ClCH2-Ph-CO-Ph
ClCH2-Ph-CO-Ph
CH3-Ph-CO-Ph-CH2OH
ClCH2-Ph-CO-Ph-COOH
HO-CH2-CH2-Ph-CO-Ph
HO-CH2-CH2-Ph-CO-Ph-CO-OCH3
CH3-Ph-CO-Ph
ClCH2-Ph-CO-Ph-Ph
ClCH2-Ph-CO-Ph-COOH
ClCH2-Ph-CO-Ph-CO-OPh
CH3-Ph-CO-Ph
BrCH2-Ph-CO-Ph
BrCH2-Ph-CO-Ph-Ph
CH3-Ph-CO-Ph-Ph
CH3-Ph-CO-Ph-CH2OH
CH3-Ph-CO-Ph-COOH
HO-CH2-CH2-Ph-CO-Ph
HO-CH2-CH2-Ph-CO-Ph-CO-OCH3
CH3-Ph-CO-Ph-COOH
BrCH2-Ph-CO-Ph-Br
CH3-Ph-CO-Ph-COOOCMe3
a
Solvent ) benzene; during photolysis, biphenyl, phenyl tert-butyl ether, acetone, and tert-butyl alcohol were also observed in all
b
cases. Solvent ) methanol.
SCHEME 1. Mech a n ism of P h otod issocia tion of P er ester s
4-(4′-methylbenzoyl)benzoyloxy radical may abstract an
H atom or combine with solvent to yield the correspond-
ing carboxylic acid or phenyl ester. Dissociation of carbon
dioxide from the 4-(4′-methylbenzoyl)benzoyloxy radical
results in the formation of the corresponding aryl radical,
which, in turn, may abstract an H atom or combine with
solvent. Since benzene is not a good H donor, it might be
possible that these radicals abstract an H atom from the
intermediate radical that forms when the aroyloxyl and
aryl radicals trap benzene, as depicted in the mechanism.
The trapping of the aroyloxyl and aryl radicals by
benzene is well established.11
Compound 3 produces the usual radical derived prod-
ucts due to cleavage of the perester bond. In addition, it
also forms 1 in an observable amount. Evidence for its
formation was obtained by NMR studies.12 The formation
of 1 during the photolysis of 3 can take place when the
C-Br bond dissociates prior to the perester bond. How-
ever, given the higher bond dissociation energy of a
benzylic C-Br bond relative to a perester bond, the
preference for dispersion of the triplet energy should be
toward the perester moiety in 3. This does not seem to
be the case. Formation of 1 as a major photolytic product
suggests that dissociation of the C-Br bond is favored.
This is further supported by the observed quantum
efficiency data described below.
4-(Bromomethyl)-4′-bromobenzophenone was observed
as one of the products during irradiation of 3. Thus, it is
(12) The chemical shift of the methyl peak of 1 in C6D6 is 1.997 ppm.
During the irradiation of 1, we observed disappearance of the peak at
1.997 ppm and appearance of a new peak at 2.020 ppm. The methyl
peak of 4-methylbenzophenone appears at 1.991 ppm. Thus, the peak
at 2.020 ppm may correspond to the methyl group of the solvent
(benzene) incorporated product, which is one of the major primary
photolytic products in all cases. During the irradiation of 3, we recorded
the appearance of two peaks exactly at 2.022 and 1.997 ppm. The peak
at 1.997 ppm corresponds to the methyl group of 1, a primary photolytic
product in this case. We observed no peak at 1.997 ppm during the
irradiation of 2. (See the Supporting Information.)
(11) Chateauneuf, J .; Lusztyk, J .; Ingold, K. U. J . Am. Chem. Soc.
1988, 110, 2886.
J . Org. Chem, Vol. 67, No. 17, 2002 6119

Shah and Neckers
TABLE 3. Qu a n tu m Yield s of Bon d Dissocia tion a
TABLE 4. Qu a n tu m Yield s of Bon d Dissocia tion in
Meth a n ol
φdiss
φdiss
compound
-O-O-
C-X
perestera
-O-O-
C-X
1
2
3
4
5
0.82
0.63
0.22
0.26
0.41
0.61
0.37
1
2
3
1.44
1.02
0.16
0.23
0.45
Concn ) 7.5 × 10-3 Μ.
a
(a) Intensity (I0) ) 9.2 × 1016 quanta L-1 s-1; (b) solvent )
a
C6D6; (c) concn ) 8.3 × 10-3 M.
been previously presented.7 The kobs of the C-Cl bond
dissociation in 4 was found to be 4.8 × 10-5 M s-1. This
is similar to the kobs (4.6 × 10-5 M s-1) of the C-Cl bond
dissociation of 2. Similarly, the kobs values of the C-Br
bond dissociations of 3 (2.7 × 10-5 M s-1) and 5 (1.9 ×
10-5 M s-1) were also found to be nearly the same. The
φdiss values which reflect the fraction of the triplet energy
utilized in dissociating the C-X bond are, as expected,
different for these compounds. Thus, the similarity in the
rate constants of the corresponding C-X bond dissocia-
tion may be a strong indication that these are the
individual rates.
also possible that the corresponding aryl radical formed
during decomposition of 3 abstracts a Br atom from the
starting material, forming the corresponding benzyl
radical and 4-(bromomethyl)-4′-bromobenzophenone. How-
ever, such an abstraction was not observed in the case
of 2. If it were one of the pathways of decomposition, an
abstraction of a Cl atom from 2 would also have been
observed.
The situation is different in the case of 2, where the
formation of 1 was not observed. Retention of the C-Cl
bond in the majority of the photoproducts indicates that
cleavage of the perester bond rather than the C-Cl bond
is a favored path. The bond dissociation energy of a
benzylic C-Cl bond is lower than the triplet energy of 2.
Despite this, the triplet energy prefers to disperse in the
direction of the perester moiety. This is not surprising,
given that the perester bond is much weaker than the
C-Cl bond.
Ra tes a n d Qu a n tu m Yield s. The quantum efficiency
of dissociation (φdiss) values of the C-X and perester
bonds were obtained by 1H NMR (Table 3). The quantum
yield of the decomposition of 1 in C6D6 was found to be
0.82, in the same range of quantum yields previously
found for decomposition of similar peresters in benzene.4
There is little difference in the φdiss values of the dis-
sociation of the perester bonds in 1 and 2. However, there
is a sharp drop in the φdiss values of the perester bonds
from 1 to 3 (0.82 to 0.22). Thus, in the case of 3, a larger
fraction of the triplet energy finds other ways to disperse
rather than toward the perester bond. This is in complete
agreement with the observed photoproducts. The φdiss
(0.41) of the C-Br bond is nearly twice the φdiss (0.22) of
the perester bond in 3, indicating the favorable cleavage
of the C-Br bond over the perester bond. The opposite
pertains to 2, in which the φdiss (0.26) of the C-Cl bond
was found to be less than half of the φdiss (0.63) of the
perester bond.
In d u ced Decom p osition . The peresters show similar
UV-vis absorptions in methanol. The observed photo-
products in methanol (Table 2) are normal radical
products that are formed, following the general mecha-
nism of homolytic dissociation of the C-X and perester
•
bonds. The CH2-OH radical was found to couple with
the corresponding phenyl radicals (formed after perester
bond dissociation followed by decarbonylation) in 1, 2,
•
and 3. Coupling of the CH2-OH radical with the corre-
sponding benzyl radicals was also observed for 2 and 3.
For simple peresters,4 a significant increase in the
quantum yield of decomposition has been found in
methanol. To test the possibility of such solvent-assisted
decomposition, we measured the φdiss in CD3OD (Table
4) and observed induced decompositions of 1 and 2.
The φdiss of the perester bond in 1 and 2 was found to
be much higher than in benzene and higher than unity.
The φdiss of the C-Cl bond, however, remained the same
in 2, indicating the solvent-induced dissociation of only
the perester bond. This may be because the triplet energy
favors dispersion toward the perester bonds in these
molecules. There was no such a solvent effect on the
decomposition of 3. The φdiss values of both the C-Br and
perester bonds were almost the same either in C6D6 or
in CD3OD. This may be because of the favoribility of the
triplet energy dispersion toward the C-Br bond, dis-
sociation of which is not induced by the solvent.
We also measured and compared the rates and φdiss
values of 4 and 5 with those of 2 and 3.13 Cleavage of the
C-X bond is the only possible photochemical dissociation
process in 4 and 5. To our knowledge, the rates and φdiss
values for the photolysis of these compounds are not
reported, although rates of polymerization of certain
monomers using these compounds as photoinitiators have
Unlike photodecomposition of the peresters in benzene
in which formation of both acetone and tert-butyl alcohol
was observed during NMR studies, formation of only tert-
butyl alcohol was observed in methanol. The correspond-
ing benzoyloxy and aryl as well as tert-butyl radicals
•
easily abstract hydrogen from methanol. The CH2-OH
radical can induce the decomposition of the peresters, as
shown in the Scheme 2.
(13) It should be noted that, in 2 or 3, when any one of the C-X or
perester bonds dissociates during irradiation, the overall concentration
of the starting material changes. Thus, it appears that the higher
observed rate and the φdiss may be the limiting rate and quantum yield
of the dissociation of each of these peresters. Even if this is the case,
we clearly see that dissociation of the C-Br bond is the favorable path
of the triplet energy dispersion in 3. The opposite is true in 2. However,
given the nature of appearance and disappearance of the peaks
observed during NMR studies and the nature of the data, we believe
that these data reflect rates and φdiss values of the individual bonds.
An analysis of the φdiss (either in benzene or methanol)
demonstrates that the decomposition of 3 is not as
efficient as that of 1 or 2. After 5 min of irradiation in
C6D6, the decomposition of 1 and 2 was almost complete,
as indicated by the disappearance of the peak belonging
to the tert-butyl group (see the Supporting Information).
However, a considerable amount of 3 was not decom-
6120 J . Org. Chem., Vol. 67, No. 17, 2002

Triplet Energy Distribution in Photoinitiators
SCHEME 2. P ossible Mech a n ism of In d u ced Decom p osition
SCHEME 3. Rea son of Low Qu a n tu m Yield of
Dissocia tion
posed, even after 5 min of irradiation. The slow decom-
position is also reflected in rates of photopolymerization
which were found to be much lower for 3 than for other
simple peresters when they were used as photoinitia-
tors.14
The more facile dissociation of the C-Br bond into a
stable benzyl radical was previously considered to be the
reason.14 For this, the rate of dissociation of the C-Br
bond has to be considerably higher than that of the
perester bond in the same molecule. However, the
observed rate constants of the C-Br bond (2.7 × 10-5
M
s-1) and the perester bond (2.0 × 10-5 M s-1) were found
to be nearly the same in 3. The bimolecular termination
rate constant for benzyl radicals in benzene at 25 °C was
coupling is well established.16 A heavy atom may enhance
the depopulation of the triplet state by inducing the
radiationless T1 f S0 decay and, hence, may also decrease
the phosphorescence quantum yield (φPh). Direct spec-
troscopic observation of a decrease in φPh values has been
made in iodo and bromo polymethine dyes at 77 K.17
In the case of peresters, a triplet radical pair is
selectively generated from a triplet precursor. Typically,
triplet-singlet ISC in a radical pair takes ∼10-8 s.18 The
time scale for the formation of secondary cages is on the
order of 10-8 s in nonviscous (n < 10 cP) homogeneous
solutions at room temperature.18,19 Thus, there is a
probability that the spin correlation of RPs (RP ) radical
pair) can change in the secondary cages. In such a
situation, a competition may arise between ISC of the
also reported to be 1.8 × 10-9 M s-1 15
. Thus, the reasoning
presented previously does not seem sufficient.
It is also noted that the φdiss (0.37) of the C-Br bond
of 5 in benzene is much lower than the φdiss (0.61) of the
C-Cl bond of 4. The excited states of these molecules
have sufficient energy to easily cleave either bond. Even
if we consider the bond dissociation energy, the opposite
should have been the case because the C-Cl bond is
stronger than the C-Br bond. However, the C-Br bond
seems to cause a decrease in the φdiss values in both 3
and 5. The reason for such a low φdiss of the C-Br bond
can be explained by taking into account the heavy atom
effect and the cage radical pair recombination.
Rea son for Low φd iss of 3 a n d 5. We propose that
there are at least two possible ways of energy dissipation
that can result in the observed low φdiss values of 3 and
5 (Scheme 3). The presence of a heavy atom in the
molecule may have an effect on the triplet state as well
as on the radical pairs that form following triplet energy
dispersion. The effect of a heavy atom in enhancing the
nonradiative decay of the triplet states by spin-orbit
3
1
type RP f RP within the solvent cage and the escape
of the RP from the cage. The heavy atom spin-orbit
interaction can have a dramatic effect on such a competi-
tion between diffusion and ISC. In many cases, an
external20 or internal21heavy atom has been found to
significantly enhance the rates of ISC in the primary and
(16) (a) Inoue, H.; Sakurai, T.; Hoshi, T. Bull. Chem. Soc. J pn. 1991,
64, 3340. (b) Cowan, D. O.; Koziar, J . C. J . Am. Chem. Soc. 1974, 96,
1229.
(17) Darmanyan, A. P. Chem. Phys. Lett. 1983, 102, 329.
(18) Koenig, T.; Fisher, H. Free Radicals; Kochi, J . K., Ed.; Inter-
science: New York, 1973; p 157.
(14) Abu-Abdoun, I. I.; Thijs, L.; Neckers, D. C. J . Polym. Sci., Polym.
Chem. Ed. 1983, 21, 3129.
(15) Burkhart, R. D. J . Phys. Chem. 1969, 73, 2703.
(19) Turro, N. J .; Weed, G. C. J . Am. Chem. Soc. 1983, 105, 1861.
J . Org. Chem, Vol. 67, No. 17, 2002 6121

Shah and Neckers
TABLE 5. P h osp h or escen ce Qu a n tu m Yield s
grade silica gel (32-63 µm, 60 Å) and silica gel plates (200
µm). Melting points are uncorrected. Elemental analyses were
performed at Atlantic Microlab, Inc., Atlanta, GA. Nuclear
magnetic resonance (NMR) spectra were recorded on a 200
MHz spectrometer. CDCl3 was the solvent for NMR unless
otherwise noted. Chemical shifts relative to TMS at 0.0 ppm
are reported in parts per million (ppm) for 1H NMR on the δ
scale. UV spectra were obtained on a UV-vis diode array
spectrometer. GC/MS data were collected using spectrometers
having 30 m × 0.25 mm × 0.25 µm columns. A phosphorimeter
coupled to a 1680 0.22m double spectrometer was used to
record phosphorescence spectra.
compound
φPh
1
2
3
4
5
0.055
0.051
0.069
0.566
0.025
secondary cage pairs. Experimental evidence of a large
influence of even a single solvent atom on the photodis-
sociation process in which a bromine atom itself is
involved has also been obtained.22
Syn th esis. The peresters were synthesized following a
literature procedure.4
(A) ter t-Bu tyl 4-(4′-Meth ylben zoyl)p er ben zoa te (1). 1,4-
Dimethyl terephthalate was selectively hydrolyzed and treated
subsequently with SOCl2.23 The Friedel-Craft acylation24
reaction of the resulting 4-(carbomethoxy)benzoyl chloride with
toluene and saponification and acidification of the product gave
4-(4′-methylbenzoyl)benzoic acid, which was refluxed with an
excess of SOCl2 to obtain 4-(4′-methylbenzoyl)benzoyl chloride.
To a solution of 4-(4′-methylbenzoyl)benzoyl chloride (7 mmol,
1.8 g) in 20 mL of dry ether was added a solution of tert-butyl
hydroperoxide (8-9.6 mmol of 5.0-6.0 M solution in decane,
1.6 mL) and triethylamine (8 mmol, 1.1 mL) in 10 mL of ether
in ∼10 min at the ice-bath temperature. The reaction mixture
was stirred for 1 h more at the same temperature and then
filtered. Crude 1 was obtained from the evaporation of the
filtrate. The crude product was chromatographed over silica
gel using CH2Cl2 as eluent and recrystallized twice from ether/
We measured the φPh of the probe molecules using 350
nm as the excitation wavelength (Table 5). It is clear that
the φPh of 4 (0.566) is comparable to that of benzophenone
(φPh ) 0.74).10 However, the φPh of 5 (0.025) is remarkably
decreased. This can be attributed to the bromine atom
induced enhancement of the T1 f S0 radiationless decay.
The deactivation of the triplet state into the ground state
causes fewer triplet states to undergo chemical decom-
position. Consequently, the quantum yield of chemical
decomposition is decreased.
We, however, found no significant difference in the φPh
values of the peresters. The φPh of 3 was found compa-
rable to the φPh of 1 that has no heavy atom. This
indicates that, in the case of 3, the heavy atom may have
little or no effect on the triplet state. This is also
supported by the nature of the triplet states of the
peresters that are assumed to be a very short-lived
species. The absence of an observable heavy atom per-
turbation on the triplet state of 3 may also be due to a
relatively larger energy gap between T1 and S0 states that
minimizes the spin-orbit interaction in the triplet state.
Thus, in the case of 3, the heavy atom probably affects
the triplet RP and induces its rapid spin inversion in the
solvent cage. As a result, the singlet RP is generated
which may recombine rapidly, causing a decrease in the
quantum yield of photodissociation processes. The mag-
nitude of this effect will be none or negligible in the case
of 1 or 2, respectively, where cage escape of the triplet
RP may take place prominently. Thus, it is the chemical
recombination of the geminate pairs that manifests in
the form of the low φdiss of 3. The bromine atom enhanced
geminate radical pair recombination may also be con-
tributing to the low φdiss of 5. However, this seems to be
the only path that is causing the low φdiss of 3.
1
hexane (1.4 g, yield 64%). H NMR (CDCl3) δ 7.23-8.09 (4d,
8H), 2.46 (s, 3H), 1.44 (s, 9H). Anal. Calcd for C11H20O4: C,
73.06; H, 6.45; O, 20.48. Found: C, 73.17; H, 6.13; O, 20.57.
Mp ) 78-79 °C (lit.4 80-81 °C).
(B) ter t-Bu tyl 4-(4′-Ch lor om eth ylben zoyl)p er ben zoa te
(2). 4-(4′-Methylbenzoyl)benzoyl chloride (7 mmol, 1.8 g) was
chlorinated25 in benzene, and 4-(4′-chloromethylbenzoyl)ben-
zoyl chloride (1.37 g) was obtained in 67% yield. This was
recrystallized from dry cyclohexane. The solid obtained after
evaporating cyclohexane still contained some dichloro product
and unreacted starting material. This was converted without
further purification to 2 by the same procedure described for
1. Chromatography of the product was carried out using
cyclohexanes/ether (3.5:1) as eluent to separate the dichloro
product. The resulting solid was recrystallized twice from
hexanes/ether to yield (49%) pure 2. 1H NMR (CDCl3) δ 7.52-
8.10 (4d, 8H), 4.66 (s, 2H), 1.45 (s, 9H). 13C NMR (APT, CDCl3)
δ 142.4, 141.44, 136.56, 130.86, 130.52 (2C), 129.9 (2C), 129.12
(2C), 128.61 (2C), 84.35, 45.25, 26.23 (3C). Anal. Calcd for
C
11H19O4: C, 65.8; H, 5.52; O, 18.45. Found: C, 64.78; H, 5.61;
O, 18.04. Mp ) 98-99 °C.
(C) ter t-Bu tyl 4-(4′-Br om om eth ylben zoyl)p er ben zoa te
(3). 4-(4′-Methylbenzoyl)benzoyl chloride (7 mmol, 1.8 g) was
brominated by refluxing it in benzene with NBS (8 mmol, 1.42
g) using benzoyl peroxide as initiator for 3 h. NBS was added
in three portions at intervals of 45 min. The resulting 4-(4′-
bromomethylbenzoyl)benzoyl chloride (1.2 g, yield 51.4%) was
recrystallized from dry cyclohexane, and the same procedure
applied for the preparation and purification of 2 was followed
Exp er im en ta l Section
Ma ter ia ls a n d Equ ip m en t. Reagents and solvents were
obtained from commercial suppliers and used as received
unless otherwise noted. Benzene and cyclohexane were dried
over sodium/benzophenone under argon before use. Solvents
were used without purification if they were either spectropho-
tometric or HPLC grade. Column chromatography and thin-
layer chromatography (TLC) were performed using standard
1
to obtain the pure perester (yield 41.6%). H NMR (CDCl3) δ
7.51-8.09 (4d, 8H), 4.54 (s, 2H), 1.44 (s, 9H). 13C NMR (APT,
CDCl3) δ 142.84, 141.47, 136.54, 130.92, 130.59 (2C), 129.88
(2C), 129.18 (4C), 84.35, 31.99, 26.25 (3C). Anal. Calcd for
C11H19O4: C, 58.33; H, 4.89; O, 16.35. Found: C, 58.79; H, 4.92;
O, 16.08. Mp ) 91-93 °C.
(20) (a) Inoue, H.; Sakurai, T.; Hoshi, T.; Okubo, J .; Ono, I. Chem.
Lett. 1990, 1059. (b) Mallik, G. K.; Pal, T. K.; Ganguly, T.; Banerjee,
S. B. Bull. Chem. Soc. J pn. 1988, 61, 3673. (c) Schloman, W. W., J r.;
Plummer, B. F. J . Chem. Soc., Chem. Commun. 1975, 705.
(21) (a) Kikuchi, K.; Hoshi, M.; Abe, E.; Kokubun, H. J . Photochem.
Photobiol., A 1988, 45, 1. (b) Asahi, Y.; Hirota, N. Bull. Chem. Soc.
J pn. 1982, 55, 1379. (c) Ferree, W. I., J r.; Plummer, B. F. J . Am. Chem.
Soc. 1973, 95, 6709.
(D) 4-(Ch lor om eth yl)ben zop h en on e (4) a n d 4-(br o-
m om eth yl)ben zop h en on e (5) were prepared by chlorina-
tion25 with SO2Cl2 and bromination with NBS in benzene of
(23) Hotten, B. W. Ind. Eng. Chem. 1957, 49, 1691.
(24) Smith, M. E. J . Am. Chem. Soc. 1922, 43, 1921.
(25) Kharasch, M. S.; Brown, H. C. J . Am. Chem. Soc. 1939, 61,
2142.
(22) Segall, J .; Wen, Y.; Singer, R.; Wittig, C.; Garcia-Vela, A.;
Gerber, R. B. Chem. Phys. Lett. 1993, 207, 504.
6122 J . Org. Chem., Vol. 67, No. 17, 2002

Triplet Energy Distribution in Photoinitiators
4-methylbenzophenone, respectively. Compound 4 was purified
by chromatography using ethyl acetate/hexane (1:15) as eluent
and recrystallized twice from ethanol; mp ) 93-94 °C (lit.26
95-96 °C). To separate 5 from the dibromo product and
unreacted starting material, the crude product was chromato-
graphed using cyclohexanes/ether (3.5:1) and recrystallized
twice from ethanol; mp ) 107-108 °C (lit.27 109-110 °C).
P r im a r y P h ot op r od u ct s. Irradiations were carried out
using a photochemical reactor equipped with a jacketed beaker
and 8 W × 16 RPR 3500 Å lamps. Quartz cuvettes (3.5 mL)
well dried and then sealed with a rubber septum were used
as photolysis vessels. All samples were degassed with argon
for at least 20 min. Disappearance of the starting material
and appearance of products were monitored by TLC and GC/
MS. The peresters undergo thermolysis when passed through
the column of the GC/MS. Thus, the GC/MS data of the
peresters were recorded before irradiation, which established
the nature of the thermolytic products. For the primary
products from photolysis, GC/MS spectra were recorded con-
secutively after 30 s, 1 min, and 2 min of irradiation. The
analysis of the GC/MS data taken before and after irradiation
in combination with the TLC observation established the
structure of the primary photoproducts. Comparison with the
GC/MS data of authentic samples was also made where it was
applicable.
dissociation were obtained from the disappearance of the peak
belonging to the tert-butyl group at 1.44 ppm. Similarly, the
disappearance of the peak corresponding to CH2-X was
monitored to obtain rates and φdiss values of C-X bond
dissociation. For rate measurement, data points collected up
to 60-90 s were considered. In this period, the change in
concentration was within 50%. For quantum yield calculations,
only the data recorded within 30-35 s (less than 15% conver-
sion) were used. The starting concentration of the samples was
chosen such that the absorption of the starting material
remained >1 at 350 nm after 20% conversion. The initial
absorptions of the molecules were 1.3-1.4, and thus, all the
light was absorbed by the starting material even after 15%
conversion. Benzophenone-benzhydrol actinometry10 was used
to determine the intensity of light. The phosphorescence
spectra were recorded in EPA glass at 77 K; the excitation
wavelength was 350 nm. Phosphorescence quantum yields
were determined according to the general procedure2 using
benzophenone as reference, and the reported φPh measure-
ments are normalized values.
Ack n ow led gm en t. The authors acknowledge in-
sightful discussions with Prof. G. S. Hammond. The
authors would also like to thank Prof. M. A. J . Rodgers
and Prof. T. H. Kinstle for their helpful discussions. One
of the authors (B.K.S.) would like to thank McMaster
Endowment for the financial support.
Ra te a n d Qu a n tu m Yield Mea su r em en ts. Quartz NMR
tubes were used for quantum yield measurements. The
solvents used were either CD3OD or C6D6, containing 0.03%
v/v TMS. Hexamethyldisiloxane was used as the internal
standard. The degassed and sealed samples were irradiated,
and spectra were recorded in 15 or 10 s intervals within 60-
90 s of the total irradiation time. The rate and φdiss of perester
Su p p or tin g In for m a tion Ava ila ble: Figures S1-2 (zero
order curve fit for rate constants), Figure S3 (NMR study of
decomposition of 4-(bromomethyl)benzophenone, Figure S4
(NMR spectra comparison of decomposition of peresters), Table
S1 (GC/MS peaks of photoproducts), and 1H and 13C NMR
spectra of 1, 2, and 3. This material is available free of charge
via the Internet at http://pubs.acs.org.
(26) Tanner, D. D.; Plambeck, J . A.; Reed, D. W. J . Org. Chem. 1980,
45, 5177.
(27) Tamiaki, H.; Kiyomori, A.; Maruyama, K. Bull. Chem. Soc. J pn.
1993, 66, 1768.
J O0202278
J . Org. Chem, Vol. 67, No. 17, 2002 6123