Photoinduced radical cleavage of bromophenyl ketones

Article abstract of DOI:10.1021/ja952782f

Several bromophenyl alkyl ketones undergo photoinduced radical cleavage of bromine atoms; the resulting acylphenyl radicals are trapped as the dehalogenated ketones either by alkane solvents or by added thiols. Quantum yields for this process decrease with increasing solvent viscosity; this behavior indicates that recoupling of the phenyl/ bromine radical pairs competes with their diffusion apart. In some cases Norrish type II elimination competes with the cleavage reaction. Steady state studies indicate room temperature rate constants for triplet state C-Br radical cleavage of 2 and 1 × 10⁸ s^-1, respectively, for m- and p-bromoacetophenone. Flash kinetics measurements of the temperature dependence for their triplet decay provided activation parameters for triplet state cleavage: ΔS^? = -5 eu for the para isomer and -3 eu for the meta; ΔH^? = 5.3 kcal/mol for both. The values for the meta isomer are the same in methanol and toluene and extrapolate to the room temperature decay rate measured by steady state techniques. The 0.002-0.003 quantum yields for conversion of m- and p-bromobenzophenone to benzophenone in cyclopentane solvent indicate triplet state cleavage rates of 0.8-1.0 × 10⁵ s^-1, which would suggest ΔH^? values near 10 kcal/mol. Consideration of triplet excitation energies and of the phenyl-Br bond strength suggests a mechanism for triplet cleavage involving activated C-Br bond stretching in the triplets until a π,π* configuration is converted into a dissociative n^π_Br,σ* state. It is suggested that the different isomers have comparable mixing of n^π_Br,π* character into their π,π* triplets and that the state interconversions occur inefficiently at surface crossings with a n^π_Br,σ* state. The o-bromo ketones show unusual behavior. The Arrhenius plot for triplet decay of o-bromoacetophenone is curved and only half the total cleavage can be quenched by naphthalene. o-Bromobenzophenone is 500 times more reactive than its isomers; direct interaction between bromine and carbonyl as well as steric depression of the C-Br dissociation energy may both contribute to this unusual reactivity.

Full text of DOI:10.1021/ja952782f

7
46
J. Am. Chem. Soc. 1996, 118, 746-754
Photoinduced Radical Cleavage of Bromophenyl Ketones
Peter J. Wagner,* James. H. Sedon, and Anna Gudmundsdottir
Contribution from the Chemistry Department, Michigan State UniVersity,
East Lansing, Michigan 48824
X
ReceiVed August 14, 1995
Abstract: Several bromophenyl alkyl ketones undergo photoinduced radical cleavage of bromine atoms; the resulting
acylphenyl radicals are trapped as the dehalogenated ketones either by alkane solvents or by added thiols. Quantum
yields for this process decrease with increasing solvent viscosity; this behavior indicates that recoupling of the phenyl/
bromine radical pairs competes with their diffusion apart. In some cases Norrish type II elimination competes with
the cleavage reaction. Steady state studies indicate room temperature rate constants for triplet state C-Br radical
8
-1
cleavage of 2 and 1 × 10 s , respectively, for m- and p-bromoacetophenone. Flash kinetics measurements of the
q
temperature dependence for their triplet decay provided activation parameters for triplet state cleavage: ∆S ) -5
q
eu for the para isomer and -3 eu for the meta; ∆H ) 5.3 kcal/mol for both. The values for the meta isomer are
the same in methanol and toluene and extrapolate to the room temperature decay rate measured by steady state
techniques. The 0.002-0.003 quantum yields for conversion of m- and p-bromobenzophenone to benzophenone in
5
-1
q
cyclopentane solvent indicate triplet state cleavage rates of 0.8-1.0 × 10 s , which would suggest ∆H values
near 10 kcal/mol. Consideration of triplet excitation energies and of the phenyl-Br bond strength suggests a
mechanism for triplet cleavage involving activated C-Br bond stretching in the triplets until a π,π* configuration
is converted into a dissociative n ,σ* state. It is suggested that the different isomers have comparable mixing of
n ,π* character into their π,π* triplets and that the state interconversions occur inefficiently at surface crossings
π
Br
π
Br
π
Br
with a n ,σ* state. The o-bromo ketones show unusual behavior. The Arrhenius plot for triplet decay of
o-bromoacetophenone is curved and only half the total cleavage can be quenched by naphthalene. o-Bromoben-
zophenone is 500 times more reactive than its isomers; direct interaction between bromine and carbonyl as well as
steric depression of the C-Br dissociation energy may both contribute to this unusual reactivity.
Introduction
with extra substituents. Very little prior work had been reported
on the photoinduced radical cleavage of halophenyl ketones.
Fluoro- and chlorobenzophenones undergo only normal pho-
The photoinduced cleavage of carbon-halogen bonds in aryl
halides is well-known.1 Direct irradiation of iodo-, bromo-,
and chlorobenzene produces phenyl and halogen radicals.
Defining the exact nature of the excited states responsible for
this cleavage has remained a challenging goal, mixing of π,π*
,2
9
,10
toreduction to halopinacols.
This result was to be expected
given that the C-X bond energies of such ketones are much
higher than their ∼68 kcal/mol triplet energies. Baum and Pitts
reported that some bromo- and iodophenyl ketones are pho-
4
states with π,σ* or σ,π* states and internal conversion to a
10
toreduced to the dehalogenated phenyl ketones. Their studies
indicate that radical cleavage reactions do occur but unfortu-
nately provided no direct measures of actual rate constants for
cleavage. Therefore we have investigated the triplet state
reactivity of various halophenyl ketones with the intent of
establishing how triplet configuration and energy affect rate
constants for cleavage of C-X bonds. This paper reports results
for bromophenyl ketones. We recently have reported the
dissociative σ,σ* or π,σ* state5 being common suggestions.
Bersohn first examined the photodissociation of several bromo
and iodo aromatics by molecular beam studies and found that
the former cleave much slower than the latter, with both
presumably involving triplet states.5
-7
Although a variety of substituted halobenzenes undergo
photoinduced cleavage, very little is known about how substit-
uents other than the halogen affect excited state reactivity. In
this regard we thought it worthwhile to study halophenyl
ketones, since the well-known rapid intersystem crossing of
phenyl ketones8 should confine reactivity to triplet states.
Moreover, excitation is confined to relatively well-defined n,π*
and π,π* triplets, with excitation energies that can be varied
11
behavior of iodobenzophenones and communicated the dif-
12
ferent behavior of iodo and bromo ketones.
X
Abstract published in AdVance ACS Abstracts, January 15, 1996.
Results
(1) (a) Sharma, R. K.; Kharasch, N. Angew. Chem., Int. Ed. Engl. 1968,
7
, 36. (b) Pinhey, J. T.; Rigby, R. D. G. Tetrahedron Lett. 1969, 1267. (c)
Fox, M.; Nichols, W. C.; Lemal, D. M. J. Am. Chem. Soc. 1973, 95, 8164.
2) The entire topic is almost entirely ignored in organic photochemistry
textbooks, with Barltrop and Coyle providing the most extensive coverage
Competing Photoreactions of Bromo Ketones. The four
ketones p-bromovalerophenone (pBrV), m-bromovalerophenone
(
(8) (a) Rentzepis, P.; Mitschele, C. J. Anal. Chem. 1970, 42, 20A. (b)
Anderson, R. W., Jr.; Hochstrasser, R. M.; Lutz, H.; Scott, G. W. Chem.
Phys. Lett. 1974, 28, 153.
(9) Wagner, P. J.; Truman, R. J.; Scaiano, J. C. J. Am. Chem. Soc. 1985,
107, 7093.
(10) (a) Baum, E. J.; Pitts, J. N. J. Phys. Chem. 1966, 70, 2066. (b)
Baum, E. J.; Wan, J. K. S.; Pitts, J. N. J. Am. Chem. Soc. 1966, 88, 2652.
(11) Wagner, P. J.; Waite. C. I. J. Phys. Chem. 1995, 99, 7388.
(12) Wagner, P. J.; Sedon, J.; Waite, C.; Gudmundsdottir, A. J. Am.
Chem. Soc. 1994, 116, 10284.
3
(
one page).
3) Barltrop, J. A.; Coyle, J. D. Excited States in Organic Chemistry;
Wiley: London, 1975.
4) Nagaoka, S.; Takemura, T.; Baba, H. Bull. Chem. Soc. Jpn. 1985,
8, 2082.
(
(
5
(
(
(
5) Dzvornik, M.; Yang, S.; Bersohn, R. J. Chem. Phys. 1974, 61, 4408.
6) Hwang, H. J.; El-Sayed, M. A. J. Chem. Phys. 1992, 96, 856.
7) Nakaoka, S.; Takemura, T.; Baba, H.; Koga, N.; Morokuma, K. J.
Phys. Chem. 1986, 90, 759.
0
002-7863/96/1518-0746$12.00/0 © 1996 American Chemical Society

Photoinduced Radical CleaVage of Bromophenyl Ketones
J. Am. Chem. Soc., Vol. 118, No. 4, 1996 747
Figure 1. Effect of added butanethiol on quantum yields for dehalo-
genated phenyl alkyl ketone from (O) pBrV and (b) pBrB in benzene.
Figure 2. Effect of added thiol in solvents of different viscosity on
quantum yields for formation of acetophenone from (b) pBrA in
n-hexane, (O) pBrA in benzene, (2) mBrA in cyclohexane, and (9)
oBrA in cyclooctane.
(mBrV), p-bromobutyrophenone (pBrB), and p-chlorobutyro-
phenone (pClB) were synthesized as described in the Experi-
mental Section. Irradiation of degassed benzene solutions of
each halo ketone gave mixtures of products from Norrish type
II hydrogen abstraction and from radical C-X bond cleavage:
the haloacetophenones and minor amounts of what are assumed
to be cyclobutanols; the corresponding dehalogenated phenyl
ketones; and a fourth product B, shown to be p-phenylvale-
rophenone in the case of pBrV. As already reported,1 pBrB
undergoes primarily radical cleavage and produces negligible
type II product. When the reactions were run in the presence
of added butanethiol, no biphenylyl ketone B was formed and
the yields of dehalogenated phenyl ketone were enhanced.
Likewise the addition of tert-butyl alcohol enhanced yields of
the Norrish type II products, except for pBrB which forms no
measurable type II product.
0b
Figure 3. Effect of added alcohol on quantum yield for p-bromo-
acetophenone formation from pBrV.
The three isomeric bromoacetophenones o-, m-, and pBrA
were irradiated in benzene with added thiol and in several alkane
solvents with and without added thiol. In all cases acetophenone
was the predominant organic product, representing >96% of
the BrA reacted. The fate of the other radical species was not
scrutinized quantitatively; but combined GC-mass spectrometric
analysis showed that cyclohexyl bromide (30%) and bicyclo-
hexyl were formed when pBrA was irradiated in cyclohexane
Table 1. Photokinetics for Competitive Halogen Cleavage and
Norrish Type II Elimination or Photoreductiona
ketone Φ-rctnt
ΦII
ΦIImax b
Φcl
Φclmax
ΦB
kqτ, M-1
17 ( 2
mBrV
pBrV
pBrB
pClB
oBrA
pBrBzP
0.17
0.10
0.087 0.25
0.033 0.072 0.014 0.14 0.021
0.007 0.07c 0.038
c
d
34 ( 3
58 ( 6
0.12 <0.001 <0.002 0.02
0.17c 0.034
c
e
0.43
<0.001 0.005 0.0013 2000
f
0.38g
0.16
24
f
0.002 0.003h
solvent. HBr was also detected. Bromine atoms, unlike iodine
f
0.0013 0.002h
atoms,11 apparently abstract hydrogen atoms from solvent13 in
mBrBzP
f
g,h
oBrBzP
0.09
0.38
110
competition with their dimerization. When the BrA’s were
irradiated in benzene alone, the solutions took on the reddish
tinge of Br2. The three bromobenzophenones o-, m-, and
pBrBzP were irradiated in benzene with added octanethiol and
in cyclopentane with and without added thiol. In all cases
benzophenone was formed, as already reported for the ortho
a
In benzene unless otherwise noted. b 2 M tert-butyl alcohol. c 0.5
M butanethiol. d 31 for quenching of p-bromoacetophenone formation;
3
6 for quenching of valerophenone formation. e Calculated from data
f
for butyrophenone and p-chlorovalerophenone, ref 21. In cyclopentane.
g
In benzene. h 0.03 M n-C
8
H17SH added.
1
0a
and para isomers, together with the expected photoreduction
formation from bromoacetophenones in various alkane solvents.
Figure 3 shows the usual effect of added alcohol
9
products from the meta and para isomers. The ortho isomer,
15,16
on
however, undergoes primarily dehalogenation, forming ben-
zophenone in 92% yield in cyclohexane, together with bro-
mocyclohexane and a small amount of fluorenone.
acetophenone yields from pBrV.
Table 1 lists the quantum yields for product formation and
reactant disappearance for all the halo ketones under different
conditions. It also lists kqτ values described by the slopes of
linear Stern-Volmer quenching plots. These were obtained by
measuring product yields from 365-nm excitation in the presence
of various concentrations of naphthalene. Decreases in both
dehalogenation and type II elimination were measured, with
good agreement between kqτ values measured both ways for
pBrV when butanethiol was included to trap all of the
acylphenyl radicals. Only half of the dehalogenation of oBrA
is quenched by naphthalene; the value of Φ°/Φ plateaus at 2 at
Quantum yields for the various products were determined
by parallel irradiation (313 nm) of degassed, sealed tubes
containing solutions 0.02-0.05 M in halo ketone. Actinometer
1
4
solutions containing 0.1 M valerophenone in benzene were
irradiated at the same time. Product yields were then determined
by GC or HPLC analysis. Figure 1 shows the effect of added
thiol on the quantum yields for reduction of two of the ketones
in benzene, the high concentration limit representing complete
scavenging of photoproduced aryl radicals.15 Figure 2 shows
the effect of added thiol on quantum yields for acetophenone
(15) (a) Wagner, P. J.; McGrath, J. M. J. Am. Chem. Soc. 1972, 94,
(
13) Scaiano, J. C.; Barra, M.; Krzywinski, M.; Sinta, R.; Calabrese, G.
J. Am. Chem. Soc. 1993, 115, 8340.
14) Wagner, P. J.; Kochevar, I. E.; Kemppainen, A. E. J. Am. Chem.
Soc. 1972, 94, 7489.
3849. (b) Lewis, F. D.; Magyar, J. G. J. Am. Chem. Soc. 1973, 95, 5973.
(16) (a) Wagner, P. J. J. Am. Chem. Soc. 1967, 89, 5898. (b) 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.
(

7
48 J. Am. Chem. Soc., Vol. 118, No. 4, 1996
Wagner et al.
Table 3. Spectroscopic Features of Halophenyl Ketones
phosp 0,0
max (ꢀ)a
n,π* λmax (ꢀ)a
(kcal/mol)
b
ketone
L
a
λ
pBrV
pBrA
pClBe
252 nm (19 000) 325 nm (68)
249 nm (18 500) 325 nm
399 nm (71.7)c
d,h
405 nm (70.6)
406 nm (70.5)
c
d
d
h
3
96 nm (72.3)
mBrV
mBrA
mClB
239 nm (7 200)
326 nm (47)
396 nm (72.2)
396 nm (72.2)
239 nm (10 000) 326 nm
405 nm (70.6)f
d
3
89 nm (73.6)
h
oBrA
237 nm (8 000)
316 nm (sh) (∼80) 400 nm (( 67 81 .. 95 ))
c,g
c,g
h
pBrBzP
mBrBzP
oBrBzP
(68.3)
412 nm (69.5)
Figure 4. Quenching by naphthalene of benzophenone formation from
oBrBzP (O) and acetophenone formation from oBrA (9), both in
cyclohexane.
a
In heptane. b At 77 K. c In 4:1 methylcyclohexane/isopentane. In
d
f
g
h
4
2
:1 methanol/ethanol. e From ref 21. In isopentane. From ref 19. In
-methyltetrahydrofuran.
Table 2. Maximized Quantum Yields for Reductive
Dehalogenation of Bromophenyl Ketones in Various Solvents
viscosity, cP
(at 25 °C)
solvent
hexane
oBrA mBrA pBrA oBrBzP
0.29
0.63
0.63
0.88
2.16
0.25
0.16
0.38
0.15
0.08
0.23
0.03
0.28
0.11
0.07
0.24
0.01
0.17
0.12
0.04
0.12
0.09
0.38
0.07
0.03
benzene
benzene
a
cyclohexane
cyclooctane
a
With 0.1 M dodecanethiol or octanethiol.
Figure 6. Phosphorescence at 77 K in methyltetrahydrofuran glasses
of (s) mBrA and (- - -) pBrA.
Figure 5. Arrhenius plots of triplet decay for (0) pBrA and (2) oBrA
in toluene and (O) mBrA in methanol.
high concentrations of 1-methylnaphthalene. Figure 4 compares
the curved Stern-Volmer plot for oBrA to the linear plot for
oBrBzP. The kqτ value for oBrA was derived by nonlinear
least-squares analysis of the data.17 No such curvature was
evident in quenching plots of m- and p-bromo ketones.
Table 2 lists maximum quantum yields for formation of
acetophenone from the bromoacetophenones in different alkane
solvents. As shown in Figure 2 added thiol generally has only
small effects on these values, although it appears to generally
raise quantum yields somewhat in hexane but not in the more
viscous solvents.
Figure 7. Phosphorescence at 77 K in methyltetrahydrofuran glasses
of (s) oBrA and (- - -) oBrBzP.
plateaus at low temperatures. The plots are extrapolated to room
temperature for comparison of decay rates with those determined
by quenching studies.
Spectroscopy. Table 3 lists the major spectroscopic features
of these bromophenyl ketones. Two chloro ketones are included
for the sake of comparison; it appears that bromo and chloro
substitution have nearly identical effects on both UV and
phosphorescence spectra. Figure 6 compares the phosphores-
cence spectra of the m- and p-bromoacetophenones and Figure
Flash Kinetics. Laser flash excitation of toluene solutions
∼
0.005 M in o-, m-, or p-bromoacetophenone produced strong
T-T absorption between 400 and 450 nm together with an
absorption maximum at ∼550 nm characteristic of the bromine
atom-benzene complex.18 Decay lifetimes of the triplet signals
were measured between 190 and 290 K in methanol. The
resulting Arrhenius plots are shown in Figure 5. Those for the
meta and para isomers are linear (an identical plot was obtained
for the meta isomer in toluene), whereas that for the ortho isomer
7
those of the o-bromo ketones.
Discussion
Two facts indicate that the photochemical reactions of all
these bromophenyl ketones (with the exception of oBrA) arise
from their triplet states: (1) both cleavage and γ-hydrogen
(
17) Wagner, P. J. In Handbook of Organic Photochemistry; Scaiano, J.
C., Ed.; CRC Press: Boca Raton, FL, 1989; Vol. 2, p 251.
18) McGimpsey, W. G.; Scaiano, J. C. Can. J. Chem. 1988, 66, 1474.
(

Photoinduced Radical CleaVage of Bromophenyl Ketones
J. Am. Chem. Soc., Vol. 118, No. 4, 1996 749
Table 4. Rate Constants (10 s-1) for Triplet Halo Ketonesa
8
abstraction are quenchable by normal triplet quenchers; and (2)
the decay kinetics of the transients observed by flash spectros-
copy reveal the same cleavage rate constants as those gleaned
from the steady state quenching studies.
1/τb
model k c
ketone
H
k
H
k
-X
mBrVd
3.5
1.8
1.0
2.5
1.35
1.0
d
pBrVd
1.75
1.0
0.4
0.4
0.02e
0.02
0.02
0.02
Nature of Triplets. That halobenzophenones have n,π*
pBrBd
pClB
lowest triplets is well established.19 Even oBrBzP displays
0.02
e
0.00015
3.0
f
g
oBrA
3.0
phosphorescence with sharp features and typical n,π* vibronic
mBrA
pBrA
2.5 (2 0)h
2.0
structure. Such is not the case for many substituted phenyl alkyl
h
1.0 (0.7)
0.7
ketones.2
0,21
The major broadening of the phosphorescence
f
g
oBrBzP
0.6
<0.001
0.076
0.063
0.6
i
j
spectra of the p-bromo phenones suggests that their lowest
triplets are π,π* in nature, as is the case for p-chloroacetophe-
mBrBzP
0.076
0.076
0.063
0.0008
0.0010
i
j
pBrBzP
0.063
2
2
a
b
none. Their UV absorption reveals CT stabilization of the
At room temperature. Derived from k τ values, with k ) 6 ×
-1 -1 c d e
q
q
1
9
La state and thus of the triplet. This effect is slightly larger
10 M
s . Analogous chloroketone, ref 21. In benzene. Extrapo-
lated from the value for butyrophenone by the effect of p-Cl on
than for the p-chloro ketones. There is no difference between
the absorption of the m-chloro- and m-bromophenones and they
all display nicely structured phosphorescence spectra; but the
bromo ketones emit at lower energy. It is likely that a meta
bromine can stabilize the π,π* triplet by charge transfer such
f
g
9
-1 -1
valerophenone. In cyclohexane. Based on k
q
) 7 × 10 M
s ,
ref 23b. h Value extrapolated from Arrhenius plots in toluene and/or
i
j
methanol. In cyclopentane. Value for the corresponding chloroben-
zophenone, ref 9.
and meta substitution producing a small inductive rate enhance-
21
as happens with a methoxy substituent, but emission still
occurs from a nearby n,π* triplet.22 The oBrA emission spec-
trum is broadened but still shows n,π* vibronic structure.
Together the spectra suggest that n,π* and π,π* triplets of the
bromophenones have similar energies for all three substitution
patterns.
9
,21
ment for the n,π* triplet.
The tiny type II quantum yield
1
0b
for pBrB, noted by Baum et al., is consistent with the low
8
-1
kH value for pClB and sets k-Br at 1 × 10 s . The faster
triplet decay of pBrV is consistent with a similar value for k-Br
and a kH value the same as that for p-chlorovalerophenone. The
measured type II quantum yield for mBrV is 28% that for
m-chlorovalerophenone and indicates the kH and k-Br values
listed. The kH value is only half that for the model chloro ketone
and suggests some lowering of the π,π* triplet by electron
donation from the bromine, as suggested above.
It is noteworthy that a meta bromine cleaves faster than a
para bromine, a fact disguised by the relative product quantum
yields from mBrV and pBrV. (The type II yield from the latter
is low enough that it was missed in an earlier study. ) The
deceptively high meta/para type II ratio is caused by the large
depression of the hydrogen abstraction rate caused by a para
Reactions of Triplets. In some cases radical C-Br bond
cleavage competes with hydrogen abstraction either from a
γ-carbon or from solvent; in others it is the sole reaction. Thus
the quantum yields for Norrish type II elimination from the
bromo ketones are lower than those of the corresponding chloro
2
1
ketones and are accompanied by typical radical trapping
products, especially the dehalogenated phenyl ketone. The
bromine atoms were not specifically trapped but were detected
by flash spectroscopy as the benzene complex. The very low
quantum yields for benzophenone formation from mBrBzP and
pBrBzP reveal that cleavage is much slower than the known
rapid hydrogen abstraction from cyclopentane.9 In contrast,
oBrBzP undergoes cleavage very efficiently, as do the bro-
moacetophenones.
21
2
1
bromine.
The very low cleavage rate constant for pClB was estimated
by multiplying the extrapolated triplet lifetime by estimated
quantum efficiencies for cleavage. The values for the bro-
mobenzophenones required similar calculation. The triplet
decay rate for pBrBzP has been measured in alkane solvent24
and closely matches the decay rate of p-chlorobenzophenone
in cyclopentane, which represents hydrogen abstraction from
9
solvent. Thus the triplet lifetimes of both isomers were
assumed to equal those of the chloro ketones. These lifetimes
9
were first multiplied by the maximum cleavage quantum yields
and then, in order to correct for revertible cleavage (see below),
divided by 0.18, the quantum yields of acetophenone formation
from the corresponding bromoacetophenones in cyclopentane.
The cleavage rates of the bromoacetophenones estimated from
quenching of the bromovalerophenones match those measured
by direct flash kinetics well.
Radiationless Decay to Hot Ground States? A referee has
suggested that the bromine atoms might induce sufficent spin-
orbit coupling in the triplets to promote radiationless decay to
vibrationally excited ground state reactant, which could then
undergo bond cleavage by rapid redistribution of vibrational
energy into the relatively weak C-Br bond. This generic
possibility has been discussed over the years for most photo-
chemical cleavage reactions and rarely found to be important
in solution because of the rapid transfer of vibrational energy
Table 4 dissects the triplet lifetimes gleaned from the kqτ
values and quantum yields into competing rate constants for
9
hydrogen abstraction and radical cleavage. A value of 6 × 10
-1
-1
23
M
s
was used for kq. We assume that kH values for the
bromo ketones should be similar to those for the corresponding
chloro ketones already studied, with para substitution slowing
hydrogen abstraction by stabilizing the unreactive π,π* triplet
(
19) Arnold, D. R. AdV. Photochem. 1968, 6, 301.
(20) (a) Kearns, D. R.; Case, W. A. J. Am. Chem. Soc. 1966, 88, 5087.
(
b) Yang, N. C.; McClure, D. S.; Murov, S. L.; Houser, J. J.; Dusenbery,
R. J. Am. Chem. Soc. 1967, 89, 5466.
21) Wagner, P. J.; Kemppainen, A. E.; Schott, H. N. J. Am. Chem. Soc.
973, 95, 5604.
22) Wagner, P. J.; May, M. J.; Haug, A.; Graber, D. R. J. Am. Chem.
Soc. 1970, 92, 5269.
(23) (a) Clark, W. D. K.; Litt, A. D.; Steel, C. J. Am. Chem. Soc. 1969,
91, 5413. (b) Wagner, P. J.; Kochevar, I. J. Am. Chem. Soc. 1968, 90, 2232.
(c) Scaiano, J. C.; Leigh, W. J.; Meador, M. A.; Wagner, P. J. J. Am. Chem.
Soc. 1985, 107, 5806.
(24) Jacques, P.; Lougnot, D. J.; Fouassier, J. P.; Scaiano, J. C. Chem.
Phys. Lett. 1986, 127, 469.
(
1
(

7
50 J. Am. Chem. Soc., Vol. 118, No. 4, 1996
Wagner et al.
to solvent. Thus we doubt that such radiationless decay is
responsible for efficient C-Br bond cleavage; but it could
explain quantum yields <50% for radical formation. Our var-
ious kinetics measurements provided rate constants, listed as
k-X in Table 4, for a triplet decay process that leads to C-Br
bond cleavage. Although these measured values actually equal
the rate constant for chemical bond cleavage k-X plus the rate
constants for any competing physical decay processes, we have
assumed that they represent only direct bond cleavage. Phos-
phorescence rates definitely are too slow to contribute to the
observed rapid decays; but how do we justify neglect of radia-
tionless decay? There is ample evidence that the spin-orbit
coupling induced by bromine substitution increases the radia-
Figure 8. Arrhenius plot for decay of triplet mBrA corrected for
competing hydrogen abstraction from methanol: (b) actual data; (- - -)
25
tionless decay rates kd of triplet aromatics such as naphthalene;
but such enhanced rate constants remain orders of magnitude
9
-3200/RT
R
H
) (1 × 10 )e
H
; (s) ) 1/τobs - R .
8
-1
lower than the measured 10 s decay rates of these bromoac-
Studies on the reactivity of phenyl radicals indicate that they
abstract hydrogen atoms from alkanes with rate constants on
etophenones. The value of kd for triplet 1-bromonaphthalene
-1
increases from 50 to 4000 s upon raising the temperature from
6
-1 -1 28
the order of 10 M s . Such values are high enough for
most or all of the acylphenyl radicals that have diffused away
from their sibling bromine atoms to be scavenged by alkane
solvent. The fact that added thiol produces slight or no increases
in acetophenone yields (Figure 2) in alkane solvents, unlike the
large increases produced in benzene, verifies that the acetophe-
none yields represent the efficiency for escape of acylphenyl
radicals from the initial solvent cage. In these solvents quantum
yields for reductive dehalogenation decrease with increasing
viscosity in a manner consistent with recoupling of the initial
caged radical pair being competitive with its diffusive separation.
It is interesting how much higher the quantum yields seem to
be in benzene than in alkanes of similar viscosity. The ability
of benzene to complex with bromine atoms apparently can
suppress recoupling of the acylphenyl radicals with bromine,
certainly for those that diffuse apart and perhaps even for the
original caged pair.
2
5
7
7 K to ambient, indicating a maximum activation energy of
only 0.9 kcal/mol, far lower than the 5.3 kcal measured for these
bromo ketones. We thus rule out radiationless decay as a
significant competitive triplet decay process.
Quantum Yields. The model chlorovalerophenones undergo
type II elimination with quantum yields in the 35-43% range
common for phenyl ketones in which γ-hydrogen abstraction
accounts for 100% of triplet decay. The lower values for the
bromo ketones gauge the extent of competing decay reactions,
which varies from 70 to 98%. The quantum yield of trapped
acylphenyl radicals from pBrB and pBrV is only 17% in
benzene. This value being so much lower than the percent of
triplet reaction competing with γ-hydrogen abstraction means
that most of the cleavage process involves reversion to the
ground state of the reactant. For mBrV, the maximum type II
quantum yield of only 25% (compared to 90% for valerophe-
none26 ) indicates that some 72% of the triplets cleave. The
It is worth mentioning that thiols are known to quench triplet
ketones by a charge-transfer process that normally does not
7
% yield of valerophenone again indicates great inefficiency
in the cleavage reaction. As described above, we believe that
these inefficiencies do not reflect direct radiationless decay of
an excited state but rather represent revertible cleavage, as is
common for triplet reactions and particularly noticeable for the
radical cleavage of halo ketones.27 Even though the radical pair
formed by cleavage is initially in a triplet state, the spin-orbit
coupling promoted by the bromine atom apparently is sufficient
to allow most of the radical pairs to convert to singlets and
couple before they diffuse apart to free, trappable radicals. We
have noted the same effect for iodophenyl ketones.1 In the
absence of a good hydrogen source, the acylphenyl radicals add
to the benzene solvent. Subsequent disproportionation of
various radicals produces the acylbiphenyls that were observed
as well as some dehalogenated phenyl ketone.
29
generate products efficiently. Fortunately, the concentrations
needed to trap acylphenyl radicals are too low to measurably
affect the short-lived bromoacetophenone triplets.
Cleavage Rate Constants. Both steady-state and time-
resolved measurements agree that m-bromoacetophenone and
its homologs undergo triplet state cleavage slightly faster than
do the para isomers. As Figure 5 shows, both isomers of
bromoacetophenone have similar activation energies; the ortho
isomer appears to be similar, at least at higher temperatures.
We did not determine the cause for the apparent unactivated
decay process of oBrA at low temperature. We discuss possible
causes below.
1
As discussed above for the bromovalerophenones, hydrogen
abstraction from the solvent should be much faster for mBrA
9
,30
than for pBrA.
Figures 8 and 9 present Arrhenius plots for
the two ketones corrected for competing hydrogen abstraction.
The observed triplet decay rates were decreased by calculated
rates for hydrogen abstraction from 10 M solvent, those values
9
-(3200/RT)
8
-(4000/RT)
being 10kH ) (1 × 10 )e
and (5 × 10 )e
for
meta and para, respectively. The activation parameters for
hydrogen abstraction were derived from measurements by
Steel, adjusted by our measurements of the effects of chloro
3
1
9
substituents. As the figures show, no correction is really needed
(
25) (a) Ermolaev, V. L.; Svitashev, K. K. Opt. Spectrosc. 1959, 7, 399.
(28) Scaiano, J. C.; Stewart, L. C. J. Am. Chem. Soc. 1983, 105, 3609.
(29) Zepp, R. G.; Wagner, P. J. J. Chem. Soc., Chem. Commun. 1972,
167. Guttenplan, J. B.; Cohen, S. G. J. Org. Chem. 1973, 38, 2001.
(30) Wagner, P. J.; Truman, R. J.; Puchalski, A. E.; Wake, R. J. Am.
Chem. Soc. 1986, 108, 7727.
(31) Berger, M.; McAlpine, E.; Steel, C. J. Am. Chem. Soc. 1978, 100,
5147.
See: Birks, J. B. Photophysics of Aromatic Molecules; Wiley: London,
970; p 264. (b) Hoffman, M. Z.; Porter, G. Proc. R. Soc. Londond Ser. A
962, 288, 46.
1
1
(
26) 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.
27) McGimpsey, W. G.; Scaiano, J. C. Can. J. Chem. 1988, 66, 1474.
(

Photoinduced Radical CleaVage of Bromophenyl Ketones
J. Am. Chem. Soc., Vol. 118, No. 4, 1996 751
Figure 9. Arrhenius plot for decay of triplet pBrA corrected for
competing hydrogen abstraction from toluene: (O) actual data; (- - -)
8
-4000/RT
R
H
) (5 × 10 )e
H
; (s) ) 1/τobs - R .
Figure 10. Natural correlation diagram between excited triplets of
halophenyl ketones and radical pair cleavage products; (s) connects
for the para isomer, but the activation parameters for the meta
isomer are somewhat higher than the uncorrected data would
suggest. For pBrA, Ea ) 5.7 kcal/mol, log A ) 12.0; for
mBrA, Ea ) 5.7 kcal/mol, log A ) 12.4. Eyring plots produced
values of ∆H ) 5.3 kcal/mol for both isomers and ∆S ) -3
and -5 eu for mBrA and pBrA, respectively. Thus the actual
differences in room temperature rates, with which the extrapo-
lated Arrhenius plots agree well (see Table 4), seem to be due
to small entropic variations, with no measurable difference in
activation energies.
symmetric states; (- - -) connects antisymmetric states. n
X
orbital
π
designation is equivalent to n in the text.
Br
q
q
with the chlorophenyl ketone component undergoing cleavage
36
akin to the halide ion elimination observed for radical anions.
Mechanism of Cleavage. These compounds undergo con-
certed homolytic cleavage; what must be explained is why the
bromo ketones show so little positional dependence, in contrast
to the 100-fold para/meta rate ratio shown by the iodoben-
Thermodynamics. The latest estimates for the bond dis-
sociation energy of bromobenzene range from 78 to 80 kcal/
mol.3 These values would appear to require activation
enthalpies of at least 6-8 kcal/mol for the bromoacetophenone
triplets, somewhat greater than those observed. The discrepancy
probably is due mainly to the Stokes shift that we have discussed
previously; the actual triplet energy of n,π* triplets is usually
11,12
zophenones.
The standard interpretation of how such C-X
radical cleavages occur from molecules whose lowest excited
2
states have no σ* character invokes an activated internal
4-7
conversion from a π,π* triplet to a dissociative state.
This
conversion occurs on the C-X bond stretching coordinate, along
which the energy of the dissociative σ* states decrease. The
standard approach to determining a reaction pathway involves
first constructing a correlation diagram between excited states
and products and then calculating or estimating where avoided
crossings between the lowest excited states and the dissociative
states are the most likely. We shall do so and then examine
some specific features of these reactions that favor one particular
pathway for cleavage.
33
1
-2 kcal/mol higher than the phosphorescence 0,0 band. There
also may be a small error in the latest but oft-changed enthalpy
of formation for phenyl radicals; or acyl substitution may lower
the C-X bond energy. The latter seems unlikely, since the
effect would have to be identical for ortho, meta, and para
substitution. In any event, it appears that the transition state
has nearly the same energy as that of the dissociated radical
pair, in which case there is no activation energy for coupling
of the triplet radical pair.
Figure 10 shows how the various excited states of these bromo
ketones correlate by symmetry with cleavage product states.
The excited states are drawn so as to show the orbitals on which
The triplet cleavage rates for the m- and p-bromobenzophe-
nones are some three orders of magnitude slower than that for
the corresponding bromoacetophenones. This difference amounts
to an additional 4-4.5 kcal/mol of activation energy, which is
exactly what would be predicted from the lower triplet energies
of the benzophenones. It is noteworthy that the bromoben-
zophenones do not exhibit the marked isomeric differences in
triplet reactivity shown by the iodobenzophenones.11,12
The small but measurable amount of cleavage for pClB is
unexpected. The 10 s rate constant deduced from the data
would represent an activation energy of 11 kcal and thus a
phenyl-Cl bond energy of 83 kcal/mol. This value is close to
that formerly considered the standard.34 One thus could view
our results as evidence for lower phenyl-X bond energies. It
is more likely that cleavage may be due to a small fraction of
charge transfer quenching by the thiol35 added to trap radicals,
the unpaired electrons are mainly localized. The nX,σ* and σ,σ*
4-7
states of halobenzenes are dissociative;
the phenyl radical
can have σ or π symmetry while the bromine atom can have
its unpaired electron in any of three orthogonal orbitals.
Consequently the reaction is pentatopic, involving π and σ
π
orbitals on the benzene ring and n, σ, and n orbitals on bromine.
Our results obviously provide no way to distinguish among the
three bromine atom symmetries. Since one p orbital on bromine
is orthogonal to the σ and π systems involved in the reaction,
its importance is negligible; states involving it have been omitted
and we treat the reaction as tetratopic.
4
-1
The ketones have either n,π* or π,π* lowest triplets, which
lie within a few kcal/mol of each other in the aryl alkyl ketones;
so both must be considered as “reactants”. We first consider
whether one is more likely than the other to convert into a
dissociative state. States with π,π symmetry are connected by
solid lines, and those with σ,π symmetry, by dotted lines. It
was originally suggested that radical cleavages of excited states
proceed by avoided crossings between states of the same overall
(32) (a) Benson, S. W. Thermochemical Kinetics, 2nd ed.; Wiley: New
York, 1976. (b) Egger, K. W.; Cocks, A. T. HelV. Chem. Acta 1973, 56,
1
516. (c) McMillen, D. F.; Golden, D. M. Annu. ReV. Phys. Chem. 1982,
3
3, 493.
3
7,38
(
33) Wagner, P. J.; Thomas, M. J.; Harris, E. J. Am. Chem. Soc. 1976,
symmetry.
In that case the π,π* triplet, which actually
9
8, 7675.
(
(
67.
34) Benson, S. W. Thermochemical Kinetics; Wiley: New York, 1968.
35) Zepp, R. G.; Wagner, P. J. J. Chem. Soc., Chem. Commun. 1972,
(36) Andrieux, C. P.; Saveant, J.-M.; Su, K. B. J. Phys. Chem. 1986,
90, 3815.
(37) Michl, J. Fortschr. Chem. Forsch. 1974, 46, 1.
1

7
52 J. Am. Chem. Soc., Vol. 118, No. 4, 1996
Wagner et al.
correlates with a radical pair containing an excited π-phenyl
radical, would couple with a σ,σ* state that leads to a triplet
Scheme 1
pair of σ radicals. Likewise, the n,π* triplet would then couple
π
with the n ,σ* state that forms a σ phenyl radical and a π
Br
3
bromine atom. However, the actual cleavage of π,π* haloben-
zenes has been attributed to coupling with a n ,σ* state,
while we have invoked coupling of the n,π* state with the σ,σ*
state to explain the behavior of the iodobenzophenones.
π
Br
4-7
3
1
1,12
Several aspects of our results have a strong bearing on state
coupling possibilities. The nearly identical activation parameters
for m- and p-bromoacetophenones reflect their similar triplet
excitation energies and thus indicate very similar coupling
between unreactive (bound) and reactive (dissociative) states.
Moreover, their n,π* and π,π* triplets have such similar 0,0
energies, both lower than that of the radical pair, that whichever
one couples better with a dissociative state should drive the
cleavage in all cases. (It is well established that a facile
equilibrium between π,π* and n,π* triplets allows efficient
meta.42 Why then does the radical cleavage of these triplet
bromophenyl ketones show no positional effect? If the unpaired
electron density resided primarily at the para carbon in all
isomers of the bromo ketones, and if it were mainly responsible
for cleavage probability, it is not clear how both meta and para
isomers could cleave with such similar A values and rates. Thus
the cleavage of these π,π* triplets does not appear to be
dominated by π* electron density as it is for the radical anions
of the same compounds and for the n,π* triplets of the
iodobenzophenones. A major difference between n,π* and π,π*
triplets is that both meta and para electron-donating substituents
hydrogen abstraction by the latter when the former are lower
in energy.2
1,39
) The fact that the n,π* triplets of iodoben-
11,12
21
zophenones show a large meta/para rate difference,
whereas
interact very strongly with the benzoyl π system in the latter;
none of the bromophenyl ketones show such a difference,
strongly suggests that it is the π,π* triplets of these bromophenyl
ketones that dominate the dissociation process. Put another way,
it is difficult to comprehend the endothermic debromination of
the lone pair p orbital on the halogen that is parallel to the
benzene π system couples with the π system such that a degree
of electron transfer from halogen to a π* orbital is mixed into
the π,π* configuration to produce changes in electron density
8
-1
that have not been well characterized. Therefore we suggest
mBrA (k ) 2 × 10 s ) being faster than the exothermic
π
Br
7
-1
that there is a common component of n f π* character
deiodination of mIBzP (k ) 3 × 10 s ) if they both react
3
mixed into the π,π* triplet of all isomers and that it is this
component that allows comparable mixing with a dissociative
state.
from their n,π* states, which are the lowest triplets for each.
The fact that A values for the bromoacetophenones are 1012
s- indicates that there is some inefficiency in the state
interconversion, since the fragmentation itself inherently pro-
duces a positive entropy change. By comparison, A values near
1
If we now return to Figure 10, the symmetry-allowed
conversion of a π,π* triplet into a σ,σ* state would require a
double electron transfer between orthogonal orbitals, so that
coupling of the two states should be weak. Such orthogonality
has been deemed responsible for the low rate constants for
exothermic energy transfer from n,π* triplet acetone to other
13
-1
39
10
s
have been measured for R-cleavage of triplet ketones,
in which mixing of adjacent n and σ orbitals makes the n,π*
states inherently dissociative. We suggest that the ∼10-fold
inefficiency for these bromo ketones should be interpreted as a
low transmission coefficient κ at the transition state, caused by
weak coupling and thus inefficient interconversion between
electronic states of different symmetry.
43
ketones and is well recognized as a cause of inefficient state
37,38
π
Br
conversions.
Interconversion with a n ,σ* state involves
only a single orthogonal electron transfer and forms the state
thought to be responsible for the photocleavage of iodobenzene
6
In unsubstituted and most para-substituted phenyl ketones,
the symmetry of the π orbitals is broken such that the lowest
π* orbital has the highest coefficient at the para carbon. Thus
the radical anions of benzenes with strong electron-withdrawing
substituents have most of their spin and charge density at the
and simple alkyl halides. Exact π-σ orthogonality and
symmetry restrictions may be broken by the halogen moving
off the axis of the phenyl-Br bond as it departs; but the final
phenyl radical is σ, so the extent of such wagging may be low.
Schemes 1and 2 depict these concepts as electron transfers
between states and between orbitals, respectively. The surface
crossings between solid lines displayed in Figure 11 are best
described as only partially avoided. Although the less than
optimal orbital overlap involved in both excited state intercon-
40
para position, with very little meta, just as observed in the
n,π* triplet of benzophenone.41 The radical anions of bro-
moacetophenones undergo rapid cleavage into phenyl radicals
and bromide ion. This cleavage, which must be driven by the
odd electron in the π* orbital, is some 100 times faster for para
versions could explain a low κ value, our suggestion that the
π
Br
bromines than for meta.36 The radical cleavage of 3n,π*
n ,σ* state is the one involved would mean that the bro-
iodobenzophenones shows the same para/meta ratio.1
1,12
mophenyl ketone triplets react by the same mechanism sug-
Be-
6
gested for iodobenzene itself, both involving irreversible state
cause the n orbital of the n,π* triplet is orthogonal to the π
system and thus mimics an independent donor, π* electron
density in the n,π* triplet is very similar to that in the radical
anion.
π
conversions of π,π* triplets to a n ,σ* state, which shares
Br
π
Br
with the n f π* state an electron-deficient bromine. In that
case the lack of any isomeric differences may be ascribed to
this partial positive charge on bromine, which may drive the
internal electron transfer strongly enough to override or to
The lowest π,π* triplets of phenyl ketones also have high
unpaired electron density at the para position with very little
(
42) (a) Hirota, N.; Wong, T. C.; Harrigan, E. T.; Nishimoto, K. Mol.
(
38) Dauben, W. G.; Salem, L.; Turro, N. J. Acc. Chem. Res. 1975, 8,
Phys. 1973, 29, 903. (b) Wagner, P. J.; May, M. J. Chem. Phys. Lett. 1976,
39, 350. (c) Wagner, P. J.; Siebert, E. J. J. Am. Chem. Soc. 1981, 103,
7329.
4
1.
(
39) Encina, M. V.; Lissi, E. A.; Lemp, E.; Zanocco, A.; Scaiano, J. C.
(43) (a) Turro, N. J.; Schore, N. E.; Steinmetzer, H. C.; Yekta, A. J.
Am. Chem. Soc. 1974, 96, 1936. (b) Mirbach, M. F.; Ramamurthy, V.;
Mirbach, M. J.; Turro, N. J.; Wagner, P. J. NouV. J. Chim. 1980, 4, 471.

Photoinduced Radical CleaVage of Bromophenyl Ketones
J. Am. Chem. Soc., Vol. 118, No. 4, 1996 753
3
energy. Thus, as shown in Figure 11, π,π* states can be
populated before the triplets acquire the energy required for
cleavage.
We include as dotted lines in Figure 11 the reaction path
1
1,12
previously suggested for iodobenzophenones,
in order to
depict why the activation energies for endothermic cleavage of
the bromoacetophenones and exothermic cleavage of the iodo-
benzophenones do not differ much. The endothermicity of the
bromine-phenyl bond cleavage means that the lowest excited
states do not intercept the dissociative state until the bond is
nearly broken, such that the transition state has nearly the same
energy as the product radical pair state. The iodobenzophenones
intercept a dissociative state at an energy well above that of
the product, so there is a significant activation energy in both
directions. The behavior of the bromo ketones raises an
interesting question akin to that posed by back electron transfer
between radical ion pairs: what is the probability that recom-
bination of bromine and acylphenyl radicals could generate
product in a triplet state?
High Reactivity of Ortho Isomers. Each of the o-bromo
ketones shows reactivity much higher than that of its meta and
para isomers, but in different ways. The high reactivity of
oBrBzP is entirely triplet derived and presumably reflects a
steric lowering of the C-Br bond energy. If the 600-fold rate
difference relative to the meta and para isomers is entirely
enthalpic, it would reflect a 4 kcal/mol decrease in bond energy.
The rate enhancement also could reflect a direct, through space
charge transfer interaction between the bromine atom and the
carbonyl, a process that is very rapid for o-thiyl ketones.44
Whatever the process is, it probably is not responsible for the
slow, temperature-independent decay of oBrA.
Figure 11. Potential energy diagram for activated conversion of n,π*
and π,π* triplets of halophenyl ketones into (s) C-Br and (- - -) C-I
dissociative states.
Scheme 2
The unquenchable reaction of oBrA most likely represents
some rapid cleavage from the excited singlet competing with
intersystem crossing. There is precedent for such behavior in
the photoenolization of o-alkylphenyl ketones; o-methylben-
zophenone reacts entirely from its triplet, whereas o-methyl-
acetophenone undergoes significant decay from its singlet.45 The
1
0-11
singlet rate constant would need to be in the range of 10
s
-1
and would indeed be higher than that estimated for the triplet
because of the higher energy of the excited singlet and thus
lower endothermicity of cleavage.
Summary. A combination of steady-state and time-resolved
studies indicates that m- and p-bromophenyl alkyl ketone triplets
undergo C-Br bond homolysis with 25 °C rate constants of 2
change regioelectronic differences in electron density. In terms
of orbital overlap, Scheme 2 indicates that significant spatial
overlap occurs between the π* and σ* orbitals on the carbon
8
8
-1
× 10 and 1 × 10 s , respectively. The two isomers appear
bearing bromine, because of the nπ f π* character in the
to have the same 5.3 kcal/mol activation enthalpy for cleavage
Br
1
2
q
and A values of 10 . The ∆H value barely equals the
endothermicity of cleavage as judged by the currently accepted
value of 79 kcal/mol for the phenyl-bromine bond dissociation
energy. The A values indicate some intrinsic inefficiency in
π,π* triplet for both m- and p-bromo ketones. This overlap
would facilitate conversion of the π,π* state to either dissocia-
tive state. In contrast, there is very little overlap between the
π
Br
σ and πA orbitals but strong overlap between the n and πA
triplet state cleavage, which is concluded to involve activated
orbitals; this difference is what we believe would favor
π
π
Br
conversion of a π,π* triplet into a dissociative n ,σ* state,
conversion to the n ,σ* state.
Br
π
Br
mixing of some n ,π* character into the π,π* state providing
Figure 11 depicts a potential energy diagram for both
the requisite coupling between electronic states. The same
mechanism would explain the even slower, more endothermic
cleavage of triplet bromobenzophenones. In both cases the
lowest n,π* triplets are assumed to convert into π,π* triplets
before cleavage. The isomeric o-bromo ketone triplets cleave
much more rapidly; a combination of steric acceleration and
singlet state participation is suggested.
processes, with avoided crossings between a π,π* triplet and
3
8
the two dissociative states indicated by dashed curves. The
various states of the bromo ketones are connected by solid lines,
with the interconversion of n,π* and π,π* triplets shown by
the former merging into the latter before intersection with a
dissociative state. We could have included separate curve
crossings for n,π* triplets, but did not for the reasons stated
above and because we do not know how the energy surfaces
for the two bound triplets differ with respect to CBr bond
stretching. Even though the bromobenzophenones have lowest
n,π* triplets, their π,π* triplets have energies similar to those
of the bromoacetophenones, lower than the phenyl-Br bond
The strong dependence of cleavage product quantum yields
on alkane solvent viscosity, together with the low absolute
(
44) Cao, Q. Ph.D. Thesis, Michigan State University, 1991.
(45) (a) Wagner, P. J.; Chen, C.-P. J. Am. Chem. Soc. 1976, 98, 239. (b)
Wagner, P. J. Pure Appl. Chem. 1977, 49, 259.

7
54 J. Am. Chem. Soc., Vol. 118, No. 4, 1996
Wagner et al.
+
+
values, indicates that large fractions of the initially formed
acylphenyl/bromine radical pairs recouple in the solvent cage
in competition with diffusing apart, despite their triplet origin.
Benzene minimizes this cage recombination, presumably by
complexing the bromine atoms.
130.0, 129.8, 128.5, 128.4, 122.6; MS m/z 262 (M , 30), 260 (M ,
3
2
1), 182.9 (23), 184.9 (21), 105 (100); UV (cyclopentane) λmax 214 (ꢀ
0 000), 249 (ꢀ 17 000), 346 nm (ꢀ 113).
o-Bromobenzophenone. o-Bromobenzoic acid (5.1 g) was con-
verted to it acid chloride, which was used to acylate benzene as
described for the meta isomer. Normal workup yielded a white solid
which was recrystallized from diethyl ether/hexane to afford prisms of
o-bromobenzophenone (3.96 g, 60%), mp 40-41 °C (lit. mp 41.5
) δ 7.82 (dd, J ) 8,
Hz, 2H), 7.63 (dd, J ) 8, 1 Hz, 1H), 7.59 (tt, J ) 8, 1 Hz, 1H), 7.45
t, J ) 8 Hz, 2H), 7.32-7.43 (m, 3H); C-NMR (CDCl
40.6, 136.0, 133.6, 133.1, 131.1, 130.1, 128.9, 128.5, 127.1, 119.4;
MS m/z 262 (M , 13), 260 (M , 14), 184.9 (19), 182.9 (21), 105 (100);
UV (cyclopentane) λmax 208 (ꢀ 18 000), 248 (ꢀ 16 000), 343 nm (ꢀ
2).
All three bromoacetophenones and p-bromobenzophenone were
purchased from Aldrich and purified by recrystallization (para) or
vacuum distillation (ortho and meta). These and the synthesized ketones
were judged to be >99% pure from their NMR spectra and GC analysis.
Identification of Photoproducts. Preparative scale irradiation of
pBrV in benzene and preparative GC isolation provided three products.
Valerophenone and p-bromoacetophenone were identified by compari-
son to authentic samples; p-phenylvalerophenone was also identified
by comparison of its spectroscopic features with those of the authentic
Experimental Section
4
8
Chemicals. Benzene solvent was commercial thiophene-free reagent
grade purified by washing with sulfuric acid and then distillation from
phosphorus pentoxide. Alkanes used as internal standards for GC
analysis had previously been vacuum distilled and passed through
alumina; the solvents were acid washed and then distilled. Naphthalene
was purified by sublimation. 2,5-Dimethyl-2,4-hexadiene and 1,3-
pentadiene were obtained from Chemical Samples Co.; the former was
allowed to sublime in the refrigerator.
-1 1
°
C); IR (KBr) 1663, 1292 cm ; H-NMR (CDCl
3
1
(
13
3
) δ 195.7,
1
+
+
9
m-Bromovalerophenone was prepared in 70% yield from the
Grignard reagent of m-dibromobenzene and valeronitrile, following
standard procedures including isolation of the imine salt in cold water
46
before its hydrolysis. It was purified by recrystallization from hexane,
-
1
1
mp 43 °C: IR (KBr) 2985, 1685, 1450, 1210, 780 cm ; H-NMR
CDCl ) δ 0.94 (t, J ) 7.5 Hz, 3 H), 1.50 (m, 4 H), 2.86 (t, J ) 8 Hz,
H), 733 (t, J ) 7.82 Hz, 1 H), 7.67 (ddd, 8.0, 1.92, 1.08 Hz, 1 H),
(
3
2
7
.86 (dd, J ) 7.68, 1.42, 1.20 Hz, 1 H), 8.06 (t, J ) 1.78 Hz, 1 H);
MS m/z 242, 240, 200, 198, 185, 183, 157, 153.
1
material synthesized from biphenyl and valeryl chloride: H-NMR
CDCl ) δ 0.94 (t, 3 H), 1.52 (m, 4 H), 2.94 (t, 2 H), 7.1-8.2 (9 H);
3
p-Chlorobutyrophenone was prepared by Friedel-Crafts acylation
of chlorobenzene with butyryl chloride. The acid chloride was added
to a slurry of aluminum chloride in chlorobenzene and allowed to stir
at 0-5 °C for 1.5 h. Normal aqueous workup and distillation, followed
by recrystallization from hexane, provided an 85% yield of white
(
MS m/z 238, 196, 181, 152. A fourth and minor product peak was
apparent on the GC traces but not isolated; it was assumed to be the
cyclobutanol that always accompanies Norrish type II elimination.
Photoproducts from the other ketones were readily identified by
comparison to authentic samples. Dibromine was revealed by its red
color. In the case of pBrA in cyclohexane, the reaction was followed
by GC-mass spectral analysis, which revealed the presence of both
cyclohexyl bromide and bicyclohexyl. GC analysis indicated a 96%
yield of acetophenone and 30% of cyclohexyl bromide. The presence
of HBr in the gas above the liquid sample was deduced by the color
change produced on wet litmus paper. A similar experiment with
o-bromobenzophenone gave a 92% chemical yield of benzophenone.
Fluorenone as well as cyclohexyl bromide were identified by GC-MS.
Procedures. Samples were prepared and irradiated as usual.33
Yields of products from the halo butyro- and valerophenones were
measured by GC analysis. Yields of reduced ketones were measured
by HPLC analysis. Phosphorescence spectra were obtained on a Perkin-
Elmer MPF-44A spectrophotometer on samples 10 M in ketone, with
313-nm excitation. Flash kinetics were performed on samples ∼0.005
M (o.d. ∼ 0.6 cm) in 7 × 7 mm square Supracil cells fitted with serum
caps that were first deaerated with nitrogen. Spectroscopic grade
methanol or reagent grade toluene were used as solvents. Excitation
was with a Molectron U-24 nitrogen laser (337 nm, 8 ns pulse, ∼8
mJ) on J. C. Scaiano’s apparatus.49 Triplet decay was monitored at
400 or 410 nm where there was little residual absorption.
-
1
crystals, mp 37.0 °C: IR (KBr) 2980, 1685, 1590, 1400, 1310 cm
;
1
H-NMR (CDCl ) δ 1.00 (t, J ) 7.5 Hz, 3 H), 1.78 (quin, J ) 7.8 Hz,
H), 2.88 (t, J ) 8 Hz, 2 H), 7.60 (AB quar, 4 H); MS m/z 184, 182,
56, 154, 147, 141, 139, 113, 11.
3
2
1
p-Bromobutyrophenone was prepared similarly from bromobenzene
in 80% yield, mp 36 °C: IR (KBr) 2980, 1685, 1400, 825 cm-1; H-
1
NMR (CDCl ) δ 1.00 (t, J ) 7.5 Hz, 3 H), 1.90 (quin, J ) 7.8 Hz, 2
3
H), 2.84 (t, J ) 8 Hz, 2 H), 7.50, 7.72 (AB quar, J ) 8.46 Hz, 4 H);
MS m/z 228, 226, 200, 198, 185, 183, 157, 155, 147.
p-Bromovalerophenone was prepared similarly from bromobenzene
and valeryl chloride in 80% yield, mp 36 °C: IR (KBr) 2980, 1685,
-
1 1
1
(
4
3
600, 1010 cm ; H-NMR (CDCl ) δ 0.92 (t, J ) 7.5 Hz, 3 H), 1.50
m, 4 H), 2.92 (t, J ) 8 Hz, 2 H), 7.51, 7.73 (AB quar, J ) 8.46 Hz,
H); MS m/z 242, 240, 200, 198, 185, 183, 157, 155.
-3
m-Bromobenzophenone. A suspension of m-bromobenzoic acid
(5.1 g) and thionyl chloride (8 mL) in chloroform (50 mL) was refluxed
for 3 h. The crude acyl chloride was dissolved in benzene (20 mL)
and added dropwise to a solution of aluminum chloride (3.8 g) in
benzene (40 mL). After 5 h of reflux the reaction was quenched by
water and concentrated HCl. Extraction with ethyl acetate and normal
workup yielded an oil which was chromatographed on a silica gel
column (10% ethyl acetate in hexane) and then crystallized from
47
methanol to give white crystals (5.3 g, 80%), mp 75-76 °C (lit. mp
-
1 1
7
6-77 °C): IR (KBr) 1659, 1278 cm ; H-NMR (CDCl
3
) δ 7.92 (t,
Acknowledgment. This work was supported by continuing
grants from the National Science Foundation. We thank Prof.
J. C. Scaiano for providing time on his flash apparatus at the
NRCC and at the University of Ottawa.
J ) 1 Hz, 1H), 7.77 (dd, J ) 8, 1 Hz, 2H), 7.70 (dd, J ) 8, 1 Hz, 2H),
7
1
.60 (tt, J ) 8, 1 Hz, 1H), 7.48 (t, J ) 8 Hz, 2H), 7.36 (t, J ) 8 Hz,
13
H); C-NMR (CDCl
3
) δ 195.0, 139.5, 136.9, 135.2, 132.8, 132.7,
(
46) (a) Hauser, C. R.; Humphlett, W. J.; Weiss, M. J. J. Am. Chem.
JA952782F
Soc. 1948, 70, 426. (b) Wagner, P. J.; Kemppainen, A. E. J. Am. Chem.
Soc. 1972, 94, 7495.
(
47) Allen, R. E.; Schumann, E. L.; Day, W. C. J. Am. Chem. Soc. 1958,
(48) Ogata, Y.; Tsuchida, M. J. Org. Chem. 1955, 20, 1631.
(49) Scaiano, J. C. J. Am. Chem. Soc. 1980, 102, 7747.
8
0, 591.