Comparative study of the reactivities of substituted 3-(benzoyl)benzyl carbanions in water and in DMSO

Article abstract of DOI:10.1021/jo049224g

Benzyl-substituted carbanions produced by photodecarboxylation of ketoprofen derivatives have been examined in basic aqueous and DMSO solutions. Product studies, combined with kinetic measurements from laser flash photolysis, have allowed the determinatio

Full text of DOI:10.1021/jo049224g

Com p a r a tive Stu d y of th e Rea ctivities of Su bstitu ted
3-(Ben zoyl)ben zyl Ca r ba n ion s in Wa ter a n d in DMSO
Laura Llauger,† Miguel A. Miranda,‡ Gonzalo Cosa,† and J . C. Scaiano*,†
Department of Chemistry, University of Ottawa, 10 Marie-Curie, Ottawa, Ontario K1N 6N5, Canada, and
Departamento de Quı´mica/ Instituto de Tecnologı´a Quı´mica UPV-CSIC, Universidad Polite´cnica de
Valencia, Avenida de los Naranjos s/ n, 46022-Valencia, Spain
tito@photo.chem.uottawa.ca
Received May 8, 2004
Benzyl-substituted carbanions produced by photodecarboxylation of ketoprofen derivatives have
been examined in basic aqueous and DMSO solutions. Product studies, combined with kinetic
measurements from laser flash photolysis, have allowed the determination of absolute rate constants
for protonation and intra-SN2 reactions leading to five- and six-membered ring cyclizations; the
former are significantly faster. Many of the well-known trends in carbanion reactivity are placed
on an absolute rate basis; thus, intra-SN2 are favored in polar nonprotic solvents, and the effect is
larger for the more hindered carbanion centers. Protonation by water is slightly dependent on the
nature of the carbanion center and is ∼400 times faster in nonhydroxylic solvents, compared with
bulk water. As expected, the reactivity for halide leaving groups follows the usual order of decreasing
bond strengths, i.e., I- > Br- > Cl-.
SCHEME 1. Mech a n ism of th e Heter olytic
P h otod eca r boxyla tion
In tr od u ction
The possibility of generating carbanions or other
reactive intermediates within the duration of a nano-
second laser pulse provides a very powerful method to
perform mechanistic studies and learn about their
reactivity.1-10 Laser excitation of suitable precursors
leads to essentially instantaneous carbanion formation
in an irreversible process, thus making their transient
decays, and the formation of photoproducts, a kinetically
controlled reaction. Employing picosecond and nano-
second flash photolysis techniques, Craig et al.1,2 inves-
tigated the dynamics of the photodecarboxylation of the
p-nitrophenylacetate in aqueous media. Similarly, Wan
et al.7,11 studied the photodecarboxylation of aroyl-
substituted phenylacetic and diarylacetic acids, giving in
most cases the corresponding carbanion. In recent re-
ports,8,9 we explored the reactivities of aroyl-substituted
benzyl carbanions formed following photodecarboxylation
of a series of ketoprofen-derived compounds in protic and
aprotic solvents.
Mechanistically, these photodecarboxylations involve
C-C bond heterolysis from the excited state, giving rise
to a benzylic carbanion intermediate (see Scheme 1). We
have proposed that this reaction involves the singlet
state.6,12
Following decarboxylation, this ionic intermediate
reacts promptly with its microenvironment. In the sys-
tems studied here, this leads to protonation or cyclization,
although in the presence of other scavengers it can lead
to the formation of a wide variety of final products.13 In
this paper, we examine the ability of photogenerated
aroyl-substituted benzyl carbanions to cyclize in protic
and aprotic media, via an intramolecular nucleophilic
* To whom correspondence should be addressed. Phone: 1 613
5625896. Fax: 1 613 5625170.
† University of Ottawa.
‡ Universidad Polite´cnica de Valencia.
(1) Craig, B. B.; Weiss, R. G.; Atherton, S. J . J . Phys. Chem. 1987,
91, 5906-5912.
(2) Craig, B. B.; Atherton, S. J . J . Chem. Soc., Perkin Trans. 2 1988,
1929-1935.
(3) Kresge, A. J . Science 1991, 253, 395-400.
(4) McClelland, R. A. Tetrahedron 1996, 52, 6823-6858.
(5) Mart´ınez, L. J .; Scaiano, J . S. J . Am. Chem. Soc. 1997, 119,
11066-11070.
(6) Cosa, G.; Mart´ınez, L. J .; Scaiano, J . C. Phys. Chem. Chem. Phys.
1999, 1, 3533-3537.
(7) Xu, M.; Wan, P. Chem. Commun. 2000, 2147-2148.
(8) Cosa, G.; Llauger, L.; Scaiano, J . C.; Miranda, M. A. Org. Lett.
2002, 4, 3083-3085.
(9) Llauger, L.; Cosa, G.; Scaiano, J . C. J . Am. Chem. Soc. 2002,
124, 15308-15312.
(12) Laferrie`re, M.; Sanrame´, C. M.; Scaiano, J . C. Org. Lett. 2004,
6, 873-875.
(13) Chre´tien, M. N.; Cosa, G.; Garc´ıa, H.; Scaiano, J . C. Chem.
Commun. 2002, 2154-2155.
(10) Borsarelli, C. D.; Braslavsky, S. E.; Sortino, S.; Marconi, G.;
Monti, S. Photochem. Photobiol. 2000, 72, 163-171.
(11) Krogh, E.; Wan, P. J . Am. Chem. Soc. 1992, 114, 705-712.
10.1021/jo049224g CCC: $27.50 © 2004 American Chemical Society
Published on Web 09/10/2004
7066
J . Org. Chem. 2004, 69, 7066-7071

Reactivities of Substituted 3-(Benzoyl)benzyl Carbanions
SCHEME 2. Rea ction of P h otogen er a ted Ar oyl-Su bstitu ted Ben zyl Ca r ba n ion s
SCHEME 3. Syn th esis of th e Ar oyl-Su bstitu ted Ben zylca r boxylic Acid s
TABLE 1. Qu a n tu m Yield s of P h otod eca r boxyla tion of
CHART 1. Set of Molecu les Stu d ied
1-4 a n d Deca y Ra te Con sta n ts for th e P h otogen er a ted
Ca r ba n ion s of 1-4 in 0.1 M KOH Solu tion s
compd
Φphotodec
kobs/s-1
1a (X ) Cl; n ) 4)
1b (X ) Br; n ) 4)
1c (X ) I; n ) 4)
2a (X ) Cl; n ) 4)
2b (X ) Br; n ) 4)
2c (X ) I; n ) 4)
3a (X ) Cl; n ) 5)
3b (X ) Br; n ) 5)
3ca (X ) I; n ) 5)
4a (X ) Cl; n ) 5)
4b (X ) Br; n ) 5)
4ca (X ) I; n ) 5)
0.78
0.70
0.82
0.82
0.77
0.78
0.68
0.62
0.5
(6.99 ( 0.04) × 106
(6.9 ( 0.1) × 106
(7.61 ( 0.07) × 106
(2.57 ( 0.02) × 106
(2.77 ( 0.02) × 106
(3.20 ( 0.04) × 106
(5.98 ( 0.05) × 106
(6.25 ( 0.08) × 106
(6.73 ( 0.05) × 106
(2.76 ( 0.01) × 106
(2.62 ( 0.01) × 106
(2.59 ( 0.01) × 106
0.6
0.75
0.8
attack of the carbanion to an eletrophilic center, as
illustrated in Scheme 2.
a
From ref 9.
When the leaving group (X) was iodide (I) and n ) 5,
we observed that the intra-SN2 reaction was favored in
aprotic media, such as dimethyl sulfoxide (DMSO), while
protonation was the dominant reaction in basic aqueous
media; elimination was not observed.9 The goal of this
paper is to quantify structure-reactivity trends in car-
banion chemistry in systems where the direct determi-
nation of kinetic data is feasible; our studies are based
on the photodecarboxylation for a series of ketoprofen-
derived compounds (Chart 1).
To study the nucleophilic behavior of the photogener-
ated carbanions, and given the fast protonation of these
species in water,9 we compared hindered (by a methyl
group) to unhindered nucleophilic sites to evaluate steric
effects and different leaving groups (I, Br, and Cl)
attached at the end of the alkyl chain, which are likely
to have a major influence on the reaction dynamics.
Further, we analyzed the effect of the length of the lateral
alkyl chain on the cyclization reaction. All the experi-
ments were carried out in aqueous solutions containing
0.1 M potassium hydroxide (to ensure only the carboxy-
late is present) or in dry, basic DMSO.
F IGURE 1. Transient absorption spectra following 308 nm
laser excitation of N2O-saturated 0.1 M KOH solution of 0.4
mM 1b (+), 2a (0), and 4b ([) 64 ns after the laser pulse and
1c (b) 32 ns after the pulse.
of compounds 1b, 1c, 2a , and 4b were similar to that for
the well-characterized carbanion from ketoprofen (see
Figure 1).6,8,10,14 The visible part of the spectrum was
common to all species, with a broad band located at ca.
590 nm.
Resu lts
Carboxylic acids 1-4 were prepared in high yields in
two steps: alkylation of esters 5a ,b with the correspond-
ing 1,n-dihalocompound (1,n-DHC, n ) 4 or 5) to give
the intermediate tert-butylcarboxylic esters 6-9, followed
by removal of the tert-butyl group by titanium tetrachlo-
ride (see Scheme 3).9
The decay for the resulting carbanions from 1 to 4,
recorded at 600 nm, followed first-order kinetics and led
to the rate constants listed in Table 1. Quantum yields
for photodecarboxylation for compounds 1-4 were esti-
mated using the reaction of ketoprofen as relative acti-
The transient absorption spectra obtained following
308 nm laser excitation of aqueous 0.1 M KOH solutions
(14) Monti, S.; Sortino, S.; De Guidi, G.; Marconi, G. J . Chem. Soc.,
Faraday Trans. 1997, 93, 2269-2275.
J . Org. Chem, Vol. 69, No. 21, 2004 7067

Llauger et al.
CHART 2. P h otop r od u cts of Ca r boxylic Acid s 1a -c, 2a -c, 3a -c, a n d 4a -c
TABLE 2. Absolu te Ra te Con sta n ts for P r oton a tion (kW) a n d Cycliza tion (kC) of P h otogen er a ted Ca r ba n ion s fr om 1-4
in DMSO a n d Qu en ch in g Ra te Con sta n ts by Wa ter (kQu en ) in Solu tion s of 1-4 in DMSOa
c
c
d
compdb
fC
fW
kobs (s-1
)
kC (s-1
)
kW (s-1
)
kquen (M-1 s-1
3.3 × 107
)
1a (X ) Cl; n ) 4)
1b (X ) Br; n ) 4)
1c (X ) I; n ) 4)
2a (X ) Cl; n ) 4)
2b (X ) Br; n ) 4)
2c (X ) I; n ) 4)
3a (X ) Cl; n ) 5)
3b (X ) Br; n ) 5)
3cf (X ) I; n ) 5)
4a (X ) Cl; n ) 5)
4b (X ) Br; n ) 5)
4cf (X ) I; n ) 5)
13 (14)
100 (14)
100 (14)
43 (15)
100 (15)
100 (15)
87 (10a )
57 (11a )
(2.2 ( 0.1) × 106
2.8 × 105
3.6 × 107
>5 × 107
4.0 × 105
>5 × 107
>5 × 107
<6.0 × 104e
6.1 × 105
4.9 × 106
<1.1 × 104e
3.1 × 105
1.9 × 106
(1.9 × 106)
(3.6 ( 0.1) × 107
(<7.3 × 105)d
(9.2 ( 0.1) × 105
(5.2 × 105)
7.8 × 106
100 (11a )
83 (11b)
18 (11a )
100 (13a )
58 (13b)
2 (13c)
(3.0 ( 0.1) × 106
(3.6 ( 0.1) × 106
(6.0 ( 0.1) × 106
(5.5 ( 0.1) × 105
(7.4 ( 0.1) × 105
(2.0 ( 0.1) × 106
(3.0 × 106)
(3.0 × 106)
(1.1 × 106)
5.5 × 105
6.9 × 107
5.5 × 107
5.1 × 107
1.0 × 107
8.3 × 106
8.8 × 106
17 (16)
82 (16)
-
42 (17)
98 (17)
(4.3 × 105)
(3.9 × 104)
a
b
Dried with CaH2, distilled under low pressure and kept in 4 Å sieves (base used NaH). In general, the samples were measured as
soon as possible following preparation, given their reduced stability in the presence of NaH. c Percent fraction of cyclization and protonation.
d
The values of kW are given in parentheses since some variation is anticipated due to day-to-day fluctuations in the water content of
“dried” DMSO. e Estimated values, considering that a 2% of photoproduct is the lower limit detectable by our HPLC. Data from ref 9.
f
nometer (Φ ) 0.75 at pH 7.415), and the photolysis was
limited to less than 10% conversion to minimize light
absorption and reaction of photoproducts. The high
values of these quantum yields (Table 1) confirm that
decarboxylation of ketoprofen and derivatives is a very
efficient photochemical reaction. The decay rate constants
can be essentially regarded as protonation rate constants,
since this is the dominant carbanion decay process.
When 4 mM 1-4 in 0.1 M KOH aqueous solutions were
irradiated, we quantitatively obtained the corresponding
protonated photoproducts 10-13 (see Chart 2), with the
exception of compounds 3c and 4c, which also gave the
cyclic photoproducts 16 (7%) and 17 (3%), respectively,
formed by intramolecular nucleophilic attack (intra-SN2).9
F IGURE 2. Transient absorption spectra following 355 laser
excitation of Ar-saturated 10 mM solutions in NaH/DMSO of
2a (b), 3c (×), and ketoprofen (4), recorded at 200, 72, and 72
ns after the laser pulse.
Taking into consideration the overall rate constant for
carbanion decay, the rate constants for cyclization for 3c
and 4c were estimated as 4.7 × 105 and 7.7 × 104 s-1
,
respectively (vide infra).
observed in this solvent following laser flash photolysis
of ketoprofen and 3c9 (see Figure 2). There is literature
precedent for solvent-induced shifts in the carbanion
spectra.9
The decay for carbanions photogenerated from 10 mM
1-4 solutions in NaH/DMSO, recorded at 390 nm,
followed first-order kinetics (see Table 2). They were
recorded at this wavelength since the intensity of the
signal at 800 nm was very weak. A few control experi-
ments showed that the 390 and 800 nm signals followed
the same kinetics.
Kinetics studies of compounds 1-4 in DMSO solutions,
containing an excess (8:1 equiv) of sodium hydride (NaH),
were measured following 355 nm excitation using a front-
face geometry, since the signal associated with the
carbanion, generated in an aprotic medium, was better
resolved under these exposure conditions. Considering
both the structural similarity and reactivity of carboxylic
acids 1-4 (shown above), we used compound 2a as a
model. Laser excitation of a solution of 2a in DMSO gave
a transient absorption spectrum with bands around 390
nm and at approximately 800 nm (weak), similar to those
Product studies were carried out under extremely dry
conditions (water < 0.03%9). The short lifetimes associ-
ated with transient carbanions, generated from photo-
(15) Costanzo, L. L.; De Guidi, G.; Condorelli, G.; Cambria, A.; Fama,
M. Photochem. Photobiol. 1989, 50, 359-365.
7068 J . Org. Chem., Vol. 69, No. 21, 2004

Reactivities of Substituted 3-(Benzoyl)benzyl Carbanions
decarboxylation of compounds 1-4, usually include a
contribution from reaction with traces of water, in
addition to the expected cyclization. Photolysis of 4 mM
1-4 in DMSO to high conversion (10%), in the presence
of an excess of NaH, gave mixtures of cyclic and pro-
tonated photoproducts (see Table 2). It is likely that the
excess NaH also behaves as a sacrificial drying agent
helping reduce the concentration of traces of water.
Considering that cyclization and protonation are the only
two decay pathways available to the photogenerated
carbanion, the cyclized and protonated absolute rate
constants, kC and kW, respectively, were calculated con-
sidering cyclized and protonated carbanion fractions (fC
and fW) determined by HPLC and the observed rate
constant (kobs), as defined eqs 1 and 2:
spectroscopy of the carbanions was largely unaffected by
remote halogen substitution, and for precursors 3c and
4c, that cyclization could be favored over protonation by
use of dry DMSO in the presence of NaH as a strong base.
Our results in this paper greatly expand the reactions
examined, in particular by examining the role of different
leaving groups and by exploring the formation of five-
and six-membered rings. The results are consistent with
our earlier reports where the data overlap.
Photodecarboxylation in aqueous solutions gave pre-
dominantly the protonated product. As observed in
earlier studies,8,9 protonation rate constants were influ-
enced by steric effects. Cyclization yields were very low,
reflecting rapid carbanion protonation. The only cases
where cyclization was detected in aqueous solution were
compounds 3c and 4c, which yielded the corresponding
cyclic products in a 7% (16) and 3% (17), respectively. In
contrast, when photodecarboxylation was carried out in
DMSO, the main process was cyclization through an
intra-SN2 reaction.
kobs ) kC + kW
(1)
(2)
kC ) fCkobs kW ) fWkobs
We have estimated the rate constants for cyclization
in DMSO using eqs 1 and 2. In those cases where
cyclization was too fast to prevent detection of the
carbanion, a lower limit of >5 × 107 s-1 was estimated
for the rate constant. In the cases of compounds 3c and
4c, it was possible to determine and compare cyclization
rate constants in both aqueous and DMSO solutions. For
Values of kW in Table 2 are given in parentheses, since
this value is influenced by variations in the trace
amounts of water contained in the “dry” solvent.
As illustrated in Table 2, formation of cyclic products
14-17 was enhanced in DMSO, confirming that SN2-type
reactions are favored in aprotic polar solvents, compared
to protic solvents (see values for water in Table 1). In
cases such as 1c, 2b, and 2c, the cyclization took place
so fast that the nanosecond laser system response was
too slow to measure the kinetics for the corresponding
carbanion reactions.
To evaluate the water effect on the photolysis of 1-4
in DMSO, we examined the quenching of the carbanions
by water, expanding earlier work9 where we also had
determined the kinetic isotope solvent effect when we
measured the quenching rate constant for compound 3c
by H2O and D2O.
3c, the values were 4.7 × 105 and 4.9 × 106 s-1
,
,
respectively, and for 4c 7.7 × 104 and 1.9 × 106 s-1
respectively. These values showed a reduction of the
cyclization rate constant in hydrogen-bonding solvents.
The effect was larger for 4c, the more sterically encum-
bered carbanion. Thus, the decrease in the importance
of cyclization in protic media is not just the result of the
competition with the solvent (55 M water in our case)
but also the result of a significant reduction in the value
of the cyclization rate constant.
The ease with which carbanions were scavenged even
by trace amounts of water provided an indication that
water was a much better quencher in DMSO than in bulk
water, as already determined in comparison of water-
quenching rates in bulk water and in nonprotic media.9
Considering the cyclization as an intra-SN2 reaction
(i.e., while an intramolecular reaction, mechanistically
equivalent to the SN2 process), the nature of the leaving
group is expected to have a significant effect on the rate
constants. When the intra-SN2 reaction involved chloride
as the leaving group, protonation was systematically
favored, despite our effort to obtain truly dry DMSO; in
earlier work we estimated the water content of “dry”
DMSO at around 0.03%, or approximately 16 mM, under
our experimental conditions. Remarkably, the photo-
reaction of compounds 1a and 2a yielded the five-
membered ring products 14 and 15 in a 13% and 43%,
respectively. Not surprisingly, when bromide or iodide
were the leaving groups, the cyclization reaction was
favored with iodide giving highest cyclization yields.
Thus, the reactivity order for the halide leaving groups
was I- > Br- > Cl-, with differences of approximately 1
order of magnitude between pairs (see Table 2).
Discu ssion
It is well-known that factors such as the nature of the
nucleophile, solvent, and leaving group directly affect the
rate of the bimolecular nucleophilic substitution (SN2)
reactions; yet, in the case of carbanions, little has been
documented with absolute rate constants. The rapid and
efficient decarboxylation of ketoprofen and related sub-
stituted benzophenones offers a simple photochemical
route to benzyl carbanions in solution. While there has
been some discussion in the literature as to whether the
decarboxylation pathway is singlet- or triplet-mediated,
there is no doubt that the carbanion is formed in the
subnanosecond time scale and the observable species is
the singlet ground-state carbanion. In addition to fast
time-resolved experiments, recent studies show that it
is possible to lengthen the carbanion lifetime to many
minutes in tetrahydrofuran,12 thus confirming that the
species observed is the ground state.
Laser flash photolysis of substrates 1-4 takes advan-
tage of the fast decarboxylation of the corresponding
carboxylates under strong basic conditions; this provides
a nearly-instantaneous source of carbanions suitable for
detailed kinetic and mechanistic studies.
Considering the data from Table 2, cyclization to form
a five-membered ring (compounds 1 and 2) was at least
1 order of magnitude faster than for six-membered rings
(compounds 3 and 4). Entropy and stereoelectronic effects
In earlier studies,9 we showed that steric effects were
important in carbanion protonation reactions, that the
J . Org. Chem, Vol. 69, No. 21, 2004 7069

Llauger et al.
flash columns, 230-400 mesh. Melting points were measured
probably favors five-membered ring closure. According
to Ruzicka’s hypothesis,16 (a) the probability of end-to-
end interaction in a bifunctional linear precursor, which
diminishes as the chain grows longer, and (b) ring strain
are two independent factors in determining the ease of
cyclization. In terms of entropy, these processes involve
negative ∆Sq contributions due to reduction of internal
rotation freedom around the single bonds of the molecular
backbone when the open-chain precursor converts to the
ring-shaped transition state. In our case, there was an
additional loss of one internal rotation in order to achieve
the transition state for the six-membered ring, compared
with a five-membered ring. Ring strain was not a key
factor in our reaction system since cyclization of the five-
membered ring, which should have a more strained ring,
was favored over the six-membered ring.
In a previous paper,9 we pointed out that the presence
of a methyl group attached directly to the carbanion
center hindered the cyclization reaction. Interestingly,
in the case of a five-membered ring closure addition of a
methyl group at the carbanion center led to a modest
increase in the intra-SN2 reaction rate constant (compare
kC of series 1 and 2). Clearly, the enhanced nucleophi-
licity of the carbanion center is more important than
steric contributions in some cases. In contrast, the
presence of the methyl group in the formation of a six-
membered ring caused the cyclization rate constant to
decrease by approximately a factor of 2. We speculate
that this may force the halide to leave from a near-axial
orientation.
on
a
MELT-TEMP II apparatus. 1H NMR spectra were
obtained in CDCl3 at 300 MHz, with TMS as an internal
standard, and 13C NMR spectra were obtained at 75.5 MHz,
unless otherwise noted. IR spectra were recorded as films on
NaCl plate. Elemental analyses were performed by MHW
laboratories (Phoenix, AZ).
Na n osecon d La ser F la sh P h otolysis. Similar laser flash
photolysis systems have been previously described.17,18 The
laser flash photolysis systems used a Lumonics EX-530 laser
(Xe-HCl-Ne mixture, 308 nm, ca. 6 ns and <100 mJ /pulse), or
the third harmonic of a Surelite Nd:YAG laser generating 355
nm pulses of 8 ns duration and 20 mJ /pulse output. The signals
from the monochromator/photomultiplier system were initially
captured by a Tektronix 2440 digitizer and transferred to a
Power Macintosh computer that controlled the experiment
with software developed in the LabVIEW 5.1 environment
from National Instruments. All the transient spectra and
kinetics were recorded by employing 7 × 7 mm2 Suprasil
quartz cells. Samples were usually deaerated by bubbling Ar
(DMSO) or N2O (KOH); the latter (N2O) was used to eliminate
possible interference from hydrated electrons that are other-
wise long-lived in aqueous systems. Samples that contained
NaH were first flushed for 15 min with argon just before
DMSO was added. To have a good signal in laser experiments,
the concentration of the samples was ca. 8 mM in DMSO (355
nm excitation) and 0.4 mM in aqueous solution (308 nm
excitation). The quenching rate constants were obtained using
static samples.
The stability of samples of substituted benzophenones in
the presence of NaH has been discussed in a recent publica-
tion.12 Briefly, while sample aging eventually leads to reduc-
tion, these samples are sufficiently stable for laser and product
studies to be conducted.
Ma ter ia ls. Commercially available organic compounds
(from Aldrich unless otherwise indicated) were used without
further purification except for the solvents. Tetrahydrofuran
(THF, Omnisolv) was freshly distilled from sodium, CH2Cl2
from calcium hydride, and DMSO from calcium hydride at
reduced pressure and stored on 4 Å molecular sieves. Starting
materials (3-benzoylphenyl)acetic acid (Karl Industries) and
ketoprofen [2-(3-benzoylphenyl)propionic acid] (Sigma) were
commercially available. Compounds 3c, 4c, 5a , 5b, 7c, and
9c were prepared according to reported methods.9 Photoprod-
ucts 12c, 13c, 16, and 17,1 as well as 14,8 were known. Details
of synthesis and characterization are provided in the Support-
ing Information.
P r od u ct Stu d ies. Steady-state photolysis studies were
carried out in a Luzchem photoreactor, employing 7 LZC-UVB
lamps (300 nm).19 The Ar-deaerated 4mM NaH/DMSO and N2-
deaerated 4 mM 0.1 M KOH solutions of 1a -c, 2a -c, 3a ,b,
and 4a ,b were irradiated for 15 min at ∼300 nm. Argon or N2
were used to eliminate oxygen; the choice of either one was
based on availability in different laboratories. To identify the
starting material, as well as the photoproducts produced, the
eluent absorbances at 254 nm were compared before and after
photolysis, given that both products and starting materials
have the same molar absorptivity. The photolysis mixture was
analyzed by a Varian HPLC system, equipped with a reversed-
phase 4.6 × 250 mm analytical Zorbax SB-C18 column. The
mobile phase was 15:85 water/methanol and the flow rate 0.5
mL/min. The detection used a Varian 9065 Polychrom diode
array detector.
Con clu sion
The kinetic and mechanistic studies presented here
provide absolute kinetic parameters for intra-SN2 and
protonation reactions of benzyl carbanions. Many of the
concepts that are normally taught in introductory organic
chemistry can now be illustrated with absolute rate
constants. Thus, intra-SN2 reactions are favored in polar
nonprotic solvents, and the effect is larger for the more
hindered carbanion centers. Protonation by water is
slightly dependent on the alkyl groups attached directly
on the carbanion center. In addition, the water protona-
tion rate constant is about 400 times faster in nonhy-
droxylic solvents, compared with bulk water. As expected,
the reactivity for halide leaving groups follows the usual
order of decreasing bond strengths, i.e., I- > Br- > Cl-.
Photoinduced decarboxylation of suitable substituted
carbanions provides a route for the formation of substi-
tuted cycloalkanes that proceeds in high yields in non-
hydroxylic solvents and with good leaving groups such
as bromide and iodide.
Exp er im en ta l Section
Gen er a l Meth od s. All reactions were conducted in an inert
atmosphere (N2 or Ar). Photoreactions were carried out in
Pyrex glassware. Combined organic extracts were dried over
anhydrous Na2SO4. Solvents were removed from the reaction
mixture or combined organic extracts by using a rotary
evaporator. Reactions were monitored by thin-layer chroma-
tography (TLC) on precoated E. Merck silica gel 60 F254 plates.
Column chromatography was carried out with silica gel for
Qu a n tu m Yield Deter m in a tion . The N2-deaerated 4 mM
0.1 M KOH solutions of 1a -c, 2a -c, 3a ,b, and 4a ,b were
irradiated at 300 nm for 1 min to ensure conversions lower
than or equal to 10%. A 0.1 M KOH solution of ketoprofen
(17) Scaiano, J . C. J . Am. Chem. Soc. 1980, 102, 7747.
(18) Scaiano, J . C.; Tanner, M.; Weir, D. J . Am. Chem. Soc. 1985,
107, 4396.
(19) For detailed lamp spectral information, see: http://
www.luzchem.com/handbook/exp_standards.html.
(16) Galli, C.; Mandolini, L. Eur. J . Org. Chem. 2000, 3117-3125.
7070 J . Org. Chem., Vol. 69, No. 21, 2004

Reactivities of Substituted 3-(Benzoyl)benzyl Carbanions
was employed as a standard (the quantum yield of photode-
carboxylation is 0.75 at pH 7.4).15 To quantify the starting
material, as well as the photoproducts, a calibration curve was
made in the HPLC with different KP standard solutions,
registering the signal at 254 nm.
Ontario Graduate Scholarship Program for a post-
graduate scholarship.
Su p p or tin g In for m a tion Ava ila ble: Details of synthetic
1
methodology. H and 13C NMR spectra of 1a -c, 2a -c, 3a ,b,
4a ,b, 6a -c, 7a ,b, 8a -c, 9a ,b, 10a -c, 11a -c, 12a ,b, 13a ,b,
14, and 15 and additional data for some known compounds.
This material is available free of charge via the Internet at
http://pubs.acs.org.
Ack n ow led gm en t. J .C.S. thanks NSERC (Canada)
and Ontario’s Premier’s Research Excellence Award
program for generous financial support. G.C. thanks the
J O049224G
J . Org. Chem, Vol. 69, No. 21, 2004 7071