Kinetic mechanism for dimerization of an α-thioamide substituted benzyl carbocation in aqueous solution

Article abstract of DOI:10.1002/(SICI)1099-1395(1998100)11:10<701::AID-POC26>3.0.CO;2-4

The products of reaction of the α-(N,N-dimethylthiocarbamoyl)-4-methoxybenzyl carbocation (1⁺) intermediate of solvolysis of α-(N,N-dimethylthiocarbamoyl)-4-methoxybenzyl benzoate esters (1-O₂CAr) show a strong dependence on solvent. The only product from reaction in 1,1,1,3,3,3-hexafluoro-2-propanoI (HFIP) is 2-dimethylamino-6-methoxybenzothiophene (2) from intramolecular cyclization of 1⁺. The reaction of 1⁺ in 50:50 (v/v) methanol-water (I = 0.50, NaClO₄) gives mainly the adducts to solvent. In 50:50 (v/v) trifluoroethanol-water (I = 0.50, NaClO₄), 1⁺ partitions between reaction with solvent (k_s, 27% yield), cyclization to form 2 (k_c, 3% yield) and nucleophilic addition of 2 to 1⁺ (k_alk, 70% yield) to form dimeric product 3. The yield of 3 in 50:50 (v/v) trifluoroethanol-water (I = 0.50, NaClO₄) is independent of the leaving group at 1-O₂CAr and remains constant as the concentration of the substrate is increased fourfold. These data show that the rate-determining step for dimerization of 1-O₂CAr is ionization of substrate to form 1⁺ and that the products of the reaction are determined by the rate constant ratio for partitioning of 1⁺ between addition of solvent and cyclization. The rate constant ratios determined for the partitioning of 1⁺ in 50:50 (v/v) trifluoroethanol-water (I = 0.50, NaClO₄) were k_s/k_c = 0.8 M^-1 and k_alk/k_s = 540 000 M^-1.

Full text of DOI:10.1002/(SICI)1099-1395(1998100)11:10<701::AID-POC26>3.0.CO;2-4

JOURNAL OF PHYSICAL ORGANIC CHEMISTRY, VOL. 11, 701–706 (1998)
Kinetic mechanism for dimerization of an a-thioamide
substituted benzyl carbocation in aqueous solution
Kathleen B. Williams and John P. Richard*
Department of Chemistry, University at Buffalo, SUNY, Buffalo, New York 14260-3000, USA
Received 20 October 1997; revised 29 December 1997; accepted 2 January 1998
‡
ABSTRACT: The products of reaction of the a-(N,N-dimethylthiocarbamoyl)-4-methoxybenzyl carbocation (1 )
intermediate of solvolysis of a-(N,N-dimethylthiocarbamoyl)-4-methoxybenzyl benzoate esters (1-O CAr) show a
2
strong dependence on solvent. The only product from reaction in 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP) is 2-
‡
‡
dimethylamino-6-methoxybenzothiophene (2) from intramolecular cyclization of 1 . The reaction of 1 in 50:50
(
(
v/v) methanol–water (I = 0.50, NaClO ) gives mainly the adducts to solvent. In 50:50 (v/v) trifluoroethanol–water
4
‡
I = 0.50, NaClO ), 1 partitions between reaction with solvent (k , 27% yield), cyclization to form 2 (k , 3% yield)
4
s
c
‡
and nucleophilic addition of 2 to 1 (kalk, 70% yield) to form dimeric product 3. The yield of 3 in 50:50 (v/v)
trifluoroethanol–water (I = 0.50, NaClO ) is independent of the leaving group at 1-O CAr and remains constant as the
4
2
concentration of the substrate is increased fourfold. These data show that the rate-determining step for dimerization of
‡
1
-O CAr is ionization of substrate to form 1 and that the products of the reaction are determined by the rate constant
2
‡
ratio for partitioning of 1 between addition of solvent and cyclization. The rate constant ratios determined for the
partitioning of 1 in 50:50 (v/v) trifluoroethanol–water (I = 0.50, NaClO ) were k /k = 0.8 M and kalk/k =
40 000 M .  1998 John Wiley & Sons, Ltd.
‡
� 1
4
s
c
s
�
1
5
KEYWORDS: carbocation; thioamide; kinetics
‡
INTRODUCTION
partitioning of 1 between intramolecular cyclization to
, nucleophilic addition of solvent and nucleophilic
2
We recently reported that the reaction of a-(N,N-
dimethylthiocarbamoyl)-4-methoxybenzyl pentafluoro-
benzoate (1-O C F ) in mixed aqueous organic solvents
addition of alkene 2.
2
6 5
proceeds by a D ‡ A (S 1) mechanism through an a-
N
N
N
thioamide-substituted 4-methoxybenzyl carbocation in-
‡
1
termediate 1 . We were very surprised to observe
‡
conversion of the carbocation 1 to 2 and then to the
dimeric product 3 in a total yield of ca 70% because, to
the best of our knowledge, there is no precedent in
organic chemistry for dimerization of the extremely low
�
8
steady-state concentrations [<10 M, see Discussion] of
‡
the strongly electrophilic species 1 in a nucleophilic
2
aqueous solvent.
We report here the results of a study of the effects of
changing solvent, substrate concentration and substituted
benzoate leaving group on the products of the reaction of
EXPERIMENTAL
1
-O CAr and the time course for formation of the
2
products in 50:50 (v/v) trifluoroethanol–water (I = 0.50,
Materials. HPLC-grade methanol was purchased from
Fisher and HPLC-grade acetonitrile and Gold label-grade
NaClO ) (see Scheme 1). The experimental results define
4
the rate-determining and product-determining steps for
2,2,2-trifluoroethanol from Aldrich. Water was purified
‡
the dimerization of 1 and the rate constant ratios for
by distillation and passage through a Milli-Q purification
system. All other organic chemicals used for syntheses
were of reagent grade and were used without further
purification.
*
Correspondence to: J. P. Richard, Department of Chemistry,
University at Buffalo, SUNY, Buffalo, New York 14260-3000, USA.
E-mail: jrichard@lacsu.buffalo.edu
Contract/grant sponsor: National Institutes of Health; contract grant
number: GM 39754.
Syntheses. The methods for the preparation of a-(N,N-

1998 John Wiley & Sons, Ltd.
CCC 0894–3230/98/100701–06 $17.50

7
02
K. B. WILLIAMS AND J. P. RICHARD
�
1
� 1
dimethylthiocarbamoyl)-4-methoxybenzyl pentafluoro-
e = 11200 and e = 22000 M cm are the extinction
2 3
2
benzoate (1-O CC F ) and a-(N,N-dimethylthiocarba-
coefficients of these products at 271 nm. The fraction of
the theoretical yield of the dimer 3 was calculated using
Eqn (2). The terms in the numerator of Eqns (1) and (2)
are proportional to the moles of product formed, while
2
6 5
moyl)-4-methoxybenzyl 4-nitrobenzoate (1-O CC H -
2
6
4
1
4
-NO ) were described in a recent paper.
2
HPLC product analyses. Product analyses were carried
out at room temperature (22 Æ 2°C). The reactions were
initiated by making a 100-fold dilution of substrate in
acetonitrile to give a final concentration of 0.05 –0.2 mM.
The procedures for HPLC analyses were as described in
the term in the denominator [A1-OSolv/e1-OSolv‡A /
2
e ‡2(A /e )] is proportional to the original number of
2
3 3
moles of substrate. Values of A /e for the dimeric
3
3
product have been multiplied by two in order to account
for the fact that 2 mol of substrate are consumed for each
mole of dimeric product 3 formed.
3
,4
earlier work, except that the mixed methanol–water
solvent used to elute the products of the reaction of
Ap="P
1
-O CC F in 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP)
fP ˆ A1
ꢀ1†
ꢀ2†
2
6 5
="
� OSolv 1� OSolv
‡ A =" ‡ 2ꢀA =" †
2
2
3
3
was prepared by mixing methanol and water which
contained 0.1 M sodium carbonate (pH 9) to give a final
concentration of 10 mM buffer. The buffered aqueous
phase was used because 2 appears to be unstable under
these conditions to elution with unbuffered solutions of
methanol–water. The fractional yield of 2 observed for
reaction in other solvents was not affected by addition of
this buffer.
2ꢀA3="3†
‡ A =" ‡ 2ꢀA =" †
f3 ˆ A1
="
� OSolv 1� OSolv
2
2
3
3
In one experiment, the time course for the reaction of
1
-O CC H -4-NO in 50:50 (v/v) trifluoroethanol–water
2
6
4
2
(
I = 0.50, NaClO ) was monitored. The fractional yields
4
of products during this reaction were calculated using
Eqns 1 and 2, where the term in the numerator refers to
the area of the product peak observed for analysis of a
fixed volume of the reaction mixture at a given reaction
time, and the terms in the denominator refer to the areas
of these product peaks observed after more than 10
reaction half-times. These experiments were conducted
using 9-hydroxy-9-methylfluorene as a stable internal
standard in order to correct for small variations in the
injection volume for different HPLC analyses.
The products of the reaction of 1-O CAr were moni-
tored at 271 nm, which is ꢀmax for 1-OH. The procedures
2
for identification of the reaction products have been
1
,2
described in earlier work. The products of the reaction
of 1-O CC F were determined after completion of the
2
6 5
reaction (ca 10 half-times), and the products of the
reaction of 1-O CC H -4-NO were determined after a
2
6
4
2
reaction time of 30 hours (ca six half-times). The ratios of
the yields of products were calculated from the ratios of
their peak areas from HPLC analysis and the ratios of
2
their extinction coefficients. In cases where 1 mol of
substrate 1-O CAr reacts to form 1 mol of product
2
1
-OSolv or 2, the fraction of the maximum theoretical
product yield was calculated using Eqn 1 where A1-OSolv
A and A refer to the areas of the respective product
RESULTS
,
Five products were observed for the solvolysis of
2
3
peaks from HPLC analysis and e1-OSolv = 13000,
1-O CC F in 50:50 (v/v) trifluoroethanol–water
2
6 5
a
Table 1. Yield of the products of reaction of 1-O CAr observed in different solvents
2
b
Product yield (%)
Substrate and reaction conditions
1-OSolv
2
3
c
1
1
1
1
-O CC F in 50:50 (v/v) methanol–water
98
2
6 5
(
(
1-OMe) 30
1-OH) 68
27
1-OTFE) 3
1-OH) 25
<1
2
-O CC F in 50:50 (v/v) trifluoroethanol–water
2
6 5
(
(
3
69
-O CC H -4-NO in 50:50 (v/v) trifluoroethanol–water
27
2
6
4
2
(
(
1-OTFE) 3
1-OH) 24
<1
3
71
d
-O CC F in HFIP
100
<1
2
6 5
a
b
c
d
At room temperature (22 Æ 2°C) and I = 0.50 (NaClO
Product yields were determined by HPLC analysis.
Data from Ref. 2.
4
), unless noted otherwise.
For reaction in a solution of neat HFIP (1,1,1,3,3,3-hexafluoro-2-propanol) which contains no added salt.

1998 John Wiley & Sons, Ltd.
JOURNAL OF PHYSICAL ORGANIC CHEMISTRY, VOL. 11, 701–706 (1998)

DIMERIZATION OF a-THIOAMIDE SUBSTITUTED BENZYL CARBOCATION
703
Figure 2. Time course for reaction of 1-O CC H -4-NO
2
2
6
4
Figure 1. Dependence of the yields of the products of
reaction of 1-O CC F in 50:50 (v/v) tri¯uoroethanol±water
on the concentration of substrate for a reaction at room
temperature (22 Æ 2°C) and I = 0.50 (NaClO4)
� 4
(
(
1 Â 10 M) in 50:50 (v/v) tri¯uoroethanol±water at 25°C
2
6 5
I = 0.50, NaClO ). The product yields, f , are equal to the
P
4
� 4
molar fraction of the original substrate (1 Â 10
M
1
-O CC H -4-NO ) consumed in the formation of the
2
6
4
2
product. The ®nal yields of 1-OSolv and 3 are indicated on
the ordinate. (&) 1-OSolv; (~) 2; (*) 3. The inset omits the
experimental points from the time course for formation of
the dimeric product 3 in order to show more clearly the lag in
the formation of this product
(
1
I = 0.50, NaClO ): the solvent adducts 1-OH and
-OTFE, the alkene 2, the dimeric product 3 and an
4
unknown product, whose structure was not determined.
The peak area from HPLC analysis at 271 nm for this
unknown product accounts for less than 2% of the total
peak areas for the other reaction products. It was not
possible to calculate an absolute yield for this product,
whose structure and extinction coefficients could not be
determined. This unknown reaction product was ignored
when calculating the absolute yields for the other
products [Eqns (1) and (2)], because the sum of the
yields of these other products was not significantly
different from 100%.
DISCUSSION
It was shown in earlier work that (1) the solvolysis of
1-O CAr in aqueous solvents occurs by a stepwise
2
mechanism through the a-thioamide-stabilized carboca-
‡
1
tion intermediate 1 ; (2) this carbocation intermediate
shows a very large selectivity for reaction with azide
1
ion; (3) dimeric 3 is a major product of the reaction of
2
1-O CC F in 50:50 (v/v) trifluoroethanol–water; and
2
6 5
The observed yields of the products of reaction of
(4) the dimeric product 3 forms by nucleophilic addition
of 2 to 1 (Scheme 1).
‡
2
1
-O CC F and 1-O CC H -4-NO in several different
2
6
5
2
6
4
2
solvents are given in Table 1. The yields reported in
Table 1 for reaction in 50:50 (v/v) trifluoroethanol–water
should replace the values from an earlier report (42%,
Rate-determining and product-determining steps
2
1-OSolv; 5%, 2; 53%, 3), which were calculated as the
for reaction of 1-O CAr
2
simple ratios of the moles of individual product formed
divided by the sum of the moles of all reaction products.
The earlier calculations of fractional product yields are
incorrect because they failed to take into account that
Figure 1 shows that the yields of the products of the
reaction of 1-O CC F in 50:50 (v/v) trifluoroethanol–
2
6 5
water, including the dimer, are independent of the
concentration of substrate. This observation requires that
these products form by reactions that are first-order in the
concentration of 1-O CC F , and is consistent with the
2
mol of substrate are required to form 1 mol of the
dimeric product 3 [see Eqns (1) and (2) in the Experi-
mental section).
2
6 5
Figure 1 shows the effect of a fourfold increase in the
concentration of 1-O CC F on the yield of products of
solvolysis in 50:50 (v/v) trifluoroethanol–water (I = 0.50,
NaClO4).
conclusion that the rate-determining step for all of these
reaction pathways is ionization of the substrate to form
2
6 5
‡
1 and the product-determining step is the partitioning of
‡
1 .
Figure 2 shows the time course for formation of the
The observed first-order rate constant for solvolysis of
�
4
� 3 � 1
products of the reaction of 1 Â 10 M 1-O CC H -4-
1-O CC F (k = 4.2 Â 10 s ) is 110-fold larger
2
6
4
2
6
5
solv
NO in 50:50 (v/v) trifluoroethanol–water (I = 0.50,
than for solvolysis of 1-O CC H -4-NO (ksolv =
2 6 4 2
2
�
5
1
NaClO4).
3.6 Â 10 ) in 50:50 (v/v) trifluoroethanol–water, which

1998 John Wiley & Sons, Ltd.
JOURNAL OF PHYSICAL ORGANIC CHEMISTRY, VOL. 11, 701–706 (1998)

7
04
K. B. WILLIAMS AND J. P. RICHARD
Scheme 1
corresponds to a 110-fold faster rate of formation of the
carbocation reaction intermediate. This requires the
accumulation of a 150-fold larger concentration of the
vation of dimerization of this dilute electrophilic reagent
in a nucleophilic solvent seems, at first glance, incredible.
However, dimerization of 1 can be shown to be a logical
consequence of the unique chemical reactivity of this
compound (Scheme 1).
‡
‡
intermediate 1 at the steady state during reaction of
1
-O CC F compared with 1-O CC H -4-NO , because
2 6 5 2 6 4 2
‡
the rate constants for the disappearance of 1 generated
by the ionization of these two substrates are the same.
‡
Once again, the yield of products of the reaction of 1 are
Reaction mechanism
seen to be independent of the steady-state concentration
of this carbocation during solvolysis of 1-O CAr (Table
We were at first confused by the observations that forma-
2
1). The observation of identical yields of products from
tion of the dimeric product 3 is first order in the concen-
reaction of 1-O CC F and 1-O CC H -4-NO requires
tration of 1-O CAr, and that the yield of this product is
2
6
5
2
6
4
2
2
‡
similar relative velocities for partitioning of 1 (v =
independent of the concentration of the steady-state
‡
‡
‡
‡
k [1 ][2], v = k [1 ] and v = k [1 ], Scheme 1) so that
reaction intermediate 1 , because it was expected that the
alk
s
c
the concentration of alkene 2 must be the same during the
course of reaction of these two substrates.
dependence of the yields of the products of simple
nucleophilic substitution at 1-O CAr on substrate con-
2
The following expression can be derived by making
the steady-state assumption for the concentration of the
centration would be different from that for the formation
of a dimeric product. However, these results can be
‡
‡
carbocation intermediate 1 ([1 ] ) during the initial
simply explained provided that k is the product-deter-
ss
c
stages of the reaction of 1-O CC H -4-NO in 50:50
mining step for dimerization, so that essentially every
mole of 2 formed by cyclization reacts with 1 to form
dimer, that is, if 2 does not accumulate to a large extent
2
6
4
2
‡
� 5 � 1
‡
(
v/v) trifluoroethanol–water: [1 ] = (3.6 Â 10 s )
ss
�
4
� 5 � 1
(
1  10 M)/(k ‡ k ), where k = 3.6  10
s
is
s
c
solv
the observed first-order rate constant for solvolysis of
during the reaction but instead reacts nearly quantita-
�
4
‡
1
-O CC H -4-NO and 1 Â 10 M is the initial con-
tively with 1 to form dimeric product 3. With this
2
6
4
2
1
1
� 1
centration of the substrate. It is known that k ꢁ 1 s ,
constraint, the alkene 2 becomes a steady-state reaction
s
so that the concentration of the species undergoing di-
intermediate with equal rates of formation by cyclization
‡
‡
‡
merization, [1 ] , is much less than 36 nM. The obser-
of 1 and disappearance by nucleophilic addition to 1 to
ss

1998 John Wiley & Sons, Ltd.
JOURNAL OF PHYSICAL ORGANIC CHEMISTRY, VOL. 11, 701–706 (1998)

DIMERIZATION OF a-THIOAMIDE SUBSTITUTED BENZYL CARBOCATION
705
form 3. The following experimental observations can be
accounted for by treatment of 2 as a steady-state reaction
intermediate:
kalk
ks
‰3Š
ˆ
ˆ 540 000 M� 1
ꢀ5†
‰2Š ‰1 � OSolvŠ
ss
Table 1 shows the curious result that a substantial yield
of the dimeric product 3 is observed for the reaction of
-O CAr in 50:50 (v/v) trifluoroethanol–water, but that
little or no 3 is observed for the reaction in 50:50 (v/v)
methanol–water, a solvent significantly more nucleophi-
lic than aqueous trifluoroethanol, or for the reaction in
HFIP, which is much less nucleophilic than aqueous
1
. Figure 2 shows that there is a lag in the formation of
the dimeric product 3 at relatively early times during
the reaction of 1-O CC H -4-NO while the concen-
1
2
2
6
4
2
tration of 2 increases to a roughly constant steady-state
concentration.
2
. The expression for the rate constant ratio for par-
titioning of the carbocation intermediate 1 between
‡
7,8
trifluoroethanol. These data show that stringent condi-
addition of solvent and 2 [eqn (3)] was derived by
making the assumption that the concentration of 2
tions must be met in order to observe the formation of 3.
Dimerization is unimportant in 50:50 (v/v) methanol–
water because the intramolecular cyclization reaction is
unimportant in this strongly nucleophilic solvent
‡
remains constant with time (d[2] /dt = 0 = k [1 ] �
ss
c
‡
k [1 ][2] ) and then solving for [2] [Eqn (4)]. This
alk
ss
ss
equation predicts that the relative yields of dimeric
(
k ꢁ k ). The observation that dimerization is also
s
c
product 3 and the solvent adducts will depend only
unimportant in HFIP shows that the change from
trifluoroethanol–water to this weakly nucleophilic but
strongly acidic solvent results in a large increase in kc
‡
upon the rate constant ratio for the partitioning of 1
between nucleophilic addition of solvent (k ) and
s
intramolecular cyclization to form 2 (k ). This
c
compared with k . The change probably reflects the
opposite effects of hydrogen bonds between solvent and
the dimethylamino group on the nucleophilic reactivity of
(decrease in k , see 4), and the electrophilic reactivity
of the sulfur cation towards addition of the 4-methoxy-
alk
provides a simple rationalization for the observation
of identical product yields from partitioning of the
common intermediate of reaction of 1-O CC F and
2
6 5
2
alk
1
-O CC H -4-NO .
2 6 4 2
phenyl ring (increase in k , see 5).
c
‰
1 � OSolvŠ
ks
k
s
ˆ
ˆ k ˆ 0:8
ꢀ3†
ꢀ4†
‰
3Š
k ‰2Š
alk
ss
c
kc
‰
2Š ˆ
ˆ 2:3  10� 6
M
ss
kalk
3
. Equation (4), which follows directly from the
assumption of a constant steady-state concentration
for 2, requires that the concentration of the alkene 2 at
the steady state be equal to the rate constant ratio for
partitioning of 2 between intramolecular cyclization
and reaction with solvent. The concentration of 2 at the
‡
steady state is approximately equal to [2] =
The value of k /k = 0.8 [Eqn (3)] for partitioning of 1
ss
s c
�
4
[
2
1-O CC H -4-NO ] f = (1 Â 10 M)
(0.023) =
shows that the reactivity of 50:50 (v/v) trifluoroethanol–
water in a bimolecular nucleophilic addition reaction is
approximately equal to the electrophilic reactivity of the
-thioamide group in an intramolecular reaction. This
2
6
4
2
P
�
6
.3 Â 10 M, where f = 0.023 is an approximate
P
average fractional conversion of substrate to 2 (Fig.
2
�
1
). The rate constant ratio k /k = 540 000 M for
alk s
‡
9
partitioning of 1 between nucleophilic addition of
0:50 (v/v) trifluoroethanol–water and 2 can then be
probably represents a large effective molarity of this
5
functional group in an intramolecular reaction since we
are not aware of any reports of bimolecular electrophilic
addition of a-thioamide-substituted carbocations to a
phenyl ring.
In summary, the dimerization of 1 is the result of (1)
the approximately equal rates of cyclization of 1 and
nucleophilic addition of 50:50 (v/v) trifluoroethanol–
water to 1 (k /k = 0.8, Scheme 1), (2) the large change
in the polarity of the benzylic carbon of 1 , from
obtained from Eqn (5), which is derived by combina-
tion of Eqns (3) and (4). This value is significantly
�
1
larger than k /k = 70000 M determined from the
alk
s
‡
product yields for reaction of 1-O CC F in the2
2
6 5
‡
presence of added 2 in 50:50 (v/v) methanol–water.
The difference in these rate constant ratios is primarily
the result of the larger nucleophilicity of aqueous
‡
s
c
2,3
‡
methanol compared with aqueous trifluoroethanol.
To the best of our knowledge, this is the highest
observed nucleophilic reactivity for an alkene com-
pared with the nucleophilic solvent water.
electrophilic to nucleophilic, which occurs on cyclization
‡
of 1 to form the carbon nucleophile 2, and (3) the
5
,6
5
5 Â 10 times greater reactivity of 2 than the solvent of

1998 John Wiley & Sons, Ltd.
JOURNAL OF PHYSICAL ORGANIC CHEMISTRY, VOL. 11, 701–706 (1998)

7
06
K. B. WILLIAMS AND J. P. RICHARD
‡
‡
5
0:50 (v/v) trifluoroethanol–water towards addition to 1
REFERENCES
which results in the nearly quantitative addition of 2 to 1
1
. J. P. Richard, S.-S. Lin, J. M. Buccigross and T. L. Amyes. J. Am.
Chem. Soc. 118, 12603–12613 (1996).
to form 3. We view the circumstances here to be so
improbable that the dimerization of low concentrations of
the steady-state reaction intermediate 1 will probably
‡
2. J. P. Richard, S.-S. Lin and K. B. Williams. J. Org. Chem. 61, 9033–
034 (1996).
. J. P. Richard, M. E. Rothenberg and W. P. Jencks. J. Am. Chem. Soc.
06, 1361–1372 (1984).
9
remain a completely unique organic reaction.
3
1
4. J. P. Richard. J. Am. Chem. Soc. 111, 1455–1465 (1989).
5. H. Mayr. Angew. Chem., Int. Ed. Engl. 29, 1371–1385 (1990).
6. H. Mayr and M. Patz. Angew. Chem., Int. Ed. Engl. 33, 938–957
(
1994).
. C. S. Q. Lew and R. A. McClelland. J. Am. Chem. Soc. 115, 11516–
1520 (1993).
. R. A. McClelland, N. Mathivanan and S. Steenken. J. Am. Chem.
Soc. 112, 4857–4861 (1990).
7
8
Acknowledgment
1
This work was supported by grant GM 39754 from the
National Institutes of Health.
9. A. J. Kirby. Adv. Phys. Org. Chem. 17, 183–278 (1980).

1998 John Wiley & Sons, Ltd.
JOURNAL OF PHYSICAL ORGANIC CHEMISTRY, VOL. 11, 701–706 (1998)