Alkyl Substituent Effect in the Deprotonation of Unsymmetrical Ketones

Article abstract of DOI:10.1021/jp953249p

The effect of various degrees of alkyl substitution on the relative rates of deprotonation from the two distinct sites in several unsymmetrical ketones in the gas phase is examined.The infrared multiple photon activation of an appropriately deuterium-labeled alkoxide ion generates the ion-molecule complex for the half-reaction of the bimolecular proton transfer process between an alkyl anion and an unsymmetrical ketone with one deprotonation site selectively deuterated.The resulting products are enolate ions generated by the removal of either a deuteron or a proton and, thus, are distinguishable by mass.The measurement of the enolate ion product ratios, along with an independent measurement of the kinetic isotope effect, allowed the kinetic effect of the alkyl environment on the relative proton transfer rates to be determined.The primary and secondary isotope effects are also estimated from the enolate ion product ratios.By examining the magnitude of the kinetic alkyl effect, the primary isotope effect, and the secondary isotope effect, we learn about the transition state for proton transfer.

Full text of DOI:10.1021/jp953249p

J. Phys. Chem. 1996, 100, 8827-8835
8827
Alkyl Substituent Effect in the Deprotonation of Unsymmetrical Ketones
Cris E. Johnson, Kristin A. Sannes, and John I. Brauman*
Department of Chemistry, Stanford UniVersity, Stanford, California 94305-5080
X
ReceiVed: NoVember 2, 1995; In Final Form: January 17, 1996
The effect of various degrees of alkyl substitution on the relative rates of deprotonation from the two distinct
sites in several unsymmetrical ketones in the gas phase is examined. The infrared multiple photon activation
of an appropriately deuterium-labeled alkoxide ion generates the ion-molecule complex for the half-reaction
of the bimolecular proton transfer process between an alkyl anion and an unsymmetrical ketone with one
deprotonation site selectively deuterated. The resulting products are enolate ions generated by the removal
of either a deuteron or a proton and, thus, are distinguishable by mass. The measurement of the enolate ion
product ratios, along with an independent measurement of the kinetic isotope effect, allowed the kinetic
effect of the alkyl environment on the relative proton transfer rates to be determined. The primary and
secondary isotope effects are also estimated from the enolate ion product ratios. By examining the magnitude
of the kinetic alkyl effect, the primary isotope effect, and the secondary isotope effect, we learn about the
transition state for proton transfer.
Introduction
O–
C
O
C
nhν
Enolate ions, generated by deprotonation of ketones, are
important intermediates in the synthesis of complex molecules.
In the case of unsymmetrical ketones, there are two distinct
deprotonation sites which generate two distinct enolate ions.
The relative rates of deprotonation from the two distinct sites
in an unsymmetrical ketone depend on the alkyl substituents.
The ability to predict and control which site will be deprotonated
in an unsymmetrical ketone is critical in the design of effective
synthetic strategies. The effect of alkyl substitution on the
relative rates of deprotonation is studied in the gas phase in
order to understand the effect of the solvent and examine the
intrinsic relationship between structure and reactivity.
CH3
CH3
CH3
CH3
CH3–
CH3
O
C
–
CH4 + CH3
CH2
(1)
deuterium labeling could be used to distinguish the isomer
enolate ions. Except for the internal energy distributions, the
alkyl anion-ketone complex formed by infrared multiple photon
activation of an alkoxide ion is identical to the complex formed
in the bimolecular proton transfer reaction between an alkyl
anion and a ketone (Figure 1). Thus, it is possible to learn about
the dynamics of a bimolecular proton transfer reaction by
studying a photochemically induced unimolecular reaction.
Extensive studies by other groups have confirmed that the
dissociation of alkoxide ions is stepwise and occurs through an
The direct study of deprotonation of unsymmetrical ketones
is difficult in the gas phase because the isomeric enolate ions
formed are not distinguishable by mass. Selective deuteration
of one deprotonation site would generate isomeric enolate ions
that are distinguishable by mass, but the R-hydrogens in ketones
are labile and proton exchanges result in the loss of integrity of
the deuterium label. In addition, multiple proton exchanges
might occur in an enolate ion-ketone complex which would
4-10
alkyl anion-ketone complex.
In addition, Haib and Stahl
examined the positional effects on deprotonation in unsym-
metrical ketones by studying the collision-induced dissociation
of 2-butoxide ion and its deuterated analogs: 3,3-d2, 1,1,1-d3,
1
scramble the hydrogens and deuteria. Thus, the deuterium label
11
and 1,1,1,3,3-d5. The ethyl hydrogens were shown to react
would no longer be associated with a specific site in the
unsymmetrical ketone and the number of deuteria in the enolate
ion could not be used to distinguish between the two isomeric
enolate ions. Alternatively, the reaction chemistry of isomeric
enolate ions with alkyl nitrites can be used to determine the
relative concentrations of isomers formed by deprotonation of
unsymmetrical ketones.2a
Deprotonation of ketones can be investigated indirectly by
studying the infrared multiple photon activation of alkoxide ions.
Tumas et al. showed that the unimolecular dissociation of
alkoxide ions proceeds through a stepwise mechanism and not
a concerted mechanism (eq 1).3 The first step is a heterolytic
cleavage of a carbon-carbon bond to form an alkyl anion-
ketone complex. The second step is a rapid and irreversible
proton transfer to yield an alkane and an enolate ion. Because
the proton transfer step is irreversible and there are no other
labile hydrogens available to exchange with the enolate ions,
2
.7 times faster than the methyl hydrogens. Although our raw
data are similar to the raw data of Haib and Stahl, we have
chosen a different analysis of the kinetic isotope effect.
In this paper, we study the effect of various degrees of alkyl
substitution on the relative rate of deprotonation from the two
distinct proton transfer sites in several unsymmetrical ketones.
Infrared multiple photon activation of an appropriately deuterium-
labeled alkoxide ion generates the ion-molecule complex for
the half-reaction of the bimolecular proton transfer process
between an alkyl anion and an unsymmetrical ketone with one
proton transfer site selectively deuterated. The resultant prod-
ucts, enolate ions, are generated by proton or deuteron abstrac-
tion and, thus, are distinguishable by mass. The measured
product ratios of enolate ions are due to a combination of the
kinetic alkyl group effect and the kinetic isotope effect as well
as a statistical factor. Using an independent measurement of
the kinetic isotope effect, the kinetic alkyl group effect can be
calculated from the measured product ratios of enolate ions.
The kinetic effect of the alkyl environment on the relative rates
X
Abstract published in AdVance ACS Abstracts, March 1, 1996.
S0022-3654(95)03249-7 CCC: $12.00 © 1996 American Chemical Society

8
828 J. Phys. Chem., Vol. 100, No. 21, 1996
Johnson et al.
2
-butanol-1,1,1-d3 (4), 2,4-dimethyl-2-pentanol-1,1,1-d3 (11),
and 2,6-dimethyl-4-(methyl-d3)-4-heptanol (12) were synthe-
sized by the Grignard reaction of methyl-d3-magnesium iodide
(Aldrich, 99+% atom D, 1.0 M solution in diethyl ether) with
acetaldehyde, propionaldehyde, 1,1,1-trifluoro-2-butanone, 4-meth-
yl-2-pentanone, and 2,6-dimethyl-4-heptanone, respectively. All
carbonyl compounds with the exception of 1,1,1-trifluoro-2-
butanone were purchased from Aldrich and used without further
purification. 1,1,1-Trifluoro-2-butanone was prepared from
trifluoroacetic anhydride and ethylmagnesium bromide in a 3:1
mole ratio.
Alkylmagnesium Iodide (Bromide) + Acetone-d6. 2-(Methyl-
d3)-2-butanol-1,1,1-d3 (6), 2-(methyl-d3)-3-methyl-2-butanol-
1
,1,1-d3 [7], 2-(methyl-d3)-2-pentanol-1,1,1-d3 (8), 2-(methyl-
d3)-4-methyl-2-pentanol-1,1,1-d3 (9), and 4,4-dimethyl-2-(methyl-
d3)-2-pentanol-1,1,1-d3 (10) were prepared by the reaction of
acetone-d6 (99.5% atom D) with the appropriate alkylmagnesium
halide. The alkyl magnesium halide was generated by reaction
of 2-iodopropane, 1-iodopropane, 1-iodo-2-methylpropane, and
1-bromo-2,2-dimethylpropane, respectively, with magnesium
turnings in diethyl ether. All compounds were purchased from
Aldrich.
Figure 1. Diagrammatic representation of the reaction surfaces of an
unimolecular infrared multiple photon induced dissociation of an
alkoxide ion and a bimolecular proton transfer between an alkyl anion
and a ketone. The complex formed from the photodissociation of the
alkoxide ion is located on the proton transfer surface.
Deuterated Alkylmagnesium Bromide + Carbonyl Compound.
2
-Butanol-3,3-d2 (3) was prepared from the reaction of ethyl-
of proton transfer from a carbonyl compound was determined
for 2° vs 1°, 3° vs 1°, and 2° vs 3°. The overall order for proton
transfer rates from fastest to slowest is 2° > 3° > 1°. The
primary and the secondary kinetic isotope effects are also
estimated from the data. By examining the magnitude of the
kinetic alkyl group effect, the primary isotope effect, and the
secondary isotope effect, we learn about the transition state for
proton transfer.
d2-magnesium bromide and acetaldehyde (Aldrich). Ethyl-d2-
magnesium bromide was generated by reacting bromoethane-
d2 (MSD Isotopes, 99.9% atom D) with magnesium turnings in
ether. 2-(Trifluoromethyl)-2-butanol-3,3,4,4,4-d5 (5) was pre-
pared from the reaction of ethyl-d5-magnesium bromide and
1,1,1-trifluoroacetone (Aldrich). 3-Ethyl-3-pentanol-d10 (14)
was prepared from the reaction of methyl propionate and two
equivalents of ethyl-d5-magnesium bromide. In both reactions,
ethyl-d5-magnesium bromide was generated by reacting bro-
Experimental Section
moethane-d with magnesium turnings in ether. The bromo-
5
ethane-d5 was prepared by refluxing ethanol-d6 in a mixture of
Materials. Nitrogen trifluoride was purchased from Ozark
Mahoning and used without further purification. Dimethyl
peroxide was prepared from the reaction of basic hydrogen
peroxide with dimethyl sulfate.12 The deuterated alcohols
studied are listed in Table 1 as the appropriate alkoxide ions.
Depending on the position of the deuteration, the alcohols were
prepared by standard Grignard reactions between either a
deuterated alkylmagnesium halide and a carbonyl compound
or an alkylmagnesium halide and a perdeuterated carbonyl
HBr and H2SO4, followed by distillation. 4-(Isopropyl-d7)-4-
heptanol (13) was prepared from the reaction of 2-propyl-d7-
magnesium bromide and 4-heptanone (Aldrich). 2-Propyl-d7-
magnesium bromide was generated by reaction of 2-bromo-
propane-d7 with magnesium turnings in ether.
Ion Formation. Primary negative ions were generated by
the dissociative electron capture of appropriate precursors at
-
7
-7
-
pressures from 1 × 10 to 5 × 10 Torr. F was generated
-
1
3
from NF , and CH O was generated from CH OOCH . The
compound.
The specific Grignard reactions used can be
3
3
3
3
choice of precursor for primary ion formation was determined
by the acidity of the alcohol studied. Using the appropriate
primary negative ion, the desired alkoxide ions were then formed
by the deprotonation of the appropriate alcohol present at
pressures from 5 × 10 to 2 × 10 Torr. The proton transfer
reaction was complete within 100-150 ms. All liquid samples
were degassed prior to use with several freeze-pump-thaw
cycles.
grouped into three types and are described below. The
compound numbers refer to Table 1. Each Grignard reaction
was followed by hydrolysis with saturated aqueous ammonium
1
4
chloride.
The deuterated alcohol was extracted from the
-7
-6
aqueous layer into ether. The ether layer was dried with sodium
carbonate and fractionally distilled. The collected alcohol was
further purified by preparative gas chromatography (Hewlett-
Packard 5790 equipped with a thermal conductivity detector;
1
0% SE30 on Chromosorb WHP 80/100 mesh column (Al-
Instrumentation. All experiments were carried out in a
Fourier transform ion cyclotron resonance (FT-ICR) mass
spectrometer consisting of an electromagnet (1.0-1.2 T), a
vacuum system, and a trapped-ion analyzer cell modified for
photochemical experiments, interfaced to an FTMS-2000 (Ion-
Spec Corp., Irvine, CA) data acquisition system. Details of the
ltech)). The product yields were typically 50-70% of the
theoretical yield. NMR spectra were consistent with those
reported in the literature, with the appropriate changes due to
deuterium substitution. The isotopic purity of the alcohols was
determined by negative-ion formation in the ion cyclotron
resonance mass spectrometer. Only one peak corresponding
to the (M - 1) ion appeared in the negative ion mass spectrum
of each alcohol. No isotopic impurities could be detected above
the background noise, indicating that the isotopic purity was
greater than or equal to 99%.
15
Fourier transform instrumentation can be found elsewhere. The
2.54 cm cubic analyzer cell was modified by replacing the back
plate with a gold-coated copper mirror and replacing the solid
front plate with a plate containing a 2.0-2.3 cm diameter hole.
All pressures were measured using a vacuum ionization gauge
(Varian, model 844) and were not calibrated. The background
Methyl-d3-magnesium Iodide + Carbonyl Compound. 2-Pro-
panol-1,1,1-d3 (1), 2-butanol-1,1,1-d3 (2), 2-(trifluoromethyl)-
pressures obtained were typically 9 × 10- to 3 × 10 Torr,
9
-8

Substituent Effect in the Deprotonation of Ketones
J. Phys. Chem., Vol. 100, No. 21, 1996 8829
TABLE 1: Enolate Ion Photoproduct Yields from the Infrared Multiple Photon Dissociation of Alkoxide Ions
compound
number
parent
alkoxide ion
anionic
product (mass)
photoproduct
yield
neutral product
O–
1
2
3
4
5
6
7
8
9
60
H
HD
2
66%
34%
5
9
CD3
CD3
CH3
CD3
CH3
CD3
CD3
CD3
CD3
CD3
CH3
C
H
CH3
O–
C
74
H
2
79%
21%
7
3
HD
CH2CH3
H
O–
C
73
H
2
43%
57%
7
2
HD
CD2CH3
H
O–
C
74
CF
CF
CD
3
H
D
62%
7%
31%
7
6
3
9
3
CH2CH3
3
COCH
2
CH
3
3
CF3
O–
C
76
CF
CF
CH
3
H
D
51%
16%
33%
7
6
5
9
3
CD2CD3
3
COCD
2
CD
CF3
O–
C
74
CD
CD
3
H
75%
25%
7
3
4
CH2CH3
CD3
O–
C
88
CD
CD
CH
3
4
3
H
49%
49%
2%
8
6
7
2
CH(CH3)2
CH2CH2CH3
CH2CH(CH3)2
CH2C(CH3)3
CH2CH(CH3)2
CHDCH
3
CD3
O–
C
88
CD
CD
CH
3
4
3
H
60%
25%
15%
8
6
7
2
CH
2 2
CH D
CD3
O–
C
102
01
CD
CD
CH
3
4
2
H
33%
13%
54%
1
6
2
DCH(CH
)
3 2
CD3
O–
C
1
0
1
116
15
CD
CD
CH
3
4
2
H
0%
0%
100%
1
6
2
DC(CH
)
3 3
CD3
O–
C
1
102
01
CH
CH
CD
CH
CH
4
3
3
3
2
23%
6%
12%
36%
23%
1
D
H
99
60
59
CD3
CH(CH
DCH(CH
3
)
2
3
)
2
2
O–
C
1
1
1
2
3
4
141
02
01
CD
CH
CH
3
3
2
H
12%
61%
27%
1
1
CH(CH
DCH(CH
3 2
)
3
CD3
CH2CH(CH3)2
)
CH2CH(CH3)2
O–
120
19
13
CH
CH
CD
3
3
3
CH
CH
CHDCD
2
CH
CH
3
75%
18%
7%
1
1
2
2
D
CH3CH2CH2
C
CD(CD3)2
3
CH3CH2CH2
O–
94
CH
CD
CD
3
CH
3
CD
3
CD
2
D
2
H
3
39%
37%
25%
9
8
0
9
CH3CH2
C
CD2CD3
CD2CD3
-
7
-6
while operating pressures were typically 5 × 10 to 2 × 10
output centered around the 9.6 and 10.6 mm transitions. Details
1
6
Torr. Pyrolysis of the parent alcohol on the filament often led
to the formation of a small amount of ketone which then
produced enolate ions. Since these enolate ions are identical
to the expected photoproduct resulting from the irradiation of
the alkoxide ions, the enolate ions were ejected prior to the laser
pulse. The photoproducts were detected a short time after the
laser pulse to avoid subsequent reactions of the photoproduct
with the neutral precursors present in the cell (typically 10 ms).
A tunable pulsed CO2 laser (Lumonics 103-2 CO2 TEA
laser) was used for the infrared multiple photon activation
experiments. The pulsed CO2 laser was operated multimode
with a diffraction grating as the rear optic to allow line tunable
of the laser operation have been provided previously. The
shot-to-shot reproducibility of the laser was quite high (within
a few percent). However, the reliability of the laser did limit
the practical range of signal averaging to 20-30 scans. If
several laser misfires occurred during data collection, the entire
data set was discarded. The laser beam spot size is reduced by
an iris before entering the analyzer cell through a potassium
chloride window. Passing through a potassium chloride window
does not appreciably decrease the intensity of the laser beam.
The intensity of the laser beam can be attenuated by passing
the beam through CaF2 flats of varying thicknesses. Because
the laser beam passes through the ion cloud twice, the effective

8
830 J. Phys. Chem., Vol. 100, No. 21, 1996
Johnson et al.
O
C
–
O
C
CD H + CH
3 2
CHCH(CH )
3
3
CH3
CH2CH(CH3)2
O
C
CD3–
–
CD3H + CH2
CH2CH(CH3)2
O
–
O–
CH + CD
4
3
C
3 2
CHCH(CH )
O
C
CH3
C
CH2CH(CH3)2
CD3
CH2CH(CH3)2
(2)
O
C
CH3–
CD3
–
CH3D + CD2
CH2CH(CH3)2
O
–
O
C
(CH3)2CHCH3 + CD3
–
C
CH2
CD3
CH3
O
C
CH3)2CHCH2–
(
(
CH3)2CHCH2D + CD2
CH3
fluence is twice the measured fluence. The energy of the laser
pulse was measured using a Scientech 365 power and energy
meter with a Scientech 38-0102 volume-absorbing disk calo-
rimeter. The area of the laser beam spot is calculated from the
anion-ketone complex is illustrated in eq 2 for compound 11.
In these cases, only the product ratios from the alkyl anion-
ketone complex that gave isomeric enolate ions that are
distinguishable by mass are examined. For compound 11, this
corresponds to the product ratios from the elimination of CH4
3
image left by the laser pulse on thermal paper. The fluence
2
(
J/cm ) is calculated by dividing the laser pulse energy (in joules)
and CH D (eq 2). The relative ease of alkyl group loss is
17
by the area (in square centimeters) of the laser beam spot and
then multiplying by two. The range of laser fluences measured
discussed elsewhere. For compound 10, only neopentane loss
is observed; no methane loss is observed. Thus, the relative
rates of proton transfer from a methyl group vs a neopentyl
group could not be determined.
2
was typically 1-4 J/cm .
Reproducibility. The quantitative results expressed in Table
1
are based on a small number of working days (1-2) in which
Except for compounds 4 and 5, the photoproduct yields
showed little or no fluence dependence over the fluence range
studied. Compounds that showed a slight fluence dependence
were treated as fluence independent, and the photoproduct yields
at the different fluences were averaged without introducing an
appreciable error. The infrared absorption cross sections remain
approximately constant for each compound. Thus, a valid
comparison can be made between the photoproduct yields of
different compounds. For compounds 4 and 5, the amount of
quantitative data with a high degree of reproducibility were
obtained. Several different data sets of multiple scans were
obtained at each different fluence value. The reproducibility
of these photoproduct yields was within (1-2%. Additional
attempts at data acquisition were made for each compound, and
qualitative agreement (photoproduct yields within 5-10% of
reported values) was observed. The low degree of reproduc-
ibility was due to instability in the ion signal prior to irradiation
and/or misalignment of the laser beam. Because of possible
systematic errors and experimental difficulties associated with
the interaction of the high-intensity laser pulse with the analyzer
cell, an estimate of the absolute accuracy of the photodissocia-
tion yield measurements is taken to be within approximately
-
CF3 lost is strongly fluence dependent. In addition, the ratio
of the isomeric enolate ions formed is fluence dependent. The
infrared absorption cross sections for compounds 4 and 5 are
different. Thus, a comparison of the photoproduct ratios of
isomeric enolate ions for compounds 4 and 5 was made at the
fluences that gave approximately the same amount of CF3- loss.
(
5%. However, some cancellation of errors should occur for
the photoproduct branching ratios reported in Table 1. It should
be noted that small changes in photoproduct yields result in
significant changes in the photoproduct branching ratio, espe-
cially for very large or very small values of the photoproduct
branching ratio. It is important that the basic conclusions of
this paper are not strongly dependent on the quantitative results.
The basic conclusions are drawn from the relative abundances
of the photoproducts, which are very reproducible.
TABLE 2: Observed Kinetic Isotope Effects: (A) in the
Infrared Multiple Photon Induced Proton Transfer
Reactions from Acetone-1,1,1-d as a Function of Base and
3
(B) in the Infrared Multiple Photon Induced Proton
Transfer Reaction from 3-Pentanone-d
5
A. From Acetone-1,1,1-d3
kCH3/kCD3
∆H° a
base
-
CF
3
6.0 ( 0.5
2.5
1.9 ( 0.3
1.6 ( 0.1
1.6 ( 0.1
1.6 ( 0.1
-8
-32
-32
-44
-48
-48
Results
-
C
H
6
H
5
-
Alkoxide Ion Photoproduct Yields. The photoproduct
yields from the alkoxide ions are summarized in Table 1. The
masses of the product enolate ions and the neutral product
formed are shown. The neutral product is inferred from mass
balance and thermochemical considerations. For compounds
(
CH
3
)
2
CHCH -
2
-
-
CH
CD3
3
B. From 3-Pentanone-d
kCH3CH2/kCD3CD2
1.5
5
base
∆H° b
-52
1
-6, the cleavage step is highly selective for a single group
-
2
(i.e. H, CF3, CD3) and only one anion-ketone complex is
CD CD
3
-
formed. For compounds 4 and 5, CF3 loss is observed as well.
For the other compounds, the cleavage step is competitive for
two or more groups and more than one alkyl anion-ketone
complex can be formed. The formation of more than one alkyl
a
Reaction enthalpy for the proton transfer reaction between a base
b
and acetone in kcal/mol. Reaction enthalpy for the proton transfer
reaction between a base and 3-pentanone-d in kcal/mol. The EA for
5
-
CD
3
CD
2
is negative.

Substituent Effect in the Deprotonation of Ketones
J. Phys. Chem., Vol. 100, No. 21, 1996 8831
This insured that the internal energy distributions of the two
compounds were approximately equal.
TABLE 3: Alkyl Group Effect as a Function of
r-Substitution
Kinetic Isotope Effects. The inherent preference for depro-
tonation of H versus D from methyl substituents was determined
from the measurement of isomeric enolate ion product ratios
alkyl group
effect
derived from
compounds
derived from a ratio
of alkyl group effects
CH3CH2 CH3
k
k
k
k
k
/k
2.8
2.3
1.9
CH3CH2CH2 CH3
(M - H)/(M - D) resulting from the infrared multiple photon
/k
(
CH3)2CH CH3
dissociation of compound I (eq 3). Cleavage of the R group
results in the formation of an ion-molecule complex between
/k
CH3CH2 (CH3)2CH
/k
1.5
1.2
CH3CH2CH2
/k(CH3)2CH
1.3
-
R and acetone-1,1,1-d3, as shown in eq 3. Proton transfer then
occurs from either a CH3 or a CD3 substituent, so that the
resulting enolate ion product ratio corresponds to a kinetic
product formed when the deuterated ketone loses a deuteron.
The last term on the right side of eqs 5-8 accounts for the
kinetic isotope effect. Three quantities are required to specify
the alkyl effect in these experiments; the experimental ratio of
products from the deuterated butanones is not sufficient.
CH3 CD3
isotope effect, k /k . Such an isotope effect has been
3
,18
previously determined for R ) CF3, phenyl, CH3, and CD3.
This study reports the additional measurement of the isotope
effect for R ) H (from 1) and isobutyl (from 11). These values
are collectively presented in Table 2A.
CH3 CD3
Because the kinetic isotope effect k /k had been measured
independently and as a function of base, eqs 5 and 7 were used
to calculate the kinetic alkyl group effects. Haib and Stahl use
equation 6 to analyze their data, but they do not use an
independent measure for the kinetic isotope effect. Although
their values thus calculated for the alkyl effect differ from ours,
their conclusions are qualitatively the same as ours.
O–
nhν
CD3
C
R
I
CH3
O
C
–
RCH vs CH from RCH C(O)CD Compounds
2
3
2
3
R
R
D + CD2
CH3
O
(
R)CH2
CH3
CD3
C
CH3
(3)
k
3
M - H
M - D
k
k
O
C
)
×
2
÷
(5)
CH3
CD3
R–
–
k
H + CD3
CH2
RCH vs CH from RCD C(O)CH Compounds
2
3
2
3
In addition, the inherent preference for deprotonation of H
vs D from ethyl substituents was determined from the measure-
ment of the isomeric enolate ion product ratios (M - H)/(M -
D) resulting from the infrared multiple photon dissociation of
(
R)CH2
(R)CH2
k
3
2
M - D
×
M - H
k
k
)
×
(6)
CH3
(R)CD2
k
3
-(ethyl-d5)-3-pentoxide-d5 ion. Cleavage of an ethyl-d5 group
results in the formation of an ion-molecule complex between
ethyl-d5 anion and 3-pentanone-1,1,1,2,2-d5 (eq 4). Proton
transfer then occurs from either a CH3CH2 or a CD3CD2
(R)(R′)CH vs CH
3
from (R)(R′)CHC(O)CD
3
Compounds
(
R)(R′)CH
CH3
CD3
k
M - H
M - D
k
k
)
3 ×
÷
(7)
substituent. The resulting enolate ion product ratio corresponds
CH3
k
CH3CH2 CD3CD2
to the kinetic isotope effect k
B.
/k
and is shown in Table
2
RCH vs (R)(R′)CH from
2
O–
RCH C(O)CD(R)(R′) Compounds
2
nhν
CD3CD2
CD3CD2
C
CH2CH3
(
R)CH2
(R)(R′)CH
k
1
2
M - H
M - D
k
k
CD2CD3
O
)
×
÷
(8)
(
R)(R′)CH
(R)(R′)CD
O
k
–
CD CD + CD CD
3
3
3
C
2 3
CH CH
Alkyl Group Effect as a Function of R-Substitution. Using
C
CH2CH3
(4)
CHCH3
CH3CH2 CH3
CH3CH2CH2
eq 5, the kinetic alkyl group effects k
k
/k and k
/
O
CD3CD2–
CH3
–
were calculated from the product ratios of compounds 6
CD3CD2H + CD3CD2
C
(CH3)2CH
and 8, respectively. The kinetic alkyl group effect k
/
CH3
k
was calculated from the product ratio of compound 7 using
Data Analysis
CH3CH2 (CH3)2CH
eq 7. The kinetic alkyl group effect k
calculated as the ratio of k
kinetic alkyl group effect k
/k
was
CH3CH2 CH3
(CH3)2CH
/kCH3. The
Kinetic Alkyl Group Effect. The kinetic alkyl group effect
is the inherent difference in deprotonation rates for a single
proton from different alkyl environments and can be determined
from the product ratios of the isomeric enolate ions. The
product ratios are assumed to be due to a combination of the
kinetic alkyl group effect, the kinetic isotope effect, and a
statistical factor. The equation relating the kinetic alkyl group
effect to the product ratio will depend on the site of the
deuteration in the compound. Equations 5-8 calculate the alkyl
group effect for the four types of compounds studied. These
equations are derived in appendix A. The statistical correction
factor is the first term on the right side of eqs 5-8. In the
second term on the right side of eqs 5-8, M - H represents
the abundance of the product formed when the deuterated ketone
loses a proton and M - D represents the abundance of the
/k
to k
CH3CH2CH2 (CH3)2CH
/k
was calculated as
CH3CH2CH2 CH3
(CH3)2CH CH3
the ratio of k
/k
to k
/k . For comparison,
the kinetic alkyl group effect kCH3CH2CH2/k
(CH3)2CH
was calculated
from the product ratio of compound 13 using equation 8. We
(
CH3)2CH (CH3)2CD
assume that the kinetic isotope effect k
/k
is
identical to the kinetic isotope effect k /k with CD3- acting
as the base. It should be noted that the former kinetic isotope
effect involves only a primary isotope effect while the later
kinetic isotope effect has contributions from both primary and
secondary isotope effects. These results are shown in Table 3.
Alkyl Group Effect as a Function of â-Substitution. The effect
of substitution at the â-carbon on the relative rates of depro-
tonation of the R-carbon was calculated from the product ratios
of compounds 6, 8, 9, and 10 using eq 5. In each case CD3-
CH3 CD3

8
832 J. Phys. Chem., Vol. 100, No. 21, 1996
Johnson et al.
CH3CH2
/kCD3CD2 is the
TABLE 4: Alkyl Group Effect as a Function of
Assuming that the kinetic isotope effect k
same as the kinetic isotope effect k
â-Substitution
CH3CH2
/kCH3CD2 and that CD3
-
R
k(R)CH2/kCH3
and CD3CD2- are similar bases, the secondary isotope effect
can be calculated from eq 11 using the kinetic isotope effects
CH
CH
3
2.8 ( 0.5
2.3 ( 0.5
2.4 ( 0.1
a
CH3 CD3
CH3CH2 CD3CD2
19
k
/k
and k
/k
from Table 2.
The primary
3
CH
2
(CH
CH
3
)
)
2
CH
C
isotope effect, p, then can be calculated from eqs 9 and 10.
The secondary isotope effect, s, is 1.1. Using this value for s,
the range of the primary isotope effect, p, is 1.3-1.4.
(
3
3
a
No methane loss was observed.
The second method allows the isotope effects to be evaluated
as a function of base. The same assumptions leading to eq 11
TABLE 5: Alkyl Group Effect kCH3CH2/kCH3 as a Function of
Base
in the first method are made in the second method. In eq 11,
base
kCH3/kCD3 a
kCH3CH2/kCH3
∆H° b
the kinetic isotope effect k /k
CH3 CD3
has been experimentally
-
measured as a function of base, but an expression for the kinetic
CF
H
CD
3
6.0
1.9
1.6
2.4
3.0
2.8
-10
-30
-50
-
CH3CH2 CH3CD2
isotope effect k
/k
in terms of known quantities is
-
3
needed. Since compounds that are identical except for the site
of deuterium substitution will have the same kinetic alkyl group
a
The value of the kinetic isotope effect from Table 2. b Range of
the reaction enthalpy for the proton transfer reaction between a base
and 2-butanone in kcal/mol.
effect, equating eqs 5 and 6 and rearranging gives an expression
CH3CH2 CH3CD2
for the kinetic isotope effect k
/k
(eq 12).
TABLE 6: Alkyl Group Effect k CH3)2CHCH2/kCH3 as a
(
CH3CH2
CH2/CD3 CH3/CD2
k
k
P
P
a
Function of Base
)
(12)
CH3CD2
CH3 CD3
kCH3/kCD3 b
k(CH3)2CCH2/kCH3
k
/k
base
-
CD
CH
3
1.6
1.6
2.4
2.1
In eq 12, PCH2/CD3 represents the (M - H)/(M - D) product
-
(
3
)
2
CHCH
2
CH3/CD2
ratio in eq 5 and P
represents the (M - H)/(M - D)
a
No data available to estimate the reaction enthalpy. b The value of
the kinetic isotope effect from Table 2.
product ratio in eq 6. Substituting eq 12 into eq 11 gives an
expression for the secondary isotope effect, s, in terms of
quantities that have been experimentally measured (eq 13).
acted as the base. The kinetic alkyl group effect k(R)CH2/kCH3
for the different R groups is shown in Table 4.
Alkyl Group Effect As a Function of Base. The kinetic alkyl
CH3 CD3 2
(k
/k
)
s )
(13)
CH2/CD3 CH3/CD2
CH3CH2 CH3
P
P
group effect k
/k
was calculated as a function of base
from the product ratios of compounds 2, 4, and 6 using eq 5.
(CH3)2CHCH2 CH3
Once s is estimated, the primary isotope effect, p, can be
estimated from eq 9. Using compounds 2 and 3 where H is
The kinetic alkyl group effect k
/k
was calculated
-
as a function of base from the product ratios of compounds 9
CH3
the base, s ) 1.3 and p ) 1.1. Using compounds 4 and 5 where
and 12 using eq 5. In each case, the kinetic isotope effect k
k
/
-
CD3
-
-
-
CF3 is the base, s ) 1.3 and p ) 3.5.
determined with the appropriate base (CF3 , H , CD3 ) is
substituted into eq 5. These results are shown in Tables 5 and
To summarize, the first method estimates the secondary
isotope effect to be 1.1 and the primary isotope effect to be
6
.
1
.3-1.4. The second method estimates the secondary isotope
Kinetic Isotope Effects. The observed kinetic isotope effect
effect to be 1.3 and the primary isotope effect to be 1.1 when
H
is a combination of primary and secondary isotope effects. It is
not possible to measure independently the primary and second-
ary isotope effects from these data. Using the data presented
here, two methods are derived to estimate the contributions of
the primary and the secondary isotope effects to the overall
kinetic isotope effect.
-
is the base. The second method estimates the secondary
isotope effect to be 1.3 and the primary isotope effect to be 3.5
when CF3- is the base. Thus, the primary isotope effect is
dependent on the reaction exothermicity while the secondary
isotope effect is insensitive to the reaction exothermicity.
In the first method, we first assume that the kinetic isotope
effect is the product of the primary isotope effect and the
secondary isotope effect for each nonreacting deuterium. This
assumption leads to the following equations, where p represents
the primary isotope effect and s represents the secondary isotope
effect:
Discussion
We are interested in understanding the details of proton
transfer in alkane loss from a vibrationally activated tertiary
alkoxide ion. If the mechanism is concerted, only the kinetic
alkyl group effect on the rate of cleavage can be determined.
On the other hand, if the mechanism is stepwise, the kinetic
alkyl group effect on the rate of proton transfer can be
determined. Tumas et al. showed that the isotope effect
associated with the cleavage of the alkyl group is dependent
on the internal energy of the system, while the isotope effect
associated with the proton transfer reaction is relatively insensi-
CH3
k
k
2
)
ps
(9)
CD3
CH3CH2
CH3CD2
k
k
)
ps
(10)
3
tive to the internal energy of the system. Because the isotope
effects have different internal energy dependences, the cleavage
step and the proton transfer step cannot occur together and must
have different transition states. Thus, the mechanism is stepwise
and not concerted. The isotope effect for the proton transfer
step is relatively independent of the internal energy of the system
because the internal energy of the system is above the reaction
threshold for this step. If the proton transfer step occurred near
the reaction threshold, unusually large isotope effects would
If we also assume that the primary and secondary isotope effects
are the same for abstraction from both CH2 and CH3, then eq 9
can be divided by eq 10 to give an expression for s, the
secondary isotope effect (eq 11).
CH3 CD3
k
/k
s )
(11)
CH3CH2 CH3CD2
k
/k

Substituent Effect in the Deprotonation of Ketones
J. Phys. Chem., Vol. 100, No. 21, 1996 8833
be expected.20 Since the proton transfer step occurs at an
internal energy above the reaction threshold for this step,
however, the kinetic alkyl group effect corresponds to a ratio
of averaged rate constants, 〈k(E)〉. If the isotope effect in the
proton transfer step depended strongly on the internal energy
of the system, only the relative ease of proton transfer from the
different alkyl groups could be determined, since the magnitude
of the alkyl group effect would be strongly dependent on the
reaction conditions.
groups. Thus, we do not expect the differences in vibrational
frequencies to have a large effect on the sums of states for the
two transition states. Second, steric hindrance can, in principle,
limit the number of orientations at which proton transfer can
occur and thus decrease the sum of states. We believe that the
differences in critical energies for the two transition states mostly
determine the kinetic order for proton transfer rates in the gas
phase.
Kinetic Isotope Effects. We can learn more about the
transition state for proton transfer by studying the kinetic isotope
In solution, the deprotonation of unsymmetrical ketones
generates enolate ion mixtures that have compositions that
depend on whether the reaction is run under thermodynamic or
effect. From the infrared multiple photon dissociation of
CH3 CD3
compound I, the kinetic isotope effect k /k
was measured
2
1
as a function of base (see eq 3 and Table 2). With increasing
kinetic control. Under thermodynamic control, the deproto-
CH3 CD3
reaction exothermicity, the kinetic isotope effect k /k
nation step is reversible and the thermodynamically more stable
enolate ion dominates.22 Under kinetic control, the deproto-
decreases, as is expected from Marcus theory. In the case of
phenyl anion and hydride ion, the reaction exothermicity is the
nation step is not reversible, and the products depend on the
rate of deprotonation. The rate of deprotonation of unsym-
metrical ketones is frequently faster at the least substituted
R-carbon, which is often the position of the less acidic hydrogen.
Faster deprotonation rates for the less acidic hydrogen are also
observed in nitroalkanes. Thus, the rate of deprotonation
frequently does not correlate with the equilibrium acidity.
CH3 CD3
same, but the kinetic isotope effect k /k
is different. A
‘mass effect’ in the symmetric stretch of the transition state for
proton transfer may contribute to this difference. In general, a
transition state for proton transfer (A‚‚‚H‚‚‚B) will have a
symmetric stretch in which A and B move in opposite directions.
If the A-H bond strength is similar to the B-H bond strength,
then the contribution that the symmetric stretch makes to the
total zero point energy of the transition state depends on the
In the gas phase, the thermodynamic order of proton transfer
from a carbonyl compound is 2° > 1° > 3°, as is the case in
2
3
2
a
relative masses of A and B.
solution. The effect is modest; at equilibrium there is 1.13
times more secondary than primary enolate ion. The order of
kinetic acidities appears, however, to be quite sensitive to the
experiment. Squires and co-workers2a find that removal of a
If A and B have similar masses, the proton that is being
transferred remains essentially motionless during the symmetric
stretch of the transition state.
3° proton is both thermodynamically and kinetically unfavorable,
in contrast to our observation that removal of a 3° proton is
faster than removal of a 1° proton. The deprotonation reactions
are carried out under rather different conditions, which may
account for some of the difference.
The vibrational frequency of the symmetric stretch in the
transition state will be independent of whether a hydrogen or a
deuterium is present and will make the same contribution to
the zero point energy of both transition states. The transition
state in which a deuterium is being transferred (II) will have a
zero point energy contribution from the remaining C-H bond
while the transition state in which a hydrogen is being
transferred (III) will have a zero point energy contribution from
the remaining C-D bond.
Our gas phase kinetic order for proton transfer is different
from both the thermodynamic order and the kinetic order in
solution. We find the kinetic order for proton transfer from a
carbonyl compound is 2° > 3° > 1° (see Table 3). Substitu-
tion at the â-carbon does not change the kinetic alkyl group
effect (see Table 4). The kinetic alkyl group effect is also
relatively insensitive to the reaction exothermicity (see Tables
5
and 6).
The kinetic alkyl group effect arises from the difference in
the stabilities of the transition states for proton transfer. The
relative stabilities of the transition states are determined by the
differences in critical energies and the sums of states for the
transition states. There are two factors that affect the critical
energies for the transition states. First, the difference in critical
energies for the two proton transfer reaction channels can be
influenced by the difference in the thermodynamic stabilities
of the products. If there is some degree of product formation
in the transition state, then the differences in the thermodynamic
stabilities of the products will be mirrored in the difference of
the critical energies of the two transition states. The estimated
secondary isotope effect (s ) 1.1-1.3) suggests that significant
rehybridization from sp to sp is occurring in the transition
state and that the transition state must resemble the products to
some degree. Second, the critical energy for the more sterically
hindered proton transfer site will be increased relative to the
transition state energy of the less sterically hindered proton
transfer site.
II
III
Since the zero point energy of a C-H bond is larger than the
zero point energy of a C-D bond, the total zero point energy
for the transition state in which the deuterium is transferred (II)
will be greater than the total zero point energy for the transition
state in which the hydrogen is transferred (III). Thus, a large
kinetic isotope effect is expected.
If A and B have very different masses, the symmetric stretch
of the transition state will involve the motion of the proton being
transferred.
3
2
The vibrational frequency of the symmetric stretch will depend
on whether a hydrogen or a deuterium is present. The
symmetric stretch will make a different contribution to the total
zero point energy of the transition state depending on whether
a hydrogen or a deuterium is present. This will decrease the
difference in total zero point energy between the transition state
for hydrogen transfer and the transition state for deuterium
There are two factors that affect the sum of states for the
transition state at a given energy above the critical energy. The
vibrational frequencies and their differences can be important,
CH3CH2 CH3
but in the case of k
/k , there is only a small difference
in the C-H stretching and bending frequencies of CH2 and CH3

8
834 J. Phys. Chem., Vol. 100, No. 21, 1996
Johnson et al.
transfer. Thus, a smaller kinetic isotope effect will be seen.
When phenyl anion acts as the base, the symmetric stretch of
the transition state can be approximated by the first case. On
the other hand, if hydride ion acts as the base, the symmetric
stretch of the transition state can be approximated by the second
case. Thus, the difference in kinetic isotope effect k /k
for phenyl anion and hydride ion can be explained in part by
the differences in their masses.
transfer from the secondary position of 2-butanone would be
expected to be faster than the rate of proton transfer from the
primary position.
To summarize, no charge delocalization and no rehybridiza-
tion in the transition state favor proton transfer from the methyl
group in 2-butanone. However, charge delocalization and
rehybridization in the transition state favor proton transfer from
the ethyl group in 2-butanone. Thus, in each proton transfer
transition state for 2-butanone, the effect of charge transfer
competes with the effects of rehybridization and charge delo-
calization such that these effects almost cancel each other out.
In solution, transition state imbalances have been observed
in proton transfer reactions involving carbon acids. Bernasconi
proposes that charge delocalization into the nitro group lags
behind proton transfer in nitroalkanes and, thus, that the central
CH3 CD3
Conclusions
We have three pieces of a puzzle that initially appear to give
conflicting pictures about the transition state for proton transfer.
First, we have a small kinetic alkyl group effect that does not
follow the thermodynamic effect. This implies that the transition
state is closer to the reactants than the products. If the transition
state were closer to the products, we would expect the kinetic
alkyl group effect to be larger and more influenced by the
thermodynamic stabilities of the products. Second, the primary
isotope effect is base dependent, becoming smaller as the
reaction exothermicity increases. This also implies that the
transition state is close to the reactants. Third, we have a
secondary isotope effect that is greater than one. In contrast to
3
24
carbon is still essentially sp hybridized in the transition state.
Saunders has observed a carbon isotope effect, and Bordwell
has observed a secondary isotope effect in the proton transfer
reactions of nitroalkanes, both of which suggest that some
2
6,27
rehybridization must be occurring in the transition state.
Thus, proton transfer, rehybridization, and charge delocalization
have all made different progress at the transition state. These
2
8,29
issues have also been addressed computationally.
3
2
the first two points, this implies that rehybridization (sp f sp )
is occurring in the transition state and that the transition state
resembles the products. In order for the secondary isotope effect
to be as small as unity, the error in our data would have to be
very large (+15%), and we believe our data to be accurate to
In summary, the kinetic alkyl group effect on the proton
transfer rate is due to a combination of factors. In the transition
state, the competing effects of proton transfer (correlates with
charge transfer), charge delocalization, and rehybridization
almost cancel each other out and lead to a small alkyl group
effect. Steric hindrance does not appear to be an important
factor in determining our kinetic alkyl group effect. Thus, the
order for proton transfer from a carbonyl group is 2° > 3° >
(
5%.
The principle of nonperfect synchronization, developed by
Bernasconi, can help explain these observations.24 In nonperfect
synchronization, transition state imbalances occur when the
various processes involved in the reaction (e.g. proton transfer,
charge delocalization, and rehybridization) do not progress at
the same rate. At the transition state, proton transfer, which
correlates with charge transfer, generally has made more
progress than charge delocalization, which leads to resonance
stabilization. Rehybridization can still have progressed signifi-
cantly at the transition state even if the charge remains localized.
In the case of proton transfer in ketones, charge is transferred
1
° in our system.
Acknowledgment. We are grateful to the National Science
Foundation for support of this research.
Appendix A: Derivation of Alkyl Group Effects
RCH2 vs CH3 from RCH2C(O)CD3 Compounds. The rate
of abstraction of a proton from the above deuterated ketone is
equal to twice the rate constant for deprotonation of a single
3
to the R-carbon; the R-carbon undergoes rehybridization (sp
(R)CH2
2
proton from the (R)CH substituent, k
, while the rate of
f sp ); and the charge is delocalized with the charge mostly
2
abstraction of a deuteron from the above deuterated ketone is
three times the rate constant for deprotonation of a single
on the oxygen in the final product.
If, in the transition state, all the transferred charge is localized
on the R-carbon instead of being delocalized and the R-carbon
CD3
deuteron from the CD3 substituent, k . Since both enolate
3
ion products come from the same transition state, the internal
energy distribution will be the same for both proton transfers
and the isomeric enolate ion product ratio can be expressed as:
is sp hybridized, the transition state for proton transfer from
the primary position in 2-butanone would be more stable than
the transition state for proton transfer from the secondary
position in 2-butanone. The electron affinity of the primary
enolate ion of 2-butanone is 2 kcal/mol greater than the electron
(
R)CH2
M - H 2 k
product ratio )
)
2
5
affinity for the secondary enolate ion of 2-butanone, which
indicates that the primary enolate ion is the more stable ion
relative to the neutral enolate radical and that, relative to a
hydrogen, the methyl group destabilizes a negative charge on a
carbon. Thus, in this case, the rate of proton transfer from the
primary position of 2-butanone would be expected to be faster
than the rate of proton transfer from the secondary position.
On the other hand, if, in the transition state, the transferred
charge is delocalized so that there is no charge on the R-carbon
M - D
3
CD3
k
where M - H represents the abundance of the product formed
when the deuterated ketone loses a proton and M - D represents
the abundance of the product formed when the deuterated ketone
loses a deuteron. The desired alkyl group effect (eq 5 in the
text) then can be related to the measured isomeric enolate ion
product ratio and the primary isotope effect as
3
2
and the R-carbon has rehybridized (sp f sp ), the transition
state for proton transfer from the secondary position of 2-bu-
tanone would be more stable than the transition state for proton
transfer from the primary position of 2-butanone because alkyl
groups, relative to hydrogen, favor rehybridization. In general,
the more highly substituted double bond is more stable than
the less substituted double bond. In this case, the rate of proton
(
R)CH2
(R)CH2
CH3
CD3
CH3
CD3
k
k
k
k
3
2
M - H
÷
M - D
k
k
)
÷
)
×
CH3
CD3
k
k
The alkyl group effect refers to the relative rate of deprotonation
of a single proton form two different alkyl substituents.
Equations 6-8 can be derived in an analogous fashion.

Substituent Effect in the Deprotonation of Ketones
J. Phys. Chem., Vol. 100, No. 21, 1996 8835
References and Notes
(15) Frieser, B. S. In Techniques in Chemistry Volume XX: Techniques
for the Study of Ion-Molecule Reactions; Farrar, J. M., Saunders, W. H.,
Jr., Eds.; Wiley and Sons: New York, 1988; pp 61-118.
(16) Johnson, C. E.; Brauman, J. I. In Techniques of Chemistry Volume
XX: Techniques for the Study of Ion-Molecule Reactions; Farrar, J. M.,
Saunders, W. H., Jr., Eds.; Wiley and Sons: New York, 1988; pp 563-
589.
(17) Johnson, C. E.; Sannes, K. A.; Brauman, J. I. In preparation.
(18) Tumas, W.; Foster, R. F.; Brauman, J. I. J. Am. Chem. Soc. 1988,
110, 2714-2722.
(19) The â-deuterium isotope effect is small, on the order of 1.1.
Streitwieser, A., Jr.; Van Sickle, D. E. J. Am. Chem. Soc. 1962, 84, 254-
258.
(
1) Multiple proton exchanges have been observed in long-lived ion-
molecule complexes of allylic type anions (Squires, R. R.; Bierbaum, V.
M.; Grabowski, J. J.; DePuy, C. H. J. Am. Chem. Soc. 1983, 105, 5185-
5
192.
(2) (a) Chyall, L. J.; Brickhouse, M. D.; Schnute, M. E.; Squires, R.
R. J. Am. Chem. Soc. 1994, 116, 8681-8690. (b) S u¨ rig, T.; Gr u¨ tzmacher,
H.-Fr. Org. Mass Spectrom. 1989, 24, 851.
(3) Tumas, W.; Foster, R. F.; Pellerite, M. J.; Brauman, J. I. J. Am.
Chem. Soc. 1987, 109, 961-970.
(
(
4) Sulzle, D.; Schwarz, H. HelV. Chim. Acta 1989, 72, 320-322.
5) Mercer, R. S.; Harrison, A. G. Can. J. Chem. 1988, 66, 2947-
2
953.
(20) Derrick, P. J. Mass Spectrom. ReV. 1983, 2, 285.
(21) House, H. O. Modern Synthetic Reactions, 2nd ed.; W. A. Benjamin,
Inc.: Menlo Park, CA, 1972; pp 499-500.
(
6) Chowdhury, S. K.; Mercer, R. S.; Harrison, A. G.; Barlow, S. E.;
DePuy, C. H. Int. J. Mass Spectrom. Ion Processes 1989, 95, 55-69.
7) Raftery, M. J.; Bowie, J. H.; Sheldon, J. C. J. Chem. Soc., Perkin
(
(
(
(
(
22) See ref 2, however, for some interesting counterexamples.
Trans. 2 1988, 563-569.
23) Westheimer, F. H. Chem. ReV. 1961, 61, 265-273.
24) Bernasconi, C. F. AdV. Phys. Org. Chem. 1992, 27, 119-238.
25) Zimmerman, A. H.; Reed, K. J.; Brauman, J. I. J. Am. Chem. Soc.
(
8) Sheldon, J. C.; Bowie, J. H. New J. Chem. 1988, 12, 269.
(9) Hayes, R. N.; Sheldon, J. C.; Bowie, J. H.; Lewis, D. E. Aust. J.
Chem. 1985, 38, 1197-1208.
10) Hayes, R. N.; Sheldon, J. C.; Bowie, J. H.; Lewis, D. E. J. Chem.
Soc., Chem. Commun. 1984, 1431.
11) Haib, J.; Stahl, D. Int. J. Mass Spectrom. Ion Processes 1990, 95,
1
1
3
5
977, 99, 7203.
26) Wilson, J. C.; K a¨ llsson, I.; Saunders, W. H., Jr. J. Am. Chem. Soc.
980, 102, 4780-4784.
27) Bordwell, F. G.; Boyle, W. J., Jr. J. Am. Chem. Soc. 1975, 97,
447-3452.
28) Bernasconi, C. F.; Wenzel, P. J. J. Am. Chem. Soc. 1994, 116,
405-5413.
29) Saunders, W. H., Jr. J. Am. Chem. Soc. 1994, 116, 5400-5404.
(
(
(
(
2
89-297.
(
12) Hanst, P. L.; Calvert, J. G. J. Phys. Chem. 1959, 63, 104.
13) Organic Syntheses, CollectiVe Volume 2; Blatt, A. H., Ed.; Wiley
(
(
and Sons: New York, 1943; pp 406 and 602.
14) Fieser, L. F.; Fieser, M. Reagents for Organic Synthesis; Wiley
and Sons: New York, 1967; pp 418 and 419.
(
(
JP953249P