Reactions of aryl phenylacetates with secondary amines in MeCN. Structure-reactivity relationship in the ketene-forming eliminations and concurrent E2 and E1cb mechanisms

Article abstract of DOI:10.1021/ja961294k

Elimination reactions of aryl esters of arylacetic acids 1 and 2 promoted by R₂NH in MeCN have been investigated kinetically. The reactions are second-order and exhibit β = 0.44-0.84, β(lg) = 0.41-0.50, and ρ(H) = 2.0-3.6. Bronsted β and β(lg) decrease with the electron-withdrawing ability of the β-aryl substituent. Hammett ρ(H) values remain nearly the same, but the β(lg) value increases as the base strength becomes weaker. Both ρ(H) and β decrease with the change of the leaving group from 4-nitrophenoxide to 2,4-dinitrophenoxide. The results are consistent with an E2 mechanism and a reaction coordinate with a large horizontal component corresponding to proton transfer. When the base-solvent system is changed from R₂NH-MeCN to R₂NH/R₂NH₂⁺-70 mol% MeCN(aq), the Bronsted β, ρ(H), and β(lg) decrease. Finally, the ketene-forming elimination reactions from p-nitrophenyl p-nitrophenylacetate promoted by R₂NH/R₂NH₂⁺ buffers in 70 mol% MeCN(aq) have been shown to proceed by concurrent E2 and E1cb mechanisms.

Full text of DOI:10.1021/ja961294k

J. Am. Chem. Soc. 1997, 119, 691-697
691
Reactions of Aryl Phenylacetates with Secondary Amines in
MeCN. Structure-Reactivity Relationship in the
Ketene-Forming Eliminations and Concurrent E2 and E1cb
Mechanisms
Bong Rae Cho,* Yong Kwan Kim, and Choon-Ock Maing Yoon
Contribution from the Department of Chemistry, Korea UniVersity,
1-Anamdong, Seoul 136-701, Korea
ReceiVed April 19, 1996. ReVised Manuscript ReceiVed October 29, 1996X
Abstract: Elimination reactions of aryl esters of arylacetic acids 1 and 2 promoted by R2NH in MeCN have been
investigated kinetically. The reactions are second-order and exhibit â ) 0.44-0.84, |âlg| ) 0.41-0.50, and FH )
2.0-3.6. Brønsted â and |âlg| decrease with the electron-withdrawing ability of the â-aryl substituent. Hammett FH
values remain nearly the same, but the |âlg| value increases as the base strength becomes weaker. Both FH and â
decrease with the change of the leaving group from 4-nitrophenoxide to 2,4-dinitrophenoxide. The results are consistent
with an E2 mechanism and a reaction coordinate with a large horizontal component corresponding to proton transfer.
When the base-solvent system is changed from R2NH-MeCN to R2NH/R2NH2+-70 mol % MeCN(aq), the Brønsted
â, FH, and |âlg| decrease. Finally, the ketene-forming elimination reactions from p-nitrophenyl p-nitrophenylacetate
+
promoted by R2NH/R2NH2 buffers in 70 mol % MeCN(aq) have been shown to proceed by concurrent E2 and
E1cb mechanisms.
Extensive studies of structure-reactivity relationships in
elimination reactions have led to a qualitative understanding of
the relationship between the reactant structure and the E2
transition state.1-8 Most of the results from these studies have
been interpreted with More-O’Ferrall-Jencks energy diagram
by assuming similar effects for the parallel and perpendicular
motion.5-8 In contrast, much less is known about the effects
of the reactant structure on the E1cb-like transition state, where
the effects in the two directions are different.9-11
Perhaps an even more important problem in elimination
mechanism studies may be the mechanistic borderlines. It is
conceivable that the change of the elimination mechanism may
occur by “merging” of the transition states, i.e., the concerted
mechanism is enforced by the change in the lifetime of the
intermediate. The mechanistic changes from E2 to E1cb for
eliminations reactions of (2-arylethyl)ammonium ions9 and 9-(2-
chloro-2-propyl)fluorene12 have been concluded to be of this
type. Alternatively, the change in mechanism may also involve
two concurrent mechanisms having different transition state
structures. A change in experimental condition or reactant
structure may lower the energy of one of the transition states,
which may in turn induce a shift in the major reaction pathway.
Such examples have been reported in the E2 and E1 borderline
for the methoxide-promoted eliminations from 2-chloro-2-
methyl-1-phenylpropane13 and N-(arylsulfonoxy)-N-alkylben-
zylamines14 as well as for the solvolytic elimination of 9-(1-
X-ethyl)fluorenes.15 However, no clear evidence for the
concurrent E2 and E1cb mechanisms has been reported.
Although the small isotope effect observed in eliminations from
1-(2-chloro-2-propyl)indene has been interpreted with parallel
E2 and E1cb mechanisms, the result could also be attributed to
a reverse stepwise preassociation mechanism.16
One of the promising compounds that may be useful to
investigate both of these problems is the aryl esters of phenyl-
acetic acids. It has been reported that the base-catalyzed
hydrolysis of aryl p-nitrophenylacetates17-22 and other activated
esters22 proceed by an E1cb elimination to afford the ketene
X Abstract published in AdVance ACS Abstracts, December 15, 1996.
(1) Saunders, Jr. W. H.; Cockerill, A. F. Mechanism of Elimination
Reactions; Wiley: New York, 1973.
(13) Thibblin, A. J. Am. Chem. Soc. 1989, 111, 5412-5416.
(14) Cho, B. R.; Pyun, S. Y. J. Am. Chem. Soc. 1991, 113, 3920-3924.
(15) Meng, Q.; Thibblin, A. J. Am. Chem. Soc. 1995, 117, 9399-9407.
(16) Olwegard, M.; McEwen, I.; Thibblin, A.; Ahlberg, P. J. Am. Chem.
Soc. 1985, 107, 7494-7499.
(2) Bartsch, R. A.; Zavada, J. Chem. ReV. 1980, 80, 453-494.
(3) Lowry, T. H.; Richardson, K. S. Mechanism and Theory in Organic
Chemistry; Harper and Row: New York, 1987; (a) pp 214-218, (b) pp
591-560, (c) pp 640-644.
(17) Holmquist, B.; Bruice, T. C. J. Am. Chem. Soc. 1969, 91, 3003-
3006.
(4) Hoffman, R. V.; Bartsch, R. A.; Cho, B. R. Acc. Chem. Res. 1989,
22, 211-217.
(18) Tagaki, W.; Kobayashi, S.; Kurihara, K.; Kurashima, K.; Yoshida,
Y.; Yano, J. J. Chem. Soc., Chem. Commun. 1976, 843-845.
(19) Chandrasekar, R.; Venkatasubramanian, N. J. Chem. Soc., Perkin
Trans. 2 1982, 1625-1631.
(20) Broxton, T. J.; Duddy, N. W. J. Org. Chem. 1981, 46, 1186-1191.
(21) Chung, S. Y.; Yoh, S. D.; Choi, J. H.; Shim, K. T. J. Korean Chem.
Soc. 1992, 36, 446-452.
(22) (a) Holmquist, B.; Bruice, T. C. J. Am. Chem. Soc. 1969, 91, 2993-
3002. (b) Pratt, R. F.; Bruice, T. C. J. Am. Chem. Soc. 1970, 92, 5956-
5964. (c) Inoue, M.; Bruice, T. C. J. Am. Chem. Soc. 1982, 104, 1644-
1653. (d) Inoue, M.; Bruice, T. C. J. Org. Chem. 1982, 51, 959-963. (e)
Isaac, N. S.; Najem, T. S. J. Chem. Soc., Perkin Trans. 2 1988, 557-562.
(f) Douglas, K. T.; Alborz, M.; Rullo, G. R.; Yaggi, N. F. J. Chem. Soc.,
Chem. Commun. 1982, 242-246. (g) William, A. J. Chem. Soc., Perkin
Trans. 2 1972, 808-812. (h) Williams, A.; Douglas, K. T. Chem. ReV.
1975, 627-649.
(5) Cho, B. R.; Kim, K. D.; Lee, J. C.; Cho, N. S. J. Am. Chem. Soc.
1988, 110, 6145-6148.
(6) Cho, B. R.; Jung, J.; Ahn, E. K. J. Am. Chem. Soc. 1992, 114, 3425-
3458 and references cited therein.
(7) Gandler, J. R. In The Chemistry of Double Bonded Functional
Groups; Patai, S., Ed.; Chichester: New York, 1989; Vol. 2, Part 1, pp
734-797.
(8) Jencks, W. P. Chem. ReV. 1985, 85, 511-527.
(9) Gandler, J. R.; Jencks, W. P. J. Am. Chem. Soc. 1982, 104, 1937-
1951.
(10) Gandler, J. R.; Yokohama, T. J. Am. Chem. Soc. 1984, 106, 130-
135.
(11) Gandler, J. R.; Storer, J. W.; Ohlberg, D. A. A. J. Am. Chem. Soc.
1990, 112, 7756-7762.
(12) Thibblin, A. J. Am. Chem. Soc. 1988, 110, 4582-4586.
S0002-7863(96)01294-2 CCC: $14.00 © 1997 American Chemical Society

692 J. Am. Chem. Soc., Vol. 119, No. 4, 1997
Cho et al.
Table 1. Rate Constants for Aminolysis of Aryl Alkanoatesa
Promoted by R2NH in MeCN at 25.0 °C
clean isosbestic points were noted in the range 286-318 nm.
Excellent pseudo-first-order kinetic plots which covered at least
3 half-lives were obtained. For reactions of 1a-d and 2a-e
with R2NH, the plots of kobs vs base concentration were straight
lines passing through the origin. The slopes were the overall
second-order rate constants k2.
To probe whether the aminolysis reaction may compete with
the ketene-forming elimination, the rates of aryloxides produc-
tion from the reactions of aryl alkanoates 3 and 4 with R2NH
were determined (eq 2). Except for the reaction of 3a with
S
102k2 , M-1 s-1 d
baseb
pKac
4a
4b
4c
4d
2ae
Bz(i-Pr)NH 16.8
0.358
56.2
1.54
0.115
0.060 0.020 0.0004
i-Bu2NH
i-Pr2NH f
18.2
18.5
24.2
1.04
47.3
15.1
0.620 1.04
29.2 8.50
5.40
0.1
0.006
0.175
2,6-DMPg,h 18.9 113
a [Substrate] ) 2.0 × 10-5 M. b [Base] ) 8.5 × 10-3 to 5.0 × 10-1
M. c Reference 36. d Estimated uncertainty, (5%. e Estimated from the
Charton plot and the steric effect parameter 0.7 for the benzyl group
(see text).24 f 104k2S values for the reactions of 3a, 3b, 3c, and 3d with
i-Pr2NH/i-Pr2NH2+ buffer in 70 mol % MeCN(aq) are 1.6, 1.08, 6.16,
5.87 M-1 s-1, respectively. g cis-2,6-Dimethylpiperidine. h The k2S value
for the reaction of 3a with 2,6-DMP is 2.03 × 10-4 M-1 s-1
.
intermediate followed by the addition of water under various
conditions. It occurred to us that the ketene-forming elimina-
tions may proceed by the E2 mechanism via an E1cb-like
transition state if the reactions were conducted under conditions
where the carbanion intermediate is destabilized. Concurrent
E2 and E1cb mechanisms may also be observed if the reaction
conditions are varied so that the energies of the two transition
states become almost the same. Accordingly, we have inves-
tigated the reactions of aryl esters of substituted phenylacetatic
acids 1 and 2 with secondary amines in acetonitrile (eq 1).
Sterically bulky amines have been employed to avoid the
complications that may be caused by the aminolysis of the
substrates.
2,6-DMP, 3 did not react with R2NH under the reaction
condition. However, 4 underwent aminolysis with all of the
R2NH employed. The second-order rate constants k2 for the
S
aminolysis reactions are summarized in Table 1. The rate
increased as the basicity of the amine increased but decreased
as the steric bulk of the alkyl group increased. The slower rate
of aminolysis from 4 with i-Pr2NH than with i-Bu2NH as the
base can be attributed to the greater steric requirement of the
former.
The plots of log k2 vs Charton’s steric effect parameter ν24
for the aminolysis of 4 with R2NH in MeCN are shown in Figure
S
S1 in the Supporting Information. The k2 values for the
reactions of 2 with R2NH were estimated from the Charton plots
by using the steric effect parameter of 0.7 for the benzyl group24
S
and are listed in Table 1. In all cases, the k2 values were
negligible compared with the k2 values (Tables 1 and 2).
Therefore, no correction was necessary to obtain the second-
order rate constants for eliminations reactions k2E from k2. The
E
k2 values for the ketene-forming eliminations from 1 and 2
promoted by R2NH in MeCN are summarized in Table 2.
Brønsted plots for eliminations from 1 and 2 are linear with
excellent correlation (plots not shown). The â values are in
the range 0.44-0.84 and decrease as the electron-withdrawing
ability of the substituent and the leaving group ability of the
aryloxides increase (Table 3).
The Hammett plots for eliminations from 1a-d are curvi-
linear and show downward curvature (Figure 1). Similar plots
were observed for the reactions of 2a-e (plots not shown). The
rate data could be correlated with eq 3 using a nonlinear
In this work, we have studied the effects of reactant structures
upon the ketene-forming eliminations. This reaction provided
us with an unusual opportunity to look into the structure-
reactivity correlations for the E1cb-like transition state. The
concurrent E2 and E1cb reactions have also been observed. The
results of these studies are reported here.
log(k/k0) ) FHσ- + τσ-2 ) (FH + τσ-)σ-
(3)
Results
regression analysis program,25 where FH and τ values are the
slope of the Hammett plot when Y ) H and the proportionality
factor by which the F value is influenced by the electron-
withdrawing ability of the substituents, respectively. For all
reactions the correlations between the experimental and the
calculated data were excellent, demonstrating the validity of
this analysis. Utilizing the computer program, the FH and τ
values that best fit with eq 3 have been calculated. The FH and
τ values are in the range 2.0-3.6 and -0.43 to -2.3 (Table 4).
The values are independent of the base strength but decrease
as the pKa value of the leaving group is decreased (Table 4).
The âlg values are calculated from the rate data and the pKlg
values of the leaving group. The |âlg| values are in the range
Aryl phenylacetates 1 and 2 and aryl alkanoates 3 and 4 were
prepared by the literature method.23 The reactions of 1a and
2a with diisopropylamine in MeCN produced N,N-diisopropyl-
benzamide in 88 and 92% yields, respectively. From the
+
reactions of 1a and 2a with i-Pr2NH/i-Pr2NH2 in 70 mol %
MeCN(aq) were obtained N,N-diisopropylbenzamide and phe-
nylacetic acid as the only products. The yields of the aryloxides
as determined by comparison of the UV absorption of the
infinity sample of the kinetic runs with those for the authentic
samples were in the range 92-98%.
The rates of the reactions of 1 and 2 with R2NH-MeCN
were followed by monitoring the increase in the absorption for
the aryloxides at 400 and 426 nm, respectively. In all cases,
(24) Charton, M. J. Am. Chem. Soc. 1975, 97, 1552-1556.
(25) MicroCal Origin Version 3.5, MicroCal Software, Inc., Northampton,
MA, 1991.
(23) Saigo, K.; Usui, M.; Kikuchi, K.; Shimada, E.; Mukaiyama, T. Bull.
Chem. Soc. Jpn. 1977, 50, 1863-1866.

Reactions of Aryl Phenylacetates
J. Am. Chem. Soc., Vol. 119, No. 4, 1997 693
Table 2. Rate Constants for Ketene-Forming Eliminations from YC6H4CH2CO2Ara Promoted by R2NH in MeCN at 25.0 °C
E
E
102k2 , M-1 s-1 c
k2 , M-1 s-1 c
baseb
1a
1b
1c
1d
2a
2b
2c
2d
2e
Bz(i-Pr)NH
i-Bu2NH
0.260
3.80
6.58
0.040
0.575
1.14
2.81
38.9
59.4
12.4
97.7
166
350
0.264
3.02
4.17
7.10
0.0874
0.995
1.39
2.32
13.4
18.8
35.6
10.5
60.3
77.6
60.0
260
346
508
i-Pr2NH
2,6-DMPd
14.2
2.21
146
2.96
135
a [Substrate] ) 2.0 × 10-5 M. b [Base] ) 1.0 × 10-3 to 2.0 × 10-1 M. c Estimated uncertainty, (5%. d cis-2,6-Dimethylpiperidine.
Table 3. Effect of Aryl Substituents upon Brønsted â and âlg
Values for Eliminations from YC6H4CH2CO2Ar Promoted by R2NH
in MeCN at 25.0 °C
the kinetic data cannot be analyzed accurately because the R2-
NH2 concentration increases as the reaction proceeds (eq 4).
+
â
Ar )
4-nitrophenyl
Ar )
2,4-dinitrophenyl
a
Y
âlg
p-CH3O 0.84 ( 0.01
0.72 ( 0.02
0.69 ( 0.04
0.56 ( 0.02
0.52 ( 0.02
0.44 ( 0.01
-0.50
-0.43
-0.41
-0.41
-
H
0.82 ( 0.01 (0.77 ( 0.03)b
m-Cl
m-NO2
p-NO2
0.81 ( 0.02
0.68 ( 0.03
-
(0.66 ( 0.07)b
a The base was Bz(i-Pr)NH. b Calculated with the k2E values in Table
5 for eliminations from 1a and 1e promoted by R2NH/R2NH2+ buffers
in 70 mol % MeCN(aq) at 25.0 °C.
+
Hence the base-solvent system was changed to R2NH/R2NH2
buffer in 70 mol % MeCN(aq). Large excess amount of the
buffer was used to ensure that the R2NH2 concentration
+
remained constant. The kinetic experiments were carried out
as described above. A typical plot of the kobs vs buffer
+
concentration for the reaction of 1a with i-Pr2NH/i-Pr2NH2
buffer in 70 mol % MeCN(aq) is shown in Figure S3. For all
reactions, the plots showed downward curvature. The curvature
is essentially independent of the substrate structure, indicating
that it does not represent a change in rate-limiting step but may
be caused by the buffer association. The buffer association
constant Kassoc was estimated from the negative deviation of
the rate data for reactions of 1a with R2NH/R2NH2+ buffers by
assuming that it occurs due to the formation of kinetically
inactive hydrogen-bonded complexes, R2NH2+‚‚‚NHR2.26 The
Kassoc values are 6.5, 6.0, 7.3, and 20 M-1 for benzylisopropyl-
amine, diisobutylamine, diisopropylamine, and 2,6-dimethylpi-
peridine, respectively. The much larger Kassoc value for the latter
may in parts be attributed to the smaller steric requirement and
lower solubility of its buffer in 70 mol % MeCN(aq). The
second-order rate constants k2E for eliminations from 1a-d and
2a promoted by R2NH/R2NH2+ buffer in 70 mol % MeCN(aq)
were obtained from the slopes of the plots of the kobs vs “free”
buffer concentration calculated with the Kassoc values. The plots
are linear for all reactions. No correction for the competing
Figure 1. Hammett plots for the ketene-forming eliminations from
p-nitrophenyl arylacetate 1 [YC6H4CH2COOC6H4-p-NO2] promoted by
R2NH in MeCN at 25.0 °C. The data are from Table 2, and the lines
were calculated by nonlinear regression analysis by using eq 3 (see
text) [R2NH ) Bz(i-Pr)NH (9), i-Bu2NH (b), i-Pr2NH (2), 2,6-DMP
(1)].
E
aminolysis reaction has been made to the k2 values because
the rates of aminolysis of 3a-d are negligible (see footnote in
E
Table 1). The k2 values are summarized in Table 5.
Table 4. Effect of Base Strength upon the FH, τ, and âlg Values
for Eliminations from YC6H4CH2CO2C6H3-2-X-4-NO2 Promoted by
R2NH in MeCN at 25.0 °C
Figure 2 shows the plot of the kobs for the reactions of 1e
with i-Pr2NH/i-Pr2NH2+ buffer vs free buffer concentration. The
data were analyzed with a nonlinear regression analysis
program27 by assuming that the reaction proceeds by concurrent
E2 and E1cb mechanisms. Utilizing the computer program, the
X ) H
X ) NO2
a
base
FH
τ
FH
τ
âlg
E
k2 , k1, and k-1/k2 values that best fit with eq 4 were calculated,
Bz(i-Pr)NH 3.5 ( 0.1 -2.2 ( 0.1 2.0 ( 0.1 -0.44 ( 0.07 -0.43
and the plots were dissected into the E2 and E1cb reaction
components (Figure 2). In all cases, the correlations between
the calculated and the experimental data are excellent (Figures
i-Bu2NH
i-Pr2NH
3.3 ( 0.1 -1.9 ( 0.1 2.1 ( 0.1 -0.43 ( 0.10 -0.40
3.6 ( 0.1 -2.3 ( 0.1 2.1 ( 0.1 -0.43 ( 0.08 -0.38
2,6-DMPb 3.5 ( 0.1 -1.6 ( 0.2 2.1 ( 0.1 -0.48 ( 0.10 -0.37
a Y ) H. b cis-2,6-Dimethylpiperidine.
E
S4-S6). Calculated k2 , k1, and k-1/k2 values are summarized
in Table 5.
0.37-0.50 and decrease as the electron-withdrawing ability of
the aryl substituent and the pKa value of the base increase
(Tables 3 and 4).
In contrast to the reactions of 1a-d and 2a-e, the kobs for
the reactions of 1e with R2NH in MeCN showed a curvilinear
relationship with the base concentration (Figure S2). However,
Brønsted plots for eliminations from 1a and 1e promoted by
R2NH/R2NH2+ buffer in 70 mol % MeCN(aq) are straight lines
(26) Keeffee, J. R.; Jencks, W. P. J. Am. Chem. Soc. 1983, 105, 265-
279.
(27) A data analysis and graphical plotting program “GENPLOT”,
Computer Graphic Service, Lansing, New York, 1989.

694 J. Am. Chem. Soc., Vol. 119, No. 4, 1997
Cho et al.
E
Table 5. The k2 , k1, and k-1/k2 Values for Elimination from
Scheme 1
a
+
YC6H4CH2COOC6H4-p-NO2 Promoted by R2NH/R2NH2 Buffers in
70 mol % MeCN(aq) at 25.0 °C
Y ) H
Y ) p-NO2
E c
E c,d
102k2 ,
k2 ,
k1,c,d
c,d
R2NHb
M-1 s-1
M-1 s-1
M-1 s-1
k-1/k2
Bz(i-Pr)NH
i-Bu2NH
0.173
2.00
0.58
3.46
6.41
16.0
1.81
18.0
37.0
55.7
387
240
202
142
i-Pr2NH
4.00e
6.60
2,6-DMP f
a [Substrate] ) 2.0 × 10-5
using eq 4 by nonlinear regression analysis (see text). e The k2E values
.
b [Buffer] ) 1.0 × 10-3 to 2.0 × 10-1
Figure S7. In contrast to the reactions of 1 and 2 with R2NH
in MeCN, the plot is linear, i.e., τ ) 0 in eq 3. The calculated
M. c Estimated uncertainty, (5%. d Calculated from the kobs values
E
k2 value for 1e falls on a single line with the experimental
+
for reactions of 1b-d and 2a with i-Pr2NH/i-Pr2NH2 buffer in 70
data for 1a-d with excellent correlation. The Hammett F value
is 1.8 ( 0.1 (Table 6).
mol % MeCN(aq) are 0.016, 0.200, 0.635, and 0.676 M-1 s-1 for 1b,
1c, 1d, and 2a, respectively. f cis-2,6-Dimethylpiperidine.
+
For reactions of 1e with R2NH/R2NH2 buffer in 70 mol %
MeCN(aq), the k1 values increased but the k-1/k2 decreased as
the basicity of the buffer increased. The plots of log k1 and
log k-1/k2 vs the pKa values of the buffers were linear with
excellent correlation (Figure S8). The slopes of the plots are
0.73 ( 0.04 and -0.19 ( 0.03, respectively.
To provide an additional evidence for the competing (E1cb)R
mechanism, the H-D exchange experiment was carried out by
mixing the reactants with i-Pr2NH/i-Pr2NH2+ buffer in 70 mol
% MeCN-30% D2O at -10 °C. The reactant was recovered
immediately after mixing. The NMR spectrum of the recovered
reactant indicated that approximately half of benzylic C-H bond
was converted to the C-D bond.
Discussion
Reaction Mechanism and Transition State Structure for
Eliminations from 1a-d and 2a-e Promoted by R2NH in
MeCN. Reactions of 1a-d and 2a-e with R2NH in MeCN
produced benzamides and aryloxide in nearly quantitative yields.
The reactions proceed by the amine-promoted elimination to
afford the ketenes followed by the addition of the R2NH to the
intermediate to afford the products. Addition-elimination
mechanism (BAC2) is ruled out by the negligible rates of
aminolysis compared with the overall rates (Tables 1 and 2). A
similar mechanism has been reported in closely related
reactions.17-21
Figure 2. Dependence of kobs vs free buffer concentration for the
ketene-forming eliminations from p-nitrophenyl p-nitrophenylacetate
1e [p-O2NC6H4CH2COOC6H4-p-NO2] promoted by i-Pr2NH/i-Pr2NH2
in 70 mol % MeCN(aq) at 25.0 °C. The closed circles show
experimental data and the solid line shows the computer-fitted curve
by using eq 4 (see text). The curve is dissected into the E2 and E1cb
reaction components (dashed lines).
+
Table 6. Solvent Effect on the Ketene-Forming E2 Reaction of
YC6H4CH2COOC6H4-p-NO2
Under the reaction condition reported here, the ketene-forming
eliminations from 1 and 2 must proceed either via a one-step
concerted mechanism or via carbanion intermediates (free or
ion-paired) in a stepwise mechanism as depicted in Scheme 1.7
The observed general base catalysis with the Brønsted â
values ranging from 0.44 to 0.84 rule out the (E1cb)R, (E1cb)ip,
and internal return mechanisms because these mechanisms
would exhibit either a specific base catalysis or Brønsted â
values near unity.3,7 The possibility that the relatively large â
values for eliminations from 1a-c and 2b are caused by the
change to an E1cb mechanism is negated by the interaction
coefficients, i.e., pxy′ < 0, pxy > 0, and pyy′ < 0 (Vide infra). In
addition, the values of |âlg| ) 0.41-0.50 rule out a mechanism
in which either k1 [(E1cb)irr] or kd is rate limiting, for which a
small or negligible leaving group effect is expected.3,7 On the
other hand, the results are consistent with an E2 mechanism in
which there is partial cleavage of the Câ-H and CR-OAr bonds
in the transition state.
solvent
MeCN
70 mol % MeCN(aq)
rel ratea,b
1
0.6
b
FH
3.6 ( 0.1
1.8 ( 0.1
0.77 ( 0.03
-0.26
âc
âlg
0.82 ( 0.01
a,c
-0.38
a Y ) H. b Base-solvent ) i-Pr2NH in MeCN and i-Pr2NH/i-
Pr2NH2+ buffer in 70 mol % MeCN(aq), respectively. c Base-solvent
) R2NH in MeCN and R2NH/R2NH2+ buffer in 70 mol % MeCN(aq),
respectively.
with excellent correlations (plots not shown). Since the pKa
+
values of R2NH2 in 70 mol % MeCN(aq) are not known in
the literature, the (pKa)MeCN values have been used to calculate
the â value.28,29 The â values for the k2E processes are 0.77 (
0.03 and 0.66 ( 0.07 for 1a and 1e, respectively (Tables 3 and
6).
A Hammett plot for eliminations from 1a-e promoted by
+
The structure of the transition states may be assessed by the
Brønsted â, Hammett F, and âlg values. The Brønsted â values
indicate the extent of proton transfer in the transition state. For
R2NH-promoted eliminations from 1a-d and 2a-e, values of
â ) 0.44-0.84 were determined (Table 3). This indicates
R2NH/R2NH2 buffer in 70 mol % MeCN(aq) is depicted in
(28) Coetzee has shown that the ∆pKa ) (pKa)MeCN - (pKa)H O values
2
for the structurally similar amines are nearly the same.29 Therefore, the
differences in the pKa values of the R2NH employed in this study should
be nearly the same in both MeCN and 70 mol % MeCN(aq). Accordingly,
no correction has been made to the (pKa)MeCN values in the Brønsted plots
of the rate data measured in 70 mol % MeCN(aq).
(29) Coetzee, J. F. Prog. Phys. Org. Chem. 1965, 4, 45-92.

Reactions of Aryl Phenylacetates
J. Am. Chem. Soc., Vol. 119, No. 4, 1997 695
moderate to extensive cleavage of the Câ-H bonds in the
transition states.
The extent of negative charge development in the transition
state may be assessed with the Hammett F values. Surprisingly,
however, the Hammett plots are curvilinear and show downward
curvature (Figure 1). Causes for the curved Hammett plots have
been attributed to (i) the change in mechanism, (ii) a single
mechanism with a different extent of bond formation and
cleavage in the transition state, and (iii) a different balance of
polar and resonance effects by different substituents in the
transition state.30-32 Since the reactions of 1a-d and 2a-e
with R2NH in MeCN proceed by a common E2 mechanism
(Vide supra), possibility (i) can be ruled out. Possibility (iii)
can also be negated because the resonance effect is already taken
into account in the σ- constants. In addition, the rate data
showed excellent correlations with eq 3, i.e., F ) FH + τσ-.
Therefore, the observed FH values of 3.5-2.0 for the ketene-
forming eliminations can be interpreted with a significant
negative charge development at the â-carbon in the transition
states, which decreases gradually as the electron-withdrawing
ability of the substituent increases (Vide infra).
Figure 3. Reaction coordinate diagram for ketene-forming eliminations.
The effects of the change to a stronger electron-withdrawing substituent,
a better leaving group, and a weaker base are shown by the shift of the
transition state from A to B, A to C, and A to D in I, II, and III,
respectively.
The âlg values are usually taken as a qualitative measure of
the extent of the leaving group cleavage. For Bz(i-Pr)NH-
promoted eliminations from 1a-d, the |âlg| values of 0.41-
0.50 have been determined (Table 3). The results are consistent
with limited extents of CR-OAr bond cleavage in the transition
states. Additional support for this conclusion is provided by
the interaction coefficients, i.e., pxy′ < 0, pxy > 0, and pyy′ < 0
(Vide infra).
The combined results reveal that the ketene-forming elimina-
tions from 1a-d and 2a-e proceed by the concerted E2
mechanism via an E1cb-like transition state with extensive Câ-H
bond cleavage, significant negative charge development at the
â-carbon, and limited CR-OAr bond cleavage (Vide infra). On
the other hand, the reactions of aryl esters of phenylacetic acid
with OH- in 80% DMSO(aq) have been shown to proceed by
the (E1cb)R or (E1cb)ion-pair mechanism.18,20 The change in the
elimination reaction mechanism for 1a from (E1cb)R or
(E1cb)ion-pair to E2 with the base-solvent system variation from
OH--80% DMSO(aq) to R2NH-MeCN can be attributed to a
solvent effect. Since the negative charge that developed on the
benzylic carbon in the E1cb intermediate cannot be stabilized
by solvation in poorly anion-solvating MeCN, the activation
energy for the E1cb mechanism should increase and the E2
mechanism may become the favored alternative.
be described on energy surface of More-O’Ferall-Jencks
diagram.7,8 An energy surface for the elimination reactions of
R2NH-promoted eliminations from 1a-d and 2a-e in MeCN
is shown in Figure 3.
Table 3 shows that the Brønsted â values for R2NH-promoted
eliminations from 1a-d and 2a-e decrease as the electron-
withdrawing ability of the â-phenyl substituent is increased. The
results can be described by a negative pxy′ interaction coefficient,
pxy′ ) ∂â/∂σ ) ∂F/∂pKBH, that describes the interaction between
the base catalyst and the â-aryl substituent.7,8 Negative pxy′
coefficients are consistent with an (E1cb)irr mechanism, which
has been ruled out by the substantial values of the |âlg| (Vide
supra) and an E2 mechanism and a reaction coordinate with a
large horizontal component corresponding to proton transfer.7-11
These changes in the â values can be described on More-
O’Ferrall-Jencks energy diagram (Figure 3). An electron-
withdrawing â-aryl substituent will lower the energy of the
carbanion intermediate in the upper left corner of the energy
diagram. The transition state on the horizontal reaction
coordinate will then move toward the right, with less proton
transfer and a smaller â, as depicted by a shift from A to B in
Figure 3I.8
For all â-aryl substituents the Brønsted â values decrease as
the leaving group is changed from 4-nitrophenoxide to 2,4-
dinitrophenoxide (Table 3). This effect corresponds to a positive
pxy interaction coefficient, pxy ) ∂â/∂pKlg ) ∂âlg/∂pKBH, that
describes the interaction between the base catalyst and the
leaving group.7,8 The observed increase in the |âlg| values as
the catalyst is made less basic (Table 4) is another manifestation
of this effect, i.e., pxy ) ∂âlg/∂pKBH > 0.7,8 On the More-
O’Ferrall-Jencks reaction coordinate diagram in Figure 3, a
change to a better leaving group will raise the energy of the
top edge of diagram. The transition state on the horizontal
reaction coordinate would then move slightly toward the right
as depicted by a shift from A to C on the energy diagram,
resulting in a small decrease in â (Figure 3II).8 Similarly, a
weaker base will raise the energy of the left side of the diagram
and shift the transition state from A to D to increase in the extent
of CR-OAr bond cleavage (Figure 3III).8 The positive pxy
coefficients are not consistent with an E1cb mechanism for
which pxy ) 0 is expected, but provide additional support for
the concerted E2 mechanism.7-11
Structure-Reactivity Coefficients. Changes in structure-
reactivity parameters that reflect changes in the transition state
structure can provide additional evidence for the mechanism
of the elimination reactions.7-11 These changes can usefully
(30) Young, P. R.; Jencks, W. P. J. Am. Chem. Soc. 1979, 101, 3288-
3293.
(31) Drago, R. S.; Zoltewicz, J. A. J. Org. Chem. 1994, 59, 2824-2830
and references cited therein.
(32) (a) Yoshida, T.; Yano, Y.; Oae, S. Terahedron 1971, 27, 5343-
5351. (b) Bunnett, J. F.; Sridharan, S.; Cavin, W. P. J. Org. Chem. 1979,
44, 1463-1471. (c) Hasan, T.; Sims, L. B.; Fry, A. J. Am. Chem. Soc.
1983, 105, 3967-3975.
(33) Cho, B. R.; Maing Yoon, C.-O.; Song, K. S. Tetrahedron Lett. 1995,
36, 3193-3196.
(34) A reviewer has made an interesting suggestion that the change of
the solvent to more protic one would stabilize the upper left corner and the
lower left products in Figure 3. This would produce perpendicular and
parallel shifts whose resultant shows a marginal decrease in â and a decrease
in |âlg| as is seen. However, we believe that the stabilization of the carbonyl
+
oxygen atom by hydrogen bonding with R2NH2 in the buffer solution
would have more profound influence on the transition state structure than
can be described by the energy diagram.
(35) Cho, B. R.; Suh, Y. S.; Lee, S. J.; Cho, E. J. J. Org. Chem. 1994,
59, 3681-3682.
As shown in Table 3, there is a progressive decrease in the
|âlg| values as the â-aryl substituent is changed from p-MeO to

696 J. Am. Chem. Soc., Vol. 119, No. 4, 1997
Cho et al.
H to m-Cl to m-NO2. This result can be described by a negative
pyy′ interaction coefficient, pyy′ ) -∂âlg/∂σ- ) -∂F/∂pKlg, that
describes the interaction between the leaving group and the
â-aryl substituent.7,8 The decrease in the FH values with a better
leaving group (Table 4) provides additional evidence for this
effect, i.e., pyy ) -∂FH/∂pKlg < 0.7,8 The negative pyy′
coefficients observed in these reactions are consistent an E2
mechanism and a reaction coordinate that has a large component
of proton transfer so that stabilization of the carbanion inter-
mediate result in a net shift from A to B in Figure 3I, in the
direction of decreased CR-OAr bond cleavage and smaller |âlg|
values.7-11 Moreover, a better leaving group would shift the
transition state from A to C in Figure 3II to decrease the extent
of negative charge development and the FH values.8
Solvent Effect. The observed second-order kinetics with the
Brønsted â ) 0.77 and âlg) -0.26 for the amine-promoted
eliminations from 1a in 70 mol % MeCN(aq) (Table 6) reveal
that the reactions also proceed by the concerted E2 mechanism.
Moreover, the decrease in the â values with a stronger electron-
withdrawing â-aryl substituent (Table 3), i.e., pxy′ < 0, is
consistent with an E2 mechanism and a reaction coordinate with
a large component of proton transfer (Vide supra).
For eliminations from 1 promoted by i-Pr2NH, the rate
decreases only slightly as the base-solvent system is changed
from i-Pr2NH-MeCN to i-Pr2NH/i-Pr2NH2+-70 mol % MeCN-
30% H2O, apparently because of the decreased basicity of the
promoting base in more protic solvent.25,31 On the other hand,
the extent of Câ-H bond cleavage decreases slightly and the
degrees of negative charge development at the â-carbon and
the CR-OAr bond cleavage decrease considerably as revealed
by the small to significant decrease in the Brønsted â, F, and
|âlg| values in R2NH/R2NH2+-70 mol % MeCN(aq) as the
base-solvent system (Table 6). The changes in the transition
state structures with the base-solvent system variation may be
attributed to a solvent effect. Due to the electronegativity
difference, the negative charge developed at the â-carbon in
the ketene-forming transition state may be delocalized more on
the carbonyl oxygen than on the â-carbon. Moreover, this effect
would be amplified in the buffer solution because the former
can be better stabilized by forming a stronger hydrogen bond
with R2NH2+. This would predict that less negative charge
would remain at the â-carbon to decrease both FH and |âlg|
values, as observed.36
proceed by the common E2 mechanism. In addition, the
Brønsted â value determined with the calculated k2 value for
E
1e is 0.66, which is smaller than that of 0.77 for 1a (Table 3).
This corresponds to a negative pxy′ interaction coefficient and
provides additional support for the operation of the E2 mech-
E
anism in the k2 process for 1e and a reaction coordinate that
has a large component of proton transfer (Vide supra).7-11
The calculated k1 and k-1/k2 values provide strong evidence
for the existence of the stepwise E1cb mechanism. The k1 value
increases with a stronger base apparently because it involves
the cleavage of the Câ-H bond leading to the carbanion
intermediate. The large â value of 0.73 for the k1 step is also
reasonable because the Câ-H bond should be extensively broken
in the transition state. Moreover, the slope of the plot of log
k-1/k2 vs pKa values of the buffer is -0.19 (Figure S8).
Although the k2 value should be independent of the basicity
because it is an intramolecular process, the k-1 should decrease
+
as the R2NH2 becomes a weaker acid. Hence the slope of
this plot could be taken as a measure of the extent of proton
transfer from the R2NH2+ to the carbanion in the k-1 step. The
small negative slope is consistent with a reactant-like transition
state for the k-1 step and provides additional support for the
stepwise E1cb mechanism. Finally, the H-D exchange experi-
ment reveals that approximately half of the benzylic C-H bond
in the recovered starting material has been converted to the C-D
bond in 70 mol % MeCN-30% D2O. All of these results are
in excellent agreement with concurrent E2 and E1cb mecha-
+
nisms for the reaction of 1e with R2NH/R2NH2 buffers in 70
mol % MeCN(aq).
It has been argued that the change in mechanism from E2 to
E1cb for eliminations from (2-arylethyl)ammonium ions and
9-(2-X-2-propyl)fluorenes occurs by the merging of the transi-
tion states because it is difficult for the energy maximum for
the transition state of the E2 reaction and an energy well for
the E1cb reaction to coexist for a single compound at almost
the same position on the energy diagram.9,12 However, if the
energies of the transition states for the competing reactions are
similar and if the positions of transition states on the energy
diagram are significantly different and/or if there is sufficient
energy barrier between them, they should be able to coexist.
For example, clear evidences for the concurrent E2 and E1
mechanisms have been reported for the methoxide-promoted
eliminations from 2-chloro-2-methyl-1-phenylpropanes and N-
(arylsufonoxy)-N-alkylbenzylamines as well as for the solvolytic
eliminations of 9-(1-X-ethyl)fluorenes.13-15 The mechanistic
borderline observed in the ketene-forming elimination from 1e
is concluded to be of the second type. Thus the elimination
reactions of 1a-d with R2NH in MeCN proceed by an E2
mechanism (Vide supra). The mechanism changes from E2 to
the parallel E2 and E1cb mechanisms by the introduction of
the p-NO2 group at the â-aryl ring and the R2NH/R2NH2+ buffer
in 70 mol % MeCN(aq) as the base-solvent system, probably
because the E1cb intermediate is stabilized by such changes so
that the two mechanisms can coexist. The transition states for
the two reactions appear to be appreciably separated on the
energy surface as indicated by the significant increase in the â
value from 0.66 for the E2 to 0.73 for the k1 step of the E1cb
mechanism, respectively (Table 3). A further stabilization of
the carbanion intermediate by utilizing OH--80% DMSO(aq)
as the base-solvent system induces a change in the reaction
mechanism to an E1cb (Table 7). To our knowledge, this is
the first example that clearly demonstrates that the E2 mech-
anism changes to an E1cb via concurrent E2 and E1cb
mechanisms.
Appearance of Concurrent E2 and E1cb Mechanisms.
Figure 2 shows a plot of kobs vs free buffer concentration for
+
eliminations from 1e promoted by i-Pr2NH/i-Pr2NH2 buffer
in 70 mol % MeCN(aq). The plots are curvilinear at low buffer
concentration but become straight lines at [buffer] > 0.01 M
for all buffer systems employed (Figure S4-S6). This result
cannot be explained with the change in the rate-limiting step
because the rate should have reached a plateau at a higher buffer
concentration in such a case. On the other hand, the rate data
can readily be analyzed with a nonlinear regression analysis by
assuming that the reactions proceed by concurrent E2 and E1cb
mechanisms (eq 4). The excellent correlations of the experi-
mental data with the computer-fitted curves demonstrate the
reliability of the kinetic analysis. In addition, the shapes of
the dissected lines are typical for the E2 and E1cb reactions,
respectively (Figure 2).
Figure S7 shows that the calculated k2E value for eliminations
from 1e falls on a single line with the experimental data for
1a-d in the Hammett plot. This result provides a strong support
E
that the k2 pathway for 1e and elimination reactions of 1a-d
(36) Cho, B. R.; Lee, S. J.; Kim, Y. K. J. Org. Chem. 1995, 60, 2072-
2076.

Reactions of Aryl Phenylacetates
J. Am. Chem. Soc., Vol. 119, No. 4, 1997 697
Table 7. Kinetic Parameters for Various Elimination Reactions
reaction
kinetic order
â
|âlg|
mechanism
ref
C6H5CH2CO2C6H4-p-NO2 + R2NH in MeCN
p-O2NC6H5CH2CO2C6H4-p-NO2 + R2NH/R2NH2
in 70 mol % MeCN(aq)
2
1-2
0.82
0.66-0.77
0.4-0.5
E2
this work
this work
+
0.26
E2 and E1cb
C6H5CH2CO2C6H4X + OH- in 80% DMSO(aq)
1-2
1
1
0.65
(E1cb)R or (E1cb)ion pair
(E1cb)anion
(E1cb)anion
18
p-O2NC6H5CH2CO2C6H4X + OH- in 80% DMSO(aq)
C6H5NHCO2C6H4X + OH- in H2O
1.25, 1.34a
1.34
18, 20
22g
a Calculated with the rate data in ref 20 and the estimated pKlg values in ref 18.
98%. To determine the yields of N,N-diisopropylbenzamide by GC,
1a and 2a (0.08 mmol) were reacted with 10 equiv of i-Pr2NH in 5
mL of MeCN for 2 h. After evaporation of the solvent, the product
was taken up in CH2Cl2 and washed thoroughly with water until all of
the amine, ammonium salt, and the aryloxides were completely
removed. The products were analyzed by GC as described before.35
The yields of N,N-diisopropylbenzamide from the reactions of 1a and
2a were 88 and 92%, respectively. The reaction of 1a and 2a with
i-Pr2NH/i-Pr2NH2+ in 70 mol % MeCN(aq) were conducted by the same
procedure except that smaller amount of the reactants (ca. 0.01 mmol)
were used. The products were identified by TLC and NMR. The
products were N,N-diisopropylbenzamide and phenylacetic acid.
H-D Exchange Experiment. To determine whether 1e may
undergo H-D exchange reaction, 1e (0.059 g, 0.19 mmol) was added
to 3.4 mL of i-Pr2NH/i-Pr2NH2+ buffer in 70 mol % MeCN-30% D2O
at -10 °C. The reaction was quenched by adding dilute D2SO4 in
D2O immediately after mixing. The recovered reactant was isolated
by extracting the products with ethyl acetate followed by the silca gel
column chromatography using ethyl acetate/hexane ) 1/4 as eluent.
The proton NMR spectrum of the recovered reactant was identical to
that of 1e except that the integration at δ 4.08 was decreased by
approximately 50%.
Experimental Section
Materials. Aryl esters of arylacetic acids 1 and 2 and of alkanoic
acids 3 and 4 were prepared by reactions of appropriately substituted
arylacetic acids and alkanoic acids with phenols in toluene or CH2Cl2
in the presence of 2-chloro-1-methylpyridinium iodide.23 The physical,
spectral, and analytical data for these compounds were consistent with
the proposed structures. The melting point (°C), NMR (ppm), IR
(cm-1), and combustion analysis data of the new compounds are listed
in the Supporting Information.
Acetonitrile was purified, and solutions of R2NH in MeCN were
prepared as described before.5 The buffer solutions of R2NH2+/R2NH
in 70 mol % MeCN(aq) were prepared by adding equivalent amounts
of R2NH2+ and R2NH to 70 mol % MeCN(aq). In all cases, the buffer
ratio was maintained to be 1.0.
Kinetic Studies. Reactions of 1-4 with R2NH in MeCN or R2-
+
NH/R2NH2 buffer in 70 mol % MeCN(aq) were followed by
monitoring the increase in the absorption of the aryl oxides at 400 and
426 nm with a UV-vis spectrophotometer as described before.5,6,35
Reactions of 2e with R2NH in MeCN and of 1e with R2NH2+/R2NH
buffer in 70 mol % MeCN(aq) were too fast to be measured by this
method, so the rate studies utilized a stopped-flow spectrophotometer.
Buffer Association. For reactions of 1e with R2NH/R2NH2+ buffer
in 70 mol % MeCN(aq), the plots of kobs vs buffer concentration showed
downward curvature. The negative deviation may be accounted for
by buffer association forming kinetically inactive hydrogen-bonded
complexes, R2NH2+‚‚‚NHR2. The buffer association constants Kassoc
were determined as described in the literature26 by using eq 5, where
[B-]ST and [BH]ST are the stoichiometric concentrations of the buffer
components and [B-] represents the concentration of the free amine
base. The Kassoc values are 6.5, 6.0, 7.3, and 20 for Bz(i-Pr)NH, (i-
Control Experiments. The stability of 1-2 were determined as
reported earlier.5,6 They were stable for at least 3 weeks in MeCN
solution at room temperature.
Acknowledgment. This research was supported in part by
OCRC-KOSEF and Basic Science Research Institute Program,
Ministry of Education, 1995 (Project No. BSRI-95-3406).
Supporting Information Available: List of melting point
(°C), NMR (ppm), IR (cm-1), and combustion analysis data of
the new compounds, Charton plot for aminolysis of 4a-d with
i-Pr2NH in MeCN, plot of kobs vs [buffer] for eliminations from
1a promoted by i-Pr2NH/i-Pr2NH2+ in 70 mol % MeCN(aq) at
25.0 °C, plots of kobs vs [buffer] for ketene-forming eliminations
Kassoc ) [B-‚‚‚BH]/[B-][BH] ) ([B-]ST - [B-])/([BH]ST
-
[B-‚‚‚BH])([B-]ST - [B-‚‚‚BH]) (5)
Bu)2NH, i-Pr2NH, and 2,6-dimethylpiperidine buffers, respectively.
Utilizing these values, [B-]ST, and [BH]ST the free buffer concentration
was calculated for all buffers prior to analysis of the kinetic data.
Calculation of the k2E, k1, and k-1/k2 Values. Utilizing the kobs
+
from 1e promoted by R2NH in MeCN and R2NH/R2NH2 in
70 mol % MeCN(aq) at 25.0 °C, Hammett plot for eliminations
+
from 1a-e promoted by i-Pr2NH/i-Pr2NH2 buffer in 70 mol
E
% MeCN(aq) at 25.0 °C, plot of log k1 and log k-1/k2 vs pKa of
the buffers for eliminations from 1e promoted by R2NH/R2-
NH2+ in 70 mol % MeCN(aq) at 25.0 °C (10 pages). See any
current masthead page for ordering and Internet access
instructions.
values and the free buffer concentration, the k2 , k1, and k-1/k2 values
that best fit with eq 4 have been calculated with the NLSFIT function
in the GENPLOT program, which is a nonlinear square fitting routine
based on the gradient search and function linearization method.27
Product Studies. The yields of the aryloxides determined by
comparing the absorbance of the infinity samples from the kinetic
reactions with those of authentic aryloxides were in the range 92-
JA961294K