Ketene-forming eliminations from aryl phenylacetates promoted by R2NH/R2NH2+ in aqueous MeCN. Mechanistic borderline between E2 and E1cb

Article abstract of DOI:10.1021/jo025555m

Elimination reactions of 2-X-4-NO₂C₆H₃CH₂C(O) OC₆H₃-2-Y-4-NO₂ [X = H (1), NO₂ (2)] promoted by R₂NH/R₂NH₂⁺ in 70 mol % MeCN(aq) have been studied kinetically. The base-promoted eliminations from 1 proceeded by the E2 mechanism when Y = Cl, CF₃, and NO₂. The mechanism changed to the competing E2 and E1cb mechanisms by the poorer leaving groups (Y = H, OMe) and to the E1cb extreme by the strongly electron-withdrawing β-aryl group (2, X = NO₂). The values of β = 0.14 and |β_1g| = 0.10-0.21 calculated for elimination from 1 (Y = NO₂) indicate a reactant-like transition state with small extents of proton transfer and C_α-OAr bond cleavage. The extent of proton transfer increased with a poorer leaving group, and the degree of leaving group bond cleavage increased with a weaker base. Also, the changes in the k₁ and k_-1/k₂ values with the reactant structure variation are consistent with the E1cb mechanism. From these results, a plausible pathway of the change of the mechanism from E2 to the E1cb extreme is proposed.

Full text of DOI:10.1021/jo025555m

Keten e-F or m in g Elim in a tion s fr om Ar yl P h en yla ceta tes P r om oted
+
by R2NH/R2NH2 in Aqu eou s MeCN. Mech a n istic Bor d er lin e
betw een E2 a n d E1cb
Bong Rae Cho,*,† Hyun Cheol J eong,† Yoon J e Seung,‡ and Sang Yong Pyun*,‡
Departments of Chemistry, Korea University, 1-Anamdong, Seoul, 136-701, Korea,
and Pukyong National University, Pusan 608-737, Korea
chobr@korea.ac.kr
Received J anuary 25, 2002
Elimination reactions of 2-X-4-NO2C6H3CH2C(O)OC6H3-2-Y-4-NO2 [X ) H (1), NO2 (2)] promoted
+
by R2NH/R2NH2 in 70 mol % MeCN(aq) have been studied kinetically. The base-promoted
eliminations from 1 proceeded by the E2 mechanism when Y ) Cl, CF3, and NO2. The mechanism
changed to the competing E2 and E1cb mechanisms by the poorer leaving groups (Y ) H, OMe)
and to the E1cb extreme by the strongly electron-withdrawing â-aryl group (2, X ) NO2). The values
of â ) 0.14 and |âlg| ) 0.10-0.21 calculated for elimination from 1 (Y ) NO2) indicate a reactant-
like transition state with small extents of proton transfer and CR-OAr bond cleavage. The extent
of proton transfer increased with a poorer leaving group, and the degree of leaving group bond
cleavage increased with a weaker base. Also, the changes in the k1 and k-1/k2 values with the
reactant structure variation are consistent with the E1cb mechanism. From these results, a plausible
pathway of the change of the mechanism from E2 to the E1cb extreme is proposed.
In tr od u ction
E1 mechanisms has been reported for the MeONa-
promoted eliminations from 2-chloro-2-methyl-1-phenyl-
propane5 and N-(arylsulfonoxy)-N-alkylbenzylamine,6 as
well as the solvolytic elimination of 9-(1-X-ethyl)fluorene.7
Recently, we have initiated a series of investigation on
the ketene-forming elimination reactions to gain an
insight into the E2 and E1cb borderline.9-11 It is well
established that the base-catalyzed hydrolysis of aryl
p-nitrophenylacetates and other activated esters proceed
by an E1cb mechanism to afford the ketene intermediate
followedbytheaddition ofwater under variousconditions.12-25
When the reaction condition was changed from OH- in
One of the most interesting problems in elimination
reaction is the mechanistic borderline.1-8 There are two
possibilities by which the mechanism changes from one
to another. The first is the “merging” of the transition
states. The change in the mechanism from E2 to E1cb
in eliminations from (2-arylethyl)ammonium ions3 and
9-(2-chloro-2-propyl)fluorene4 has been concluded to be
of this type. It has been argued that it would be difficult
for the energy maximum of the E2 mechanism and an
energy well of the E1cb mechanism to coexist for a single
compound at almost the same position on the energy
diagram, and the “merging” becomes the favored alterna-
tive. On the other hand, if the energies of transition
states for the competing reactions are similar, and if the
positions of the transition states are significantly differ-
ent, they should be able to coexist. Under this condition,
the mechanism change could occur via a competing
mechanism. Clear evidence in favor of concurrent E2 and
(9) Cho, B. R.; Kim Y. K.; Maing Yoon, C. O. J . Am. Chem. Soc.
1997, 119, 691-697.
(10) Cho, B. R.; Kim Y. K.; Yoon, J . S.; Kim, J . C.; Pyun, S. Y. J .
Org. Chem. 2000, 65, 1239-1241.
(11) Cho, B. R.; Kim, N. S.; Kim, Y. K.; Son, K. H. J . Chem. Soc.,
Perkin Trans. 2 2000, 1419-1422.
(12) Holmquist, B.; Bruice, T. C. J . Am. Chem. Soc. 1969, 91, 2993-
3002.
(13) Holmquist, B.; Bruice, T. C. J . Am. Chem. Soc. 1969, 91, 3003-
3006.
(14) Pratt, R. F.; Bruice, T. C. J . Am. Chem. Soc. 1970, 92, 5956-
* E-mail for S.Y.P.: sypyun@pknu.ac.kr.
† Korea University.
5964.
(15) Inoue, M.; Bruice, T. C. J . Am. Chem. Soc. 1982, 104, 1644-
1653.
‡ Pukyong National University.
(1) Gandler, J . R. The Chemistry of Double Bonded Functional
Groups; Patai, S., Ed.; J ohn Wiley and Sons: Chichester, 1989; Vol.
2, part 1, pp 734-797.
(16) Inoue, M.; Bruice, T. C. J . Org. Chem. 1982, 47, 959-963.
(17) William, A. J . Chem. Soc., Perkin Trans. 2 1972, 808-812.
(18) William, A.; Douglas, K. T. Chem. Rev. 1975, 75, 7-649.
(19) Tagaki, W.; Kobayashi, S.; Kurihara, K.; Kurashima, K.;
Yoshida, Y.; Yano, J . J . Chem. Soc., Chem. Commun. 1976, 843-845.
(20) Broxton, T. J .; Duddy, N. W. J . Org. Chem. 1981, 46, 1186-
1191.
(2) J encks, W. P. Chem. Rev. 1985, 85, 511-527
(3) Gandler, J . R.; J encks, W. P. J . Am. Chem. Soc. 1982, 104, 1937-
1951.
(4) Thibblin, A. J . Am. Chem. Soc. 1988, 110, 4582-4586.
(5) Thibblin, A. J . Am. Chem. Soc. 1989, 111, 5412-5416.
(6) Cho, B. R.; Pyun, S. Y. J . Am. Chem. Soc. 1991, 113, 3920-
3924.
(7) Meng, Q.; Thibblin, A. J . Am. Chem. Soc. 1995, 117, 9399-9407.
(8) Olwegard, M.; McEwen, I.; Thibblin, A.; Ahlberg, P. J . Am. Chem.
Soc. 1985, 107, 7494-7499.
(21) Chandrasekar, R.; Venkatasubramanian, N. J . Chem. Soc.,
Perkin Trans. 2 1982, 1625-1631.
(22) Douglas, K. T.; Alborz, M.; Rullo, G. R.; Yaggi, N. F. J . Chem.
Soc., Chem. Commun. 1982, 242-246.
(23) Isaac, N. S.; Najem, T. S. J . Chem. Soc., Perkin Trans. 2 1988,
557-562.
10.1021/jo025555m CCC: $22.00 © 2002 American Chemical Society
Published on Web 06/18/2002
5232
J . Org. Chem. 2002, 67, 5232-5238

E2 and E1cb Mechanistic Borderline
TABLE 1. Ra te Con sta n ts for Keten e-F or m in g
H2O to R2NH in MeCN the reaction proceeded by the E2
mechanism via an E1cb-like transition state.9-11 Curved
Hammett plots observed in this study were attributed
to the highly carbanionic transition state.11 Competing
E2 and E1cb mechanisms were noted in eliminations
from p-nitrophenyl p-nitrophenylacetates promoted by R2-
NH/R2NH2+ in 70 mol % MeCN(aq).9 By comparing with
the literature data, it was concluded that the mechanism
change could occur via the competing stage. However, it
was not unambiguously established that the latter was
indeed the intermediate stage between the two mecha-
nisms.
a
Elim in a tion s fr om 4-NO2C6H4CH2CO2C6H3-2-Y-4-NO2
+
P r om oted by R2NH/R2NH2 in 70 m ol % MeCN(a q) a t 25.0
°Cb,c
k2E, M-1 s-1 f,g
R2NHd
pKa
1a h
1bh
1c
1d
1e
e
Bz(i-Pr)NH
i-Bu2NH
i-Pr2NH
16.8
18.2
18.6
18.9
0.170i
0.530
0.130
0.390
15.5
34.8
30.9
52.1
27.1
60.4
45.0
70.9
43.6
70.1
61.0
85.8
2,6-DMPi
1.77
1.15
[Substrate] ) 4.0 × 10-5 M. [R2NH]/[R2NH2+] ) 1.0. c µ )
a
b
0.10 M (Bu4N+Br-). [R2NH] ) 9.6 × 10-4 to 6.0 × 10-2 M.
d
e Reference 9. f Average of three or more rate constants. Esti-
g
To further expand our understanding on the E2 and
E1cb borderline, we have investigated the reactions of
aryl aryl acetates (1 and 2) with R2NH/R2NH2+ in 70 mol
% MeCN(aq) (eq 1). The reaction mechanism and struc-
h
mated uncertainty, ( 5%. Calculated from the kobs by using eq
i
E
2. k2 ) 0.160 and 0.230 M-1 s-1when [R2NH]/[R2NH2+] ) 0.5
j
and 2.0, respectively. cis-2,6-Dimethylpiperidine.
The rates of elimination reactions were followed by
monitoring the increase in the absorption at the λmax for
the aryloxides in the range of 400-434 nm. In all cases,
the ionic strength was maintained to be 0.10 M with
Bu4N+Br-. Excellent pseudo-first-order kinetics plots,
which covered at least three half-lives, were obtained.
The rate constants are summarized in Tables S1-6 in
Supporting Information. The plots of kobs versus base
concentration for the reaction of 1c were straight lines
passing through the origin (Figures 1 and S1). This result
indicates that (i) the reaction is overall second order, first
order to the substrate and first order to the base
concentration, (ii) OH-- or solvent-promoted elimination
does not contribute to the observed rates, (iii) the slopes
of the straight lines are the overall second-order rate
constants k2E. Hence, the rate constants for the base-
promoted eliminations from 1d and 1e were determined
ture-reactivity relationship of the ketene-forming elimi-
nations have been determined. Here we report that (i)
elimination reactions from 1c-e proceed by the E2
mechanism, (ii) the mechanism changes to a competing
E2 and E1cb mechanism by the change in the leaving
group to a poorer one (1a and 1b) and (iii) to the E1cb
extreme by a strongly electron-withdrawing â-aryl group
(2a ). This result provides the first unambiguous evidence
that the elimination reaction mechanism changes gradu-
ally from E2 to E1cb via the concurrent mechanism by
the systematic change of the reactant structure.
E
at a single base concentration (Table S2). The k2 values
were obtained by dividing the kobs by the base concentra-
E
tion. Values of k2 for eliminations from 1c-e are
Resu lts
summarized in Table 1. Except for the rate data deter-
mined with i-Pr2NH, the rate increased with the basicity
and the leaving group ability. The slower rate of elimina-
tions from 1c-e with i-Pr2NH than with i-Bu2NH as the
base can be attributed to the greater steric requirement
of the former. On the other hand, the same plot for the
reaction of 2a showed a saturation curve typical for the
E1cb mechanism. The rate data showed excellent cor-
relation with eq 2 when fitted without the first term.
Aryl p-nitrophenylacetates (1) were synthesized by the
reaction between p-nitrophenylacetic acid, substituted
phenols, 2-chloro-1-methylpyridinum iodide, and Et3N in
CH2Cl2 as reported previously.9-11,25 4-Nitrophenyl 2,4-
dinitrophenyl acetate (2a ) was prepared by reacting 2,4-
dinitrophenylacetyl chloride with 1 equiv of the sodium
4-nitrophenoxide in CH2Cl2 by the literature procedure.26
The spectral and analytical data for the compounds were
consistent with the proposed structures.
The product of the reaction between 1a and i-Pr2NH/
i-Pr2NH2+ in 70 mol % MeCN(aq) was previously identi-
fied as N,N-diisopropylbenzamide and p-nitrophenoxide.9
Similarly, N,N-diisobutyl-2,4-dinitrobenzamide was ob-
tained as the only product from the reaction of 2a i-Bu2-
+
NH/i-Bu2NH2 in 70 mol % MeCN(aq). For elimination
reactions from 1a -e and 2a , the yields of the aryloxides
as determined by comparison of the UV absorption of the
infinity sample of the kinetic runs with those of the
authentic samples were in the range 93-99%. The
possibility of a competing aminolysis had been ruled out
as previously reported.9
Figure 1 shows the plots of kobs versus base concentration
for the elimination reactions of 1c and 2a with Bz(i-Pr)-
NH/Bz(i-Pr)NH2 in 70 mol % MeCN(aq). Similar plots
were obtained for the reactions of 1c and 2a with other
bases (Figures S1,2).
In contract to the reactions of 1c and 2a , the plots of
kobs versus base concentration for the reactions of 1a and
1b were curves at low base concentration but became
straight lines at higher base concentration. A typical
+
(24) Chung, S. Y.; Yoh, S. D.; Choi, J . H.; Shim, K. T. J . Korean
Chem. Soc. 1992, 36; 446-452.
(25) Saigo, K.; Usui, M.; Kikuchi, K.; Shimada, E.; Mukaiyama, T.
Bull. Chem. Soc. J pn. 1977, 50, 1863-1866.
(26) Chandrasekar, R.; Venkatasubramanian, N. J . Chem. Soc.,
Perkin Trans. 2 1982, 1625-1631.
J . Org. Chem, Vol. 67, No. 15, 2002 5233

Cho et al.
TABLE 2. k1 a n d k-1/k2 for Elim in a tion fr om
a
2-X-4-NO2C6H3CH2CO2C6H3-2-Y-4-NO2 P r om oted by
R2NH/R2NH2 in 70 m ol % MeCN(a q) 25.0 °Cb,c
+
k1, M-1 s-1 e
k-1/k2, M-1 e
R2NHd
1a
1b
2a
1a
1b
2a
Bz(i-Pr)NH
i-Bu2NH
11.7f
20.2
31.1
8.10
16.3
22.8
52.3
68.8
92.0
560g
225
56.0
677
327
73.5
1187
289
133
2,6-DMPh
[Substrate] ) 4.0 × 10-5 M. [R2NH]/[R2NH2+] ) 1.0 except
a
b
otherwise noted. c µ ) 0.10 M (Bu4N+Br-). [R2NH] ) 9.6 × 10-4
d
to 6.0 × 10-2 M. e Calculated from the kobs by using eq 2. f k1
)
11.7 and 10.9 M-1 s-1 when [R2NH]/[R2NH2+] ) 0.5 and 2.0,
g
+
respectively. k-1/k2 ) 551 and 536 M-1 when [R2NH]/[R2NH2
]
h
) 0.5 and 2.0, respectively. cis-2,6-Dimethylpiperidine.
program, the k2E, k1, and k-1/k2 values that best fit with
eq 2 were calculated, and the plots were dissected into
the E2 and E1cb components. Figure 2 shows that the
correlation between the calculated and the experimental
data is excellent, indicating the reliability of this kinetic
analysis. To assess the effect of buffer ratio, the kobs
values for the ketene-forming eliminations from 1a have
F IGURE 1. Plots of kobs versus base concentration for
eliminations from 4-NO2C6H4CH2CO2C6H3-2-Cl-4-NO2 (1c, 9)
and 2,4-(NO2)2C6H3CH2CO2C6H4-4-NO2 (2a , b) promoted by
+
Bz(i-Pr)NH/Bz(i-Pr)NH2 in 70 mol % MeCN(aq) at 25.0 °C,
[Bz(i-Pr)NH]/[Bz(i-Pr)NH2+] ) 1.0, µ ) 0.10 M (Bu4N+Br-).
+
been measured at Bz(i-Pr)NH/Bz(i-Pr)NH2 ) 0.5 and
2.0, respectively (Table S5). When Bz(i-Pr)NH/Bz(i-Pr)-
NH2+ ) 0.5, the kobs were determined up to [base] ) 0.03
M because of the limited solubility. At a given base
concentration, the kobs increases gradually as the buffer
ratio is changed from 0.5 to 1.0 to 2.0, as expected from
eq 2 (Tables S4 and S5). In both cases, the rate data
showed excellent correlations with eq 2 (Figures S3 and
S4). Similar correlation was obtained for the reactions
of 1a -b with other buffer systems (Figures S5-9). These
results provide additional evidence in support of the
kinetic analysis given above. However, the rate data for
the reactions of 1a , 1b, and 2a promoted by i-Pr2NH/i-
Pr2NH2+ were not included in the analysis because of the
complications due to the steric effect (vide supra). Cal-
culated values of k2E, k1, and k-1/k2 for the eliminations
from 1a , 1b, and 2a are summarized in Tables 1 and 2.
E
Except for the slightly larger k2 calculated at [B]/[BH+]
F IGURE 2. Plot of kobs versus base concentration for elimina-
tions from 4-NO2C6H4CH2CO2C6H4-4-NO2 (1a ) promoted by
) 2.0, the values of k2E, k1, and k-1/k2 for the Bz(i-Pr)-
NH-promoted eliminations from 1a are nearly the same
at all buffer ratios (Tables 1 and 2).
+
Bz(i-Pr)NH/Bz(i-Pr)NH2 in 70 mol % MeCN(aq) at 25.0 °C,
[Bz(i-Pr)NH]/[Bz(i-Pr)NH2+] ) 1.0, µ ) 0.10 M (Bu4N+Br-).
The closed circles are the experimental data, and the solid line
shows the computer fitted curve by using eq 2 (see text). The
curve is dissected into the E2 and E1cb reaction components
(dashed lines).
The k2E values for eliminations from 1a -e showed good
correlation with the pKa values of the promoting bases
on the Bro¨nsted plot (Figure 3). The â values are in the
range of 0.14-0.47 and increase as the leaving group
ability of the aryloxide decreases (Table 3). Values of both
k1 and k-1/k2 increased as the electron-withdrawing
ability of the â-aryl group increased (Table 2). On the
other hand, the k1 increased and k-1/k2 decreased with a
stronger base and a sterically less hindered leaving
group.
plot of kobs versus base concentration for the reaction of
+
1a with Bz(i-Pr)NH/Bz(i-Pr)NH2 in 70 mol % MeCN-
(aq) is shown in Figure 2. The possibility of the buffer
association as the cause of the curvature was ruled out
by the straight lines observed in the plot of kobs versus
base concentration for the reaction of 1c (vide supra).
Also, the possibility that the salt effect may have caused
the steady increase in the kobs at higher base concentra-
tion is negated because the ionic strength is maintained
to be 0.10 M with Bu4N+Br-. Hence the data were
analyzed with a nonlinear regression analysis program
by assuming that the reaction proceeds by concurrent E2
and E1cb mechanisms.27 By utilizing the computer
E
The plots of the k2 values for 1a -e against the pKlg
values of the leaving group are depicted in Figure 4. The
rate data showed good correlations on the straight lines,
if the data for 1a ,b were excluded. Therefore, the âlg
values were calculated from the straight lines without
the date for 1a ,b. The |âlg| values are in the range 0.21-
0.10 and decrease as the pKa value of the base increases
(Table 4).
To provide additional evidence for the competing E1cb
(27) MacroCal Origin, Version 3.5; MacroCal Software, Inc.:
Northamption, MA, 1991.
mechanism, the H-D exchange experiment was carried
5234 J . Org. Chem., Vol. 67, No. 15, 2002

E2 and E1cb Mechanistic Borderline
F IGURE 3. Bro¨nsted plots for the ketene-forming elimina-
tions from 4-NO2C6H4CH2CO2C6H3-2-Y-4-NO2 promoted by R2-
+
NH/R2NH2 in 70 mol % MeCN(aq) at 25.0 °C, [R2NH]/
[R2NH2+] ) 1.0, µ ) 0.10 M (Bu4N+Br-). Y ) H (1a , 9), OMe
(1b, b), Cl (1c, 2), CF3 (1d , 1), NO2 (1e, ().
TABLE 3. Br o1n sted â Va lu es for Elim in a tion fr om
E
F IGURE 4. Plots of log k2 versus pKlg values of the leaving
4-NO2C6H4CH2CO2C6H3-2-Y-4-NO2 P r om oted by R2NH/
group for eliminations reactions of 4-NO2C6H4CH2CO2C6H3-
+
R2NH2 in 70 m ol % MeCN(a q) a t 25.0 °Ca ,b
+
2-Y-4-NO2 (1a -e) promoted by R2NH/R2NH2 in 70 mol %
1a
1b
1c
1d
1e
MeCN(aq) at 25.0 °C, [R2NH]/[R2NH2+] ) 1.0, µ ) 0.10 M
(Bu4N+Br-). R2NH ) Bz(i-Pr)NH (9), i-Bu2NH(b), 2,6-DMP-
(2).
c
pKlg 20.7
â
20.6
18.1
17.0
16.0
0.47 ( 0.01d 0.44 ( 0.08d 0.25 ( 0.01 0.21 ( 0.04 0.14 ( 0.01
a
b
[R2NH]/[R2NH2+] ) 1.0. µ ) 0.10 M (Bu4N+Br-). c Reference
TABLE 4. âlg for Elim in a tion fr om
d
E
30. k2 values calculated from the kobs by using eq 2 have been
used.
4-NO2C6H4CH2CO2C6H3-2-Y-4-NO2 (1a -e) P r om oted by
+
R2NH/R2NH2 in 70 m ol % MeCN(a q) a t 25.0 °Ca ,b
R2NH
out by mixing 1a and 2a with i-Bu2NH/i-Bu2NH2+ in 70
mol % MeCN-30% D2O at -10 °C. The reactant was
recovered immediately after mixing. The NMR spectrum
indicated that approximately 30% of benzylic C-H bond
of 1a was converted to the C-D bond. On the other hand,
2a underwent complete H-D exchange.
Bz(i-Pr)NH
i-Bu2NH
2,6-DMPc
âlg
-0.21 ( 0.01
-0.15 ( 0.04
-0.10( 0.01
[R2NH]/[R2NH2+] ) 1.0. µ ) 0.10 M (Bu4N+Br-). c cis-2,6-
Dimethylpiperidine.
a
b
consistent with an E2 mechanism in which there is
partial cleavage of the Câ-H and CR-OAr bonds in the
transition state.
Discu ssion
In contrast to the reactions of 1c-e, the plots of the
kobs versus the base concentration for the elimination
reaction of 2a with R2NH/R2NH2 showed saturation
Mech a n ism of Elim in a tion s fr om 1c-e a n d 2a .
Results of kinetic investigations and product studies
+
+
reveal that the reactions of 1c-e with R2NH-R2NH2
curves, reaching at a plateau at [base] > 0.01 M (Figure
1). The result can most reasonably be explained with an
E1cb mechanism, i.e., kobs ) k1k2[B]/(k-1[BH+] + k2).1 This
predicts that the plot of kobs should change from a straight
line at a low base concentration, i.e., kobs ) k1[B] when
k-1[BH+] , k2, to a plateau at a higher base concentra-
tion, i.e., kobs ) k1k2/k-1 when k-1[BH+] . k2. The
excellent agreement between the experimental data and
the theoretically fitted curve provides strong evidence for
this mechanism (Figures 1 and S2). Moreover, the
calculated k1 and k-1/k2 values are consistent with the
trends observed for 1a and 1b (vide infra). Finally, the
NMR spectrum of the recovered reactant for the reaction
of 2a with i-Bu2NH/i-Bu2NH2+ in 70 mol % D2O in MeCN
indicated complete absence of the benzylic C-H bond at
δ 4.43, providing additional support for this mechanism.
Mech a n ism of Elim in a tion s fr om 1a ,b. The plots
of the kobs values for eliminations from 1a and 1b against
base concentration are curves at low buffer concentration
in 70 mol % MeCN(aq) proceed by the E2 mecha-
nism.1,2,28,29 The reactions produced only elimination
products and exhibited second-order kinetics, negating
all but bimolecular elimination pathways. The (E1cb)R,
(E1cb)ip, and internal return mechanisms are ruled out
by the observed general base catalysis with the Bro¨nsted
â values ranging from 0.14 to 0.25 because these mech-
anisms would exhibit either a specific base catalysis or
Bro¨nsted â values near unity.1 In addition, the possibility
that the relatively small values of |âlg| ) 0.21-0.10 might
be due to the (E1cb)irr mechanism29 is negated by the
interaction coefficients, i.e., pxy ) ∂â/∂pKlg ) ∂âlg/∂pKBH
> 0 (vide infra). On the other hand, the results are
(28) Saunders, W. H., J r.; Cockerill, A. F. Mechanism of Elimination
Reactions; Wiley: New York, 1973.
(29) 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.
(30) Coetzee, J . F. Prog. Phys. Org. Chem. 1965, 4, 45-92.
J . Org. Chem, Vol. 67, No. 15, 2002 5235

Cho et al.
and become straight lines at [base] > 0.01 M (Figures 2
and S3-9). This result cannot be explained with the
change in the rate-limiting step because the rate should
have reached a plateau at a higher base concentration
in such a case (vide supra). In contrast, the rate data can
readily be explained if one assumes that the reactions
proceed by the concurrent E2 and E1cb mechanisms, i.e.,
cates that the transition state structures for 1a ,b do not
fit in the trends observed for 1c,e. Although the values
of |âlg| for 1a ,b are not available, the negative deviation
could be explained if the extents of CR-OAr bond cleav-
age are assumed to be smaller than those of 1c-e. The
transition state would then be destabilized as a result of
the difficulty associated with the charge transfer from
the â- to the R-carbon, to decrease the degree of double
bond character and the rate. Also, a close examination
of Table 3 reveals that the Bro¨nsted â values increase
abruptly from 0.14-0.25 to 0.44-0.47 as the substrate
structures are changed from 1c-e to 1a ,b, indicating a
sharp increase in the extent of proton transfer. These
results seem to indicate that the transition state struc-
tures for the E2 pathway of the buffer-promoted elimina-
tions from 1a ,b are more E1cb-like with greater extent
of proton transfer and limited degree of the leaving group
bond cleavage, compared with 1c-e.
kobs ) k2 + k1k2[B]/(k-1[BH+] + k2) (eq 2). As shown in
E
Figures 2 and S3-9, all of the rate data for the reactions
+
of 1a ,b with R2NH/R2NH2 show excellent correlation
with eq 2. Note that the shapes of the dissected lines are
typical for the E2 and E1cb mechanisms. In addition, the
kobs values for the reactions of 1a with Bz(i-Bu)NH/Bz-
+
(i-Bu)NH2 increases gradually with the change in the
buffer ratio from 0.5 to 1.0 to 2.0, as expected from eq 2
(Tables S4 and S5).
The calculated rate constants for the E2 and E1cb
pathways provide additional support for this mechanism.
(i) The k2E, k1, and k-1/k2 values for the reactions of 1a
with Bz(i-Bu)NH/Bz(i-Bu)NH2+ are very similar regard-
less of the buffer ratio, indicating that the same mech-
anism is operating under these conditions (Tables 1 and
Effect of Bu 4N+Br - on th e Elim in a tion fr om 1a .
The relative rates of the E2 and E1cb pathways of the
elimination reactions of 1a change dramatically by the
addition of Bu4N+Br-. Figure 2 shows that the rate of
E1cb reaction is always faster than the E2 reaction in
the presence of Bu4N+Br- for all concentration ranges
employed in this study. In contrast, when the reaction
was carried out without Bu4N+Br-, the E1cb reaction was
faster when [base] < 0.02 M, after which the E2 reaction
became the major reaction pathway.9 The faster rate of
the E1cb reaction in the presence of Bu4N+Br- can be
attributed to the salt effect. As the ionic strength is
increased by the addition of Bu4N+Br-, the dipolar ionic
atmosphere would stabilize the anionic E1cb transition
state more than the E2 transition state, which would in
turn result in a preferential enhancement of the E1cb
rate.
E
2). (ii) The k2 values for 1a and 1b nicely fit in the trend
observed for 1c-e, i.e., the k2E of 1a -e increase gradually
with a stronger base and a better leaving group (Table
E
1). The only exception to this trend is the larger k2 of
1a compared to that of 1b, which may be attributed to
the leaving group steric effect. Because the p-nitrophe-
noxide in 1a is the only leaving group without an ortho
substituent, it should be sterically less hindered than
other substrates in the series, to increase the rate.11 (iii)
The observed increase in the k1 with a stronger base and
a more acidic benzylic C-H bond (2a ), as well as a small
decrease in the k1 with the leaving group steric effect
(1b), is consistent with what would be expected for the
deprotonation step (Table 2). (iv) The increase of the k-1
/
The structures of E2 transition states are also changed
by the Bu4N+Br-. Comparison with earlier results reveals
that the Bro¨nsted â and |âlg| values decrease from 0.66
to 0.47 and 0.26 to 0.10-0.21, respectively (Tables 3 and
4).9 This indicate small decrease in the extents of Câ-H
and CR-OAr bonds in the presence of Bu4N+Br-.
Ma p p in g of th e E2 Tr a n sition Sta te. The structures
of the E2 transition states may be assessed by the
Bro¨nsted â and |âlg| values. The Bro¨nsted â values
indicate the extent of proton transfer in the transition
state. For elimination from 1e, the value of â ) 0.14 was
determined. This indicates a small degree of proton
transfer in the transition state. The |âlg| values are
usually taken as the extent of the leaving group bond
cleavage. Hence the observed values of |âlg| ) 0.21-0.10
are consistent with limited extent of the CR-OAr bond
rupture. The combined results reveal that the ketene-
forming eliminations from 1e proceed by the E2 mech-
anism via a reactant-like transition state with limited
extents of proton transfer and CR-OAr bond cleavage.
Therefore, it seems reasonable to locate the transition
state near the right corner of the More-O’Ferall-J encks
diagram (Figure 5).
k2 values in the order 1a < 1b < 2a is also consistent
(Table 2). If the k1 of 1b is smaller than that of 1a , the
k-1 of 1b should be larger than that of 1a . On the other
hand, the values of k2 for 1a and 1b are expected to be
similar because a fraction of the difference in the leaving
group pKlg values (∆pKlg ) 0.1) will be reflected in the
k2. Also, the ortho nitro group is expected to stabilize the
carbanion intermediate to increase k-1 and decrease k2
of 2a , and this effect should be much larger for k-1 than
k2 because the second step involves a higher energy
barrier. A combination of these factors would be to
increase the k-1/k2 values in the order 1a < 1b < 2a , as
observed. (v) The increase in the k-1/k2 ratio with a
weaker base can also be explained similarly (Table 2).
+
The k-1 should increase with the acidity of R2NH2
,
although little change is expected for k2 by the base
strength variation. This predicts an increase in the k-1
k2 ratio with a weaker base.
/
The NMR spectrum of the recovered reactant from the
+
reaction of 1a with i-Bu2NH/i-Bu2NH2 in 70 mol %
MeCN-30% D2O at -10 °C indicated that approximately
30% of the benzylic C-H bond underwent H-D ex-
change. All of these results provide strong evidence that
eliminations from 1a ,b promoted by R2NH/R2NH2+ in 70
mol % MeCN(aq) proceed by the concurrent E2 and E1cb
mechanism. 1,29
This conclusion is supported by the interaction coef-
ficients. Table 3 shows that the â values of the k2E process
for 1a -e decrease, as the leaving groups are made less
basic. This effect corresponds to a positive pxy interaction
coefficient, pxy ) ∂â/∂pKlg, which describes the interaction
between the base catalyst and the leaving group.13,14 The
E
It should be noted that the k2 values for 1a ,b show
large negative deviations in Figure 4. This result indi-
5236 J . Org. Chem., Vol. 67, No. 15, 2002

E2 and E1cb Mechanistic Borderline
reaction can no longer compete, the E1cb mechanism
becomes the predominant reaction pathway.
As stated in the Introduction, the change of the
mechanism from E2 to E1cb could occur by the merging
of the transition state, if it is difficult for the energy
maximum of the E2 mechanism and an energy well of
the E1cb mechanism to coexist for a single compound at
almost the same position on the energy diagram.1-4
However, if the positions of the transition states are
significantly different, they should be able to coexist, as
observed in the E2 and E1 borderline.5-7 The change in
elimination mechanism observed in the present study is
concluded to be of this type. Because the E2 transition
state is located near the center of the horizontal reaction
coordinate, which is far from the position of the E1cb
transition state, there seems to be no reason for why the
two transition states cannot coexist. Furthermore, the
mechanism changes systematically from the concurrent
E2 and E1cb mechanisms to E2 as the E2 transition state
is stabilized by the better leaving group and to the E1cb
extreme when the strongly electron-withdrawing â-aryl
group stabilizes the latter transition state. To our knowl-
edge, this is the first example that shows a gradual
change in the elimination reaction mechanism from E2
to E1cb via a competing E2 and E1cb mechanism
wrought by the systematic variation of the reactant
structure.
F IGURE 5. Reaction coordinate diagram for the ketene-
forming elimination. The effect of the change to a poorer
leaving group and a weaker base are shown by the shift of
the transition state from A to B and A to C, respectively.
Exp er im en ta l Section
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. On the More-O’Ferall-
J encks diagram in Figure 5, a change to a poorer leaving
group will raise the energy of the bottom edge of diagram
shifting the transition state toward the product and the
E1cb intermediate. The transition state on the horizontal
reaction coordinate will then move toward the left, with
more proton transfer and a larger â, as depicted by a shift
from A to B on the energy diagram (Figure 5).1-3
Similarly, a weaker base will raise the energy of the left
side of the diagram and shift the transition state from A
to C to increase in the extent of CR-OAr bond cleavage
(Figure 5).1-3 The positive pxy coefficients are not consis-
tent with an E1cb mechanisms for which pxy ) 0 is
expected but provide additional support for the concerted
E2 mechanism.1-3
Ch a n ge of th e Mech a n ism . The kinetic results
described above clearly indicate that the elimination
mechanism changes from E2 to a competing E2 and E1cb
to E1cb as the reactant structure is changed from 1c-e
to 1a ,b to 2a . The most reasonable explanation for this
result is as follows. As discussed above, the position of
the E2 transition state for the reactions from 1e could
be located at A in Figure 5. As the leaving group is made
poorer (1c,d ), the reactant-like E2 transition state gradu-
ally changes to the E1cb-like, as indicated by a shift from
A to B (vide supra). When the leaving group ability is
further decreased to Y ) H and OMe (1a ,b), the E2
transition state is destabilized and the E1cb mechanism
emerges simultaneously. It is interesting to note that the
E2 transition states for eliminations from 1a ,b are more
E1cb-like than 1c-e (vide supra). Finally, when the
carbanion intermediate is stabilized by the strongly
electron-withdrawing â-aryl group (2a ), so that the E2
Ma ter ia ls. 4-Nitrophenyl p-nitrophenylacetate (1a ) and 2,4-
dinitrophenyl p-nitrophenylacetate (1e) were available from
previous study.9 Aryl p-nitrophenylacetates (1b-d ) were
prepared from 4-nitrophenylacetic acid and substituted phe-
nols in the presence of Et3N and 2-chloro-N-methylpyridinum
iodide as previously reported.9 p-Nitrophenyl 2,4-dinitrophe-
nylacetate (2a ) was prepared by treating 2,4-ditrophenylacetyl
chloride with 1 equiv of the sodium 4-nitrophenoxide in
methylene chloride as previously reported.26 The yield (%),
melting points (°C), IR (KBr, CdO, cm-1), NMR (300 MHz,
CDCl3), and combustion analysis data for the new compounds
are as follows.
2-Met h oxy-4-n it r op h en yl p-Nit r op h en yla cet a t e (1b ).
Yield 34%; mp 101 °C; IR 1759 (CdO); 1H NMR δ 3.88 (s, 3H),
4.04 (s, 2H), 7.17 (d, 1H, J ) 8.7 Hz), 7.58 (d, 2H, J ) 8.7 Hz),
7.83 (d, 1H, J ) 2.6 Hz), 7.89 (dd, 1H, J ) 8.7, 2.6 Hz), 8.25
(d, 2H, J ) 8.7 Hz). Anal. Calcd for C15H12N2O7: C, 54.22; H,
3.64; N, 8.43. Found: C, 54.27; H, 3.65; N, 8.36.
2-Ch lor o-4-n itr oph en yl p-Nitr oph en ylacetate (1c). Yield
29%; mp 128 °C; IR 1764 (CdO); NMR δ 4.09 (s, 2H), 7.33 (d,
1H, J ) 8.7 Hz), 7.60 (d, 2H, J ) 8.7 Hz), 8.20 (dd, 1H, J )
8.4, 2.7 Hz), 8.28 (d, 2H, J ) 8.4 Hz), 8.36 (d, 1H, J ) 2.7 Hz).
Anal. Calcd for C14H9N2O6Cl: C, 49.94; H, 2.69; N, 8.32.
Found: C, 49.90; H, 2.68; N, 8.23.
2-Tr iflu or om et h yl-4-n it r op h en yl p -Nit r op h en yla ce-
ta te (1d ). Yield 25%; mp 131 °C; IR 1765 (CdO); NMR δ 4.08
(s, 2H), 7.49 (d, 1H, J ) 8.7 Hz), 7.55 (d, 2H, J ) 8.7 Hz), 8.26
(d, 2H, J ) 8.7 Hz), 8.45 (dd, 1H, J ) 8.7, 2.6 Hz), 8.58 (d, 1H,
J ) 2.6 Hz). Anal. Calcd for C15H9N2O6F3: C, 48.66; H, 2.45;
N, 7.57. Found: C, 48.67; H, 2.57; N, 7.60.
p-Nitr oph en yl 2,4-Din itr oph en ylacetate (2a). Yield 32%;
mp 136 °C; IR 1759 (CdO); NMR δ 4.43 (s, 2H), 7.30 (d, 2H,
J ) 8.9 Hz), 7.70 (d, 1H, J ) 8.9 Hz), 8.30 (d, 2H, J ) 9.3 Hz),
8.54 (dd, 1H, J ) 9.3, 2.4 Hz), 9.07 (d,1H, J ) 2.4 Hz). Anal.
Calcd for C14H9N3O8: C, 48.40; H, 3.67; N, 12.10. Found: C,
48.35; H, 3.55; N, 12.05.
Acetonitrile was purified as described before.9 The solutions
+
of R2NH/R2NH2 in 70 mol % MeCN(aq) were prepared by
dissolving equivalent amount of R2NH and R2NH2+ in 70 mol
J . Org. Chem, Vol. 67, No. 15, 2002 5237

Cho et al.
% MeCN(aq). In all cases, the ionic strength was maintained
to 0.1 M with Bu4N+Br-.
in 70 mol % MeCN-30% D2O at -10 °C. The reaction was
quenched, and the reactant was isolated as described before.9-11
The proton NMR spectrum of the recovered reactant from 1a
was identical to that of the substrate except that the integra-
tion at δ 4.05 decreased by approximately 30%. From the
reaction of 2a , the NMR spectrum of the recovered reactant
indicated compete absence of the benzylic proton resonance
at δ 4.43.
Kin etic Stu d ies. Reactions of 1a -e and 2a with R2NH/
+
R2NH2 in 70 mol % MeCN(aq) were followed by monitoring
the increase in the absorbance of the aryl oxides at 400-434
nm with a UV-vis spectrophotometer as described before.9-11
In all cases, the ionic strength was maintained to be 0.10 M
with Bu4N+Br-.
Ca lcu la tion of k2E, k1, a n d k-1/k2 Va lu es. Utilizing the
kobs values and the base concentration, the k2E, k1, and k-1/k2
values that best fit with eq 2 have been calculated with the
Origin program, which is a nonlinear square fitting routine
based on the gradient search and function linearization
method.27
Con tr ol Exp er im en ts. The stabilities of 1a -e and 2a were
determined as reported earlier.9-11 They were stable for at
least 3 weeks in MeCN solution at room temperature.
Ack n ow led gm en t. This work was supported by
NRL-MOST and CRM-KOSEF. H.C.J . was supported
by BK21 scholarship. We thank Prof. D. Kim at PNU
for the computer fitting.
P r od u ct Stu d ies. The product of elimination from 2a
promoted by i-Bu2NH/i-Bu2NH2+ in 70 mol % MeCN(aq) were
identified as described previously.9-11 From this reactions,
N,N-diisobutyl-2,4-dinitrobenzamide was obtained in 70%
yield. Also, the yields of the aryloxides from the reactions of
1a -e and 2a determined by comparing the UV absorptions of
the infinity samples with those for the authentic aryloxides
were in the range of 93-99%.
Su p p or tin g In for m a tion Ava ila ble: Rate constants for
+
eliminations from 1a -e and 2a promoted by R2NH/R2NH2
in 70 mol % MeCN(aq) and plots of kobs versus base concentra-
tion. This material is available free of charge via the Internet
at http://pubs.acs.org.
H-D Exch a n ge Exp er im en t. To determine whether 1a
and 2a may undergo H-D exchange reaction, the substrates
(0.19 mmol) were added to i-Bu2NH/i-Bu2NH2+ (0.1 M, 3.4 mL)
J O025555M
5238 J . Org. Chem., Vol. 67, No. 15, 2002