Remarkable catalytic effect of H+ in Michael-type additions of anilines to 3-butyn-2-one

Article abstract of DOI:10.1016/S0040-4020(01)00981-4

Second-order rate constants (k_N) for the Michael-type reaction of 3-butyn-2-one (1) with a series of anilines in H₂O have been determined spectrophotometrically. The k_N values are dependent on the free aniline fraction (F<

Full text of DOI:10.1016/S0040-4020(01)00981-4

TETRAHEDRON
Pergamon
Tetrahedron 57 #2001) 9585±9589
Remarkable catalytic effect of H1 in Michael-type additions
of anilines to 3-butyn-2-one
Ik-Hwan Um,p Eun-Ju Lee and Ji-Sook Min
Department of Chemistry, Ewha Womans University, Seoul 120-750, South Korea
Received 20 June 2001; accepted 2 October2001
AbstractÐSecond-order rate constants #kN) forthe Michael-type reaction of 3-butyn-2-one # 1) with a series of anilines in H2O have been
determined spectrophotometrically. The kN values are dependent on the free aniline fraction #FN). The plot of log kN vs pKa of the conjugate
acid of the anilines is linear for the reactions run at FNˆ1.00. However, the Brùnsted-type plot for the reactions performed at FNˆ0.50 is
nonlinear, suggesting a change in the reaction mechanism as the basicity of anilines changes. The analysis of the kinetic results has revealed
that the reaction of anilines proceeds through the protonated form of the substrate 1 as well as the unprotonated. The protonated form of 1 is
of the order of 109 times more reactive than the unprotonated form toward anilines. The contribution of the reaction with the protonated
species to the kN is suggested to be responsible for the nonlinear Brùnsted-type plot obtained for the reactions run at FNˆ0.50. q 2001
Elsevier Science Ltd. All rights reserved.
1. Introduction
study to reactions of 1 with a series of aromatic amines as
shown in Eq. #1). We found the addition reaction is remark-
ably catalyzed by H1 and report herein the detailed reaction
mechanism.
Nucleophilic additions of amines to carbon±carbon double
bonds conjugated with a strong electron withdrawing group
#EWG) have been intensively investigated, and their
reaction mechanisms are fairly well understood.1±4 The
corresponding reactions of carbon±carbon triple bonds
with amine nucleophiles have also attracted a great deal
of attention.5±8 However, most studies have been focused
on the stereochemistry of the reaction products #e.g. the E-
and Z-isomers of enamines) due to synthetic interests.5±8
Kinetic studies of addition reactions of amines to electron
de®cient acetylenes have been little investigated, and there-
fore, the reaction mechanism is not clearly understood.9
ꢀ1†
ArNH2ˆZ-C6H4NH2; Zˆ4-CN, 3-Cl, 4-Cl, 3-Me, 4-MeO,
4-OH.
2. Results and discussion
All the reactions obeyed pseudo-®rst-order kinetics, and
pseudo-®rst-order rate constants #kobs) were obtained from
the well-known equation: ln#A12At)ˆ2kobst1c. The corre-
lation coef®cients of the linear regressions were always
higherthan 0.9995. Plots of kobs vs total aniline concen-
tration were linear and passed through the origin for all
the anilines studied. Therefore, the rate law for the present
reactions can be given as Eq. #2), in which S is the substrate
1, kN is the second-order rate constant, FN is the free aniline
fraction, and [N]tot is the concentration of total aniline #free
aniline and its conjugate acid). The kN values were obtained
by dividing the slope of linearplots of kobs vs [N]tot by FN.
The kN values determined in this way are summarized in
Table 1. It is estimated from replicate runs that the uncer-
tainty in any particular measured rate constant is less than
^3%.
We recently performed Michael-type addition reactions of a
series of aliphatic primary amines including a-effect
nucleophiles such as methoxylamine and hydrazine to an
electron de®cient acetylene, 3-butyn-2-one #1).9 We
expected a large a-effect forthe reaction of methoxylamine
and hydrazine with the sp hybridized electrophile #1), since
the magnitude of the a-effect has been reported to increase
with increasing s character of the electrophile.10±12
However, we found that the a-effect is unexpectedly
small, and gave some explanations forthe small a-effect.9
It appeared to us as highly informative to investigate the
reaction mechanism for additions of amines to electron
de®cient acetylenes. Therefore, we have expanded our
Keywords: Michael-type reaction; reaction mechanism; Brùnsted-type plot;
acid catalysis; acetylene; a-effect.
Corresponding author. Tel.: 182-2-3277-2349; fax: 182-2-3277-2844;
d‰PŠ=dt ˆ kobs‰SŠ; where kobs ˆ kNFN‰NŠ
One can see that the kN values are dependent on FN, i.e. the
ꢀ2†
tot
p
e-mail: ihum@mm.ewha.ac.kr
0040±4020/01/$ - see front matter q 2001 Elsevier Science Ltd. All rights reserved.
PII: S0040-4020#01)00981-4

9586
I.-H. Um et al. / Tetrahedron 57 +2001) 9585±9589
Table 1. Summary of second-order rate constants, kN #M21 s21) for
Michael-type addition reactions of substituted anilines #Z-C6H4NH2) to 1
in H2O at 25.0^0.18C
large slope to a small one as the basicity of nucleophilic
amines increases.13±16 Such a break or curvature in a
Brùnsted-type plot has been interpreted as due to a change
in the rate-determining step #RDS) as the basicity of amines
increases.13±16 Interestingly, as shown in Fig. 1, the slope of
the Brùnsted-type plot for the reactions run at FNˆ0.50 is
smallerforthe less basic anilines than forthe moer basic
ones, which is opposite to the nonlinearBrùnsted-type plot
obtained formost aminolysis eractions. Such a nonlinear
Brùnsted-type plot as in the present system has been rarely
reported.17,18 Cevasco et al. obtained a nonlinearBrùnsted-
type plot for the base catalyzed hydrolysis of aryl 4-hydroxy-
benzoates.18 The rate of hydrolysis was found to be faster
for esters with good leaving groups #low basic aryloxides)
than was predicted from the Brùnsted-type plot of esters
with poor leaving groups #highly basic aryloxides). The
nonlinearBrùnsted-type plot was suggested to be due to a
change in the reaction mechanism from a simple nucleo-
philic substitution reaction for esters of strongly basic aryl-
oxides to an E1cb reaction mechanism for esters of weakly
basic aryloxides.18 Therefore, one might consider that the
nonlinear Brùnsted-type plot obtained in the present system
is indicative of a change in the reaction mechanism as the
basicity of ArNH2 changes.
1
Z
pKa #Z-C6H4NH3
)
kN #M21 s21
)
FNˆ1.00 FNˆ0.75 FNˆ0.50 FNˆ0.25
4-CN
3-Cl
4-Cl
H
3-Me
4-OMe 5.29
4-OH 5.50
1.74
3.32
3.81
4.58
4.67
0.0151
0.162
0.258
0.376
0.536
1.36
0.0562
0.181
0.254
0.410
0.536
1.35
0.161
0.219
0.296
0.415
0.560
1.40
0.475
0.348
0.421
0.593
0.654
1.56
1.49
1.50
1.51
1.68
kN value increases as the FN value decreases for a given
aniline. However, the dependence of kN values on FN is
greater for the less basic anilines. For example, the kN values
for the least basic 4-cyanoaniline increases from
0.0151 M21 s21 to 0.0562, 0.161 and 0.475 M21 s21 as FN
decreases from 1.00 to 0.75, 0.50 and 0.25, respectively,
while the kN value forthe most basic 4-hydorxyaniline
increases from 1.49 to 1.68 M21 s21 as FN decreases from
1.00 to 0.25.
The effect of basicity on the kN values is demonstrated in
Fig. 1. The Brùnsted-type plot is linear for the reaction run
at FNˆ1.00. However, interestingly, the corresponding plot
for the reactions performed at FNˆ0.50 is nonlinear, i.e.
anilines with an electron withdrawing substituent exhibit
positive deviations from the linearity. #The plots for the
reactions run at FNˆ0.75 and 0.25 are also nonlinear, but
®gures are not shown.)
Since the positive deviation from the linear Brùnsted-type
plot in Fig. 1 is greater for the less basic ArNH2 and forthe
reaction run at lower FN #lowerpH) fora given ArNH 2, the
present reaction is considered to be speci®c acid catalyzed.
Thus, one can propose that ArNH2 reacts with the unproton-
ated substrate 1 and with the protonated species 2 as shown
in Eqs. #1) and #4), respectively.
Aminolysis of esters with a good leaving group have often
shown a break or a curvature in Brùnsted-type plots from a
ꢀ3†
ꢀ4†
Let [S] and [SH1] be the concentrations of the substrate 1
and its protonated form, respectively. Then, the total
concentration of the substrate [S]tot is the sum of [S] and
[SH1]. One can derive Eqs. #5)±#7) from Eq. #3) and the
relationship [S]totˆ[S]1[SH1].
Ka ˆ ‰SŠ‰H1Š=‰SH1Š
‰SŠ ˆ ‰SŠtot=ꢀ1 1 ‰H1Š=Ka†
‰SH1Š ˆ ‰SŠtot=ꢀ1 1 Ka=‰H1Š†
Then, the rate law can be expressed as Eqs. #8) and #9).
ꢀ5†
ꢀ6†
ꢀ7†
Rate ˆ k2‰NŠ‰SŠ 1 k3‰NŠ‰SH1Š
Rate ˆ ꢀk2Ka 1 k3‰H1Š†‰NŠ‰SŠtot=ꢀKa 1 ‰H1Š†
ꢀ8†
ꢀ9†
The Ka value of the conjugate acid of acetone has been
reported to be 107 in H2O.19 The resonance effect due to
the alkynyl±carbonyl conjugation may increase the base
strength of 1. However, one can expect an sp carbon is
more electronegative than an sp3 carbon and the Hammett
sp constant is much larger for the ethynyl group
Figure 1. A Brùnsted-type plot for Michael-type additions of substituted
anilines #Z-C6H4NH2) to 1 in H2O at 25.0^0.18C.

I.-H. Um et al. / Tetrahedron 57 +2001) 9585±9589
9587
strated graphically in Fig. 3. As shown in Table 2, the k3/Ka
value is much larger than the k2 value fora given aniline,
e.g. the k3/Ka values are about 5.4£102 and 1.4£104 times
larger than the k2 values for4-cyanoaniline and 4-hydroxy-
aniline, respectively. Since the Ka value of the protonated
species 2 is considered to be of the order of 107, as
mentioned above, the protonated species 2 will be of the
order of 109±1011 times more reactive than the unprotonated
species 1, indicating that the catalytic effect of H1 on the
addition of ArNH2 to the substrate 1 is remarkable.
There has been no previous report on the acid catalyzed
addition of amines to activated acetylenes as in the present
study. However, additions of amines to the carbonyl group
have been suggested to proceed through acid catalyzed as
well as uncatalyzed processes.20 Foracetoxime formation at
258C in water, Jencks reported that the rate constant for the
acid catalyzed process is 1.7£105 M22 s21, while the one for
. Since the Ka
the uncatalyzed process is 2.3£103 M21 s21 20
value for the conjugate acid of acetone has been reported
to be 107 in H2O,19 the acid catalyzed process is of the
order of 108±109 times fasterthan the uncatalyzed process.
Therefore, the catalytic effect of H1 in the present system is
more remarkable than the one in the acetoxime formation
reaction.
Figure 2. A plot of kobs/[N] vs [H1] forthe Michael-type addition of
4-cyanoaniline to 1 in H2O at 25.0^0.18C.
In Fig. 3 are shown linear Brùnsted-type plots for both the k2
and k3 processes. Since log k3/Kaˆlog k32log Ka and Ka 1is
constant, the slope of the plot of log k3/Ka vs pKa #ArNH3 )
#spˆ10.23) than forthe methyl group # spˆ20.17). Since
the resonance effect in 1 is opposite to the inductive effect in
1, the Ka value forthe conjugate acid of 1 would be of the
order of 107.
is the same as that of log k3 vs pKa #ArNH31). The bnuc
values in the present system have been calculated to be
0.51 and 0.93 forthe k2 and k3 process, respectively. We
recently reported that the bnuc value is 0.32 for the corre-
sponding reactions of 1 with a series of aliphatic primary
amines in H2O9 and 0.28 forthe eractions of an activated
acetylene, methyl propiolate #HCuCCO2Me) with a series
of alicyclic secondary amines in H2O.21 Therefore, the bnuc
value of 0.93 forthe k3 process in the present system is
unusually large. The magnitude of bnuc values has been
suggested to represent the degree of relative bond formation
between the nucleophile and the substrate at the rate-deter-
mining transition state #TS).22 Therefore, one can suggest
that the degree of bond formation for the acid catalyzed
process is signi®cantly more advanced than the uncatalyzed
process and analogous reactions.
Since the present reactions were performed under pseudo-
®rst-order conditions #i.e. [N]@[S]tot), Eq. #10) can be
derived from Eq. #9). The pH of the reaction medium in
the present system was above 1.0 even for the reaction
with the least basic 4-cyanoaniline #i.e. [H1],1021),
while Ka would be much larger than [H1] in the present
system. Thus, Eq. #10) can be simpli®ed to Eq. #11):
kobs ˆ ꢀk2Ka 1 k3‰H1Š†‰NŠ=ꢀKa 1 ‰H1Š†
ꢀ10†
ꢀ11†
k
obs=‰NŠ ˆ kN ˆ k2 1 k3‰H1Š=Ka
If the present reaction proceeds as described above, the plot
of kobs/[N] vs [H1] should be linear. In fact, as exempli®ed
in Fig. 2, the plot of kobs/[N] vs [H1] is linearwith a positive
intercept for all the anilines studied. Accordingly, the k2 and
k3/Ka values were determined from the intercepts and slopes
of the linearplots, erspectively. The k2 and k3/Ka values
obtained in this way are summarized in Table 2 and demon-
As shown in Fig. 3, the Brùnsted-type plot for the present
reactions is linear for both the k2 and k3 processes. Such a
linear Brùnsted-type plot clearly suggests that the reaction
mechanism does not change upon changing the basicity of
ArNH2 in the present system. The nonlinear Brùnsted-type
plot shown in Fig. 1 for the reactions run at FNˆ0.50 is
de®nitely not due to a change in the reaction mechanism
but due to the contribution of the k3[H1]/Ka term to the kN
value. As shown in Eq. #11), the kN value consists of two
terms, k2 and k3[H1]/Ka. The contribution of the latter term
to the kN value is negligible for the reactions run at
FNˆ1.00, since the pH of the reaction medium was higher
than 7 when FNˆ1.00 #i.e. [H1],1027). This argument can
be supported by the fact that the k2 value in Table 2 and the
kN value for the reactions run at FNˆ1.00 in Table 1 are the
same within experimental error. However, the contribution
of the k3[H1]/Ka term to the kN value becomes signi®cant as
Table 2. Summary of k2 and k3/Ka values forMichael-type addition reac-
tions of substituted anilines #Z-C6H4NH2) to 1 in H2O at 25.0^0.18C
1
Z
pKa #Z-C6H4NH3
)
k2 #M21 s21
)
k3/Ka #M22 s21
)
4-CN
3-Cl
4-Cl
1.74
3.32
3.81
4.58
4.67
5.29
5.50
0.0154
0.163
0.246
0.375
0.530
1.34
8.34
127
368
2630
1980
13800
20600
H
3-Me
4-Ome
4-OH
1.47

9588
I.-H. Um et al. / Tetrahedron 57 +2001) 9585±9589
before use. Doubly glass distilled water was further boiled
and cooled under nitrogen just before use.
3.2. Kinetics
The kinetic study was performed with a Sinco S-1130
UV±Vis spectrophotometer equipped with a constant
temperature circulating bath to keep the reaction mixture
at 25.0^0.18C. The reactions were followed by monitoring
the appearance of the addition product at a ®xed wavelength
corresponding to the maximum absorption of the enamine.
All the reactions were carried out under pseudo-®rst-order
conditions in which the concentration of free ArNH2 was at
least 20 times greater than that of the substrate 1. The kinetic
study was performed under 4 different free aniline fractions
#e.g. FNˆ1.00, 0.75, 0.50 and 0.25). The FN values were
controlled by adjusting the ratio of [ArNH2]/
{[ArNH2]1[ArNH31Cl2]}. At least ®ve different ArNH2
concentrations were used to obtain the second-order rate
constant #kN) foreach ArNH 2 at a given FN. Otherdetailed
kinetic methods were reported previously.16a
3.3. Product analysis
1
Product analysis was performed by H NMR spectroscopy.
The 1H NMR spectrum of the reaction mixture for the reac-
tion of 1 with aniline revealed that no 1,2-addition product
but only the Michael-type addition product, #E)-4-anilino-3-
buten-2-one, was formed.
Figure 3. Plots of logarithmic k2 and k3/Ka vs pKa of the conjugate acid of
anilines forMichael-type additions of substituted anilines # Z-C6H4NH2) to
1 in H2O at 25.0^0.18C.
the pH of the reaction medium decreases especially for
weakly basic anilines. This is consistent with the fact that
the kN value fora given aniline increases with decreasing the
FN value #i.e. increasing [H1]), and that the positive devia-
tion from the linear Brùnsted-type plot in Fig. 1 is more
Acknowledgements
The authors are grateful for the ®nancial support from the
Korea Institute of S&T Evaluation and Planning through the
Women's University Research Fund #KISTEP-00-B-WB-
06-A-07). Helpful discussions with Professors Erwin
Buncel and Howard Maskill are also acknowledged.
signi®cant forthe less basic ArNH
.
2
We recently reported a similar result for the nucleophilic
substitution reaction of 4-nitrophenyl X-substituted benzo-
ates with a series of alicyclic secondary amines.16a The
Hammett plots of log kN vs the s constant of the benzoyl
substituent X were nonlinear for all the amines studied due
to negative deviations by the p-electron donor substituents
such as 4-MeO, suggesting a change in the RDS. However,
the corresponding Yukawa±Tsuno plots resulted in good
linear correlation with s1r#s12s), indicating that the
nonlinearHammett plots weer not due to a change in
the RDS but due to changes in the resonance supply of the
p-electron donor substituent.16a Therefore, one can suggest
that deductions of a reaction mechanism based just on a
linearora nonlinearBùrnsted orHammett plot can be
misleading.
References
1. Reviews: #a) Bernasconi, C. F. Acc. Chem. Res. 1987, 20,
301±308. #b) Bernasconi, C. F. Tetrahedron 1989, 45,
4017±4090. #c) Kutyrev, A. A.; Moskva, V. V. Russ. Chem.
Rev. 1991, 60, 72±106.
2. #a) Bernasconi, C. F.; Leyer, A. E.; Rappoport, Z. J. Org.
Chem. 1999, 64, 2897±2902. #b) Bernasconi, C. F.; Zitomer,
J. L.; Schuck, D. F. J. Org. Chem. 1992, 57, 1131±1139.
3. #a) Oh, H. K.; Yang, J. H.; Lee, H. W.; Lee, I. J. Org. Chem.
2000, 65, 5391±5395. #b) Oh, H. K.; Yang, J. H.; Lee, H. W.;
Lee, I. J. Org. Chem. 2000, 65, 2188±2191. #c) Oh, H. K.;
Yang, J. H.; Sung, D. D.; Lee, I. J. Chem. Soc., Perkin Trans. 2
2000, 101±105.
3. Experimental
3.1. Materials
4. #a) Varghese, B.; Kothari, S.; Banerji, K. K. J. Chem. Res. +S)
1998, 422. #b) Jalani, N.; Kothari, S.; Banerji, K. K. Can. J.
Chem. 1996, 74, 625±629.
5. Perlmutter, P. Conjugate Addition Reactions in Organic
Synthesis; Pergamon: Oxford, 1992, and references cited
therein.
3-Butyn-2-one 1 was prepared by oxidizing 3-butyn-2-ol
with chromic acid as described in the literature.23 Other
chemicals including all the anilines used in the present
study were of the highest quality available from Aldrich
or Tokyo Kasei, and generally recrystallized or distilled
6. #a) Truce, W. E.; Tichenor, G. J. J. Org. Chem. 1972, 37,
2391±2396. #b) Truce, W. E.; Gorbaty, M. L. J. Org. Chem.
1972, 35, 2113±2117.

I.-H. Um et al. / Tetrahedron 57 +2001) 9585±9589
9589
7. #a) Crisp, G. T.; Millan, M. J. Tetrahedron 1998, 4, 637±648.
#b) Sinsky, M. S.; Bass, R. G. J. Heterocyclic Chem. 1984, 21,
759±768.
15. #a) Koh, H. J.; Han, K. L.; Lee, I. J. Org. Chem. 1999, 64,
4783±4789. #b) Koh, H. J.; Han, K. L.; Lee, I. J. Org. Chem.
1998, 63, 983±9839. #c) Lee, I. J. Org. Chem. 1997, 62, 5780±
5784.
8. #a) Ma, S.; Lu, X.; Li, Z. J. Org. Chem. 1992, 57, 709±713.
#b) Ma, S.; Lu, X. J. Chem. Soc., Chem. Commun. 1990,
1643±1644.
16. #a) Um, I. H.; Min, J. S.; Ahn, J. A.; Hahn, H. J. J. Org. Chem.
2000, 65, 5659±5663. #b) Um, I. H.; Kim, M. J.; Lee, H. W.
Chem. Commun. 2000, 2165±2166. #c) Um, I. H.; Min, J. S.;
Lee, H. W. Can. J. Chem. 1999, 77, 659±666.
9. Um, I. H.; Lee, J. S.; Yuk, S. M. J. Org. Chem. 1998, 63,
9152±9153.
17. #a) Cevasco, G.; Thea, S. J. Org. Chem. 1999, 64, 5422±5426.
#b) Broxton, T. J.; Dubby, N. W. J. Org. Chem. 1981, 41,
1186±1191.
10. #a) Um, I. H.; Buncel, E. Chem. Commun. 2001, 27±28.
#b) Um, I. H.; Park, Y. M.; Buncel, E. Chem. Commun.
2000, 1917±1918. #c) Um, I. H.; Buncel, E. J. Org. Chem.
2000, 65, 577±582. #d) Um, I. H.; Yon, H. W.; Lee, J. S.;
Moon, H. J. J. Org. Chem. 1997, 62, 5939±5944.
11. #a) Fountain, K. R.; Tad-Y, D. B.; Paul, T. W.; Golynskiy,
M. V. J. Org. Chem. 1999, 64, 6547±6553. #b) Fountain, K. R.;
Dunkin, T. W.; Patel, K. D. J. Org. Chem. 1997, 62, 2738±
2741. #c) Fountain, K. R.; Hutchinson, L. K.; Mulhearn,
D. C. L.; Xu, Y. B. J. Org. Chem. 1993, 58, 7883±7890.
12. #a) Buncel, E.; Chuaqui, C.; Wilson, H. J. Am. Chem. Soc.
1982, 104, 4896±4900. #b) Buncel, E.; Chuaqui, C.; Wilson,
H. J. Org. Chem. 1980, 45, 3621±3626.
18. #a) Page, M. I.; Williams, A. Organic and Bio-Organic
Mechanism; Longman: Essex, UK, 1997; pp 35±36.
#b) Cevasco, G.; Guanti, G.; Hopkins, A. R.; Thea, S.;
Williams, A. J. Org. Chem. 1985, 50, 479±484.
19. Carroll, F. A. Perspectives on Structure and Mechanism in
Organic Chemistry; Brooks/Cole: Tokyo, 1998; p 40.
20. #a) Jencks, W. P. J. Am. Chem. Soc. 1959, 81, 475±481.
#b) Jones, R. A. Y. Physical and Mechanistic Organic
Chemistry; Cambridge University: Cambridge, 1984; p 258.
21. Um, I. H.; Kim, K. H.; Kwon, D. S. Bull. Korean Chem. Soc.
1993, 14, 307±309.
13. #a) Jencks, W. P. Chem. Soc. Rev. 1985, 85, 511±527.
#b) Jencks, W. P.; Gilchrist, M. J. Am. Chem. Soc. 1968, 90,
2622±2637. #c) Gresser, M. J.; Jencks, W. P. J. Am. Chem.
Soc. 1977, 99, 6963±6970.
22. #a) Buncel, E.; Um, I. H.; Hoz, S. J. Am. Chem. Soc. 1989, 111,
971±975. #b) Bernasconi, C. F. 4th ed; Techniques of Organic
Chemistry; Vol. 6; Wiley: New York, 1986.
23. Bowden, K.; Heilbron, I. M.; Jones, E. R. H.; Weddon, B. C. L.
J. Chem. Soc. 1946, 39±45.
14. #a) Castro, E. A. Chem. Rev. 1999, 99, 3505±3524. #b) Castro,
E. A.; Ruiz, M. G.; Salinas, S.; Santos, J. G. J. Org. Chem.
1999, 64, 4817±4820. #c) Castro, E. A.; Saavedra, C.; Santos,
J. G.; Umana, M. I. J. Org. Chem. 1999, 64, 5401±5407.