Scheme 3
1H NMR and 31P NMR analyses in D2O (Scheme 3). The
NMR spectra of the reaction mixture were almost identical
with the solution of [PdCl(η3-allyl)]2 in D2O.5
Initially, we speculated that carbon dioxide from sodium
carbonate might act as the activator, since Yamamoto et al.
have already reported a formation of π-allylpalladium from
allyl alcohol under a pressure of CO2 (30 atm) via a half-
carbonate ester.1f However, the fact that the reaction also
proceeds in the absence of sodium carbonate clearly rules
out this possibility. Involvement of allyl acetate via trans-
esterification with AcOEt is not probable because the reaction
occurred in an Et2O-water biphasic system.6 At this time,
we propose the reaction mechanism in which water activates
allyl alcohol via hydration of the hydroxy group and
stabilizes the resulting hydroxide ion by strong solvation with
water, as outlined in Scheme 4. In organic solvents, elimina-
Scheme 4
tion of naked hydroxide is a very unfavorable process. In
water, however, the hydroxide ion can leave with hydrating
water molecules and the negative charge can be delocalized
in the water cluster.
To elucidate the effect of water, we have conducted
theoretical calculations.7 The geometry of palladium com-
plexes PC and the transition state TS was optimized with
the B3LYP functional. The 6-31G* basis set was employed
except for Pd, with which LANL2DZ+ECP were used
(denoted as BS1). During optimization, solvent effects were
taken into account by means of the Onsager method.8 At
the optimized geometries, energies were calculated with the
PCM method at the B3LYP/BS1 level.9
Figure 1. Calculated structures, Mulliken charges (italics), and
activation energies.
one water hydration, whereas Ea ) 26.8 kcal/mol for two
water hydration). Electron population analyses (Mulliken
charge) indicate that negative charge delocalizes in the
solvating waters in TS, and an increased degree of hydration
results in decreasing each charge on oxygen atoms (Figure
1). In the real system, a greater number of water molecules
The calculated activation energy with the PCM method
was very high (Ea ) 60.7 kcal/mol).10 We then included
hydrating waters coordinating to the hydroxy group. Addition
of hydrating waters changed the structure of transition state
depending on the number of waters during reoptimization.
Introduction of hydrating waters significantly lowered the
activation energy, and increasing the number of waters
resulted in decrease of the activation energy (Ea ) 36.0 for
(7) For theoretical calculations involving π-allylpalladium, see: (a)
Delbecq, F.; Lapouge, C. Organometallics 2000, 19, 2716. (b) Branchadell,
V.; Moreno-Manas, M.; Pleixats, R. Organometallics 2002, 21, 2407. (c)
Solin, N.; Szabo´, K. J. Organometallics 2001, 20, 5464. (d) Szabo´, K. J.
Chem. Eur. J. 1997, 3, 592. (e) Szabo´, K. J. Chem. Eur. J. 2000, 6, 4413.
(f) Solin, N.; Narayan, S.; Szabo´, K. J. J. Org. Chem. 2001, 66, 1686. (g)
Szabo´, K. J.; Hupe, E.; Larsson, A. L. E. Organometallics 1997, 16, 3779.
(h) Szabo´,K. J. Organometallics 1996, 15, 1128. (i) Szabo´, K. J. J. Am.
Chem. Soc. 1996, 118. 7818. (j) Sakaki, S.; Takeuchi, K.; Sugimoto, M.;
Kurosawa, H. Organometallics 1997, 16, 2995. (k) Sakaki, S.; Satoh, H.;
Shono, H.; Ujino, Y. Organometallics 1996, 15, 1713. (l) Biswas, B.;
Sugimoto, M.; Sakaki, S. Organometallics 1999, 18, 4015. (m) Branchadell,
V.; Moreno-Man˜as, M.; Pajuelo, F.; Pleixats, R. Organometallics 1999,
18, 4934; (n) Hagelin, H.; Åkermark, B.; Norrby, P.-O. Chem. Eur. J. 1999,
5, 902; (m) Dedieu, A. Chem. ReV. 2000, 100, 543.
(5) See the Supporting Information.
(6) The reaction of acetylacetone (1b) with allyl alcohol in an Et2O-
water biphasic system provided 3b in 53% yield under otherwise identical
conditions as entry 2 in Table 1.
Org. Lett., Vol. 6, No. 22, 2004
4087