Communication
ly, the role of the second molecule of 1 is to act as a hydrogen
source to generate
(Scheme 2, path a).
a
palladium hydride intermediate
2
ð1Þ
rate ¼ k½Pdꢁ½1ꢁ
To get a better mechanistic understanding, deuterium kinet-
ic isotope effect measurements for the allylation of 2 by
1 were carried out.[11] A deuterium kinetic isotope effect (KIE)
was determined by comparing the rate of allylation of 2 by
[D2]-1 and 1. A large secondary deuterium KIE (kCH/kCD =1.34ꢂ
0.01) was observed (Table 2). The large secondary deuterium
Figure 2. Energy profile of the OꢀH bond and the CꢀO bond cleavage reac-
tions.
Table 2. Kinetic deuterium isotope effects on the allylation of 2 with 1 or
5 by Pd[P(OPh)3]4.
viously been proposed but never observed.[17] Attempts to
1/[D2]-1
1/[D1]-1
1/[D3]-1
[D1]-1/[D3]-1
5/[D2]-5
kCH/kCD
kOH/kOD
kCHOH/kCDOD
kCDOH/kCDOD
kCH/kCD
1.34ꢂ0.01
2.06ꢂ0.08
2.05ꢂ0.02
1.00ꢂ0.05
1.11ꢂ0.04
1
detect a palladium hydride complex by H NMR failed, proba-
bly due to a fast conversion (Figure 2).[17a] To overcome this
problem, ESI-MS was used to analyze the hydride complex for
the oxidative addition step. Attempts to detect the palladium
hydride complex of 1 and [D2]-1 also failed. This is consistent
with the negligibly lower energy barrier for the OꢀH bond
cleavage, as compared to the CꢀO bond cleavage (Figure 2).
Therefore, the more sterically crowded crotyl alcohol was sub-
jected to the same reaction in the presence of Pd[P(OPh)3]4.
The ESI-MS spectrum revealed a signal at m/z 899, which was
consistent with the Pd hydride complex (PdH(OC4H7)NEt3[P-
(OPh)3]2). The structure of the ion at m/z 899 was further con-
firmed by the MS/MS collision-induced-dissociation (CID),
where ions of complexes with the masses that correspond to
PdH(OC4H7) [P(OPh)3]2 and PdH(OC4H7)P(OPh)3 were also ob-
KIE is consistent with the proposal that the CꢀO bond cleav-
age occur either before or during the rate-determining step.
For comparison, we carried out additional deuterium KIE ex-
periments with 5 having a good leaving group. A secondary
KIE for allyl benzoate (kCH/kCD =1.11ꢂ0.04) was determined by
preparing allyl [D2]-1,1-benzoate ([D2]-5) and comparing its rate
of allylation to 5 at 258C.[12] The use of CH2 =CHCH2OD ([D1]-1)
for the allylation of 2 with Pd[P(OPh)3]4 gave a primary deuteri-
um KIEs (kOH/kOD =2.06ꢂ0.08) indicating that an OꢀH bond
cleavage occurs either before or during the rate-determining
step (Table 2).[13] To determine whether the cleavage of the Oꢀ served (Figure 3).
H proceeds before, simultaneously, or after the CꢀO bond
cleavage, the allylation of 2 was carried out with doubly la-
beled CH2 =CHCD2OD ([D3]-1). Comparison of the rate con-
stants for the reaction of amination with 1 and [D3]-1 gave
deuterium KIEs of kCHOH/kCDOD =2.05ꢂ0.02 (Table 2). Thereby,
the doubly labeled allyl alcohol ([D3]-1) showed a similar KIE
(2.05) as observed for [D1]-1 (2.06).[14]
The absence of a product isotope effect rules out the possi-
bility that the second molecule of 1 activates the hydroxyl
group of 1, which is similar to what has been proposed for
water or ammonium ions (Scheme 2, paths b and c).[15] The re-
sults are consistent with a mechanism in which the OꢀH bond
cleavage occurs in a separate step, prior to the CꢀO bond
cleavage, with either a negligibly lower (Figure 2, Ea’) or with
a similar (Figure 2, Ea’’) activation energy (Figure 2). This would
explain why the secondary KIE is not observed in the presence
of the primary KIE.[16] These data are also consistent with the
observed second-order dependence in 1. Thus, the primary
deuterium KIE of 2.06, may indeed indicate an insertion by pal-
ladium to the OꢀH bond of 1 to generate a palladium hydride
intermediate in the rate-determining step. Similar KIE have
been observed in rate-determining OꢀH bond cleavage by
other transition metal catalysts.[13]
Both[18] an inner-sphere mechanism,[19] in which the nucleo-
phile coordinates the metal prior to attack, and an outer-
sphere[2e,20] mechanism, in which the nucleophile attacks the
p-allyl without prior coordination to the metal, have been sug-
gested for the palladium-catalyzed allylation of amines. To de-
termine whether the allylation of aromatic amines proceeds
through an inner or outer-sphere mechanism, enantiomerically
enriched allyl alcohol (R)-6 was prepared and converted to the
corresponding piperidine in the presence of Pd[P(OPh)3]4
(Scheme 5). The allylic substitution of (R)-6 proceeded with an
overall retention of stereochemistry (double inversion), as de-
termined by chiral HPLC and single-crystal X-ray analysis of hy-
drochloride salt of the piperidine (R)-7 (Figure 4). The forma-
tion of (R)-7 indicated that the reaction proceeded through an
outer-sphere mechanism. Moreover, the palladium-catalyzed
intramolecular amination of enantioenriched allyl alcohol (R)-6
to (R)-7 proceeded with an excellent chirality transfer.
A mechanism for the palladium-catalyzed direct amination
of allyl alcohols is proposed (Scheme 6). Palladium inserts into
the OꢀH bond of 1 to generate palladium hydride intermedi-
ate A. This step has an activation barrier which is either negli-
gibly lower or equal to the rate-determining step (k2 ꢃk1). This
is consistent with the observed primary KIE for the OꢀH bond
cleavage (kOH/kOD =2.06). The ion of the corresponding palladi-
um hydride complex from crotyl alcohol was observed by ESI-
A palladium hydride intermediate that promotes the CꢀO
bond cleavage to generate a p-allypalladium complex has pre-
Chem. Eur. J. 2014, 20, 1520 – 1524
1522
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim