Mechanism of palladium-catalyzed allylic acetoxylation of cyclohexene

Article abstract of DOI:10.1002/(sici)1521-3765(19980615)4:6<1083::aid-chem1083>3.0.co;2-f

The mechanism of the quinone-based palladium-catalyzed oxidation of cyclohexene to allylic acetate was studied under various reaction conditions with 1,2-dideuterocyclohexene (62-74% D) as the substrate. The reactions gave a 1:1 mixture of 1,2-dideutero-

Full text of DOI:10.1002/(sici)1521-3765(19980615)4:6<1083::aid-chem1083>3.0.co;2-f

FULL PAPER
Mechanism of Palladium-Catalyzed Allylic Acetoxylation of Cyclohexene
Helena Grennberg* and Jan-E. Bäckvall*
Abstract: The mechanism of the qui-
none-based palladium-catalyzed oxida-
tion of cyclohexene to allylic acetate was
studied under various reaction condi-
tions with 1,2-dideuterocyclohexene
(62 ± 74% D) as the substrate. The re-
actions gave a 1:1 mixture of 1,2-dideu-
tero- and 2,3-dideutero-1-acetoxy-2-cy-
clohexene (2 and 3, respectively), as
determined by 1H NMR spectroscopy.
Control experiments with 1-deutero-1-
acetoxy-2-cyclohexene showed that no
1,3-allylic transposition of the acetoxy
group occurred under the reaction con-
ditions employed for the allylic acetox-
ylation. These results provide evidence
for a (p-allyl)palladium intermediate in
the quinone-based palladium-catalyzed
allylic acetoxylation. With added chlor-
ide ligands and extended reaction times,
bis-allylic acetates were formed. The
presence of a stoichiometric amount of
CH3SO3H led to the formation of a
homoallylic acetate.
Keywords: allylic oxidation ´ homo-
geneous catalysis
´
palladium
´
pi-allyl intermediates
labeling
´
isotopic
Introduction
Allylic oxidation of olefins is an important reaction in organic
chemistry because it leads to synthetic intermediates that can
be used for further functionalizations. Direct allylic function-
alization of readily available olefins has already been inves-
tigated.[1] Apart from radical-initiated reactions,[2] reactions
based on selenium[3] and transition metals[4] have attracted
considerable interest. Among all these methods, the quinone-
based palladium-catalyzed allylic acetoxylation stands out as
a practical and highly useful process for the synthesis of allylic
alcohol derivatives [Eq. (1)].[5±10] In particular, simple cyclic
Scheme 1. The two proposed mechanisms for the palladium-catalyzed
allylic acetoxylation: (p-allyl) route (A) versus the oxypalladation route
(B).
acetate[16] would then give the observed allylic acetate. The
pathway of the (s-alkyl) mechanism[11, 13] (Scheme 1, Path B)
would involve attack of the olefin ± palladium complex
by acetate. The (s-alkyl)palladium intermediate obtained
would then undergo a b-elimination to give the observed
product.
The latter pathway was proposed by Winstein et al.,[12] in a
study of the stoichiometric oxidation of 1- and 2-butene. The
involvement of an acetoxypalladation in the catalytic oxida-
tion of cyclohexene to complex mixtures of products with a
PdCl2 ± CuCl2 oxidation system was demonstrated by Henry,
although a competing (p-allyl)palladium pathway was not
completely excluded.[11, 13] The (p-allyl)palladium pathway
was favored by Wolfe et al.[14] based on results from the
catalytic oxidation of isotopically labeled cyclohexene with a
PdII ± HNO3 oxidation system, in which the main by-product
was a homoallylic acetate. Frankel and co-workers also
proposed a (p-allyl)palladium intermediate in the palladi-
olefins are oxidized to their corresponding allylic carboxylates
in excellent yields and high selectivity with p-benzoquinone
(BQ) as a stoichiometric oxidant or electron-transfer medi-
ator.
Two different mechanisms for this palladium-catalyzed
process have been suggested.[11±14] These mechanisms involve
different organometallic intermediates: a (p-allyl)palladium
complex or a (s-alkyl)palladium complex. Both pathways
begin with the coordination of the alkene to PdII. The pathway
of the (p-allyl) mechanism (Scheme 1, Path A) then involves
the formation of a p complex, followed by the cleavage of the
allylic C ± H bond.[15] Subsequent nucleophilic attack by
[*] Ass. Prof. Dr. H. Grennberg, Prof. Dr. J.-E Bäckvall
Department of Organic Chemistry, University of Uppsala
Box 531, SE-75121 Uppsala (Sweden)
Fax: (‡46)18-512524
Chem. Eur. J. 1998, 4, No. 6
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
0947-6539/98/0406-1083 $ 17.50+.25/0
1083

H. Grennberg and J.-E. Bäckvall
FULL PAPER
um-catalyzed oxidation of methyl ole-
ate with CuCl2 or LiNO3 as stoichio-
metric oxidants, and various Pd cata-
lysts, such as palladium chloride, palla-
dium acetate, and supported zero-
valent palladium.[17] Therefore, it seems
that the mechanism depends upon both
the substrate and the oxidation system.
The quinone-based palladium-cata-
lyzed allylic acetoxylation has been
assumed to proceed via the (p-allyl)-
palladium pathway (Path A), although
evidence for such a mechanism has
been lacking until recently. In a pre-
liminary communication,[18] we provid-
ed evidence for the involvement of a
(p-allyl)palladium intermediate in the
palladium(ii)-benzoquinone-catalyzed
allylic acetoxylation of cyclohexene. In
Scheme 2. Expected outcome of the (p-allyl) route (A) versus the oxypalladation route (B) in the
allylic acetoxylation of 1,2 dideuterocyclohexene.
the present paper, we give a full account of these results,
discuss the mechanism, and also present additional results
that support the (p-allyl)palladium route.
of an allylic C ± H bond[15] to yield a (p-allyl)palladium
intermediate 5, or a trans attack[13, 24] by acetate to give 6 takes
place. Both pathways would produce
a dideuterated
1-acetoxy-2-cyclohexene[25], which is fully deuterated at C2.
Assuming that the secondary isotope effect in the nucleophilic
attack by acetate on the (p-allyl)palladium intermediate is
negligible, the (p-allyl) pathway would give equal amounts of
the products 2 and 3. The acetoxypalladation pathway would,
on the other hand, yield only the allylic acetate 2. For an olefin
containing less deuterium, the reasoning is analogous since
the amount of deuterium at C2 in the product would reflect
the degree of deuteration in the starting olefin.[23, 26]
Results and Discussion
Choice and preparation of the substrate: In order to
distinguish between the two possible mechanisms (Scheme 1),
a substrate was required in which the position of the original
double bond can be easily determined in the product.[19]
Olefins such as 3- and 4-methylcyclohexene would not be
useful candidates for this type of study since asymmetrical
olefins have been found to yield mixtures of regioisomeric
allylic acetates.[7] Although studies with the objective of
improving the selectivity by the addition of ligands,[6] strong
acids,[20] or different palladium ± oxidant combinations[21] have
been carried out, the problem of regiocontrol has not yet been
solved and interpretation of the results from the oxidation of
such olefins would not be unambiguous. This problem would,
however, not arise for an isotopically labeled substrate such as
1,2-dideuterocyclohexene (1), which can form only one
regioisomer of each proposed intermediate, thus minimizing
the number of regioisomeric allylic acetates generated. Addi-
tionally, deuterium substituents would not affect the reaction
rates as much as would carbon substituents.
Cyclohexene deuterated in the vinylic position was ob-
tained from hexahydrophthalic anhydride by an a-proton ±
deuterium exchange, with D2SO4 or DCl as the deuterium
source, followed by decarboxylation of the corresponding acid
with Pb(OAc)4.[22] If commercial D2SO4 (99.5% D) was used,
then several cycles of dehydration ± a-proton ± deuterium
exchange ± hydrolysis of the hexahydrophthalic acid formed
in the first step had to be performed prior to decarboxylation
in order to obtain a product sufficiently enriched (>60%) in
deuterium.[23]
The results from the allylic acetoxylation of 1 (62 ± 74% D)
are summarized in Table 1. Oxidation of 1 with Pd(OAc)2 as
the stoichiometric oxidant in acetic acid at room temperature
afforded a 1:1 mixture of the deuterated products 2 and 3, as
1
determined by H NMR spectroscopy[26] (entry 1). This is in
excellent agreement with what is expected from a mechanism
involving a (p-allyl)palladium intermediate. An increase in
the temperature did not affect the outcome of the reaction
(entry 2). The same result was obtained with catalytic
amounts (1 ± 5%) of Pd(OAc)2 and p-benzoquinone (BQ)
as the stoichiometric oxidant (entries 3 ± 6 and 9). Neither the
addition of LiOAc (entry 7), nor the use of Pd/C[17] (entry 10),
[27]
PdCl2 (entry 8), nor Pd(OCOCF3)2 in the presence of 2'-
methoxyacetophenone as ligand[6] (entry 11) affected the
ratio of compounds 2 and 3.
Several different oxidation systems were tried in the allylic
acetoxylation of 1 without any significant change in the ratio
of compounds 2 and 3. Thus, systems such as cat. Pd(OAc)2/
cat. BQ/MnO2, cat. Pd(OAc)2/cat. BQ/cat. iron phthalocya-
nine (Fe(Pc)/O2) and cat. Pd(OAc)2/CuCl2 all afforded a 1:1
ratio of 2 and 3 (entries 12 ± 16). All oxidation systems, apart
from PdII ± CuCl2[28] and PdCl2 ± BQ,[27] yielded clean reactions
with allylic acetate as the main or only observed product.
Control experiments with isolated (h3-allyl)palladium chlor-
ide and acetate complexes[29] did verify that allylic acetate
does form from such complexes in the presence of BQ,
however, much more rapidly for an acetoxy complex than for
a chloro complex.[30]
Allylic acetoxylation: In Scheme 2, the expected outcome of
an allylic acetoxylation of 1,2-dideuterocyclohexene is shown
for the two pathways. In both cases, the initial step would be
the formation of a p-olefin complex 4. Then, either cleavage
1084
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
0947-6539/98/0406-1084 $ 17.50+.50/0
Chem. Eur. J. 1998, 4, No. 6

Palladium-Catalyzed Oxidations
1083±1089
Table 1. Allylic oxidation of 1,2-dideuterocyclohexene (62 ± 74% D).[a, b]
Entry
PdII [%]
Oxidant[d]
(Additive)
T [8C] (time [h])
% D[b]
1H NMR (H3:H2:H1)[c]
Predicted via
Predicted acetoxy- Observed
palladation[e]
p-allyl[e]
ratio[f]
1
2
3
4
5
6
7
8
Pd(OAc)2 (100)
Pd(OAc)2 (100)
Pd(OAc)2 (1)
Pd(OAc)2 (2)
Pd(OAc)2 (5)
Pd(OAc)2 (5)
Pd(OAc)2 (5)
PdCl2 (5)
Pd(OAc)2 (5)
10%Pd/C (3)
Pd(TFA)2 (5)
Pd(OAc)2 (0.5)
Pd(OAc)2 (1)
Pd(OAc)2 (2)
Pd(OAc)2 (5)
Pd(OAc)2 (5)
none
none
A
A
A
A
25 (4h45)
60 (2h40)
60 (26h45)
60 (25h)
60 (20h)
26 ± 28 (25h)
60 (21h)
60 (25h)
60 (20h)
60 (6h)
62
73
73
73
73
73
73
74
73
62
74
63
70
70
74
74
1.81:1:1.81
2.35:1:2.35
2.63:1:1
3.70:1:1
1.43:1:1.50
1.64:1:1.64
1.96:1:2.0
2.0:1:1.98
1.89:1:2.18
1.88:1:1.84
1.88:1:1.75
1.84:1[g]
1.97:1:1.79
1.61:1:1.7
1.57:1:1.51
1.88:1:1.88
2.17:1:2.08[i]
2.23:1:2.23
2.04:1:2.07[i]
2.02:1:2.04[i]
A (0.5m LiOAc)
A
2.42:1:2.42
2.35:1:2.35
1.81:1:1.81
2.42:1:2.42
1.85:1:1.85
2.17:1:2.17
3.85:1:1
3.70:1:1
2.63:1:1
3.85:1:1
2.72:1:1
3.33:1:1
9
B
B
10
11
12
13
14
15
16
B (ligand)[h]
25 (6h)
C
C
C
D
E
60 (50h)
60 (45h)
60 (38h)
60 (5h30)
50 (22h)
2.42:1:2.42
3.85:1:1
[a] The reactions were carried out on a 0.25 ± 0.5 mmol scale. [b] See the Experimental Section. [c] The integral of H2 of 1-acetoxy-2-cyclohexene was used to
normalize the integrals of H3 and H1. See the Experimental Section. [d] A: 200% BQ and B: 100% BQ. See refs. [6, 20, and 35], C: 10% BQ and 110%
MnO2, see refs. [5] and [7], D: 10% hydroquinone, 5% Fe(Pc), st. O2, see ref. [8a], E:10% hydroquinone, 5% Cu(OAc)2, 1 atm O2, see ref. [8b].
[e] Calculated values. [f] Estimated error Æ 0.02. [g] Formation of diacetate with resonance at the same chemical shift as H1. [h] 2'-Methoxyacetophenone,
see ref. [6]. [i] Averaged values (nexp ˆ 2).
To completely verify the (p-allyl)palladium pathway in
allylic oxidation, the potential involvement of a 1,3-allylic
transposition of the acetoxy substituent[31] [Eq. (2)] was
This was also observed in the experiments run in the
presence of BQ and hydroquinone (100 mol% of each), or
excess hydroquinone (150 ± 200 mol%). In some experiments,
a palladium mirror formed, without any significant effect on
the distribution of deuterium in the allylic acetate. Not even
the treatment of 7 with PdCl2(PhCN)2 in THF at room
temperature for 2 h[31a] resulted in a rearrangement of more
than 1 ± 2%. It is most likely that longer reaction times are
required for this kind of substrate.[34] Thus, even if the 1,3-
allylic shift of the acetate can occur under the conditions of
the allylic acetoxylation, the process is apparently not fast
enough to account for the results observed for the allylic
oxidation of 1 (Table 1), and therefore the acetoxypalladation
pathway can be excluded.
investigated. This type of rearrangement, proposed to in-
volve a cyclic s-palladium intermediate, would, if rapid
enough, lead to the observed 1:1 ratio of allylic
acetates 2 and 3 (Table 1) for the acetoxypalladation pathway
too.[32]
Thus, 1-acetoxy-1-deutero-2-cyclohexene (7) (90% D),
prepared by CeCl3-catalyzed NaBD4 reduction of 2-cyclo-
hexenone[33] and subsequent esterification [Eq. (3)], was
Further oxidation of the primary product: 1-Acetoxy-2-cyclo-
hexene, the primary product of an allylic acetoxylation of
cyclohexene, is still a possible substrate for allylic acetoxyla-
tion. Thus, a second oxidation leading to diacetoxycyclohex-
ene may occur. The regiochemistry of the diacetates that may
form in the second oxidation will depend on the mechanistic
pathways involved in that step.
subjected to various reaction conditions. When allylic acetate
7 was treated with catalytic amounts of Pd(OAc)2 (5 mol%)
in the presence of excess BQ (150 ± 200 mol%) in acetic acid
at 608C for 16 h, no rearrangement to 1-acetoxy-3-deutero-2-
cyclohexene (8) [Eq. (4)] could be detected by H NMR
spectroscopy (integration of H1 and H3).
Oxidation of 1-acetoxy-2-cyclohexene (Scheme 3) via a (p-
allyl)palladium intermediate 9 (Path A) could yield 1,2-
diacetoxy-3-cyclohexene (11) and 1,4-diacetoxy-2-cyclohex-
ene (12). It is known that complex 9 only yields 1,4-substituted
products in the presence of benzoquinone,[16b] thus, the bis-
allylic diacetate 12 would be the only product from Path A.
1
Chem. Eur. J. 1998, 4, No. 6
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
0947-6539/98/0406-1085 $ 17.50+.25/0
1085

H. Grennberg and J.-E. Bäckvall
FULL PAPER
1.83:1 for the (p-allyl) mechanism.[26] The
observed ratio of these integrals is in good
agreement with this mechanism (Table 2).
Thus, the regiochemistry of the product
formed in the secondary oxidation of 1-ace-
toxy-2-cyclohexene and the distribution of the
deuterium label in the two-step oxidation of 1
provide further evidence for the involvement of
(p-allyl)palladium intermediates in the palla-
dium-catalyzed allylic acetoxylation reaction.
Isomerization of cyclohexene: In allylic acetox-
ylation, the rate-determining step is most likely
the formation of the (p-allyl)palladium inter-
mediate, since nucleophilic attack on (p-allyl)-
palladium acetate complexes has been found to
Scheme 3. Oxidation of 1-acetoxy-2-cyclohexene.
be very rapid in the presence of BQ.[16] Because of the slow
allylic C ± H bond cleavage, double bond isomerization[37] of
the substrate olefin 1 may compete with formation of (p-
allyl)palladium complex 5. Although this isomerization has
been reported to be slow,[15b, 38] it does seem to occur to a small
extent since the integral ratio of H2 to H3 (Table 1) indicates a
loss of deuterium at C2.[39] As mentioned above, independent
of which mechanism operates in the allylic acetoxylation,
there should be no loss of deuterium at C2. The slight
On the other hand, an oxidation involving acetoxypalladation
(Path B) could yield 1,2-diacetoxy-3-cyclohexene (11), 1,3-
diacetoxy-1-cyclohexene (13), and 1,2-diacetoxy-2-cyclohex-
ene (14). Therefore, it is possible to determine which of the
two mechanistic pathways is operating by GLC and 1H NMR
spectroscopic analyses of the oxidation products.[35]
When 1-acetoxy-2-cyclohexene reacted with Pd(OAc)2 and
BQ under allylic acetoxylation conditions, but in the presence
of chloride ligands,[5, 7, 8] a slow reaction occurs in which 1,4-
diacetate 12[36] was isolated as a 1:1 mixture of cis and trans
isomers, together with unreacted starting material. This is the
expected outcome from a mechanism involving a (p-allyl)
intermediate (Scheme 3, Path A).
The position of the deuterium label in the diacetate
obtained from 1 via secondary in situ oxidation of the allylic
monoacetate (Scheme 4) was also consistent with a (p-allyl)
1
mechanism for both oxidation steps. The predicted H NMR
integration ratio of (H1 ‡ H4) and (H2 ‡ H3) of 1,4-diace-
toxy-2-cyclohexene (15 ‡ 16) obtained from 1 (74% D) is
Scheme 5. Allylic acetates arising from intermediates in the rearrange-
Scheme 4. Diallylic oxidation of cyclohexene 1 via allylic acetates 2 and 3.
ment of cyclohexene 1.
Table 2. Diallylic oxidation of 1,2-dideuterocyclohexene (74% D).[a, b]
Entry
PdII [%]
Oxidant [%] Additive
T [8C] (time) Oxidation product[c]
1H NMR of (15 ‡ 16) (H1 ‡ H4:H2 ‡ H3)[b, c, d, e]
1
2
3
Pd(OAc)2 (5)
Pd(OAc)2 (5)
PdCl2 (5)
BQ(200)
BQ(200)
BQ(200)
150% LiCl
150% LiCl
no additive
60 (26h)
60 (26h)
60 (25h)
(15 ‡ 16)
1.62:1
1.54:1
1.84:1
(15 ‡ 16)
(15 ‡ 16):(2 ‡ 3) 1:1
[a] The reactions were carried out on a 0.25 mmol scale. [b] See the Experimental Section. [c] Determined by 1H NMR spectroscopy of the product mixture.
[d] The integrals of H2 and H3 of 1,4-diacetoxy-2-cyclohexene was used to normalize the integals of H1 and H4. [e] Acetoxypalladation ‡ p-allyl would give
an 1H NMR integration ratio of (2 xD):(2 xD) (i.e. 1:1), whereas the corresponding ratio for p-allyl ‡ p-allyl would be (4 xD)/(4 3xD):1 (i. e. 1.83:1 for
xD ˆ 0.74).
1086
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
0947-6539/98/0406-1086 $ 17.50+.50/0
Chem. Eur. J. 1998, 4, No. 6

Palladium-Catalyzed Oxidations
1083±1089
isomerization may be due to a hydride elimination ± readdi-
tion mechanism, or to a reversible formation of a (p-
allyl)palladium intermediate.
acetate may form from allylic acetate in a rearrangement
reaction.[42] Such a (p-allyl)-acetate attack ± double bond
migration pathway would yield homoallylic acetates 21 and
22 (Scheme 6, Path A).
The relative integral of H3 in the 1-acetoxy-2-cyclohexene
product should remain constant, regardless of the amount of
isomerization of 4 to 17, whereas that of H2 should increase as
the amount of product from the nucleophilic attack on 18
increases (Scheme 5). The distribution of allylic acetates
arising from 5 (2 and 3) to those from 18 (19 and 20) was
calculated as the ratio of the observed 1H NMR integral of H2
to that expected from the amount of deuterium label expected
if no rearrangement of 4 to 17 had occurred.[26, 40] It was found
that under standard conditions, the estimated amount of
product from 18 was 5 ± 10%, but in reactions run with a
stoichiometric amount of Pd(OAc)2, or with catalytic amounts
of Pd/C or Pd(OCOCF3)2, it was as much as 15 ± 20%.
When the oxidation of protic cyclohexene under homo-
allylic oxidation conditions was monitored by 1H NMR
spectroscopy, it was found that signals corresponding to a
(p-allyl)palladium complex was observed prior to the appear-
ance of any other product, and that the initally formed
product (<10 min at 608C) was allylic acetate. The amount of
homoallylic acetate increased with time, as the signals for the
allylic acetate disappeared.[43] This strong indication of a
rearrangement from allylic to homoallylic acetate was verified
in a separate 1H NMR experiment, in which 1-acetoxy-2-
cyclohexene was treated with Pd(OAc)2 and CH3SO3H in
deuterated acetic acid.
Oxidation leading to the homoallylic acetate: Škermark
et al.[20] reported that reaction conditions selective for allylic
acetoxylation, in the presence of a strong acid, such as
CH3SO3H, gave good yields of homoallylic acetate [Eq. (5)]
Conclusion
The quinone ± palladium-catalyzed allylic acetoxylation of
cyclohexene proceeds via a (p-allyl)palladium intermediate,
as shown by the use of 1,2-dideuterocyclohexene as the
substrate. The choice of stoichiometric oxidant, reaction
temperature, and reaction time did not significantly affect
the reaction pathway. Control experiments showed that no
significant isomerization of the allylic acetate via 1,3-allylic
transposition occurred under the normal reaction conditions
for the allylic acetoxylation. These results eliminate acetoxy-
palladation as a possible pathway in the palladium-catalyzed
allylic acetoxylation of cyclohexene. A slight olefin isomer-
ization was observed under allylic acetoxylation conditions.
The observations presented in this paper explain, to some
extent, the results previously observed in the allylic acetox-
ylation of substituted olefins, and could be useful for the
development of new, more selective reaction conditions for
allylic acyloxylation of these substrates.
[41]
.
In a study of 3,3,6,6-tetradeuterocyclohexene, Wolfe and
Campbell observed that small amounts of homoallylic ace-
tates were formed in addition to the allylic acetates that
constituted the major isolated products.[14] The observed
position of the deuterium labels in the homoallylic acetate
indicated that an acetoxypalladation ± elimination ± double
bond migration reaction pathway might be operating parallel
to the p-allyl pathway leading to the allylic product.
For 1,2-dideuterocyclohexene, such an acetoxypallada-
tion ± elimination ± double
bond
migration
sequence
(Scheme 6, Path B) would lead to homoallylic acetate 21.
In order to test this hypothesis, we subjected 1 to
homoallylic oxidation conditions [Eq. (5)].[20] The 1-acetoxy-
3-cyclohexene thus obtained did not show the 1H NMR
integral ratio expected for 21,[26] but rather one with the
deuterium label distributed over C1, C2, C3, and C4. The
integral ratio observed agrees more closely with a mechanism
involving a (p-allyl)palladium intermediate: the homoallylic
Experimental Section
1H NMR spectra were recorded at 300 and 400 MHz in CDCl3 and
[D4]acetic acid solutions. Analytical GLC was performed on a SE54
column (25 m, 250 mm). Chemicals and solvents were from commercial
sources and were used as received. Bis[cyclohexenyl(h3-allyl)palladium]
complexes were prepared according to literature procedures.[29, 44]
1-Acetoxy-1-deutero-2-cyclohexene (7) (90% D): This was pre-
pared by CeCl3-catalyzed NaBD4 reduction of 2-cyclohexenone[33]
and subsequent esterification.
1,2-Didieuterocyclohexene (1): Prepared from hexahydrophthalic
acid as described by Škermark et. al[22] but with commercial D2SO4
(99.5% D). Several cycles of dehydration ± a-proton ± deuterium
exchange-hydrolysis of the hexahydrophthalic acid formed in the
first step had to be performed prior to the decarboxylation in order
to obtain a product sufficiently enriched (>60%) in deuterium. The
amount of deuterium (xD) in the two vinylic positions was
determined by 1H NMR spectroscopy from the ratio of the relative
integral of the vinylic signal (Irel) to that of the nondeuterated
compound, which is 2. Thus, xD can be defined by [1-(Irel/2)], and
varies between 0 and 1, with the latter figure representing complete
deuteration (i. e. 1,2-dideuterocyclohexene).
Scheme 6. Expected outcome of the oxypalladation route (A) versus the (p-allyl)
route (B) in the homoallylic acetoxylation of 1,2-dideuterocyclohexene.
Chem. Eur. J. 1998, 4, No. 6
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
0947-6539/98/0406-1087 $ 17.50+.25/0
1087

H. Grennberg and J.-E. Bäckvall
FULL PAPER
Allylic acetoxylations of 1: These reactions were carried out on a 0.25 mmol
scale in acetic acid (1 mL). The reaction product was, after extractive
workup[8] and removal of most of the solvent, compared with an authentic
sample. The mechanism involving acetoxypalladation would give an 1H
NMR integration ratio of 1-acetoxy-2-cyclohexene for protons H3, H2, and
H1 (at d ˆ 5.94, 5.69 and 5.25, respectively) of 1:(1 xD):(1 xD), where xD
is the relative amount of deuterium in the vinylic position of the starting
cyclohexene. For a (p-allyl)palladium mechanism, the corresponding ratio
would be 0.5(2 xD):(1 xD):0.5(2 xD).
[5] a) H. Grennberg, J. E. Bäckvall, in Transition Metals for Fine
Chemicals and Organic Synthesis (Eds.: C. Bolm, M. Beller), VCH
Weinheim, 1998; b) A. Heumann, B. Škermark, Angew Chem. 1984,
96, 443; Angew. Chem. Int. Ed. Engl. 1984, 23, 453; c) A. Heumann, B.
Škermark, S. Hansson, T. Rein, Organic Synthesis, Vol. 68, pp. 109;
d) A. Heumann, M. Reglier, B. Waegell, Angew. Chem. 1982, 94, 397;
Angew. Chem. Int. Ed. Engl. 1982, 18, 366; e) B. Principato, M. Maffei,
C. Siv, G. Buono, G. Peiffer, Tetrahedron, 1996, 52, 2087.
Ï
Â
[6] J. E. McMurry, P. Kocovsky, Tetrahedron Lett. 1984, 25, 4187.
[7] S. Hansson, A. Heumann, T. Rein, B. Škermark, J. Org. Chem. 1990,
55, 975.
Allylic acetoxylation of 1-acetoxy-2-cyclohexene and diallylic acetoxyla-
tions: These reactions were carried out under allylic acetoxylation
conditions with either 1-acetoxy-2-cyclohexene, protic cyclohexene, or
deuterated cyclohexene 1 as the substrate. The reaction products were,
after extractive workup[8] and removal of most of the solvent, compared
with authentic samples of 1-acetoxy-2-cyclohexene and 1,4-diacetoxy-2-
cyclohexenes. In the oxidation of 1, a mechanism involving two (p-allyl)
[8] a) J. E. Bäckvall, R. B. Hopkins, H. Grennberg, M. M. Mader, A. K.
Awasthi, J. Am. Chem. Soc. 1990, 112, 5160; b) S. E. Byström, E. M.
Larsson, B. Škermark, J. Org. Chem. 1990, 55, 5674; c) E. M. Larsson,
B. Škermark, Tetrahedron Lett. 1993, 34, 2523; d) H. Grennberg, K.
Bergstad, J. E. Bäckvall, J. Mol. Catal. A 1996, 113, 355; e) K.
Bergstad, H. Grennberg, J. E. Bäckvall, Organometallics, 1998, 17, 45.
[9] C. Jia, P. Müller, H. Mimoun, J. Mol. Catal. A 1995, 101, 127.
[10] An extension to allow carboxylate nucleophiles other than acetate has
further increased the utility of the reaction. See: B. Škermark, E. M.
Larsson, J. D. Oslob, J. Org. Chem. 1994, 59, 5729.
1
intermediates would give an H NMR integration ratio for protons (H1 ‡
H4) to (H3 ‡ H4) (at d ˆ 5.91 (cis) ‡ 5.89 (trans) and 5.32 (trans) ‡ 5.23
(cis), respectively) of 0.5(4 xD):0.5(4 3xD), [i. e. (4 xD)/(4 3xD):1],
whereas the corresponding ratio for a mechanism involving an acetoxy-
palladation ± b-elimination followed by a (p-allyl) intermediate would be
(2 xD):(2 xD), (i. e. 1:1).
[11] P. M. Henry, in Palladium-Catalyzed Oxidation of Hydrocarbons,
Reidel Publishing Co, Dordrecht, 1980, pp. 103.
Isomerization of 1 during allylic acetoxylation: This was observed as a loss
of deuterium at C2 in the product relative to what was expected from either
mechanism. The sum of the integrals of H1, H2, and H3 had a value less
than that expected from the amount of deuterium in the substrate. The
observed integrals is the sum of the contributions from intermediate 5 [A, 2
and 3, IH2 ˆ 1 xD, IH3 ˆ 0.5(2 xD)], and from the (p-allyl) 18 formed
[12] W. Kitching, Z. Rappoport, S. Winstein, W. G. Young, J. Am. Chem.
Soc. 1966, 88, 2054.
[13] P. M. Henry, G. A. Ward, J. Am. Chem. Soc. 1971, 93, 1494.
[14] a) S. Wolfe, P. G. C. Campbell, J. Am. Chem. Soc. 1971, 93, 1497;
b) ibid. 1971, 93, 1499.
[15] For example: a) R. G. Brown, R. V. Chaudhar, J. M. Davidsson, J.
Chem. Soc. Dalton Trans. 1977, 176; b) B. M. Trost, P. J. Metzner, J.
Am. Chem. Soc. 1980, 102, 3572; c) J. E. Bäckvall, K. Zetterberg, B.
Škermark, in Inorganic Reactions and Methods (Ed.: A. P. Hagen),
VCH, Weinheim 1991, Vol. 12A, pp. 123; d) D. R. Chrisope, P. Beak,
W. H. Saunders, J. Am. Chem. Soc. 1988, 110, 230.
[16] a) J. E. Bäckvall, R. E. Nordberg, E. Björkman, C. Moberg, J. Chem.
Soc. Chem. Commun. 1980, 943; b) J. E. Bäckvall, R. E. Nordberg, D.
Wilhelm, J. Am. Chem. Soc. 1985, 107, 6892; c) H. Grennberg, V.
Langer, J. E. Bäckvall, J. Chem. Soc. Chem. Commun. 1991, 1190.
[17] E. N. Frankel, W. K. Rohwedder, W. E. Neff, D. Weisleder, J. Org.
Chem., 1975, 40, 3272.
from the rearranged starting material [B, 10 and 11, IH2 ˆ 1 and IH3
ˆ
0.5(2 xD)]. Also, A ‡ B ˆ 1 (100% product), thus A ˆ (1 IH2, rel)/xD
and IH2, rel ˆ IH2, obs [0.5(2 xD)]/IH3, obs, since IH3 ˆ 0.5(2 xD) for all values
of A and B.
Homoallylic acetoxylation: This was performed as for the allylic acetox-
ylations with protic cyclohexene or 1 as the substrate in acetic acid or in
[D4]acetic acid, but in the presence of CH3SO3H (100 mol%). The reaction
product was, after extractive workup[8] and removal of most of the solvent,
compared with an authentic sample. For 1,
a mechanism involving
acetoxypalladation and a 1,3-hydrogen shift would give a integration ratio
for hydrogens 1:2:(3 ‡ 4) (at d ˆ 5.0, 2.4, and 5.7 ‡ 5.6) of (1 xD):(2
xD):2, (i. e. 0.27:1.27:2 for xD ˆ 0.73). A mechanism involving a (p-allyl)
intermediate would lead to a ratio of [0.5(2 xD)]:(2 xD):[0.5(4 xD)]
(i. e. 0.63:1.27:1.63). The results are in greater accord with the latter
mechanism, however with the integral for H3 ‡ H4 smaller than expected.
This is probably due to allylic acetate rearrangement.[42]
[18] H. Grennberg, V. Simon, J. E. Bäckvall, J. Chem. Soc. Chem.
Commun. 1994, 265.
[19] This is of importance since a (p-allyl) intermediate would yield allylic
acetates from the attack of the acetate nucleophile at C1 and C3,
whereas in acetoxypalladation the acetoxylation occurs at C1 and C2
(see Scheme 2).
Ê
[20] B. Škermark, S. Hansson, T. Rein, J. Vagberg, A. Heumann J. E.
Acknowledgments: Financial support from the Swedish National Science
Research Council is gratefully acknowledged.
Bäckvall, J. Organomet. Chem. 1989, 369, 433.
[21] a) See refs [5 ± 10]; b) H. Grennberg, unpublished results from our
laboratory.
[22] G. Ahlgren, B. Škermark, K. I. Dahlquist, Acta Chem. Scand. 1968,
22, 1129. The concentration of SO3 in commercial concentrated D2SO2
is probably insufficient for complete enolization of the substrate.
[23] The amount of deuterium (xD) in the two vinylic positions was
determined by 1H NMR spectroscopy.
[24] A. Gogoll, H. Grennberg, Magnetic Resonance in Chemistry, 1993, 31,
954.
[25] For a related example of secondary isotope effects in copper chemistry
see: H. L. Goering, V. D. Singleton, J. Am. Chem. Soc. 1976, 98, 7854.
[26] See the Experimental Section.
[27] A 1:1 mixture of allylic acetate (2 and 3) and bis-allylic diacetate (15
and 16) was obtained.
[28] With this oxidation system, numerous additional products were
observed by capillary GLC. No attempt at isolation or identification
of these side-products was made.
[29] B. M. Trost, P. E. Strege, Tetrahedron Lett. 1974, 2603.
[30] Wolfe et al. suggested that a chloro complex reacts by different
reaction pathways than does the corresponding acetato complex. See
ref. [14b].
[31] For example: a) L. E. Overman, F. M. Knoll, Tetrahedron Lett. 1979,
321; b) J. Clayden, E. W. Collington, S. Warren, ibid, 1992, 33, 7039;
c) P. M. Henry, J. Am. Chem. Soc. 1972, 94, 5200; d) L. E. Overman,
November 17, 1997 [F894]
[1] Encyclopedia of Reagents for Organic Synthesis (Ed.: L. A. Paquette),
Wiley, New York 1995.
[2] For example: L. M. Stephenson, M. R. Grdina, M. Orfanopoulos, Acc.
Chem. Res. 1980, 13, 419.
[3] a) M. A. Umbreit, K. B. Sharpless, J. Am. Chem. Soc. 1977, 99, 5526;
b) L. M. Stevenson, D. R Speth, J. Org. Chem. 1979, 44, 4683;
c) K. B.Sharpless, R. F. Lauer, J. Am. Chem. Soc. 1972, 94, 7154;
d) K. B. Sharpless, R. P. Lauer, J. Org. Chem. 1974, 39, 429; e) H. J.
Reich, J. Org. Chem. 1974, 39, 428; f) H. J. Reich, S. Wollowitz, J. E.
Trend, F. Chow, D. F. Wendelborn, ibid. 1978, 43, 1697; g) T. Hori,
K. B. Sharpless, ibid. 1978, 43, 1689; h) L. Engman, ibid. 1989, 54, 889.
[4] a) For a review, see: J. Muzart, Bull. Soc. Chim. Fr. 1986, 65; b) For
recent enantioselective applications see: A. S. Gokhale, A. B. E.
Minidis, A. Pfaltz, Tetrahedron Lett. 1995, 36, 1831; M. N. Andrus,
A. B. Argade, X. Chen, M. G. Pamment, ibid. 1995, 36, 2945; C.
Zondervan, B. L. Feringa, Tetrahedron Asymmetry 1996, 7, 1898; A. D.
Gupta, V. K. Sing, Tetrahedron Lett. 1996, 37, 2633; M. J. Södergren,
P. G. Andersson, ibid. 1996, 37, 7577.
1088
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
0947-6539/98/0406-1088 $ 17.50+.50/0
Chem. Eur. J. 1998, 4, No. 6

Palladium-Catalyzed Oxidations
1083±1089
Angew. Chem. 1984, 96, 565; Angew. Chem. Int. Ed. Engl. 1984, 23,
579.
[32] Wolfe and Campbell reported a slow rearrangement in the presence of
PdII ± HNO3 or stoichiometric PdCl2 (see ref. [14a]).
[33] a) J.-L. Luche, J. Am. Chem. Soc. 1978, 100, 2226; b) A. Hassner, V.
Alexanian, Tetrahedron Lett. 1978, 4475.
genous Catalysis; the Application and Chemistry of Catalysis by
Soluble Transistion Metals. 2nd ed, Wiley Interscience, NY 1992,
pp. 14; b) H. G. Tange, D. G. Sherrington, J. Mol. Catal. 1994,
94, 7.
[38] R. Cramer, R. V. Lindsey, Jr, J. Am. Chem. Soc. 1966, 88, 3534.
[39] Or loss of proton at C1 and C3. The sum of the 1H NMR integrals of
H1, H2, and H3 has a value less than that expected from the amount of
deuterium in the substrate.
[34] For example: J. V. Allen, J. M. J. Williams, Tetrahedron Lett. 1996, 37,
1859.
[35] Also non-palladium-mediated pathways to afford saturated diacetates
could be operating. However, no reaction was observed in the absence
of palladium, with one exception: in the control experiment in the
presence of CuII, 1-acetoxy-2-cyclohexene was transformed into a
large number of compounds. Many of these had the same GLC
retention time as products observed in the presence of CuII and PdII
(see ref. [28]).
[36] J. E. Bäckvall, S. E. Byström, R. E. Nordberg, J. Org. Chem. 1984, 49,
4619.
[37] Such rearrangements with a number of metal catalysts have been
discussed. See for example: a) G. W. Parshall, S. D. Ittel, in Homo-
[40] The results do not suggest the involvement of an acetoxypalladation.
[41] J. B. Lambert, D. E. Marko, J. Am. Chem. Soc. 1985, 107, 7978.
[42] Under these conditions, that is in the presence of strong acid, a control
experiment with 7 showed that some isomerization to 8 by an 1,3-
allylic transposition occurs parallel to the rearrangement of allylic to
homoallylic acetate. Therefore, the labeling experiments under these
conditions can not be used as strong evidence for the (p-allyl)
pathway.
[43] Minor formation of 1,4-diacetate 12 was observed.
[44] F. Bökman, A. Gogoll, L. G. M. Pettersson, O. Bohman, H. O. G.
Siegbahn, Organometallics 1992, 11, 1784.
Chem. Eur. J. 1998, 4, No. 6
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
0947-6539/98/0406-1089 $ 17.50+.25/0
1089