Mechanistic investigations of bipyrimidine-promoted palladium-catalyzed allylic acetoxylation of olefins

Article abstract of DOI:10.1139/V08-133

Several pyridine-like ligands were found to improve Pd(OAc) ₂-catalyzed allylic oxidation of allylbenzene to cinnamyl acetate by p-benzoquinone in acetic acid. The best ligand examined, bipyrimidine, was used to identify the catalyst precursor for this system, (bipyrimidine)Pd(OAc) ₂, which was fully characterized. Mechanistic studies suggest the reaction takes place through disproportionation of (bipyrimidine)Pd(OAc) ₂ to form a bipyrimidine-bridged dimer, which reacts with olefin to form a Pd^II-olefin adduct, followed by allylic C-H activation to produce (η³-allyl)Pd^II species. The (η³-allyl)Pd^II intermediate undergoes a reversible acetate attack to generate a Pd⁰-(allyl acetate) adduct, which subsequently reacts with p-benzoquinone to release allyl acetate and regenerate (bipyrimidine)Pd(OAc)₂. No KIE is observed for the competition experiment between allylbenzene-d₀ and allylbenzene-d₅ (CD₂=CDCD₂C₆H₅), suggesting that allylic C-H activation is not rate-determining. Catalytic allylic acetoxylations of other terminal olefins as well as cyclohexene were also effected by (bipyrimidine)Pd(OAc)₂.

Full text of DOI:10.1139/V08-133

264
Mechanistic investigations of bipyrimidine-
promoted palladium-catalyzed allylic acetoxylation
of olefins1
Bo-Lin Lin, Jay A. Labinger, and John E. Bercaw
Abstract: Several pyridine-like ligands were found to improve Pd(OAc)2-catalyzed allylic oxidation of allylbenzene to
cinnamyl acetate by p-benzoquinone in acetic acid. The best ligand examined, bipyrimidine, was used to identify the
catalyst precursor for this system, (bipyrimidine)Pd(OAc)2, which was fully characterized. Mechanistic studies suggest
the reaction takes place through disproportionation of (bipyrimidine)Pd(OAc)2 to form a bipyrimidine-bridged dimer,
3
which reacts with olefin to form a PdII-olefin adduct, followed by allylic C–H activation to produce (η -allyl)PdII spe-
3
cies. The (η -allyl)PdII intermediate undergoes a reversible acetate attack to generate a Pd0-(allyl acetate) adduct, which
subsequently reacts with p-benzoquinone to release allyl acetate and regenerate (bipyrimidine)Pd(OAc)2. No KIE is ob-
served for the competition experiment between allylbenzene-d0 and allylbenzene-d5 (CD2=CDCD2C6H5), suggesting that
allylic C–H activation is not rate-determining. Catalytic allylic acetoxylations of other terminal olefins as well as
cyclohexene were also effected by (bipyrimidine)Pd(OAc)2.
Key words: olefin, palladium catalysis, allylic C–H oxidation, p-benzoquinone, bipyrimidine.
Résumé : On a observé que plusieurs ligands apparentés à la pyridine améliorent l’oxydation allylique de
l’allylbenzène en acétate de cinnamyle par la p-benzoquinone dans l’acide acétique et catalysée par le Pd(OAc)2. On a
utilisé le meilleur ligand étudié, la bipyrimidine, pour identifier le précurseur du catalyseur pour ce système, le (bipyri-
midine)Pd(OAc)2, qui a été complètement caractérisé. Des études mécanistiques suggèrent que la réaction se produit
par le biais d’une réaction de dismutation du (bipyrimidine)Pd(OAc)2 qui conduit à la formation d’un dimère à pont bi-
pyrimidine qui réagit avec l’oléfine pour conduire à un adduit Pd(II)-oléfine, le tout suivi de l’activation du C–H ally-
3
3
lique conduisant à la formation de l’espèce (η -allyl)Pd(II). L’intermédiaire (η -allyl)Pd(II) subit une attaque réversible
par l’ion acétate qui génère un adduit Pd°-(acétate d’allyle) qui réagit subséquemment avec la p-benzoquinone pour li-
bérer l’acétate d’allyle et régénérer le (bipyrimidine)Pd(OAc)2. On n’observe aucun effet isotopique cinétique pour
l’expérience de compétition entre l’allylbenzène-d0 et l’allylbenzène-d5 (CD2=CDCD2C6H5); ce résultat suggère que
l’activation allylique n’est pas l’étape cinétiquement déterminante. Les acétoxylations allyliques catalytiques du cyclo-
hexène et d’autres oléfines terminales sont aussi affectées par le (bipyrimidine)Pd(OAc)2.
Mots-clés : oléfine, catalyseur de palladium, oxydation de C–H allylique, p-benzoquinone, bipyrimidine.
[Traduit par la Rédaction] Lin et al. 271
Introduction
yields, and poor selectivities (2), extensive work on systems
based on Pd(II) has led to some remarkably selective allylic
C–H activations and oxidations (3, 4). The solvent often
plays a key role in achieving selective oxidation to allyl ace-
tates: in neat acetic acid, only symmetrical cyclic olefins are
selectively oxidized with Pd(OAc)2 (5), whereas terminal
olefins can be selectively oxidized in DMSO/acetic acid (6),
and yields can be further improved in dimethylacetamide
(DMA) (7). The combination of Pd(II) with sulfoxide lig-
ands also selectively oxidizes terminal olefins (8); this reac-
tion has been successfully applied for catalytic allylic
aminations (9) and macrolactonizations (10).
The selective oxidation of allylic C–H bonds offers a
valuable approach to the construction of complex molecules
(1). While earlier catalysts based on copper or selenium suf-
fer from problems, such as limited substrate scope, low
Received 17 April 2008. Accepted 2 July 2008. Published on
the NRC Research Press Web site at canjchem.nrc.ca on
15 October 2008.
This paper is dedicated to Professor Richard Puddephatt for
his great contributions in Organometallic Chemistry.
B.-L. Lin, J.A. Labinger,2 and J.E. Bercaw.2 Arnold and
Mabel Beckman Laboratories of Chemical Synthesis,
California Institute of Technology, Pasadena, CA 91125,
USA.
To date, no catalytic system has been reported to selec-
tively oxidize both terminal and internal olefins to allyl ace-
tates. Furthermore, in no case has the active catalytic species
been characterized, which complicates the study and under-
standing of reaction mechanisms and impedes further devel-
opment of new catalysts. Herein, we disclose a well-defined
new catalyst precursor, (bipyrimidine)Pd(OAc)2, that effects
selective allylic acetoxylations of terminal olefins and
1This article is part of a Special Issue dedicated to Professor
R. Puddephatt.
2Corresponding authors (e-mail: jal@caltech.edu,
bercaw@caltech.edu).
Can. J. Chem. 87: 264–271 (2009)
doi:10.1139/V08-133
© 2008 NRC Canada

Lin et al.
265
1
cyclohexene in acetic acid, and offer evidence for its mecha-
nism of operation.
yield: 76%; NMR yield of the trans isomer: 72%. H NMR
3
(CDCl3) δ: 7.20–7.45 (m, 5H), 6.60–6.71 (d, 1H, JHH
15.9 Hz), 6.24–6.35 (dt, 1H, JHH = 15.9 Hz, 6.6 Hz), 4.71–
4.76 (dd, 2H, JHH = 6.6 Hz, JHH = 1.3 Hz), 2.11 (s, 3H).
13C NMR (CDCl3) δ: 171.1, 136.4, 134.4, 128.9, 128.3,
126.9, 123.4, 65.3, 21.3. HR-MS (FAB+) m/z calcd. for
C11H12O2: 176.0837; found: 176.0832.
=
3
3
4
Experimental
All starting materials are commercially available and used
3
as received without any further purification. [(η -1-
3
phenylallyl)PdCl]2 (11) and [(η -1-phenylallyl)Pd(OAc)]2
(12) were synthesized according to literature procedures.
Allylbenzene-d5 (C6H5CD2CD=CD2) was synthesized by a
modified literature procedure (13) where allyl bromide-d5
was used instead of allyl bromide-d0. Unless otherwise
trans-p-Methylcinnamyl acetate (Table 2, entry 2)
p-Allyltoluene: 99% pure. Reaction time: 32 h 46 min.
Eluent: 1:10 (v/v) EtOAc/hexanes. The product was isolated
as a mixture of cis and trans isomers. Mass: 165 mg; overall
1
1
noted, all manipulations were carried out in air. H and
yield: 78%; NMR yield of the trans isomer: 75%. H NMR
13C NMR spectra were recorded at room temperature on
Varian Mercury 300 (1H, 299.8 MHz; 13C, 125.9 MHz) spec-
trometers. NMR spectra were referenced to residual solvent
3
(CDCl3) δ: 7.26–7.32 (d, 2H, JHH = 8.1 Hz), 7.10–7.17 (d,
2H, JHH = 8.1 Hz), 6.58–6.68 (d, 1H, JHH = 15.9 Hz),
6.18–6.30 (dt, 1H, JHH = 15.9 Hz, 6.6 Hz), 4.68–4.76 (dd,
2H, JHH = 6.6 Hz, JHH = 1.3 Hz), 2.35 (s, 3H), 2.10 (s,
3H). 13C NMR (CDCl3) δ: 171.1, 138.2, 134.5, 133.6, 129.6,
126.8, 122.3, 65.5, 21.5, 21.3. HR-MS (FAB+) m/z calcd.
for C12H14O2: 190.0994; found: 190.0995.
3
3
3
1
peak (CDCl3: H, 7.26 ppm; 13C, 77.23 ppm; CD3COOD:
3
4
1H, 2.04 ppm) and reported in parts per million (ppm). Mul-
tiplicities were reported as follows: singlet (s), doublet (d),
doublet of doublets (dd), doublet of triplets (dt), triplet (t),
quartet (q), multiplet (m), and broad resonance (br). GC
measurements were taken on an Agilent 6890 series GC in-
strument with an Agilent HP-5 column. The X-ray structure
was obtained at Caltech X-ray Crystallography Laboratory.
High-resolution mass spectra were obtained at Caltech Mass
Spectrometry Laboratory. Reaction yields are corrected for
the purities of the starting materials.
trans-p-Methoxycinnamyl acetate (Table 2, entry 3)
p-Methoxyallylbenzene: 98% pure. Reaction time: 16 h.
Eluent: 1.5:10 (v/v) EtOAc/hexanes. The product was iso-
lated as a mixture of cis and trans isomers. Mass: 118 mg;
overall yield: 53%; NMR yield of the trans isomer: 50%. H
NMR (CDCl3) δ: 7.30–7.36 (d, 2H, JHH = 8.4 Hz), 6.82–
6.89 (d, 2H, JHH = 8.4 Hz), 6.55–6.64 (d, 1H, JHH
1
3
3
3
=
3
15.3 Hz), 6.10–6.21 (dt, 1H, JHH = 15.3 Hz, 6.6 Hz), 4.66–
(Bipyrimidine)Pd(OAc)2 (1)
3
4
4.74 (dd, 2H, JHH = 6.6 Hz, JHH = 1.3 Hz), 3.80 (s, 3H),
2.09 (s, 3H). 13C NMR (CDCl3) δ: 171.2, 159.8, 134.3,
129.2, 128.1, 121.0, 114.2, 65.6, 55.5, 21.3. HR-MS (FAB+)
m/z calcd. for C12H14O3: 206.0943; found: 206.0952.
A solution of 25 mg Pd(OAc)2 and 17.8 mg bipyrimidine
(BPM) in 4.5 mL glacial acetic acid was stirred for 30 min,
resulting in a clear yellow solution. After removing the sol-
vent under vacuum, (BPM)Pd(OAc)2 (1) was obtained as a
yellow powder, pure enough for NMR analyses. A yellow
crystal of [(BPM)Pd(OAc)2·H2O] was obtained by slow
evaporation of a chloroform solution of a 1:1:1 mixture of
BPM, Pd(OAc)2, and acetic acid under air. 1H NMR
1-Acetoxy-trans-2-decene (Table 2, entry 5)
Reaction time: 20 h. Eluent: 1:10 (v/v) EtOAc/hexanes.
The product was isolated as a mixture of allylic acetates
1
3
4
along with (1-acetoxy-)3-, 4-, and 5-decenes. Both H NMR
(CDCl3) δ: 9.20 (dd, 2H, JHH = 4.8 Hz, JHH = 2.2 Hz),
3
4
and GC gave the estimated ratio of allylic acetates (sum of
trans-primary allylic acetate, cis-primary allylic acetate, and
secondary allylic acetate) to other isomers to be ~3.8. Only
1-acetoxy-trans-2-decene was fully characterized. Mass:
171 mg; overall yield: 78%; NMR yield of 1-acetoxy-trans-
8.68 (dd, 2H, JHH = 5.7 Hz, JHH = 2.2 Hz), 7.73 (dd, 2H,
3JHH = 5.7 Hz, 4.8 Hz), 2.13 (s, 6H). 13C NMR (CDCl3) δ:
178.9, 160.5, 157.6, 124.2, 23.3. HR-MS (FAB+) m/z calcd.
for C10H9N4O2Pd+: (M – OAc)+ 322.9796; found: 322.9774.3
1
2-decene: 54%. H NMR (CDCl3) δ: 5.70–5.83 (m, 1H),
General procedure for allylic oxidation (Table 2, entries
1, 2, 3, 5, and 6)
3
5.49–5.62 (m, 1H), 4.47–4.53 (dd, 2H, JHH = 6.6 Hz,
4JHH = 1 Hz), 2.06 (s, 3H), 1.92–2.08 (m, 2H), 1.27 (br s,
A solution of 25 mg Pd(OAc)2 and 17.6 mg BPM in
4.5 mL glacial acetic acid was stirred for ~30 min, followed
by sequential additions of 10 equiv. of olefin and 20 equiv.
of benzoquinone. The mixture was heated at 70 °C with stir-
ring, in a 20 mL glass vial sealed with a Teflon-lined cap
and PTFE tape. After the reported time, the reaction mixture
was cooled to room temperature in a water bath and neutral-
ized with 50 mL of an aqueous solution of 6.2 g NaOH. The
resultant mixture was extracted with 3 × 100 mL CH2Cl2.
Organic portions were combined and dried over MgSO4.
Volatiles were removed under vacuum. The resultant oil was
further purified by flash silica-gel column chromatography.
10H), 0.87 (t, 3H, JHH = 6.9 Hz). 13C NMR (CDCl3) δ:
3
171.1, 137.0, 123.9, 65.6, 32.5, 32.0, 29.6, 29.3, 29.1, 22.9,
21.3, 14.3. HR-MS (FAB+) m/z calcd. for C12H14O2:
198.1620; found: 198.1625.
1-Acetoxy-3-cyclohexyl-trans-2-propene (Table 2, entry 6)
Allylcyclohexane: 96% pure. Reaction time: 20 h. Eluent:
1:10 (v/v) EtOAc/hexanes. The product was isolated as a
mixture of allylic acetates, 3-cyclohexylidenepropyl acetate,
and 3-cyclohexenylpropyl acetates. The ratio of allylic ace-
tates (sum of trans-primary allylic acetate, cis-primary al-
lylic acetate, and secondary allylic acetate) to other isomers
1
was estimated to be ~6.2 by H NMR and GC. Only 1-
trans-Cinnamyl acetate (Table 2, entry 1)
Allylbenzene: 98% pure. Reaction time: 34 h. Eluent:
1:10 (v/v) EtOAc/hexanes. The product was isolated as a
mixture of cis and trans isomers. Mass: 146 mg; overall
acetoxy-3-cyclohexyl-trans-2-propene was fully character-
ized. Mass: 150 mg; overall yield: 77%; NMR yield of 1-
acetoxy-3-cyclohexyl-trans-2-propene: 65%. 1H NMR
© 2008 NRC Canada

266
Can. J. Chem. Vol. 87, 2009
Table 2. (Bipyrimidine)Pd(OAc)2-catalyzed allylic acetoxylation
Table 1. Nitrogen ligand effects on Pd(OAc)2-catalyzed allylic
of olefins.
acetoxylation of allylbenzene.
Entry
Catalyst
Yielda,b
1
2
10 mol% Pd(OAc)2
13%
76%
10 mol% Pd(OAc)2 +
10 mol% 2,2′-bipyrimidine
10 mol% Pd(OAc)2 +
20 mol% pyridine
10 mol% Pd(OAc)2 +
10 mol% 1,10-phenanthroline
10 mol% Pd(OAc)2 +
10 mol% 2,2′-bipyridine
3
4
5
35%
47%
c
—
Note: General experimental procedure: A mixture of catalyst, 0.7 mL
glacial acetic acid, allylbenzene, 1 equiv. BQ, and 20 µL nitrobenzene (in-
ternal standard for GC) is stirred and heated in a sealed vial at 80 °C for
16 h.
Note: Entry 1–6: 10 mol% catalyst loading and 2 equiv. BQ; Entry 7: 5
mol% catalyst loading and 1.2 equiv. BQ; All reactions except 4 and 7
are near completion.
aSum of cis and trans isomers; GC yields.
bConversions of allylbenzene: entry 1–3, 100%; entry 4, 87%; entry 5, 3%.
cNo product detected.
aProduct isolated as a mixture of isomers: Entries 1–3: cis- + trans-1°
allyl acetates; Entries 5–7: C=C bond migration results in minor products
such as homoallyl acetates; Entries 5 and 6: small amounts of 2° allyl ac-
etates observed.
3
(CDCl3) δ: 5.66–5.75 (dd, 1H, JHH ≈ 15.6 Hz, 6.3 Hz),
3
4
5.44–5.55 (dtd, 1H, JHH = 15.6 Hz, 6.6 Hz, JHH = 1.1 Hz),
3
4.46–4.52 (d, 2H, JHH = 6.6 Hz), 2.05 (s, 3H), 1.84–2.05
Fig. 1. X-ray crystallographic structure and the crystal packing
of [(bipyrimidine)Pd(OAc)2·H2O] (see Supplementary data).3
(m, 1H), 0.94–1.79 (m, 10H). 13C NMR (CDCl3) δ: 171.1,
142.4, 121.5, 65.8, 40.6, 32.7, 26.3, 26.2, 21.3. HR-MS
(FAB+) m/z calcd. for C11H18O2: 182.1307; found:
182.1313.
Allyl acetate (Table 2, entry 4)
A 5 mL round-bottom flask was charged with 27.1 mg
(BPM)Pd(OAc)2 (10 mol%), 153 mg BQ (2 equiv.), 20 µL
nitrobenzene, and 2.2 mL glacial acetic acid. Propylene
(15.7 mL, 1 atm) was then condensed into the flask by liquid
nitrogen after degassing (1 atm = 101.325 kPa). The mixture
was stirred and heated at 70 °C for 24 h. An aliquot of the
reaction mixture was taken and filtered through a short pi-
pette of silica gel with Et2O as the eluent for GC analysis.
The GC-peak area of allyl acetate relative to nitrobenzene
was used to calculate the yield. Allyl acetate was the only
1
allylic oxidation product observed by H NMR. GC Yield:
45%.
Scheme 1.
Cyclohexenyl acetates (Table 2, entry 7)
The general procedure above was followed, except for the
use of 20 equiv. of cyclohexene and 24 equiv. of benzo-
quinone, and heating at 80 °C. After 24 h, the mixture was
cooled to room temperature by water bath and neutralized by
6.2 g NaOH in 50 mL H2O. The resultant mixture was ex-
tracted by 3 × 100 mL CH2Cl2. Organic portions were com-
bined and dried over MgSO4. Volatiles were removed under
vacuum. The resultant oil was further purified by reduced-
pressure distillation. A ~2.8:1 mixture of cyclohex-2-enyl
acetate and cyclohex-3-enyl acetate was obtained. Mass:
160 mg; overall yield: 51.3%.
21.6, 19.1. HR-MS (FAB+) m/z calcd. for C8H12O2:
140.0837; found: 140.0841.
Cyclohex-2-enyl acetate
1H NMR (CDCl3) δ: 5.87–5.98 (m, 1H), 5.63–5.73 (m,
1H), 5.18–5.27 (m, 1H), 2.03 (s, 3H), 1.53–2.45 (m, 6H).
13C NMR (CDCl3) δ: 171.0, 132.9, 125.9, 68.3, 28.5, 25.1,
Cyclohex-3-enyl acetate
1H NMR (CDCl3) δ: 5.62–5.67 (m, 1H), 5.51–5.60 (m,
1H), 4.93–5.03 (m, 1H), 2.03 (s, 3H), 1.53–2.45 (m, 6H).
© 2008 NRC Canada

Lin et al.
267
3
CD3COOD. 1H NMR shows that [(BPM)Pd(η -1-
Table 3. Crystal data and structure refinement for
[(BPM)Pd(OAc)2·H2O]
phenylallyl)]+(OAc)– was formed quantitatively. Unfortu-
nately, compound 2 is not isolable owing to rapid decompo-
sition to Pd metal in the absence of acetic acid. H NMR
Empirical formula
C12H12N4O4Pd·H2O
1
Formula mass
Crystallization solvent
Crystal habit
Crystal size (mm)
Crystal colour
λ Mo Kα (Å)
Temperature (K)
a (Å)
400.67
Dichloromethane
Block
0.36 × 0.31 × 0.27
Yellow
0.710 73
(CD3COOD) δ: 9.24 (br s 4H), 7.78 (br s 2H), 7.46–7.82
3
(m, 5H), 6.59 (app dt, 1H, JHH = 12.3 Hz, 7.2 Hz), 5.06 (d,
3
3
1H, JHH = 12.3 Hz), 4.68 (d, 1H, JHH = 7.2 Hz), 3.87(d,
3
1H, JHH = 12.3 Hz). HR-MS (FAB+) m/z calcd. for
C17H15N4Pd+: (M – OAc)+ 381.0332; found: 381.0343.
100(2)
(Bipyrimidine)[Pd(OAc)2]2 (3)
9.6974(3)
14.9781(4)
10.0106(3)
98.7400(10)
1437.14(7)
4
3.6 mg BPM and 10 mg Pd(OAc)2 (2 equiv.) were mixed
with 0.7 mL CD3COOD. A clear yellow solution was ob-
tained after stirring for 2 h. The solution contained an equili-
brated mixture of Pd(OAc)2, (BPM)Pd(OAc)2 (1), and
(BPM)[Pd(OAc)2]2 (3). The equilibrium constant was esti-
b (Å)
c (Å)
β (°)
V (Å3)
1
mated to be ~25 mol–1 L by H NMR. NMR yield of com-
Z
1
Crystal system
Space group
Dcalcd. (Mg/m3)
Monoclinic
P21/n
pound 3 based on BPM: 35%. H NMR (CD3COOD) δ:
3
8.75 (d, 4H, 3JHH = 5.7 Hz), 7.78 (t, 2H, JHH = 5.7 Hz). MS
+
(MALDI) m/z calcd. for C16H18N4Pd2 : (M+) 605.919;
1.852
found: 605.7228.
F(000)
800
θ range for data collection
Completeness to θ = 44.89°
Index ranges
2.47° to 44.89°
77.7%
–18 ≤ h ≤ 17,
–29 ≤ k ≤ 29,
–17 ≤ l ≤ 17
28 065
[(Bipyrimidine)Pd(␩3-2-propenyl)]+(OAc)–
To a solution of compound 1 in 0.7 mL CD3COOD was
added 2.4 µL allyl acetate (1 equiv.). After degassing, the so-
lution was heated at 80 °C for 15 h. According to the H
1
NMR spectrum, ~50% of the starting material was converted
Reflections collected
Independent refl.
Absorption coefficient (mm–1)
Absorption correction
Max. and min. transmission
Hydrogen placement
Data, restraints, parameters
Treatment of hydrogen atoms
Goodness-of-fit on F2
3
1
to [(BPM)Pd(η -2-propenyl)]+. H NMR (CD3COOD) δ:
9164 [Rint = 0.0588]
3
9.32 (br s 4H), 7.99 (br s 2H), 6.03–6.19 (tt, 1H, JHH = 6.6
1.321
None
2
3
Hz, 12.6 Hz), 4.72–4.76 (dd, 2H, JHH = 1.3 Hz, JHH = 6.6
Hz), 3.60–3.72 (d, 2H, 3JHH = 12.6 Hz). HR-MS (FAB+) m/z
calcd. for C11H11N4Pd+: (M – OAc)+ 305.0019; found: 305.0044.
0.7168 and 0.6477
Difference Fourier map
9164, 0, 255
Unrestrained
1.485
Exchange between trans-cinnamyl acetate and
CD3COOD
A mixture of compound 2 and 10 µL trans-cinnamyl ace-
tate (3 equiv.) in 0.7 mL CD3COOD was heated in a
J–Young NMR tube under nitrogen at 80 °C. After 16 h
5 min, 20% of the starting PhCH=CHCO2CH3 was con-
verted to PhCH=CHCO2CD3. No change of compound 2
was detected. Control experiments were carried out under
similar conditions except that compound 2 was absent or re-
placed by an equimolar amount of NaOAc. In neither case
Final R indices [I > 2σ(I), 7498 refl.]
R1 = 0.0305,
wR2 = 0.0664
R indices (all data)
R1 = 0.0393,
wR2 = 0.0678
Max shift/error
Average shift/error
Largest diff. peak and hole (e Å–3)
0.004
0.000
1.503 and –1.576
1
was acetate scrambling detected by H NMR.
13C NMR (CDCl3) δ: 171.0, 127.0, 123.9, 70.0, 30.9, 27.5,
Competition between allylbenzene-d0 and allylbenzene-d5
A mixture of 4.3 mg (BPM)Pd(OAc)2 (~5 mol%), 10 µL
nitrobenzene (internal standard), 15 µL allylbenzene-d0,
15 µL allylbenzene-d5, 25 mg BQ, and 0.35 mL CH3COOH
was heated in a sealed vial with stirring at 80 °C. Aliquots
(~5 µL) were taken at certain times by microsyringe for
GC analysis. The GC signals of allylbenzene-d0 and
allylbenzene-d5 are fully separated, but those of the two
isotopologs of trans-cinnamyl acetate are not.
23.5, 21.6.
[(Bipyrimidine)Pd(␩3-1-phenylallyl)]+(OAc)– (2)
Route 1
A mixture of 8.6 mg (BPM)Pd(OAc)2 and 29.5 µL allyl-
benzene in 0.7 mL CD3COOD was heated in a J–Young
NMR tube under nitrogen at 80 °C for 96 h. The yield of
compound 2 based on Pd was 89%, as estimated by 1H NMR.
Route 2
Results and discussion
3
A 2:1 ratio of AgOAc (3.8 mg) and [(η -1-phenyl-
allyl)PdCl]2 (5.5 mg) was stirred in 1 mL of acetone at room
temperature for 30 min. After filtering off the precipitate,
solvent was pumped away under vacuum. To the resultant
yellow solid was added 3.6 mg BPM and 0.7 mL
The oxidation of allylbenzene by p-benzoquinone (BQ) in
acetic acid at 80 °C, with Pd(OAc)2 as the catalyst, leads to
only a 13% yield of cinnamyl acetate (100% consumption of
allylbenzene). Several other unidentified products, probably
© 2008 NRC Canada

268
Can. J. Chem. Vol. 87, 2009
Scheme 2.
including the same Wacker-type products previously re-
ported from reaction at 40 °C (6b), are observed. The selec-
tivity for cinnamyl acetate is significantly improved by the
addition of pyridine-like nitrogen ligands (with the excep-
tion of 2,2′-bipyridine, which shuts down the reaction) (Ta-
ble 1). Among the ligands tested, bipyrimidine (BPM) is the
best, leading to a 76% yield of cinnamyl acetate in 16 h at
80 °C.
Scheme 3.
Studies of substrate scope (Table 2) indicate that both
benzylic C–H and methoxy groups are compatible with the
acetoxylation (Table 2, entries 2 and 3). Unfortunately, the
yield for olefins with alkyl groups at the allylic position
(Table 2, entries 5 and 6) suffers from olefin isomerization,
giving byproducts such as homoallylic acetates. Cyclo-
hexene is also an active substrate for the reaction, producing
a mixture of cyclohex-2-enyl acetate and cyclohex-3-enyl
acetate (~ 2.8:1) (Table 2, entry 7). Interestingly, cyclohex-
2-enyl acetate becomes the only allylic oxidation product if
(1,10-phenanthroline)Pd(OAc)2 or bis(oxazoline)Pd(OAc)2
(14) is used as the catalyst.
Scheme 4.
The reaction of bipyrimidine with Pd(OAc)2 in
CD3COOD forms (BPM)Pd(OAc)2 (1) in quantitative yield
after 30 min at room temperature, as monitored by 1H NMR.
A crystal structure for (BPM)Pd(OAc)2·H2O was obtained
(Fig. 1; Table 3).3 Heating (BPM)Pd(OAc)2 with allyl-
benzene (10 equiv.) in CD3COOD at 80 °C for ~4 day af-
3
fords [(BPM)Pd(η -1-phenylallyl)]+(OAc)– (2), the product
3 Supplementary data for this article are available on the journal Web site (canjchem.nrc.ca) or may be purchased from the Depository of Un-
published Data, Document Delivery, CISTI, National Research Council Canada, Ottawa, ON K1A 0R6, Canada. DUD 3834. For more in-
formation on obtaining material refer to cisti-icist.nrc-cnrc.gc.ca/cms/unpub_e.shtml. CCDC 634014 contains the crystallographic data for
this manuscript. These data can be obtained, free of charge, via www.ccdc.cam.ac.uk/conts/retrieving.html (Or from the Cambridge Crystal-
lographic Data Centre, 12 Union Road, Cambridge CB2 1EZ, UK; fax +44 1223 336033; or deposit@ccdc.cam.ac.uk).
© 2008 NRC Canada

Lin et al.
269
3
Fig. 2. Effect of BPM to Pd(OAc)2 ratio on allylic oxidation of
allylbenzene by benzoquinone, Pd(OAc)2 fixed to be 5 mol%.
Upper graph: initial rates of the disappearance of allylbenzene
relative to that of n(BPM)/n(Pd) = 1.1; lower graph: appearance
of cinnamyl acetate.
Fig. 3. Oxidation of (η -allyl)PdII by benzoquinone is independ-
ent of the concentration of benzoquinone.
Scheme 5.
Scheme 6.
Scheme 7.
of allylic C–H activation, in ~89% yield (Scheme 1). Fur-
thermore, the allylic C–H activation is irreversible, because
neither olefin exchange nor deuterium incorporation into the
allyl fragment is observed after heating 2 with p-allyltoluene
in CD3COOD at 80 °C for 15 h (Scheme 1). Similar irre-
versibility was also proposed for the sulfoxide–Pd(OAc)2
system (8).
The reaction between allylbenzene and (BPM)Pd(OAc)2
could occur via the displacement by the olefin of either an
acetate or a coordinated BPM nitrogen to form a Pd(II)–
olefin adduct, followed by allylic C–H activation to form 2
(Scheme 2, mechanisms 1 and 2). For either mechanism, ad-
dition of excess BPM should not have a significant effect on
the reaction rate (a mixture of Pd(OAc)2 and 1.15 equiv.
BPM in CD3COOD affords 1 equiv. (BPM)Pd(OAc)2 and
0.15 equiv. free BPM (NMR)). However, no reaction is seen
for allylbenzene or 1-decene when a slight excess (≤20 mol%
relative to Pd(OAc)2) of BPM is added (Scheme 3), clearly
ruling out either mechanism for the acetoxylation.
Full dissociation of BPM from (BPM)Pd(OAc)2, forming
Pd(OAc)2, which then catalyzes the reaction, would be con-
sistent with the inhibitory effect of excess BPM (Scheme 2,
© 2008 NRC Canada

270
Can. J. Chem. Vol. 87, 2009
Scheme 8.
mechanism 3), but this is also unlikely to be the mechanism
because Pd(OAc)2 does not lead to selective allylic oxida-
tion under current catalytic conditions (Table 1, entry 1).
The nonselective oxidation catalyzed by Pd(OAc)2 must
originate in the allylic C–H activation of allylbenzene by
Pd(OAc)2 rather than a subsequent step, because the
Fig. 4. No significant KIE is observed for competition reaction
between allylbenzene-d0 and allylbenzene-d5.
3
solvolysis of [(η -1-phenylallyl)Pd(OAc)]2 in wet acetic acid
at 75 °C quantitatively forms allylic oxidation products (al-
lylic acetates and cinnamaldehyde; the latter results from
water as the nucleophile) (15).
In a further experiment, a mixture of (BPM)Pd(OAc)2 and
Pd(OAc)2 in CD3COOD (generated in situ by combining
Pd(OAc)2 and BPM in a 2:1 ratio) reacts to form a new
species, characterized by 1H NMR and MALDI as the BPM-
bridged Pd(OAc)2 dimer 3 (Scheme 4). The reaction does
not reach 100% conversion, but equilibrates with Keq
~
1
25 mol–1 L, as estimated by H NMR. The resultant mixture
reacts with allylbenzene under air at room temperature to
3
form an η -allyl species, cinnamyl acetate, and palladium
black.
We propose that dimer 3 is the active species for the reac-
tion with allylbenzene, because the Pd(OAc)2-only system
results in less selective oxidation of allylbenzene, while
heating is required for the reaction between (BPM)Pd(OAc)2
and allylbenzene. According to our proposal, in the latter re-
action, small amounts of dimer and free BPM are formed via
disproportionation of (BPM)Pd(OAc)2, explaining the inhi-
bition by excess BPM. The coordination of a second equiva-
lent of Pd(OAc)2 to BPM would be expected to weaken a
Pd–N bond and thus facilitate the coordination of olefin; the
Pd–olefin adduct subsequently undergoes allylic C–H activa-
tion to form 2 (Scheme 2, mechanism 4). This mechanism is
further supported by the observation that the initial con-
sumption rate of allylbenzene increases as the ratio of BPM
to Pd(OAc)2 decreases; the best overall yields (~82%) were
obtained by combining 3–4 mol% BPM with 5 mol%
Pd(OAc)2 (Fig. 2).
Although 2 is stable in CD3COOD at 80 °C, the addition
of benzoquinone (BQ) releases cinnamyl acetate (16) and
generates (BPM)Pd(OAc)2 and hydroquinone (HQ). It has
been proposed that coordination of benzoquinone to PdII in-
duces acetate attack (8, 17), but we find that the rate of the
reaction between 2 and benzoquinone in acetic acid is inde-
pendent of the benzoquinone concentration when at least
5.8 equiv. are used (Fig. 3), which argues against this mech-
anism.
Instead, we believe that reversible attack of acetate at the
η -allyl moiety (18) occurs in the absence of benzoquinone
3
to form the (unobservable) Pd0(allyl acetate) species 4
(Scheme 5). Such a route is suggested by the reaction be-
tween allyl acetate and 2 in CD3COOD, which generates
© 2008 NRC Canada

Lin et al.
271
3
cinnamyl acetate and [(BPM)Pd(η -propenyl)]+(OAc)–
(Scheme 6), and by the fact that 2 catalyzes exchange of la-
beled and unlabeled acetate between trans-cinnamyl acetate
and CD3COOD (Scheme 7). (BPM)Pd0(cinnamyl acetate)
complex 4 then reacts with BQ to release cinnamyl acetate
and form (BPM)Pd0(BQ) (19), followed by a redox reaction
with acetic acid to generate (BPM)Pd(OAc)2 and hydro-
quinone (HQ) to complete the catalytic cycle (Scheme 5).
Scheme 8 shows a mechanism consistent with all current
observations. Allylic C–H activation of allybenzene by
References
1. For examples, see the following review: J. Eames and M.
Watkinson. Angew. Chem., Int. Ed. 40, 3567 (2001).
2. (a) Copper: M.S. Kharasch and G. Sosnovsky. J. Am. Chem.
Soc. 80, 756 (1958); (b) Selenium: M.A. Umbreit and K.B.
Sharpless. J. Am. Chem. Soc. 99, 5526 (1977).
3. (a) Selective stoichiometric activation of allylic C–H bonds of
olefins to form (η -allyl)PdII: B.M. Trost and P.J. Metzner. J.
3
Am. Chem. Soc. 102, 3572 (1980); (b) DMSO-promoted se-
lective stoichiometric oxidation of 1-butene to but-2-enyl ace-
tate: W. Kitching, Z. Rappoport, S. Winstein, and W.G. Young.
J. Am. Chem. Soc. 88, 2054 (1966).
dimer 3 leads to a dimeric (η -allyl)PdII complex, which can
3
react either with benzoquinone to release cinnamyl acetate
and regenerate the active dimer 3 (Scheme 8, cycle 1), or
with free BPM (generated from the disproportionation) to
4. Pd-catalyzed selective oxidations of α,β-enones (or ketals) to
1,4-enediones: J.-Q. Yu and E.J. Corey. J. Am. Chem. Soc.
125, 3232 (2003).
5. (a) J.E. McMurry and P. Kocovsky. Tetrahedron Lett. 25, 4187
(1984); (b) S. Hansson, A. Heumann, T. Rein, and B.
Akermark. J. Org. Chem. 55, 975 (1990); (c) C.G. Jia, P.
Müller, and H. Mimoun. J. Mol. Catal. A: Chem. 101, 127
(1995).
6. (a) R.C. Larock and T.R. Hightower. J. Org. Chem. 58, 5298
(1993); (b) M.S. Chen and M.C. White. J. Am. Chem. Soc.
126, 1346 (2004).
7. (a) D. Clark, P. Hayden, and R.D. Smith. Discussions Faraday
Soc. 46, 98 (1968); (b) T. Mitsudome, T. Umetani, N. Nosaka,
K. Mori, T. Mizugaki, K. Ebitani, and K. Kaneda. Angew.
Chem., Int. Ed. 45, 481 (2006).
form the monomeric (η -allyl)PdII 2, followed by reaction
3
with benzoquinone to release cinnamyl acetate and regener-
ate (BPM)Pd(OAc)2 (Scheme 8, cycle 2). We assume that
the mechanistic inferences from the stoichiometric reactions
described above, which would be rigorously applicable only
to the monomeric intermediates in cycle 2, most probably
would be valid for the analogous dimeric intermediates of
cycle 1 as well.
If cycle 2 is the dominant pathway, the rate-determining
step for the whole catalytic cycle must occur before the for-
mation of (η -allyl)PdII because the only Pd species ob-
3
1
served by H NMR during the reaction is (BPM)Pd(OAc)2.
In addition, the BPM inhibitory effect suggests a pre-
equilibrium between (BPM)Pd(OAc)2 and 3. Therefore, ei-
ther the olefin coordination or allylic C–H activation would
be rate-determining. A competition experiment using a 1:1
(v/v) ratio of allylbenzene-d0 and allylbenzene-d5
(CD2=CDCD2C6H5) shows no kinetic isotope effect for the
consumption rate of allylbenzene (Fig. 4), suggesting olefin
coordination rather than C–H activation is rate-determining.
However, if cycle 1 is the dominant pathway, we cannot
8. M.S. Chen, N. Prabagaran, N.A. Labenz, and M.C. White. J.
Am. Chem. Soc. 127, 6970 (2005).
9. K.J. Fraunhoffer and M.C. White. J. Am. Chem. Soc. 129,
7274 (2007).
10. K.J. Fraunhoffer, N. Prabagaran, L.E. Sirois, and M.C. White.
J. Am. Chem. Soc. 128, 9032 (2006).
11. N. Marion, O. Navarro, J.-G. Mei, E.D. Stevens, N.M. Scott,
and S.P. Nolan. J. Am. Chem. Soc. 128, 4101 (2006).
12. W. Kitching, T. Sakakiyama, Z. Rappoport, P.D. Sleezer, S.
Winstein, and W.G. Young. J. Am. Chem. Soc. 94, 2329 (1972).
13. J. Ichihara. Chem. Commun. 1921(1997).
14. Preliminary screening of the following enantiopure ligands at
60 °C did not, however, lead to significant enantioselectivity:
(–)-sparteine, (4S)-Bz-Box, (4S)-tBu-Box, and (S)-(–)-BINAP.
15. W. Kitching, T. Sakakiyama, Z. Rappoport, P.D. Sleezer, S.
Winstein, and W.G. Young. J. Am. Chem. Soc. 94, 2329 (1972).
16. A small amount of 1-phenylallyl acetate (~9%) is also formed
during the reaction. More detailed mechanistic studies are un-
derway to reconcile this observation with the proposed mecha-
nism.
3
rule out the possibility that acetate attack on the dimeric (η -
allyl)PdII complex is rate-determining. In that case, the
3
dimeric (η -allyl)PdII complex would be the major species
within the actual catalytic cycle, but it could still be
unobservable if the equilibrium between 1 and 3 strongly fa-
vours the former. Since at present we do not know which cy-
cle is dominant, we cannot conclusively identify the rate-
determining step.
In summary, we have reported (BPM)Pd(OAc)2 as a well-
defined new catalyst for allylic C–H oxidation of olefins.
Mechanistic studies have provided insight to understand the
catalysis and led to further improvement of the reaction
yield for the oxidation of allylbenzene. More studies are un-
derway to explore the potential of the dimeric species and its
derivatives in catalytic allylic C–H oxidations.
17. J.E. Bäckvall, R.E. Nordberg, and D. Wilhelm. J. Am. Chem.
Soc. 107, 6892 (1985).
18. A
similar
equilibrium
was
also
proposed
for
(PCy3)Pd(allyl)(OAc): T. Yamamoto, O. Saito, and A.
Yamamoto. J. Am. Chem. Soc. 103, 5600 (1981).
19. (a) Mixing Pd(dba)2, BPM, and BQ in acetone forms
(BPM)Pd(BQ): R.A. Klein, P. Witte, R. van Belzen, J. Fraanje,
K. Goubitz, M. Numan, H. Schenk, J.M. Ernsting, and C.J.
Elsevier. Eur. J. Inorg. Chem. 319 (1998); (b) BQ could be
viewed as an electron-deficient olefin. For a detailed mecha-
nistic study of olefin exchange on Pd0, see: S.S. Stahl, J.L.
Thorman, N. de Silva, I.A. Guzei, and R.W. Clark. J. Am.
Chem. Soc. 125, 12 (2003).
Acknowledgement
We thank Dr. Michael W. Day and Lawrence M. Henling
for performing the X-ray crystallographic analysis. B.-L. Lin
also thanks George S. Chen for suggestions in preparing the
manuscript. This work has been generously supported by BP
through the MC2 program.
© 2008 NRC Canada