Optimization studies on the hydro-acylation of 1-alkenes to α-methylketones using a homogeneous palladium/Josiphos catalyst

Article abstract of DOI:10.1021/op100342v

The effect of process conditions (temperature, partial pressures of CO and H₂) on the product selectivity and ee of the palladium-catalysed asymmetric hydro-acylation of 1-pentene in dichloromethane with the chiral Josiphos ligand (SL-J008-1, L1) was investigated. The highest ee value (73%) for the desired product 4-methyl-5-decanone (1) was obtained at the lowest temperature in the range (30 °C). The selectivity for 1 was between 12 and 20 mol % at 30 °C, though considerably higher at 90 °C (59 mol %). Statistical modeling was applied to quantify the influence of the process conditions on the product selectivity and ee.

Full text of DOI:10.1021/op100342v

Article
pubs.acs.org/OPRD
Optimization Studies on the Hydro-acylation of 1-Alkenes to
α-Methylketones using a Homogeneous Palladium/Josiphos Catalyst
†
‡
‡
,†
Hans Heeres, Adri J. Minnaard, Bart Hessen, and Hero J. Heeres*
†
‡
Department of Chemical Engineering, ITM and Stratingh Institute for Chemistry University of Groningen, Nijenborgh 4,
9747 AG Groningen, The Netherlands
ABSTRACT: The effect of process conditions (temperature, partial pressures of CO and H ) on the product selectivity and
2
ee of the palladium-catalysed asymmetric hydro-acylation of 1-pentene in dichloromethane with the chiral Josiphos ligand (SL-
J008-1, L1) was investigated. The highest ee value (73%) for the desired product 4-methyl-5-decanone (1) was obtained at the
lowest temperature in the range (30 °C). The selectivity for 1 was between 12 and 20 mol % at 30 °C, though considerably
higher at 90 °C (59 mol %). Statistical modeling was applied to quantify the influence of the process conditions on the product
selectivity and ee.
INTRODUCTION
The enantioselective production of linear α-substituted ketones
is of particular interest as this structure element is present in
Our research activities concern the development of an asym-
metric version of the hydro-acylation reaction given in Scheme
■
2
. We have performed an extensive catalyst screening study
1
,2
with homogeneous palladium catalyst of the type (L )Pd-
many drugs and pheromones. A known method for the
asymmetric synthesis of α-methyl substituted ketones is the
reduction of α-methylene ketones by the use of Baker’s yeast
2
8
(
OTf) . A total of 16 diphosphine ligands belonging to four
2
different chiral ligand classes (Josiphos, DuPhos, Walphos, and
ferroTANE) were studied (Figure 1). The reactions were
carried out in dichloromethane at reaction temperatures be-
3
−5
(
Scheme 1). Good results were obtained in, for example, the
tween 60 and 125 °C and at initial H and CO pressures of
2
Scheme 1. Reduction of methyleneketones by Baker’s yeast
3
0 bar each. The enantio-, chemo-, and regioselectivity of the
reaction are a clear function of the diphosphine ligand (Table 1).
Low ee values (<20%) were obtained for the Duphos ligands,
albeit at a product selectivity >66 mol %. The highest ee for 1
was 39% for a Josiphos type ligand (L1) (Figure 1) at 60 °C,
though the product selectivity was low (6 mol %). Other
products were enones, higher CO−olefin oligomers resulting
from alternating CO−olefin copolymerisation aldehydes/
alcohols resulting from hydroformylation and olefin oligomer-
isation products (Scheme 3).
reduction of α-methyleneketone with R = Me and R = n-hexyl
1
2
at 30 °C. A conversion of 70% was obtained after 2 h with a
5
product ee > 99%.
An attractive alternative for the synthesis of α-methyl-
substituted ketones is the hydro-acylation of 1-alkenes with
Syngas (Scheme 2). Recently, Drent and Budzelaar reported a
In this paper, we have investigated the effects of process
conditions such as temperature and H /CO pressure on the
2
Scheme 2. Hydro-acylation of 1-alkenes to α-methylketones
hydro-acylation reaction of 1-pentene with chiral ligand L1,
with the objective to optimise the chemo-, regio-, and enan-
tioselectivity. Statistical modelling software (Design Expert) has
been used to investigate and quantify the influence of these
process conditions on reaction performance.
EXPERIMENTAL SECTION
■
Chemicals. Josiphos ligand L1 ((R,S)-SL-J008-1), was
purchased from Sigma Aldrich (≥97%). The ligand and palla-
dium acetate (Sigma Aldrich >98%) were stored in a glovebox.
Dichloromethane (Lab-Scan 99.8%) was distilled from CaH2
before use and stored under nitrogen. 1-Pentene (Sigma
nonasymmetrical version of this reaction, using palladium
diphosphine catalysts of the type L PdX (in which X is a non-
2
2
coordinating anion) for the selective production of mono-
6
,7
ketones (Scheme 2). Very promising results were obtained
Aldrich 98%) was distilled from sodium and stored under
with a catalyst prepared in situ from Pd(OAc) , an alkyl-
2
9,10
nitrogen.
Trifluoromethanesulfonic acid (HOTf) (Strem
diphosphine like DnBPP (1,3-bis(di-n-butylphosphino)-
propane), and trifluoromethanesulphonic acid (HOTf). For
9
9+%) was stored under nitrogen at 5 °C. The synthesis gas
1
-octene, a chemoselectivity of 98 mol % was obtained when
the reactions were performed in diglyme at 125 °C with a
PCO = PH2 = 30 bar.
Received: December 26, 2010
Published: January 24, 2012
6
,7
©
2012 American Chemical Society
400
dx.doi.org/10.1021/op100342v | Org. Process Res. Dev. 2012, 16, 400−408

Organic Process Research & Development
Article
temperature 200 °C. The reactor was electrically heated and the
reactor content stirred by a Parr overhead stirrer, equipped with
a gas inducing impeller. Temperature and stirring speed were
controlled by the Parr 4843 controller. The synthesis gas was
fed from a premixed (50/50) gas cylinder and additional
hydrogen was added manually from a separate gas cylinder.
General procedure. The catalyst was freshly prepared
before each experiment under a nitrogen atmosphere using
standard Schlenk techniques. Pd(OAc) (9 mg, 0.04 mmol)
2
was dissolved in dichloromethane (2 mL). After approximately
1
0 min the diphosphine ligand L1 (0.04 mmol) dissolved in
dichloromethane (2 mL) was added. The solution was stirred
for 10 min prior to the addition of HOTf (30 mg, 0.2 mmol)
and stirred for another 10 min before use. The batch autoclave
was charged with solvent (dichloromethane, 10 mL), 1-pentene
(
4.0 mL, 36.5 mmol) and the catalyst solution. The reactor was
closed and flushed with nitrogen to remove air. The reactor was
pressurised (10−100 bar) with synthesis gas (50/50 H /CO), and
2
when required, H (10−140 bar) was added. Subsequently, the
2
reactor was heated to the desired reaction temperature
(
30, 60, or 90 °C). During reaction, the stirrer speed was main-
tained at 1000 rpm. After 5 h, the reactor was cooled to room
temperature, depressurised, and flushed several times with N . The
2
liquid product was filtered over silica gel to remove the catalyst.
Product Analyses. The product composition in the liquid
phase was analysed using a GC-FID HP-5890 series II,
equipped with a 30 m HP-1 column and He as the carrier gas.
The following temperature profile was applied: 10 min at 30 °C,
from 30 to 325 °C at a rate of 10 °C/min, 15 min at 325 °C.
Product compositions were obtained by comparing product
Figure 1. Overview of chiral diphosphine ligands tested for the
hydroacylation of 1-pentene.
14
peak areas by means of the 100% method. The method was
applied for peaks belonging to the most abundant product
groups present in the samples: mono-oxygenates (ketones,
aldehydes/alcohols), olefin dimers and trimers, and diketones.
The response factors of all individual components are not known.
To compensate for this, the concept of effective carbons was
used was a 50:50 premixed CO/H gas mixture (HiQ (high
2
quality), CO ≥ 99,997 vol%, H ≥ 99.9999 vol%) and was pur-
2
chased from Hoek Loos. Pure hydrogen gas was also obtained
from Hoek Loos (HiQ, purity ≥99.9999 vol%). Racemic 4-
methyl-5-decanone was synthesised using literature proce-
15
applied to determine the mol fraction of a product (eq 1).
4
,11
dures. Enantio-enriched (80% ee) 4-methyl-5-decanone was
Fj/( ∑ C
)
ef j
Fi/( ∑ C
4
,11
synthesised in a four step procedure.
Enantio-enriched 4-
Xj =
n
∑
)
)
(
ef i
(1)
methyl-nonanone (90% ee) was prepared using a three step
i
4
,12
procedure.
The tail-to-tail monoketone, 6-undecanone (2),
Here X stands for the mol fraction of component j and F
j
j
was purchased from Sigma Aldrich (97%).
stands for the peak area of component j. F is divided by the
j
Experimental Setup. The catalytic reactions were carried
out in a Parr autoclave (50 mL), which was operated in a batch
mode with respect to the gas phase (Figure 2). The maximum
operating pressure of the reactor was 200 bar, the maximum
sum of its effective carbons (C ) to obtain the ‘effective area’ of
ef
component j. The effective area of component j is divided by
the sum of all effective areas of the relevant components in the
16
chromatogram.
Table 1. Overview of results for catalytic hydro-acylations of 1-pentene using the examples of chiral Josiphos, Duphos,
a,b
FerroTANE, and Walphos ligands
sat. MK (1 + 2 + 3)
sat. MK (1)
sat. MK (2)
enones (4)
alc/ald (8 + 9)
DK (5)
(%)
dim. (6)
(%)
olefin conv
(%)
c
d
e
f
g
e
entry ligand
(%)
(%)
SMD
ee
(%)
(%)
(%)
1
2
3
4
5
6
L1
L2
L3
L4
L5
L6
44
69
59
42
16
15
13
95
6
66
57
42
7
39
87
5
23
4
0
1
0
0
1
0
32
17
34
37
27
24
0
0
59
58
32
72
39
73
h
16
20
h
96
4
2
0
100
44
6
4
6
0
16
48
36
0
56
53
0
47
7
15
a
b
Intake: 36.5 mmol of 1-pentene; 0.04 mmol of Pd; 0.04 mmol L ; 0.2 mmol HOTf as the acid, 5 h reaction time, P = PCO = 30 bar, 60 °C.
2
H
2
c
d
Numbering of compounds is given in Scheme 3. Ligand numbering is provided in Figure 1 H-to-h regioisomer (3) was below the GC detection
e
f
g
h
limit. In the sat. MK fraction (1+2+3). Defined according to eq 3. Ee of the saturated h-to-t monoketone, 4-methyl-5-decanone (1). Actual value
likely lower due to peak overlap in GC measurements.
4
01
dx.doi.org/10.1021/op100342v | Org. Process Res. Dev. 2012, 16, 400−408

Organic Process Research & Development
Article
Scheme 3. Main- and side products from the L PdX -catalyzed reactions of 1-pentene with Syngas (only one isomer of the
2
2
products is shown)
The 1-pentene conversion was determined using the same GC
equipment. The concentration of 1-pentene in the liquid phase
after reaction was determined using calibration curves with known
β-PM column with He as the carrier gas. The following tem-
perature profile was applied: 50 to 55 °C, rate 10 °C/min, hold
time 100 min, then from 55 to 180 °C, rate 10 °C/min, final
time 15 min. A solvent change was required before the reaction
products could be injected. For this purpose, the dichloro-
methane was removed under reduced pressure (100 mbar, 30
1-pentene concentrations. The product solution was directly ana-
lyzed after reaction to avoid excessive evaporation of 1-pentene.
Product identification was performed by GC−MS and GC.
GC−MS chromatograms were recorded on a GC (HP 6890)
MS (HP-5973) equipped with a 30 m HP-5 MS column and
He as the carrier gas (from 35 to 250 °C, rate 5 °C/min, final
time 10 min.). GC chromatograms were recorded on a GC-FID
HP-5890 series II, with FID on a 30 m HP-1 column and
He as the carrier gas (vide supra). Reference compounds, either
obtained from chemical suppliers or prepared, were used to
identify relevant components in the mixture.
°C) to give a yellow oil. The oil was redissolved in diethyl ether
and analysed. The retention times of the enantiomers were 107.9
and 108.3 min. The chromatograms were compared with the
GC−MS and GC-FID analyses for peak identification and to
identify possible overlap of individual enantiomers with by-
product of the reaction. Also the chromatograms were compared
with those obtained with the reference materials.
The enantiomeric excess of the desired 4-methyl-5-decanone
1) was determined using a GC equipped with a 30 m chiral
Statistical Modeling Using Design Expert. The
experimental results for each response were analyzed using
(
4
02
dx.doi.org/10.1021/op100342v | Org. Process Res. Dev. 2012, 16, 400−408

Organic Process Research & Development
Article
software to obtain a quantitative relation between the responses
and the process variables.
Effect of Process Conditions on the ee of 1. The ee of 1
is a strong function of the temperature. Highest ee values were
obtained at 30 °C (67−73%, see Table 3). At a reaction
temperature of 90 °C (Table 3, entry 1 and 2) the ee was
negligible. A verification experiment was performed with the
other enantiomer (S,R) of Josiphos ligand L1 at 30 °C. In this
case, the opposite enantiomer of 1 was obtained with an ee of
6
7%, which is within the experimental error. The GC chroma-
tograms for both experiments are given in Figure 3.
The effect of the process conditions on the ee was quantified
using statistical modelling. It was found that only the tem-
perature and neither the hydrogen nor the CO partial pressure
has a significant effect on the ee, leading to the following
relation:
Figure 2. Schematic representation of the reactor setup.
the Design Expert 7 software package. Responses are modeled
using standard expressions (eq 2):
ee = 103 − 1.14 × T
(4)
Clearly a low temperature has a positive effect on product ee. A
parity plot for the predicted and experimental ee values using this
equation is given in Figure 4, indicating that agreement between
2
y = b + ∑ b x + ∑ b x + ∑ ∑ b x x
0
i i
ii i
jk i k
i
j
k
(2)
2
model and experimental data is good (adjusted R : 0.973).
Selectivity for 4-Methyl-5-decanone (SMD) as a
Function of Process Conditions. Statistical modeling of
the data reveals that the SMD is a function of both the reaction
temperature and the partial CO pressure (eq 5). The effect of
hydrogen pressure is not significant.
Here, i = A to C, j = A to C, and k = A to C with A, B, C
representing the independent variables; bi, bii, and bjk are the
regression coefficients which are obtained by statistical analyses of
the data. Significant factors were selected on the basis of their p
value in the ANOVA analyses. Factors with a p value lower than
0
.05 are regarded as significant and included in the response model.
S
= − 13 + 0.81 × T − 0.90 × PCO
(5)
MD
Backward elimination was applied to eliminate all statistically
insignificant terms. After each elimination step, a new ANOVA table
was generated to select the subsequent nonsignificant factor.
Thus, low partial CO pressures and higher reaction temper-
atures have a positive effect on SMD. This is also illustrated in
Figure 5, where the SMD is provided as a function of the CO
pressure and the reaction temperature. The highest selectivity
60 mol %) was obtained at a temperature of 90 °C and a
PCO of 5 bar.
RESULTS
■
(
Experimental Window and Approach. A total of 10 ex-
periments was carried at different process conditions (tempera-
ture and partial CO and hydrogen pressure) using an in situ
Agreement between the model data and experimental data is
adequate (adjusted R : 0.910), as is also illustrated by the parity
2
formed Pd-catalyst based on Pd(OAc) , Josiphos ligand L1 and
HOTf as the acid component in dichloromethane. The experi-
mental ranges for the independent variables are given in Table 2.
2
plot given in Figure 6. Evidently, it is desired to identify those
process conditions that lead to 1 with a high product selectivity
and enantioselectivity. Unfortunately, these conditions appear
to be conflicting. The highest ee is obtained at low temperature,
whereas the highest selectivity to 1 is observed at the highest
temperature in the range.
Table 2. Experimental ranges for the hydro-acylation of
1
-pentene
Chemoselectivity vs Process Conditions. The SMD
is a function of both the chemo- and regioselectivity of the
hydro-acylation reaction. To gain insight in the extent to which
the two factors influence the SMD, the effect of process conditions
on the chemo- and regioselectivity were evaluated separately.
The highest chemoselectivity with respect to saturated
monoketones was obtained at 90 °C (Table 3, entry 2: 83%
saturated monoketones). The major byproduct are enones
lowest value
highest value
T (°C)
PCO (bar)
PH2 (bar)
30
90
2.5
17.5
30
105
For all experiments, a fixed Pd intake was applied and the
substrate to catalyst ratio was set at 910 (about 0.1 mol %). For
each experiment, the olefin conversion, product distribution
and the ee of the desired 4-methyl-5-decanone (1) was deter-
mined and the results are given in Table 3. Here, the selectivity
for 1 (SMD, mol %) is defined as follows:
(
(
4, 2−37 mol %), diketones (5, 2−32 mol %) and olefin dimers
6, 0−80 mol %) and trimers (7, 0−17 mol %). Aldehydes (8)
and/or alcohols (9) formation by hydroformylation was ob-
served for only two experiments (maximum 2 mol %). The
1
-pentene dimers and trimers (Scheme 3) are monounsaturated
S
= (amount of saturated monoketones (1 + 2 + 3))
MD
(
GC−MS) though the position of the double bond and the
×
(fraction of (1) in saturated monoketones)
extent of branching could not be established.
C
1
The effect of process conditions on the desired saturated
monoketone fraction in the product mixture is best described
by the following relation (adjusted R : 0.880):
=
∑
C
all products
(3)
2
The responses (ee, conversion, SMD, chemo- and regioselectivity)
at different process conditions were analysed using statistical
F
= − 13 + 1.08 × T − 0.65 × PCO
(6)
sat.MK
4
03
dx.doi.org/10.1021/op100342v | Org. Process Res. Dev. 2012, 16, 400−408

Organic Process Research & Development
Article
Table 3. Effects of process conditions on conversion, chemo-, regio-, and enantioselectivity for the hydro-acylation of
a,b
1
-pentene
dim.
(6)
(%)
trim.
(7)
(%)
T
PCO
(bar)
P
(bar)
H
2
sat. MK (1+ 2
sat. MKd
sat. MKd
enones
alc/ald
DK (5)
(%)
olefin
conv (%)
c
e
f
entry (°C)
+ 3) (%)
(1) (%)
SMD
ee
(2) (%)
(4) (%) (8 + 9) (%)
1
2
3
4
5
6
7
8
9
1
90
90
60
60
60
60
60
30
30
30
30
5
30
105
30
61
83
44
43
55
44
20
20
18
12
56
72
13
50
50
41
55
55
62
84
35
60
6
0
44
28
87
50
50
59
45
45
38
16
29
7
1
0
0
0
1
0
0
0
0
0
8
7
0
0
0
0
64
84
59
81
50
77
90
22
59
24
0
39
29
41
32
33
67
65
73
30
5
24
22
23
18
37
7
32
21
19
15
21
11
6
0
0
50
22
27
18
11
11
11
10
10
0
0
15
15
30
5
105
145
120
50
0
20
19
43
62
80
2
2
17
10
2.5
2.5
50
2
0
17.5
2
2
4
a
b
Intake: 36.5 mmol of 1-pentene; 0.04 mmol of Pd; 0.04 mmol L ; 0.2 mmol HOTf as the acid, 5 h reaction time, dichloromethane. Numbering of
2
c
d
e
compounds is given in Scheme 3. H-to-h regioisomer (3) was below the GC detection limit. In the sat. MK fraction (1+2+3). Defined according
to eq 3. Ee of the saturated h-to-t monoketone, 4-methyl-5-decanone (1).
f
Figure 3. Chiral GC chromatograms of the experiment performed at 30 °C, PCO = 2.5 bar and PH2 = 17.5 bar with (R,S)-L1 and (S,R)-L1.
The effect of hydrogen pressure is statistically not relevant. The
relation is visualised in Figure 7, a parity plot is provided in
Figure 8. Higher reaction temperatures have clearly a positive
effect on the amount of saturated monoketone formed. The
effect of the PCO is less pronounced, though higher partial CO
pressures have a small but significant negative effect on the
saturated monoketone selectivity.
For some of the reactions, olefin dimers and trimers
were actually the main products (Table 3). The amount of
olefin dimers in the mixture was modeled as a function of
process conditions and is best described by the following
relation:
F
= 85 − 0.96 × T − 0.52 × PCO
(7)
dim
4
04
dx.doi.org/10.1021/op100342v | Org. Process Res. Dev. 2012, 16, 400−408

Organic Process Research & Development
Article
Figure 4. Experimental versus modeled ee values using eq 4.
Figure 7. Chemoselectivity towards saturated monoketones (1 + 2 + 3)
versus reaction temperature and PCO (PH2 = 80 bar).
Figure 5. Selectivity towards 1 (SMD) versus reaction temperature and
partial CO pressure (H fixed at 80 bar).
Figure 8. Experimental versus predicted values for the chemo-
selectivity of the saturated monoketone fraction (1+2+3).
2
Figure 6. Predicted selectivity for (1) versus the experimental values.
The relation is visualised in Figure 9. Clearly, the temperature
has a pronounced effect on the amount of dimerization
products, with lower temperatures leading to high amounts of
dimers. The amount of dimers is also a function of the partial
Figure 9. Modeled amount of olefin dimerization products (6) as a
function of process conditions (eq 7).
4
05
dx.doi.org/10.1021/op100342v | Org. Process Res. Dev. 2012, 16, 400−408

Organic Process Research & Development
Article
CO pressure, with high pressures leading to a reduction in the
amount of dimerization products.
Scheme 4. Proposed catalytic cycle for the Pd(II)-catalysed
reactions of 1-alkenes and syngas
Regioselectivity. The major regioisomers present in the
saturated monoketone fraction were the head-to-tail (1) and
the tail-to-tail isomer (2), whereas the head-to-head
regioisomer (3) could not be detected (GC). It turned out
to be impossible to determine statistically sound relations be-
tween the process conditions and the amount of the desired
head-to-tail regioisomer in the reaction mixture. However, in
the majority of the experiments, the desired head-to-tail isomer
is formed in the largest amounts.
Olefin Conversion. The highest 1-pentene conversion after
h was 90%, obtained at 60 °C, a partial CO pressure of 30 bar
5
and a partial H2 pressure of 120 bar (Table 3). This
corresponds to an average turnover frequency (TOF) of 160
mol/(mol Pd·h). The TOF at the initial stage of the reaction is
expected to be considerably higher as the TOF reported here is
an average over the batch time. Typical literature values for
achiral hydro-acylations using alkyldiphosphines in batch re-
actors at 115 °C are about 1000 mol/(mol Pd·h) for 1-propene
6,7
and about 100 mol/(mol Pd·h) for higher olefins like 1-octene.
Thus, particularly considering the relatively low temperature,
the TOF for the Josiphos-based catalyst is at the high end of the
range given in the literature.
A sound statistical relation for the effects of process
conditions on the 1-pentene conversion could not be obtained.
One of the possible reasons is the simultaneous occurrence of
favor the formation of saturated monoketones. Higher CO
pressures are expected to speed up the rate of CO insertion in
the Pd-alkyl bond. This reaction is known to be reversible and
only a subsequent olefin insertion will produce a stable Pd-alkyl
compound. The latter is subsequently hydrogenated to form a
ketone.
1
-pentene isomerisation to mixtures of internal olefins (GC
analyses of the reaction mixture). It is well-known that the
activity of L PdX catalysts for internal olefins is much lower
than for α-olefins. The extent of olefin isomerisation will also
be affected by process conditions, though to a different extent than
hydro-acylation, leading to a complex relation between olefin con-
version and process conditions. Furthermore, isomerisation is a
reversible process and the internal olefins may isomerise back to
2
2
6
,7
The observation that the hydrogen pressure does not affect
the chemoselectivity is rather surprising. On the basis of
Scheme 4, it is anticipated that higher hydrogen pressures
would lead to higher amounts of the saturated monoketones at
the expense of enone formation. However, this is not the case.
This is an indication that hydrogen is not directly involved in
the formation of saturated monoketones from the Pd-chelate. A
possible alternative termination mechanism without direct
hydrogen involvement is the reaction of the Pd-chelate with
HOTf, present in excess in the reaction mixture, to give a
L Pd(OTf) species and the saturated monoketone (proto-
1-pentene and converted to hydro-acylation products.
DISCUSSION
■
The hydrocarbonylation of olefins using (L )PdX catalysts
2
2
leads to the formation of the desired monoketones (hydro-
acylation) as well as to alcohols and aldehydes by hydro-
formylation and higher ketones by copolymerisation reac-
6
,7
tions. A proposed catalytic cycle including the formation of
the various product classes is given in Scheme 4. The active
species is commonly accepted to be a square-planar cationic
2
2
nolysis). Subsequent reaction of the L Pd(OTf) with hydro-
2
2
6
,7
gen then leads to the regeneration of the Pd−H and HOTf
Pd-hydride complex.
17
(
Scheme 5). Termination reactions by protonolysis are
Chemoselectivity. The major product classes formed using
Josiphos ligand L1 in dichloromethane are saturated mono-
ketones, enones, diketones and noncarbonylation products like
olefin dimers and trimers (Scheme 3 and Table 3). Aldehydes/
alcohols were not formed in significant amounts under these
conditions (<2 mol %). This suggests that the key intermediate
in the catalytic cycle is the Pd-alkyl species with a chelating
carbonyl group (top compound in catalytic cycle, Scheme 4).
This compound may either react with hydrogen to form the
desired saturated monoketones, undergo β-hydrogen elimination
to produce enones or react with CO and subsequently an olefin
to give a diketone. Apparently, all three pathways occur when
using L1. The selectivity towards the desired saturated
monoketones is shown to be a function of the temperature
and to a lesser extent the CO pressure (eq 6), whereas the
effect of the hydrogen pressure was statistically not relevant.
Higher temperatures and lower CO pressures were shown to
Scheme 5. Alternative termination mechanism: Pd-alkyl
protonolysis with HOTf
well-known in CO-olefin copolymerisations. For instance,
Vavasori and co-workers reported the active involvement of
acetic acid in the termination reaction in the ethylene/CO
18
copolymerisations catalysed by PdCl (dppf) complexes.
2
4
06
dx.doi.org/10.1021/op100342v | Org. Process Res. Dev. 2012, 16, 400−408

Organic Process Research & Development
Article
Remarkable is the formation of 1-pentene oligomers under
hydro-acylation conditions. The formation of olefin dimeriza-
tion products in the absence of CO with the L PdX catalyst
experimental ranges. This gives valuable information on the
origin of enantioselectivity of the hydro-acylation reaction. The
enantioselectivity of the reaction is most likely determined by
the second insertion of the olefin in the Pd-alkyl bond of the
2
2
used in this study is well-known in the literature. Early ex-
8
chelate (Scheme 4). To obtain enantiopure 1, the insertion
amples of the dimerization of linear α-olefins were reported by
1
9
20
step should be highly stereoselective. Lower ee values may be
explained by either an intrinsically limited chiral induction of
the Josiphos ligand or by partial epimerization of the stereo-
genic centre after the second olefin insertion step (Scheme 7).
In the case the ee is only affected by the intrinsic properties
of the Josiphos ligand, the partial pressures of CO and hydro-
gen are not expected to affect the product ee, as was observed
experimentally. However, this does not exclude the occurrence
of epimerization. Epimerization can occur by (reversible)
β-hydrogen elimination followed by conjugate hydride addition
Drent (1990) and Jiang et al. (1993). In 2003, Tsuchimoto
2
1
et al. demonstrated that palladium−indium triflate catalysts
were highly active for the dimerization of vinylarenes. More
recently, Bedord et al. developed an (asymmetric) version
22
by using Pd(OAc) , and a chiral diphosphine ligand in the
2
presence of In(OTf) (Scheme 6). Josiphos ligands were also
3
shown to be active for the reaction.
Scheme 6. Styrene oligomerisation catalysed by Pd(AcO)2/
diphosphine ligand/In(OTf) catalyst systems
23
3
and formation of the palladium enolate. In this case, the pro-
duct ee is not only determined by the intrinsic chiral induction
during the olefin insertion step but is also a function of the rate
of β-hydrogen elimination/enolisation versus the rate of the
termination reaction of the Pd-chelate with either HOTf or
hydrogen (Scheme 7). The termination reaction with hydrogen
is less likely as the amount of saturated monoketones in the
mixture was found to be independent of the hydrogen pressure
A possible explanation for olefin oligomerisation is the
strong) acid-catalysed (here HOTf) oligomerisation of
-pentene. If this would be valid, however, olefin oligomerisa-
(
1
tion is expected to occur for all reactions and this is certainly
not the case (Table 3). An alternative explanation is the
involvement of cationic Pd catalysts, even though CO is
present. In this case, olefin insertion has to compete with CO
insertion in a Pd-alkyl bond (Scheme 4). When this explanation
is valid, higher amounts of olefin dimers are expected at low
CO pressures. This was indeed observed experimentally and
supported by statistical modelling (Figure 9). Though not
conclusive, a mechanism involving Pd-compounds for 1-
pentene oligomerisation is likely. Further research will be
required to provide definite conclusions on the formation of
olefin oligomers.
Regioselectivity. The major regioisomers present in the
saturated monoketone fraction were the head-to-tail (1) and
the tail-to-tail isomers (2), whereas the head-to-head
regioisomer (3) could not be detected (GC). It proved not
possible to determine statistically sound relations between the
process conditions and the amount of the desired head-to-tail
regioisomers in the reaction mixture. However, in the majority
of the experiments, the desired head-to-tail isomer is formed in
the largest amounts.
(
vide supra). Thus, the rate of termination by protonation
versus the rate of β-hydrogen elimination likely affects the rate
of epimerization. When this mechanism is valid, the product ee
is expected to be independent of the hydrogen pressure and
this was indeed the case. Further research including molecular
modeling is required to gain further insight into the occurrence
of epimerization and its role in the enantioselectivity of the
reaction.
CONCLUSIONS
■
The effect of process conditions (temperature, partial CO and
hydrogen pressure) on the synthesis of enantio-enriched 1 by
the Pd-Josiphos L1-catalysed reaction of 1-pentene with syngas
was established. The selectivity of the reaction (SMD) was
shown to be a function of the temperature and the highest
amounts of 1 were obtained at the highest temperature in the
range (90 °C). The product ee is also strongly temperature
depending, with lower temperatures leading to higher ee values.
The highest ee of 73% was found at 30 °C. As the requirements
for high chemoselectivity and enantioselectivity are clearly
conflicting, a high yield synthesis of enantiopure 1 with the
Josiphos ligand used in this study remains elusive. Such studies
Enantioselectivity. The enantioselectivity was shown to be
a strong function of the temperature (eq 4), whereas the PCO
and PH2 have no significant effect on the product ee within the
Scheme 7. Epimerization of the stereogenic centre leading to ee reductions
4
07
dx.doi.org/10.1021/op100342v | Org. Process Res. Dev. 2012, 16, 400−408

Organic Process Research & Development
Article
with alternative bidentate chiral ligands are in progress and will
be reported in due course. The effects of process conditions
were quantified using statistical modelling. The results provide
interesting mechanistic insights in the effects of process
conditions on the chemo-, regio-, and enantioselectivity of
(asymmetric) hydro-acylation reactions.
AUTHOR INFORMATION
Corresponding Author
H.J.Heeres@rug.nl
■
*
ACKNOWLEDGMENTS
■
The author would like to thank Dr. F. Gini for her help with the
chemistry and analyses and Prof. dr. E. Drent from the Leiden
University, The Netherlands for stimulating discussions and
valuable input.
REFERENCES
■
(
(
(
1) Loughlin, W. A. Bioresources Technology 2000, 74, 49−62.
2) Mori, K.; Harashima, S. Tetrahedron Lett. 1991, 32, 5995−5998.
3) Conceicao, G. J. A.; Moran, P. J. S.; Rodrigues, J. A. R. Arkivoc
2
(
003, 500−506.
4) Gini, F. PhD thesis University of Groningen, 2007.
(
5) Siqueira, E. P.; Rodrigues, J. A. R.; Moran, P. J. S. Tetrahedron-
Asymmetry 2001, 12, 847−852.
6) Drent, E.; Budzelaar, P. H. M. J. Organomet. Chem. 2000, 594,
11−225.
7) Drent, E.; Mul, W. P.; Budzelaar, P. H. M. Comments in Inorg.
Chem 2002, 23, 127−147.
8) Heeres, H.; Minnaard, A. J.; Hessen, B.; Heeres, H. J.;
unpublished results
9) Perrin, D. D.; Armarego, W. L. F. Purification of laboratory
chemicals, Pergamon Press, Oxford, 3rd ed. 1988.
(
2
(
(
(
(
(
10) van Leeuwen, P. W. N. M. Appl. Catal., A 2001, 212, 61−81.
11) Oldenziel, O. H.; van Leusen, D.; van Leusen, A. M. J. Org.
Chem. 1977, 42, 3114−3118.
12) Wipf, P.; Kim, Y.; Fritch, P. C. J. Org. Chem. 1993, 58, 7195−
203.
(
7
(
(
13) Corey, E. J.; Enders, D. Tetrahedron Lett. 1976, 17, 3−6.
14) Braam, W. G. M. Chromatografie, Educatieve Partners
Nederland BV, Culemborg, The Netherlands, 5, 1994.
15) Svojanovsky, V.; Krejí, M.; Tesaík, K.; Janak, J. Chromatogr. Rev.
966, 8, 90−171.
16) Jorgensen, D.; Picel, K. C.; Stamoudis, V. C. Anal. Chem. 1990,
2, 683−689.
17) Robertson, R. A. M.; Cole-Hamilton, D. J. Coord. Chem. Rev.
002, 225, 67−90.
18) Vavasori, A.; Bellieni, A.; Ronchin, L.; Dall’Acqua, F.; Toniolo,
(
́
́
1
(
6
(
2
(
L.; Cavinato, G. J. Mol. Catal. A: Chem. 2007, 263, 9−14.
(
(
(
19) Drent, E. Pure Appl. Chem. 1990, 62, 661−669.
20) Jiang, Z.; Sen, A. Organometallics 1993, 12, 1406−1415.
21) Tsuchimoto, T.; Kamiyama, S.; Negoro, R.; Shirakawa, E.;
Kawakami, Y. Chem. Commun. 2003, 852−853.
22) Bedford, R. B.; Betham, M.; Blake, M. E.; Garce’s, A.; Millara,
S. L.; Prashar, S. Tetrahedron 2005, 61, 9799−9807.
23) Zuideveld, M. A.; Kamer, P. C. J.; van Leeuwen, P. W. N. M.;
Klusener, P. A. A.; Stil, H. A.; Roobeek, C. F. J. Am. Chem. Soc. 1998,
20, 7977−7978.
(
(
1
4
08
dx.doi.org/10.1021/op100342v | Org. Process Res. Dev. 2012, 16, 400−408