An Empirical Study of Phosphine Ligands for the Methoxycarbonylation of Medium-Chain Alkenes

Article abstract of DOI:10.1002/cctc.201500464

The methoxycarbonylation reaction provides a route to the synthesis of esters from medium-chain alkenes that may be used as fuel supplements. However, the known productive catalytic systems are expensive and/or unstable at elevated temperatures. Most of the data available on the methoxycarbonylation of alkenes is derived from ethylene and styrene as substrates. To broaden the scope, we conducted a comparative study of a range of phosphine ligands under comparable conditions for the methoxycarbonylation of 1-octene. The results demonstrate that a number of ligand structural motifs facilitate the process effectually. Furthermore, the critical importance of alkene isomerization and the acid/ligand and Pd/ligand ratios are presented.

Full text of DOI:10.1002/cctc.201500464

DOI: 10.1002/cctc.201500464
Communications
An Empirical Study of Phosphine Ligands for the
Methoxycarbonylation of Medium-Chain Alkenes
Cedric Holzapfel and Tyler Bredenkamp*[a]
The methoxycarbonylation reaction provides a route to the
synthesis of esters from medium-chain alkenes that may be
used as fuel supplements. However, the known productive cat-
alytic systems are expensive and/or unstable at elevated tem-
peratures. Most of the data available on the methoxycarbony-
lation of alkenes is derived from ethylene and styrene as sub-
strates. To broaden the scope, we conducted a comparative
study of a range of phosphine ligands under comparable con-
ditions for the methoxycarbonylation of 1-octene. The results
demonstrate that a number of ligand structural motifs facilitate
the process effectually. Furthermore, the critical importance of
Scheme 2. Proposed mechanism for the hydromethoxycarbonylation reac-
tion.
alkene isomerization and the acid/ligand and Pd/ligand ratios
are presented.
Carbonylation reactions provide an efficient route to generate
a variety of functional groups from CÀC unsaturation by using
readily available carbon monoxide and other ancillary re-
agents.[1] Amongst others, these metal (Pd, Co, Ni, Rh, etc.)[2]
catalyzed carbonylation reactions include the syntheses of al-
dehydes (hydroformylation),[3] carboxylic acids (hydrocarboxyla-
tion),[4] and esters (hydromethoxycarbonylation or methoxycar-
bonylation for short, Scheme 1).[5]
As South Africa has no natural source of liquid fuel, there is
an opportunity and a need to convert its (Sasol) Fischer–
Tropsch medium-chain olefins into their corresponding methyl
esters as substitutes for diesel fuels. Although a number of cat-
alytic systems have been developed for the required methoxy-
carbonylation, industrial implementation with regard to
medium-chain alkenes has not yet been realized. In addition to
patent protection, the known catalysts suffer from a number
of disadvantages, including high cost. Most often, these sys-
tems comprise Pd2(dba)3 (dba=dibenzylideneacetone) and the
ligands (and their analogues) shown in Figure 1.[8,9] The use of
ligand A in the methoxycarbonylation of ethylene is a well-
known industrial process (Lucite process).[10i]
Scheme 1. The Pd-catalyzed methoxycarbonylation reaction.
Two mechanisms, namely, PdÀH and PdÀOMe, have been
proposed for the methoxycarbonylation reaction, each with
much study and evidence in its favor.[5b] However, most evi-
dence collected thus far indicates a preference for the PdÀH
mechanism, as shown in Scheme 2.[5c,d] Despite the fact that
long-chain esters play a crucial role in the manufacture of lu-
bricants and surfactants[6] and are eligible candidates for petro-
leum derivatives, these higher-carbon linear aliphatic esters are
typically produced by further manipulations of the products of
the hydroformylation reaction.[7] However, much effort is being
devoted to catalytic methoxycarbonylation with respect to the
utilization of renewable resources.[8]
Figure 1. Bulky diphosphine ligands for the methoxycarbonylation of al-
kenes.
The requirements for a potential industrial catalyst would in-
clude low cost, efficiency in terms of reaction rates and selec-
tivity, high temperature stability, and the capability of recy-
cling. The development of a catalyst system of this kind can be
expedited by a better understanding of the critical factors, in
particular, ligand structure and reaction conditions, that affect
catalyst efficiency and productivity.
[a] Prof. C. Holzapfel, Dr. T. Bredenkamp
Research Centre for Synthesis and Catalysis
Department of Chemistry
University of Johannesburg
PO Box 524, Auckland Park, 2006 (South Africa)
E-mail: tylerb@uj.ac.za
ChemCatChem 2015, 7, 2598 – 2606
2598
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communications
The identification of these factors is difficult on the basis of
the published literature. Most of these studies are typically per-
formed under different conditions, which makes direct com-
parisons impossible.[7] Additionally, a large majority of reports
on methoxycarbonylation have been limited to ethylene and
styrene.[7,9,10] As expected, the same difficulty pertains to com-
paring ligands in the hydroformylation reaction,[11] a subject
presently under investigation in our laboratory. We now report
a comparative study of the methoxycarbonylation of medium-
chain alkenes by using commercially available ligands under
comparable conditions. We believe that this study contributes
to the identification of the above-mentioned critical parame-
ters.
with cost considerations, this was replaced by Pd(OAc)2 (cheap-
er per Pd equivalent) with the added benefit of a reduction in
induction period. The larger induction period noted with
Pd2(dba)3 may be due to the slow displacement of the dba
ligand from Pd.[12] Alternatively, it could involve the oxidation
of Pd0 to PdII as a precatalytic step. PdCl2 could also be used as
the source of Pd. However, the reaction rates are reduced.
By employing 1-octene as the substrate, initial studies were
performed by using monodentate ligands that required ligand/
Pd ratios ꢀ8:1 to ensure catalyst stability. Reactions were per-
formed under the conditions shown in Table 1. However, some
catalyst complexes were inactive (Table 1, entries 4–6, 10, and
11), and for those cases in which activity was observed, low
turnover frequencies (TOFs) were obtained.
In the hope of obtaining higher reaction rates, attention was
then directed to bidentate ligands. In the interest of cost re-
duction, initial studies were performed by using a Pd/ligand
ratio of 1:2 in contrast to the 1:5 ratio used for the benchmark
ligand [i.e., 1,2-bis(di-tert-butylphosphinomethyl)benzene, li-
gand A].[9a] However, some catalysts required a higher Pd/
ligand ratio (Table 2, entries 5
A wide range of phosphine ligands (including monodentate
and bidentate) were selected for evaluation in the Pd-catalyzed
methoxycarbonylation of 1-octene (Figure 2). The selection
was chosen to span a range of steric and electronic properties.
The known published[8,9] catalytic systems that provide rea-
sonable rates of reaction most often make use of expensive
and air-sensitive Pd2(dba)3 as the source of Pd. In compliance
and 18).
The results obtained for the
bidentate ligands are shown in
Table 2. Results for ligands F
(bite angle 858)[13] and G (bite
angle 918)[13] are omitted, as
these yielded other products in
addition to esters.
In line with the literature,[9a]
employing ligand A (provided
that a Pd/ligand ratio of 1:5 was
used) furnished excellent results,
and near-quantitative conversion
to the linear ester product was
observed. Ligand A and some
structurally related compounds
(see Figure 1) remain unique in
their ability to furnish linear
products. All other catalysts eval-
uated in this investigation gave
lower
linear/branched
(l/b)
ratios. However, this is irrelevant
in the context of liquid fuels. In
addition, in terms of our stated
objectives ligand A suffers from
catalyst decomposition at lower
Pd/ligand ratios (Table 2, com-
pare entries 1–3). Notably, with
the exception of ligands A, B,
and P, 2 equivalents of added
ligand was sufficient to maintain
catalyst stability under the reac-
tion conditions.
The results indicate that mod-
erate to good productivities
were obtained by using relative-
Figure 2. Phosphine ligands for the methoxycarbonylation of 1-octene.
ChemCatChem 2015, 7, 2598 – 2606
www.chemcatchem.org
2599
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communications
groups.[13c] In addition, cationic Pd complexes of this conforma-
tion may be locked in by coordination of Fe to Pd.[15] The same
situation may pertain to the corresponding cationic complex
of ligand L with O replacing Fe. As the results suggest (Table 2,
entries 14, 16–18), in the absence of the extreme steric require-
ments of ligand N, the Pd complexes may be able to adopt
the required cis conformation. This is supported by the obser-
vation that ligand O bearing the less bulky cyclohexyl group
(Table 2, entry 16) affords a modicum of activity in the reaction.
This rationalization is based on the accepted mechanism for
the methanolysis of Pd–acyl species, which is the rate-deter-
mining step[9a,10,13c] requiring cis coordination of the phosphine
ligands (Scheme 3).[13c]
Table 1. Methoxycarbonylation of 1-octene by using monodentate
ligands.[a]
Entry
Ligand
Pd/L/acid
Conversion
[%][b]
TOF[c]
l/b
1
2
3
4
5
6
7
8
A
1
2
3
4
5
6
7
8
9
1:5:10
1:8:10
1:8:10
1:8:10
1:8:10
1:8:10
1:8:10
1:8:10
1:8:10
1:8:10
98
49
50
n.r.
n.r.
n.r.
17
52
n.r.
n.r.
600
140
100
–
–
–
85
105
–
98
74
80
–
–
–
73
77
–
10
11
–
–
[a] Reaction conditions: Pd(OAc)2 (0.045 g, 0.2 mmol), ligand, MsOH
(130 mL, 2 mmol), 1-octene (20 mL), methanol (100 mL), 808C, CO
(3.0 MPa), 3 h. n.r.=no reaction. [b] As determined by GC-FID analysis
(esters were the only products observed); toluene (2 mL) was used as an
internal standard. [c] TOF=mol CO absorbed/mol Pd/h, based on the
linear part (first 20–30 min) of the CO uptake curve (average of two runs).
Scheme 3. Methanolysis of the Pd–acyl species.
Table 2. Methoxycarbonylation of 1-octene by using bidentate ligands.[a]
With regard to chemoselectivity (esters vs. oligo-
Entry
Ligand[14]
Pd/L/acid
Conversion
[%][b]
TOF[c]
Bite angle
[8]
l/b
mers), the results also indicate that the steric require-
ments of the ligands may be of lesser importance in
1
2
3
4
5
6
7
8
A
A
A
B
B
C
D
E
H
I
J
K
L
1:2:10
1:4:10
1:5:10
1:2:10
1:5:10
1:2:10
1:2:10
1:2:10
1:2:10
1:2:10
1:2:10
1:2:10
1:2:10
1:2:10
1:2:10
1:2:10
1:2:10
1:4:10
52[d]
77[d]
98
252
550
600
399
65
74
65
297
241
46
–
594
–
362
n.r.
46
1220
1160
104[9c]
104
104
98
98
92
92[13b]
ꢁ96
98
97
our case in comparison to the methoxycarbonylation
of ethylene. Both M[13e] and K[13c] yield oligomers in
the case of ethylene. However, in the case of 1-
octene only ester products were observed. This indi-
cates that the relative size of the alkene contributes
to the total steric environment of the Pd–acyl com-
plex. For the carbonylation of ethylene, increasing
steric bulk of the bidentate ligands initially increases
the rate of insertion, which gives more productive
polymerization catalysts. Beyond a certain alkene
size, alkene insertion becomes unfavorable and only
carbon monoxide or alcohol can coordinate cis to an
acyl group; this accelerates elimination and thus
makes the bulkiest ligands the best choice for the
production of methyl propanoate (MeP).[10a] In the
case of octene, alkene insertion is undesirable owing
to its steric contribution to the environment, and
hence, esters are the only products observed.
97
98
77
78
72
72
82
75
78
–
77
–
74
–
87
77
77
65[d]
40
18
21
70
63
12
<5
82
n.r.
84
n.r.
20
84[d]
98
9
10
11
12
13
14
15
16
17
18
111
111
103
ꢁ105
M
N
O
P
P
96
104[13c]
102[13d]
ꢁ100
ꢁ100
[a] Reaction conditions: Pd(OAc)2 (0.045 g, 0.2 mmol), ligand, MsOH (130 mL, 2 mmol),
1-octene (20 mL), methanol (100 mL), 808C, CO (3.0 MPa), 3 h. n.r.=no reaction. [b] As
determined by GC-FID analysis (esters were the only products observed); toluene
(2 mL) was used an as internal standard. [c] TOF=mol CO absorbed/mol Pd/h, based
on the linear part (first 20–30 min) of the CO uptake curve (average of two runs).
[d] Pd black was observed.
The efficacy of benchmark ligand A has been large-
ly attributed to steric factors.[10a] However, this work
indicates that flexibility and electronic properties of
the ligand may be of equal or even greater impor-
ly flexible bidentate ligands with moderate to large bite
angles. Comparing structurally related ligands I, J, K, and L
(Table 2, compare entries 10–13) revealed that the Xantphos-
type ligands (i.e., I and J) offered very poor rates, whereas K
offered a rate similar to that of A. According to van Leeu-
wen,[13c] this is due to the fact that Xantphos ligands form inac-
tive trans complexes in the catalytic cycle.
tance. The importance of the electronic properties of the
ligand is demonstrated upon comparing ligands D and E
(Table 2, compare entries 7 and 8). They have approximately
the same steric properties, yet electron-rich analogue E pro-
vides the better rate. Comparing ligands K and L shows that
electron-poor diphosphine K affords higher conversions. The
reversal in the electronic effects of these two types of ligands
may be due to the fact that ligands D and E are bidentate,
whereas K and L can be (hemilabile) tridentate. For the pur-
pose of this study, on the basis of the work of van Leeu-
Similarly, 1,1-bis(di-tert-butylphosphino)ferrocene (N) was
not active in the reaction, as it is unable to adopt the required
cis configuration as a result of the steric repulsion of the tBu
ChemCatChem 2015, 7, 2598 – 2606
www.chemcatchem.org
2600
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communications
wen[13c,15b] and Braunstein,[16] we suggest that ligands I–P have
the ability to act as hemilabile tridentate ligands.
A noteworthy result of the comparative study was the pro-
ductivity obtained with the unsymmetrical bidentate phos-
phine 1-(diphenylphosphino)ferrocene-1-(di-tert-butylphosphi-
no)ferrocene (ligand P; Table 2, entry 18), which clearly outper-
formed the other ligands (Figure 3). This ligand was first de-
Figure 5. Effective unsymmetrical bidentate ligands for the methoxycarbony-
lation of ethylene.
onstration of the efficacy of unsymmetrical ligands was pre-
sented the following year, for which Pringle[20d] et al. conducted
a study by using a variety of unsymmetrical phosphines based
on o-diphosphinoxylenes. Employing unsymmetrical ligand T
(Figure 5) in the methoxycarbonylation of ethylene afforded
unparalleled rates for the formation MeP, and this ligand out-
performed all other ligands evaluated. However, regardless of
substituents on this particular backbone, it is always unsym-
metrical.
Synergistic effects may play a significant role in the final
methanolysis step of the reaction. This step can be formally
considered a reductive elimination[10a] (not necessarily the clas-
sical mechanism) step as in the seminal publication of van
Leeuwen.[13c] Relatively few Pd catalysts containing unsymmet-
rical bidentate ligands have been described to be involved in
reductive elimination. However, the few known examples indi-
cate that unsymmetrical ligands can significantly enhance reac-
tion rates.[22] Tongi[22c] and co-workers, for example, demon-
strated a highly active system comprising Pd(OAc)2, unsymmet-
rical ligand U (R1 =3,5-(F3C)2C6H3, R2 =cyclohexyl) and BF3·OEt2
for the copolymerization of carbon monoxide with propene to
afford a regioregular and stereo-
Figure 3. Amount of CO absorbed after 20 min (based on the CO uptake
curve, moles CO absorbed=esters formed) for selected active bidentate
phosphine ligands. Pd/ligand/acid=1:2:10 for ligands K and M; Pd/ligand/
acid=1:5:10 for ligand A, and Pd/ligand/acid=1:4:10 for ligand P.
scribed by Cullen[17] et al. and Butler[18] et al. in 1985 and has
received attention in other contexts.[15,19] To the best of our
knowledge, this is the first time this ligand has been employed
for the methoxycarbonylation reaction. Remarkably, corre-
sponding symmetrical analogues M and N are less reactive and
unreactive, respectively (Table 2, compare entries 14, 15, and
18).
regular isotactic copolymer.
These results can be compared with those obtained by Prin-
gle[20a] et al. by using ligands A, Q, and R (Figure 4) for the me-
thoxycarbonylation of ethylene. The first two were very active
The high electronic differentia-
tion between the two phosphine
atoms was posited as the reason
for the high activity. It is believed
that the electron-withdrawing aryl-
phosphine decreases the electron
density at the Pd center, which
leads to a higher rate of reaction.
Hartwig[22d] used the same backbone (R1 =Ph or cyclohexyl,
R2 =tBu) for the Pd-catalyzed cross-coupling of aryl tosylates
with Grignard reagents. That work demonstrated that the
more effective system was that with R1 =Ph.
Figure 4. Bis(di-tert-butylphosphinomethyl)benzene analogues for the me-
thoxycarbonylation of ethylene.
for the production of methyl propionate, and the latter result-
ed in a sluggish reaction rate and afforded ethene/CO co-olig-
omers and polyketone.[10j,20] Notably, the Pd catalyst based on
ligand Q had a longer lifetime than the Pd complex derived
from ligand A.[20a] This is another demonstration that both
steric and electronic properties of the ligand are important.
The electronic differences between ligands M, N, and P were
demonstrated in electrochemical studies.[19d,21]
Later, Pringle’s[20b] group demonstrated that unsymmetrical
ligand S (Figure 5) also outperformed its symmetrical ana-
logues in the methoxycarbonylation of ethylene. In this case,
the electron-poorer phosphine also has an extreme steric re-
quirement that favors the formation of MeP.[10a] A further dem-
The efficacy of these unsymmetrical ligands is difficult to ra-
tionalize. However, it is tempting to associate the effectiveness
of these unsymmetrical ligands (containing both a good elec-
tron donor and a good electron acceptor) to the captodative
effect observed in the stabilization of radicals[23] and in cyclo-
addition reactions.[24] For the methoxycarbonylation reaction, it
is apparent that an effective catalyst should contain one bulky
electron-donating group and one bulky electron-withdrawing
group. Notably, the efficacy of unsymmetrical ligands has also
been demonstrated in the Ni-catalyzed amination of chloro-
benzene.[25]
A significant observation is that the initial rates for ligands A
and K are almost identical (Table 2, compare entries 3 and 12).
ChemCatChem 2015, 7, 2598 – 2606
www.chemcatchem.org
2601
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communications
However, the conversion after 3 h is noticeably lower in the
case of K. This difference could not be ascribed to differences
in catalyst stability. Following the progress of the reaction by
GC analysis for the two ligands showed substantial isomeriza-
tion of the 1-alkene into internal isomers. With K, virtually all
of 1-octene was converted into internal olefins at approximate-
ly 40% conversion to esters, and in the case of A less than
20% conversion of 1-octene into esters occurred before com-
plete isomerization of 1-octene into internal isomers. Isomeriz-
ing 1-octene at a pressure of 3.0 MPa He under otherwise iden-
tical conditions used in the methoxycarbonylation revealed
that in the case of K, isomerization was slower than in the case
of A (Figure 6). At approximately 40 min the amount of 1-
Figure 8. Amount of CO absorbed for the methoxycarbonylation of 1-octene
by using Pd complexes derived from ligands A and K (moles CO absorb-
ed=moles esters formed).
fins. In the case of A, the isomerization was faster, yet the me-
thoxycarbonylation rate was not greatly affected. Given that
both reactions proceed to completion, the isomerization must
be reversible in both cases. Clearly, the rate of isomerization in
either direction is greater for ligand A than for ligand K.
The difference in catalyst efficiency was also demonstrated if
a mixture of internal octene isomers was carbonylated by
using the two different catalytic systems (Figure 9).
Figure 6. Isomerization of 1-octene under 3.0 MPa He pressure.
octene is minute in the case of ligand K. The corresponding
GC traces for A are shown in Figure 7. Clearly, the isomeriza-
tion of 1-alkenes into internal isomers plays a significant role in
the methoxycarbonylation of these alkenes. The role of isomer-
ization in the methoxycarbonylation of internal alkenes has re-
ceived separate attention.[8a–e]
Figure 9. Amount of CO absorbed for 1-octene versus internal octene iso-
mers by employing ligands A and K (moles CO absorbed=moles esters
formed). Isomerization mixture formed by using conditions described in
Table 2, except for the use of a 3.0 MPa He atmosphere. Exact composition
of octene isomers was not determined.
Similar results were obtained upon comparing 3-octene and
5-decene. Interestingly, in both cases ester products identical
to those obtained with the corresponding 1-alkenes were
formed, whereas no intermediate alkene isomers were ob-
served by GC analysis. This is in line with the elegant mecha-
nistic studies of Walther[8c] et al. that showed low energy barri-
ers in either direction for the isomerization of alkenes if using
A as ligand. Similarly, the only other catalyst compared to the
benchmark catalyst that offered a high rate and conversion, li-
gand P, showed a methoxycarbonylation rate essentially unaf-
fected by isomerization.
Figure 7. GC traces corresponding to octene isomers formed by using the
Pd complex derived from ligand A under 3.0 MPa He.
Against this background, the plot for CO uptake for these
two ligands is demonstrated in Figure 8. Starting at essentially
identical rates, A surpasses K after approximately 35 min, at
which stage 1-octene is essentially converted into internal ole-
In the case of ligands A and P, methoxycarbonylation is
seemingly determined by the rate of methanolysis of the Pd–
acyl species throughout the reaction. In other cases, the rate-
determining step of the methoxycarbonylation may shift to
ChemCatChem 2015, 7, 2598 – 2606
www.chemcatchem.org
2602
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communications
the rate of isomerization. It appears that a requirement for
a good methoxycarbonylation catalyst for longer-chain alkenes,
in addition to being efficient in the final methanolysis step, is
that it should also be a good isomerization catalyst.
The role of the acid co-catalyst has been debated for some
time.[7,26] It is undoubtedly involved in the formation of the
active PdÀH species. A definite requirement is that a balance
between the molar amounts of added ligand and acid must be
maintained, given that the protonation of the ligand decreases
its availability for complexation with Pd, which results in cata-
lyst instability. This may explain the failure of basic ligand A to
support the methoxycarbonylation if the acid/ligand ratio is
higher than 2:1 (Table 3, entries 2 and 3). Though not shown in
Figure 10. Amount of CO absorbed for the methoxycarbonylation of 1-
octene by using different acid concentrations by employing dppf [1,1’-bis(di-
phenylphosphino)ferrocene] as the ligand (moles CO absorbed=moles
esters formed).
Table 3. The effect of acid concentration on the rate of the methoxycar-
bonylation of 1-octene by using selected bidentate ligands.[a]
obtained with other strong Brønsted acids [e.g., trifluorome-
thanesulfonic acid (TfOH) and anhydrous HBF4 (see the Experi-
mental Section)]. However, in the interest of cost efficiency,
methanesulfonic acid (MsOH) was selected as the acid of
choice. As expected,[27] further rate enhancements (ꢀ50%)
could also be achieved by replacement of the Brønsted acid
with the Lewis acid Al(OTf)3. The enhanced catalytic effect of
Al(OTf)3 and other Lewis acids in this reaction[27] and in other
reactions of oxygenate electrophiles is well known.[28]
Entry
Ligand[14]
Pd/L
Acid
TOF[b]
Conversion[c]
[%]
[equiv.]
1
2
3
4
5
6
7
8
A
A
A
K
K
K
1:5
1:5
1:5
1:2
1:2
1:2
1:2
1:2
1:2
1:2
1:2
1:4
1:4
1:4
1:4
10
20
30
3
10
20
30
3
10
20
30
3
10
20
30
600
510
n.d.
316
594
743
920
204
362
483
594
956
1160
1189
1260
98 (34)
85[d]
<10[d]
80 (17)
82 (34)
85 (41)
90 (51)
80 (11)
84 (20)
89 (26)
96 (33)
90 (51)
98 (64)
97 (68)
97 (70)
K
M
M
M
M
P
P
P
P
The dramatic effect of acids on the rate, particularly in the
early stages of the reaction for which little isomerization of the
alkene has occurred, may at least in part be ascribed to the
promotion of PdÀH formation, but its direct participation in
the final methanolysis step (also believed to be the rate-limit-
ing step)[9a,10a,13c] cannot be excluded. A possible mechanistic
explanation for this rate enhancement is shown Scheme 4. This
is in line with the stepwise mechanism for methanolysis, as
suggested by Walther[8c] et al. but with the added promotional
role of a proton.
9
10
11
12
13
14
15
[a] Reaction conditions: Pd(OAc)2 (0.2 mmol), ligand, MsOH, 1-octene
(20 mL), MeOH (100 mL), 808C internal temperature, 3 h. n.d.=not deter-
mined. [b] TOF calculated based on the first 20 min of the reaction (aver-
age of two runs); toluene (2 mL) was used as an internal standard. [c] Re-
sults in parentheses are the conversions after 20 min (based on CO
uptake curve at 808C, partial pressure of methanol was not considered
and values are uncorrected). [d] Pd black was observed.
the table, basic ligand E was also sensitive to increases in the
acid concentration, which resulted in catalyst decomposition.
The requirement for catalyst stability is paramount, and it
was therefore necessary to establish the performance of the
promising ligands in the presence of additional acid (Table 3).
In the case of ligand P, an increase in the amount of acid
above 20 equivalents did not offer any significant added bene-
fit (Table 3, compare entries 14 and 15). This could not be at-
tributed to catalyst decomposition at higher acid loadings. The
other ligands evaluated also showed rate enhancements with
an increase in acid concentration even beyond 20 equivalents.
These results show the remarkable stability of the evaluated
catalysts in the presence of additional acid. A noteworthy
point of interest is that high rates and conversions can be ach-
ieved simply by the addition of more acid to cheaper ligand M
(Table 3, entries 8–11 and Figure 10).
Scheme 4. Possible acid promoted mechanism for the final methanolysis
step.
Though not shown in Table 3, the l/b ratio remained un-
changed with increasing acid concentrations and increasing re-
action rates. Furthermore, no changes in the ratios were ob-
served with the different acids employed. This indicates that
the l/b ratio is independent of the rate-determining step.
Increasing the CO pressure from 3.0 to 5.0 MPa resulted in
decreased rates (ꢁ10%) for the ligands (i.e., A, K, M, N, and P)
evaluated in this regard. This is likely due to a competition be-
tween the alkene and CO in which CO suppresses activation of
the alkene and in turn facilitates a coordinatively saturated cat-
alyst complex in the early stages of the catalytic cycle. Addi-
tionally, CO can compete with methanol for coordination to Pd
As in our earlier studies[27a] on the methoxycarbonylation of
alkenes, even better results (up to 30% rate increase) could be
ChemCatChem 2015, 7, 2598 – 2606
www.chemcatchem.org
2603
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communications
in the final stage of the reaction. However, decreasing the
pressure to 1.5 MPa showed little change in the reaction rate.
With cost efficiency in mind, attention was given to the
effect of the concentration of Pd on the reaction rates and
TOF. Preliminary studies indicated that in all cases, including
those of the monodentate ligands, the TOF increased with
lower Pd concentrations. This may be a result of a shift in the
equilibrium away from dimeric resting states of Pd at lower
concentrations.[29] The possible role of Pd nanoparticles in this
effect was considered unlikely in the absence of a demonstra-
ble Tyndall effect. Nevertheless, the role of nanoparticles, as
demonstrated, for example, in the Heck[30] and Suzuki[31] reac-
tions, cannot be excluded without further investigation. How-
ever, in the case of the methoxycarbonylation reaction, nano-
particle involvement could simply involve leeching of Pd from
the particles into solution, as shown in Equation (1):
to limitations of equipment, accurate TOF values could not be
obtained. Catalyst stability under these reaction conditions
was indicated by the observation that after completion of the
first conversion of the alkene into ester, an additional amount
of 1-octene (10–20 mL) was added and subsequently effective-
ly methoxycarbonylated (the formation of Pd black was not
observed in the reactor postreaction). This is a further indica-
tion of the potential for catalyst recycling.
The development of an efficient catalytic system for the me-
thoxycarbonylation of medium-chain alkenes cannot be solely
reliant on the vast amount of knowledge that has been ac-
quired for ethylene and styrene substrates. This is because of
the crucial importance of alkene isomerization in the case of
medium-chain alkenes. Our results demonstrate that alkene
isomerization (in both directions) and methoxycarbonylation
itself are equally important.
In addition, our results indicate that there are a number of
ligand structural types that can be used to improve the effica-
cy of the methoxycarbonylation of medium-chain alkenes.
Some of these structural motifs, as found, for example, in
ligand K, could not support the methoxycarbonylation of eth-
ylene or lower alkenes because of competitive alkene inser-
tions into the intermediate Pd–acyl species. However, increas-
ing the size of the alkene led to exclusive selectivity for ester
products.
þ
Pdn þ 2 L þ Hþ ! PdnÀ1 þ ½HPdL2Š
n ¼ nanoparticles, L ¼ ligand
ð1Þ
From a cost point of view, this could be a significant out-
come. The results of a quantitative study with ligand P are
shown in Figure 11.
The importance of acid concentration as well as the acid/
ligand ratio on the reaction rate and catalyst stability was dem-
onstrated. In addition, the ligands used in this study that con-
ferred both thermal stability and high activity to the Pd-cata-
lyzed methoxycarbonylation reactions can be regarded as
(hemilabile) tridentate ligands. Excellent results in terms of re-
action rates and stability were obtained with the unsymmetri-
cal ligand 1-(diphenylphosphino)ferrocene-1-(di-tert-butylphos-
phino)ferrocene, a ligand that has not previously been used in
the methoxycarbonylation of alkenes. In this regard, we believe
that the potential of unsymmetrical bidentate ligands (contain-
ing one bulky electron-poor phosphine and one bulky elec-
tron-rich phosphine) should receive further attention.
Figure 11. The effect of the Pd concentration on the TOF for the methoxy-
carbonylation of 1-octene.
From an industrial point of view, catalyst recyclability is also
of paramount importance. The preliminary results indicate that
this may be possible for ligands M and P. This is indicated by
the fact that after complete conversion to the esters, additional
alkene (10 mL) was subsequently added (by a pressurized re-
servoir connected to the reactor), and it was also fully convert-
ed into esters under the standard conditions. This important
aspect is currently under further investigation.
Experimental Section
A Hastelloy Parr “Bench Top Reactor” (volume=300 mL, maximum
temperature=2508C, and maximum pressure of 30.0 MPa) pur-
chased from Labotec was used during this study. The reactor was
fitted with a mechanical overhead stirrer and its own heating
mantle. The reactor was equipped with a “nonreturn” valve, which
prevented the reaction mixture from pushing back into the pipes.
This valve had a backpressure of 2 kPa. The reactor was also fitted
with a safety valve that was set at a max pressure of 12.0 MPa. A
80 mL gas burette was fitted to the reactor to give a total volume
of 380 mL.
In the fine-chemical industry, reaction temperatures are criti-
cal and lower temperatures are favored. However, the heavy-
chemical industry often generates large amounts of energy,
which can be utilized to drive other energy-intensive opera-
tions. For this reason, the stability of the catalytic system at
higher temperatures is critical for potential applications in the
production of diesel fuels. With ligands M and P, catalyst stabil-
ity up to 1208C (conditions as described in Table 2) was at-
tained with complete conversion into esters at enhanced rates
(the formation of Pd black was not observed). However, owing
A typical experiment involved the addition of a Pd source and
specified amounts of ligand and acid to the reactor. The substrate
(20 mL) and MeOH (100 mL) were added to the reactor. The reactor
was flushed with CO (3”) to remove any remaining air and was
pressurized to 3.0 MPa CO at room temperature. The mixture was
then heated to the specified temperature and was stirred at
600 rpm with a mechanical stirrer device.
ChemCatChem 2015, 7, 2598 – 2606
www.chemcatchem.org
2604
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communications
Martin, A. Kçckritz, J. Catal. 2013, 297, 44–55; d) J. T. Christl, P. R. Roelse,
F. Stemfle, P. Wucher, I. Gottker-Scnetmann, G. Müller, S. Mecking, Chem.
Eur. J. 2013, 19, 17131–17140; e) P. Roesle, C. J. Dürr, H. M. Mçller, L.
Cavallo, L. Caporaso, S. Mecking, J. Am. Chem. Soc. 2012, 134, 17696–
17703; f) G. Walther, J. Deutsch, A. Martin, F. E. Baumann, D. Fridag, R.
Franke, A. Kçckritz, ChemSusChem 2011, 4, 1052–1054; g) S. Chikkali, F.
Stempfle, S. Mecking, Macromol. Rapid Commun. 2012, 33, 1126–1129;
see reference [20].
It is recognized that the partial pressure of methanol may play
a role in the total pressure at a given temperature. This contribu-
tion would be approximately the same in all reactions and would
remain more or less consistent during each run. In all cases, semi-
quantitative data were used in determining the TOF values and
the partial pressure of methanol was not considered. The TOF
values were calculated on the basis of the linear part of the CO
uptake curve during the initial part of the reaction (first 20–
30 min). Regression analysis of the linear section of the CO curve
provided R2 >0.99 in all cases.
[9] a) C. Jimenez Rodriguez, D. F. Foster, G. R. Eastman, D. J. Cole-Hamilton,
Chem. Commun. 2004, 1720–1721; b) R. I. Pugh, E. Drent, P. G. Pringle,
Chem. Commun. 2001, 1476–1477; c) W. Clegg, G. R. Eastham, M. R. J.
Elsegood, R. P. Tooze, X. L. Wang, K. Whiston, Chem. Commun. 1999,
1877–1878.
With the exception of 1-octene (obtained from Sasol R&D) and
HBF4 (see below), all other chemicals were purchased from Sigma–
Aldrich with no further purification. Anhydrous HBF4 was prepared
by adding NaHBF4(s) (1 g) to MeOH (8 mL) at room temperature.
The solution was subsequently cooled to À108C whilst bubbling
HCl(g) through the solution to precipitate the resulting NaCl(s). The
precipitate was filtered under an argon atmosphere, and the solu-
tion was dried under reduced pressure to afford anhydrous HBF4 in
87% yield.
[10] a) L. Kóllµr, Modern Carbonylation Methods, Wiley-VCH, Weinheim, Ger-
many, 2008 and references therein; b) E. Drent, P. H. M. Budzelaar,
Chem. Rev. 1996, 96, 663–681; c) R. A. M. Robertson, D. J. Cole-Hamil-
ton, in Catalytic Synthesis of Alkene-Carbon Monoxide Copolymers and
Co-oligomers (Ed.: A. Sen), Kluwer Academic Publishers, Dordrecht, The
Netherlands, 2003 and references therein; d) I. R. Butler, P. K. Baker, G. R.
Eastman, K. M. Fortune, P. N. Horton, M. B. Hursthouse, Inorg. Chem.
Commun. 2004, 7, 1049–1052; e) A. Balµzs, C. Benedek, S. Tçrçs, J. Mol.
Catal. A 2006, 244, 105–106; f) C. Bianchini, A. Meli, W. Oberhauser, Or-
ganometallics 2003, 22, 2409–2421; g) W. Mägerlein, A. F. Indolese, M.
Beller, J. Organomet. Chem. 2002, 641, 30–40; h) I. del Río, P. W. N. M.
van Leeuwen, Eur. J. Inorg. Chem. 2001, 2719–2738; i) T. Fanjul, G. R
Eastman, P. G. Pringle, M. Waugh (Lucite Int. UK Ltd), WO2010001174,
2010; j) O. E. Sielcken, F. P. W. Agterberg, N. F. Haasen (DSM N. V. E. I Du
Pont De Nemours and company), US Patent 5495041, 1996; see refer-
ence [20].
GC analysis was used to routinely monitor the progress and out-
come of the reactions. Accordingly, a small aliquot of the reaction
mixture was filtered through silica gel and analyzed by using a gas
chromatograph (Shimadzu GC 2010). A flame-ionization detector
(FID) was used and was maintained at 3008C, with the following
gas flows: make up flow 30 mLminÀ1, H2 flow 40 mLminÀ1, and air
flow 400 mLminÀ1
.
[11] P. W. N. M. van Leeuwen, Homogeneous Catalysis: Understanding the art,
Kluwer Academic Publishers, Dordrecht, The Netherlands, 2004,
pp. 145.
[12] P. W. N. M. van Leeuwen, J. C. Chadwick, Homogeneous Catalysts, Wiley-
VCH, Weinheim, Germany, 2011, pp. 279–346.
Acknowledgements
[13] a) P. W. N. M. van Leeuwen, P. C. J. Kamer, J. N. H. Reek, P. Dierkes, Chem.
Rev. 2000, 100, 2741–2769; b) M. Ogasawara, K. Yoshida, T. Hayashi, Or-
ganometallics 2000, 19, 1567–1571; c) P. W. N. M. van Leeuwen, M. A.
Zuideveld, B. H. G. Swennenhuis, Z. Freixa, P. C. J. Kramer, K. Goubitz, J.
Fraanje, M. Lutz, A. L. Spek, J. Am. Chem. Soc. 2003, 125, 5523–5539;
d) L. E. Hagopian, A. N. Campbell, J. A. Golen, A. L. Rheingold, C. Naturo,
J. Organomet. Chem. 2006, 691, 4890–4900; e) O. V. Gusev, A. M. Kaslin,
M. G. Peterleiter, P. V. Petrovskii, K. A. M. G. Lyssenko, N. G. Akhmedov, C.
Bianchini, A. Meli, W. Oberhauser, Organometallics 2002, 21, 3637–
3649.
We thank Sasol, the National Research Foundation, and the Uni-
versity of Johannesburg for financial assistance with this project.
Keywords: alkenes · captodative · carbonylation · diesel ·
isomerization
[1] a) L. Wang, Y. Wang, C. Liu, A. Lei, Angew. Chem. Int. Ed. 2014, 53, 5657–
5661; Angew. Chem. 2014, 126, 5763–5767; b) H. M. Colquhoun, D. J.
Thompson, M. V. Twigg, Carbonylation—Direct Synthesis of Carbonyl
Compounds, Plenum Press, New York, 1991; c) New Synthesis with
Carbon Monoxide (Ed.: J. Falbe), Springer, Berlin, 1980; d) A. L. Lapidus,
S. D. Pirozhkov, Russ. Chem. Rev. 1989, 58, 117–137.
[14] B=(R,R)-DIOP, (À)-2,3-O-Isopropylidene-2,3-dihydroxy-1,4-bis(diphenyl-
phosphino)butane; C=BINAP, 2,2’-bis (diphenylphosphino)-1,1’-binaph-
thalene; D=BIPHEP, {2-[2-(diphenylphosphanyl)phenyl]phenyl}diphenyl-
phosphane;
E=dicyclohexyl-({2-[2-(dicyclohexylphosphanyl)phenyl]-
phenyl})-phosphane; F=dppe, 1,2-bis-(diphenylphosphino)ethane, G=
dppp, 1,3-bis(diphenylphosphino)propane; H=dppb, 1,4-bis(di-phenyl-
phosphino)butane; I=Xanthphos, 4,5-bis(di-phenylphosphino)-9,9-di-
methylxanthene; J=tBuXantphos, 9,9-Dimethyl-4,5-bis(di-tert-butyl-
phosphino)-xanthene; K=DPEPHos, (oxydi-2,1phenylene)bis(diphenyl-
[2] a) M. Beller, B. Cornils, C. D. Frohning, C. W. Kohlpaintner, J. Mol. Catal. A
˝
1995, 104, 17–85; b) P. Pongrµcz, G. Petocz, M. Shaw, D. B. G. Williams,
L. Kollµr, J. Organomet. Chem. 2010, 695, 2381–2384; c) G. Petçcz, G.
Rangits, M. Shaw, H. de Bod, D. B. G. Williams, L. Kollµr, J. Organomet.
Chem. 2009, 694, 219–222.
phosphine);
L=dicyclohexyl({2-[2(dicyclohexylphosphanyl)phenoxy]-
phenyl})-phosphane; M=dppf, 1,1’-bis(diphenylphos-phino)ferrocene;
N=dtbf, 1,1’-Bis(di-tert-butyl phosphino) ferrocene O=1,1’bis(dicyclo-
hexyl phosphino)ferrocene.
[3] a) M. Vilches-Herrera, L. Domke, A. Bçrner, ACS Catalysis 2014, 4, 1706–
1724; b) C. L. Joe, T. P. Blaisdell, A. F. Geoghan, K. L. Tan, J. Am. Chem.
Soc. 2014, 136, 8556–8559.
[15] a) K. M. Gramigna, J. V. Oria, C. L. Mandell, M. A. Tiedemann, W. G.
Dougherty, N. A. Piro, W. S. Kassel, B. C. Chan, P. L. Diaconescu, C.
Nataro, Organometallics 2013, 32, 5966–5979; b) M. A. Zuideveld,
B. H. G. Swennenhuis, P. C. J. Kramer, P. W. N. M. van Leeuwen, J. Organo-
met. Chem. 2001, 637–639, 805–808.
[16] P. Braunstein, F. Naud, Angew. Chem. Int. Ed. 2001, 40, 680–699; Angew.
Chem. 2001, 113, 702–722.
[17] W. R. Cullen, T. J. Kim, F. W. B. Einstein, T. Jones, Organometallics 1985, 4,
346–351.
[4] P. Pino, F. Piacenti, M. Bianchi, I. Wender, Organic Synthesis via Metal Car-
bonyls, Wiley, New York, 1968.
[5] a) V. Elango, M. A. Murphy, G. N. Mott, E. G. Zey, B. L. Smith, G. L. Moss
(Hoechst Celanse), EP 400892, 1990; b) G. Cavinato, L. Toniolo, A. Vava-
sori, J. Mol. Catal. A 2004, 219, 233–240; c) R. P. Tooze, K. Whiston, A. P.
Malyan, M. J. Taylor, N. W. Wilson, J. Chem. Soc. Dalton Trans. 2000,
3441–3444; d) G. R. Eastham, R. P. Tooze, M. Kilner, D. F. Foster, D. J. Co-
le Hamilton, J. Chem. Soc. Dalton Trans. 2002, 1613–1617.
[6] T. O. Vieira, M. J. Green, H. Alper, Org. Lett. 2006, 8, 6143–6145.
[7] G. Kiss, Chem. Rev. 2001, 101, 3435–3456, and references cited therein.
[8] a) G. Walther, A. Martin, A. Kçckritz, J. Am. Oil Chem. Soc. 2013, 90, 141–
145; b) M. R. L. Furst, R. Le Goff, D. Quinzler, S. Mecking, C. H. Botting,
D. J. Cole-Hamilton, Green Chem. 2012, 14, 472–477; c) G. Walther, L. R.
[18] I. R. Butler, W. R. Cullen, T. J. Kim, S. J. Rettig, J. Trotter, Organometallics
1985, 4, 972–980.
[19] a) P. Ste
˘pnicˇka, I. Císarˇovµ, J. Schulz, Organometallics 2011, 30, 4393–
ˇ
˘
ˇ
ˇ
4403; b) P. Stepnicka, K. Skoch, I. Císarovµ, Organometallics 2013, 32,
ˇ
´
ˇ
˘
ˇ
Knçpke, J. Reekah, M. P. Che˛cinski, H. Jiao, U. Bentrup, A. Brückner, A.
623–635; c) K. Skoch, I. Císarovµ, P. Stepnicka, Inorg. Chem. 2014, 53,
ChemCatChem 2015, 7, 2598 – 2606
www.chemcatchem.org
2605
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communications
568–577; d) S. L. Kahn, M. K. Breheney, S. L. Martinak, S. M. Fosbenner,
A. R. Seibert, W. S. Kassel, W. G. Dougherty, C. Nataro, Organometallics
2009, 28, 2119–2126; e) A. Acosta-Ramírez, M. Munoz-Hernandez, W. D.
Jones, J. J. Garcia, Organometallics 2007, 26, 5766–5769.
[25] R. A. Green, J. F. Hartwig, Angew. Chem. Int. Ed. 2015, 54, 3768–3772;
Angew. Chem. 2015, 127, 3839–3843.
[26] a) D. B. G. Williams, T. M. Bredenkamp, ChemCatChem 2012, 4, 206–208;
b) I. del Río, N. Ruiz, C. Claver, Inorg. Chem. Commun. 2000, 3, 166–168;
c) E. Drent, P. Arnoldy, P. H. M. Budzelaar, J. Organomet. Chem. 1994,
475, 57–63; d) S. Oi, M. Nomura, T. Aiko, Y. Inoue, J. Mol. Catal. A 1997,
115, 289–295; e) M. Rosales, I. Pacheco, J. Medina, J. Fernµndez, . Gon-
zµlez, R. Izquierdo, L. G. Melean, P. J. Baricelli, Catal. Lett. 2014, 144,
1717–1727.
[27] a) D. B. G. Williams, M. L. Shaw, M. J. Green, C. W. Holzapfel, Angew.
Chem. 2008, 120, 570–573; b) J. Yang, Y. Yuan, Catal. Lett. 2009, 131,
643–648.
[20] a) T. Fanjul, G. Eastman, N. Fey, A. Hamilton, A. G. Orpen, P. G. Pringle,
M. Waugh, Organometallics 2010, 29, 2292–2305; b) T. Fanjul, G. R. East-
ham, M. F. Haddow, A. Hamilton, P. G. Pringle, A. G. Orpen, T. P. W.
Turner, M. Waugh, Catal. Sci. Technol. 2012, 2, 937–950; c) V. de La
Fuente, M. Waugh, G. R. Eastham, J. A. Iggo, S. Castillon, C. Claver,
Chem. Eur. J. 2010, 16, 6919–6931; d) T. Fanjul, G. Eastman, J. Floure,
S. J. K. Forrest, M. F. Haddow, A. Hamilton, P. G. Pringle, A. G. Orpen, M.
Waugh, Dalton Trans. 2013, 42, 100–115.
[28] Acid Catalysis in Modern Organic synthesis (Eds.: H. Yamamoto, K. Ishi-
hara), Wiley-VCH, Weinheim, Germany, 2008, pp. 109–500 and referen-
ces cited therein.
[29] P. W. N. M. van Leeuwen, Appl. Catal. A 2001, 212, 61–81.
[30] a) M. T. Reetz, J. G. de Vries, Chem. Commun. 2004, 1559–1563; b) M. T.
Reetz, E. Westermann, Angew. Chem. Int. Ed. 2000, 39, 165–168; Angew.
Chem. 2000, 112, 170–173.
[21] M. Trivedi, S. K. Uijain, G. Singh, A. Kumar, S. K. Dubey, N. P. Rath, J. Or-
ganomet. Chem. 2014, 772, 202–209.
[22] a) J. F. Hartwig, Inorg. Chem. 2007, 46, 1936–1947; b) M. Yamashita, J. V.
Cuevas, J. F. Hartwig, J. Am. Chem. Soc. 2003, 125, 16347–16347; c) C.
Gambs, S. Chaloupka, G. Consiglio, A. Tongi, Angew. Chem. Int. Ed. 2000,
39, 2486–2488; Angew. Chem. 2000, 112, 2602–2604; d) A. H. Roy, J. F.
Hartwig, J. Am. Chem. Soc. 2003, 125, 8704–8705; e) R. C. J. Atkinson,
V. C. Gibson, N. J. Long, Chem. Soc. Rev. 2004, 33, 313–328 and referen-
ces therein.
[31] K. Arvela, N. E. Leadbeater, M. S. Sangi, V. A. Williams, P. Granados, R. D.
Singer, J. Org. Chem. 2005, 70, 161–168.
[23] H. G. Viehe, J. Zdenek, R. Merenyi, L. Stella, Acc. Chem. Res. 1985, 18,
148–154.
Received: April 28, 2015
[24] R. L. Domingo, E. Chamorro, P. PØrez, J. Org. Chem. 2008, 73, 4615–
4624.
Revised: May 25, 2015
Published online on July 31, 2015
ChemCatChem 2015, 7, 2598 – 2606
www.chemcatchem.org
2606
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim