Catalyst activity and selectivity in the isomerising alkoxycarbonylation of methyl oleate

Article abstract of DOI:10.1002/chem.201301124

The synthesis of unsymmetrical diphosphine ligands (3 a-g) with an o-tolyl backbone and tert-butyl, adamantyl, cyclohexyl and isopropyl substituents on the phosphorus moiety is described (1,2-(CH₂PR₂) (PR′₂)C₆H<

Full text of DOI:10.1002/chem.201301124

FULL PAPER
DOI: 10.1002/chem.201301124
Catalyst Activity and Selectivity in the Isomerising Alkoxycarbonylation of
Methyl Oleate
Josefine T. Christl, Philipp Roesle, Florian Stempfle, Philipp Wucher,
Inigo Gçttker-Schnetmann, Gerhard Mꢀller,* and Stefan Mecking*[a]
Abstract: The synthesis of unsymmetri-
cal diphosphine ligands (3a–g) with an
o-tolyl backbone and tert-butyl, ada-
mantyl, cyclohexyl and isopropyl sub-
stituents on the phosphorus moiety is
linear a,w-diester (L) with 70–80% se-
lectivity. The products of this catalytic
reaction with the known [{1,2-
complexes 5a–g varied with the steric
bulk of the alkyl substituent on the
phosphorus. Ligands with more bulky
groups, like tert-butyl or adamantyl
(e.g., 5a, 5d, 5g), were more produc-
tive systems. The formation of the cata-
(tBu2PCH2)2C6H4}PdACTHUNRTGNEUNG(OTf)2] complex
(5h) were fully analysed, and revealed
the formation of the linear a,w-diester
(L, 89.0%), the methyl-branched die-
ster B1 (4.3%), the ethyl-branched die-
ster B2 (1.0%), the propyl-branched
diester B3 (0.6%) and all diesters from
butyl- to hexadecyl-branched diesters
B4–B16 (overall 4.8%) at 908C and
20 bar CO. The productivity of the cat-
alytic conversion of methyl oleate with
described
(1,2-(CH2PR2)ACHTNGUTERNN(UG PR’2)C6H4;
3a: R=tBu, R’=tBu, 3b: R=tBu,
R’=Cy, 3c: R=tBu, R’=iPr, 3d: R=
Ad, R’=tBu, 3e: R=Ad, R’=Cy, 3 f:
R=Cy, R’=Cy, 3g: R=Ad, R’=Ad).
The corresponding diphosphine–PdII
lytically
active
hydride
species
[(P^P)Pd(H)
N
(6-MeOH)
was investigated and observed directly
for complexes 5a–e and 5g, respective-
ly. These hydride species were isolated
as the corresponding triphenylphos-
phine complexes (6-PPh3) and fully
characterised, including by X-ray crys-
tallography. The catalytic productivity
of 6a-PPh3 was virtually identical to
that of 5a, thereby confirming the effi-
cient hydride formation of 5a under
catalytic conditions.
ditriflate complexes [(P^P)PdACTHNURTGNENG(U OTf)2]
(5a–g) were prepared and structurally
characterised by X-ray crystallography.
These new complexes were studied as
catalyst precursors in the isomerising
Keywords: carbonylation · isomeri-
sation · palladium · phosphines · re-
newable resources
methoxycarbonylation
of
methyl
oleate, and were found to convert
methyl oleate into the corresponding
Introduction
ladium(II) complexes of 1,2-(tBu2PCH2)2C6H4 (dtbpx;
“Lucite ligand”).[7]
The use of renewable resources as a source of chemicals re-
quires their efficient transformation into useful building
blocks. Fatty acids are attractive feedstocks due to their
unique, long-chain methylene sequences.[1] Their transforma-
tion into linear, a,w-functionalised compounds is of interest,
for example, for the generation of semi-crystalline aliphatic
polyesters and hydrophobic polyamides. The terminal func-
tionalisation of fatty acids is a synthetic challenge, however.
Biotechnological[2,3] as well as entirely chemical catalytic ap-
proaches[1b,4] are studied to this end.
Remarkably, palladium(II) catalysts modified with meso/
rac-1,3-bis(phospha-oxa-adamantyl)propane[8a] or the afore-
mentioned diphosphine dtbpx[8b] (Figure 1) convert the
Alkoxycarbonylation of ethylene to methyl propionate
has recently been implemented as a part of a new industrial
process to methyl methacrylate developed by Lucite Inter-
national.[5,6] This reaction is catalysed with high rates by pal-
[a] M.Sc. J. T. Christl, M.Sc. P. Roesle, M.Sc. F. Stempfle,
Dipl.-Chem. P. Wucher, Dr. I. Gçttker-Schnetmann,
Prof. Dr. G. Mꢀller, Prof. Dr. S. Mecking
Department of Chemistry, University of Konstanz
78464 Konstanz (Germany)
Figure 1. Bulky diphosphine ligands that convert internal octenes into
linear esters.
double bond of internal olefins into a terminal ester group.
Thus, the double bond deep in the chain of unsaturated
fatty acids is carbonylated to a terminal ester group with
high selectivity, yielding a,w-diesters.[9–11] These a,w-diesters
can be converted into a,w-diols, a,w-diamines and a,w-ace-
Fax : (+49)7531-88-5152
E-mail: gerhard.mueller@uni-konstanz.de
stefan.mecking@uni-konstanz.de
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201301124.
Chem. Eur. J. 2013, 19, 17131 – 17140
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
17131

tals, all obtained in polyconden-
sation-grade purity (>99%).[12]
Beyond these findings, no fur-
ther insights into the require-
ment of diphosphine ligands to
promote this unusual catalytic
Figure 2. Unsymmetrical
di-
phosphine ligands synthesised
in this work. 3a: R=tBu, R’=
tBu, 3b: R=tBu, R’=Cy, 3c:
R=tBu, R’=iPr, 3d: R=Ad,
R’=tBu, 3e: R=Ad, R’=Cy,
3 f: R=Cy, R’=Cy, 3g: R=
Ad, R’=Ad.
isomerising
alkoxycarbonyla-
tion have been reported. To
this end, we chose the o-tolyl
backbone[13,26] (Figure 2) as this
allows for more straightforward
and less hazardous preparation
of diphosphines in comparison
with
1,2-(tBu2PCH2)2C6H4.
More importantly, a large range of unsymmetrically substi-
tuted, bulky electron-rich diphosphines are also accessible
selectively.[13,14] We report herein on the corresponding neu-
tral complexes suitable as single-component catalyst precur-
sors, and their selectivity in isomerising methoxycarbonyla-
tion reactions.
Figure 3. Comparison of the gas chromatograms of the crude reaction
mixture (top) and after crystallisation from methanol (bottom) of the
products obtained in the isomerising methoxycarbonylation of methyl
oleate with [(dtbpx)PdACTHNUTRGNEUNG(OTf)2] (5h) (MO/Pd=500, 120 h, 20 bar CO,
908C).
monoesters as side-products (methyl stearate) were ob-
served, we concluded that MO is converted exclusively into
diesters by reaction of its double bond with CO and MeOH.
Besides the desired product (dimethyl nonadecane-1,19-
dioate, L), formed in a total yield of 71.8%, 16 branched
diester (by-)products (B1–B16) are conceivable with MO as
the starting material (Scheme 1). In these byproducts, the
Results and Discussion
The isomerising methoxycarbonylation of methyl oleate
(MO) or the corresponding triglyceride has been reported
to yield dimethyl nonadecane-1,19-dioate (L) as the major
product, which can be obtained
cleanly (see above).[10–12] How-
ever, a comprehensive picture
of the selectivity and the prod-
ucts formed remains to be es-
tablished. By crystallisation of
the crude reaction mixture
from methanol (used as the re-
action medium in catalysis), the
desired linear product L is ob-
tained in >99% purity (Fig-
ure 3).[12a,b] As
a prerequisite
for a correlation of diphosphine
ligand structure with selectivity,
a detailed analysis of the prod-
ucts of isomerising methoxycar-
bonylation with [(dtbpx)Pd-
ACHTUNGTRENNUNG(OTf)2] (5h) was performed.
Scheme 1. Possible products formed during the isomerising methoxycarbonylation of methyl oleate.
Products of the isomerising me-
thoxycarbonylation of methyl
oleate: The product spectrum formed was determined from
a reaction in which [(dtbpx)Pd(OTf)2] (5h) was used as the
newly formed ester function is not at the w position, but at
all positions from wÀ1 (B1) to the a (B16) position. GC
analysis (Figure 3) of the crude reaction mixture showed
that at least seven branched esters are formed (see the Sup-
porting Information). To identify the byproducts, these were
enriched by removing the linear dimethyl nonadecane-1,19-
dioate (L) from the crude reaction mixture by crystallisation
from methanol and column chromatography of the superna-
tant to remove the starting material (MO; for details, see
AHCTUNGTRENNUNG
defined catalyst precursor[12b] with a ratio of methyl oleate/
Pd of 500 and a prolonged reaction time of 120 h at 908C at
a CO pressure of 20 bar using a mechanically stirred
(500 rpm) 200 mL pressure reactor. Pure methyl oleate
(99%) was used to exclude impurities from the starting ma-
terial and facilitate identification of the products. GC analy-
sis revealed a MO conversion of 80.9%. As no saturated
17132
www.chemeurj.org
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 17131 – 17140

Alkoxycarbonylation of Methyl Oleate
FULL PAPER
1
the Supporting Information). NMR analysis (1H, 13C, H,1H
COSY, H,13C HSQC, H,13C HMBC) of this purified reac-
tion mixture revealed the presence of dimethyl 2-methyloc-
tadecane-1,18-dioate (B1), dimethyl 2-ethylheptadecane-
1,17-dioate (B2) and dimethyl 2-propylhexadecane-1,16-
dioate (B3). Longer-chain branched diesters, dimethyl 2-bu-
tylpentadecane-1,15-dioate (B4) and dimethyl 2-pentyltetra-
decane-1,14-dioate (B5), were also found. For diesters with
even longer branches, B6–B15, individual compounds could
not be assigned unambiguously by NMR spectroscopy.
Formation of the malonic ester MeOOC-CHR-COOMe
(R=C16H33; B16) was evidenced by comparison with the 13C
carbonyl NMR shift of a genuine sample of the compound
prepared independently, and also by enrichment of the puri-
fied reaction mixture with this genuine sample in the GC
analysis (see the Supporting Information).
1
1
Scheme 2. Synthetic route to the unsymmetrical diphosphines. 3a: R=
tBu, R’=tBu; 3b: R=tBu, R’=Cy, 3c: R=tBu, R’=iPr, 3d: R=Ad,
R’=tBu, 3e: R=Ad, R’=Cy, 3 f: R=Cy, R’=Cy, 3g: R=Ad, R’=Ad.
Table 1 gives the product distribution as determined by
GC. The selectivity for the formation of the linear diester
dimethyl nonadecane-1,19-dioate (L) was found to be
mercially available o-bromobenzyl bromide (1) as the start-
ing material (Scheme 2). Compound 1 was treated with sec-
ondary bulky phosphines such as HPACHTUNGRTENUNG(tBu)2, HP(Ad)2 or
HP(Cy)2, leading to the corresponding monosubstituted in-
termediate derivatives as hydrobromic acid salts, which pre-
cipitate as white solids. The best yields were obtained with
acetonitrile as a solvent. After the subsequent release of hy-
drobromide, the resulting neutral monophosphines 2a, 2d
and 2 f could be used directly in the next step without fur-
Table 1. Product distribution observed in the isomerising methoxycarbo-
nylation of methyl oleate with [(dtbpx)PdACTHNUTRGNEUNG
(OTf)2].[a]
Product
L
B1
B2
B3
B4–B15
B16
Fraction [%]
89.0
4.3
1.0
0.6
4.8
0.3
[a] Reaction conditions: 96 mmol (28.5 g) MO (99%), 0.19 mmol
(153 mg) [(dtbpx)Pd(OTf)2], 130 mL MeOH, 20 bar CO, 908C, 120 h.
1
ACHTUNGTRENNUNG
ther purification (purity by H NMR, >98%).
Total conversion of MO was 80.9%. Product distribution calculated from
GC data.
Compound 2a was lithiated in pentane by the addition of
nBuLi at room temperature over 2 h. The corresponding iso-
lated lithium aryl 2a-Li was treated with chlorophosphines
in THF at À808C to afford the desired unsymmetrical di-
phosphines 3a and 3b, which were recrystallised from etha-
nol.
89.0%. Amongst the branched products, dimethyl 2-methyl-
octadecane-1,18-dioate (B1) predominates with 4.3% con-
tent. The portion of the respective branched products de-
creases with increasing length of the branch. The malonic
diester (B16) is also formed, but to a very small extent only.
This is in accordance with our recent mechanistic studies[9]
by low-temperature NMR spectroscopy and DFT methods
in which a preference for linear insertion products along
Alternatively, the monophosphines 2a, 2d and 2 f were
lithiated in THF solution with tBuLi at À808C. The in situ
generation of the lithium aryls of the monophosphines
turned out to be advantageous due to the limited stability of
the isolated lithium aryls. By using tBuLi as the lithiation
agent rather than nBuLi, the in situ formation of the lithium
aryl occurred more rapidly. The target compounds 3c–f
were obtained in moderate-to-high yields (34–87%) by
quenching the in situ generated lithium aryls with chloro-
phosphines followed by recrystallisation from ethanol, lead-
ing to suitable crystals of diphosphines 3a, 3d and 3 f for X-
ray diffraction (see the Supporting Information). To further
explore the potential of bulky alkyl substituents, another
chelating ligand with two adamantyl groups on both phos-
phorus atoms was synthesised. The general synthetic route
with
a roughly similar amount of the branched acyl
[(P^P)PdC(=O)CHRCOOMe]+ was observed. The forma-
tion of the latter is promoted by the stabilisation incurred
by chelation of the ester group of the MO substrate. Howev-
er, this branched acyl is subject to a relatively slow metha-
nolysis, and thus only a very small amount of B16 was
formed.
Synthesis of the diphosphine ligands: PdII complexes bearing
the sterically demanding 1,2-(tBu2PCH2)2C6H4 ligand are
known to be very productive in the isomerising alkoxycarbo-
nylation of plant oils[10,11,12a,b] but the diphosphine is some-
what hazardous to synthesise, especially when sodium alkyl
intermediates are involved.[15] Following the rationale of
using bulky substituents on phosphorus and a relatively rigid
backbone, we chose the o-tolyl backbone to vary the sub-
stituents at both phosphorus atoms independently in non-
symmetric diphosphines. At the same time, the target li-
gands are conveniently accessible in three steps from com-
led to diphosphine 3g but by using the corresponding bro-
[23]
mophosphine BrPAd2
instead of the chlorophosphine.
BrPAd2 was prepared by the reaction of HPAd2 with CBr4
in CH2Cl2. The reaction took place within several minutes at
room temperature. The desired bromo compound was isolat-
ed in 95% yield by removing the solvent in vacuum. In situ
lithiation of 2d and addition of BrPAd2 led to the formation
of 3g, which was isolated after several crystallisation steps
from MeOH and pentane.
Chem. Eur. J. 2013, 19, 17131 – 17140
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
17133

G. Mꢀller, S. Mecking et al.
Table 2. 31P{1H} NMR chemical shifts of the diphosphine ligands 3a–f.[a]
The coordination geometries contain the two phos-
phorus atoms of the chelating phosphines with the
other two coordination sites occupied by triflate
oxygen atoms in 5a–c and 5e,f. In complexes 5b,
5c, 5e and 5 f, both triflate counter anions are coor-
dinated to PdII in a monodentate fashion, whereas
in 5a (with the tBu2/tBu2-substituted diphosphine)
only one triflate anion coordinates to palladium as
Ligand
3a
3b
3c
3d
3e[b]
3 f
3g[b]
Ad/Ad
tBu/tBu
tBu/Cy
tBu/iPr
Ad/tBu
Ad/Cy
Cy/Cy
d(P2) [ppm]
d(P1) [ppm]
35.0
15.0
36.4
À17.6
36.4
À8.7
33.8
13.6
38.1
À17.7
7.9
35.9
16.1
À16.1
[a] In C6D6 at 258C, unless noted otherwise. [b] In CDCl3.
The 31P NMR shifts of the diphosphine ligands (Table 2)
show good correlation with the nature of the substituents R
and R’, with d decreasing in the order tBuꢀAd>iPr>Cy.
a bidentate chelating ligand. In 5d (Ad2/tBu2-substituted di-
phosphine), only one triflate anion is coordinated as a mono-
dentate ligand, with the fourth coordination site occupied by
a water molecule, probably originating from traces of water
in the solvent or the trifluoromethanesulfonic acid reagent.
Complexes 5d and 5e show the most severe deviation from
the square-planar geometry (see the Supporting Informa-
tion). Although the water oxygen atom in 5d is roughly co-
planar with the phosphorus and palladium atoms, the triflate
oxygen is situated markedly above this plane. Note that the
water oxygen is located cis to the -CH2P2Ad2 phosphorus
atom, whereas the triflate ligand is cis to the ArP1tBu2 phos-
phorus atom. This geometrical arrangement seems to be dic-
tated by steric pressure although an intramolecular hydro-
gen bond between one water hydrogen and a triflate oxygen
À
Synthesis and molecular structures of the diphosphine–PdII
complexes: The desired PdII complexes were obtained by
a two-step protocol starting from the diphosphine ligands
and [Pd
[Pd(dba)2] were mixed in THF and stirred at room tempera-
ACHTUNGTRENNUNG
(dba)2] (Scheme 3).[9,27] The diphosphine (3a–g) and
ACHTUNGTRENNUNG
ture for 12 h. After filtration of small amounts of palladium
atom may also play a role (dACTHNUGRTNEUNG(O O)=2.7(2) ꢂ, a O-H-O
157.1(3)8). Bond lengths and angles (see the Supporting In-
formation) are in the typical range for this type of complex.
Interestingly, no clear-cut correlation between steric bulk of
À
the phosphine ligands and the Pd P bond lengths is evident.
In some cases significant differences in aryl/alkyl substitu-
ents can be found but no general trend can be easily ascer-
tained. The steric bulk seems to be relieved by a tilt of the
P2Pd and O2Pd molecular planes with respect to each other
(5a (tBu2/tBu2 substitution at P): ca. 8.48, 5b (tBu2/Cy2): ca.
21.08, 5c (tBu2/iPr2): ca. 25.58, 5d (Ad2/tBu2): ca. 31.68, 5e
(Ad2/Cy2): ca. 30.98, 5 f (Cy2/Cy2): ca. 9.98). The bite
angles[24] of the diphosphine ligands at PdII range between
89.35(2) and 94.65(5)8 (see the Supporting Information).[16]
These values are similar to those of dppp and dppf PdII com-
plexes (91 and 968, respectively).[28] The complex with the
“Lucite ligand” with its more flexible xylyl-type backbone
(seven-membered chelate ring) has a slightly higher bite
angle of 99.38.[12b]
Scheme 3. Synthesis of diphosphine–PdII complexes 5a–g.
black and removal of the solvent, the reddish residue ob-
tained was washed with pentane to remove residual diphos-
phine. The obtained diphosphine–Pd0 complexes were used
in the next step without characterisation. Crystals of 4d suit-
able for X-ray crystallography were obtained by the addition
of pentane to a solution of 4d in CH2Cl2 (see the Supporting
Information).
Dissolution of the [(P^P)Pd0
ACTHNUTRGNE(UNG dba)] complexes in diethyl
ether and addition of trifluoromethanesulfonic acid in the
presence of benzoquinone as oxidising agent afforded the
desired diphosphine–PdII ditriflate complexes (5a–g) as off-
white to yellow solids within 12 h. The complexes were ana-
lysed by NMR spectroscopy, X-ray crystallography
(Figure 4),[29] mass spectrometry and elemental analysis.
Single crystals suitable for X-ray diffraction were obtained
by slow diffusion of pentane into methylene chloride solu-
tions of the complexes. Complexes 5a and 5 f crystallised
with 1 equivalent of CH2Cl2, and complex 5e includes half
an equivalent of pentane per formula unit in the crystal. All
the complexes 5a–f have four-coordinate PdII centres that
are essentially square planar with tetrahedral distortion in
some cases (deviations from the least-root mean-square
planes defined by the central metal atom and the four coor-
dinating atoms are given in the Supporting Information).
Methoxycarbonylation of methyl oleate (MO): The catalytic
properties of the new complexes 5a–g in the isomerising
methoxycarbonylation of methyl oleate were then studied
(Scheme 4). A constant pressure of CO (20 bar), n(Pd)=
0.048 mmol, n(MO)=6 mmol, 10 mL methanol, 908C and
a reaction time of 120 h were chosen as the experimental
parameters to evaluate the catalytic performance of com-
Scheme 4. Methoxycarbonylation of methyl oleate (MO).
17134
www.chemeurj.org
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 17131 – 17140

Alkoxycarbonylation of Methyl Oleate
FULL PAPER
Figure 4. X-ray crystal structures of complexes 5a–f. Hydrogen atoms except for coordinated H2O, non-coordinating triflate counter ions, and interstitial
solvent molecules have been omitted for clarity. Displacement ellipsoids are shown at the 50% probability level.[29]
plexes 5a–g. To study the “real-life” catalytic performance,
technical-grade Dakolub MB 9001 high oleic sunflower oil
methyl ester with a methyl oleate content of 92.5% was
used[21] rather than highly pure methyl oleate (Table 3).
Catalysts 5a–g converted methyl oleate into diesters in
yields of 18–98%. Surprisingly,
the steric demand of the substituents on the phosphorus
moieties of the ligands. To this end, we quantified the avail-
able space around the metal centre from the available X-ray
crystal structure data (see above). For this purpose we deter-
mined the angles between the vector V defined by the
the selectivity towards the de-
Table 3. Results of the isomerising methoxycarbonylation of methyl oleate using PdII complexes 5a–g and the
reference complex 5h.[a]
sired 1,19-diester was similar
(70–80%) for all catalysts stud-
ied, however, the conversion
varied strongly. Complex 5 f
showed a very low conversion
(1.3%) of methyl oleate, which
is in line with the observation
that no hydride formation oc-
curred upon addition of metha-
nol to this complex (see below).
The differences in the conver-
PdII
complex
Conversion of
MO [%]
Relative distribution [%]
1,19-diester branched products
Selectivity for
linear 1,19-diester [%]
5a (tBu/tBu)
5b (tBu/Cy)
5c (tBu/iPr)
5d (Ad/tBu)
5e (Ad/Cy)
5 f (Cy/Cy)
5g (Ad/Ad)
5h (dtbpx)
94.2
17.8
18.7
77.4
22.2
1.3
64.0
13.5
14.0
61.0
17.7
–
30.1
4.3
4.7
16.3
4.5
–
68.0
76.0
75.0
79.0
80.0
–
97.1
95.8
76.0
86.8
20.8
9.0
79.0
90.6
[a] Reaction conditions: n(Pd)=0.048 mmol, 2 mL Dakolub MB 9001 (92.5% MO), 10 mL MeOH, 20 bar CO,
sions of 5a–g can be related to 908C, 120 h.
Chem. Eur. J. 2013, 19, 17131 – 17140
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
17135

G. Mꢀller, S. Mecking et al.
centre between the two phosphorus atoms and the PdII
centre, and four vectors R1, R2, R3 and R4 defined by the
PdII centre and the substituents on the phosphorus moiety
(cf. Figure 5). The arithmetic average of these angles is the
Time dependency of the methyl oleate conversion: Com-
plexes 5a, 5g and the reference complex 5h afforded nearly
complete conversion under the studied reaction conditions
(120 h). To further evaluate the productivity of these com-
plexes, the conversion with complexes 5a and 5h as catalyst
precursors was monitored over time. Reactions were carried
out in a 200 mL pressure reactor equipped with a sampling
valve at the bottom of the reactor from which samples were
taken at regular intervals (for details see the Supporting In-
formation). A constant 20 bar pressure of CO, n(Pd)=
0.77 mmol, n(MO)=96 mmol, 110 mL methanol and 908C
were employed with technical-grade Dakolub MB 9001 used
as the MO source. Figure 6 shows the consumption of
methyl oleate versus time for both complexes 5a and 5h.
Figure 5. Plot of the conversion of methyl oleate versus the calculated
half-cone angle.
“half-cone angle” (for details see the Supporting Informa-
tion). Note that the angles were calculated from the corre-
sponding X-ray structures without considering the rotation
À
of alkyl substituents around the P C bond in solution.
Larger values indicate increasing available space around the
metal centre and thus a smaller steric constraint imposed by
the diphosphine ligand. Figure 5 shows the conversion of
methyl oleate versus the increasing steric constraint imposed
by the diphosphine ligand: increasing conversion accompa-
nies increasing steric demand (5 f<5b<5c<5e<5d<5a).
The most active unsymmetrical catalysts (5a and 5d, 89 and
908, respectively) have similar half-cone angles to 5h (868).
A more detailed analysis of the relative distribution of the
branched ester byproducts (Table 4) shows that catalysts
with similar productivity (5a, 5d, 5g and 5h versus 5b, 5c
and 5e) also show similar distribution of branched products.
The less active catalysts have a stronger preference for the
methyl-branched product (B1), which is the predominant
branched ester in all cases, whereas the more bulky substi-
tuted and more active catalysts (5a, 5d, 5g and 5h) have
a lower preference for this methyl-branched product com-
pared with the other branched byproducts.
Figure 6. Time dependency of methyl oleate conversion with complexes
5a (*) and 5h (~). Reaction conditions: n(Pd)=0.77 mmol, 32.5 mL
Dakolub MB 9001 (92.5% MO), 110 mL MeOH, 20 bar CO, 908C.
Although the consumption of methyl oleate with catalyst 5a
is around three times slower than with 5h, both catalysts
eventually completely converted methyl oleate into diesters.
The selectivity towards the desired 1,19-diester was 68% for
5a, and 91% for 5h, respectively, which agrees with the re-
sults of the aforementioned screening of the catalysts under
slightly different reaction conditions.
Interestingly, in samples drawn from the reaction with 5h
as a catalyst, considerable amounts of palladium black were
observed after a reaction time of 1 h. In contrast, in reac-
tions with 5a, almost no palladium black was observed. This
suggests that catalyst 5a is more stable towards degradation.
Note that the selectivity towards the 1,19-diester is constant
over time for both complexes 5a and 5h. This implies that
any degradation does not impact the nature of the active
species.
Table 4. Relative distribution of branched ester byproducts with different
catalysts.
PdII complex
B1
B2
B3
B4–B16
Formation and isolation of palladium(II) hydride species:
The catalytically active species of the methoxycarbonylation
reaction is considered to be a PdII hydride species that is
formed within several minutes upon addition of methanol to
5a (tBu/tBu)
5b (tBu/Cy)
5c (tBu/iPr)
5d (Ad/tBu)
5e (Ad/Cy)
5g (Ad/Ad)
5h (dtbpx)
21
49
49
28
51
24
34
3
5
4
3
4
3
9
3
5
4
3
4
3
6
63
41
43
66
41
70
51
the [(P^P)PdACTHNUTRGNEUNG
(OTf)2] complexes at room temperature.[9,22,27a]
To elucidate whether the new PdII complexes studied here
form hydride species upon addition of methanol, complexes
17136
www.chemeurj.org
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 17131 – 17140

Alkoxycarbonylation of Methyl Oleate
FULL PAPER
1
5a–h were dissolved in CD2Cl2/MeOH (2:3, v/v) and H and
31P{1H} NMR spectra were recorded. Complete conversion
of 5a, 5d, 5g and 5h was revealed by 31P{1H} NMR spec-
troscopy. Evidence for the formation of hydride species
1
arises from the appearance of H resonances at around d=
À10 ppm with typical coupling patterns of diphosphine–PdII
hydride species. Figure 7 exemplarily illustrates the forma-
tion of the hydride 6d-MeOH from 5d in CD2Cl2/MeOH so-
1
lution. In the H NMR spectrum, a hydride resonance at d=
Figure 8. 31P{1H} NMR spectrum of the in situ formation of the PdII hy-
dride species from 5d in [D4]MeOH at 258C.
31P NMR spectrum showed P2 to be cis (doublet of triplet
2
2
with JPPcis =24.2 Hz and JPDcis =4.5 Hz) and P1 to be trans
(doublet of triplet with 2JPPcis =24.2 Hz and JPDtrans
=
2
29.0 Hz) to the deuteride. The corresponding observations
were made with complexes 5a and 5g. The hydride species
formed from 5a, 5d and 5g are remarkably stable represen-
tatives of PdII hydrides with a chelating diphosphine ligand
and a weakly coordinated solvent molecule. For example,
these hydrides can be stored in MeOH at room temperature
for 4 or 5 days before decomposition occurs.
Figure 7. 31P{1H} NMR spectrum of the in situ formation of the PdII hy-
À
dride species from 5d in CD2Cl2/MeOH (2:3, v/v) at 258C. Inset: Pd H
resonances.
In contrast, the addition of methanol to complexes 5b, 5c
and 5e did not result in complete conversion of the triflate
complexes into single hydride species. Instead, up to 70%
(5e) of the unreacted triflate complex was still detected.
The observation of several 1H NMR resonances in the
region d=À5 to À10 ppm with different coupling patterns
for all three complexes suggests the formation of different
hydride species. For complexes 5b and 5c, a quintet is ob-
served amongst others, which could indicate the formation
of binuclear hydride species.[9] No hydride species were ob-
served from complex 5 f under these conditions.
To isolate the palladium hydrides, we introduced PPh3 as
a strongly coordinating ligand. Addition of PPh3 to com-
plexes 5a–g and 5h in methanol led to the formation of the
corresponding [(P^P)PdII(H)PPh3]+ species (6-PPh3). Nota-
bly, in contrast to the corresponding solvent-coordinated
species 6-MeOH, a mixture of cis and trans isomers was ob-
served for all complexes, the ratios varying with time. A typ-
ical 31P{1H} NMR spectrum is shown in Figure 9; all the sig-
nals could be assigned to palladium hydrides. In many cases,
the predominant isomer is the species with the hydride lo-
cated cis to P2.
2
À9.53 ppm (doublet of doublets with JPHtrans =187.6 Hz and
2JPHcis =28.7 Hz) is observed. The 31P{1H} NMR spectrum
shows two doublets at 34.2 and 82.3 ppm for the two non-
equivalent phosphorus atoms with JPPcis =24.2 Hz (Table 5).
Notably, only a single isomer of the hydride was observed in
MeOH solution. 2D 1H,31P NMR correlation spectroscopy
showed that the signal at d=34.2 ppm could be assigned to
the aryl phosphorus atom (P1), whereas the signal at d=
2
2
82.3 ppm displays JPH coupling to the benzyl protons and
was therefore identified as the benzyl phosphorus atom (P2).
With [D4]MeOH as solvent, the respective deuteride com-
plexes were formed (Figure 8). The coupling patterns in the
Table 5. 31P{1H} NMR data for PdII hydride complexes 6-MeOH in
CD2Cl2/MeOH at 258C.
For complexes 5a, 5d and 5h, virtually full conversion to
the PPh3-coordinated hydrides was observed. Although in
the cases of 5b, 5c, 5e and 5g it was possible to form the
corresponding PPh3 hydrides (in yields of 40–90%), the con-
version was not enhanced. Signals in the NMR spectra were
assigned to the starting triflate complexes.
PdII hydride
d(P2)
[ppm]
d(P1)
[ppm]
ACTHNGUTERNNUG
2J(P2–P1)
[Hz]
G
E
6a-MeOH (tBu/tBu)
6d-MeOH (Ad/tBu)
6g-MeOH (Ad/Ad)
83.8
82.3
82.1
34.6
34.2
33.0
24.5
24.2
24.3
The hydrides 6a-PPh3 to 6h-PPh3 were isolated by extrac-
tion of the methanol solution with pentane, which resulted
Chem. Eur. J. 2013, 19, 17131 – 17140
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
17137

G. Mꢀller, S. Mecking et al.
À
with the formation of a Pd C bond (2.120(2) ꢂ) and net ex-
pulsion of molecular hydrogen. The preference of a tBu sub-
2
stituent at P1 for C H activation over that of a tBu2P CH2
À
group might again reflect the larger steric rigidity of the
R2P1 moiety compared with the more flexible R2P2CH2
backbone. Probably more important, however, is the cis ori-
entation, and thus closer spacial proximity, of the palladium-
bound hydrogen atom in 6a with respect to the activated
À
methyl group of the tBu substituent at P1 (d
ACTHNUTRGNEU(GN Pd C15)=
2.120(2) ꢂ). This suggests the direct involvement of the pal-
À
ladium-bound hydrogen atom in the C H activation process.
This formation of a pallada-phospha-heterocycle might also
be a part of the deactivation of the catalyst that occurs
during the catalytic reaction.
Figure 9. 31P{1H} NMR spectrum of 6d-PPh3 formed by dissolution of 5d
Methoxycarbonylation of methyl oleate with the PdII hy-
À
and 1 equiv of PPh3 in CD2Cl2/MeOH (2:3, v/v) at 258C. Inset: Pd H res-
onances.
dride species as
a catalyst precursor: We conducted
methoxycarbonylation experiments of methyl oleate with
AHCTUNGTRENNUNG
isolated 6a-PPh3 and 6h-PPh3 as catalyst precursors to
probe the possible catalytic productivity of these hydrides.
Similar reaction conditions to those used for the methoxy-
carbonylation reactions with 5a and 5h were employed
(20 bar CO, ratio (MO/Pd)=125, 10 mL methanol, 908C,
120 h). The catalytic performances of the isolated hydride
species are virtually identical to those of the corresponding
triflate complexes (Table 6). Although this result may have
in precipitation of the desired complexes. Crystals suitable
for X-ray diffraction were obtained from 6a-PPh3, 6e-PPh3,
6d-PPh3 and 6h-PPh3 by layering a methylene chloride so-
lution of the respective compound with pentane at room
À
temperature (Figure 10). The bond lengths (dACTHNUGRTNEUNG(H Pd)=
1.51(5)–1.54(3) ꢂ) and angles (see the Supporting Informa-
tion) for 6a-PPh3, 6e-PPh3 and 6d-PPh3 are in the typical
range for this type of complex.
In 6a-PPh3 and 6d-PPh3, the
palladium-bound hydrogen is
Table 6. Comparison of the isomerising methoxycarbonylation of methyl oleate with complexes 5a, 6a-PPh3,
5h and 6h-PPh3.
oriented trans with respect to
P2, whereas in 6e-PPh3 it is sit-
uated cis. The different stereo-
chemistry most probably re-
flects the minimisation of steric
hindrance between the PPh3
ligand and the different steric
demands of the R’2P1 and
R2P2CH2 donors. With R=R’=
tBu (6a-PPh3) the PPh3 ligand
is oriented cis with respect to
PdII hydride
Conversion of
methyl oleate [%]
Relative distribution [%]
1,19-diester branched products
Selectivity for
1,19-diester [%]
5a (tBu/tBu)[a]
94.2
91.0
95.8
97.0
64.0
60.0
86.8
88.6
30.1
30.0
9.0
68.0
60.0
90.6
91.0
[b]
6a-PPh3
5h (dtbpx)[a]
[b]
6h-PPh3
8.7
[a] Reaction conditions: n(Pd)=0.048 mmol, 2 mL Dakolub MB 9001 (92.5% MO), 10 mL MeOH, 20 bar CO,
908C, 120 h. [b] n(Pd)=0.024 mmol, 1 mL Dakolub MB 9001 (92.5% MO), 10 mL MeOH, 20 bar CO, 908C,
120 h.
the more flexible tBu2P2CH2 substituent. The same applies
to 6d-PPh3 (tBu2P1 versus Ad2P2CH2), but even a simple in-
spection of the molecular structure (Figure 10) reveals no-
ticeable distortions arising from the larger steric bulk of the
adamantyl substituents. Finally, in 6e-PPh3, the PPh3 ligand
prefers a cis orientation with respect to the (less bulky)
Cy2P1 group rather than to the Ad2P2CH2 moiety, again with
severe distortions reflecting the large steric bulk of the
been anticipated, it is very relevant to confirming the essen-
tial issue that hydrides are formed efficiently from the tri-
flate complexes [(P^P)PdACTHNUTRGNE(UGN OTf)2] under catalytic conditions.
Conclusion
Bulky diphosphines 3a–g with o-tolyl backbones have been
prepared in moderate-to-good yields. These ligands were
transformed into the corresponding diphosphine–PdII ditri-
flate complexes 5a–g. The bite angles of the ligands, as de-
termined by single-crystal X-ray diffraction, are all in the
range of 89.35(2)–94.65(5)8 for all complexes and similar to
the prototypical symmetrical “Lucite ligand” 1,2-
(tBu2PCH2)2C6H4 (99.38) in 5h. All the complexes 5a–g cat-
alysed the isomerising methoxycarbonylation of methyl
oleate to yield the linear dimethyl nonadecane-1,19-dioate
À
groups involved. In 6h-PPh3, the Pd H bond length is
slightly longer (1.59(2) ꢂ) than in 6a-PPh3, 6e-PPh3 and
6d-PPh3.
Upon crystallisation of 6a-PPh3, which led to the molecu-
lar structure shown in Figure 10, crystals of a second species,
7a, formed. Its structure determination (Figure 10) revealed
the formation of a four-membered pallada-phospha-hetero-
[25]
À
cycle formed by C H activation of the methyl group of
a tBu substituent at P1 by the adjacent palladium centre
17138
www.chemeurj.org
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 17131 – 17140

Alkoxycarbonylation of Methyl Oleate
FULL PAPER
À
Figure 10. X-ray crystal structures of complexes 6a-PPh3, 6d-PPh3, 6e-PPh3, 6h-PPh3, and 7a. Hydrogen atoms (except for Pd H) and triflate counter
ions have been omitted for clarity. Displacement ellipsoids are shown at the 50% probability level.[29]
in moderate-to-good yields. This clearly shows that catalyst
5h is not unique in transforming the internal double bond
deep in the methyl oleate chain into the linear 1,19-diester.
A thorough analysis of the products formed revealed that
likely all possible branched side-products are formed to
a very small extent with the methyl-branched diester pre-
dominating significantly compared with the other branched
isomers. Remarkably, the selectivity for the linear 1,19-die-
ster over the branched byproducts did not vary strongly for
the different diphosphine structures. Also, for less bulky sec-
ondary alkyl substituents at the phosphorus, a strong prefer-
ence for isomerising methoxycarbonylation to the linear
product was observed. Structure–productivity relationships
were derived from crystallographic and catalytic data. Cata-
lyst productivity clearly increases with steric bulk of the sub-
stituents on the phosphorus moiety. Thus, the nature of the
bulky substituents at phosphorus appears to influence pro-
ductivity more distinctly than selectivity. Stoichiometric
studies of hydride formation revealed that for the less bulky
substituted complexes, formation of the active species is
much less efficient than for the tert-butyl- and adamantyl-
substituted diphosphine. As a guideline for further improv-
ing this remarkable and useful reaction, an efficient conver-
sion of a given catalyst precursor into the active species, as
well as keeping the latter in the catalytic cycle appear deci-
sive key issues.
Experimental Section
For further details, please see the Supporting Information.
Acknowledgements
Financing by the Stiftung Baden-Wꢀrttemberg (Programm Umwelttech-
nologie-forschung, U14) is gratefully acknowledged. P.R. gratefully ac-
knowledges support from the Carl-Zeiss-Foundation through a graduate
fellowship. We thank Dako AG for the donation of Dakolub MB 9001.
Dr. Etienne Grau is acknowledged for providing complex 6h-PPh3.
Chem. Eur. J. 2013, 19, 17131 – 17140
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
17139

G. Mꢀller, S. Mecking et al.
[16] They are rather small compared with other diphosphine–PdII com-
plexes (DPEphos 102.78, Xantphos 1108) used, for example, in allyl-
ic alkylation reactions (see Refs. [17–20]). DPEphos=[oxybis(2,1-
[1] a) U. Biermann, U. Bornscheuer, M. A. R. Meier, J. O. Metzger,
H. J. Schꢃfer, Angew. Chem. 2011, 123, 3938–3956; Angew. Chem.
Int. Ed. 2011, 50, 3854–3871; b) S. Chikkali, S. Mecking, Angew.
Chem. Int. Ed. 2012, 51, 5802–5808.
phenylene)]bis(diphenylphosphine);
phosphino)-9,9-dimethylxanthene.
Xantphos=4,5-bis(diphenyl-
[2] a) S. Picataggio, T. Rohrer, K. Deanda, D. Lanning, R. Reynolds, J.
Mielenz, L. D. Eirich, Nat. Biotechnol. 1992, 10, 894–898; b) U.
Schçrken, P. Kempers, Eur. J. Lipid Sci. Technol. 2009, 111, 627–
645.
[17] P. Dierkes, P. W. N. M. van Leeuwen, Dalton Trans. 1999, 1519–
1530.
[18] G. P. C. M. Dekker, C. J. Elsevier, K. Vrieze, P. W. N. M. van Leeu-
wen, Organometallics 1992, 11, 1598–1603.
[3] W. Lu, J. E. Ness, W. Xie, X. Zhang, J. Minshull, R. A. Gross, J. Am.
Chem. Soc. 2010, 132, 15451–15455.
[19] J. Ledford, C. S. Schultz, D. P. Gates, P. S. White, J. M. DeSimone,
M. Brookhart, Organometallics 2001, 20, 5266–5276.
[20] P. W. N. M. van Leeuwen, P. C. J. Kamer, J. N. H. Reek, Pure Appl.
Chem. 1999, 71, 1443–14452.
[21] Typical distribution of Dakolub MB 9001: 92.5% methyl oleate,
2.5% methyl linoleate, 2.5% methyl palmitate, 1.5% methyl stea-
rate, 1.0% methyl esters of longer chain (>C18) fatty acids.
[22] G. R. Eastham, B. T. Heaton, J. A. Iggo, R. P. Tooze, R. Whyman, S.
Zacchini, Chem. Commun. 2000, 609–610.
[4] a) F. Stemple, P. Roesle, S. Mecking, ACS Symp. Series 1105 Bio-
based Monomers, Polymers and Materials (Ed.: R. Gross, P. B.
Smith), 2012, pp. 151–163; b) M. B. Dinger, J. C. Mol, Adv. Synth.
Catal. 2002, 344, 671–677; c) A. Behr, D. Obst, A. Westfechtel, Eur.
J. Lipid Sci. Technol. 2005, 107, 213–219.
[5] a) M. McCoy, J.-F. Tremblay, Chem. Eng. News 2009, 87, 9; b) B.
Harris, Ingenia 2010, 45, 19–23.
[6] Now: Mitsubishi Rayon; cf. Ref. [5a].
[23] J. Goerlich, R. Schmutzler, Phosphorus Sulfur Silicon Relat. Elem.
1994, 88, 241–244.
[24] P. W. N. M. van Leeuwen, Homogenous catalysis, understanding the
art, Kluwer, 2004, 125, pp. 16–19.
[25] D. J. Tempel, L. K. Johnson, R. L. Huff, P. S. White, M. Brookhart, J.
Am. Chem. Soc. 2000, 122, 6686–6700.
[7] a) W. Clegg, G. R. Eastham, M. R. J. Elsegood, R. P. Tooze, X. L.
Wang, K. Whiston, Chem. Commun. 1999, 1877–1878; for a prior
report of the coordination chemistry of dtbpx, see: b) N. Carr, B. J.
Dunne, A. G. Orpen, J. L. Spencer, J. Chem. Soc. Chem. Commun.
1988, 926–928.
[8] a) R. I. Pugh, E. Drent, P. G. Pringle, Chem. Commun. 2001, 1476–
1477; b) C. Jimꢄnez-Rodriguez, D. F. Foster, G. R. Eastham, D. J.
Cole-Hamilton, Chem. Commun. 2004, 1720–1721. Also cf; c) V. de
La Fuente, M. Waugh, G. R. Eastham, J. A. Iggo, S. Castillꢅn, C.
Claver, Chem. Eur. J. 2010, 16, 6919–6932.
[9] P. Roesle, C. J. Dꢀrr, H. M. Mçller, L. Cavallo, L. Caporaso, S.
Mecking, J. Am. Chem. Soc. 2012, 134, 17696–17703.
[10] C. Jimꢄnez-Rodriguez, G. R. Eastham, D. J. Cole-Hamilton, Inorg.
Chem. Commun. 2005, 8, 878–881.
[11] a) G. Walther, J. Deutsch, A. Martin, F. E. Baumann, D. Fridag, R.
Franke, A. Kçckritz, ChemSusChem 2011, 4, 1052–1054; 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.
[26] a) E. Drent, W. W. Jager, R. I. Pugh (Shell Oil Company) US
7265242 B2, 2007; b) M. Boutain, S. B. Duckett, J. P. Dunne, C.
Godard, J. M. Hernꢆndez, A. J. Holmes, I. G. Khazal, J. L. Serrano,
Dalton Trans. 2010, 39, 3495–3500; c) A. K. Ghosh, D. R. Nicponski,
J. Kass, Tetrahedron Lett. 2012, 53, 3699–3702; d) M. Davis, F. G.
Mann, J. Chem. Soc. 1964, 3786–3790; e) E. Drent, R. Ernst, W. W.
Jager, (Shell Internationale Research Maatschappij) WO2004/
103948A1.
[27] a) W. Clegg, G. R. Eastham, M. R. J. Elsegood, B. T. Heaton, J. A.
Iggo, R. P. Tooze, R. Whyman, S. Zacchini, Organometallics 2002,
21, 1832–1840; b) J. J. R. Frew, K. Damian, H. Van Rensburg,
A. M. Z. Slawin, R. P. Tooze, M. L. Clarke, Chem. Eur. J. 2009, 15,
10504–10513.
[12] a) D. Quinzler, S. Mecking, Angew. Chem. 2010, 122, 4402–4404;
Angew. Chem. Int. Ed. 2010, 49, 4306–4308; b) F. Stempfle, D.
Quinzler, I. Heckler, S. Mecking, Macromolecules 2011, 44, 4159–
4166; c) S. Chikkali, F. Stempfle, S. Mecking, Macromol. Rapid
Commun. 2012, 33, 1126–1129; d) F. Stempfle, P. Ortmann, S. Meck-
ing, Macromol. Rapid Commun. 2013, 34, 47–50.
[13] a) G. R. Eastham, P. I. Richards (Lucite International UK Limited)
WO 2009/010782 A1; b) T. Fanjul, G. Eastham, 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] dppp=1,3-bis(diphenylphosphino)propane; dppf: 1,1’-bis(diphenyl-
phosphino)ferrocene.
[29] CCDC-931019 (3a), CCDC-931020 (3d), CCDC-931021 (3 f),
CCDC-931022 (4d), CCDC-931025 (5a), CCDC-931023 (5b),
CCDC-931024 (5c), CCDC-931012 (5d), CCDC-931013 (5e),
CCDC-931014 (5 f), CCDC-931015 (6a-PPh3), CCDC-931016 (6d-
PPh3), CCDC-931017 (6e-PPh3), CCDC-937890 (6h-PPh3), CCDC-
931018 (7a) contain the supplementary crystallographic data for this
paper. These data can be obtained free of charge from The Cam-
bridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_-
request/cif. Further details are also given in the Supporting Informa-
tion.
[14] The unsymmetrical diphosphine ligand with four tert-butyl substitu-
ents (3a) has been reported previously, cf. Ref. [13].
[15] A. J. Rucklidge, G. E. Morris, A. M. Z. Slawin, D. J. Cole-Hamilton,
Helv. Chim. Acta 2006, 89, 1783–1800.
Received: March 23, 2013
Published online: November 20, 2013
17140
www.chemeurj.org
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 17131 – 17140