Single-step access to long-chain α,ω-dicarboxylic acids by isomerizing hydroxycarbonylation of unsaturated fatty acids

Article abstract of DOI:10.1021/acscatal.6b02622

Dicarboxylic acids are compounds of high value, but to date long-chain α,ω-dicarboxylic acids have been difficult to access in a direct way. Unsaturated fatty acids are ideal starting materials with their molecular structure of long methylene sequences and a carboxylate functionality, in addition to a double bond that offers itself for functionaliza-tion. Within this paper, we established a direct access to α,ω-dicarboxylic acids by combining isomerization and selective terminal carbonylation of the internal double bond with water as a nucleophile on unsaturated fatty acids. We identified the key elements of this reaction: a homogeneous reaction mixture ensuring sufficient contact between all reactants and a catalyst system allowing for activation of the Pd precursor under aqueous conditions. Experiments under pressure reactor conditions with [(dtbpx)Pd(OTf)₂] as catalyst precursor revealed the importance of nucleophile and reactant concentrations and the addition of the diprotonated diphosphine ligand (dtbpxH₂)(OTf)₂ to achieve turnover numbers >120. A variety of unsaturated fatty acids, including a triglyceride, were converted to valuable long-chain dicarboxylic acids with high turnover numbers and selectivities for the linear product of >90%. We unraveled the activation pathway of the Pd^II precursor, which proceeds via a reductive elimination step forming a Pd⁰ species and oxidative addition of the diprotonated diphosphine ligand, resulting in the formation of the catalytically active Pd hydride species. Theoretical calculations identified the hydrolysis as the rate-determining step. A low nucleophile concentration in the reaction mixture in combination with this high energetic barrier limits the potential of this reaction. In conclusion, water can be utilized as a nucleophile in isomerizing functionalization reactions and gives access to long-chain dicarboxylic acids from a variety of unsaturated substrates. The activity of the catalytic system of hydroxycarbonylation ranks as one of the highest achieved for isomerizing functionalizations in combination with a high selectivity for the linear product.

Full text of DOI:10.1021/acscatal.6b02622

This is an open access article published under an ACS AuthorChoice License, which permits
copying and redistribution of the article or any adaptations for non-commercial purposes.
Research Article
pubs.acs.org/acscatalysis
Single-Step Access to Long-Chain α,ω-Dicarboxylic Acids by
Isomerizing Hydroxycarbonylation of Unsaturated Fatty Acids
Verena Goldbach,† Laura Falivene,‡ Lucia Caporaso,*,§ Luigi Cavallo,‡ and Stefan Mecking*,†
†Chair of Chemical Materials Science, Department of Chemistry, University of Konstanz, 78464 Konstanz, Germany
‡KAUST Catalysis Center, Physical Sciences and Engineering Division, King Abdullah University of Science and Technology, Thuwal,
23955-6900, Kingdom of Saudi Arabia
§Department of Chemistry, University of Salerno, Via Ponte Don Melillo, 84084-Fisciano, Salerno, Italy
S
* Supporting Information
ABSTRACT: Dicarboxylic acids are compounds of high value,
but to date long-chain α,ω-dicarboxylic acids have been
difficult to access in a direct way. Unsaturated fatty acids are
ideal starting materials with their molecular structure of long
methylene sequences and a carboxylate functionality, in
addition to a double bond that offers itself for functionaliza-
tion. Within this paper, we established a direct access to α,ω-
dicarboxylic acids by combining isomerization and selective
terminal carbonylation of the internal double bond with water
as a nucleophile on unsaturated fatty acids. We identified the
key elements of this reaction: a homogeneous reaction mixture
ensuring sufficient contact between all reactants and a catalyst
system allowing for activation of the Pd precursor under
aqueous conditions. Experiments under pressure reactor
conditions with [(dtbpx)Pd(OTf)2] as catalyst precursor revealed the importance of nucleophile and reactant concentrations
and the addition of the diprotonated diphosphine ligand (dtbpxH2)(OTf)2 to achieve turnover numbers >120. A variety of
unsaturated fatty acids, including a triglyceride, were converted to valuable long-chain dicarboxylic acids with high turnover
numbers and selectivities for the linear product of >90%. We unraveled the activation pathway of the PdII precursor, which
proceeds via a reductive elimination step forming a Pd0 species and oxidative addition of the diprotonated diphosphine ligand,
resulting in the formation of the catalytically active Pd hydride species. Theoretical calculations identified the hydrolysis as the
rate-determining step. A low nucleophile concentration in the reaction mixture in combination with this high energetic barrier
limits the potential of this reaction. In conclusion, water can be utilized as a nucleophile in isomerizing functionalization reactions
and gives access to long-chain dicarboxylic acids from a variety of unsaturated substrates. The activity of the catalytic system of
hydroxycarbonylation ranks as one of the highest achieved for isomerizing functionalizations in combination with a high
selectivity for the linear product.
KEYWORDS: isomerizing functionalization, fatty acid functionalization, water as nucleophile, Pd-catalyzed carbonylation, Pd hydride,
aqueous reaction medium
maintain the entire sequence of methylene groups originating
from the fatty acid substrates are of particular interest, but their
synthesis often involves multiple steps or extensive workup.5−7
There are several options for functionalizing a fatty acid in an
atom-efficient way by completely utilizing the feedstock and
retaining their unique molecular structure with the long
hydrocarbon backbone. Biotechnological approaches offer one
possibility to access those valuable compounds in one step via
enzymatic pathways from a variety of fatty acids,8−11 but the
potential of these approaches is limited due to complex
INTRODUCTION
■
Fatty acids from plant and algae oils are attractive substrates
with a unique molecular structure that can be used to generate
higher-value chemicals.1,2 Unsaturated fatty acids with their
double bonds can be functionalized by chemical catalysis or
enzymatic approaches, which yield a second functionality in
addition to the carboxylic acid group that is already present in
those fatty acids.3
Linear α,ω-dicarboxylic acids are important organic com-
pounds. For example, one of the most important commodity
dicarboxylic acids is adipic acid, which is used as a monomer for
polyamide nylon-6,6 production. A typical midchain-length
dicarboxylic acid, such as sebacic acid, is easily accessible from
ricinoleic acid.4 However, long-chain α,ω-dicarboxylic acids that
Received: September 13, 2016
Revised: October 20, 2016
© XXXX American Chemical Society
8229
DOI: 10.1021/acscatal.6b02622
ACS Catal. 2016, 6, 8229−8238

ACS Catalysis
Research Article
downstreaming processes and the necessity of feeding costly
glucose.3 Transition-metal catalysis can generate a range of
functional groups from unsaturated fatty acids: in particular,
carbonyl functionalities such as aldehydes and ester groups that
could be further transformed into carboxylic acid moieties.
Hydroformylation, as the most prominent carbonylation
reaction, can be carried out in an isomerizing fashion, in
which the internal double bond of a fatty acid is converted to a
terminal carbonyl group with high activities (average turnover
frequency (TOF) for conversion of methyl oleate 35 h−1).12
Isomerizing hydroformylation can also be combined with an in
situ reduction, forming valuable ω-hydroxy esters, even though
this occurs with lower activities (average TOF for conversion of
methyl oleate 2.6 h−1).13 However, isomerizing hydroformyla-
tion, also as a tandem hydroformylation−reduction approach,
suffers from a low selectivity toward the desired difunctional-
ized product (40% toward linear aldehyde ester, 56% toward
linear ω-hydroxy ester) due to the formation of thermodynami-
cally slightly favored α,β-unsaturated esters in the isomerization
step and their propensity for hydrogenation. In comparison,
isomerizing alkoxycarbonylation with an alcohol as nucleophile
has a slightly lower reaction rate than isomerizing hydro-
formylation (initial TOF 12 h−1 for conversion of methyl
oleate) but a selectivity toward the linear diester of up to 96%
can be achieved.14−16
A carbonylation of an olefin with water as a nucleophile can
form carboxylic acids directly. Such hydroxycarbonylation
reactions are well established for the generation of short-
chain substrates, mostly from terminal olefins.17−21 Concerning
an envisioned isomerizing carbonylation with water as a
nucleophile, however, the aforementioned related alkoxycarbo-
nylation necessarily requires an alcohol for formation of the
active species.22,23 In addition, also with other catalysts, there
are only a few examples of hydroxycarbonylation of terminal
olefins longer than C8 due to their poor solubility in water or a
low solubility of the nucleophile water in the reaction medium,
both resulting in a low reactivity.18,24 To date, no isomerizing
functionalization reaction has used water as a reactant or was
conducted in the presence of water and the desired direct
access toward linear long-chain dicarboxylic acids via chemical
catalysis is missing.
monounsaturated oleic acid (93%; see the Supporting
Information for details on composition). Oleic acid was then
converted to 1,19-nonadecanedioic acid with water as a
nucleophile, at high pressures of CO catalyzed by a Pd
complex with dtbpx as diphosphine ligand.
A homogeneous reaction mixture allowing for sufficient
contact between all reactants and the catalyst is crucial to
achieve high activities of the Pd catalyst. The immiscibility of
fatty acids and water prohibits a direct hydroxycarbonylation of
oleic acid with CO and water without additional solvent. To
obtain a homogeneous reaction mixture, as assessed under
ambient conditions for practical reasons,25 polar, mainly aprotic
solvents having good miscibility properties with water and fatty
acids were screened for isomerizing hydroxycarbonylation of
oleic acid (Table 1). We chose a reaction temperature of
Table 1. Solvent Screening for Isomerizing
a
Hydroxycarbonylation of Oleic Acid
c
c
entry
solvent
conversion (%)
selectivity (%)
1
2
3
4
5
6
7
8
9
ethyl acetate
2-methyl-THF
DMSO
2.6
14.8
2.9
72.7
42.8
66.9
91.0
89.7
93.2
85.6
92.5
91.3
91.2
92.6
diglyme
34.7
42.3
33.9
41.7
33.9
31.3
52.7
61.0
1,2-dimethoxyethane
1,4-dioxane
methyl ethyl ketone
tBuOH
THF
b
10
1,2-dimethoxyethane
THF
b
11
a
Reaction conditions unless specified otherwise: 25 μmol of
[(dtbpx)Pd(OTf)2], 75 μmol of (dtbpxH2)(OTf)2, 5 mmol of
technical grade oleic acid (Dakolub MB6098, 93% oleic acid content),
1 mL of H2O, 10 mL of solvent, 20 bar of CO (initial pressure), 125
b
°C, 18 h. 50 μmol of [(dtbpx)Pd(OTf)2], 120 μmol of (dtbpxH2)-
c
(OTf)2, 10 mmol of technical grade oleic acid. Conversion and
selectivity for linear products determined by gas chromatography from
crude reaction mixture.
125 °C to overcome potential barriers and added a slight excess
of trifluoromethanesulfonic acid (TfOH) and diphosphine
ligand dtbpx in a 2:1 ratio, as the diprotonated diphosphine
ligand (dtbpxH2)(OTf)2, to stabilize the Pd catalyst under
these conditions.
We now report on a selective formation of long-chain α,ω-
dicarboxylic acids in one step through isomerizing functional-
ization of fatty acids. Key features of this transformation are a
matched solvent and catalyst system, including an activation
pathway specific to water.
With ethyl acetate or 2-methyl-THF as solvent, the reaction
mixtures were not entirely homogeneous under the chosen
conditions, giving turbid suspensions upon mixing of all
components. Consequently, low conversions were observed
(Table 1, entries 1 and 2). By using DMSO, which gave a
homogeneous reaction mixture, oleic acid was only converted
to a minor extent (<3% conversion, entry 3). Since DMSO is
presumably coordinating to the Pd center, active sites on the
catalyst were blocked, causing a low conversion rate. Using
solvents bearing oxygen functionalities that do not coordinate
strongly or react in the carbonylation reaction (entries 4−9)
increased conversions above 30% were observed. With THF as
a solvent the best activities and selectivities toward the desired
linear dicarboxylic acid under pressure reactor conditions were
achieved (entry 9). THF also gave best results regarding activity
and selectivity of the catalyst in experiments with higher oleic
acid concentrations in comparison to 1,2-dimethoxyethane
under identical conditions (entries 10 and 11). THF was also
RESULTS AND DISCUSSION
■
Generation of Dicarboxylic Acids under Pressure
Reactor Conditions. The isomerizing hydroxycarbonylation
of an unsaturated fatty acid, such as oleic acid, with water as a
nucleophile brings about several challenges, such as overcoming
the immiscibility of fatty acids and water and achieving high
reaction rates with high selectivities at the same time. Pd
complexes bearing sterically demanding diphosphine ligands,
such as 1,2-(CH2PtBu2)2C6H4 (dtbpx), have shown excellent
activities and selectivities in other isomerizing carbonylation
reactions in the past14,15,23 and were therefore chosen for this
approach of hydroxycarbonylation. To elucidate the key
features of the catalytic system under practically relevant
conditions, we used technical grade oleic acid as the substrate,
which also contains saturated and multiunsaturated fatty acids
of various chain lengths in addition to the desired
8230
DOI: 10.1021/acscatal.6b02622
ACS Catal. 2016, 6, 8229−8238

ACS Catalysis
Research Article
Table 2. Variation of Temperature, Amount of Solvent, Substrate, and Nucleophile Concentration in Isomerizing
a
Hydroxycarbonylation in Homogeneous Reaction Mixtures
b
b
c
entry
n(OA) (mmol)
V(THF) (mL)
V(H2O) (mL)
c(H2O) (mol L−1
)
T (°C)
conversion (%)
selectivity (%)
TON
6.9
21
1
2
3
4
5
6
7
8
2.5
2.5
2.5
2.5
2.5
5
10
10
10
10
10
10
7
1
1
1
1
1
1
1
1
4.7
4.7
4.7
4.7
4.7
4.4
5.7
3.9
90
110
125
135
150
125
125
125
13.8
41.5
71.1
60.8
28.8
71.5
87.6
61.0
91.8
93.0
92.1
89.5
84.9
93.1
92.8
92.6
36
30
14
72
5
88
10
10
122
a
Reaction conditions: 50 μmol of [(dtbpx)Pd(OTf)2], 120 μmol of (dtbpxH2)(OTf)2, technical grade oleic acid (Dakolub MB6098, 93% oleic acid
b
content), 20 bar of CO (initial pressure), 18 h. Conversion and selectivity for linear products determined by gas chromatography from the crude
c
reaction mixture. Turnover number in units of (mol of products) (mol of Pd)−1.
the solvent of choice for the following experiments because of
its volatility and easy removal from the reaction mixture as well
as miscibility properties with all components, in particular the
dissolution of the long-chain reaction products.
The effect of temperature, amount of solvent, substrate, and
nucleophile concentration was studied, and the results are
shown in Table 2.
In addition to the homogeneity of the reaction mixture, the
reaction temperature also turned out to be a key factor in
achieving a high turnover number (TON) in the conversion of
oleic acid. Experiments under typical carbonylation conditions
with [(dtbpx)Pd(OTf)2] as a catalyst precursor (90 °C, 20 bar
of CO, 18 h) with a homogeneous reaction mixture containing
oleic acid and water in THF surprisingly gave very low activities
and a low conversion (Table 2, entry 1). Increasing the reaction
temperature to 125 °C resulted in a higher TON (Table 2,
entries 1−3). Between 90 and 125 °C the desired linear
product is obtained with a high selectivity of around 92%
(Scheme 1).
dicarboxylic acids, which typically have a higher energetic
barrier,15 also become energetically more feasible under those
conditions, relative to the energetic barrier for the linear
product.
As outlined, a homogeneous reaction mixture25 is crucial to
achieve a satisfactory conversion for this reaction. At the same
time, high concentrations of the reactants are desirable to
maximize the reaction rates. Increasing the amount of oleic acid
and decreasing the amount of THF resulted in higher turnover
numbers (Table 2, entries 6 and 7). Under these conditions, a
high selectivity for the linear dicarboxylic acid and, in particular,
a high substrate conversion were achieved. However, a further
reduction of the amount of THF was limited by the
immiscibility of the fatty acid substrate with water. Reaction
mixtures with larger quantities of water showed phase
separation under ambient conditions, and subsequently,
conversions were significantly lower (see Table S1 in the
Supporting Information). The concentration of the nucleophile
is limited and greatly affects the reaction rates of the
conversion. When even lower concentrations of water were
applied (c(H2O) < 3.3 mol L−1), significant amounts of
palladium black were observed, indicating catalyst instability
(see Table S1). From these observations, we assume that a
certain concentration of H2O is necessary to obtain a stable
catalyst and a satisfactory conversion of the substrate.
Scheme 1. Isomerizing Hydroxycarbonylation of Oleic Acid
To Generate Linear, Long-Chain α,ω-Dicarboxylic Acids
The best TON with a high selectivity toward the desired
linear product was achieved with a further increase in the oleic
acid concentration while the reaction mixture was kept
homogeneous by adapting THF and water ratios (Table 2,
entry 8). The concentration of the nucleophile H2O is relatively
low for this reaction mixture but is still in a significant excess to
the oleic acid substrate.
An evaluation of the CO pressure and its influence on the
activity of the catalyst revealed that the turnover numbers did
not change significantly between 20 and 40 bar of CO and
decreased for lower CO pressures (see Table S2 in the
Supporting Information).
With an average turnover frequency (TOF) of 6.8 h−1 the
isomerizing hydroxycarbonylation of long-chain unsaturated
fatty acids ranks as one of the most active transformations in
isomerizing functionalization reactions. This value is lower than
a typical average TOF in Rh-catalyzed isomerizing hydro-
formylation (35 h−1 for methyl oleate),12 but hydroxycarbony-
lation has a significantly higher selectivity toward the desired
linear product (92% for linear dicarboxylic acid in hydrox-
ycarbonylation vs 40% for linear aldehyde ester in hydro-
formylation). Isomerizing alkoxycarbonylation of methyl oleate
At even higher reaction temperatures lower activities were
observed (Table 2, entries 4 and 5). Substantial amounts of Pd
black were observed for temperatures above 135 °C, suggesting
catalyst decomposition at higher temperatures. At the same
time, selectivity toward the desired linear product decreased
slightly, as competing pathways forming alkyl-branched
8231
DOI: 10.1021/acscatal.6b02622
ACS Catal. 2016, 6, 8229−8238

ACS Catalysis
Research Article
with an initial TOF of 12 h−1 is the only isomerizing
functionalization outranking hydroxycarbonylation of fatty
acids regarding catalyst activity with a comparable selectivity
for the linear products.3,15 However, it is important to take into
account that the concentration of the nucleophile affects these
reaction rates directly. A typical nucleophile concentration, e.g.
of methanol, that serves as solvent and nucleophile in
alkoxycarbonylation, is five times higher in comparison to the
water concentration in a solvent−water−oleic acid mixture as
needed for hydroxycarbonylation, which highlights the
potential of hydroxycarbonylation to functionalize unsaturated
substrates with excellent catalyst activities.
Key Features of the Catalytic System under Pressure
Reactor Conditions. The role of the excess of diphosphine
ligand and strong, but weakly coordinating acid TfOH, also as
the diprotonated diphosphine ligand (dtbpxH2)(OTf)2, was
further elucidated under pressure reactor conditions. Previously
optimized reaction conditions of temperature, nucleophile and
substrate concentrations, solvent, CO pressure, and [(dtbpx)-
Pd(OTf)2] as a catalyst precursor were applied for the
conversion of oleic acid (Table 3).
entries 1−4), it becomes evident that the reaction rates are
significantly enhanced in the presence of (dtbpxH2)(OTf)2.
The addition of 1 equiv of (dbtpxH2)(OTf)2 to the precursor
[(dtbpx)Pd(OTf)2] already resulted in a 2-fold increase of the
conversion of oleic acid. An excess greater than 2.5 equiv in
relation to the palladium precursor did not result in further
substantial increases in TONs.
In contrast to this, the addition of the acid TfOH or the
diphosphine dtbpx alone to the catalyst precursor [(dtbpx)Pd-
(OTf)2] decreased the conversion of oleic acid significantly in
all experiments. In the presence of TfOH, substantial amounts
of palladium black were observed after the carbonylation
reaction, originating from catalyst decomposition (Table 3,
entries 5 and 6). The presence of additional dtbpx to the
defined diphosphine-coordinated precursor [(dtbpx)Pd(OTf)2]
gave rise to even lower conversions (<5%) to the
difunctionalized product (Table 3, entries 7 and 8). It can be
assumed that an excess of diphosphine blocks free coordination
sites on the Pd center and therefore deactivates the catalyst. We
believe that a strong acid alone, such as TfOH, is not acting as a
cocatalyst or promoter in this catalytic system, in contrast to
what is often assumed for other carbonylation reactions.26
However, the 2:1 addition of acid TfOH and ligand dtbpx as
the protonated ligand (dtbpxH2)(OTf)2 obviously has a major
beneficial impact on the activity of the Pd catalyst.
On the basis of these observations, we investigated whether
the protonated ligand (dtbpxH2)(OTf)2 can form a catalytically
active species in situ with a Pd precursor and how activities
compare to those for a catalyst generated from a Pd precursor
with separate dtbpx and TfOH (Table 4). In situ generation of
the active catalyst with Pd(OAc)2 or PdCl2 as a metal
precursor, the diphosphine ligand dtbpx, and the acid TfOH
resulted in low conversions and turnover numbers for both
cases, while the selectivity for the linear products remained
above 90% (Table 4, entries 1 and 2). It is noteworthy that a
larger excess of TfOH with these in situ systems did not have
any effect on the activity of the catalyst (entry 3), in contrast to
observations made with the defined precursor [(dtbpx)Pd-
(OTf)2] (vide supra). Remarkably, a 3-fold higher TON was
observed when using (dtbpxH2)(OTf)2 instead of separate
dtbpx and TfOH (Table 4, entries 4−6). These results suggest
that the catalytically active species, most likely a Pd hydride, is
formed very effectively in the presence of the protonated ligand.
In addition to catalysis experiments with PdII precursors,
[Pd(dba)2] as a Pd0 precursor was also applied under the same
reaction conditions (Table 4, entry 6). This reaction with a Pd0
Table 3. Variation of Excess Ligand dtbpx and Acid TfOH in
a
Isomerizing Hydroxycarbonylation of Oleic Acid
b
b
n(dtbpx)
(μmol)
n(TfOH)
(μmol)
conversion
(%)
selectivity
(%)
c
entry
TON
1
2
3
4
5
6
7
8
0
0
20.0
49.1
61.0
64.4
7.4
89.7
92.6
92.6
92.7
92.8
93.7
90.9
n.d.
40
98
50 (as (dtbpxH2)(OTf)2)
120 (as (dtbpxH2)(OTf)2)
170 (as (dtbpxH2)(OTf)2)
122
129
15
0
0
120
550
0
2.2
4.4
4.2
n.d.
120
550
2.1
0
<1
a
Reaction conditions: 50 μmol of [(dtbpx)Pd(OTf)2], 10 mmol of
oleic acid (technical grade, Dakolub MB6098, 93% oleic acid content),
10 mL of THF, 1 mL of H2O, 20 bar of CO (initial pressure), 125 °C,
b
18 h. Conversion and selectivity for linear products determined by
c
gas chromatography from crude reaction mixture. Turnover number
in units of (mol of products) (mol of Pd). n.d. = not determined.
By a comparison of conversions and TONs of isomerizing
hydroxycarbonylation of oleic acid with varying amounts of
additional acid TfOH and diphosphine ligand dtbpx, in a 2:1
ratio as the diprotonated ligand (dtbpxH2)(OTf)2 (Table 3,
Table 4. Comparison of Catalytic Performance of in Situ Generated Catalysts in Isomerizing Hydroxycarbonylation of Oleic
a
Acid
b
b
c
entry
Pd precursor
n(dtbpx) (μmol)
n(TfOH) (μmol)
conversion (%)
selectivity (%)
TON
d
e
1
2
3
4
5
6
Pd(OAc)2
PdCl2
170
240
18.0
17.3
18.1
53.3
40.7
53.9
92.0
90.6
91.2
91.7
93.1
91.1
36
35
d
e
170
240
d
Pd(OAc)2
Pd(OAc)2
PdCl2
170
670
36
170 (as (dtbpxH2)(OTf)2)
107
81
170 (as (dtbpxH2)(OTf)2)
170 (as (dtbpxH2)(OTf)2)
[Pd(dba)2]
108
a
Reaction conditions: 50 μmol of Pd precursor, 10 mmol of oleic acid (technical grade, Dakolub MB6098, 94.3% oleic acid content), 10 mL of THF,
b
1 mL of H2O, 20 bar of CO (initial pressure), 125 °C, 18 h. Conversion and selectivity for linear products determined by gas chromatography from
c
d
crude reaction mixture. Turnover numbers in units of (mol of products) (mol of Pd)−1. A 170 μmol amount of dtbpx corresponds to the amount
e
of diphosphine ligand present in previous experiments with 50 μmol of [(dtbpx)Pd(OTf)2] and 120 μmol of (dtbpxH2)(OTf)2. A 240 μmol
amount of TfOH corresponds to the molar amount of protons that are present when using 120 μmol of (dtbpxH2)(OTf)2 in addition to
[(dtbpx)Pd(OTf)2] (2 equiv of TfOH for generation of 120 μmol of (dtbpxH2)(OTf)2).
8232
DOI: 10.1021/acscatal.6b02622
ACS Catal. 2016, 6, 8229−8238

ACS Catalysis
Research Article
Scheme 2. Generation of Long-Chain Dicarboxylic Acids from Fatty Acid Derivatives via Pd-Catalyzed Isomerizing
Hydroxycarbonylation
precursor showed a high turnover number that is very similar to
the performance of the PdII precursor Pd(OAc)2, indicating
that the protonated ligand can also activate a Pd0 precursor to
generate the catalytically active PdII hydride species. This
observation is important and will be discussed in more detail in
the following section on the formation of the catalytically active
Pd hydride species. Other Pd0 precursors, such as [Pd(PPh3)4]
and [Pd(PtBu3)2], were also studied under similar conditions
and showed lower turnover numbers (see Table S3 in the
Supporting Information). These observations can be accounted
for by the presence of coordinating phosphine ligands from the
precursors, blocking coordination sites of the active catalyst.
When the defined diphosphine-coordinated precursor
[(dtbpx)Pd(OTf)2] is used, additional coordinating ligands or
anions are not present and, therefore, this precursor gives the
best catalyst performance in isomerizing hydroxycarbonylation
under the given conditions. The various experiments with
different Pd precursors illustrated that the protonated ligand
(dtbpxH2)(OTf)2 has a crucial role in achieving high activities
in hydroxycarbonylation. Pressure reactor studies alone do not
allow for a conclusive picture of the mechanistic function of the
protonated diphosphine ligand, and NMR studies are outlined
below.
conditions for this catalytic system. These results are in unison
with observations of the related isomerizing alkoxycarbonyla-
tion of internal alkenes, including different octenes, and a low
energetic barrier for the isomerization, which was confirmed by
DFT calculations.15,27,28
Various fatty acids and derivatives were used in isomerizing
hydroxycarbonylation to generate valuable linear, long-chain
dicarboxylic acids (Scheme 2 and Table 5).
Table 5. Substrate Scope of Isomerizing
Hydroxycarbonylation for the Production of Linear Long-
a
Chain Dicarboxylic Acids
b
b
entry
substrate
conversion (%)
selectivity (%)
1
2
3
4
10-undecenoic acid
oleic acid
94.6
61.0
56.9
47.4
91.3
92.6
88.8
88.1
high oleic sunflower oil
erucic acid
a
Reaction conditions: 50 μmol of [(dtbpx)Pd(OTf)2], 120 μmol of
(dtbpxH2)(OTf)2, 10 mmol of fatty acid derivative (entries 1, 2, 4) or
3.33 mmol of high oleic sunflower oil (oleic acid triglyceride 93%), 10
mL of THF, 1 mL of H2O, 20 bar of CO (initial pressure), 125 °C, 18
b
h. Conversion and selectivity for linear products determined by gas
chromatography from crude reaction mixture.
Substrate Scope of Isomerizing Hydroxycarbonylation. A
selection of unsaturated substrates was further utilized in
isomerizing hydroxycarbonylation to explore the scope of the
reaction and to shed light on its mechanism.
From the results in Table 5, it is evident that the conversion
for different substrates from C10 up to C22 decreased with an
increasing chain length of the substrate. This observation can
be accounted for by our previous mechanistic studies on
isomerizing alkoxycarbonylation:15 the longer the hydrocarbon
chain is, the more isomeric metal species are formed and,
therefore, the concentration of the reactive terminal Pd alkyl
and acyl species is lower. The rate-determining step, solvolysis,
has a lower energy barrier for the terminal Pd acyl species in
comparison to all other possible branched Pd acyls. With a
lower concentration of the terminal species, the reaction rate
decreases consequently. The selectivity also decreases to a
minor extent for the conversion of the longer-chained erucic
acid in comparison to oleic acid due to a larger amount of
possible branched byproducts resulting from branched Pd acyl
To investigate if the position of the double bond within the
substrate molecule has an influence on activity or selectivity of
the catalyst in this reaction, 1- and 4-octene were applied to
typical isomerizing hydroxycarbonylation reaction conditions
(0.5 mol % of [(dtbpx)Pd(OTf)2], 1.2 mol % of (dtbpxH2)-
(OTf)2, 10 mL of THF, 1 mL of H2O, 20 bar of CO, 125 °C,
18 h). Both substrates were converted with virtually identical
rates (73% conversion in 3 h) and high selectivities (95%
selectivity with 1- and 4-octene for linear nonanoic acid).
Therefore, it can be concluded that isomerization to the
thermodynamic equilibrium of all double-bond isomers is fast,
and the original position of the double bond in the substrate
has no influence on activity or selectivity under pressure reactor
8233
DOI: 10.1021/acscatal.6b02622
ACS Catal. 2016, 6, 8229−8238

ACS Catalysis
Research Article
1
Figure 1. H NMR (left) and 31P{1H} NMR spectra (right) of relevant signals monitored over time for observation of Pd hydride formation.
intermediates. Nevertheless, it is noteworthy that the C22 erucic
acid was converted to the valuable long-chain 1,23-dicarboxylic
acid with an excellent TON of 95 and a selectivity of 89% under
the same conditions as for the C19 oleic acid.
insertion that rapidly undergoes elimination of CO2 and forms
a Pd hydride species.19 However, literature is not conclusive on
the formation mechanism of the active Pd hydride species,
suggesting various pathways depending on the catalytic
system.19,22,23,30
It is remarkable that not only fatty acids themselves but also
triglycerides, such as high oleic sunflower oil, can be used
directly in isomerizing hydroxycarbonylation, forming the
desired 1,19-dicarboxylic acid (Table 5, entry 3). Under the
conditions identified for isomerizing hydroxycarbonylation
(Tables 1 and 2), the ester bonds of the triglyceride are
hydrolyzed, and carbonylation occurs on the terminal chain end
simultaneously, generating the long-chain α,ω-dicarboxylic
acids in one step. Although the reaction mixture showed
phase separation upon mixing all components under ambient
conditions, there is only a minor difference between conversion
values of oleic acid and its triglyceride (Table 5, entry 2 vs 3).
After 18 h of reaction time, the reaction mixture was completely
homogeneous, and no separated phases were observed. The
slightly lower conversion rate of high oleic sunflower oil could
be a result of the inhomogeneous reaction mixture at the
beginning of the reaction29 and a minimally lower concen-
tration of water over the course of the reaction, as some water is
consumed by the ester hydrolysis of the triglycerides.
With these experiments, we showed that hydroxycarbonyla-
tion is not limited to short-chain or terminal alkenes but can
also be utilized in combination with isomerization to generate
high-value (di)carboxylic acids from a range of unsaturated
substrates with a Pd catalyst and a bulky diphosphine ligand.
Alkenes with terminal or internal double bonds, unsaturated
fatty acids of various chain lengths, and technical grade plant
oils are converted with remarkable selectivities to the desired
linear (di)carboxylic acids.
Formation of the Catalytically Active Hydride Species. To
detect and observe the formation of the key compound of the
catalytic cycle, the active Pd hydride complex, we transferred
our catalytic system of the isomerizing hydroxycarbonylation
from the pressure reactor to an NMR tube experiment. Upon
dissolution of the catalyst precursor [(dtbpx)Pd(OTf)2] in the
presence of an excess of (dtbpxH2)(OTf)2 in THF and water
(0.6 mL of THF, 0.15 mL of H2O, sealed capillary filled with
DMSO-d6 as “lock stick”) the diaqua complex [(dtbpx)Pd-
(H2O)2]2+ was formed rapidly, as observed via 31P{1H} NMR
spectroscopy (singlet at 54.3 ppm). From this Pd species we
were able to follow the slow formation of the hydride
1
[(dtbpx)PdH(H2O)]+ via H and 31P{1H} NMR spectroscopy
at room temperature (Figure 1). The 1H NMR spectra reveal a
characteristic doublet of doublets at −10.93 ppm (2JPH,cis = 22.5
2
Hz, JPH,trans = 188.4 Hz) corresponding to the water-
coordinated Pd hydride species [(dtbpx)PdH(H2O)]+. The
corresponding signals of two inequivalent phosphorus atoms
were detected via 31P{1H} NMR spectroscopy at 23.0 and 73.5
2
ppm (both JPP = 19.0 Hz). These characteristic signals match
well with data reported by Clegg et al., who were also able to
observe the formation of the Pd hydride [(dtbpx)PdH(H2O)]+
in a THF/water mixture.22 Further, we observed a decline in
the singlet of the initially formed [(dtbpx)Pd(H2O)2]2+ in the
31P{1H} NMR spectra (Figure 1, orange boxes), while the
intensities of the two doublets corresponding to [(dtbpx)PdH-
(H2O)]+ increase (Figure 1, blue boxes). Additional hydride
Mechanistic Considerations for Isomerizing Hydrox-
ycarbonylation. For hydroxycarbonylation reactions, it is
proposed that the catalytic cycle starts with a palladium hydride
species. A “hydroxy-cycle”, starting from a Pd−OH species in
the presence of water, is unlikely, as this would involve the
formation of a highly unstable Pd−COOH species by CO
1
signals of low intensity around −18.3 ppm in the H NMR
spectra and singlets between 60 and 65 ppm in the 31P{1H}
NMR spectra can be assigned to reaction products from
hydride formation that will be discussed later on.
8234
DOI: 10.1021/acscatal.6b02622
ACS Catal. 2016, 6, 8229−8238

ACS Catalysis
Research Article
+
Relating the hydride signals against dimethyl terephthalate as
an internal standard in the 1H NMR spectra, we concluded that
a maximum of approximately 12% of the initial amount of the
precursor [(dtbpx)Pd(OTf)2] was transformed into the
observed hydride [(dtbpx)PdH(H2O)]+ (see Figures S2 and
S3 in the Supporting Information). The hydride formation is
comparably slow at room temperature, and after 12 h at 300 K
the maximum amount of Pd hydride species was detected with
no further significant increase at longer reaction times. At
elevated temperatures, the hydride formation is significantly
faster. About 10% of Pd hydride was already formed within 20
min from [(dtbpx)Pd(H2O)2]2+ and (dbtpxH2)(OTf)2 at 50
°C. Nevertheless, under NMR conditions, even at elevated
temperatures, the Pd precursor is virtually consumed
completely, but the amount of generated Pd hydride never
surpassed the aforementioned maximum of hydride present in
this reaction mixture (see Figure S4 in the Supporting
Information).
Note that in the presence of atmospheric pressure of CO in
the NMR tube no Pd hydride species were detected at room
temperature with [(dtbpx)Pd(OTf)2] and (dtbpxH2)(OTf)2 in
THF/H2O. van Leeuwen et al. have also observed this issue in
their mechanistic investigations on hydroxycarbonylation of
styrene with various Pd phosphine catalyst systems.19 However,
we believe that the Pd hydride [(dtbpx)PdH(L)] (L = H2O,
CO, THF) as the catalytically active species is formed in
sufficient amounts in the presence of CO at higher temper-
atures and at higher pressures, since isomerizing hydroxycarbo-
nylation of oleic acid was successfully run at 125 °C and 20 bar
of CO in a pressure reactor (vide supra).
be identified by ESI-MS (m/z calcd for C24H45OP2 411.29
[dtbpxO + H]+, found 411.28), as well as various other dtbpx-,
dtbpxO-, and dtbpxO2-coordinated Pd species (see Figure S7).
After a longer reaction time, ESI-MS shows only two major sets
of signals at m/z 427.28 and 449.26, both corresponding to the
+
diphosphine dioxide dtbpxO2 (m/z calcd for C24H45O2P2
427.29 [dtbpxO + H]+, found 427.28; m/z calcd for
C24H44NaO2P2+ 449.27 [dtbpxO2 + Na]+, found 449.26; Figure
S7). These observations confirm the oxidation of the
diphosphine ligand via reductive elimination from the PdII
precursor [(dtbpx)Pd(OTf)2] in the presence of water. The
resulting Pd0 species, which is inevitably formed during this
process, is then a crucial reactant in the next step.
While the experiment mentioned above allowed for a clear
analysis of the diphosphine oxidation, we were not able to
detect the hydride [(dtbpx)PdH(H2O)]+ via the characteristic
1
signals (dd around −11 ppm) in H NMR spectra in the
absence of the protonated ligand (dtbpxH2)(OTf)2. This
confirms the active role of the protonated diphosphine ligand
in the hydride formation. NMR experiments with an internal
standard proved that the amount of protonated ligand
(dtbpxH2)(OTf)2 declined over time during hydride formation
(see Figure S3 in the Supporting Information). Since a Pd0
species is present from the preceding reductive elimination
reaction, we assume that (dtbpxH2)(OTf)2 undergoes an
oxidative addition, forming a Pd hydride species. The role and
importance of the protonated ligand are further emphasized by
pressure reactor experiments, for which a lower amount of
(dtbpxH2)(OTf)2 resulted in significantly lower conversion of
oleic acid to the 1,19-dicarboxylic acid (cf. Table 3, entries 1−
4). With a lower amount of protonated ligand, less active Pd
hydride is formed and consequently results in a lower
conversion rate.
As previously mentioned, in addition to signals originating
from the Pd hydride or starting materials, the intensity of
singlets between 60 and 65 ppm in the 31P{1H} NMR spectra
increased during hydride formation. These singlets can be
assigned to phosphine oxides that are formed via reductive
elimination from the diaqua complex [(dtbpx)Pd(H2O)2]2+
(see the Supporting Information for comparative NMR data of
phosphine oxides dtbpxO and dtbpxO2, see also Figure S5),
with the singlet at 65.4 ppm corresponding to the diphosphine
dioxide dtbpxO2. By an ESI-MS measurement of this NMR
reaction mixture, the formation of dtbpxO2 was confirmed (m/
z calcd for C24H45O2P2+ 427.29 [dtbpxO2 + H]+, found 427.28;
The feasibility of an oxidative addition of the protonated
diphosphine ligand to a Pd0 species was demonstrated by using
[Pd(PtBu3)2] as a molecular, soluble Pd0 precursor with
(dtbpxH2)(OTf)2 in THF. We observed the formation of the
hydride [(dtbpx)PdH(THF)]+ by 1H and 31P{1H} NMR
spectroscopy (see Figure S8 in the Supporting Information).
An oxidative addition of a protonated phosphine ligand with a
Pd0 species has also been reported for other Pd catalysts.35
On the basis of this compiled evidence, we propose a two-
step process as the major pathway for the formation of the
catalytically active Pd hydride species in water. First, the diaqua
complex [(dtbpx)Pd(H2O)2]2+ undergoes reductive elimina-
tion, forming a Pd0 species and dtbpx oxides. These Pd0 species
are then accessible for the second step, an oxidative addition of
the protonated ligand, which is usually present in excess in the
reaction mixtures, to generate the desired Pd hydride species
[(dtbpx)PdH(H2O)]+ (Scheme 3).
+
m/z calcd for C24H44NaO2P2 449.27 [dtbpxO2 + Na]+, found
449.26). The additional hydride signals around −18.3 ppm in
1
the H NMR spectra (Figure 1) correlate to the small singlets
(61.4, 64.4 ppm) in the 31P{1H} NMR spectrum, indicating the
coordination of a phosphine oxide to a Pd hydride species. The
amount of these phosphine oxide Pd hydride species is usually
below 2% (see Figure S2 in the Supporting Information) and,
therefore, is negligible in terms of the fate of the metal.
This process of a reductive elimination generating a
phosphine oxide and a Pd0 species in the presence of water
was already investigated for a number of Pd systems with
mono- and diphosphines as ligands31−33 and can be expanded
to our catalytic system starting from [(dtbpx)Pd(OTf)2] as a
precursor. NMR spectroscopy did not allow for a complete
assignment of all phosphorus species in this reaction mixture
(see the Supporting Information for more details). Therefore,
we followed the formation of the diphosphine oxides in the
absence of the protonated ligand (dtbpxH2)(OTf)2 in THF and
water by ESI-MS (see Figure S7 in the Supporting
Information).34 At an early stage of the oxidation (after 3 h
at room temperature), the diphosphine monoxide dtbpxO can
Scheme 3. Proposed Pathway for Formation of the
Catalytically Active Pd Hydride Species via Reductive
Elimination (1) and Subsequent Oxidative Addition (2)
8235
DOI: 10.1021/acscatal.6b02622
ACS Catal. 2016, 6, 8229−8238

ACS Catalysis
Research Article
This two-step activation of the Pd precursor with the
protonated ligand as a crucial reaction partner differs from
other suggested pathways of hydride formation of a Pd
precursor in carbonylation reactions. A β-H elimination
pathway, which is assumed for hydride formation of a Pd
precursor with a primary or secondary alcohol,22 is not feasible
for water. As already mentioned, Clegg et al. observed the
formation of a hydride species in THF/H2O as the reaction
solvent and suggested an oxidation of water to generate the
observed Pd hydride and O2.22 We did not observe any gas
evolution during hydride formation, and therefore, this reaction
pathway was ruled out for our catalytic system with the PdII
precursor and excess protonated ligand. A different approach
proposes that a Pd hydride species can be formed by a water-
gas shift reaction in the presence of CO and water.30 As
elucidated, we were not able to detect any Pd hydride in the
presence of CO under ambient conditions, indicating that CO
does not play an essential role in activation of our Pd precursor
for isomerizing hydroxycarbonylation. However, under pressure
reactor conditions CO can act as another reducing agent for the
PdII precursor,19,36 generating Pd0 that is available for the
oxidative addition of the protonated ligand to form the active
Pd hydride species.
quantitative, due to other competing reaction pathways,
which is decisive for the catalyst performance.
Hydrolysis as the Rate-Determining Step. To gain an
insight into the rate-limiting factors of hydroxycarbonylation,
we investigated the crucial step of hydrolysis in isomerizing
hydroxycarbonylation. This choice is based on the finding that
alcoholysis is the rate-determining step in the related
alkoxycarbonylation.15,23 The decisive step of hydrolysis was
examined by density functional theory (DFT) calculations, as
NMR experiments alone do not allow for a comprehensive
access to this important step and its overall energy profile.37
The dtbpx PdII acyl species [(dtbpx)PdC(O)-
(CH2)6COOH]+ has been considered as a zero point energy
reference (for computational details, see the Supporting
Information). The coordination of H2O to Pd and the
following insertion into the Pd−C bond leading to the
formation of the linear dicarboxylic acid, plus the Pd hydride
species, were calculated (Figure 2). For comparison, the
This proposed pathway for the hydride formation also
explains the high activities observed for the in situ approaches
with Pd0 precursors (cf. Table 4). By using a Pd0 precursor, the
entire amount of Pd is accessible for the direct oxidative
addition of (dtbpxH2)(OTf)2 and reduction of the PdII
precursor, which is possibly incomplete, is not required. This
could eventually result in a larger amount of catalytically active
PdII hydride species that participates in the catalytic cycle and
the transformation of the unsaturated fatty acids to linear
dicarboxylic acids. The utilization of Pd(OAc)2 as a PdII
precursor resulted in a high activity, comparable to results
obtained with [Pd(dba)2] as precursor in isomerizing
hydroxycarbonylation (TON for Pd(OAc)2 106.6 and for
[Pd(dba)2] 107.8). This high activity is a result of the very
effective reductive elimination and presumably complete
generation of Pd0 from Pd(OAc)2, as this precursor is known
to effectively undergo reductive elimination and oxidation of
phosphines with water.33
The formation of the catalytically active Pd hydride species
with a mechanism that is specific to the reaction with water is
not rate-limiting in the isomerizing hydroxycarbonylation. The
reaction forming the hydride is comparably slow at room
temperature, but significantly faster at elevated temperatures
under NMR conditions. We assume that the hydride formation
occurs rapidly under pressure reactor conditions, especially at
125 °C. Pressure reactor studies showed no delay or induction
period of conversion of oleic acid (see Figure S1 in the
Supporting Information). This observation was further
confirmed by isomerizing hydroxycarbonylation experiments
of oleic acid with an additional period allowing for hydride
formation in the absence of CO by stirring at room
temperature or elevated temperatures before pressurizing.
These experiments showed almost identical conversion in
comparison to experiments without additional reaction time for
hydride formation (see Table S4 in the Supporting
Information). Note that, under NMR tube conditions, only a
small amount of catalyst precursor was converted to the desired
catalytically active Pd hydride species. Presumably, the hydride
formation under pressure reactor conditions is also not
Figure 2. Energy profile (THF, in kcal mol−1) of the hydrolysis (blue)
and methanolysis (orange) of the linear dtbpx Pd acyl species (P P =
dtbpx).
̑
energetic profile for the methanolysis of the Pd acyl species is
also reported in Figure 2. For MeOH,15 the energy profile
involving a cluster of three molecules of water is favored versus
a coordination of a single H2O molecule (the difference in
energetic barriers is almost 3 kcal mol−1). From calculations,
the energetic barrier for the hydrolysis is 30.1 kcal mol−1 and is
1.5 kcal mol−1 higher than that for the corresponding
methanolysis reaction reaction (this preference for methanol-
ysis over hydrolysis is confirmed by using different functionals;
see Table S6 in the Supporting Information).
In detail, both hydrolysis and methanolysis involve a cluster
of three hydrogen-bonded molecules in which one ROH
species is coordinated to the metal and another is oxygen-
bonded to the acyl chain. The reductive elimination of the
product occurs through a transition state in which the −OR
moves from the coordinated ROH molecule to the acyl moiety
and the excess of the proton is transferred to the Pd center via
proton shuttle across the cluster, forming the Pd−H species
(Figure 3).
A detailed analysis of the TS geometries reveals that for the
hydrolysis an additional Pd−H interaction, involving the H2O
molecule bonded to the acyl moiety, occurs (see the short
distance in Figure 3a) so that two different H2O molecules
interact with the Pd center in the TS. During the elimination
8236
DOI: 10.1021/acscatal.6b02622
ACS Catal. 2016, 6, 8229−8238

ACS Catalysis
Research Article
generated Pd0, resulting in the catalytically active Pd hydride
species. The hydride formation is presumably not quantitative,
limiting the reaction rates to some extent. Further, the
importance of the protonated diphosphine ligand was
demonstrated by pressure reactor studies, as well as NMR
experiments. Theoretical calculations showed that the rate-
determining step of this isomerizing hydroxycarbonylation is
the hydrolysis of the linear Pd acyl species, contributing to a
significant extent to the limitation of the rates of the
isomerizing hydroxycarbonylation in addition to the low
concentration of the reactive terminal species. Experimentally,
the limitation of this functionalization reaction is also given by a
low nucleophile concentration caused by the immiscibility of
fatty acids and water and the necessity of using an organic
solvent. Nevertheless, the activity of the catalytic system of
hydroxycarbonylation with an average TOF of 6.8 h−1 ranks as
one of the highest achieved for isomerizing functionalization
reactions.
Figure 3. Transition state (TS) geometries of hydrolysis (a) and
methanolysis (b). All hydrogen atoms of the Pd acyl species are
omitted for clarity, and the CH3 groups of methanol are shown in
green.
ASSOCIATED CONTENT
* Supporting Information
■
process, electron density moves from the eliminating moieties
to the metal; this additional Pd−H interaction makes it more
difficult to transfer the proton from the H2O-coordinated
molecule to the Pd center. Vice versa, for the methanolysis,
only one methanol molecule of the cluster interacts with the
metal.38
S
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acscatal.6b02622.
Details on experimental procedures, additional pressure
reactor studies, NMR experiments, ESI-MS, and
computational details (PDF)
The high energetic barrier of hydrolysis limits the rate of
isomerizing hydroxycarbonylation significantly. Note that, in
addition to this energetic barrier of 30 kcal mol−1, the
concentration of the reactive linear Pd acyl species is low in a
mixture of all possible branched acyl species,15 also limiting the
reaction rates to a certain extent. Nevertheless, the calculated
energetic barrier matches well with the experimental conditions
identified for isomerizing hydroxycarbonylation. This energetic
barrier can be overcome by a high reaction temperature of 125
°C. The reaction rates obtained with our optimized catalytic
system also match well with the theoretical data (see the
Supporting Information for a detailed description and
comparison of isomerizing hydroxy- and methoxycarbonyla-
tion).
AUTHOR INFORMATION
Corresponding Authors
*E-mail for L.C.: lcaporaso@unisa.it.
*E-mail for S.M.: stefan.mecking@uni-konstanz.de.
Notes
■
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
V.G. gratefully acknowledges support by the Carl-Zeiss-
Foundation through a graduate fellowship. We thank Adrian
Donner for ESI-MS measurements.
CONCLUSION
REFERENCES
■
■
Linear long-chain dicarboxylic acids are valuable compounds,
for example as building blocks for polyamide synthesis, but a
direct access in an atom-efficient approach has not been
achieved to date. Carbonylation with water as a nucleophile
offers the possibility of forming such carboxylic acid
functionalities from alkenes but has not been applied to
internal double bonds. Unsaturated fatty acids with their
internal double bonds and carboxylate functionalities are ideal
starting materials for an isomerizing functionalization reaction
generating long-chain α,ω-dicarboxylic acids.
Within this work, we demonstrated that water can be applied
as a nucleophile in isomerizing hydroxycarbonylation and
turnover numbers above 120 (for oleic acid) were achieved.
This approach is suitable for various substrates with different
positions of the double bond or different chain lengths and
gives access to a range of long-chain α,ω-dicarboxylic acids in a
single step with a selectivity for the linear products of above
90%. With water as a reactant, we unraveled a specific activation
mechanism of our catalytic system consisting of a reductive
elimination step from a PdII precursor forming a Pd0 species
and subsequent oxidative addition of a protonated ligand to the
(1) Biermann, U.; Bornscheuer, U.; Meier, M. A. R.; Metzger, J. O.;
Schafer, H. J. Angew. Chem., Int. Ed. 2011, 50, 3854−3871.
̈
(2) Chikkali, S.; Mecking, S. Angew. Chem., Int. Ed. 2012, 51, 5802−
5808.
(3) Goldbach, V.; Roesle, P.; Mecking, S. ACS Catal. 2015, 5, 5951−
5972.
(4) Naughton, F. C. J. Am. Oil Chem. Soc. 1974, 51, 65−71.
(5) Stempfle, F.; Ortmann, P.; Mecking, S. Chem. Rev. 2016, 116,
4597−4641.
(6) Vilela, C.; Silvestre, A. J. D.; Meier, M. A. R. Macromol. Chem.
Phys. 2012, 213, 2220−2227.
(7) Stempfle, F.; Quinzler, D.; Heckler, I.; Mecking, S. Macro-
molecules 2011, 44, 4159−4166.
(8) Picataggio, S.; Rohrer, T.; Deanda, K.; Lanning, D.; Reynolds, R.;
Mielenz, J.; Eirich, L. D. Nature Biotechnol. 1992, 10, 894−898.
(9) Yang, Y.; Lu, W.; Zhang, X.; Xie, W.; Cai, M.; Gross, R. A.
Biomacromolecules 2010, 11, 259−268.
(10) Zibek, S.; Huf, S.; Wagner, W.; Hirth, T.; Rupp, S. Chem. Ing.
Tech. 2009, 81, 1797−1808.
(11) Lu, W.; Ness, J. E.; Xie, W.; Zhang, X.; Minshull, J.; Gross, R. A.
J. Am. Chem. Soc. 2010, 132, 15451−15455.
(12) Behr, A.; Obst, D.; Westfechtel, A. Eur. J. Lipid Sci. Technol.
2005, 107, 213−219.
8237
DOI: 10.1021/acscatal.6b02622
ACS Catal. 2016, 6, 8229−8238

ACS Catalysis
Research Article
(13) Yuki, Y.; Takahashi, K.; Tanaka, Y.; Nozaki, K. J. Am. Chem. Soc.
2013, 135, 17393−17400.
́
(14) Jimenez-Rodriguez, C.; Eastham, G. R.; Cole-Hamilton, D. J.
Inorg. Chem. Commun. 2005, 8, 878−881.
(15) Roesle, P.; Caporaso, L.; Schnitte, M.; Goldbach, V.; Cavallo, L.;
Mecking, S. J. Am. Chem. Soc. 2014, 136, 16871−16881.
(16) Christl, J. T.; Roesle, P.; Stempfle, F.; Muller, G.; Caporaso, L.;
̈
Cavallo, L.; Mecking, S. ChemSusChem 2014, 7, 3491−3495.
(17) El Ali, B.; Vasapollo, G.; Alper, H. J. Org. Chem. 1993, 58,
4739−4741.
(18) Goedheijt, M. S.; Reek, J. N. H.; Kamer, P. C. J; van Leeuwen, P.
W. N. M. Chem. Commun. 1998, 2431−2432.
(19) del Río, I.; Claver, C.; van Leeuwen, P. W. N. M. Eur. J. Inorg.
Chem. 2001, 2001, 2719−2738.
́ ́
(20) Tortosa Estorach, C.; Orejon, A.; Ruiz, N.; Masdeu-Bulto, A.
M.; Laurenczy, G. Eur. J. Inorg. Chem. 2008, 2008, 3524−3531.
(21) Low, C. H.; Nobbs, J. D.; van Meurs, M.; Stubbs, L. P.; Drent,
E.; Aitipamula, S.; Pung, M. H. L. Organometallics 2015, 34, 4281−
4292.
(22) Clegg, W.; Eastham, G. R.; Elsegood, M. R. J.; Heaton, B. T.;
Iggo, J. A.; Tooze, R. P.; Whyman, R.; Zacchini, S. J. Chem. Soc., Dalton
Trans. 2002, 3300−3308.
(23) Roesle, P.; Durr, C. J.; Moller, H. M.; Cavallo, L.; Caporaso, L.;
̈
̈
Mecking, S. J. Am. Chem. Soc. 2012, 134, 17696−17703.
(24) Verspui, G.; Feiken, J.; Papadogianakis, G.; Sheldon, R. A. J. Mol.
Catal. A: Chem. 1999, 146, 299−307.
(25) If all components of a reaction mixture gave a clear solution
without visible phase separation in the Schlenk tube under ambient
conditions, prior to transferring the reaction mixture into the pressure
reactor, we referred to this as a “homogeneous reaction mixture”. As
shown in Table S1 in the Supporting Information, reaction mixtures
that had separate phases upon mixing under ambient conditions
consequently gave lower conversions under pressure reactor
conditions.
(26) Brennfuhrer, A.; Neumann, H.; Beller, M. ChemCatChem 2009,
̈
1, 28−41.
́
(27) Jimenez-Rodriguez, C.; Foster, D. F.; Eastham, G. R.; Cole-
Hamilton, D. J. Chem. Commun. 2004, 1720−1721.
(28) Pugh, R. I.; Drent, E.; Pringle, P. G. Chem. Commun. 2001,
1476−1477.
(29) Note that this reaction mixture turns homogeneous over the
course of the reaction, presumably as a result of the ester hydrolysis,
forming oleic acid (or its carbonylation products) and glycerol that are
more soluble in the given reaction mixture. In addition, pressure
reaction conditions (elevated temperature and high pressure) could
also enhance the homogeneity of the reaction mixture.
(30) Bianchini, C.; Meli, A.; Oberhauser, W. Dalton Trans. 2003,
2627−2635.
(31) Ozawa, F.; Kubo, A.; Hayashi, T. Chem. Lett. 1992, 21, 2177−
2180.
(32) Amatore, C.; Carre, E.; Jutand, A.; M’Barki, M. A. Organo-
metallics 1995, 14, 1818−1826.
(33) Grushin, V. V. Organometallics 2001, 20, 3950−3961.
(34) Note that [(dtbpx)Pd(OTf)2], dissolved in dry THF, is stable
and no oxidation of dtbpx occurs in the absence of H2O.
(35) Runzi, T.; Tritschler, U.; Roesle, P.; Gottker-Schnetmann, I.;
̈
̈
Moller, H. M.; Caporaso, L.; Poater, A.; Cavallo, L.; Mecking, S.
̈
Organometallics 2012, 31, 8388−8406.
(36) Drent, E.; Budzelaar, P. H. M. J. Organomet. Chem. 2000, 593−
594, 211−225.
(37) A model reaction of the hydrolysis of the Pd acyl species
[(dtbpx)Pd(COMe)Cl] producing acetic acid was rapid in an NMR
tube experiment at room temperature (see Scheme S1 and Figure S9
in the Supporting Information).
(38) From the NPA charge analysis, the charges on Pd in the
hydrolysis and methanolysis TS are −0.22 e and −0.16 e, respectively.
8238
DOI: 10.1021/acscatal.6b02622
ACS Catal. 2016, 6, 8229−8238