Palladium-catalyzed 1,3-butadiene telomerization with methanol. Improved catalyst performance using bis-o-methoxy substituted triarylphosphines

Article abstract of DOI:10.1016/j.jorganchem.2011.02.007

New, improved phosphine ligands for palladium-catalyzed butadiene telomerization with methanol have been discovered. Using high throughput experimentation and an Electrospray Ionization-Mass Spectrometry (ESI-MS) investigation of telomerization catalysts

Full text of DOI:10.1016/j.jorganchem.2011.02.007

Journal of Organometallic Chemistry 696 (2011) 1677e1686
Contents lists available at ScienceDirect
Journal of Organometallic Chemistry
journal homepage: www.elsevier.com/locate/jorganchem
Palladium-catalyzed 1,3-butadiene telomerization with methanol. Improved
catalyst performance using bis-o-methoxy substituted triarylphosphines
,
John R. Briggs a, Henk Hagen b, Samir Julka a, Jasson T. Patton a
*
a The Dow Chemical Company, Corporate R&D, 1776 Building, Lab D-5, Midland, MI 48674, USA
b Dow Benelux B.V., Herbert H. Dowweg 5, P.O. Box 48, 4530 AA Terneuzen, The Netherlands
a r t i c l e i n f o
a b s t r a c t
Article history:
New, improved phosphine ligands for palladium-catalyzed butadiene telomerization with methanol have
been discovered. Using high throughput experimentation and an Electrospray Ionization-Mass Spec-
trometry (ESI-MS) investigation of telomerization catalysts solutions, we have identified phosphines of
the type P(C6H4-2-OMe)2(C6H5ꢀn(X)n), where X is an electron-withdrawing group, as high selectivity,
high activity phosphine ligands for butadiene telomerization with methanol to 1-methoxy-2,7-octadiene
(MOD-1). These ligands were designed to mitigate anaerobic oxidation of phosphines by alkylation
which was shown by Electrospray Ionization-Mass Spectrometry (ESI-MS) studies to correlate with
catalyst death and palladium precipitation in working telomerization catalyst solutions. The best
phosphine-promoted catalysts achieve selectivities to desired 1-methoxy-2,7-octadiene of 94% at high
butadiene conversion, significantly improved over those achieved commercially by triphenylphosphine.
Ó 2011 Elsevier B.V. All rights reserved.
Received 1 December 2010
Received in revised form
3 February 2011
Accepted 7 February 2011
Keywords:
1,3-Butadiene
Telomerization
Electrospray ionization-mass spectrometry
(ESI-MS)
1-Methoxy-2,7-octadiene
1-Octene
1. Introduction
The commercially desired product of the palladium-catalyzed
telomerization of 1,3-butadiene with methanol is 1-methoxy-2,7-
A number of different transition metals have been reported to
catalyze the telomerization of conjugated dienes, but the most
active appears to be palladium [1]. Since the initial reports of its
performance [2], the palladium-catalyzed telomerization of 1,3-
dienes with nucleophiles has become a popular area of research.
The telomerization of 1,3-butadiene with a nucleophile such as
methanol or water has been of particular interest due to the
commercial significance of the resulting products [3]. Ligand
promoters for the palladium-catalyzed reaction include phos-
phines [4], isonitriles [5], and N-heterocyclic carbenes [6]. A large
variety of phosphines have been reported to be effective ligands for
palladium telomerization catalysts, but only a limited study of the
structure-performance correlations has been undertaken [7]. The
mechanism of palladium-catalyzed butadiene telomerization has
been examined by Keim [8], Jolly [9], and Beller [10]. The critical
step required to form methoxy-terminated telomers, carbon-
oxygen bond formation, is proposed to occur either by intra- or
intermolecular nucleophilic attack by methanol or methoxide on
octadiene (MOD-1, 1), (Fig. 1). A significant use of this material is its
ultimate conversion to 1-octene, a monomer in great demand for
a variety of polyolefin applications. In recent years, The Dow
Chemical Company has sought performance improvements in
order to consider commercial production of 1-octene from buta-
diene [11]. The first step in this process is the telomerization of
butadiene with methanol using a homogeneous triphenylphos-
phine-promoted palladium catalyst with a sodium methoxide
promoter to enhance the reaction rate.The most significant by-
products of this reaction are 3-methoxy-1,7-octadiene (MOD-3, 2)
and 1,3,7-octatriene (3) (Fig. 2).
Increased conversions and greater selectivity to
1 would
increase the throughput of the commercial plant and give better
raw material efficiency. While significant improvements in selec-
tivity have been reported in the literature using N-heterocyclic
carbenes as ligands for palladium telomerization catalysts, the
N-heterocyclic carbene ligands are costly, and the resulting cata-
lysts require high concentrations of base promoter for their optimal
performance (typically 1 mol % base relative to butadiene).
Economic constraints generally favor the retrofit of an improved
catalyst into the existing plant, since any new catalyst is required to
perform under conditions that are fully compatible with the
existing plant configuration, hardware, and catalyst handling
procedures.
a
p-allyl complex derived from a C8 hydrocarbyl ligand that is
formed by the oxidative coupling of two butadiene molecules at
palladium (Fig. 1).
* Corresponding author. Tel.: þ1 989 636 1978; fax: þ1 989 638 6225.
E-mail address: jpatton@dow.com (J.T. Patton).
0022-328X/$ e see front matter Ó 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.jorganchem.2011.02.007

1678
J.R. Briggs et al. / Journal of Organometallic Chemistry 696 (2011) 1677e1686
Fig. 1. Proposed intermolecular nucleophilic attack by methanol or methoxide on a palladium
p-allyl complex.
In this study, we have focused our efforts on developing a more
selectivity and productivity, expressed as turnover number in moles
of1permoles ofpalladium ina 2h reaction, arehigherat 5.0M initial
methanol concentration than at 2.0 M methanol. Interestingly, data
for tris(2-methoxyphenyl)phosphine, L8, appears to fall on the same
lines as the isosteric 4-substituted triarylphosphines, suggesting
there may be little steric effect manifested with this phosphine in
this reaction. However, the data for tris-(2,4,6-trimethoxyphenyl)
phosphine, L9, (not shown in Fig. 3) do not correlate well with the
rest of the data, and for this ligand, steric effects (Chi ¼ ꢀ8.1, cone
angle ¼ 184ꢁ) may well be at play. Significantly, both selectivity and
productivity are maximized at low ligand Chi and at the higher
methanol concentration for L1eL8. Under these experimental
conditions the benchmark triphenylphosphine-promoted catalyst
(L4) demonstrated only moderate productivity of 1232 g 1/g Pd, and
a selectivity of only 74% to 1. All catalysts promoted by triar-
ylphosphines with electron-releasing para-substituents exceeded
the performance of the triphenylphosphine-promoted catalyst in
both selectivity and productivity Not only was the catalyst system
derived from L8 the most productive, but it gave among the highest
selectivities observed (93% to 1), making it one of the most selective
phosphine-promoted catalysts reported to date. The very basic and
sterically encumbered tris(2,4,6-trimethoxyphenyl)phosphine also
performed very well under these conditions, although, in contrast to
the other ligands listed in Table 1, its performance decreased at
higher methanol concentration.
Although high throughput (HTP) experimentation offers dramatic
increases in experimental productivity, it also suffers from some
limitations. To fully capture the experimental productivity increases
that HTP experimentation offers, reagent and product transfer is
achieved robotically, preferably by syringing liquids and solutions.
With volatile substrates, however, such as butadiene, this procedure
limits the concentration that can be effectively added without
encountering losses due to expulsion of liquid contents from the
robotic syringes due to the high vapor pressure of such volatile
reagents. We found that we could reproducibly syringe solutions of
butadiene in methylcyclohexane solvent at concentrations no greater
than 2 M, which gave a concentration of butadiene in the reactor of
only about 1 M after dilution with the required minimum amount of
methanol. This butadiene concentration is significantly lower
than that employed in the commercial telomerization process.
Additionally, because of reactor volume limitations of the HTP
selective monodentate triarylphosphine ligand to improve catalyst
performance that would more likely operate within the constraints
of the existing commercial plant. High-throughput experimenta-
tion was utilized on different ligands as depicted in Fig. 3. In
addition, Electrospray Ionization-Mass Spectrometry (ESI-MS)
investigation was carried out on telomerization catalysts solutions
to understand the role of ligand basicity in palladium-catalyzed
telomerization reactions.
2. Results and discussion
2.1. High throughput catalyst screening
While a number of studies indicate a considerable influence of the
phosphine ligand on the performance of palladium catalysts during
butadiene telomerization, a thorough understanding of the effect of
the phosphine structure on catalyst performance has not appeared in
the literature [4]. We began this study to improve catalyst perfor-
mance using high throughput (HTP) methodology to investigate
a diverse library of monodentate-phosphine ligands looking for
improvements in either selectivity to 1 or catalyst productivity. The
general approach and equipment used was similar to that reported
previously from this laboratory [12]. Selected results of our initial HTP
experiments that examined some 50 ligands under a number of
reaction conditions are shown in Table 1. Plots of this data in Fig. 4
show the selectivity to 1 and total productivity of a series which
includes “isosteric” para-substituted triarylphosphines as a function
of Tolman’s ligand electronic parameter (Chi values) [13]. This set of
ligands was included in the experimental design so that we might
explore catalyst performance as a function of ligand electronic
properties in the absence of concurrent changes in ligand steric
properties. Nevertheless, the complete HTP library included sterically
and electronically diverse ligand structures and was undertaken as
a means to identify promising ligand classes for further study. A class
of ligands that performed particularly well was 2-methoxyaryl
phosphines, as they exhibited both high selectivity and catalyst
productivity under the HTP conditions.
The plots in Fig. 4 show a strong correlation between catalyst
productivity and selectivity to 1 with the Tolman ligand Chi value,
with more basic ligands (lower Chi) performing better. Both
Pd(AcAc)2
2 PPh3
OMe
+
+
MeOH/NaOMe
OMe
1
2
3
Fig. 2. Major products of triphenylphosphine/palladium-catalyzed 1,3-butadiene telomerization with methanol.

J.R. Briggs et al. / Journal of Organometallic Chemistry 696 (2011) 1677e1686
1679
X
X
X
P
Y
Y
X
OMe
OMe
OMe
OMe
OMe
P
P
X
X
X
X
X
X
L1 X = NMe2
L2 X = OMe
L3 X = Me
L4 X =H
L5 X = F
L6 X = Cl
L7 X = CF3
L8 X = H
L9 X = OMe
L10 X = H, Y = H
L11 X = H, Y = CF3
L12 X = Cl, Y = H
L13 X = F, Y = H
L14 X = CF3, Y = H
L15 X = H, Y = F
Fig. 3. Substituted triarylphosphines examined.
equipment initial methanol concentrations of 2.0 and 5.0 M were
used in these experiments. Furthermore, a palladium concentration
of 3.2 ꢂ 10ꢀ4 M (40 ppm), somewhat higher than that used
commercially, was employed to ensure that the catalysts were not
unduly affected by catalyst poisons. These choices resulted in the
maximum turnover numbers achievable by the catalyst in these HTP
experiments being ca. 1660, which is much lower than that which is
required commercially. Having employed high throughput experi-
mentation to expediently identify the better-performing ligands, we
turned our efforts to examining the performance of the most prom-
ising ligands under more relevant conditions (i.e. at higher initial
methanol (12.7 M) and butadiene (2.5 M) concentrations, and lower
palladium concentrations 7.7 ꢂ 10ꢀ5 M (10.9 ppm), and at a higher
temperature of 90 ꢁC). These conditions forced us to employ
a different experimental procedure using a Fischer-Porter bottle
which allowed us to work at similar concentrations used commer-
cially. To our surprise, under these new conditions, the most basic
ligands which had performed so well under the high throughput
conditions now performed poorly, exhibiting profoundly depressed
productivities.This data is shown in Table 2.
reaction through a mechanism that mirrors a step in the formation
of the desired reaction product, 1, we sought to determine if this
process was involved in the early loss of catalyst activity when
using the more basic phosphines. It is well known that phosphines
are competent nucleophiles and they have been shown to react
with metal
p-allyls [17]. Thus, we sought to confirm that phosphine
ligands were being converted to phosphonium cations of the type
[Ar3PCH2CH]CHCH2CH2CH2CH]CH2]þ under telomerization con
ditions, and determine if this process could be related to the loss of
catalytic activity observed with the more basic ligands at higher
methanol concentrations. It is known that such phosphonium salts
can be solvolyzed in basic alcoholic media to phosphine oxides [18],
which we had previously seen in catalyst solutions after the reac-
tion, although we could not rule out aerobic oxidation as an
explanation for their formation.
Table 1
Performance data of selected triarylphosphines for butadiene telomerization
obtained from high throughput studies.
The relationship of the ligand basicity, quantified by plotting the
catalysts TON values versus ligand Chi values under the reaction
conditions employed commercially is shown in Fig. 5. The
increasing activity with decreasing ligand Chi value seen at 5.0 M
methanol and 1.0 M butadiene concentrations is not seen under
these conditions. Of particular note was the dramatic decline in the
performance of the catalyst promoted by ligand L2. Under these
conditions, the catalyst promoted by this ligand exhibited poor
activity while maintaining good MOD-1 selectivity.
Ligand
Substituent
Chi13
(
c
)
[MeOH]initial
(M/l)
TON
(mol 1/
mol Pd)
Selectivity
to 1 (mol %)
L1
L2
L3
L4
L5
L6
L7
L8
L9
4-NMe2
4-OMe
4-Me
5.6
10.2
10.5
12.9
15
2
5
2
5
2
5
2
5
2
5
2
5
2
5
2
5
2
5
898
1224
554
1143
259
1012
138
933
33
560
5
409
0
76
87
57
83
38
81
23
74
16
60
11
51
NA
48
87
93
94
93
4-H
4-F
2.2. Decomposition of phosphines during telomerization
4-Cl
16.8
20.6
2.7
Phosphonium salts have been reported as promoters for telo-
merization catalysts by Kuraray [15] suggesting that such species
play a significant role in the phosphine telomerization catalyst
solutions. Indeed, Behr has suggested that phosphonium salts may
be acting as a reservoir of free ligand in telomerization catalysts [3].
Additionally, in studies of the telomerization of isoprene using
palladium catalysts promoted by the very bulky and very basic
4-CF3
91
2-OMe
2,4,6-(OMe)3
1145
1342
1264
862
ꢀ8.1a
Conditions: [Pd] ¼ 3.2 ꢂ 10ꢀ4 M (40 ppm), [NaOMe] ¼ 1.6 ꢂ 10ꢀ3, L/Pd ¼ 2,
T ¼ 80 ꢁC, [butadiene]initial ¼ 1.0 M, reaction time ¼ 2 h. Methylcyclohexane was
added to each reactor to equalize the total reaction volume (5.0 mL). Productivity
and selectivity values are averages of two runs, except in cases where one value was
suspect. In these cases, the data from the higher productivity run is reported.
phosphine tris(2,4,6-trimethoxyphenyl)phosphine,
a phospho-
nium salt of this type has been isolated when the reaction is per-
formed in the absence of other nucleophiles [16]. Based on this
earlier work which had shown that phosphine ligands can be
alkylated to give phosphonium cations during the telomerization
Maximum TON for 100% yield of 1 is 1660. NA e not applicable.
a
Value for L9 was estimated from a value of n(CO) for [Ni(CO)3L] [14].

1680
J.R. Briggs et al. / Journal of Organometallic Chemistry 696 (2011) 1677e1686
TON to MOD-1 versus Ligand Chi
MOD-1 Selectivity versus Ligand Chi
100
90
80
70
60
50
40
30
20
10
0
1600
1400
1200
1000
800
600
400
200
0
R2 = 0.9109
R2 = 0.9118
R2 = 0.9543
R2 = 0.9345
0
5
10 15
Ligand Chi
20
25
0
5
10
Ligand Chi
15
20
25
Fig. 4. Plots of selectivity and productivity to 1 for ligands shown in Table 1 (L9 excluded). ꢃ e [MeOH]initial ¼ 5.0 M; : e [MeOH]initial ¼ 2.0 M; [Pd] ¼ 3.2 ꢂ 10ꢀ4 M (40 ppm);
[NaOMe] ¼ 1.6 ꢂ 10ꢀ3; L/Pd ¼ 2; T ¼ 80 ꢁC, [butadiene]initial ¼ 1.0 M; reaction time ¼ 2 h.
The formation of phosphonium cations by attack of free phos-
phine on the -allyl palladium complex to generate a triar-
in and persists throughout the remainder of the spectra taken at
longer reaction times. This behavior is qualitatively consistent with
the expected concentration profiles predicted by the mechanism
of Fig. 6.
p
ylphosphonium octadienyl fragment is analogous to formation of
1-methoxy-2,7-octadiene,1, by attack of methoxide (or methanol) on
the same complex, in the mechanism proposed by Beller [10]. This
phosphonium cation can react with methoxide to form a five-coor-
dinate phosphorane, which itself can then react with methoxide to
form phosphine oxide by rupture of one of the phosphorus-carbon
bonds of the original phosphonium cation. The PeC bond that is
ruptured is known to be that which gives the more stable anion [19].
In this case, the allyl anion is displaced preferentially to the phenyl
anion, so the phosphine oxide formed would be expected to be tri-
arylphosphine oxide.
These data clearly show that phosphonium cations derived from
addition of a triphenylphosphine to a C8 fragment are formed under
telomerization reaction conditions. This results in the depletion of
triphenylphosphine from the catalyst solution over time. This
ligand decomposition occurs concurrently with the loss of palla-
dium observable by ESI-MS throughout the reaction. It also is
correlated with the visible appearance of palladium black in
working catalyst solutions e the more basic ligands manifest the
formation of palladium black earlier than less basic ligands under
these conditions. We suggest that the reduction of phosphine
ligand concentration by this mechanism is a primary initiator of
palladium agglomeration, and ultimately, the complete loss of
catalyst activity in working telomerization catalyst solutions.
To distinguish between structures (B) and (D) of Fig. 6 as the
assignment for the 403.2 m/z peak, additional MS/MS experiments
were performed. The MS/MS spectrum of this species depicts
a fragmentation pattern that is very similar to that observed for
species (A) of Fig. 7, confirming that the two are closely related. This
would be expected for structure (B) which could form (A) simply by
dissociation of methoxide from phosphorus. We believe the frag-
mentation pattern of structure (D) would be significantly different
from (A) or (B), suggesting that (D) is not a viable structure for the
403.2 m/z species. However, in the absence of a reference standard
for structure (D), we have further attempted to distinguish between
these possible structures by examining the ESI-MS spectrum of this
species in the presence of CD3OD after it has formed from MeOH.
Our rationale is that if the structure is that of (B) in Fig. 6, then in
the presence of deuterium, the parent ion should also incorporate
a Dþ to make the species cationic and give at least some M þ 2
2.3. Electrospray Ionization-Mass Spectrometry
In order to determine why catalysts promoted by more basic
phosphine ligands, such as L8, showed poorer activity at elevated
initial methanol concentrations than those formed from less basic
phosphines, such as L4, Electrospray Ionization-Mass Spectrometry
(ESI-MS) analyses was carried out working telomerization catalyst
solutions. The primary goal of ESI-MS analyses was to observe the
evolution and course of triarylphosphine ligand degradation
throughout the telomerization reaction. The ESI-MS analyses were
carried out in the positive ion mode, enabling facile detection of
phosphonium cations. Neutral species, such as triarylphosphine
oxide, or the phosphorane species, would be expected to be
observed as a protonated (M þ 1) or sodiated cation (M þ 23).
Spectra of samples run within 15 min after removal from
working telomerization reactions using triphenylphosphine-
promoted palladium catalysts depict a peak in the ESI-MS spectrum
at 371.2 m/z. We assign this to the phosphonium cation (A) in Fig. 7.
This peak initially grows and then declines in intensity as the
reaction progresses. This is consistent with an intermediate species
that is both being formed and consumed throughout the telome-
rization reaction, as we have proposed. Additionally, a peak at
403.2 m/z, corresponding to the M þ 1 peak of phosphorane (B) of
Fig. 7, is also observed. This intermediate also initially grows and
then declines during the reaction. An alternate structure for this
species could be that formed by addition of methanol across one of
the double bonds to form another phosphonium salt (species D in
Fig. 6). We discuss later our rationale for assignment of the 403.2 m/
z peak as structure (B) and not as (D).Finally, a peak at 301.0 m/z,
corresponding to the sodium adduct of triphenylphosphine oxide
(C) of Fig. 7, although not present in the 15 or 30 min spectra, grows
Table 2
Performance of Selected Triarylphosphines under Relevant Conditions and
Concentrations.
Ligand
Chi (c
)
TON
Selectivity
to 1 (mol %)
L2
L3
L4
L6
L8
L9
5.6
10.2
12.9
16.8
2.7
720
84
88
89
80
86
6
5772
10834
6935
676
ꢀ8.1
5

J.R. Briggs et al. / Journal of Organometallic Chemistry 696 (2011) 1677e1686
1681
with CD3O- under these conditions so this species would be
expected to exchange a maximum of 2 deuterons. The observed
fragmentation pattern and the deuterium experiment rules out
species (D) as a possible structure for the 403.2 m/z cluster of peaks.
In the case of tris(2-methoxyphenyl)phosphine, the analogous
experiment showed incorporation of deuterium into the alpha
positions of the octadienyl group of the phosphonium cation, but
not into the phosphorane. While we do not have a completely
satisfactory explanation of this, we do know that the rate of proton
exchange into the alpha positions of phosphonium salts becomes
slower as the phosphine becomes more basic [22]. If capture of
methoxide by the phosphonium in this case is faster than H/D
exchange, and is irreversible, this is the result we would expect.
At low methanol concentrations, the performance of tris(2-
methoxyphenyl)phosphine, L8, is exceptional, giving selectivities of
around 95% with good rates. The catalyst activity decays very
rapidly, however, particularly under commercial conditions where
the methanol concentration is higher. We attribute these effects to
the decomposition reaction shown above to occur with triphenyl-
phosphine, and propose that more basic phosphines, such as tris
(2-methoxyphenyl)phosphine, would undergo this reaction more
rapidly than less basic phosphines. Further, since this decomposi-
tion reaction proceeds through a phosphonium cation, and the
conversion presumably involves a charge-separated transition
state, we suggest that this should be faster in a more polar solvent,
such as at high methanol concentration.
To examine these proposals, additional ESI-MS analyses was
carried out on tris(2-methoxyphenyl)phosphine-promoted palla-
dium-catalyzed telomerization reaction in comparison to the tri-
phenylphosphine data discussed above. In addition to examining
ligand decomposition chemistry, we also hoped to identify differ-
ences in the species present under working conditions that might
explain the higher linear regiochemistry observed with this ligand.
The ESI-MS spectrum of a solution of tris(2-methoxyphenyl)
phosphine (L8), acetic acid and [Pd(acac)2] in 12.7 M methanol
showed the formation of [Pd(acac)(P(C6H4-2-OMe)3)2]þ (909.1 m/z)
as the predominant product, along with lesser amounts of a mono-
phosphine complex, [Pd(acac)(P(C6H4-2-OMe)3)]þ (557.1 m/z). That
triphenylphosphine (L4) forms an analogous bisephosphine
complex, but not the corresponding mono-phosphine complex,
can be attributed to the larger steric requirements of the tris
(2-methoxyphenyl)phosphine (L8) compared to triphenylphosphine,
which destabilizes the bisephosphine complex relative to the mono-
phosphine complex. There is little change in the spectrum upon
addition of sodium methoxide, with only low concentrations of
additional palladium-containing species being formed. There are,
TON to MOD-1 Versus Ligand Chi
16000
14000
12000
10000
8000
6000
4000
2000
0
-10
-5
0
5
10
15
20
Ligand Chi
Fig. 5. TON versus Ligand Chi at [methanol] ¼ 12.7 M and [butadiene] ¼ 2.5 M.
peaks in the parent ion. The alternate structure (D), inherently
being a cation, would be expected to give an unchanged parent ion
in the presence of CD3OD.
The deuterium-labeled methanol experiment was carried out
with both triphenylphosphine and tris(2-methoxyphenyl)phos-
phine-ligated catalysts. Some differences in the data in the two
cases were noted. With the triphenylphosphine based catalysts, we
noted that the cation [Ph3PCH2CH ¼ CHCH2CH2CH2CH]CH2]þ
showed the presence of an Mþ cluster of peaks that indicated the
presence of one and two substitutions of deuterium for hydrogen in
the molecule. This was initially surprising, since, as it is already
a cation, no protonation was expected. Review of the literature,
however, indicated that exchange of the two octadienyl protons
alpha to the phosphorus atom is to be expected in the basic catalyst
solution which contains sodium methoxide [20].
The species formulated as [Ph3P(OMe)(CH2CH ¼ CHCH2CH2CH2-
CH]CH2)] (structure (B) in Fig. 7; also Supplementary Material
Figures I and II) showed an Mþ cluster of peaks that indicated it
had incorporated from 1 to 6 deuteriums. We conclude that these
arise from incorporation of one or two deuteriums alpha to the
phosphorus, since this species is derived from the phosphonium
cation (structure (A) of Fig. 7), with an additional three deuteriums
from exchange of CD3O- for CH3O- bound to phosphorus, and one
deuterium to render it cationic. The exchange of alkoxy groups in
phosphoranes is known to be facile [21]. This explanation accounts
for the observed experimental spectrum. It is not likely that the
methyl ether of proposed structure (D) of Fig. 6 would exchange
PAr3
Ar3P
(L2)Pd
Pd(L2)
OMe
Ar3P
Ar3P
OMe
OMe
Ar3P
O
OMe
+
+
MeOMe
PAr3
H
Fig. 6. Proposed mechanism for the formation of phosphonium cation, and subsequent solvolysis to phosphine oxide.

1682
J.R. Briggs et al. / Journal of Organometallic Chemistry 696 (2011) 1677e1686
Fig. 7. Possible decomposition products of triphenylphosphine in telomerization catalyst solutions.
however, significant changes in the spectrum of the reaction solution
taken15minafterthe additionofbutadiene. Althoughsome[Pd(acac)
(P(C6H4-2-OMe)3)2]þ remains, and the formation of some [Pd
(CH2CH]CHCH2CH2CH2CH]CH2)(P(C6H4-2-OMe)3)]þ is observed,
analogous to the complex seen in the triphenylphosphine-promoted
catalyst, the intensity of the peaks from these palladium-containing
complexes is very low, suggesting that most of the palladium has lost
its supporting ligands and is no longer observable in the ESI-MS
experiment. The predominant feature in this 15 min spectrum is an
ion corresponding to [(C6H4-2-OMe)3P(CH2CH ¼ CHCH2CH2CH2CH]
CH2)]þ (461.2 m/z), the phosphonium cation corresponding to struc-
ture (A) of Fig. 7. After 30 min and thereafter, essentially no palladium
complexes are seenin the spectra. The peaks correspondingto[(C6H4-
2-OMe)3PCH2CH ¼ CHCH2CH2CH2CH]CH2]þ remain, while a new
peak, corresponding to species (B) of Fig. 7, grows in.
at lower MeOH concentration since these catalysts retain activity
longer under such conditions.
These data clearly show that the more basic ligand undergoes the
same decomposition reaction as triphenylphosphine, but at a much
more rapid rate. This observation also suggests that the loss of ligand
in telomerization catalyst solutions correlates with loss of catalytic
activity, loss of observable palladium in the positive ion ESI-MS
spectrum, and the appearance of palladium black in the working
catalyst solutions. Fig. 8 shows the evolution of the ESI-MS spectra of
the triphenylphosphine and tris(2-methoxyphenyl)phosphine-
ligated catalysts with time. It clearly demonstrates that the more
basic ligand L8 decomposes more rapidly than triphenylphosphine,
which coincides with a loss of ESI-MS detectable palladium.
2.4. Design, synthesis and evaluation of new ligands
The data presented above is consistent with rapid formation of
a phosphonium salt of the tris(o-methoxyphenyl)phosphine very
early in the telomerization reaction, with the resulting loss of
palladium from solution. This is also consistent with the short
lifetime of this catalyst system under telomerization conditions in
concentrated methanol solution. We have not yet examined the
effect of methanol concentration on the rate of phosphine ligand
degradation. We predict that this rate would be significantly lower
The correlations of increased intrinsic activity and selectivity, and
of decreased stability, with increasing ligand basicity means that the
most basic ligands are unsuitable for commercial use. The remainder
of the work reported here focuses on efforts to capture the high
selectivity achieved by ligands possessing the o-methoxyphenyl
substitution motif, by combining it with increased ligand stability.
Our approach to this goal was to moderate the phosphine basicity,
403.2
493.2
100
%
100
4 hrs
4 hrs
a
b
417.3
461.2
494.2
495.3
%
0
477.1
220.2
371.2
575.4
221.2
499.3
308.3
0
30min
30min
477.1
476.1 479.1
461.2
100
%
100
%
461.7
461.8
493.2
481.1
475.1
418.4
403.2
371.2
220.2
482.1
494.2
575.4
236.1
0
0
15min.
15min.
477.1
461.2
100
%
100
%
481.1
482.1
475.1
462.2
463.2
403.2
400
567.1
600
259.3
0
0
m/z
600
m/z
200
300
500
200
300
400
500
Fig. 8. ESI-MS Spectra (200e600 amu) focusing on the degradation products from working (a) triphenylphosphine and (b) tris(2-methoxyphenyl) phosphine catalyst solutions as
a function of time.

J.R. Briggs et al. / Journal of Organometallic Chemistry 696 (2011) 1677e1686
1683
multiple recrystallizations required to remove residual color and
obtain highly pure product. The 31P and 13C NMR spectra of the
products L10eL15 are consistent with the proposed structures. The
31P NMR analysis of the phosphorus-based ligands showed that the
chemical shifts range from ꢀ26 to ꢀ21 ppm. Spectra of ligands L11,
L13, and L14 are complex due to the splitting of some 13C resonances
by phosphorus and fluorine.
L10 (6.1)
L11 (11.1)
L12 (7.4)
The
c parameters of each ligand included in this study are shown
in Fig. 9. These values were estimated using published substituent
electronic contributions for phosphine ligands. In the case of ligand
L10, the electronic contribution of each m-CF3 substituent is not
reported. In order to estimate the electronic effectof this substitution,
the contribution of a p-CF3 substituent and related substituents was
used to assign this Chi value [24]. By maintaining the bis(2-methox-
yphenyl)phosphine framework and varying the donoreacceptor
properties of each phosphine ligand by changing the substituents
only in the meta- and para-positions of the third aryl group, the steric
differences between these ligands was expected to be small so that
performance differences within this subset of phosphine ligands
(L9eL15) were essentially electronic in nature.
L13 (6.8)
L14 (8.6)
L15 (9.5)
Fig. 9. Bis-o-alkoxy substituted aryl phosphines evaluated in this study. The estimated
ligand Chi values are shown in parentheses.
e.g., by substituting an arene ring with electron-withdrawing
groups.We anticipated that this might slow the decomposition
reaction sufficiently to allow such ligands to survive long enough to
achieve commercial productivities, while maintaining improved
selectivities compared to triphenylphosphine. While o-methoxyaryl
phosphine ligands used as promoters for telomerization reactions of
biomass-derived alcohols and polyols has been reported earlier [23],
the incorporation of additional electron-withdrawing groups to
improve performance and catalyst stability has not been reported.
The synthetic approach taken described here focuses on moderating
the basicity of one aryl ring while maintaining the o-methoxy
substitution of the remaining two aryl rings (Fig. 3).Ligands which
were examined to test this hypothesis are shown in Fig. 9, along with
the estimated Tolman Chi values in parentheses.
2.6. Catalyst screening
The catalytic performance of these new ligands was monitored
with time as a function of initial MeOH concentration (12.7 and
5.1 M) at 90 ꢁC by sampling at 30 min, 1 h, 2 h, and 4 h. Details of
these experiments can be found in the experimental section.The
performance data of palladium catalysts promoted by these ligands
is collected in Table 3.
The resulting time series plots are shown in Fig. 9.Excellent
selectivities to MOD-1 are achieved by each of the new catalysts
compared to that of incumbent ligand L4, triphenylphosphine, at
each of the MeOH concentrations examined. At a methanol
concentration of 5.1 M, the selectivity to MOD-1 observed by the
incumbent triphenylphosphine system was 77.9% after 4 h. Under
the same conditions, the selectivities of the catalysts derived from
the bis(2-methoxyphenyl)phosphine based ligands ranged from
88.0 to 94.9%. At the higher methanol concentration of 12.7 M, the
selectivity improvements are less significant, yet the best bis
(2-methoxyphenyl)phosphine ligands, L8 and L9, continue to
demonstrate improved MOD-1 selectivities of 95.1 and 94.8%
respectively, compared to 89.6% for L4.
2.5. Ligand synthesis and characterization
The synthesis of ligands L11, L12, and L13, was accomplished by
reacting the commercially available starting material chlorobis(2-
methoxyphenyl)phosphine with the appropriate Grignard reagent.
The synthesis of ligand L14 was accomplished by performing
a lithium-halogen exchange byreacting n-BuLi with the precursor 1-
bromo-4-(trifluoromethyl)benzene. The in-situ generated 1-lithio-
4-(trifluoromethyl)benzene was then reacted with chlorobis(2-
methoxyphenyl)phosphine to produce the desired product. While
the crude yields were generally high, some loss occurred during the
The most striking features of the graphs shown in Fig. 10 are the
early flattening of some of the TON versus time curves, which we
equate with catalyst death. The plots of reactions carried out in
5.1 M methanol concentration clearly show that L11, the least basic
ligand, becomes inactive within the first 30 min of the reaction, and
TON Versus Time (12.7 M Methanol)
TON Versus Time (5.1 M Methanol)
14000
14000
12000
12000
10000
10000
L10, Chi = 6.1, Sel = 94 %
L10, Chi = 6.1, Sel = 89 %
L13, Chi = 6.8, Sel = 93 %
8000
L13, Chi = 6.8, Sel = 91 %
8000
L12, Chi = 7.4, Sel = 92 %
L12, Chi = 7.4, Sel = 92 %
L14, Chi = 8.4, Sel = 94 %
L15, Chi = 9.1, Sel = 93 %
L11, Chi = 11.1, Sel = 87 %
6000
4000
2000
0
L14, Chi = 8.4, Sel = 95 %
6000
L15, Chi = 9.1, Sel = 95 %
4000
2000
0
L11, Chi = 11.1, Sel = 92 %
0
1
2
3
4
5
0
1
2
3
4
5
Time (hrs)
Time (hrs)
Fig. 10. Performance of the new ligand-promoted palladium catalysts with time at 12.7 M methanol, 5.1 M methanol. [Pd] ¼ 8.8 ꢂ 10ꢀ5 M (11 ppmw), [NaOMe] ¼ 3.2 ꢂ 10ꢀ4
L/Pd ¼ 2, T ¼ 90 ꢁC, [butadiene]initial ¼ 2.5 M, reaction time ¼ 2 h. Maximum TON ¼ 16,600.
,

1684
J.R. Briggs et al. / Journal of Organometallic Chemistry 696 (2011) 1677e1686
Table 3
TON (mol 1/mol Pd) of Selected Bis(2-methoxyphenyl)arylphosphine-Promoted Palladium Catalysts in Butadiene Telomerization with Time and Initial Methanol
Concentration.
Ligand (Tolman Chi Value)
Time (hr)
L10 (6.1)
L13 (6.8)
L12 (7.4)
L14 (8.4)
L15 (9.1)
L11 (11.1)
Methanol ¼ 12.7 M
Methanol ¼ 5.1 M
0
0.5
1
2
4
0
0
0
0
7094
10169
11779
13022 (95)
0
0
1831
1940
1924
1949 (89)
5478
5682
5924
5922 (91)
6440
7283
7654
7663 (92)
5653
7814
9273
9967 (95)
5174
5470
5490
5487 (92)
0
0.5
1
2
4
0
0
9811
10295
11278
11574 (93)
0
8588
9845
10886
11305 (92)
0
8731
10158
11356
12079 (94)
0
0
5808
6659
6881
6974 (94)
4834
7086
8926
10405 (93)
2062
2200
2212
2206 (87)
Chi values are in parentheses following the ligand label.
Values in parentheses at 4 h are the final selectivity to 1 in mol %.
L10, the most basic ligand, after about 1 h. Ligands L13, L12 and L14
have very similar profiles, and appear to achieve maximum TON
productivity between one and 2 h L15 remains active for the entire
4 h reaction time. In 12.7 M methanol, catalysts promoted by L10,
L13, and L11 appear to lose all activity in the first 30 min, L12
somewhat later in the first hour, while L14 and L15 retain activity
into the second or third hour of the reactions. Although we do not
have a complete understanding of the factors determining the form
of these TON versus time plots, we suggest that the anaerobic
decomposition reaction described above, coupled with the differ-
ential binding of the ligands to the palladium catalyst throughout
the catalytic telomerization cycle, likely plays an important role.
Fig. 11 shows the activity of the ligand-stabilized catalysts,
of the resulting catalyst systems appears to be ligand instability, as
a steady decline in the TON performance over the 4 h period is
observed even though initial rates may be high. This is also supported
by the visual observation of earlier and increasing amounts of palla-
dium black precipitation during the course of the runs as the ligand
basicity is increased. As expected (Fig. 6), these more basic ligands
(L10, L13, and L12) perform better at lower methanol concentrations,
but experience a decline in performance as the methanol concen-
trations reach 12.7 M. The selectivities to 1 of the highest productivity
ligand-promoted catalysts, which use L14 and L15, are also very high
under both methanol concentrations. Consequently, catalysts based
on this ligand class, particularly L14, are considered viable candidates
for commercial use under a wider variety of process conditions.
expressed as TONs achieved after
4 h reactions at initial
[MeOH] ¼ 12.7 and 5.1 M and initial [butadiene] ¼ 2.5 M as a function
of ligand basicity. As can be seen, the least basic ligand L11 with a Chi
value of 11.1, gives only moderate productivity to 1 over the 4 h
reaction period. As the ligand basicity increases, first to a Chi value of
8.6 for L14, the TON performance of each catalyst system steadily
improves. L14 appears to perform close to the optimum for this class
of bis(2-methoxyphenyl)phosphine ligands where the effects of
increased activity and decreased stability with ligand basicity
compete to give a performance optimum in the range of Chi z 8e9.
Astheligandbasicityisfurtherincreased progressing through ligands
L10, L13, and L12, the primary factor governing the TON performance
3. Conclusions
We have shown that catalyst systems containing basic triaryl-
based phosphines, especially ortho-methoxy-substituted triar-
ylphosphines at low methanol concentration, give enhanced catalyst
performance in the process catalytically telomerizing two equiva-
lents of 1,3-butadiene to 1-methoxy-2,7-octadiene (MOD-1) in
methanol.Improvements are seen in both selectivity and activity.At
higher methanol concentrations, however, competing effects of
increased activity/selectivity and decreased stability with increasing
ligand basicity generate a performance optimum at moderate ligand
basicity under these more commercially-relevant conditions.Loss of
ligand from telomerization solutions was shown to be a result of
alkylation by an octadienyl group, forming phosphonium salts,
phosphoranes, and in one case, phosphine oxide, which were
observed by ESI-MS. Concurrent loss of ESI-MS-observable palla-
dium was also demonstrated. These observations lead us to the
suggestion that phosphine decomposition is the triggering event
that causes catalyst deactivation in phosphine-promoted telomeri-
zation catalysts. Improved performance was achieved by moderating
the phosphine ligand basicity, e.g., by substituting one arene ring
with electron-withdrawing groups, while maintaining the ortho-
methoxyphenyl structural motif of the remaining two aryl rings.
TON to 1 Versus Ligand Chi
14000
12000
10000
8000
6000
4000
Methanol = 12.7 M
2000
0
Methanol = 5.1 M
4. Experimental
4.1. Preparation of solvents and reagents
5
6
7
8 9
Ligand Chi
10
11
12
All solvents and reagents were handled and purified (if appli-
cable) in a glovebox under nitrogen. Anhydrous methylcyclohexane
(MCH) and methanol (CH3OH) purchased from Aldrich were purified
further by flowing each solvent through a short column of activated
Fig. 11. Plot of Ligand Chi versus TON (at 4 h reaction time) of the new ligand set
demonstrating the effects of increased activity and decreased stability with ligand
basicity competing to give a ligand optimum. Maximum TON ¼ 16,600 possible, 90 ꢁC,
[MeOH] ¼ 12.7 M or 5.1 M, L/Pd ¼ 2.

J.R. Briggs et al. / Journal of Organometallic Chemistry 696 (2011) 1677e1686
1685
alumina and a layer of silica gel to keep the alumina fines from
coming through the filter. Dibutyl ether, which is used as an internal
standard for gas chromatography analysis, was purchased from
Aldrich, and prepared by first stirring it over sodium (Na)/potassium
(K) alloy overnight and then flowing it through a short column of
activated alumina and a layer of silica gel to keep the alumina fines
from coming through the filter. Chlorobis(2-methoxyphenyl)phos-
phine was purchased from Alpha Aesar and used as is (3,5-Bis(tri-
fluoromethyl)phenyl) magnesium bromide, (4-chlorophenyl)
magnesium bromide, (4-fluorophenyl)magnesium bromide, (3,5-
difluorophenyl)magnesium bromide, L10, and 1-bromo-4-(tri-
fluoromethyl)benzene, as well as all other reagents, were used as
purchased from Aldrich unless otherwise noted.
d
ꢀ21.1. 19F NMR (C6D6, externally referenced using CCl3F): ꢀ63.0.
Anal. Calcd for C22H17F6O2P: C, 57.65; H, 3.74, found: C, 57.57; H, 3.83.
ESI-TOF-MS (Positive Mode) calcd for (M þ H)þ ion of C22H17F6O2P:
459.095 amu, found 459.095 amu.
4.1.3. Preparation of (4-Chlorophenyl)bis(2-methoxyphenyl)
phosphine, L12
(4-Chlorophenyl)magnesium bromide (2.51 mmol, 2.51 mL of
a 1.0 M solution in diethylether) was added dropwise to a solution
of chlorobis(2-methoxyphenyl)phosphine (0.640 g, 2.28 mmol) in
THF (50 mL) at 0 ꢁC. This mixture was then heated to reflux over-
night. After the reaction period the volatiles were removed and the
mixture was extracted and filtered using toluene. Removal of the
toluene resulted in the isolation of a solid which was then dissolved
in warm acetonitrile (20 mL), filtered, and placed in the freezer
(ꢀ10 ꢁC) overnight during which time a white crystalline solid
formed. This material was isolated by decanting the liquid away
and then dried under vacuum resulting in the isolation of the
4.1.1. Catalytic telomerization screening
Precatalyst stock solutions were prepared by dissolving Pd
(acac)2 (0.0294 g, 0.0966 mmol), phosphine (0.1932 mmol), and
acetic acid (0.0966 mmol 0.50 mL of 0.193 M HOAc in CH3OH) in
MeOH to give a total volume of 50.00 mL. The two equivalents of
acetic acid are used as a stabilizing additive in the plant catalyst
preparation maintained in this study to best mimic commercial
catalyst handling. Into an open Fischer-Porter bottle were syringed
dibutyl ether, MeOH, methylcyclohexane, the precatalyst stock
solution (1.00 mL), and the NaOMe solution. In a typical experi-
ment four reactors are run in parallel with only the methanol
concentrations varying (5.1 and 12.7 M) and the initial volume
being 25.1 mL. The Fischer-Porter vessel was then attached to
a manifold equipped with a pressure gauge and a sampling valve.
Approximately 3.5 g (5.5 mL) of 1,3-butadiene was measured into
a gas-tight syringe. The butadiene had been previously distilled
from the container in which it was received into a Fischer-Porter
bottle to remove inhibitor. The 1,3-butadiene was then injected
into the reactor via the septum with the needle placed just
below the surface of the liquid to facilitate dissolution. The initial
concentrations of the components are dibutyl ether ¼ 1.17 M,
methylcyclohexane ¼ 9.4 M or 0.50 M depending upon 5.1 M or
12.7 M MeOH conditions respectively, [Pd] ¼ 7.7 ꢂ 10ꢀ5 M,
NaOMe ¼ 3.8 ꢂ 10ꢀ4 M, and 1,3-butadiene ¼ 2.54 M. The reactor
was then placed into the preheated oil bath with stirring and
allowed to react for a period of 4 h.The reactor was sampled peri-
odically using a gas-tight syringe for gas chromatographic analysis.
Gas chromatographic analysis was performed on an HP-6890 gas
chromatograph using a DB-1701 column at constant gas flow.
Dibutyl ether was employed as internal standard. Response factors
were determined using samples of known composition.
desired product (0.626 g, 77.0%). 1H NMR (C6D6):
d 3.17 (s, 6 H), 6.49
(dd, 3JHH ¼ 8.2 and 4.5 Hz, 2 H), 6.75 (t, 3JHH ¼ 7.4 Hz, 2 H), 6.93e7.29
(m, 8 H). 13C NMR (C6D6): 55.3, 110.6, 121.4, 125.8 (d, 15.0 Hz), 130.4,
134.1, 134.9, 135.9 (d, 21.8 Hz), 136.4 (d, 14.2 Hz), 161.8 (d, 16.5 Hz).
31P NMR (C6D6, externally referenced using H3PO4):
d
ꢀ26.1. Anal.
Calcd for C20H18ClO2P: C, 67.33; H, 5.09, found: C, 66.35; H, 5.18. ESI-
TOF-MS (Positive Mode) calcd for (M þ H)þ ion of C20H18ClO2P:
357.081 amu, found 357.081 amu.
4.1.4. Preparation of (4-Fluorophenyl)bis(2-methoxyphenyl)
phosphine, L13
(4-Fluorophenyl)magnesium bromide (3.14 mmol, 1.57 mL of
a 2.0 M solution in diethylether) was added dropwise to a solution of
chlorobis(2-methoxyphenyl)phosphine (0.880 g, 2.85 mmol) in THF
(50 mL) at 0 ꢁC. This mixture was then heated to reflux overnight.
After the reaction period the volatiles were removed and the residue
was extracted and filtered using toluene. Removal of the toluene
resulted in the isolation of a yellow/brown solid which was extracted
using a minimum amount of acetonitrile, filtered and placed in the
freezer (ꢀ10 ꢁC) overnight during which time the desired product
crystallized asawhite solid which wasisolatedby decanting away the
solution and drying under vacuum (0.560 g, 54.5%). 1H NMR (C6D6):
d
3.18 (s, 6 H), 6.50 (dd, 3JHH ¼ 9.0 and 6.0 Hz, 2 H), 6.68e6.81 (m, 4 H),
6.93e6.99 (m, 2 H), 7.08e7.15 (m, 2 H), 7.26e7.35 (m, 2 H). 13C NMR
(C6D6): 55.2,110.5,115.5 (d, 7.6Hz),115.8 (d, 7.7Hz),121.3,130.3,134.0,
136.4 (d, 7.6 Hz), 136.7 (d, 8.4 Hz), 161.7 (d, 16.0 Hz), 165.3. 31P NMR
(C6D6, externally referenced using H3PO4):
d
ꢀ26.0. 19F NMR (C6D6,
externally referenced using CCl3F): ꢀ113.5. Anal. Calcd for
C20H18FO2P: C, 70.58; H, 5.33, found: C, 69.55; H, 5.51. ESI-TOF-MS
(Positive Mode) calcd for (M þ H)þ ion of C20H18FO2P: 341.111 amu,
found 341.111 amu.
4.1.2. Preparation of (3,5-Bis(trifluoromethyl)phenyl)bis
(2-methoxyphenyl)phosphine, L11
(3,5-Bis(trifluoromethyl)phenyl)magnesium bromide (2.02 mm
ol, 4.01 mL of a 0.5 M solution in THF) was added dropwise to
a
solution of chlorobis(2-methoxyphenyl)phosphine (0.515 g,
4.1.5. Preparation of Bis(2-methoxyphenyl)(4-(trifluoromethyl)
phenyl)phosphine, L14
1.84 mmol) in THF (50 mL) at 0 ꢁC. This mixture was then heated to
refluxovernight. Afterthereactionperiodthevolatileswereremoved
and the mixture was extracted and filtered using toluene. Removal of
the toluene resulted in the isolation of a yellow/brown solid which
was extracted using a minimum amount of acetonitrile, filtered, and
placed in the freezer (ꢀ10 ꢁC) over the weekend during which time
the desired product crystallized as a while solid which was isolated
by decanting away the solution and drying under vacuum (0.502 g,
1-Bromo-4-(trifluoromethyl)benzene (0.593 g, 2.63 mmol) was
stirred in diethylether (30 mL) at 0 ꢁC as n-BuLi (2.63 mmol, 1.32 mL
of a 2.0 M solution in cyclohexane) was added dropwise. This
mixture was allowed to stir for an additional hour. Chlorobis(2-
methoxyphenyl)phosphine (0.739 g, 2.63 mmol) was then added as
a solid and the mixture allowed to stir for an additional 4 h at room
temperature. After the reaction period the volatiles were removed
and the residue was dissolved in benzene and filtered to remove
the salts. Removal of the volatiles from the filtrate resulted in the
isolation of a yellow residue which was dissolved in a minimum
amount of acetonitrile, filtered, and placed in the glovebox freezer
(ꢀ10 ꢁC) overnight during which time crystals precipitated which
were isolated by decanting off the liquid and drying under vacuum.
59.7%). 1H NMR (C6D6):
d
3.08 (s, 6 H), 6.41 (dd, 3JHH ¼ 7.8 and 4.5 Hz,
2 H), 6.65 (t, 3JHH ¼ 7.5 Hz, 2 H), 6.91 (dt, 3JHH ¼ 6.4 and 1.5 Hz, 2 H),
7.03 (dt, 3JHH ¼ 7.7 and 1.2 Hz, 2 H), 7.67 (s,1 H), 7.88 (d, 3JHH ¼ 5.7 Hz,
2 H).13C NMR (C6D6): 55.2,110.8,121.6,123.6 (d,14.0 Hz),125.8,131.3,
131.7 (d, 6.1 Hz), 133.8 (d, 20.6 Hz), 134.4 (d, 6 Hz), 142.9 (d, 21.2 Hz),
161.7 (d, 14.2 Hz). 31P NMR (C6D6, externally referencedusingH3PO4):

1686
J.R. Briggs et al. / Journal of Organometallic Chemistry 696 (2011) 1677e1686
This recrystallization process was performed three times resulting
in the isolation of the desired product as a white microcrystalline
Conditions: [Pd] ¼ 7.7 ꢂ 10ꢀ5 M (10.9 ppm), L/Pd ¼ 2 mol/mol,
[butadiene]initial ¼ 2.5 M, [MeOH]initial ¼ 12.7 M, T ¼ 90 ꢁC, reaction
time ¼ 2 h. Methylcyclohexane was added to bring the total reac-
tion volume to the same value. Reactions were conducted in sealed
Fischer-Porter tubes immersed in a constant temperature bath.
Experimental details are reported in the experimental section.
Maximum TON for 100% yield of 1 is w16,600.
solid (0.419 g, 49.3%). 1H NMR (C6D6):
d 3.16 (s, 6 H), 6.47e6.52 (m,
2 H), 6.71e6.77 (m, 2 H), 6.90e6.95 (m, 2 H), 7.08e7.15 (m, 2 H),
7.21e7.25 (m, 2 H), 7.31e7.37 (m, 2 H). 13C NMR (C6D6): 55.2, 110.6,
121.4, 125.1 (m), 130.7, 134.23, 134.25, 134.4, 134.7, 143.4, 161.8 (d,
15.7 Hz). 31P NMR (C6D6, externally referenced using H3PO4):
d
ꢀ24.6. 19F NMR (C6D6, externally referenced using CCl3F): ꢀ62.7.
Anal. Calcd for C21H18F3O2P: C, 64.62; H, 4.65, found: C, 64.85; H,
4.68. ESI-TOF-MS (Positive Mode) calcd for (M þ H)þ ion of
C21H18F3O2P: 391.107 amu, found 391.108 amu.
Appendix. Supplementary data
Supplementary data associated with the article can be found in
online version, at doi:10.1016/j.jorganchem.2011.02.007.
4.1.6. Preparation of Bis(2-methoxyphenyl)(3,5-difluorophenyl)
phosphine, L15
Chlorobis(2-methoxyphenyl)phosphine (0.666 g, 2.37 mmol)
was stirred in diethylether (30 mL) at 0 ꢁC as (3,5-difluorophenyl)
magnesium bromide (2.37 mmol, 4.74 mL of a 0.5 M solution in
diethylether) was added dropwise. This mixture was allowed to stir
overnight at room temperature. After the reaction period the
volatiles were removed and the residue was dissolved in benzene
and filtered to remove the salts. Removal of the volatiles from the
filtrate resulted in the isolation of a dark yellow residue which was
dissolved in a minimum amount of acetonitrile, filtered, and placed
in the glovebox freezer (ꢀ10 ꢁC) overnight during which time
crystals precipitated which were isolated by decanting off the
liquid and drying under vacuum. This recrystallization process was
repeated until the product was isolated as a white microcrystalline
References
[1] J.M. Takacs, in: E.W. Abel, G. Stone, G. Wilkinson, Series Eds, L.S. Hegadus, Vol.
Ed (Eds.), Comprehensive Organometallic Chemistry II, vol. 12, Chapt. 8.1,
Elsevier Science Ltd, Oxford, New York, Tokyo, 1995, p. 785.
[2] (a) E.J. Smutny, J. Am, Chem. Soc. 89 (1967) 6793;
(b) S. Takahashi, T. Shibano, N. Hagihara, Tetrahedron Lett. 8 (1967) 2451.
[3] A. Behr, M. Becker, T. Beckmann, L. Johnen, J. Leschinski, S. Reyer, Angew.
Chem. Int. Ed. 48 (2009) 3598.
[4] (a) B.J. Schaart, H.L. Pelt, G.B. Jacobsen, Continuous Process for the Telomeri-
zation of Conjugated Dienes U. S. Patent 5,254,782 (October 19, 1973);
(b) B. Estrine, S. Bouquillon, F. Hénin, J. Muzart, Eur. J. Org. Chem. (2004)
2914e2922;
(c) M. Basato, L. Crociani, F. Benvenuti, A.M.R. Galletti, G. Sbrana, J. Mol, Cat. A:
Chem. 145 (1999) 313;
(d) C.L. Edwards, Process for Producing Detergent Molecules Comprising an
Ether Linkage from Butadiene U. S. Patent 7,141,539 (November 28, 2006).
[5] E. Singer, R. Landgraf, D.J. Slamon, D. Eisenberg, Process for Production of
Ethers U. S. Patent 7,314,916 (January 1, 2008).
[6] R. Jackstell, S. Harkal, H. Jiao, A. Spannenberg, C. Borgmann, D. Röttger,
F. Nierlich, M. Elliot, S. Niven, K. Cavell, O. Navarro, M.S. Viciu, S.P. Nolan,
M. Beller, Chem. Eur. J. 10 (2004) 3891.
[7] F. Benvenuti, C. Carlini, M. Lami, M. Marchionna, R. Patrini, A.M.R. Galletti,
G. Sbrana, J. Mol, Cat. A: Chem. 144 (1999) 27.
[8] W. Keim, in: G.N. Schrauzer (Ed.), Transition Metals in Homogeneous Catal-
ysis, Marcel Dekker, New York, 1971.
[9] (a) P.W. Jolly, R. Mynott, B. Raspel, K.-P. Schick, Organometallics 5 (1986) 473;
(b) R. Benn, P.W. Jolly, R. Mynott, B. Raspel, G. Schenker, K.-P. Schick,
G. Schroth, Organometallics 4 (1985) 1945.
[10] F. Vollmüller, J. Krause, S. Klein, W. Mägerlein, M. Beller, Eur. J. Inorg. Chem.
(2000) 1825.
[11] M.J.-L. Tschan, E.J. Garcia-Suarez, Z. Freixa, H. Launay, H. Hagen, J. Benet-
Buchholz, P.W.N.M. van Leeuwen, J. Am. Chem. Soc. 132 (2010) 6463.
[12] D. Beigzadeh, Macro. React. Eng. 1 (2007) 331.
solid (0.350 g, 41.2%). 1H NMR (C6D6):
d 3.14 (s, 6 H), 6.39e6.49 (m,
2 H), 6.67e6.75 (m, 2 H), 6.90e7.17 (m, 7 H). 13C NMR (C6D6): 55.3,
104.0 (t, 25 Hz),110.7,116.5 (m),121.5,124.7 (d,14.2 Hz),130.8,134.1
(d, 2.2 Hz), 140.7, 161.8 (d, 16.0 Hz), 165.0 (m). 31P NMR (C6D6,
externally referenced using H3PO4):
d
ꢀ22.7. 19F NMR (C6D6,
externally referenced using CCl3F): ꢀ110.3 (m). Anal. Calcd for
C20H17F2O2P: C, 67.04; H, 4.78, found: C, 66.61; H, 4.72. ESI-TOF-MS
(Positive Mode) calcd for (M
359.101 amu, found 359.102 amu.
þ
H)þ ion of C20H17F2O2P:
4.1.7. ESI-MS analyses
All samples were transferred in Hamilton Gas-Tight Syringes to
a drybox adjacent to the mass spectrometer. The syringes were
subsequently removed from the drybox and infused directly into
a syringe needle adapter union on the Z-spray Electrospray probe of
a Waters-Q-ToF II Mass Spectrometer. This allowed the sample to be
manually infused from the syringe to the ESI probe without any
intervening tubing. In order to avoid carry-over between different
time-point injections, the probe was cleaned by infusing dry
methanol after each analysis. The Time of Flight (ToF) analyzer was
[13] (a) C.A. Tolman, Chem. Rev. 77 (1977) 313;
(b) F. Benvenuti, C. Carlini, M. Lami, M. Marchionna, R. Patrini, A.M.R. Galletti,
G. Sbrana, J. Mol, Cat. A: Chem. 144 (1999) 27.
[14] S.J. Chen, K.R. Dunbar, Inorg. Chem. 29 (1990) 588.
[15] US Patent 7,041,619, May 9, 2006, assigned to Kuraray.
[16] S.M. Maddock, M.G. Finn, Organometallics 19 (2000) 2684.
[17] K.L. Granberg, J.-E. Backwall, J. Am. Chem. Soc. 114 (1992) 6858.
[18] M. Grayson, P.T. Keough, J. Am, Chem. Soc. 82 (1960) 3919.
[19] J.R. Corfield, S. Trippett, J. Chem. Soc. D: Chem. Comm. 19 (1970) 1267.
[20] (a) S.E. Cremer, R.J. Chorvat, Tetetrahedron Lett. 4 (1966) 419;
(b) M.-C. Tseng, H.-C. Kan, Y.-H. Chu, Tetrahedron Lett. 48 (2007) 9085.
[21] B.C. Chang, W.E. Conrad, D.B. Denney, D.Z. Denny, R. Edelmaq, R.L. Powell,
D.W. White, J. Am. Chem. Soc. 93 (1971) 4004.
calibrated prior to analysis using a calibrant solution (2 l NaI/
mg/
m
0.05 g/ l CsI in 1/1 isopropanol/water) at a flow rate of 10 mL/min.
m
m
Data acquisition was performed with a cycle time of 1 scan/sec
(scan acquisition time: 0.88 s; interscan delay: 0.1 s) in the MS
[22] (a) S.E. Cremer, R.J. Chorvat, Tet. Lett. 4 (1966) 419e421;
(b) C. M- Tseng, H.-C. Kan, Y.-H. Chu, Tet, Lett. 48 (2007) 9085e9089.
[23] (a) PJ.C. Hausoul, P.C.A. Bruijnincx, R.J.M. Klein Gebbink, B.M. Weckhuysen,
Chem. Sus. Chem. 2 (2009) 855;
mode. The infusion rate was approximately 20e50
ml/min. The
following mass spectral conditions were utilized: MCP Detector:
2150 V; Resolution: 9000; Data Acquisition Mode: Continuum
Positive Ion Mode; Mass Scan: 50e2100 amu; Capillary Voltage:
2500 V; Cone Voltage: 20 V; Source Temperature: 110 ꢁC; Des-
olvation Temperature: 300 ꢁC; Desolvation Gas: Nitrogen at 60 psig.
In MS/MS mode, fragmentation of each component was carried out
separately with collision energies set to 25 V, 30 V, 35 V and 40 V.
The instrument was scanned at 0.5 scan per seconds and the mass
range was set to 100e1900 m/z.
(b) R. Palkovits, I. Nieddu, C.A. Kruithof, R.J.M. Klein Gebbink, B.M. Weckhuysen,
Chem. Eur. J. 14 (2008) 8995;
(c) R. Palkovits, I. Nieddu, R.J.M. Klein Gebbink, B.M. Weckhuysen, Chem. Sus.
Chem. 1 (2008) 193;
(d) R. Palkovits, A.N. Parvulescu, P.J.C. Hausoul, C.A. Kruithof, R.J.M. Klein Geb-
bink, B.M. Weckhuysen, Green Chem. 11 (2009) 1155.
[24] (a) M.R. Wilson, H. Liu, A. Prock, W.P. Giering, Organometallics 12 (1993) 2044;
(b) A. Fernandez, C. Reyes, M.R. Wilson, D.C. Woska, A. Prock, W.P. Giering,
Organometallics 16 (1997) 342.