Studies on palladium-bisphosphine catalyzed alternating copolymerization of CO and ethylene

Article abstract of DOI:10.1016/S1381-1169(99)00258-7

Palladium(II)-bisphosphine catalyzed copolymerization of CO and ethylene was studied in detail. The results showed that DPPPr/Pd(II) mole ratio, CF₃COOH/Pd(II) mole ratio and solvent greatly influence the catalytic activity. Solvent effects were studied and in detail through end group analysis. With high pressure in situ ¹H-NMR technique, the signals of coordinated ethylene at 4.4 ppm and methylene linked with Pd(II) (Pd - CH₂-) at 0.6 ppm were observed. With high-pressure in situ IR technique three palladium carbonyl absorptions were observed generated at 1638 cm^-1, 1616 cm^-1 and 1970 cm^-1, which may be assigned to three intermediates (3) (5) (6). With the above results, the copolymerization mechanisms were discussed. High-pressure in situ ³¹P-NMR experimental result showed that only mono- chelate ring complex (1) was produced when DPPPr/Pd(II) = 1; only bis-chelate ring complex (2) was produced when DPPPr/Pd(II) = 2; equimolar complex (1) and (2) were produced when DPPPr/Pd(II) = 1.5. Extended X-ray absorption fine structure (EXAFS) was used to calculate the coordination number (CN) and shell radius (R, in ?) of complexes (DPPPr)Pd(OCOCF₃)₂ (a), (DPPBu)Pd(OCOCF₃)₂ (b), (DPPEt)Pd(OCOCF₃)₂ (c), (DPPPr)₂Pd(OCOCF₃)₂ (e), and two catalyst solutions of methanol freshly taken before and in the middle of copolymerization (named S1 and S2, respectively). The Pd - P bond length of complexes (a) (b) (c) are 2.25(3) ?, 2.33(3) ?, and 2.38(2) ?, respectively, Pd - O bond length of complexes (a) (b) (c) are 2.07(3) ?, 2.06(2) ? and 2.05(3) ?, respectively. In the order of complexes (a) (b) (c), the catalytic activity increases with Pd - P bond decreasing and Pd - O bond increasing, which may show that the chelate ring is more stable, the anions can go away more easily, and the corresponding catalyst is more efficient. The coordination number of two catalyst solutions S1 and S2 are increased by 1.9 and 0.9 compared with solid complex (a), which may show that 18-electron and five-coordinated pyramidal intermediates were produced and existed in the system besides 16-electron and four-coordinated square planar intermediates. (C) 2000 Elsevier Science B.V.

Full text of DOI:10.1016/S1381-1169(99)00258-7

Ž
.
Journal of Molecular Catalysis A: Chemical 151 2000 91–113
www.elsevier.comrlocatermolcata
Studies on palladium-bisphosphine catalyzed alternating
copolymerization of CO and ethylene
b
c
c
He-Kuan Luo a,), Yuan Kou , Xi-Wen Wang , Da-Gang Li
a
c
State Key Laboratory for Oxo Synthesis and SelectiÕe Oxidation, Lanzhou Institute of Chemical Physics, Chinese Academy of Sciences,
Lanzhou 730000, China
College of Chemistry and Molecular Engineering, Beijing UniÕersity, Beijing 100871, China
b
State Key Laboratory for Oxo Synthesis and SelectiÕe Oxidation, Lanzhou Institute of Chemical Physics, Chinese Academy of Sciences,
Lanzhou 730000, China
Received 15 October 1998; accepted 3 June 1999
Abstract
Ž .
Palladium II -bisphosphine catalyzed copolymerization of CO and ethylene was studied in detail. The results showed that
Ž .
Ž .
DPPPrrPd II mole ratio, CF3COOHrPd II mole ratio and solvent greatly influence the catalytic activity. Solvent effects
were studied and in detail through end group analysis. With high pressure in situ 1H-NMR technique, the signals of
Ž . Ž
.
coordinated ethylene at 4.4 ppm and methylene linked with Pd II Pd—CH2- at 0.6 ppm were observed. With
1
high-pressure in situ IR technique three palladium carbonyl absorptions were observed generated at 1638 cmy1, 1616 cmy
1
and 1970 cmy , which may be assigned to three intermediates 3
5
6 . With the above results, the copolymerization
Ž . Ž . Ž .
mechanisms were discussed. High-pressure in situ 31P-NMR experimental result showed that only mono-chelate ring
Ž .
Ž .
Ž .
Ž .
complex 1 was produced when DPPPrrPd II s1; only bis-chelate ring complex 2 was produced when DPPPrrPd II s2;
Ž . Ž . Ž .
equimolar complex
1
and
.
EXAFS was used to calculate the coordination number CN and shell radius R, in A of complexes DPPPr Pd OCOCF3
2
2 were produced when DPPPrrPd II s1.5. Extended X-ray absorption fine structure
˚
Ž
.
Ž
.
Ž
.
Ž
.
Ž
Ž .
Ž
.
Ž
.
Ž .
Ž
.
Ž
.
Ž . . Ž .
c , DPPPr 2Pd OCOCF3 e , and two catalyst solutions of
2
Ž
.
Ž
a , DPPBu Pd OCOCF3
b , DPPEt Pd OCOCF3
2
2
Ž
.
methanol freshly taken before and in the middle of copolymerization named S1 and S2, respectively . The Pd—P bond
length of complexes a c are 2.25 3 A, 2.33 3 A, and 2.38 2 A, respectively, Pd—O bond length of complexes a
Ž . Ž .
b
˚
˚
˚
Ž . Ž . Ž .
Ž .
Ž .
Ž .
Ž .
Ž . Ž . Ž .
b c , the catalytic activity
b
˚
˚
˚
Ž .
Ž .
Ž .
c are 2.07 3 A, 2.06 2 A and 2.05 3 A, respectively. In the order of complexes a
increases with Pd—P bond decreasing and Pd—O bond increasing, which may show that the chelate ring is more stable, the
anions can go away more easily, and the corresponding catalyst is more efficient. The coordination number of two catalyst
Ž .
solutions S1 and S2 are increased by 1.9 and 0.9 compared with solid complex a , which may show that 18-electron and
five-coordinated pyramidal intermediates were produced and existed in the system besides 16-electron and four-coordinated
square planar intermediates. q 2000 Elsevier Science B.V. All rights reserved.
Keywords: Palladium-bisphosphine; CO; Ethylene
)
Corresponding author. Beijing Research Institute of Chemical Industry, Beijing 100013, China.
1381-1169r00r$ - see front matter q 2000 Elsevier Science B.V. All rights reserved.
Ž
.
PII: S1381-1169 99 00258-7

(
)
92
H.-K. Luo et al.rJournal of Molecular Catalysis A: Chemical 151 2000 91–113
1. Introduction
The copolymerization of CO with ethylene to prepare alternating polyketones has attracted great
Ž .
interest in recent 20 years, especially after the highly efficient palladium II catalyst was developed by
w
x
SHELL 1,2 , a large number of researchers were absorbed to study the copolymerization in order to
find a cheap and excellent catalyst or modify the highly efficient palladium II -bisphosphine catalyst.
Ž .
About the chain propagation mechanism, Sen first proposed an alternative insertion chain propaga-
w x
tion mechanism according to the copolymer analysis 3 , and was accepted widely. Drent also
w x
proposed a similar but more perfect mechanism also according to the copolymer analysis 4 . But
w x
Batistini and Consiglio 5 proposed a new chain propagation mechanism involving cationic Pd—
carbene species in order to account for the formation of COrpropylene copolymer with spiroketal
repeating units. About this mechanism, Sen 6 and Wong et al. 7 showed strong disagreement with
Consiglio. They all believed that polyspiroketal was formed after polyketone was produced under
certain conditions, and Sen proved that there existed an interconversion relationship between
polyspiroketal and polyketone in certain conditions without metallic palladium.
w x
w x
Chain initiation and termination mechanisms, up to now, were all proposed according to the
end-group analysis. Sen first proposed a Pd—H initiation in CHCl3 solvent. H is coming from trace
w x
protic impurity of the system, for example H2O and CH3OH. Later, Sen 8 proposed that chain
initiation was mainly due to Pd—H and secondarily due to insertion of CO into Pd—OCH3 bond in
w x
CH3OH solvent. Drent et al. 4 analyzed the end groups of low molecular weight COrethylene
co-oligomers with GC-MS technique, and found three kinds of general formula polyketones
Ž
.
Ž
.
Ž
.
Ž
.
CH3CH2 COCH2CH2 nCOOCH3 keto-esters , CH3O COCH2CH2 nCOOCH3 diesters and
Ž
.
Ž
.
CH3CH2 COCH2CH2 nCOCH2CH3 diketiones . According to this result, Drent proposed that,
chain initiation was firstly and mainly due to insertion of CO into Pd—OCH3 bond, less Pd—H was
produced when chain was terminated as alcoholysis of the palladium–acyl bond, Pd—H can also
initiate the copolymerization.
In order to support the alternative chain propagation mechanism, Pd—H initiation and Pd—OCH3
w x
w
x
w x
initiation mechanism, Brumbaugh et al. 9 , Dekker et al. 10–12 and Rix et al. 13 have prepared
many model complexes, with which olefin insertions into Pd—COR bond and Pd—COOR bond, and
Ž
.
Ž
.
CO insertions into Pd—R bond and Pd—OR bond were studied. With complex BPY Pd CH3 2 as
13
w
x
model catalyst, and with C-NMR technique, Brookhart et al. 25 studied the copolymerization of
13CO with p-tertbutyl styrene, observed the carbonyl resonances of the intermediates in the initial
several steps of 13CO and p-tertbutyl styrene alternative insertions, provided more direct evidence for
this copolymerization mechanism.
From the above statements, we can demonstrate that, the real polyketone catalytic systems are
highly active, and intermediates are usually too reactive to be isolated or even detected; up to now,
w
x
mechanisms proposed by researchers are all according to the copolymer analysis 3–8,14 , and
evidence is not adequate. Therefore, several groups have studies model systems using other ligands
Ž
.
Ž
.
mainly dinitrogen systems and olefins styrene, norborene, norboradiene for which intermediates
Ž
.
can be detected by, e.g., NMR or IR or sometimes isolated. While such model studies provide
valuable background information, it is often difficult to translate the results to the real catalytic
systems.
Ž
Ž .
.
Here, we selected real catalytic systems palladium II -bisphosphine catalyst to study COrethyl-
ene copolymerization reaction and mechanisms under real reaction conditions. Influence of
Ž .
Ž .
DPPPrrPd II mole ratio, CF3COOHrPd II mole ratio and solvent were studied in details. End
1
groups of polyketones prepared from various solvents were analyzed. High-pressure in situ H-NMR,

(
)
H.-K. Luo et al.rJournal of Molecular Catalysis A: Chemical 151 2000 91–113
93
31P-NMR and IR techniques were used to study the mechanism under real reaction conditions.
˚
Ž .
Finally, with EXAFS techniques, CN and R A of palladium in catalytic systems were analyzed.
These studies provided some more direct and valuable evidence for studying the mechanisms, some
results demonstrated from model studies by other groups were translated to the real catalytic systems,
and therefore, made the copolymerization mechanisms more prefect.
2. Results and discussion
( )
2.1. Studies on palladium II -bisphosphine catalyzed copolymerization of CO with ethylene
( )
2.1.1. InÕestigations on the effect of DPPPrrPd II mole ratio on COrethylene copolymerization
Ž .
Ž
.
DPPPrrPd II mole ratio has great influence on this copolymerization see Table 1 . Without
Ž .
Ž .
DPPPr, only palladium 0 black was observed to generate. When DPPPrrPd II s2.5, only light
yellow solution was obtained; under these two conditions copolymerization could not be performed.
Ž .
When DPPPrrPd II is between 0 and 1.5, the catalytic activity was increased rapidly with the
Ž .
Ž .
increasing of DPPPrrPd II mole ratio. When DPPPrrPd II is between 1.5 and 2.5, the catalytic
Ž .
activity decreased rapidly with the increasing of the DPPPrrPd II mole ratio. This result was
rationalized in the following high pressure in situ 31P-NMR studies.
( )
2.1.2. InÕestigations on the effect of CF3COOHrPd II mole ratio on COrethylene copolymerization
Ž .
Ž
.
CF3COOHrPd II mole ratio also has great influence on this copolymerization see Table 2 .
Ž .
Ž .
Without CF3COOH, palladium II -bisphosphine catalyst has no activity, only palladium 0 black was
observed generated, clearly indicating CF3COOH is an essential component of this catalyst. When
Ž .
Ž .
CF3COOHrPd II mole ratio is between 0 and 10, with the increase in CF3COOHrPd II mole ratio,
Ž .
the catalytic activity increased quickly. When CF3COOHrPd II mole ratio is above 10, with the
Ž .
increase in CF3COOHrPd II mole ratio, the catalytic activity increased slowly. The effect of acid or
Ž .
anions may be displayed through its stabilization on cationic palladium II , and will be further probed
in the following EXAFS studies.
2.1.3. Copolymerization in Õarious solÕents
2.1.3.1. Copolymerization in Õarious alcoholic solÕents. In methanol, the catalyst has the most
efficient activity, and with the increase in the carbon number of solvent from methanol to 1-pentanol,
Ž
.
the reactivity decreased significantly see Table 3 . One reason may be mass transference rate was
lowered with the increase in the degree of solvent viscosity from methanol to 1-pentanol. The other
reason may be that the insertion rate of CO into Pd—OR bond was lowered with R group becoming
Ž
.
bigger from –CH3 to – CH2 3CH3, which may be the main chain initiation and will be discussed in
the following.
Table 1
Ž .
Effect of DPPPrrPd II mole ratio on COrethylene coplymerization
Ž
.
2
Reaction conditions: Pd OAC 0.075 mmol, CF3COOH 0.1 mmol, CH3OH 40 ml, Pressure 6.0 MPa, COrethylene 1:1, 908C, 1 h.
Ž .
DPPPrrPd II mole ratio
0
1.25
1.5
2
2.5
Ž .
Yield g
Reaction rate Kgr gPd h
Ž .
Pd 0
0
15.1
5.7
18.5
7.0
10.0
3.8
0
0
Ž
.

(
)
94
H.-K. Luo et al.rJournal of Molecular Catalysis A: Chemical 151 2000 91–113
Table 2
Ž .
Effect of CF3COOHrPd II mole ratio on COrethylene copolymerization
Ž
.
2
Reaction conditions: Pd OAC 0.025 mmol, DPPPr 0.0375 mmol, CH3OH 40 ml, pressure 6.0 MPa, CO:ethylene 1:1, 908C, 1 h.
Ž .
Ž .
Yield g
Ž
.
CF3COOHrPd II mole ratio
Reaction rate kgr gPd h
0
1
1.25
1.5
2
0
0
;0
0.38
0.75
3.2
Trace
1.0
2.0
8.5
3
4
5
10
20
50
62
10.5
12.2
13.2
13.8
14.0
18.5
18.5
3.9
4.6
5.0
5.2
5.3
7.0
7.0
Ž
.
2.1.3.2. Copolymerization in Õarious nonalcoholic solÕents. The experimental results see Table 4
showed that, besides methanol and acetone, in propionaldehyde, N,N-dimethylformamide, N,N-di-
methylacetamide, glacial acetic acid, dimethylsulfoxide, the catalyst exhibited moderate efficient
activity, especially in glacial acetic acid, the catalytic reaction can proceed without CF3COOH. But in
Ž .
ethyl ether and acetic ether, the catalyst deactivated completely, palladium II was reduced to
Ž .
palladium 0 black observed and mostly deposited on the inner autoclave wall.
Solvent effect will be studied further in the following.
( )
2.2. Mechanistic studies on palladium II -bisphosphine catalyst
2.2.1. Chain initiation, termination and solÕent effect
Solvent is an important component of the catalytic system, for example, methanol participates the
chain initiation and termination; therefore, studies on solvent effect is very useful to probe the
catalytic mechanisms, especially for chain initiation and termination mechanisms. Here, end groups of
polyketones prepared from different solvents were analyzed with high-resolution NMR technique. The
1H-NMR data and assignments are listed in Table 5, the 13C-NMR data and assignments are listed in
Tables 6 and 7, respectively.
w
x
From Tables 5–7, we can conclude that the backbone of these polyketones is – CH2CH2CO –, in
n
which only very few fragments of two ethylene monomers are linked together. When these spectra are
enlarged, the resonances due to end groups are very clear. Because end groups are produced in chain
initiations and terminations, chain initiation and termination mechanisms can be induced from the
end-group analysis. In this paper, end-group analysis showed that polyketones produced from
methanol have two approximately equimolar –COOCH3 and –CH2CH3. But it is not clear which
group is the ‘‘head’’ and which is the ‘‘tail’’ of the polymer; for this reason, high-pressure in situ
Table 3
Copolymerization results in various alcoholic solvents
Ž
.
Reaction conditions: Pd OAC 0.025 mmol, DPPPr 0.0375 mmol, CF3COOH 0.1 ml, solvent 40 ml, pressure 6.0 MPa, COrethylene 1:1,
2
908C, 1 h.
Solvent
Methanol
Ethanol
1-Propanol
1-Butanol
1-Pentanol
Ž .
Yield g
Reaction rate kgr gPd h
13.0
4.9
10.0
3.8
9.1
3.4
4.3
1.6
1.9
0.7
Ž
.

(
)
H.-K. Luo et al.rJournal of Molecular Catalysis A: Chemical 151 2000 91–113
95
Table 4
Copolymerization results in various nonalcoholic solvents
Reaction conditions: pressure 6.0 MPa, COrethylene 1:1, 1 h.
Ž .
Yield g
Ž
.
Pd OAC
2
Solvent
Amount of
Reaction rate
DPPPr
CF3COOH
Temperature
Ž
.
Ž
.
Ž
.
Ž
.
Ž
.
Ž .
8C
solvent ml
gr gPd h
mmol
mmol
mmol
Ž
.
Methanol for comparison
Acetone
40
40
50
50
50
50
50
50
50
18.5
6.0
9.6
8.7
4.0
3.5
2.0
0
7000
2200
914
828
381
333
210
0
0.025
0.025
0.10
0.10
0.10
0.10
0.10
0.025
0.025
0.0375
0.0375
0.15
0.15
0.15
0.15
0.15
0.0375
0.0375
1.25
1.25
5.0
5.0
5.0
5.0
5.0
1.25
1.25
90
100
95
95
95
95
95
90
90
Propionaldehyde
N,N-Dimethylformamide
N,N-Dimethylacetamide
Glacial acetic acid
Dimethylsulfoxide
Ethyl ether
Acetic ether
0
0
1H-NMR and IR techniques were used to study this catalytic copolymerization in methanol and
1
Ž
.
2-ethyl hexanol, respectively see below . Pd—H signal was not observed in the in situ H-NMR
studies near y7 ppm, and in the in situ IR studies near 2050 cmy1. But three IR absorptions due to
palladium carbonyl intermediates were trapped successfully, suggested that chain initiation was
Ž
Ž ..
mainly due to CO insertion into Pd—OR bond, produced end group –COOCH3 see Eq. 1 chain
termination was mainly due to the protolysis of palladium–acyl bond, produced end group –CH2CH3
Ž
Ž ..
see Eq. 2 . End-group analysis also showed that polyketones prepared from nonalcoholic solvents
have predominant end group –CH2CH3, and less –CH5CH2 which was produced by reaction of
Ž
Ž ..
Ž
b-H elimination of palladium–alkyl bond see Eq. 3 ; chain initiation was only from –CH2CH3 see
Ž ..
Eq. 4 . Therefore, in nonalcoholic solvent, chain initiation was due to the insertion of ethylene into
Pd—H bond see Eq. 5 . H came from protic impurity of solvent, or trace H2 in CO comonomer.
Chain termination was mainly due to hydrogen activation see Eq. 6 .
Ž
Ž ..
Ž
Ž ..
Pd–ORqCO Pd–COOR
1
Ž .
Pd–CH2CH2COyP ROH Pd–ORqCH3CH2CO–P
2
Ž .
b-H
Pd–CH2CH2CO–P
Pd–HqCH2 sCHCO–P
3
Ž .
elimination
CO
Pd–CH2CH3
Pd–COCH2CH3
4
Ž .
Pd–HqC2 H4 Pd–CH2CH3
5
Ž .
H 2
Pd–CH2CH2CO–P
Pd–HqCH3CH2CO–P
6
Ž .
Table 5
1H-NMR data and assignments of polyketones prepared from various solvents
Ž .
CH3CH2CO– – CH2CH2CO –n –COOCH3 –CH2CH2OCH3 CH2sCHCO– CH2 sCHCO–
Solvent
Assignment
Ž
.
main
CH3CH2CO–
Ž .
1.15 3
Ž .
2.68 4
Ž
.
.
.
.
Methanol
Acetone
3.00 main
3.84
4.08
Ž .
Ž .
Ž
Ž .
6.07 3
Ž .
6.4 2
1.04 3
2.55 4
2.87 main
Ž .
Ž .
2.56 4
Ž .
2.56 4
Ž
Propionaldehyde 1.04 3
2.87 main
Ž .
1.05 3
Ž
Ž .
6.08 3
Ž .
6.4 2
N,N-Dimethyl-
formamide
Glacial acetic
acid
Dimethyl-
sulfoxide
2.88 main
Ž .
1.45 3
Ž .
2.97 4
Ž
.
.
3.29 main
Ž .
1.65 3
Ž .
3.20 4
Ž
3.48 main
Ž .
7.01 3
Ž .
6.35 2

(
)
96
H.-K. Luo et al.rJournal of Molecular Catalysis A: Chemical 151 2000 91–113
Table 6
13C-NMR data of polyketones prepared from various solvents
Solvent
No.
1
2
3
4
5
6
7
8
9
10
11
12
Ž
.
Ž
.
main
main
Methanol
9.53 24.97 30.57 38.77
39.63 43.55 55.60
217.62 219
222.04
Acetone
9.30 24.82
9.63 25.14
9.39 24.89
9.01 24.51
10.94 26.5
38.82
38.92
38.69
38.33
40.30
43.38
43.67
43.44
43.11
44.84
108.88 155.06 217.56 219.57
109.17 155.38 217.89 219.90
108.93 155.14 217.64 219.66
154.79 217.21 219.20
218.68
Propionaldehyde
N,N-Dimethylformamide
Glacial acetic acid
Dimethylsulfoxide
223.30
The above discussion demonstrated that alcoholic solvent participated in the chain initiation and
termination. Although copolymerization can proceed in acetone, propionaldehyde, N,N-dimethyl-
formamide, N,N-dimethylacetamide, glacial acetic acid and dimethylsulfoxide; in fact, these solvents
did not participate the reaction. It seems that non-alcoholic solvent has no influence on this catalytic
Ž
.
reaction, only acting as diluent and has no difference. But in fact, it is not see Table 6 . Under the
same conditions, but in different non-alcoholic solvent, the catalyst showed significantly different
activity, especially in ethyl ether and acetic ether, etc.; the catalyst deactivated completely,
Ž .
Ž .
palladium II was reduced to palladium 0 black, showed clearly that solvent not only act as diluent,
Ž .
Ž .
but also had suitable stabilization on cationic palladium II to assist palladium II showing highly
efficient catalytic activity. This kind of stabilization may be displayed through coordination of solvent
Ž .
with palladium II , and will be studied further in the following EXAFS studies.
In conclusion, a solvent has at least two kind of effects. The first is acting as diluent, the second is
Ž .
suitable stabilization on cationic palladium II , alcoholic solvent has the third one to participate the
chain initiation and termination.
2.2.2. High-pressure in situ 31P-NMR studies
( )
Ž .
Ž .
2.2.2.1. InÕestigations on DPPPrrPd II mole ratio. When a DPPPrrPd II s0.5, only one signal
Ž
.
Ž .
Ž .
was observed at 13.26 ppm 1a . When b DPPPrrPd II s1.0, only one signal was observed at
Table 7
Assignments of 13C-NMR data of polyketones prepared from various solvents
Ž
.
No.
Chemical shift d
Assignments
Experiment
Calculated
1
2
3
9.53
24.97
30.57
38.77
39.63
43.55
55.60
108
9.41
24.90
30.41
39.41
39.41
43.32
51.4
–COCH2CH3
–COCH2CH2CH2CH2CO–
–CH2CH2COOCH3
Ž
.
Ž
.
4 main
– CH2CH2CO –, –COCH2CH3
n
5
6
7
–CH2CH2COOCH3
–COCH2CH2CH2CH2CO–
–COOCH3
Ž
Ž
.
.
8 impurity
–OCOCF3
9 impurity
155
–OCOCF3
Ž
.
Ž
.
– CH2CH2CO –
n
10 main
11
12
217.62
219
222
–COCH2CH2CH2CH2CO–
–COCH2CH3

(
)
H.-K. Luo et al.rJournal of Molecular Catalysis A: Chemical 151 2000 91–113
97
Table 8
In situ 31 P-NMR studies on DPPPrrPd mole ratio
Ž
.
Conditions: pressure 2.0 MPa, COrethylene 1:1 , Variation of temperature 40–908C, 124.5 ppm of 85% H3PO4 was used as external
standard.
DPPPrrPd mole ratio
31 P-NMR 1a
Chemical 2a
Shift 3a
Activity
0.5
1.0
1.5
2.0
3.0
4.0
13.26
13.44
14.05
low
high
high
low
none
none
2.27;1.18
1.93
1.81
y13.89
y14.94
1.76
Ž
.
Ž .
13.44 ppm 1a ; showing that only mono-chelate ring complex 1 was produced and had highly
Ž
.
Ž .
Ž .
efficient catalytic activity see Table 8 . When c DPPPrrPd II s1.5, two signals were observed at
Ž
.
Ž
.
14.05 ppm 1a and 2.27–1.18 ppm 2a whose integral ratio was 1:2, showing that two equimolar
Ž .
Ž .
Ž .
d
mono-chelate ring complex
1
and bis-chelate ring complex
2
were produced. When
Ž .
Ž
.
DPPPrrPd II s2.0, only one signal was observed at 1.93 ppm 2a , showing that only bis-chelate
Ž .
Ž .
Ž .
Ž .
ring complex 2 was produced. When e DPPPrrPd II s3.0, two signals were observed at 1.81
Ž
Ž
.
.
Ž
.
Ž .
ppm 2a and y13.89 ppm 3a . When f DPPPrrPd II s4.0, two signals were observed at 1.76
Ž
.
Ž .
Ž .
Ž .
Ž .
ppm 2a and y14.94 ppm 3a . Under conditions of e and f , coordination vacant of palladium II
were saturated by DPPPr. Besides bis-chelate ring complex 2 , free DPPPr existed in the system.
Copolymerization result showed that, under these two conditions, the activity deactivated completely.
Ž .
Table 1 indicates that the catalyst had the most efficient activity when DPPPrrPd II s1.5. Table
Ž .
Ž .
8 indicates that two equimolar mono-chelate ring complex 1 and bis-chelate ring complex 2 were
produced. In order to rationalize this result, we proposed a interconversion relationship between these
Ž
Ž ..
two complexes see Eq. 7 . During COrethylene copolymerization, DPPPr would be decomposed
Ž .
Ž .
andror exhausted slowly, then Eq. 7 would shift to the left to ensure the activity of palladium II .
Therefore, although bis-chelate ring complex 2 had no activity, it could increase the total activity
and prolong the catalyst life, so palladium II exhibited the most efficient activity.
Ž .
Ž .
Ž .
7
Ž .
In order to probe the existence of a mono-chelate ring complex 1 and bis-chelate ring complex
Ž .
Ž
.
Ž
. Ž .
2 , we separately prepared model complexes DPPPr Pd OCOCF3 2 a having mono-chelate ring,
and DPPPr Pd OCOCF3 2 e having bis-chelate ring, which were studied further in the following
Ž
.
Ž
. Ž .
2
EXAFS studies.
2.2.2.2. In situ 31P-NMR studies under different ambience. No evident chemical shift was observed
Ž
. Ž
.
between under pure ethylene and mixed COrethylene 1:1 see Table 9 . Under pure CO, signal 1a
Ž .
of mono-chelate ring complex 1 had significant shift from 14.05 ppm to 5.63 ppm due to
coordination of CO with palladium II , which is coincident with in situ IR result under pure CO. This
Ž .
result suggested that in the competitive coordination of CO and ethylene with palladium, ethylene
tends to be easier than CO, or the coordination time of ethylene with palladium was much longer than
that of CO.

(
)
98
H.-K. Luo et al.rJournal of Molecular Catalysis A: Chemical 151 2000 91–113
Table 9
In situ 31 P-NMR data under different ambience
Ž
.
Ž .
Conditions: Pd OAC 0.075 mmol, DPPPrrPd II s1.5:1, CH3OH 0.8 ml, CF3COOH 0.2 ml, Pressure 2.0 MPa, 124.5 ppm of 85%
2
H3PO4 was used as external standard.
Ambience
Chemical shift
1a
2a
Corethylene
Ethylene
CO
14.06
13.58
5.63
2.27;1.18
1.92;1.02
1.89;0.98
1
2.2.3. High-pressure in situ H-NMR studies
1
2.2.3.1. Studies on coordinated ethylene. First, in situ H-NMR experiment was carried out in
CD3OD, with the temperature rising from 40 to 908C, free ethylene at 4.9 ppm was observed to
decrease rapidly; meanwhile, three signals appeared and increased at 0.6 ppm, 2.9 ppm and 4.4 ppm
Ž
.
see Table 10 . With CD3COCD3 as solvent, free ethylene at 4.7 ppm was observed to decrease
rapidly, and a new signal at 0.5 ppm appeared. In order to assign these new resonances, we designed a
Ž .
series of probe experiments; for example, under pure CO, under pure N2, and under DPPPrrPd II s
1
2.5 etc., under these unreactive conditions, no change was observed in the in situ H-NMR spectra at
different temperatures.
With CF3COOD as solvent and under pure ethylene, very strong peak was observed at 1.2 ppm
even at ambient temperature. During the temperature raising process, free ethylene at 4.9 ppm
decreased rapidly and a new peak at 0.6 ppm appeared. GC-MS analysis indicate C4 s, C6 s, C8 s
Ž .
were produced, proved that palladium II -bisphosphine exhibited high efficient activity for ethylene
oligomerization, indicating clearly there is no thermodynamic barrier to ethylene insertion into the Pd
—alkyl bonds in this system.
The above result showed that the resonances at 0.6 ppm, 2.9 ppm and 4.4 ppm were only observed
under real reaction conditions, did not appear under many unreactive conditions. Compared with the
1H-NMR spectrum of polyketone, 2.9 ppm was assigned to –COCH2– in the backbone, showing
Table 10
New resonances observed in high pressure in situ 1H-NMR
Reaction pressure 2.0 MPa, Increasing, Decreasing.
1
Ž .
DPPPrrPd II
Ž
.
Solvent
Ambience
Temperature
Ž .
8C
H-NMR data d
Activity
Ž
.
Ž
.
mol ratio
2.0 MPa
a
b
c
d
COrC2 H4
40
70
40
90
40
90
40
90
40
90
40
90
2.9
2.9
4.9
4.9
4.7
4.7
Ž
.
CD3OD
1.5:1
1.5:1
1:1
1:1 v:v
0.6
0.5
4.4
High
High
None
None
None
Higha
COrC2 H4
Ž
.
CD3COCD3
CF3COOD
CF3COOD
CF3COOD
CF3COOD
1:1 v:v
pure
N2
pure
CO
pure
C2 H4
pure
C2 H4
1:1
4.9
4.9
4.9
4.9
2.5:1
1:1
0.6
0.6
aCatalytic activity of oligmerization of ethylene.

(
)
H.-K. Luo et al.rJournal of Molecular Catalysis A: Chemical 151 2000 91–113
99
Ž .
Ž .
clearly two resonances at 0.6 ppm and 4.4 ppm 1 had close connection with copolymerization, 2
was not due to materials and products, suggesting that these two resonances are due to palladium
Ž .
intermediates. Active center is cationic palladium II , therefore, intermediates should be cationic
Ž .
Ž .
palladium II complexes, further suggest that these new resonances are due to cationic palladium II
intermediates.
w
x
To resonance at 0.6 ppm, large number of authors 15–18 showed that normal organic compounds
did not appear here, but methyl or methylene linked with palladium usually appeared. Therefore,
resonance at 0.6 ppm may be due to methyl or methylene linked with palladium contained in
intermediates.
w
x
To resonance at 4.4 ppm, which appeared at the up-field of free ethylene at 4.9 ppm. Refs. 19,20
showed that coordinated ethylene with palladium appeared near 4.4 ppm in the up-field of free
ethylene. Therefore, resonance at 4.4 ppm may be due to coordinated ethylene with palladium. This
result showed that before insertion into the Pd—acyl bond to realize chain propagation, ethylene was
activated through coordination with palladium.
1
2.2.3.2. In situ H-NMR experiments for Pd–H. Resonance of M-H usually appears in up-field,
w
x
slightly influenced by other proton resonance, therefore, is easily to be trapped 21–23 . Resonance of
w
x
Pd–H usually appears near y7 ppm 21,22 . With CD3OD and CD3COCD3 as solvents, Pd–H was
not observed, we suspected that Pd–D was produced instead of Pd–H in deuterated solvent. For this
reason, we designed and performed a series of in situ experiments in CH3OH instead of CD3OD.
With CH3OH as solvent and under real reaction conditions, copolymer was observed to produce more
and more rapidly with temperature rising from 20 to 908C, but Pd–H was not observed in up-field.
Under unreactive conditions, such as under pure CO, pure N2, Pd–H was also not observed. This
1
result indicates that Pd–H concentration was too low to be observed with H-NMR technique. Pd–H
was also not observed in in situ IR experiments near 2050 cmy1. In conclusion, chain initiation due to
Pd–H was slight in alcoholic solvents.
2.2.4. High-pressure in situ IR studies
Ž
.
Ž
.
We selected model complex DPPPr Pd p-CH3PhSO3 2 as catalyst and 2-ethyl hexanol as solvent,
because they have no evident IR absorption between 2200 and 1600 cmy
.
1
2.2.4.1. In situ IR studies under pure CO. At ambient temperature, 4.5 MPa CO was introduced, and
1
immediately, three IR absorptions were observed at 2113 cmy1, 1900 cmy and 1810 cmy1, and with
1
1
1
the temperature rising, 1900 cmy decreased gradually, 1810 cmy increased gradually, 2113 cmy
is due to free CO, and 1900 cmy and 1810 cmy were difficult to be assigned concretely, but it is no
doubt they were due to palladium carbonyl complex, showing clearly CO could coordinate with
palladium under pure CO ambience. This result is coincident with in situ 31P-NMR under pure CO.
1
1
2.2.4.2. In situ IR studies under real reaction conditions. In order to control the reaction rate to trap
high reactive intermediates, batch experiment technique was used here. First, autoclave was heated to
858C and pressure of ethylene was controlled at 4.0 MPa. Then, 0.4 MPa CO was introduced, and
Ž
.
immediately, in situ IR system was recycled and IR spectra were recorded see Fig. 1 . Before CO
1
1
was introduced, only two absorptions due to free ethylene was observed at 2013 cmy and 1860 cmy
Ž
.
see Fig. 1-a . After 0.4 MPa, CO was introduced, four new absorptions immediately appeared at
1690 cmy1, 1670 cmy1, 1638 cmy1 and 1616 cmy1 see Fig. 1-b . After reaction for 40 min, a
Ž
.
shoulder appeared at 1700 cmy1, at the same time, 1690 cmy and 1670 cmy increased, 1638 cmy
1
1
1

(
)
100
H.-K. Luo et al.rJournal of Molecular Catalysis A: Chemical 151 2000 91–113
Ž
.
Ž
.
Ž . Ž
as catalyst. DPPPr Pd p-
Fig. 1. High-pressure in situ IR spectra under real reaction conditions with DPPPr Pd p-CH3PhSO3
2
.
Ž .
Ž .
Ž .
Ž .
Ž .
CH3PhSO3 2 s1 mmol, 2EHs80 ml. a 808C, 4.0 MPa C2 H4. b 858C, introduce 0.4 MPa CO. c 858C, 4.2 MPa, after 40 min. d
858C, 3.8 MPa, after 120 min. e 858C, 3.6 MPa, after 240 min. f 858C, 2.9 MPa, introduce 1.0 MPa CO and after 20 min. g 908C, 2.8
Ž .
Ž .
MPa, introduce 1.3 MPa CO and after 30 min.
1
and 1616 cmy significantly decreased see Fig. 1-c . With reaction time being prolonged, these five
Ž
.
new absorptions all increased again and became strong, a sixth new absorption appeared at 1970
1
y
cmy see Fig. 1-d,e . Then 1.0 MPa CO was introduced, 20 min later 1700 cm 1, 1690 cmy1 and
Ž
.
1670 cmy continued increasing, 1638 cmy decreased significantly, 1616 cmy disappeared rapidly
1
1
1
see Fig. 1-f . More 1.3 MPa CO was introduced and 30 min later, 1700 cm 1, 1690 cmy1 and 1670
y
Ž
.
1
1
1
cmy became very strong to form a broad absorption, 1638 cmy disappeared, 1970 cmy increased
Ž
.
slowly see Fig. 1-g .
1
1
Among these six new absorptions, 1700 cmy1, 1690 cmy and 1670 cmy are due to ycso
absorptions in the polyketones. 1970 cmy1, 1638 cmy and 1616 cmy were likely due to palladium
1
1
carbonyl containing in intermediates, will be assigned in detail in the following.

(
)
H.-K. Luo et al.rJournal of Molecular Catalysis A: Chemical 151 2000 91–113
101
w x
w
x
Brumbaugh et al. 9 and Dekker et al. 10–12 have prepared many model complexes, studied the
insertions of CO into Pd–R and Pd–OR bonds, and insertions of olefins into Pd–COR and Pd–COOR
w
x
bonds. This kind of complex has been reported earlier 24 .
w
x
Refs. 9,10,24 showed that ycso IR absorptions of Pd–COR and Pd–COOR bonds are between
1700 and 1600 cmy1, and significantly influenced by R group, which rule that when R is bigger, the
ycso IR absorptions is lower. Therefore, 1638 cmy were likely due to Pd–COOR in which R is
1
Ž
.
Ž
.
– CH2CH2CO OCH2CH CH2CH3 CH2CH2CH2CH3, because here R is big, the ycso IR absorp-
n
1
Ž .
tions is low. Together considering the in situ H-NMR result, we proposed an intermediate 3 .
w
x
The literature 9,12 showed that, after the insertions of olefins into Pd–COR and Pd–COOR
bonds, ycso IR absorptions all significantly shifted to near 1620 cmy1, because a stable five-mem-
y
1
Ž .
w x
bered ring 4 was formed 9 . Therefore, 1616 cm was likely due to five-membered ring contained
Ž .
in intermediate 5 produced by intramolecular insertion of coordinated ethylene into Pd–COOR bond
Ž .
in intermediate 3 .
1
y1
1970 cmy was observed during the copolymerization, was different with that 1900 cm and
Ž
1810 cmy observed under pure CO, therefore, was likely due to a certain intermediate, and
1
.
Brookhart 25 has trapped the 13C-NMR signal of coordinated CO at low temperature. Here, we
w
x
suggest that 1970 cmy was due to intermediate 6 .
1
Ž .

(
)
102
H.-K. Luo et al.rJournal of Molecular Catalysis A: Chemical 151 2000 91–113
Ž .
Ž .
2.2.4.3. About chain initiation and propagation. The existence of intermediates 3 and 5 showed
that chain initiation was firstly and mainly due to CO insertion into Pd–COOR bond, and chain
propagation was alternative mechanism. According to these, we proposed a mechanism shown in
y
1
Scheme 7. With intermediates 3 and 5 , the growth and decline rule of 1638 cm and 1616 cmy1
Ž .
Ž .
could be rationalized exactly.
Ž
.
Ž
.
Fig. 2. Best fits dashed lines to the Fourier-filtered Pd K-edge EXAFS of DPPPr PdCl2 and to the Fourier transform.

(
)
H.-K. Luo et al.rJournal of Molecular Catalysis A: Chemical 151 2000 91–113
103
In Fig. 1, at the beginning of copolymerization, the pressure of ethylene was much higher than that
1
y1
of CO, 1616 cmy was stronger than 1638 cmy1 see Fig. 1-b . This is because 1638 cm appeared
Ž
.
first after CO insertion into Pd–OR bond, and then changed to 1616 cmy after ethylene insertion
1
into Pd–COOR bond, and the next insertion of CO became slow because the pressure of CO was low.
When reaction time was prolonged, most intermediates participated to long chain propagation,
Ž
.
therefore, all decreased see Fig. 1-c . With CO being almost exhausted, the pressure of CO was more
low, chain propagation decreased, chain termination increased, and chain initiation was also in-
creased; therefore, 1638 cmy and 1616 cmy all grow up gradually, and 1638 cmy became
1
1
1
1
stronger than 1616 cmy see Fig. 1-d,e . When much CO was introduced, the copolymerization rate,
Ž
.
especially the insertion rate of CO was significantly increased, 1638 cmy and 1616 cmy decreased
1
1
rapidly. 1616 cmy disappeared first see Fig. 1-f , than 1638 cm followed see Fig. 1-g .
1
y1
Ž
.
Ž
.
Ž .
Ž .
2.2.4.4. Comparison of insertion rate of CO with ethylene. Intermediates 3 and 5 were observed
together only when pressure of ethylene was much higher than that of CO. After much CO was
Ž .
introduced, intermediate 5 declined immediately. This result indicated clearly that CO insertion was
much faster than that of ethylene, according to which we predicted that reaction rate under
COrethylene -1 was higher than that of under COrethylenes1. This idea has been proved to be
w
x
correct in Shell’s recent patent 26 .
Table 11
˚
Ž
.
Ž
.
Ž
Pd K-edge EXAFS-derived CN and R A of standard sample DPPPr PdCl2
Filtered range A , 0.92;2.63; fitting range in K-space, 1.50;11.52 A
y1
3
˚
˚
Ž
.
.
with the weight of k .
Shell
EXAFS
CN
Crystal
˚
˚
.
Ž
.
Ž
R A
R-factor
CN
R A
R-factor
Ž .
2.0 2
Ž .
2.4 2
Ž .
2.24 1
Ž .
2.35 1
Pd–P
Pd–Cl
0.12
2
2
2.244, 2.249
2.351, 2.358
0.027

(
)
104
H.-K. Luo et al.rJournal of Molecular Catalysis A: Chemical 151 2000 91–113
Table 12
˚
Ž
.
Ž . .
Pd K-edge EXAFS-derived CN and R A of complexes having general formula L Pd OCOCF3
2
Ž
˚
Ž
.
L
Pd–O
CN
Pd–P
CN
R-factor
Filtered range A
Fitting range in
Activity order
K-space with K3
˚
˚
Ž .
A
Ž
.
R A
y1
˚
Ž
.
weight A
Ž .
2.0 8
Ž .
2.06 2
Ž .
2.05 3
Ž .
1.8 7
Ž .
2.25 3
DPPPr
DPPBu
DPPEt
2.07 3
Ž .
0.14
0.12
0.13
0.895;2.557
0.895;2.224
0.869;2.787
1.50;11.52
1.50;11.52
1.50;11.52
1
2
3
Ž .
Ž .
Ž .
1.4 3
1.3 5
2.33 3
Ž .
1.7 6
Ž .
1.2 4
Ž .
2.38 3
Ž .
2.2.4.5. ActiÕation of CO. The existence of intermediate 6 show that CO was activated through
coordination with palladium before its insertion. This result was coincident with that of Brookhart
w
x
25 .
2.2.5. EXAFS studies
Ž
.
Ž .
Using DPPPr PdCl2 f as standard sample for calibrating the fitting parameters in this work, CN
˚
Ž .
and R A of central palladium were calculated. Six samples were analyzed. Four solid samples
Ž
.
Ž
.
Ž .
Ž
.
Ž
.
Ž .
Ž
.
Ž
.
Ž .
Ž
.
DPPPr Pd OCOCF3 2 a , DPPBu Pd OCOCF3 2 b , DPPEt Pd OCOCF3 2 c , and DPPPr -
2
Ž
.
Ž
.
Ž
.
2
Fig. 3. Best fits dashed lines to the Fourier-filtered Pd K-edge EXAFS of DPPPr Pd OCOCF3 and to the Fourier transform.

(
)
H.-K. Luo et al.rJournal of Molecular Catalysis A: Chemical 151 2000 91–113
105
Ž
.
Ž .
Pd OCOCF3 2 e , and two solution samples freshly taken before and in the middle of the
Ž .
copolymerization at DPPPrrPd II s1, were selected.
2.2.5.1. EXAFS analysis of standard sample. Fourier transform and fitting curve in K-space of Pd
Ž
.
Ž .
K-edge EXAFS of DPPPr PdCl2 f shown in Fig. 2 showed that experimental curves were well
identical with theoretical curves, especially in low K-space. The satisfactory EXAFS-derived parame-
˚
w
x
Ž . Ž
.
ters and single-crystal data 27 were fixed very closely, especially for R A see Table 11 , provide a
reliable technique to study the relationship between catalyst structure and activity.
2.2.5.2. EXAFS studies on ligand effect and anion effect. EXAFS-derived parameters of complexes
Ž . Ž .
Ž .
a , b and c are shown in Table 12. Their Fourier transforms and fitting curves in K-space are
shown in Figs. 3–5. From Table 12, it can be seen that, with Pd–P bond becoming shorter and with
Pd–O bond becoming longer, the catalytic activity was increased significantly, suggesting that chelate
ring was more stable and the anion will more easily go away, and the corresponding catalyst had more
efficient reactivity. In order to probe the anion effect, the catalytic activity of four complexes having
1
1
different anions was evaluated and shown in Table 13. CF3COOy and p-CH3PhSOy3 belong to
Ž .
strong acids, have excellent stability to cationic palladium II , and are weakly coordinated anions,
Ž
.
Ž
.
Ž
.
2
Fig. 4. Best fits dashed lines to the Fourier-filtered Pd K-edge EXAFS of DPPBu Pd OCOCF3 and to the Fourier transform.

(
)
106
H.-K. Luo et al.rJournal of Molecular Catalysis A: Chemical 151 2000 91–113
Ž
.
Ž
.
Ž
.
2
Fig. 5. Best fits dashed lines to the Fourier-filtered Pd K-edge EXAFS of DPPEt Pd OCOCF3 and to the Fourier transform.
easy to go away to leave the coordinate vacant to comonomers; therefore, activity is high. Although
1
Cly also belongs to strong acid and has excellent stability to cationic palladium II , it coordinates
Ž .
with palladium strongly, is very difficult to separate to leave the coordinate vacant to comonomers;
1
therefore, activity is low. CH3COOy belongs to weak acid, could not stabilize cationic palladium II
Ž .
Ž .
which was reduced to palladium 0 black; therefore, deactivated completely. The above results
indicate clearly that a suitable anion first belongs to strong acid, has excellent stability to cationic
Ž .
palladium II , and easy to split off to leave the coordinate vacant to comonomers.
( ) (
)
(
)
2.2.5.3. Comparison of mono-chelate ring complex a DPPPr Pd OCOCF3 2 with bis-chelate ring
( ) (
)
(
)
Ž .
Ž .
complex e DPPPr 2 Pd OCOCF3 2. EXAFS-derived parameters of complexes a and e are listed
in Table 14. Their Fourier transforms and fitting curves in K-space were respectively shown in Figs. 3
Ž .
and 6. From Table 14, it can be seen that the Pd–P bond and Pd–O bond of complex a were much
Ž .
Ž .
Ž .
shorter than that of complex e , indicating clearly that complex a was more stable than complex e ,
providing direct evidence for their interconversion relationship drawn from in situ 31P-NMR.
2.2.5.4. EXAFS studies on solution samples freshly taken before and in the middle of the copolymer-
( )
Ž .
ization at DPPPrrPd II s1. EXAFS-derived parameters of complex a and two solution samples
are listed in Table 15; their Fourier transforms and fitting curves in K-space are shown in Figs. 3, 7

(
)
H.-K. Luo et al.rJournal of Molecular Catalysis A: Chemical 151 2000 91–113
107
Table 13
Ž .
Evaluation results of complexes having general formula DPPPr Pd A
2
Ž
.
Ž
.
Conditions: complex 0.025 mmol, CH3OH 40 ml, COrethylene 1:1 6.0 MPa, 908C, 1 h.
Ay
Yield g
Reaction rate kgr gPd h
Ž .
Ž
.
CF3COOy
p-CH3PhSOy3
Cly
10.0
7.6
0.95
3.76
2.86
0.357
0
CH3COOy
Ž .
Pd 0
Ž
.
and 8. It can be found that coordination number of sample S1 CNs3.0q2.7s5.7 and sample S2
Ž
.
CNs2.6q2.1s4.7 significantly increased by 1.9 and 0.9, respectively, compared with solid
sample a which is four-coordinated. Therefore, d , 18-electron and five coordinated pyramidal
species see Scheme 8 likely existed in the system besides d , 16-electron and four-coordinated
square planar species, showing that 1 solvent could coordinate with palladium and therefore, had
8
Ž .
Ž
8
.
Ž .
Ž .
significant influence on the activity, 2 five-coordinated intermediates may have existed and acted as
important actors in the catalytic copolymerization.
3. Experimental
3.1. General considerations
Carbon monoxide is the product of Lanzhou Chemical Physics Institute of Academia Sinica.
Ž
.
Ethylene is the product of polymer grade of 303 plant of Lanzhou Petroleum Chemical. Pd OAC
2
was prepared by myself. Ethyl ether was dried and purified with standard method and freshly distilled
before use under N2. Other solvents and reagents were of commercial grade.
Table 14
Ž . Ž
.
Ž
.
2
Comparison of Pd K-edge EXAFS-derived parameters of mono-chelate ring complex
a
DPPPr Pd OCOCF3 with bis-chelate ring
Ž . Ž
.
Ž
.
complex e DPPPr 2 Pd OCOCF3
2
Complex
Pd–O
CN
Pd–P
CN
R-factor
Filtered
range
Fitting range in
K-space with K3
˚
˚
Ž
.
Ž
.
R A
R A
y1
˚
˚
Ž
.
Ž
.
A
weight A
Ž .
Ž .
e
Ž .
2.0 8
Ž .
2.0 5
Ž .
2.07 3
Ž .
2.33 2
Ž .
1.8 7
Ž .
4.9 5
Ž .
a
2.25 3
0.14
0.08
0.895;2.557
0.920;2.889
1.50;11.52
1.50;11.52
Ž .
2.38 1

(
)
108
H.-K. Luo et al.rJournal of Molecular Catalysis A: Chemical 151 2000 91–113
Ž
.
Ž
.
Ž
.
Fig. 6. Best fits dashed lines to the Fourier-filtered Pd K-edge EXAFS of DPPPr 2 Pd OCOCF3 2 and to the Fourier transform.
In situ NMR spectra and normal NMR spectra of complexes were recorded on a varian FT-80A
spectrometer. NMR spectra of polyketones were recorded on a Bruker AM400 spectrometer. In situ
IR and normal IR spectra were recorded on a Specord 75 IR spectrometer.
3.2. General procedure for COrethylene copolymerization
Copolymerization was carried out in a 200 ml stainless steel autoclave. First, catalyst precursor or
components were added to the reactor; immediately, the reactor was sealed and was purged with
Table 15
˚
Ž . Ž .
Comparison of solid sample a with solution samples of EXAFS-derived CN and R A
Code name
Pd–O
CN
Pd–P
CN
R-factor
Filtered
range
Fitting range in
K-space with K3
˚
˚
Ž .
A
Ž
.
R A
y1
˚
˚
Ž
.
Ž
.
A
weight A
Ž .
S1
S2
Ž .
2.0 8
Ž .
2.08 2
Ž .
2.08 2
Ž .
1.8 7
Ž .
2.25 3
a
2.07 3
Ž .
0.14
0.15
0.14
0.895;2.557
0.895;2.429
0.895;2.506
1.50;11.52
1.50;11.52
1.50;11.52
Ž .
Ž
.
Ž .
3.0 3
2.7 12
Ž .
2.19 4
Ž .
2.6 6
Ž .
2.22 4
2.1 9

(
)
H.-K. Luo et al.rJournal of Molecular Catalysis A: Chemical 151 2000 91–113
109
Ž
.
Fig. 7. Best fits dashed lines to the Fourier-filtered Pd K-edge EXAFS of S1 and to the Fourier transform.
ethylene at least three times, then the solvent was introduced and ethylene was charged to chosen
pressure. After the reactor was heated to the chosen temperature in 30 min, equimolar CO was
introduced and total pressure was controlled at about 6.0 MPa during the copolymerization by
introducing equimolar CO and ethylene. After reaction for 1 h, the autoclave was rapidly cooled, and
the unreacted gases were vented. The product slurry was filtered off, washed with fresh methanol, and
dried in the air to give snow-white powder. The average catalytic activity was calculated from
Ž
.
copolymer weight as unite of g polyketoner gPd h .
3.3. In situ NMR experiments
w
x
High-pressure in situ NMR technique developed by Da-Gang Li et al. 28 was used. The in situ
Ž
.
NMR tube fs10 mm , in fact, is a transparent micro-reactor of homogeneous reaction. Therefore,
NMR determination can be carried out under real reaction conditions, and the phenomenon can be
observed clearly during the copolymerization.
3.3.1. RepresentatiÕe procedure for in situ NMR experiments
Catalyst components and solvent were taken exactly and added into the in situ NMR tube;
immediately, the tube was sealed and was purged with ethylene at least three times, and was charged
with 2.0 MPa equimolar CO and ethylene. NMR spectra were recorded with the temperature rising
from 208C to 908C.

(
)
110
H.-K. Luo et al.rJournal of Molecular Catalysis A: Chemical 151 2000 91–113
Ž
.
Fig. 8. Best fits dashed lines to the Fourier-filtered Pd K-edge EXAFS of S2 and to the Fourier transform.
3.4. In situ IR experiments
w
x
High-pressure in situ IR system developed by Wu et al. 29 was used here. 200 ml stainless steel
autoclave was used to perform the copolymerization. Two CaF2 crystals were used as the high-pres-
sure IR cell. Sample space were controlled about 0.2 mm thick. Normal KBr IR cell was used to
deduct solvent absorption in sample. 2-Ethyl hexanol was used as solvent.
Ž
.
Ž
.
1.0 mmol DPPPr Pd p-CH3PhSO3 2 was added into 200 ml reactor, immediately the reactor was
sealed and was purged with ethylene at least three times. Then ethylene was introduced and reactor
was heated to 858C and ethylene pressure was controlled 4.0 MPa, and then 0.4 MPa, 1.0 MPa and
1.3 MPa CO was introduced separately in periodic order; at the same time IR spectra were recorded.
3.5. EXAFS
Ž
.
X-ray absorption spectra were obtained in fluorescence mode solid samples and transmission
Ž
.
Ž
.
Ž
.
mode liquid samples using the BL-7C facility at the Photon Factory Tsukuba, Japan . A Si 111
double-crystal Sagittal focusing monochromator with a position beam energy of 2.5 GeV and an
Ž
.
average stored current of 250 mA was employed.

(
)
H.-K. Luo et al.rJournal of Molecular Catalysis A: Chemical 151 2000 91–113
111
The data were processed on a COMPAQ 486 microcomputer with the Program Library for EXAFS
w
x
Data Analysis written by the Institute of Physics, Chinese Academy of Sciences 30 . The Fourier
transform containing the peaks of interest was filtered into K-space by a Hanning window function.
A least-squares current-fitting procedure was then performed to calculate the structural parameters
per each shell using theoretically calculated components determined by program FEFF 3.25 developed
˚
.
w
x
w
x
Ž
by Rehr et al. 31 and Mustre et al. 32 . The structural parameters CN and R in A with which the
FEFF was started were corrected repeatedly until they were identical to the fitting parameters. The
reduction factor used, So2 was 1.1. The best fit of fitting curves was first chosen by an R factor, then
was transformed into R-space weighted by K8 with the same range to give a fit of the Fourier
w
x
transform. The parameter error estimates were calculated by the recommended method 33 .
3.6. Preparation of complexes
(
)
(
) ( )
3.6.1. Preparation of DPPPr Pd OCOCF3 2
a
Ž
.
2
The following operations were under Ar ambience. To a 100 ml glass bottle, Pd OAC 224.6 mg
Ž
.
1.0 mmol , 25 ml dry ethyl ether and 0.2 ml CF3COOH was added and stirred for 10 min. Exchange
of anions was performed see Eq. 8 , and orange solution was obtained, to which 412.2 mg 1.0
mmol r25 ml dry ethyl either solution was added dropwise with stirring see Eq. 9 pale yellow
powder was precipitated. The suspension was stirred for 1 h. Pale yellow powder was filtered, washed
Ž
Ž ..
Ž
.
Ž
Ž ..
Ž
.
Ž
.
with dry ethyl either and dried in vacuum. 592.4 mg DPPPr Pd OCOCF3 2 was obtained, Yield 80%.
31
Ž
.
P-NMR: d 13.44 CD3OD
Pd OCOCH q2CF3COOH Pd OCOCF q2CH3COOH
8
Ž .
Ž
.
Ž
.
3
3
2
2
Pd OCOCF 2 qDPPPr
DPPPr Pd OCOCF
9
Ž .
Ž
.
Ž
.
Ž
.
3
3
2
Ž
.
Ž
Similar procedure was used for the preparation of other six complexes as follows: DPPBu Pd OC-
31
.
Ž .
Ž
. Ž
. Ž .
.
Ž
.
Ž .
OCF3 2 b , Yield 70.9%, P-NMR: d 29.24 CDCl3 ; DPPEt Pd OCOCF3 2 c , Yield 78.7%,
P-NMR: d 61.09 CDCl3 ; DPPPr Pd p-CH3PhSO3 2 d , Yield 85%, P-NMR: d 25.58 CDCl3 ;
31
31
Ž
. Ž
Ž .
Ž
.
Ž
Ž
.
31
Ž
.
Ž
.
Ž
. Ž
.
Ž .
DPPPr Pd OCOCF3 2 e , Yield 55.6%, P-NMR: d y2.46 CDCl3 ; DPPPr PdCl2 f , Yield
2
31
76%, P-NMR: d 10.72 DMSO-d6 ; DPPPr Pd OCOCH3 2 g , Yield 81.1%, 31P-NMR: d y9.4
. Ž
.
Ž
. Ž .
Ž
.
CDCl3
4. Conclusions
Ž .
Palladium II -bisphosphine catalyzed copolymerization of CO and ethylene was studied in detail.
Ž .
Ž .
The results showed that DPPPrrPd II mole ratio, CF3COOHrPd II mole ratio and solvent all
greatly influence the catalytic activity. Solvent effect studies showed that suitable solvent not only act
Ž .
as diluent, but also had perfect stability on cationic palladium II . Furthermore, alcoholic solvent
participated the chain initiation and termination. End-group analysis, in situ NMR, in situ IR and
EXAFS studies provided many direct and valuable evidences for mechanistic studies, with which
many questions were made more clearly, and some new viewpoints were first proposed and
explained, as follows.
Ž .
1 In alcoholic solvents, chain initiation was firstly and mainly due to CO insertion into Pd–OR
bond; in nonalcoholic solvents, chain initiation was mainly due to ethylene into Pd–H bond, deduced
1
from end group analysis, in situ H-NMR and in situ IR studies.

(
)
112
H.-K. Luo et al.rJournal of Molecular Catalysis A: Chemical 151 2000 91–113
Ž .
Ž .
Ž .
2 When DPPPrrPd II mole ratio was between 1 and 2, mono-chelate ring complex 1 had
Ž .
interconversion relationship with bis-chelate ring complex 2 in the catalytic system, deduced from in
situ 31P-NMR studies and EXAFS analysis.
Ž .
Ž .
3 CO and ethylene were activated through coordination with cationic palladium II before their
1
insertions, deduced from in situ H-NMR and in situ IR studies.
Ž .
4
Ž .
In competitive coordination to vacant sites of palladium II species, ethylene exhibited
significant superiority compared with CO, deduced form in situ 31P-NMR studies.
Ž .
5 Insertion of CO was much faster than that of ethylene, deduced from in situ IR studies.
Ž .
6 Stable five-member ring intermediate was proved to be produced during copolymerization,
deduced from in situ IR studies.
Ž .
7 Solvent molecular was in competitive coordination with comonomer to vacant site of palla-
Ž .
dium II ; therefore, significantly influence the catalytic activity, deduced from solvent effect studies
and EXAFS analysis.
Ž .
Ž . Ž .
Ž .
8 For complexes a , b and c , the chelate ring was more stable, the corresponding complex
was more efficient, deduced from EXAFS analysis.
Ž .
Ž .
9 Suitable anions are of strong acid, had excellent stability on cationic palladium II , and easy to
go away to leave coordination vacant to comonomers, deduced from solvent effect studies and
EXAFS analysis.
Ž
.
10 18-electron, five coordinated pyramidal intermediates was probably produced and existed in
the system, beside 16-electron, four-coordinated square planar intermediates, deduced from EXAFS
analysis.
Acknowledgements
We are grateful to the Photon Factory for the use of BL-7C Facilities. We thank Dr. T. Tanaka and
Professor M. Nomura for experimental assistance. We also thank Kong Suling of Lanzhou University
and Zhang Hui of Beijing University for their friendly help.
References
w x
1
w x
2
E. Drent, EP0181014, 1985.
E. Drent, US4818810, 1989.
A. Sen, T.W. Lai, J. Am. Chem. Soc. 104 1982 3520.
Ž .
E. Drent, J.A.M. Van Broeknoven, M.J. Doyle, J. Organomet. Chem. 417 1991 235.
A. Batistini, G. Consiglio, Organometallics 11 1992 1766.
Ž .
A. Sen, Acc. Chem. Res. 26 1993 303.
P.K. Wong, J.A. Van Doorn, E. Drent et al., Ind. Eng. Chem. Res. 32 1993 986.
Ž .
Z. Jiang, G.M. Dahlen, K. Houseknecht, A. Sen, Macromolecules 25 1992 2999.
J.S. Brumbaugh, R.R. Whittle, M. Parvez, A. Sen, Organometallics 9 1990 1735.
w x
Ž .
3
w x
4
w x
Ž .
5
w x
6
w x
Ž .
7
w x
8
w x
Ž .
9
w
w
w
w
w
w
w
w
w
x
Ž
.
10 G.P.C.M. Dekker, C.J. Elsevier, K. Vrieze, Organometallics 11 1992 1598.
x
Ž
.
11 G.P.C.M. Dekker, A. Buijs, C.J. Elsevier et al., Organometallics 11 1992 1937.
x
Ž
.
12 G.P.C.M. Dekker, C.J. Elsevier, K. Vrieze et al., J. Organomet. Chem. 430 1992 357.
. Ž
x
Ž
.
13 F.C. Rix, M. Brookhart, P.S. White, J. Am. Chem. Soc. 118 4746 1996 .
x
Ž
.
14 F.Y. Xu, A.X. Zhao, J.C. Chien, Makromol. Chem. 194 1993 2579.
x
Ž
.
15 W. de Graaf, J. Boersma, G. Van Koten, Organometallics 9 1990 1479.
x
Ž
.
16 A. Gille, J.K. Stille, J. Am. Chem. Soc. 102 1980 4933.
x
Ž
Ž
.
.
17 F. Ozowa, T. Ito, A. Yamamoto, J. Am. Chem. Soc. 102 1980 6457.
x
18 M. Rudler-Chauvin, H. Rudler, J. Organomet. Chem. 134 1977 115.

(
)
H.-K. Luo et al.rJournal of Molecular Catalysis A: Chemical 151 2000 91–113
113
w
w
w
w
w
w
w
w
w
w
w
w
w
w
w
x
Ž
.
19 F. Ojama, T. Ito, Y. Nakama, A. Yamamoto, J. Organomet. Chem. 168 1979 375.
x
Ž
.
20 H. Kurosawa, T. Majima, N. Asada, J. Am. Chem. Soc. 102 1980 6996.
x
Ž
. Ž
.
21 V.N. Zudin, V.D. Chinakov, V.M. Nekipelov et al., J. Organomet. Chem. 289 425 1985 .
x
Ž
.
22 V.N. Zudin, V.D. Chinakov, V.M. Nekipelov et al., J. Mol. Catal. 52 1989 27.
Ž . Ž
x
.
23 D.-G. Li, C.-G. Xia, Y.-W. Sun, WuLi HuaXue XueBao 12 1 1996 71.
x
Ž
. Ž
.
24 S. Otsuta, A. Nakamura, T. Yoshida, M. Naruto, K. Actaka, J. Am. Chem. Soc. 95 3180 1973 .
x
Ž
.
25 M. Brookhart, F.C. Rix, J.M. Desimone, J. Am. Chem. Soc. 114 1992 5894.
x
26 J.A. Bakkum, A.M. Bradfor, A. Buys et al., WO 95r29946, 1995.
x
Ž
. Ž
.
27 W.L. Steffen, G.J. Palenik, Inorg. Chem. 15 10 1976 2432.
x
28 D.-G. Li, C.-G. Xia, Y.-W. Sun, Chinese Patent NO. 28312, 1989.
x
Ž
.
29 G.-S. Wu, D.-G. Li, S.-Q. Xue, YouJi HuaXue 1 1983 29.
x
Ž
.
30 Y. Kou, H.L. Wang, M. Te, T. Tanaka, M. Nornura, J. Catal. 141 1993 660.
x
Ž
.
31 J.J. Rehr, J. Mustre de Leon, S.I. Zabinsky, R.C. Albers, J. Am. Chem. Soc. 113 1991 5135.
x
Ž
.
32 J. Mustre, J.J. Rehr, S.I. Zabinsky, R.C. Albers, Phys. Rev. B 44 1991 4146.
x
Ž
.
33 F.W. Lytle, D.E. Sayers, E.A. Stern, Physica B 158 1989 701.