Exploring electronic and steric effects on the insertion and polymerization reactivity of phosphinesulfonato pdii catalysts

FULL PAPER

DOI: 10.1002/chem.201301365

Exploring Electronic and Steric Effects on the Insertion and Polymerization

Reactivity of Phosphinesulfonato Pd^IICatalysts

Boris Neuwald,^[a]Laura Falivene,^[b]Lucia Caporaso,*^[b]Luigi Cavallo,^[c]and

Stefan Mecking*^[a]

Abstract: Thirteen different symmetric

and asymmetric phosphinesulfonato

palladium complexes ([{(^X1-Cl)-m-M}_n],

and copolymerization of ethylene and

MA regarding catalyst activity and sta-

bility over time, polymer molecular

weight, and polar co-monomer incor-

poration were investigated. DFT calcu-

lations were performed on the main in-

sertion steps for both monomers to ra-

tionalize the effect of the ligand substi-

tution patterns on the polymerization

behaviors of the complexes. Full analy-

sis of the data revealed that: 1) elec-

tron-deficient catalysts polymerize with

higher activity, but fast deactivation is

also observed; 2) the double ortho-sub-

stituted catalysts ^(MeO)21 and ^(MeO)31

allow very high degrees of MA incor-

poration at low MA concentrations in

the copolymerization; and 3) steric

shielding leads to a pronounced in-

crease in polymer molecular weight in

the copolymerization. The catalyst

properties induced by a given P-aryl

(alkyl) moiety were combined effec-

tively in catalysts with two different

non-chelating aryl moieties, such as

^cHexO/(MeO)21, which led to copolymers

with significantly increased molecular

weights compared to the prototypical

^MeO1.

M=Na, Li, 1= ACHTUNGTRENNUNG

^X(P^O)PdMe) were

prepared (see Figure 1). The solid-state

structures of the corresponding pyri-

dine or lutidine complexes were deter-

mined

for

^(MeO)21-py,

^(iPrO)21-lut,

^(MeO,Me2)1-lut, ^(MeO)31-lut, ^CF31-lut, and

^Ph1-lut. The reactivities of the catalysts

^X1, obtained after chloride abstraction

with AgBF₄, toward methyl acrylate

(MA) were quantified through deter-

mination of the rate constants for the

first and the consecutive MA insertion

and the analysis of b-H and other de-

composition products through NMR

spectroscopy. Differences in the homo-

Keywords: catalysis · coordination

modes · density-functional calcula-

tions · kinetics · reaction mecha-

nisms · theoretical chemistry

Introduction

tion with ethylene was achieved for the first time with cat-

ionic Pd^IIdiimine complexes. In this case, highly branched

copolymers are formed, which consist of ethylene as the

major component (ꢀ75 mol%) and contain acrylate units

situated preferentially at the ends of the branches.^[1,2]By

contrast, neutral arylphosphinesulfonato Pd^IIcomplexes co-

polymerize methyl acrylate (MA) and ethylene to linear

random copolymers.^[3]This catalyst system has recently at-

tracted much attention owing to the variety of polar mono-

mers that can be copolymerized with ethylene, including

acrylonitrile,^[4,5]vinyl acetate,^[6]and acrylic acid.^[7,8]Phos-

The catalytic insertion polymerization of ethylene and pro-

pylene is one of the most well-studied chemical reactions. In

terms of applications, it is employed for the production of

more than 70 million tons of polyolefins annually. However,

an insertion polymerization of polar-substituted vinyl mono-

mers had remained elusive for a long time. Copolymeriza-

[a] Dr. B. Neuwald, Prof. Dr. S. Mecking

Chair of Chemical Materials Science, Department of Chemistry

University of Konstanz

AHCTUNGTRENNUNG

phinesulfonato Pd^IImethyl complexes are well-defined cata-

lyst precursors that can provide a convenient entry into the

catalytic cycle.^[9]

78464 Konstanz (Germany)

Fax : (+49)7531-88-5152

E-mail: Stefan.Mecking@uni-konstanz.de

For such complexes as (P^O)PdMe(L) (^MeO1-L; P^O=k²-

[b] Dr. L. Falivene, Dr. L. Caporaso

Department of Chemistry and Biology

University of Salerno

P,O-(2-MeOC₆H₄)₂PACTHNUGTRNEUNG(C₆H₄SO₂O), Figure 1), the coordinat-

ing monodentate ligand L plays an important role for the

polymerization activity at low ethylene pressures. Com-

plexes exhibiting a weakly coordinating DMSO ligand al-

lowed for the synthesis of copolymers with high polar mono-

mer incorporation of up to 50%, and even homo-oligomeri-

zation of methyl acrylate, owing to their substantial activity

also at low monomer concentrations.^[10]A detailed study ex-

ploring the perspectives and limitations of such weakly coor-

dinated precursors revealed that the in situ activation of

[{(^MeO1-Cl)-m-Na}₂] with AgBF₄to species free of any addi-

Via Giovanni Paolo II, 132 - 84084 - Fisciano (SA) (Italy)

Fax : (+39)089-969603

E-mail: lcaporaso@unisa.it

[c] Prof. Dr. L. Cavallo

Physical Sciences and Engineering

Kaust Catalysis Center

King Abdullah University of Science and Technology (KAUST)

Thuwal 23955-6900 (Saudi Arabia)

Supporting information for this article is available on the WWW

under http://dx.doi.org/10.1002/chem.201301365.

Chem. Eur. J. 2013, 19, 17773 – 17788

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

17773

Furthermore, the increased steric bulk has been suggested

to destabilize the alkyl olefin complex, and thereby enhance

polymerization rates.^[20]The replacement of the ortho MeO

À

with an MeS group reduces activity drastically, but appears

to lead to the formation of polymers with high molecular

weights (M_w=5ꢁ10⁵gmol^À1, M_w/M_n=6). However, in this

case, the broad molecular weight distribution indicates the

presence of various active sites.^[17]In relation to these find-

ings, Jordan et al. reported the formation of high-molecular-

weight linear PE with a very broad molecular weight distri-

bution (M_w=10⁷gmol^À1, M_w/M_n=60) with phosphinesulfo-

nato complexes bearing ligands with two aryl sulfonate moi-

eties, leading to a tetranuclear aggregate of the catalyst.

This study revealed that PE with these characteristics is only

formed if the aggregate stays intact during polymeri-

zation,^[21]similarly to the dramatic effect of the reaction sol-

vent on PE molecular weight in polymerization of the Ni^II

analogues.^[22]Replacement of the non-chelating aryl moiet-

ies with alkyl groups (tert-butyl groups in particular) increas-

es the activity drastically in ethylene polymerization, but si-

multaneously reduces the temperature stability of the cata-

lyst and decreases MA incorporation in copolymerization

experiments.^[14]A conclusive comparison of these studies is

hampered for two reasons. First, the comparability is re-

stricted because of the different polymerization conditions

and catalyst precursors used. Furthermore, the separation of

electronic and steric influences remains difficult in most

cases.

For a deeper understanding of the relations between the

catalyst structure and the (co)polymer obtained, a broad va-

riety of (P^O)PdMe catalysts were synthesized and studied

in detail. Their reactivities toward (polar) olefins were

screened under identical conditions in stoichiometric inser-

tion studies and polymerizations, respectively, to obtain in-

sights into steric and electronic influences, and to identify

the catalysts most suitable for certain targets, for example,

the production of copolymers with higher molecular weights.

Furthermore, we show that a more directed catalyst design

becomes possible through combination of two different non-

chelating (aryl) substituents at phosphorus, so that certain

properties of two symmetric catalysts can be combined.

Figure 1. Phosphinesulfonato substitution patterns studied.

tional coordinating monodentate ligand is an easy and ac-

cessible way to very active catalysts.^[11]

However, these studies also revealed that the copolymeri-

zation activity is intrinsically limited by the formation of

k-O coordinated chelate complexes [(P^O)Pd{k²-C,O-

CH(R)CH₂CH(C(O)OMe)CH₂CH₃}] in the absence of

strongly coordinating monodentate ligands. The coordina-

tion strength of such chelates, which are inevitably formed

by (co)monomer insertion after an MA insertion, has been

shown to prevail over p-coordination of ethylene by a factor

of several hundred.^[12,13]Another important intrinsic limita-

tion of this catalytic system concerns the low molecular

weights of the copolymers produced. These limitations

cannot be overcome through variation of the monodentate

ligand L in ^MeO1-L, because this ligand is not coordinated in

the active species and does not affect the polymer micro-

structure. Accordingly, the catalyst design has to be adjusted

through steric and electronic variation of the phosphinesul-

fonato ligand.

The influence of the non-chelating aryl substituents at

phosphorus on the activity of (P^O)PdMe catalysts in poly-

merization reactions has been addressed by Claverie,^[14,15]

Goodall,^[16]and Rieger et al.^[17]The substitution of the pro-

totypical anisyl moiety for bulky naphthyl, phenanthryl, an-

thracenyl, or 1-methoxynaphthalene moieties in general ap-

pears to reduce polymerization activity, molecular weight,

and polar monomer incorporation.^[15,17]For catalysts with li-

Results and Discussion

Nomenclature: The designation of species discussed in this

work is depicted in Scheme 1. (P^O)PdMe fragments (1)

generated from a catalyst precursor react with a polar mo-

nomer to form the mono-insertion products (2), and subse-

quently, the consecutive (double) insertion products (3). De-

gands bearing very bulky 2-{2’,6’ACHTUNRGTNEUNG(MeO)₂C₆H₃}C₆H₄moieties,

polar monomer incorporation is also diminished, but the

molecular weights and catalyst activities are enhanced sig-

nificantly in ethylene homopolymerization.^[10,18]This may be

related to a favorable steric bulk in the axial position of the

palladium center in this case. For the related cationic Pd^II

diimine complexes, the blocking of the axial positions has

been found to be essential for the inhibition of chain trans-

fer and the production of high-molecular-weight polymers.^[19]

À

composition of 2 and 3 leads to intermediate Pd H species

that can insert further polar monomers to yield mono- (4)

and consecutive insertion products (5). The phosphinesulfo-

nato ligand of the corresponding complex is denoted by a su-

perscript acronym (^X1) as depicted in Figure 1.

Synthesis: Symmetric and asymmetric complexes were ob-

17774

www.chemeurj.org

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2013, 19, 17773 – 17788

Phosphinesulfonato Pd^IICatalysts

FULL PAPER

bulk and electronic properties, were studied, including the

newly synthesized catalysts ^(Me2;MeO)1, ^cHexO1, and ^Me*/(MeO3)1

(Figure 1). For the synthetic procedures and full characteri-

zation of these compounds see the Supporting Information.

Commonly, the [(P^O)PdMe] fragment was generated

in situ through halide abstraction from the cation-bridged

complexes [{(^X1-Cl)-m-M}_n] for insertion and polymerization

studies.

All the protonated ligands studied exist in a zwitterionic

form in CD₂Cl₂, with the proton situated at phosphorus,

À

leading to a P H coupling of approximately 600 Hz in the

³¹P NMR spectra. The ³¹P NMR shifts are situated in the

range 4 to À31 ppm for the zwitterionic phosphinesulfonates

(P^O)H, and between À12 and À49 ppm for the phosphine-

sulfonate salts (P^O)M (M=Na, Li; see Supporting Infor-

mation, Table S1). The ³¹P NMR shift is not an exact mea-

sure of the ligand basicity here, because electronic influen-

ces are disturbed by steric interaction, but in principle,

³¹P NMR resonances are low-field-shifted with increasing

electron deficiency.^[26]Accordingly,

^Ar(P^O)H can be considered to be relatively electron-defi-

cient compared to electron-rich ligands such as

^(MeO)2(P^O)H, ^(iPrO)2(P^O)H, and ^(MeO)3(P^O)H.

The complexes [{(^X1-Cl)-m-M}_n] form multinuclear species

^CF3(P^O)H and

ACHTUNGTRENNUNG

Scheme 1. Labeling logic of complexes.

AHCTUNGTRENNUNG

G

E

N

ACHTUNGTRENNUNG

that are bridged by intermolecular coordination of the chlor-

ine atom and various oxygen atoms of the sulfonate group

to the sodium or lithium counterion (M). For [{(^MeO1-Cl)-m-

Na}₂], X-ray structure analysis revealed a dimeric structure

in the solid state.^[27]As the stability of these compounds in

solution is limited and partial loss of MCl salts is observed,

crystals of the corresponding pyridine and lutidine com-

plexes were grown. These complexes are easily synthesized

in situ through reaction of [{(^X1-Cl)-m-M}_n] with the corre-

sponding base and evaporation of the volatiles. The crystal

structures of ^(MeO)21-py, ^(MeO)31-lut, ^(iPrO)21-lut, ^(MeO;Me2)1-lut,

^CF31-lut, and ^Ph1-lut were determined. They all show

a square-planar geometry around the palladium center, with

the methyl group trans to the sulfonate group, as is generally

observed for (P^O)PdMe complexes (Figure 2 and Support-

ing Information).

Scheme 2. Synthetic pathways to symmetric and asymmetric (P^O)PdMe

catalyst precursors.

tained through multiple synthetic routes (Scheme 2).^[16,23–25]

Asymmetric substitution at phosphorus could be achieved

through successive reaction of Cl₂PNEt₂with different lithi-

ated aryl or alkyl compounds (red route). This route allows

the precise synthesis of asymmetric ligands, whereas all aryl

substituents of the corresponding symmetric ligands can be

combined. Complexes of the type [{(^X1-Cl)-m-M}_n] can be

synthesized through the reaction of [(cod)PdMe(Cl)] with

the lithium salts ACHTUNGTRENNUNG

^X(P^O)Li obtained directly from the ligand

synthesis (green route). Alternatively, the complexes can be

synthesized by reacting [(cod)PdMe(Cl)] with the sodium

salts (P^O)Na, which are obtained after acidic work-up to

the protonated ligands and subsequent deprotonation with

NaH (orange route).

Figure 2. Solid-state structure of ^Ph1-lut. Ellipsoids represent 50% proba-

bility. Hydrogen atoms and solvent molecules are omitted for clarity.

Solid-state structures of ^(MeO)21-py, ^(iPrO)21-lut, ^(MeO;Me2)1-lut, ^(MeO)31-lut, and

^CF31-lut can be found in the Supporting Information.

For a profound analysis, ten symmetric and three asym-

metric substitution patterns at phosphorus, differing in steric

Chem. Eur. J. 2013, 19, 17773 – 17788

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

17775

L. Caporaso, S. Mecking et al.

ligand salts (P^O)M (Supporting Information, Table S1).

However,

the

cyclohexyloxy-substituted

compound

À

^cHexO/(MeO)21-lut exhibits an unusually short Pd N bond

(2.120(2) ꢂ)^[28]. For the pyridine complexes, a pronounced

À

elongation of the Pd N bond is observed for the 2,6-

(MeO)₂C₆H₃-substituted compounds ^Ar1-py and ^Ar/(MeO)21-py

compared with those of ^MeO1-py and ^(MeO)21-py (Supporting

Information, Tables S1 and S2). This may be related to the

high steric demand of the aryl substituent.

In conclusion, an exact determination of the electronic

properties of the investigated complexes through experi-

mental methods remains difficult, and the data obtained

allow only a rough classification.

Figure 3. Exo₂and exo₃configurations for (P^O)PdMe complexes.

The aryl moieties of (P^O)PdMe complexes can adopt

either an exo₂or exo₃configuration. In the exo₃configura-

tion, all equal ortho substituents are situated on the same

side as the central phosphorus above a plane defined by the

three aryl carbons bound to phosphorus. In the various exo₂

configurations, one of the ortho substituents is situated un-

derneath this plane (Figure 3). ^CF31-lut and ^Ph1-lut show an

exo₂configuration, which is observed for most of the

(P^O)PdMe structures determined.^[28]Interestingly, for

^Ar1-py, which is structurally closely related to ^Ph1-lut, an exo₃

configuration was observed, so the additional methoxy sub-

stituents in ^Ar1 appear to have a distinct structural influ-

ence.^[18]For ^(MeO;Me2)1-lut, an exo₃-like configuration is ob-

served, with the ortho methyl group always situated on the

same side as phosphorus. Because the adoption of an exo₃

configuration is associated with a decrease in steric interac-

tion between the corresponding ortho substituents, this indi-

Polar monomer insertion: For the evaluation of the reactivi-

ty of the complexes toward polar olefins, the insertion of

MA into the Pd Me bond of ^X1, generated in situ by chlo-

ride abstraction from [{(1-Cl)-m-M}_n], to 2_MAand subse-

À

X

quently to ^X3_MA, as well as the elimination and reinsertion

reactions at 258C, were monitored through ¹H NMR over

a period of 12–18 h (Figures 4 and 6, cf. Supporting Informa-

tion Figures S22–S34). The insertion products ^X2, ^X3, ^X4, and

^X5 (Scheme 1, for example, Figure 4) were identified

through the assignment of their spin systems by

1

¹H¹H gCOSY and H¹H zTOCSY experiments, as well as by

¹H¹³C gHMBC experiments if possible (cf. Supporting Infor-

mation, Figures S6–S21). All organic elimination products

1

were identified through comparison with the H NMR shifts

of the corresponding isolated substances (Figure 5, cf. Sup-

porting Information).

cates increased steric demand of the

A

For ^MeO1, ^(MeO)21, ^(MeO)31, ^cHexO1, and ^cHexO/(MeO)21, almost no

b-H elimination or other decomposition reactions of the in-

sertion products formed were observed during the experi-

framework. Whereas for ^(MeO)21-py and ^(iPrO)21-lut a differen-

tiation of exo configurations is not possible because of sym-

metry, the increased steric demand of ^(iPrO)21-lut is reflected

in the deformation of the boat configuration of the 6-mem-

bered (P^O)Pd chelate, which is observed in all other solid-

state structures, to an envelope-like ring configuration.^[28]

A

comparison of NMR data of the various complexes (Sup-

porting Information, Table S1) reveals that easily accessible

experimental parameters such as ¹H or ³¹P NMR shifts

cannot reflect the electronic properties satisfactorily because

they are sensitive to steric parameters as well as to the exact

measurement conditions.

À

Comparison of the Pd N bond lengths (Supporting Infor-

mation, Tables S1 and S2) determined by X-ray structure

analysis of pyridine and lutidine complexes revealed that

the distance can be correlated partly to the electron density

at palladium (as anticipated from the electron-donating or

-withdrawing nature of the substituents on the P-bound aryl

of the phosphinesulfonato ligand, vide supra) if the same

probe (lutidine/pyridine complex) is applied. However,

À

steric effects can still have a pronounced influence. The Pd

N bond length increases with increasing electron density at

the palladium center: ^CF31 (2.118(3) ꢂ) < ^Ph1 (2.128(3) ꢂ) <

^MeO1 (2.134(2) ꢂ)^[4]

< <

^(MeO;Me2)1 (2.138(2) ꢂ) ^(iPrO)21

Figure 4. ¹H NMR (400 MHz, CD₂Cl₂) monitoring of the reaction of ^H1

with MA to the insertion products ^H2 and subsequently to ^H3, and rein-

(2.146(2) ꢂ) ꢁ ^(MeO)31 (2.146(2) ꢂ). The correlation found is

in line with the observed trend in ³¹P NMR shifts of the

sertion into the Pd H to ^H4 at 258C.

À

17776

www.chemeurj.org

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2013, 19, 17773 – 17788

Phosphinesulfonato Pd^IICatalysts

FULL PAPER

whereas for ^CF31, ^(MeO)21, ^cHexO1, and ^Me*/(MeO)31 only trace

amounts of C were detected (cf. Supporting Information,

Figures S22–S34). The formation of C can be explained by

a 1,2-reinsertion of A into a palladium hydride and subse-

quent b-H elimination, by a closely related series of b-H

eliminations and reinsertions at a given metal center starting

X

from 2_MAÀ2,1(chain walking), or by isomerization of A

caused by formed Pd black. Considering the huge excess of

MA present in comparison to the sterically more demand-

ing, and thus, less reactive A, reinsertion in significant

amounts after separation of the substrate from the metal

center is rather implausible. Furthermore, the exclusive for-

mation of C by isomerization appears unlikely, because it is

formed simultaneously with A, and is thermodynamically

disfavored (Figure 6). The portion of C appears to increase

for sterically demanding compounds such as ^Ph1. This might

indicate a chain-walking mechanism at a given metal center,

as chain transfer is presumed to be hindered for sterically

crowded metal centers.

Figure 5. Observed decomposition products.

ment. In general, b-H elimination was found to be the pre-

dominant decomposition pathway for the insertion products

X

^X2_MAÀ2,1and 3_MA(^X5_MA). However, the intermediately

The 1,2-insertion products 2_MAÀ1,2are stable toward b-H

formed palladium hydride species could not be detected. We

assume that the palladium hydrides either reinsert methyl

acrylate instantly to form ^X4_MA, or decompose to Pd⁰. It has

been reported that strong donor ligands such as phosphines

are necessary for the stabilization of isolable (P^O)PdH

species.^[9]However, the addition of phosphines reduces any

reactivity drastically. The main b-H-elimination product

elimination under the reaction conditions, and only for ^Ph1

was the formation of trace amounts of methyl methacrylate

(D) as an elimination product observed (Figure 6). This high

stability can be related to the formation of five-membered

k-O coordinated chelate complexes (Scheme 1). However,

(iPrO)2

for

2

,

^(MeO;Me2)2_MAÀ1,2, and ^Ar2_MAÀ1,2, nonetheless

MAÀ1,2

significant decomposition and simultaneous formation of the

saturated ester G was detected (cf. Supporting Information,

Figures S23, S31, and S32). This might be caused by a net re-

ductive proton transfer: recent investigations on the decom-

X

from 2_MAÀ2,1is trans-methyl crotonate (A, Figure 5).

However, in some cases cis-methyl crotonate (B) can also

be observed in small amounts (A:B > 10:1, cf. Supporting

Information, Figures S22–S34). For ^(iPrO)21, ^(MeO;Me2)1, ^Ar1, and

^Ph1, the formation of a significant amount of methyl but-3-

enate (C) was observed (up to A:C ꢁ 3:1 for ^Ph1, Figure 6),

position reactions of phosphinesulfonato compounds re-

0

À

vealed that (P^O)Pd H species can decompose to Pd and

the protonated ligand (P^O)H through reductive elimina-

tion. It has been shown that (P^O)H can induce protolytic

À

cleavage of Pd alkyl bonds. Alternatively, a direct H-trans-

fer by palladium hydride species in a multinuclear arrange-

ment is conceivable.^[9,29]For (N^N)Pd alkyl complexes, ho-

À

molytic cleavage of Pd alkyl bonds has also been ob-

served.^[30,31]However, in the latter case, a disproportionation

reaction from the primarily formed free radical should lead

to a 1:1 mixture of D and G, which was not observed for

^(iPrO)21, ^(MeO;Me2)1, and ^Ar1.^[32]As expected, no organic prod-

ucts indicating a recombination of free radical species,

which is even less probable for steric reasons, were detected.

Minor amounts of methyl butyrate (E) were observed in

some cases (e.g., for ^Ar1, cf. Supporting Information, Fig-

ure S32). This reveals that a similar decomposition pathway

X

to that assumed for 2_MAÀ1,2also exists for ^X2_MAÀ2,1. Howev-

er, as the portion E is small compared with the mainly

formed A, a homolytic bond cleavage cannot be excluded in

this case.

In comparison, the products of a consecutive insertion

^X3_MAare generally very stable owing to the formation of six-

membered k-O coordinated chelates (Scheme 1), the only

exception being ^CF33, which eliminates H rather rapidly (cf.

Supporting Information, Figure S29).

Figure 6. ¹H NMR (400 MHz, CD₂Cl₂) monitoring of the reaction of ^Ph1

with MA, of subsequent elimination reactions of the insertion products

À

formed, and of the reinsertion into Pd H species at 258C.

Chem. Eur. J. 2013, 19, 17773 – 17788

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

17777

L. Caporaso, S. Mecking et al.

Palladium hydride species resulting from the b-H elimina-

tion reactions can reinsert MA to form ^X4 and subsequently

^X5 (Figure 4, cf. Supporting Information, Figure S29). An as-

signment of these species is difficult and subject to several

10^À3s^À1vs. k_MeOÀ1st=3ꢁ10^À3). Considering the compounds

having double ortho-substituted aryl moieties, the highest

rate constant was found for ^(MeO)31, which can be considered

as more electron-rich than the sterically similar ^(MeO)21

(k_(MeO)3=26ꢁ10^À3s^À1vs. k_(MeO)2=1ꢁ10^À3s^À1, Table 1). The

substitution of one 2,4,6-(MeO)₃C₆H₂substituent by

a methyl group in ^Me*/(MeO)31 leads to a strong reduction of

À

preconditions: pronounced b-H elimination to form Pd H

species, reasonable reactivity of the palladium hydrides

toward insertion compared to decomposition, and sufficient

stability of the newly formed insertion products, or respec-

tively, a distinct decomposition to specific products, are re-

quired. For this reason, so far, only the 1,2-insertion product

the rate constant by

a

factor of more than 200

AHCTUNGTRENNUNG

effect of a cyclohexyloxy group, the substitution of the me-

thoxy groups in ^(MeO)21 by more sterically demanding isopro-

poxy groups in ^(iPrO)21 has no negative effect, and the rate

constant is even enhanced (k_(iPrO)2À1st=3ꢁ10^À3s^À1vs.

H

CF3

^X4_MAÀ1,2could be assigned as for 4_MAÀ1,2and

4

MAÀ1,2

(Figure 4, cf. Supporting Information, Figure S29). The cor-

X

responding 2,1-insertion products 4_MAÀ2,1probably undergo

rapid b-H elimination to form MA, or in the case of

k

_(MeO)2À1st=1ꢁ10^À3). The rate constants for the consecutive

^CF34_MAÀ2,1, react to ^CF3

5

_MA, which can be detected by the b-H-

insertion are one to two orders of magnitude smaller than

those for the first insertion. This can be explained by the in-

creased steric bulk at the palladium, going along with the in-

sertion into an a-substituted carbyl instead of an n-alkyl

species, as well as the opening of the four-membered chelate

formed after the first insertion. In general, the steric

demand of the ligand has a much more pronounced influ-

ence on the second insertion, so that no consecutive inser-

tion at all could be observed for the very bulky systems ^Ar1,

^Ph1, and ^HexO1.^[28]The highest rate constants were found for

^CF31 and ^MeO31, and in general, the relation between the in-

sertion rates is similar to that discussed for the first inser-

tion.

elimination product formed, dimethyl 4-methylpent-2-ene-

dioate (I, cf. Supporting Information, Figure S29).

For a quantitative analysis of the influences of electronic

and steric properties on the reactivity of the palladium

methyl complexes toward polar olefins, the rate constants

Table 1. Rate constants of insertion and b-H elimination as well as regio-

selectivity for catalysts ^X1.^[a]

compound

k_1st

k_2nd

k_bH-1st

Regioselectivity

(2,1):k(1,2)

[ꢁ10^À3s^À1

]

[ꢁ10^À5s^À1

]

[ꢁ10^À6s^À1

]

k

N

ACHTUNGTRENNUNG

^CF31

4.8

0.1

3.2

43

0.3

9.2

100

5

4:1

5:1

>15:1

n.d.

2:1

1:1

1.6:1

6:1

>30:1

7:1

^H1

^MeO1

10

X

The rate constants for the b-H-elimination from 2_MAÀ2,1

^cHexO1

0.2

n.o.^[b]

6

ꢁ10^[c]

40

are in the range 10^À5–10^À6s^À1, and are significantly lower

^Ph1

0.9

than the rate constants for the first insertion. Larger b-H-

^Ar1

ꢁ1–5^[c]

n.d.^[d]

<4

elimination rate constants are observed for ^CF31 (k_CF3ÀbH

=

^Ar/(MeO)21

^cHexO/(MeO)21

^(MeO)21

^(iPrO)21

^(MeO)31

^Me*/(MeO)31

^(MeO;Me2)1

4

3

10^À4s^À1) and ^(MeO;Me2)1 (k_{(MeO;Me2)ÀbH}=3ꢁ10^À3s^À1). Consider-

ing the formation of ^X3, b-H elimination competes signifi-

cantly with consecutive insertion for ^CF31, ^H1, ^cHexO1, ^Ph1, ^Ar1,

^(iPrO)21, ^Me*/(MeO)31, and ^(MeO;Me2)1.

5.3

1.3

3.4

26

3.7

ꢁ1–5^[c]

5

38

n.o.^[e]

ꢁ5

>50:1

8:1

0.09

<0.1

traces

n.d.^[c]

3000

Steric influences emerge clearly in the regioselectivity, as

determined from the distribution of the 1,2- and 2,1-inser-

tion products ^X2_MA. An increase in steric constraints around

the palladium center leads to a shift of the 2,1:1,2 ratio from

>15:1 for ^MeO1, ^(MeO)21, ^(MeO)31, ^Me*/(MeO)31 to 7–6:1 for ^(iPrO)21,

^cHexO/(MeO)21, and to 2–1:1 for ^Ar1, ^Ar/(MeO)21. This can be relat-

ed to a steric interaction of the ligand with the p-bound MA

monomer leading to increased barriers for the electronically

favored 2,1-insertion mode.^[33]However, for the rather un-

constrained ^H1, a rather low selectivity of 5:1 is observed.

In summary, each catalyst exhibits an individual reaction

profile in the presence of MA, which is determined by the

relation of insertion and elimination rate constants, the rela-

tion of the insertion modes, and the relation of different de-

composition reactions. The resulting reactivity can hardly be

correlated to electronic properties. Additional effects caused

by coordination of the solvent (CD₂Cl₂) and the resulting

monomer/solvent equilibrium may also complicate the de-

tection of clear trends. However, the influence of steric bulk

is more explicit. An increase in steric bulk decreases the in-

sertion rates (with the only exception being ^(iPrO)21), disfa-

vors a 2,1-insertion mode, and enhances chain walking.

[a] Conditions: [Pd]=0.02 molL^À1, 1.1 equiv AgBF₄, Pd/MA=1:15, sol-

vent CD₂Cl₂, T=258C. [b] No consecutive insertion observed. [c] Exact

determination not possible owing to overlapping or broad resonances.

[d] No b-H elimination rate determined owing to overlapping resonances.

[e] No b-H elimination observed.

for the first and consecutive MA insertions were determined

1

through H NMR spectroscopy at 258C ([Pd]=0.02 molL^À1,

Pd:MA=1:15, Table 1). Note that the [{(^X1-Cl)-m-M}_n] pre-

cursors can contain up to one equivalent of weakly coordi-

nating solvent, for example, acetone, per formula unit, and

that this might affect the determined rate constants. A com-

parison of the insertion rates for the first insertion does not

reveal a clear trend. The rather electron-poor ^CF31 exhibits

a larger rate constant than the more electron-rich ^MeO1

(k_CF3À1st=5ꢁ10^À3s^À1vs. k_MeOÀ1st=3ꢁ10^À3s^À1). However, for

^H1 bearing no substituents at the aryl moieties, and which is

expected to have an intermediate electron density, the rate

constant is much lower: k_HÀ1st=0.1ꢁ10^À3s^À1. An increase in

steric bulk through the replacement of methyl by cyclohexyl

in ^cHexO1 leads to a smaller rate constant (k_cHexOÀ1st=0.2ꢁ

17778

www.chemeurj.org

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2013, 19, 17773 – 17788

Phosphinesulfonato Pd^IICatalysts

FULL PAPER

Single catalysts exhibit outstanding reactivities: the largest

insertion rates by far were found for ^CF31 and ^(MeO)31, espe-

cially for the second insertion. By contrast, the asymmetric

substituted compound ^Me*/(MeO)31 inserts MA most slowly,

with a rate constant around 200 times smaller than that for

^(MeO)31 for the first insertion. Concerning the stability of the

insertion products, exceptionally high b-H-elimination rate

significantly during a polymerization time of 30 min are

compared to the stable catalyst ^MeO1. A pronounced decom-

position was observed especially for those compounds

having a ligand with an aryl substituent in the ortho position

of the non-chelating aryl moieties (^Ar1, ^Ar/(MeO)21, ^Ph1) and for

the ortho-CF₃-substituted compound ^CF31 as well as for

^(MeO;Me2)1, for which no significant mass flow activity was ob-

served after 5 min. A fast decomposition in the ethylene

polymerization correlates with the high b-H-elimination rate

constants found for ^Ph1, ^(MeO;Me2)1, and ^CF31 and the observed

multifold decomposition reactions for the insertion products

of ^Ar2 and ^Ar/(MeO)22 to various saturated and unsaturated or-

ganic products. For ^(MeO;Me2)1, only very small amounts of PE

could be isolated, and only very low mass flow activity was

detected within the first five minutes. This indicates low ac-

tivity and rapid decomposition of the catalysts. Furthermore,

analysis of the formed polymers through GPC revealed a bi-

modal molecular weight distribution, with the main fraction

CF3

(MeO;Me2)

constants were observed for

2

and

2

.

MAÀ2,1

Here, the detected respective elimination products imply

that the resulting palladium hydride species have opposite

À

reactivities. Whereas

^CF3A

species ^CF35_MA, no further insertions, and hence, only decom-

position to Pd⁰was detected for

^(MeO;Me2)A

Homopolymerization of ethylene: All the catalysts were

studied in the homopolymerization of ethylene at an ethyl-

ene pressure of 5 bar in toluene at 808C (Supporting Infor-

mation, Table S27). The active catalysts were generated

in situ through chloride abstraction from [{(1-Cl)-m-M}_n]

with AgBF₄in dichloromethane. The resulting insoluble

salts were removed by filtration before addition of the cata-

lyst solution to the pressure reactor. A comparison of the

differently coordinated precatalysts ^MeO1-dmso, ^MeO1-Cl

([{(1-Cl)-m-M}_n] without chloride abstraction), and ^MeO1 re-

vealed that ^MeO1-dmso and ^MeO1 show comparable catalytic

activities (Supporting Information, Table S27, entries 27-

1 and 27-2). This is in line with previous observations that at

an ethylene pressure of 5 bar the DMSO/ethylene pre-equi-

librium is already shifted completely to ethylene coordina-

tion.^[10]In contrast, ^MeO1-Cl in the absence of AgBF₄shows

a significantly reduced turn-over frequency (TOF). This con-

firms that activation with AgBF₄is necessary for full catalyst

activity to be reached. However, the complexes [{(1-Cl)-m-

M}_n] still lead to active catalysts, and can be valuable precur-

sors if the application of silver salts has to be avoided.

being low-molecular-weight PE (ꢁ1000 gmol^À1) and

a

small amount of

a

high-molecular-weight polymer

(>30000 gmol^À1). This indicates the existence of several

active sites during the short catalyst lifetime. A tentative ex-

planation may be an alteration of the nature of the active

sites, for example, through exchange of the aryl substituents

of the ligand owing to elimination/oxidative addition, as

demonstrated recently.^[9]However, there is no obvious

reason for the pronounced occurrence of such a transforma-

tion for this catalyst. Compared with the stable catalyst

^(MeO)21 with a 2,6-methoxy substitution pattern, the electron-

ic properties for a 3,6-methyl-2-methoxy substitution should

not be exceptionally different. This might point to steric in-

fluences caused by the additional 3-methyl substituent. In

the crystal structure of ^(MeO;Me2)1-lut, it has been observed

that the arrangement of the non-chelating aryl substituents

differs from the usually observed exo₂arrangement, which

may also indicate a specific influence of the 3-alkyl substitu-

ent. Another difference in the complex conformation be-

tween ^(MeO;Me2)1 and ^(MeO)21, as observed in the solid state, is

the alignment of the methyl instead of methoxy groups

toward the palladium center in ^(MeO;Me2)1-lut. Clearly, small

changes in the substitution pattern can have a strong influ-

ence on the catalyst activity and stability.

Comparison of the activities of different catalysts is com-

plicated by the different stabilities of the catalysts over time

under polymerization conditions. Hence, the ethylene mass

flow was monitored to evaluate catalyst stability. In

Figure 7, the mass flow traces of catalysts that decompose

For the evaluation of catalyst activities and elucidation of

the actual influence of decomposition on the polymer yield,

the average activities for ethylene polymerization after poly-

merization times of 100, 200, and 300 s (as determined from

mass flow integration^[36]) and the TOF after a polymerization

time of 30 min (as determined from the isolated polymer

yield) are compared in Figure 8. For rather stable catalysts

such as ^MeO1, the activity is constant in all intervals deter-

mined, whereas for catalysts showing strong decomposition,

a pronounced decrease in activity is already evident within

the first 300 s, as, for example, for ^CF31. However, for very

low activities, the accuracy of this method is limited, as seen

in the significantly higher activities determined from the

polymer yield than the mass flow activities for the stable

catalysts ^(MeO)31 and ^(Me*/MeO3)1. In these cases, the activity de-

Figure 7. Mass-flow traces (fitted; see ref. [34]) for ethylene homopoly-

merization (p=5 bar, T=808C) with ^MeO1, ^CF31, ^Ar1, ^Ar/(MeO)21, ^Ph1, and

^(iPrO)21, showing catalyst decomposition over time (see ref. [35]).

Chem. Eur. J. 2013, 19, 17773 – 17788

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

17779

L. Caporaso, S. Mecking et al.

exhibit the highest MA insertion rates of all the catalysts,

but in contrast to ^CF31, only a moderate activity for ethylene

homopolymerization is found for ^(MeO)31. This might be

caused by a comparatively lower reactivity of ethylene for

binding and insertion for electronic reasons. In addition, a fa-

vorable pronounced preference for p-binding of MA versus

k-x binding with the ester group for the electron-rich com-

plexes might contribute to this result.

The comprehensive picture revealed by this study agrees

with previous single observations: Claverie et al. reported

a drop in activity from ^Ar1-lut over ^MeO1-py to ^H1-py. Howev-

er, in this case, catalysts exhibiting electron-deficient ligands

bearing 3,5-CF₃substituted or perfluorinated aryl moieties

showed further decreased activities.^[14]Rieger et al. observed

a drastic decrease in catalytic activity upon replacement of

the MeO groups in ^MeO1 with more electron-donating (but

also potentially coordinating) MeS groups.^[17]Jordan et al.

showed that enhancing the electron deficiency of the Pd

Figure 8. Average activity of various catalysts in ethylene homopolymeri-

zation for the first 100, 200, and 300 s of the polymerization as deter-

mined from mass flow, and TOF after 30 min as determined from isolated

PE yield.

center by binding of BACHTUNRGTNE(NUG C₆F₅)₃to the sulfonate group of

(P^O)PdMe compounds increases the activity and chain-

transfer rate.^[37]Regarding P-alkyl substitution, Nozaki et al.

reported that replacement of the 2-MeOC₆H₄moieties with

cyclohexyl groups decreases the activity.^[6]In contrast, Clav-

erie found that substitution with tert-butyl groups increases

the polymerization activity drastically, but at the same time,

the catalyst decomposes rapidly.^[14]

The polyethylene molecular weight depends on electronic

and steric factors (Supporting Information, Table S27,

Figure 9). The molecular weight decreases with increasing

termined from the polymer yield should be considered as

the peak activity.

By far the highest activities were found for ^CF31 and ^Ar1

with an initial TOF of approximately 3ꢁ10⁶mol

ACTHNUTRGEN(UNG C₂H₄)

mol(Pd)^À1h^À1. However, both catalysts decompose rapidly

(vide supra). The high peak activity and highest tendency

for decomposition found for the electron-poor ^CF31 are in

line with the stoichiometric insertion studies. Similarly to

the results found for the MA insertion rates, ^H1 shows

a lower activity than both the electron-poor and the elec-

tron-rich substituted ^CF31 and ^MeO1.

In contrast, the rather electron-rich compound ^(MeO)21 ex-

hibits significantly lower activities than ^MeO1. However, the

substitution of the methoxy group by sterically more de-

manding isopropoxy groups in ^(iPrO)21 does not further affect

the activities (Supporting Information, Table S27, entries 27-

7 vs. 27-8). The introduction of a third MeO group in the

para position of the non-chelating aryl moiety in ^(MeO)31 dou-

ACTHNUTRGNEUNG

bles the activity to 4ꢁ10⁴mol(C₂H₄) mol(Pd)^À1h^À1com-

pared to ^(MeO)21 (entries 27-7 vs. 27-9). In contrast, replace-

ment of one 2,4,6-(MeO)₃C₆H₂substituent by a methyl

group in ^Me*/(MeO)31 reduces the activity to 2ꢁ10⁴mol

ACTHNUTRGEN(UNG C₂H₄)

mol(Pd)^À1h^À1(entries 27-9 vs. 27-15). The asymmetric sub-

stituted compounds ^Ar/(MeO)21 and ^cHexO/(MeO)21 are significant-

ly more active than the symmetric catalyst ^(MeO)21. It appears

that the activity of asymmetric compounds is roughly the

average of the activities of the corresponding symmetric cat-

alysts. Rather unexpectedly, a comparison of ^Ar1 and ^Ph1 re-

veals a strong difference between these closely related cata-

lysts, as ^Ar1 possesses a peak activity nearly 15 times higher.

This might point to a special role of the additional methoxy

substituents in ^Ar1, as the complexes appear to be electroni-

cally and sterically rather similar.

Figure 9. Molecular weights (M_n) of polyethylene obtained with different

catalysts (p_C2H4=5 bar, 808C).

electron deficiency of the ligand from ^MeO1 to ^CF31 (M_n: 1.2ꢁ

10⁴vs. 1.5ꢁ10³gmol^À1). However, the introduction of bulky

cyclohexyloxy substituents in ^cHexO1 (M_n: 6ꢁ10³gmol^À1)

does not increase the polymer chain length further. With the

introduction of a second ortho-methoxy group in ^(MeO)21, the

molecular weight is decreased drastically compared to ^MeO1

(M_n: 1.6ꢁ10³vs. 1.2ꢁ10⁴gmol^À1), but can be increased

again significantly by steric shielding as in ^(iPrO)21 (M_n: 1.2ꢁ

10⁴gmol^À1) or by the introduction of another electron-rich

In general, the relative activities observed in ethylene ho-

mopolymerization for different catalysts correspond to the

ratio of the acrylate insertion rates determined from NMR

experiments, with one important exception: ^(MeO)31 and ^CF31

17780

www.chemeurj.org

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2013, 19, 17773 – 17788

Phosphinesulfonato Pd^IICatalysts

FULL PAPER

group in the para position, as in ^(MeO)31 (M_n: 1.4ꢁ

10⁴gmol^À1). The bulky-aryl-substituted ^Ar1 produces the

highest-molecular-weight PE in this study (Figure 9, M_n:

2.9ꢁ10⁴gmol^À1). Comparison with ^Ph1 (M_n: 9.5ꢁ10³gmol^À1)

shows that steric shielding by phenyl substituents alone does

not lead to an increase in chain length. Again, this might in-

dicate a special role of the methoxy groups in ^Ar1, which can

arrange in the axial position of the Pd center.^[18]In this con-

text, note that for catalysts with bulky naphthyl, phenan-

thryl, anthracenyl, or 1-methoxynaphthalene substituents, no

increase in molecular weight was observed compared to that

for ^MeO1.^[15,17]A small positive effect of steric shielding on

molecular weight is also found for the asymmetric catalysts

^Ar/(MeO)21 (M_n: 1.7ꢁ10⁴gmol^À1) and ^cHexO/(MeO)21 (M_n: 2.1ꢁ

10⁴gmol^À1) compared to ^(MeO)21 (M_n: 1.6ꢁ10³gmol^À1).

The polymers produced by the different catalysts show

very similar microstructures. Highly linear polymers are ob-

tained, with no more than five methyl branches per 1000

carbon atoms and no longer branches. It seems to be an in-

trinsic property of these (P^O) chelated palladium catalysts

either to show only a minimal amount of chain walking, or

to have a strong preference for insertions into primary car-

bons. As a result, catalysts that are prone to b-H elimina-

tion, such as ^CF31, preferentially undergo chain transfer

rather than creating branches. Within the narrow range of

0–5 methyl branches per 1000 carbon atoms, it appears that

more electron-rich compounds produce a slightly greater

number of branches (Supporting Information, Table S27, en-

tries 27-2 vs. 27-5/6) and that steric shielding reduces the

number of branches (entries 27-7 vs. 27-8 and 27-2 vs. 27-4/

11/12).

^(MeO)21 vs. ^(MeO)31). Further, the combination of two distinct

substituents at phosphorus in an asymmetric catalyst proved

to be a fruitful concept for predictable modulation of cata-

lyst properties to a desirable overall catalyst profile.

Copolymerization of ethylene with polar monomers: The

catalysts were studied in the copolymerization of ethylene

with methyl acrylate at an ethylene pressure of 5 bar and

various polar monomer concentrations at 958C. The [{(^X1-

Cl)-m-M}_n] precatalysts were activated in situ with AgBF₄as

described previously. It could be shown that polymerizations

with ^MeO1 and ^MeO1-dmso result in the formation of similar

polymers with comparable activities (Supporting Informa-

tion, Table S28, entries 28-4 vs. 28-18). For ^MeO1, ^(MeO)21, and

^(MeO)31, polymerizations were performed with methyl acry-

late concentrations from 0.15 to 5 molL^À1(entries 28-1–28-

17). With increasing methyl acrylate concentration, the MA

incorporation increases, as expected. In addition, the aver-

age polymer chain length and activity decrease with increas-

ing MA concentration and incorporation, as also observed

for Pd diimine catalysts.^[38]The copolymerization activities

(TOF ethylene/MA) cannot be compared precisely between

different catalysts because the determined TOFs do not

only depend on the nature of the active species, that is the

alkyl olefin complex. Only the activities for copolymeriza-

tions yielding polymers with similar degrees of polar mono-

mer incorporation can be compared, as the formation of six-

membered k-O coordinated chelate resting states during the

insertion hinders the subsequent monomer insertion, and af-

fects the activity significantly.^[11]Even for copolymerizations

with similar incorporation ratios, the comparison of activi-

ties might be hampered, as the monomer can block the

active site through k-X coordination, so that the monomer

concentration has to be considered. However, the effect of

k-O coordination of acrylic acid esters on the activity has

been found to be rather low for (P^O)Pd catalysts.^[39]

^MeO1, ^Ar1 and ^CF31 were also compared in polymerization

at 408C to evaluate the influence of temperature on decom-

position, chain length, and branching. At 408C, ^CF31 is the

most active catalyst, which agrees with the high peak activi-

ty of this compound in the mass flow at 808C. Interestingly,

for all three catalysts, the molecular weight and branching

are both reduced at 408C, with the relation of the molecular

weights observed for different catalysts unchanged. This is

unexpected, because usually, an increase in molecular

weights is observed for polymerizations at lower tempera-

tures. Possibly, chain growth is very slow and chain transfer

caused by impurities becomes relevant.

In comparison to the acrylate insertion rates determined

by NMR, it appears that electron-deficient compounds (^CF31,

^Ar1) (in comparison with electron-rich compounds:

^(MeO)21,^(MeO)31) possess higher activities, whereas catalyst sta-

bility is decreased with increasing electron deficiency. Con-

cerning molecular weights, steric shielding can have a posi-

tive effect if adjusted appropriately.

At similar MA concentrations ([MA]=1.2m), ^CF31 incor-

porates significantly less MA than ^H1 and ^MeO1 (c_MA,CF3

=

9 mol%, c_MA,H=19 mol%, c_MA,MeO=22 mol%; Figure 10).

Steric hindrance leads to a stronger discrimination of ethyl-

ene and the more bulky MA monomer, as concluded from

the lower incorporation ratio of ^cHexO1 (c_MA,cHexO

=

13 mol%). Steric effects also limit incorporation strongly for

the very bulky ^Ar1 (c_MA,Ar=2 mol%) and ^Ph1 (c_MA,Ph

=

2 mol%). The introduction of a second methoxy group in

the ortho position of the non-chelating aryl moiety in ^(MeO)21

increases MA incorporation significantly (c_MA,(MeO)2

=

30 mol%). The same applies for ^(MeO)31 (c_MA,(MeO)3

=

33 mol%). The very high MA incorporation ratio of ^(MeO)31

reflects the effect of the relatively high MA insertion rates

observed in the NMR experiment compared with an only in-

termediate activity in the ethylene homopolymerization

with this catalyst. Furthermore, the observed increase in the

incorporation ratio for electron-rich systems can be related

to the weaker p-donor capability of the C=C double bond

of MA compared with that of ethylene. This leads to a shift

of the MA/ethylene p-coordination pre-equilibrium for elec-

In this context, it should be noted that for ^Ar1 and ^Ph1, the

metal center appears to be more globally shielded than for

the other compounds, and that for elimination and chain

transfer, the axial positions at the Pd become more impor-

tant. Electronic influences on the polymersꢃ molecular

weights are ambiguous, but electron-deficient compounds

tend to produce lower molecular weights (^CF31 vs. ^MeO1;

Chem. Eur. J. 2013, 19, 17773 – 17788

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

17781

L. Caporaso, S. Mecking et al.

Figure 11. Relation between catalyst structure and copolymer average

chain length for ꢁ10% (black) and ꢁ5% (blue) MA incorporation (de-

1

termined by H NMR spectroscopy).

Figure 10. Relation between catalyst structure and MA incorporation in

strained systems, and especially with double ortho-substitut-

ed ligands.

ethylene/MA copolymerization experiments with p_Ethylene=5 bar, and var-

ious methyl acrylate concentrations (black: [MA]=1.2 molL^À1

[MA]=0.6 molL^À1; blue: [MA]=0.3 molL^À1).

; red:

The studies resulted in the identification of catalysts and

conditions for the production of ethylene-MA copolymers

with reasonably increased M_nvalues compared with the pro-

totypical ^MeO1 (0.3 molL^À1MA, c=6 mol%, DP_n=132,

M_n=4000 gmol^À1, yield 3.2 g) at 958C. The asymmetric cata-

lyst ^cHexO/(MeO)21 combines good incorporation ratios with

high molecular weights (0.6 molL^À1MA, c=5 mol%, DP_n=

382, M_n=12000 gmol^À1, yield 0.5 g), that is, the chain length

is tripled compared with that of a copolymer with similar

composition produced by ^MeO1. Even copolymers with de-

grees of incorporation greater than 10 mol% with molecular

weights of still around 7000 gmol^À1are accessible

tron-rich compounds toward MA coordination. That is, pre-

ferred binding of the stronger p-donor ethylene is less pro-

nounced for the more electron-rich metal center. Steric

shielding in ^(iPrO)21 again reduces the incorporation ratio

drastically (c_MA,(iPrO)2=5 mol%). Substitution of one of the

2,6-(MeO)₂C₆H₃moieties in ^(MeO)21 by bulky 2-cHexOC₆H₄

or 2,6-(MeO)₂C₆H

^Ar/(MeO)21 also results in reduced MA incorporation

ACHTUNGTRENNUNG(c_MA,Ar/(MeO)2=5 mol%, c_{MA,cHexO/(MeO)2}=13 mol%).

₃ACHTUNGTRENNUNG

(C₆H₄) substituents in ^cHexO/(MeO)21 or

ACHTUNGTRENNUNG

In contrast, the replacement of

a

2,4,6-(MeO)₃C₆H₂

(1.2 molL^À1

MA,

c=13 mol%,

DP_n=200,

M_n=

moiety of ^(MeO)31 by a rather small methyl group in ^Me*/(MeO)31

7000 gmol^À1, yield 0.7 g). However, the catalytic activity of

^cHexO/(MeO)21 is rather low. High average degrees of polymeri-

zation are accessible with good activities with ^Ar1, but in this

case, already very high MA concentrations are required for

degrees of incorporation of approximately 10 mol%, which

limits its broad usability (5 molL^À1MA, c=8 mol%, DP_n=

365, M_n=12000 gmol^À1, yield=2.6 g).

does not increase the incorporation ratio further

ACHTUNGTRENNUNG(c_{MA,Me*/(MeO)3}=24 mol%). However, so far, the exact influ-

ACHTUNGTRENNUNG

ence of P-alkyl substituents remains generally ambiguous

(vide supra).^[6,14]

For all the catalysts, chain termination is strongly pre-

ferred after the insertion of methyl acrylate, which becomes

clear in the relation between MA and ethylene-derived un-

saturated end groups in comparison to the MA incorpora-

tion ratio (Supporting Information, Table S28). Therefore,

a discussion of chain length is most meaningful for polymer-

ization runs resulting in copolymers with similar degrees of

MA incorporation. A comparison of chain lengths for de-

grees of incorporation of approximately 10% and 5% re-

vealed that steric shielding as in ^cHexO1, ^cHexO/(MeO)21, ^(iPrO)21,

or ^Ar1 leads to copolymers with higher average DP_ncom-

pared to ^MeO1 and ^(MeO)21 (Figure 11). The compounds bear-

ing a second methoxy group in the ortho position of the

non-chelating aryl moieties, ^(MeO)21 and ^(MeO)31, produce sig-

nificantly shorter polymer chains than ^MeO1, as also observed

for the ethylene homopolymers. The same applies for the

electron-poor ^CF31.

Computational studies: For further illumination of the key

À

steps of the first monomer insertion reaction into the Pd

Me bond, DFT calculations were performed. For all the sys-

tems studied (Table 2) with both monomers (ethylene and

MA), the intermediate ^X1-monomer_cis with the monomer

coordinated cis to the oxygen donor exhibits the lowest

Table 2. Energies (kJmol^À1) for ethylene (E) cis-coordination, the inser-

tion transition state, and the rate-determining barrier DE relative to the

(P^O)Pd Me species ^X1+E.

À

^X1+E

^X1-E_cis

TS_Ins

DE_barrÀE

^MeO1

^CF31

0

À87

À91

À80

À84

À92

À99

À100

À86

À9

78

70

72

70

77

81

76

À21

À8

^Ar1

In conclusion, chain length can be increased by steric

shielding and decreased by electron-poor substituents or

a 2,6-(MeO)₂substitution pattern at the non-chelating aryl

moieties of the ligand. Most notably, copolymers with high

MA incorporation are accessible with sterically uncon-

^Ph1

À13

À15

À18

À19

À9

^H1

^(MeO)21

^(MeO)31

^cHexo1

17782

www.chemeurj.org

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2013, 19, 17773 – 17788

Phosphinesulfonato Pd^IICatalysts

FULL PAPER

energy, and the insertion transition state (TS_Ins) is the TOF-

determining transition state (see Supporting Information).

Accordingly, the calculations focused on these two crucial

states.

Coordination of ethylene is remarkably favored in all

cases, with coordination energies in the range 80–

100 kJmol^À1, indicating that monomer capture is not an

issue. The energy barrier for insertion (DE_barrÀE) is rather

similar for the different catalysts, and within 70–81 kJmol^À1

in all cases (Table 2). Interestingly, ^Ar1, ^Ph1, and ^CF31 exhibit

the lowest energetic barriers of about 70 kJmol^À1. This is in

line with the very high activities of ^Ar1 and ^CF31 observed in

the ethylene polymerization studies, but cannot explain the

significantly lower activity of ^Ph1.

Steric maps^[40–46]illustrate the distribution of the steric

bulk around the palladium center. Here, the Pd atom is

À

placed at the origin with the z-axis bisecting the O Pd P

angle. The P^O ligand is placed in the xz-plane, with the O

and P atoms exhibiting negative z values. For the analysis,

the methyl group and the monomer were removed and the

first coordination sphere was investigated (defining four

quadrants to facilitate orientation). In the coordination step,

the monomer is located cis to the oxygen donor, so that

quadrant IV ((IV)_coorin Figure 12) should be considered for

a discussion for this step. In contrast, the monomer is placed

trans to the oxygen in the insertion step, and consequently,

quadrant III (III)_{ins_TS}is particularly relevant for the inser-

tion reaction.

Comparison of MA and ethylene coordination

(Table 3)^[47]indicates that coordination of the latter is fa-

vored by up to 40 kJmol^À1compared to the more electron-

deficient and bulkier MA. Further, if an unfavorable ÀTDS

contribution of approximately 42 kJmol^À1is assumed for

monomer coordination,^[2,12,48]the data suggest that MA co-

ordination for ^Ar1 and ^Ph1 is almost thermoneutral, making

MA capture a critical step for these two compounds. At the

same time, the MA insertion transition states are only 16–

35 kJmol^À1higher in energy than the TS for ethylene inser-

tion, which indicates that the energy difference between the

coordination intermediate and the insertion transition state

is smaller for MA. The relative incorporation of both mono-

mers observed experimentally is reflected by the energy dif-

ference between the two insertion transition states. The cal-

culated energy difference of 16–35 kJmol^À1is large enough

to account for the preferred incorporation of ethylene in the

copolymerization experiments with these catalysts. For ^Ar1,

^Ph1 and ^CF31, the large energy difference determined be-

tween the MA and ethylene insertion transition states (>

30 kJmol^À1in favor of ethylene) is consistent with the ex-

ceptionally low degrees of MA incorporation observed in

the copolymerization experiments. In contrast, for ^MeO1, this

gap is only 18 kJmol^À1, which is in line with the higher

degree of MA incorporation observed for this catalyst. This

analysis indicates conclusively that both steric and electronic

effects can have an impact on the relative propensity to

insert ethylene and MA, which eventually makes incorpora-

tion of MA less favorable.

Figure 12. Alignment of steric maps and numbering of quadrants.

Notably, apart from ^Ar1, ^Ph1, and ^CF31, the steric maps for

MA coordination (Figure S42) and for the corresponding in-

For ethylene coordination and for the corre-

sponding insertion transition state (Figure S41)

^MeO1, ^MeO21, ^MeO31, and ^cHexO1 exhibit comparable

Table 3. Energies (kJmol^À1) for methyl acrylate (MA) cis-coordination, the insertion

steric properties, whereas ^Ar1, ^Ph1, and ^CF31 are sub-

stantially more crowded in the cis position (cf.

quadrant IV of maps (a) in Figure S41). The trans

position in TS_Insis generally more crowded than the

cis position in ^X1-E_cis (compare quadrant III of

maps (b) with quadrant IV of maps (a) in Fig-

ure S41). This is in line with the trans position being

in proximity with the phosphorus bearing the bulki-

er substituents. Here, ^Ar1, ^Ph1, and ^CF31 are also most

crowded in the trans position TS_Ins. However, steric

differences between the catalysts are reduced in the

insertion step as compared with the coordination

step.

À

transition state, and the rate-determining barrier DE relative to [P^O]Pd Me+free

monomer.^[47][a]

^X1+MA

^X1-MA_cis

DE

TS_Ins

DE_barr-MA

DDTS_Ins(MAÀE)

(^X1-MA_cis-

^X1-E_cis)

^MeO1

^CF31

0

À53

À61

À48

À43

À59

À64

À59

34

30

33

41

33

35

27

9

62

72

71

66

60

67

65

71

18

33

31

37

16

20

19

22

12

23

1

2

0

^Ar1

^Ph1

^H1

^(MeO)21

^(MeO)31

^cHexo1

13

[a] Transition states considered are for a 2,1-MA insertion.

Chem. Eur. J. 2013, 19, 17773 – 17788

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

17783

L. Caporaso, S. Mecking et al.

sertion transition states are similar to those for ethylene

(Figure S41). This indicates that here, the P^O ligand

always adopts the conformation that minimizes the intramo-

lecular steric interactions, and that steric interactions be-

tween the monomer and the ligand are not decisive in the

first place for the conformation adopted by the phosphine-

sulfonate itself, at least in these two insertion reaction steps.

In contrast to all other compounds, the steric maps for

ethylene and MA deviate significantly from each other for

^Ar1, ^Ph1, and ^CF31. Here, the MA coordination (and insertion)

geometry appears less crowded. For all complexes except

for ^Ar1, ^Ph1, and ^CF31, the exo₂conformation is the most fa-

vorable geometry for both ethylene and MA coordination.

Conversely, ^Ar1, ^Ph1, and ^CF31 undergo a change from the fa-

vored exo₃conformation observed for ethylene coordina-

from both ^Ar1-E_cis and ^Ph1-E_cis (cf. Supporting Informa-

tion). However, the energy barriers found for ^Ar1 and ^Ph1

are rather similar. A conceivable additional deactivation

À

pathway involving the C H activation of one ortho hydro-

gen atom of the C₆H₅substituent of ^Ph1 was explored

À

(Figure 13). The C H activation proceeds through a concert-

Figure 13. Possible deactivation mechanism for ^Ph1. Energies in kJmol^À1

.

À

tion, as well as for the monomer-free Pd Me species, to an

exo₂conformation for MA coordination (Figure S43,

Figure 2). Hence, for the bulky compounds ^Ar1 and ^Ph1, the

exo₃conformation is the geometry that minimizes the steric

ed oxidative hydrogen migration mechanism, in which the

hydrogen, interacting with the Pd center, migrates from the

phenyl substituent to the methyl group bound to the palladi-

um.^[52]For this deactivation pathway, an interaction of the

activated aryl with the palladium center is essential, as is

indeed found for the lowest-energy geometry of ^Ph1

(Figure 13, Figure S44B). The overall decomposition reac-

tion to Deact-^Ph1+CH₄is exergonic by 61 kJmol^À1with

a low energy barrier of 66 kJmol^À1. These results suggest

that the observed low productivity of ^Ph1 in the ethylene

polymerization can be explained by the low stability of this

compound, which decomposes easily to CH₄and Deact-^Ph1

À

intramolecular interactions in the monomer-free Pd Me

species. The same applies to ^CF31.^[49]A detailed analysis of

the geometries reveals that in ^Ar1-E_cis and ^Ph1-E_cis, the

ligand conformation leads to an interaction between the pal-

À

ladium center and one aryl substituent (see Pd C_aryldistan-

ces indicated in Figure S43A/B). This interaction leads to

a beneficial electronic influence that further stabilizes the

exo₃conformation.^[50]This interaction can account for the

differences between the ortho aryl-substituted compounds

and other bulky substituted complexes such as ^cHex1, for

which an exo₂conformation is adopted in all cases. For ^CF31,

interactions of the fluorine atoms with the Pd center were

found, which may also stabilize the exo₃conformation

owing to a beneficial electronic effect.

À

(this reactivity of the Pd Me catalyst precursor can also be

À

expected to reflect qualitatively the behavior of higher Pd

alkyls present as the catalytic species of chain growth). This

is also in line with the observed fast decomposition of this

species from mass flow experiments. In contrast, a similar

pathway is not available for ^Ar1, because here, the ortho po-

sitions are substituted by MeO groups, and the distances be-

tween the aryl substituent and palladium center are signifi-

cantly higher (see Supporting Information, Figure S44A vs.

B). Preliminary NMR experiments were performed to con-

firm the general possibility of such a decomposition path-

way. Heating ^Ph1 (synthesized in situ from [{(^Ph1-Cl)-m-M}_n],

vide supra) in C₂D₂Cl₄at 908C in a Young tube led to a con-

version of over 90% to a new species and methane within

10 min. The ESI-MS spectra of the decomposition product

are in line with the formation of Deact-^Ph1 (cf. Supporting

Information, Figures S36–S39).

To minimize the steric interaction between the bulky MA

monomer and the phosphinesulfonato ligand in the MA cis-

coordination geometry, the aryl moieties rotate, resulting in

the exo₂conformation, which is higher in energy, because of

ligand intramolecular interactions and the loss of the benefi-

À

cial Ar Pd interaction (compare Figure S43A/B/C with D/

E/F). Note that for ^Ar1, ^Ph1, and ^CF31, the exo₃conformation

is generally stabilized compared with the exo₂geometry in

all transition states and intermediates, including TS_Insand

the b-agostic intermediate.^[51]Consequently, the barriers are

not affected by the conformational change.

In conclusion, the high reactivity of ^CF31 and ^Ar1 for ethyl-

ene polymerization, and the significantly lower reactivity

toward MA in the copolymerizations, can be traced to the

loss of a favorable exo₃ligand arrangement upon binding of

MA, which is caused by steric interactions involving the mo-

nomer and the ligand.

Conclusion

For rationalization of the different experimentally ob-

served behaviors of ^Ar1 and ^Ph1, deactivation pathways were

considered. Deactivation mechanisms for ^MeO1 (vide supra)

have been unraveled in detail,^[9]and the energies for these

identified pathways were studied for ^Ar1 and ^Ph1. Calcula-

Phosphinesulfonato palladium catalysts are unique in that

they allow the synthesis of linear copolymers of ethylene

with polar olefins such as alkyl acrylates. So far, the materi-

als accessible have been limited to low-molecular-weight co-

polymers. In addition, directed catalyst development is hin-

dered by the lack of insight into the catalyst structure–activi-

ty relationship. To this end, ten symmetric and three asym-

tions focused on the rate-determining step, namely, the re-

À

ductive elimination of CH

(P^O) (Ar

N

17784

www.chemeurj.org

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2013, 19, 17773 – 17788

Phosphinesulfonato Pd^IICatalysts

FULL PAPER

metric precatalysts were synthesized and studied in terms of

their reactivity toward (polar) olefins and their performance

in (co)polymerization.

Full characterization of the compounds and comparison

of NMR and X-ray data led to a rough classification of the

corresponding palladium centers in the complexes as rather

electron-rich (^(iPrO)21, ^(MeO)21,^(MeO)31) or rather electron-defi-

cient (^CF31, ^Ar1) compared with the prototypical ^MeO1. This

agrees with the expected effect of the different types and

numbers of substituents on the P-aryl moiety.

The results of stoichiometric MA insertion studies as well

as ethylene homopolymerization and copolymerization stud-

ies are ambiguous regarding clear trends in the effect of

electronic properties. However, certain relations can be ex-

tracted. Catalysts bearing electron-deficient ligands such as

^CF31 and ^Ar1 show high activities, but also decompose rapidly

under polymerization conditions. Comparison of the poly-

mers produced by ^(MeO)21 and ^(MeO)31 indicates that more

electron-donating ligands lead to higher polymer molecular

weights. Additionally, catalysts bearing ligands with a 2,6-

(MeO)₂-substitution pattern on the non-chelating aryl moi-

eties were found to incorporate high ratios of methyl acry-

late in the copolymerization. On comparison of all the data,

it becomes clear that for ^H1 bearing no ortho substituents,

the properties determined by NMR as well as the catalytic

performance and produced (co)polymers deviate from the

general trends found on going from the more electron-defi-

cient ^CF31 to the more electron-rich ^MeO1. A reasonable ex-

planation is that steric bulk also influences the electronic

bulky ^Ar1 and ^Ph1 systems—and simultaneously allows a fa-

vorable electronic interaction through space between the

metal and the aromatic ring (^Ar1 and ^Ph1) or the fluorine

atoms (^CF31), respectively. The exo₃conformation is adopted

with the small ethylene monomer, whereas with MA, the

less favored exo₂conformation is adopted to minimize steric

interactions between the bulkier monomer and the ligand.

The observed conformational changes rationalize that ^CF31,

^Ar1, and ^Ph1 exhibit significantly lower insertion barriers in

the ethylene insertion than the other catalysts. This is in line

with the experimentally observed high reactivity of ^CF31 and

^Ar1 toward ethylene compared with the other catalysts. Con-

sidering MA insertion, a higher discrimination of MA versus

ethylene can be expected for ^CF31, ^Ar1, and ^Ph1 on the basis

of the large energy difference between the two correspond-

ing insertion transition states. In contrast, the other com-

pounds investigated show lower energy differences, leading

to a less pronounced discrimination of MA versus ethylene.

This is in line with the significantly lower degrees of MA in-

corporation found for ^CF31, ^Ar1, and ^Ph1 in the copolymeriza-

tion experiments.

Experimentally, a marked difference was found in the cat-

alytic performance of ^Ar1 and ^Ph1. That is, ^Ph1 shows a signifi-

cantly lower activity, presumably because the catalyst de-

composes rapidly under polymerization conditions. Here,

a deactivation pathway involving the ortho-H atoms of the

phenyl substituent of ^Ph1 may account for the different sta-

bilities of ^Ar1 and ^Ph1. The aforementioned electronic inter-

action between the Pd center and the aromatic ring of the

ligand appears to be a prerequisite for this decomposition

pathway. Consequently, the DFT calculations performed on

the main insertion steps for the first ethylene and MA reac-

tions suggested that the behavior of the catalysts cannot be

fully explained by considering the energetic profile of the

main steps of the insertion reaction alone, but, as shown for

^Ph1, additional factors such as the deactivation and stability

of the compounds must also be taken into account.

From steric variations, some clear trends could be identi-

fied experimentally. If the steric bulk is adjusted correctly,

the molecular weights of homo- and copolymers are en-

hanced. This is illustrated by the doubling of the copolymer

molecular weight upon the introduction of isopropyl groups

in ^(iPrO)21 compared with ^(MeO)21. This concept is further ex-

emplified by the high molecular weight produced by ^Ar1

compared with ^MeO1 and (also electronically deficient) ^CF31,

which only produces low-molecular-weight PE. It becomes

obvious from a comparison of ^Ar1 with the closely related

^Ph1, which produces PE of intermediate molecular weight,

that steric shielding alone is insufficient to enhance molecu-

lar weights. Note that the aforementioned possible decom-

position pathway for ^Ph1 may also contribute here. Regard-

ing the activity, no unfavorable effects of steric bulk were

observed in the ethylene homopolymerization. In the co-

polymerization experiments, an increase in steric bulk clear-

ly led to enhanced discrimination of the bulkier polar mono-

mer versus ethylene.

À

properties by changing the orbital overlap of the C P bond

owing to the different angles (aSPS) of the substituents (S)

at the phosphorus atom. Such an influence would be most

pronounced for the difference between a simple phenyl sub-

stituent and ortho-substituted aryl moieties.

The steric influence of the phosphinesulfonato ligand

framework on monomer coordination and insertion was

evaluated quantitatively through DFT methods for ^MeO1,

^CF31, ^(MeO)21, ^(MeO)31, ^cHexO1, ^H1, ^Ar1, and ^Ph1. Steric maps aid

the analysis of the key steps of the ethylene and MA inser-

tion pathways. As expected, ^Ar1 and ^Ph1 show more pro-

nounced steric shielding than the other catalysts. Beyond

this, for most of the complexes no significant differences in

the first coordination ground state for both the coordination

and the insertion transition state emerge for ethylene and

MA insertion, respectively. Exceptions are found for ^CF31

and for the ortho aryl-substituted compounds ^Ar1 and ^Ph1.

Consequently, the reactivities of ^MeO1, ^(MeO)21, ^(MeO)31, ^cHexO1,

and ^H1 toward ethylene versus MA should not depend on

steric factors involving interaction between the monomer

and the ligand. Conversely, for ^CF31, ^Ar1, and ^Ph1, a conforma-

tional change from the exo₃conformation in the ethylene in-

sertion to an exo₂conformation in the MA insertion was

found, whereas for the other catalysts the exo₂conformation

was calculated to be the most stable conformation in all of

the above reaction steps with both monomers. In fact, the

exo₃conformation is favored because it minimizes the steric

intramolecular interaction of the ligand—especially for the

Chem. Eur. J. 2013, 19, 17773 – 17788

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

17785

L. Caporaso, S. Mecking et al.

ligands) [{(^CF31-Cl)-m-Li}₂], [{(^(MeO)21-Cl)-m-Na}₂], [{(^(iPrO)21-Cl)-m-Na}₂],

[{(^Ar1-Cl)-m-Na}₂], [{(^cHexO/(MeO)21-Cl)-m-Na}₂], and [{(^Ar/(MeO)21-Cl)-m-

Na}₂]^[28]were prepared according to reported procedures. NMR spectra

were recorded on

DRX 600, or a Bruker Avance III spectrometer equipped with a cryo-

AHCTUNGTRENNUNG

probe head. ¹H and ¹³C NMR chemical shifts were referenced to the sol-

a Varian Unity INOVA 400, a Bruker Avance

vent signal. ¹⁹F and ³¹P NMR chemical shifts were referenced to CFCl₃

and 85% H₃PO₄, respectively. Multiplicities are given as follows (or com-

binations thereof): s: singlet, d: doublet, t: triplet, vt: virtual triplet, m:

multiplet. Diastereomeric atoms/groups/molecules are designated as X

and X’, respectively. The identity and purity of metal complexes was es-

tablished through H, ¹³C, and ³¹P NMR and elemental analysis. NMR as-

1

Figure 14. Influences of catalyst structure on reactivity.

signments were confirmed by ¹H,¹H gCOSY, ¹H,¹³C gHSQC, and ¹H,¹³C

gHMBC experiments. For ethylene homopolymers and ethylene-methyl

acrylate copolymers, gel permeation chromatography (GPC) was per-

formed in 1,2,4-trichlorobenzene at 1608C at a flow rate of 1 mLmin^À1

on a Polymer Laboratories 220 instrument equipped with Olexis columns

with differential refractive index (RI), viscosity, and light-scattering (158

and 908) detectors. Data reported were determined with the RI detector

against the linear polyethylene standard. For copolymers, the molecular

weights were determined through ¹H NMR spectroscopy, and the polydis-

persity index was determined through GPC at 1608C. Elemental analysis

and FAB mass spectroscopy were performed by the Analytical Services

at the Department of Chemistry, University of Konstanz. Elemental anal-

yses were performed on an Elementar vario MICRO cube instrument.

The insights gained (Figure 14) can be compared to sys-

tematic investigations on acyclic Pd^IIdiimine cata-

lysts.^{[2,38,53–55]}Here, it is assumed that steric shielding of the

active center is necessary to produce high-molecular-weight

material by blocking associative chain transfer. Studies con-

cerning electronic influences revealed that catalysts with

more electron-donating ligands produce PE of higher molec-

ular weight.^[54]For more electron-deficient systems, a higher

initial activity but reduced thermal stability were found.^[55]

In the copolymerization of ethylene with polar monomers, it

was observed that catalysts bearing more electron-donating

ligands incorporate higher amounts of the electron-deficient

monomer.^[2,54]

FAB mass spectra were obtained with

a double-focusing Finnagan

MAT 8200 mass spectrometer equipped with an Ion Tech (Teddington,

U.K.) FAB Ion Source. ESI mass spectra were recorded on a Bruker Es-

quire 3000+ instrument. HR-ESI MS spectra were recorded on a Bruker

Daltonics microTOF II in the positive or negative mode.

X-Ray diffraction: X-ray diffraction analyses were performed at 100 K

on a STOE IPDS-II diffractometer equipped with a graphite monochro-

mated radiation source (MoK_a, l=0.71073 ꢂ) and an image plate detec-

tion system. Crystals were mounted on a fine glass fiber with silicon

grease. The selection, integration, and averaging procedure of the mea-

sured reflex intensities, the determination of the unit cell dimensions, and

a least-squares fit of the 2q values as well as data reduction, LP correc-

tion, and space-group determination were performed using the X-Area

software package delivered with the diffractometer.^[58]A semiempirical

absorption correction was performed. The structures were solved through

the Patterson and direct methods (SHELXS-97),^[59]completed with dif-

For phosphinesulfonato Pd^IIcatalysts, precise prediction

of the influences of single substituents in, for example, the

ortho position, and of their combination, remains challeng-

ing, as underlined by the low activity and fast decomposition

of ^(MeO;Me2)1, differing from the related ^(MeO)21. Nonetheless,

the results of this study (Figure 14) provide a guideline for

the appropriate choice of the P-aryl (or alkyl) moiety. This

concept is underlined by unsymmetric catalysts with two dif-

ferent aryl (alkyl) substituents, which effectively combine

the properties imparted by the individual substituents to

achieve a desirable overall catalyst profile. The virtues of

this approach are reflected in the tripled molecular weights

of copolymers with a similar degree of incorporation pro-

duced by ^cHexO/(MeO)21 (M_nꢁ12000 gmol^À1) compared with

the prototypical ^MeO1. In addition, conditions have also been

identified for the production of similar polymers with the

symmetric catalyst ^Ar1. However, in this case, the incorpora-

tion ratio is limited: for a copolymer containing 8 mol%

MA, an MA concentration of 5 molL^À1(neat MA=

11 molL^À1) is required at an ethylene pressure of 5 bar.

ference Fourier syntheses, and refined with full-matrix least-squares using

2

SHELXL-97^[60]minimizing w

ACHTUNGTRENNUNG

the goodness of fit GooF are based on F². All non-hydrogen atoms were

refined with anisotropic displacement parameters. All hydrogen atoms

were treated in a riding model. Structures were plotted using Dia-

mond 3.1.^[61]The drawn ellipsoids represent 50% probability.

DFT calculations: All density functional theory (DFT) calculations were

performed using the Gaussian 09 package.^[62]The GGA functional of

Becke and Perdew (BP86) with the 6-31G(d) basis set for main-group

atoms and the LANL₂DZ ECP for palladium was used.^{[63, 64]}All geome-

tries were localized in the gas phase at the BP86 level. Minima were lo-

calized by full optimization of the starting structures, whereas transition

states were approached through a linear transit procedure and then locat-

ed by a full transition-state search. All structures were confirmed as

a minimum or transition state through frequency calculations.

Experimental Section

Acknowledgements

Materials and general considerations: Unless noted otherwise, all manip-

ulations of air-sensitive compounds were performed under an inert at-

mosphere using standard glovebox or Schlenk techniques. THF, toluene,

and CH₂Cl₂were dried by using standard protocols.^[56]Pentane and Et₂O

were dried by passing through columns equipped with aluminum oxide/

molecular sieve (3 ꢂ). Ethylene (3.5 grade) supplied by Praxair and

methyl acrylate (99%) supplied by Aldrich were used as received.

[(cod)PdMeCl],^[57][2-(phenyl)phosphino]benzenesulfonic acid,^[25][{(1-

Cl)-m-Na}₂],^[27]^MeO1-dmso,^[10]as well as the complexes (and corresponding

This work was supported financially by the DFG (Me 1388/10). B.N. ac-

knowledges support by the state of Baden-Wꢄrttemberg by a Landesgra-

duiertenfçrderung-Stipend. The authors thank Anna-Lena Steck for

high-resolution ESI-MS measurements, Alexander Klaiber for participa-

tion in this research as a part of his undergraduate studies, and the HPC

team of Enea for use of the ENEA-GRID and the HPC facilities

CRESCO in Portici, Italy.

17786

www.chemeurj.org

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2013, 19, 17773 – 17788

Phosphinesulfonato Pd^IICatalysts

FULL PAPER

[33] P. Wucher, L. Caporaso, P. Roesle, F. Ragone, L. Cavallo, S. Meck-

ing, I. Gçttker-Schnetmann, Proc. Natl. Acad. Sci. USA 2011 , 108 ,

895 5 –8 959.

[1] L. K. Johnson, S. Mecking, M. Brookhart, J. Am. Chem. Soc. 1996 ,

118 , 26 7 –2 68.

[2] S. Mecking, L. K. Johnson, L. Wang, M. Brookhart, J. Am. Chem.

Soc. 1998 , 120 , 88 8 –8 99.

[3] E. Drent, R. v. Dijk, R. v. Ginkel, B. v. Oort, R. I. Pugh, Chem.

Commun. 2002 , 74 4 –7 45.

[4] T. Kochi, S. Noda, K. Yoshimura, K. Nozaki, J. Am. Chem. Soc.

2007 , 129 , 894 8 –8 949.

[34] Mass flow curves were fitted with exponential decay functions

Ae^Àx/t+B for ꢁ1600 data points, to express the rough trend of the

decay of the curves. Consequently, information concerning reaction

order cannot be extracted from these curves.

[35] Note that a mass flow below approximately 1–2 gh^À1cannot be

monitored with the experimental setup employed.

[5] K. Nozaki, S. Kusumoto, S. Noda, T. Kochi, L. W. Chung, K. Moro-

kuma, J. Am. Chem. Soc. 2010 , 132 , 1603 0 –1 6042.

[6] S. Ito, K. Munakata, A. Nakamura, K. Nozaki, J. Am. Chem. Soc.

2009 , 131 , 1460 6 –1 4607.

[36] Before integration of the mass flow, a mass flow obtained under the

same conditions in the absence of catalyst was subtracted from the

raw data, to account for saturation of the liquid phase with ethylene

in the early stages of the experiment.

[7] T. Rꢄnzi, D. Frçhlich, S. Mecking, J. Am. Chem. Soc. 2010, 132,

17690–17691.

[37] Z. Cai, Z. Shen, X. Zhou, R. F. Jordan, ACS Catal. 2012 , 2 , 118 7 –

1195.

[8] A. Berkefeld, S. Mecking, Angew. Chem. 2008 , 120 , 257 2 –2 576;

Angew. Chem. Int. Ed. 2008 , 47 , 253 8 –2 542; A. Nakamura, S. Ito,

K. Nozaki, Chem. Rev. 2009 , 109 , 521 5 –5 244; E. Y. X. Chen, Chem.

Rev. 2009 , 109 , 515 7 –5 214; D. Takeuchi, Dalton Trans. 2010 , 39 ,

31 1 –3 28.

[9] T. Rꢄnzi, U. Tritschler, P. Roesle, I. Gçttker-Schnetmann, H. M.

Mçller, L. Caporaso, A. Poater, L. Cavallo, S. Mecking, Organome-

tallics 2012 , 31 , 838 8 –8 406.

[38] S. D. Ittel, L. K. Johnson, M. Brookhart, Chem. Rev. 2000 , 100 ,

116 9 –1 204.

[39] T. Friedberger, P. Wucher, S. Mecking, J. Am. Chem. Soc. 2012 , 134 ,

101 0 –1 018.

[40] A. Poater, B. Cosenza, A. Correa, S. Giudice, F. Ragone, V. Scarano,

L. Cavallo, Eur. J. Inorg. Chem. 2009 , 175 9 –1 766.

[41] L. Cavallo, A. Correa, C. Costabile, H. Jacobsen, J. Organomet.

Chem. 2005 , 690 , 540 7 –5 413.

[10] D. Guironnet, P. Roesle, T. Rꢄnzi, I. Gçttker-Schnetmann, S. Meck-

ing, J. Am. Chem. Soc. 2009 , 131 , 42 2 –4 23.

[42] A. Poater, F. Ragone, S. Giudice, C. Costabile, R. Dorta, S. P. Nolan,

L. Cavallo, Organometallics 2008 , 27 , 267 9 –2 681.

[11] B. Neuwald, F. ꢅlscher, I. Gçttker-Schnetmann, S. Mecking, Orga-

nometallics 2012 , 31 , 312 8 –3 137.

[43] A. Poater, L. Cavallo, Dalton Trans. 2009, 41, 8885–8890.

[44] F. Ragone, A. Poater, L. Cavallo, J. Am. Chem. Soc. 2010 , 132 ,

424 9 –4 258.

[45] A. Poater, F. Ragone, R. Mariz, R. Dorta, L. Cavallo, Chem. Eur. J.

2010 , 16 , 1434 8 –1 4353.

[12] D. Guironnet, L. Caporaso, B. Neuwald, I. Gçttker-Schnetmann, L.

Cavallo, S. Mecking, J. Am. Chem. Soc. 2010 , 132 , 441 8 –4 426.

[13] Related stoichiometric studies of acrylate insertion in Pd^IIcom-

plexes: a) P. Braunstein, C. Frison, X. Morise, Angew. Chem. 2000 ,

112 , 298 9 –2 992; Angew. Chem. Int. Ed. 2000, 39, 2867–2870; b) M.

Agostinho, P. Braunstein, Chem. Commun. 2007 , 5 8 –6 0.

[14] L. Piche, J.-C. Daigle, G. Rehse, J. P. Claverie, Chem. Eur. J. 2012 ,

18 , 327 7 –3 285.

[15] L. Piche, J.-C. Daigle, R. Poli, J. P. Claverie, Eur. J. Inorg. Chem.

2010 , 459 5 –4 601.

[16] N. T. Allen, B. L. Goodall, L. H. McIntosh, EP1760086A2, 2007.

[17] T. M. J. Anselment, C. Wichmann, C. E. Anderson, E. Herdtweck,

B. Rieger, Organometallics 2011 , 30 , 660 2 –6 611.

[18] K. M. Skupov, P. R. Marella, M. Simard, G. P. A. Yap, N. Allen, D.

Conner, B. L. Goodall, J. P. Claverie, Macromol. Rapid Commun.

2007 , 28 , 203 3 –2 038.

[19] L. K. Johnson, C. M. Killian, M. Brookhart, J. Am. Chem. Soc. 1995 ,

117 , 641 4 –6 415.

[20] T. K. Woo, T. Ziegler, J. Organomet. Chem. 1999 , 591 , 20 4 –2 13.

[21] Z. Shen, R. F. Jordan, J. Am. Chem. Soc. 2010 , 132 , 5 2 –5 3.

[22] D. Guironnet, T. Rꢄnzi, I. Gçttker-Schnetmann, S. Mecking, Chem.

Commun. 2008 , 496 5 –4 967.

[23] E. Drent, R. v. Dijk, R. v. Ginkel, B. v. Oort, R. I. Pugh, Chem.

Commun. 2002 , 96 4 –9 65.

[46] A. Bondi, J. Phys. Chem. 1964 , 68 , 44 1 –4 51.

[47] Note that the energies for ^MeO1 in Table 3 differ slightly from those

reported in Ref. [12] as the values were obtained with different pro-

grams (ADF vs. Gaussian09).

[48] A. Haras, G. D. W. Anderson, A. Michalak, B. Rieger, T. Ziegler,

Organometallics 2006 , 25 , 449 1 –4 497.

[49] In the solid state, most of the (P^O)PdMe complexes show an exo₂

conformation, whereas a change to an exo₃conformation is ob-

served for bulky ligands (see ref. [28]). For ^Ar1-L it has been ob-

served that both exo₃and exo₂conformations can be adopted in the

solid state depending on L (see ref. [18]), whereas for ^Ph1-lut and

^CF31-lut an exo₂conformation was found (Figure 2).

[50] For the electronic influence of an Ar–Ru^IIinteraction through

space, see: R. Credendino, L. Falivene, L. Cavallo, J. Am. Chem.

Soc. 2012 , 134 , 812 7 –8 135. For a back-biting h⁶-arene coordination

to

(CH₂Ph)ACTHUNGTRENNNUG

a ACHTUTGNENRNUG

d⁰-Zr^IVcenter in [Cp*Zr(CH₂CHMeCH₂Ph)

T

see: C. Pellecchia, A. Immirzi, A. Zambelli, J. Organomet. Chem.

1994, 479, C9–C11.

[51] For ^CF31 the exo₃conformation is calculated to be favored over the

exo₂conformation only by ca. 8 kJmol^À1. Considering this small

energy difference, it is reasonable that the adoption of an exo₂con-

formation was observed in the solid state (see above).

[24] J. Vela, G. R. Lief, Z. Shen, R. F. Jordan, Organometallics 2007 , 26 ,

662 4 –6 635.

[25] A. K. Hearley, R. J. Nowack, B. Rieger, Organometallics 2005 , 24 ,

275 5 –2 763.

[26] W. N. Chou, M. Pomerantz, J. Org. Chem. 1991 , 56 , 276 2 –2 769.

[27] T. Rꢄnzi, D. Guironnet, I. Gçttker-Schnetmann, S. Mecking, J. Am.

Chem. Soc. 2010, 132, 16623–16630.

[52] A. G. Wong-Foy, G. Bhalla, X. Y. Liu, R. A. Periana, J. Am. Chem.

Soc. 2003 , 125 , 1429 2 –1 4293; J. Oxgaard, R. P. Muller, W. A. God-

dard, R. A. Periana, J. Am. Chem. Soc. 2004 , 126 , 35 2 –3 63; G.

Bhalla, X. Y. Liu, J. Oxgaard, W. A. Goddard, R. A. Periana, J. Am.

Chem. Soc. 2005 , 127 , 1137 2 –1 1389.

[28] B. Neuwald, L. Caporaso, L. Cavallo, S. Mecking, J. Am. Chem. Soc.

2013 , 135 , 102 6 –1 036.

[29] A. Berkefeld, S. Mecking, J. Am. Chem. Soc. 2009 , 131 , 156 5 –1 574.

[30] A. C. Albꢆniz, P. Espinet, R. Lopez-Fernandez, Organometallics

2003 , 22 , 420 6 –4 212.

[31] M. Nagel, A. Sen, Organometallics 2006 , 25 , 472 2 –4 724.

[32] In addition, free-radical-formed PMA might be observed in the case

of a free-radical pathway to D and G. However, such an effect is not

distinguishable from other processes that form free radicals, and fur-

ther, would be suppressed by BHT, which was added in the insertion

reaction with ^Ar1 (see the Supporting Information).

[53] D. P. Gates, S. A. Svejda, E. OÇate, C. M. Killian, L. K. Johnson,

P. S. White, M. Brookhart, Macromolecules 2000 , 33 , 232 0 –2 334.

[54] C. Popeney, Z. Guan, Organometallics 2005 , 24 , 114 5 –1 155.

[55] C. S. Popeney, Z. Guan, Macromolecules 2010 , 43 , 409 1 –4 097.

[56] W. L. Armarego, D. D. Perrin, Purification of Laboratory Chemicals,

4th ed., Elsevier, Burlington, 1996.

[57] E. V. Salo, Z. Guan, Organometallics 2003 , 22 , 503 3 –5 046.

[58] Stoe X-RED Ver. 1.3, Data Reduction Programm Darmstadt, Ger-

many, 2005.

[59] G. M. Sheldrick, SHELXS-97, Program for Crystal Structure Analy-

sis, Universitꢇt Gçttingen, Gçttingen, 1986.

Chem. Eur. J. 2013, 19, 17773 – 17788

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

17787

L. Caporaso, S. Mecking et al.

[60] G. M. Sheldrick, SHELXL-97, Program for Crystal Structure Re-

finement, Universitꢇt Gçttingen, Gçttingen, 1997.

[61] K. Brandenburg, Diamond Ver. 3.0a, Crystal Impact GbR, Bonn,

2005.

Knox, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts,

R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli,

J. W. Ochterski, R. L. Martin, K. Morokuma, V. G. Zakrzewski,

G. A. Voth, P. Salvador, J. J. Dannenberg, S. Dapprich, A. D. Dan-

iels, ꢅ. Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski, D. J. Fox,

Gaussian, Inc., Wallingford CT, 2009.

[62] Gaussian 09, Revision D.01, M. J. Frisch, G. W. Trucks, H. B. Schle-

gel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V.

Barone, B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X.

Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Son-

nenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa,

M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vre-

ven, Jr., J. A. Montgomery, J. E. Peralta, F. Ogliaro, M. Bearpark,

J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov, R. Kobayashi,

J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyen-

gar, J. Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene, J. E.

[63] A. D. Becke, Phys. Rev. A 1988 , 38 , 309 8 –3 100; J. P. Perdew, Phys.

Rev. B 1986 , 34 , 7406; J. P. Perdew, Phys. Rev. B 1986 , 33 , 882 2 –

8824.

[64] F. Weigend, R. Ahlrichs, Phys. Chem. Chem. Phys. 2005 , 7 , 329 7 –

3305.

Received: April 10, 2013

Revised: September 17, 2013

Published online: November 21, 2013

17788

www.chemeurj.org

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2013, 19, 17773 – 17788

Article Doi

Exploring electronic and steric effects on the insertion and polymerization reactivity of phosphinesulfonato pdii catalysts

DOI: 10.1002/chem.201301365

Source and publish data:

Authors:

Article abstract of DOI:10.1002/chem.201301365

Full text of DOI:10.1002/chem.201301365

Products guided by the article

R&D Labs maybe for 926936-65-4

Relevant to this article

Hot Product