Marked Counteranion Effects on Single-Site Olefin Polymerization Processes. Correlations of Ion Pair Structure and Dynamics with Polymerization Activity, Chain Transfer, and Syndioselectivity

Published on Web 03/17/2004

Marked Counteranion Effects on Single-Site Olefin

Polymerization Processes. Correlations of Ion Pair Structure

and Dynamics with Polymerization Activity, Chain Transfer,

and Syndioselectivity

Ming-Chou Chen, John A. S. Roberts, and Tobin J. Marks*

Contribution from the Department of Chemistry, Northwestern UniVersity,

EVanston, Illinois 60208-3113

Received May 22, 2003; E-mail: tjmarks@casbah.acns.nwu.edu

Abstract: Counteranion effects on the rate and stereochemistry of syndiotactic propylene enchainment by

the archetypal C_s-symmetric precatalyst [Me₂C(Cp)(Flu)]ZrMe₂(1; Cp ) C₅H₄; Flu ) C₁₃H₈, fluorenyl) are

probed using the cocatalysts MAO (2), B(C₆F₅)₃(3)_,B(2-C₆F₅C₆F₄)₃(4)_,Ph₃C⁺B(C₆F₅)₄(5), and Ph₃C⁺-

-

FAl(2-C₆F₅C₆F₄)₃^-(6), offering greatly different structural and ion pairing characteristics. Reaction of 1 with

3 affords [Me₂C(Cp)(Flu)]ZrMe⁺MeB(C₆F₅)₃(7). In the case of 4, this reaction leads to formation the

-

µ-methyl dinuclear diastereomers {([Me₂C(Cp)(Flu)]ZrMe)₂(µ-Me)}⁺MeB(2-C₆F₅C₆F₄)₃(8). A similar

-

reaction with 6 results in diastereomeric [Me₂C(Cp)(Flu)]ZrMe⁺FAl(2-C₆F₅C₆F₄)₃(10) ion pairs. The

-

molecular structures of 7 and 10 have been determined by single-crystal X-ray diffraction. Reorganization

pathways available to these species have been examined using EXSY and dynamic NMR, revealing that

-

the cation-MeB(C₆F₅)₃interaction is considerably weaker/more mobile than in the FAl(2-C₆F₅C₆F₄)₃

-

derived analogue. Polymerizations mediated by 1 in toluene over the temperature range of -10° to

+60 °C and at 1.0-5.0 atm propylene pressure (at 60 °C) reveal that activity, product syndiotacticity, m

and mm stereodefect generation, and chain transfer processes are highly sensitive to the nature of the ion

pairing. Thus, the complexes activated with 4 and 5, having the weakest ion pairing, yield the highest

estimated propagation rates, while with 6, having the strongest pairing, yields the lowest. The strongly

coordinating, immobile FAl(2-C₆F₅C₆F₄)₃^-anion produces the highest/least temperature-dependent product

syndiotacticity, lowest/least temperature-dependent m stereodefect abundance, and highest product

molecular weight. These polypropylene microstructural parameters, and also M_w, are least sensitive to

increased propylene pressure for FAl(2-C₆F₅C₆F₄)₃^-, but highest with MeB(C₆F₅)₃^-. In general, mm

stereodefect production is only modestly anion-sensitive; [propylene] dependence studies reveal enantiofacial

propylene misinsertion to be the prevailing mm-generating process in all systems at 60 °C, being most

dominant with 6, where mm stereodefect abundance is lowest. For 1,3-dichlorobenzene as the polymerization

solvent, product syndiotacticity, as well as m and mm stereodefects, become indistinguishable for all

cocatalysts. These observations are consistent with a scenario in which ion pairing modulates the rates of

stereodefect generating processes relative to monomer enchainment, hence net enchainment syndiose-

lectivity, and also dictates the rate of termination relative to propagation and the preferred termination

pathway. In comparison to 3-6, propylene polymerization mediated by MAO (2) + 1 in toluene reveals an

estimated ordering in site epimerization rates as 5 > 4 > 2 > 3 > 6, while product syndiotacticities rank

as 6 > 2 > 5 ∼ 4 > 3.

perfluoroarylborates,⁷and aluminates,⁸all of which undergo

reaction with metallocenes to generate highly active “cationic”

complexes as the actual agents for olefin polymerization (A;

Introduction

Cocatalysts are of great current interest as vital components

of single-site olefin polymerization catalysts.^1,2Well-known

cocatalysts include methylaluminoxane (MAO; 2),³tris(per-

fluorophenyl)borane B(C₆F₅)₃; (3),⁴and related perfluoroaryl-

boranes,⁵ammonium or trityl salts of B(C₆F₅)₄^-(5)⁶and related

eq 1). Over the past two decades, numerous elegant efforts have

(1) For recent reviews, see: (a) Pe´deutour, J.-N.; Radhakrishnan, K.; Cramail,

H.; Deffieux, A. Macromol. Rapid Commun. 2001, 22, 1095-1123. (b)

Chen, Y.-X.; Marks, T. J. Chem. ReV. 2000, 100 (4), 1391-1434. (c)

Gladysz, J. A., Ed. Chem. ReV. 2000, 100, 1167-1682. (d) Marks, T. J.;

Stevens, J. C., Eds. Top. Catal. 1999, 7, 1-208. (e) Britovsek, G. J. P.;

Gibson, V. C.; Wass, D. F. Angew. Chem., Int. Ed. Engl. 1999, 38, 428-

447. (f) Jordan, R. F.; Ed. J. Mol. Catal. 1998, 128, 1-337.

been directed at “engineering” the cationic portion of such

catalysts,^1chowever only recently has the charge-compensating

anion (X^-) begun to receive attention in regard to understanding

9

J. AM. CHEM. SOC. 2004, 126, 4605-4625

4605

A R T I C L E S

Chen et al.

and optimizing the role of the ion pairing dynamics in catalyst

system performance. Strong evidence now suggests that the

activator and the structures of the resulting ion pairs can have

a profound influence on single-site polymerization catalyst

activity, lifetime, stability, chain-transfer characteristics, and

possibly stereoregulation.^1,2As part of our continuing efforts

to characterize cocatalyst-related structure-reactivity relation-

ships for such catalysts, we are particularly interested in

fluoroarylborate and -aluminate anions and the ion pairing

behavior of complexes derived from them. Recently, we

communicated some preliminary observations on counteranion

effects on propylene enchainment stereochemistry by the

archetypal C_s-symmetric precatalyst [Me₂C(Cp)(Flu)]ZrMe₂(1;

Cp ) C₅H₄; Flu ) fluorenyl),⁹using a series of structurally/

coordinatively diverse cocatalysts/counteranions.¹⁰In principle,

the established pathway¹¹for syndiospecific propylene enchain-

ment by C_s-symmetric catalysts should be a sensitive probe of

the importance of cocatalyst/counteranion^1,2interactions since

olefin enchainment must occur in concert with “chain swinging”

(eq 2, R ) polypropylene fragment). It is known that rates of

similar reorganization/symmetrization processes are sensitive

to ion pairing strength in model metallocenium systems (R )

H, alkyl group),¹²and thought that analogous “back-skip”

processes without concomitant enchainment are a major source

of polypropylene stereodefects in C_s-symmetric systems (site

epimerization, Scheme 1B). These stereodefects, in particular

m-type stereodefects, have distinct spectroscopic signatures and

can be quantified, as has been established.^11,13

In the preliminary work,¹⁰it was observed that counteranion

effects are strikingly large, and to a significant degree qualita-

tively understandable, in terms of established trends in ion

pairing strength and dynamics. The results at that stage

suggested a mechanistic picture in which anion-specific ion

pairing effects modulate not only the enchainment and chain

transfer rates, but more importantly, the relatiVe rates of

enchainment versus m stereodefect generation. This suggested

(2) For recent cocatalyst studies, see: (a) Busico, V.; Cipullo, R.; Cutillo, F.;

Vacatello, M.; Van Axel Castelli, V. Macromolecules 2003, 36, 4258-

4261. (b) Mohammed, M.; Nele, M.; Al-Humydi, A.; Xin, S.; Stapleton,

R.; Collins, C. J. Am. Chem. Soc. 2003, 125, 7930-7941. (c) Li, L.; Metz,

M. V.; Li, H.; Chen, M.-C.; Marks, T. J. J. Am. Chem. Soc. 2002, 124,

12 725-12 741. (d) Metz, M. V.; Schwartz, D. J.; Stern, C. L.; Marks, T.

J.; Nickias, P. N. Organometallics 2002, 21, 4159-4168. (e) Metz, M. V.;

Sun, Y. M.; Stern, C. L.; Marks, T. J. Organometallics 2002, 21, 3691-

3702. (f) Wilmes, G. M.; Polse, J. L.; Waymouth, R. M. Macromolecules

2002, 35, 6766-6772. (g) Lancaster, S. J.; Rodriguez, A.; Lara-Sanchez,

A.; Hannant, M. D.; Walker, D. A.; Hughes, D. H.; Bochmann, M.

Organometallics 2002, 21, 451-453. (h) Rodriguez, G.; Brant, P. Orga-

nometallics 2001, 20, 2417-2420. (i) Kaul, F. A. R.; Puchta, G. T.;

Schneider, H.; Grosche, M.; Mihalios, D.; Herrmann, W. A. J. Organo-

metallic Chem. 2001, 621, 177-183. (j) Chen, Y.-X.; Kruper, W. J.; Roof

G.; Wilson, D. R. J. Am. Chem. Soc. 2001, 123, 745-746. (k) Zhou, J.;

Lancaster, S. J.; Walker, D. A.; Beck, S.; Thornton-Pett, M.; Bochmann,

M. J. Am. Chem. Soc. 2001, 123, 223-237. (l) Kehr, G.; Roesmann, R.;

Frohlich, R.; Holst, C.; Erker, G. Eur. I. Inorg. Chem. 2001, 535-538.

(m) Mager, M.; Becke, S.; Windisch, H.; Denninger, U. Angew. Chem.,

Int. Ed. Engl. 2001, 40, 1898-1902. (n) Chase, P. A.; Piers, W. E.; Patrick,

B. O.; J. Am. Chem. Soc. 2000, 122, 12 911-12 912. (o) LaPointe, R. E.;

Roof, G. R.; Abboud, K. A.; Klosin, J. J. Am. Chem. Soc. 2000, 122, 9560-

9561. (p) Sun, Y. M.; Metz, M. V.; Stern, C. L.; Marks, T. J. Organome-

tallics 2000, 19, 1625-1627. (q) Metz, M. V.; Schwartz, D. J.; Stern, C.

L.; Nickias, P. N.; Marks, T. J. Angew. Chem., Int. Ed. Engl. 2000, 39,

1312-1316.

(3) (a) Sinn, H.; Kaminsky, W. AdV. Organomet. Chem. 1980, 18, 99-149.

(b) Sinn, H.; Kaminsky, W.; Vollmer, H.-J.; Woldt, R. Angew. Chem., Int.

Ed. Engl. 1980, 19, 390-392.

(4) (a) Yang, X.; Stern, C. L.; Marks, T. J. J. Am. Chem. Soc. 1994, 116,

10 015-10 031. (b)Yang, X.; Stern, C. L.; Marks, T. J. J. Am. Chem. Soc.

1991, 113, 3623-3625. (c) Ewen, J. A.; Elder, M. J. Chem. Abstr. 1991,

115, 136 998g.

(5) (a) Li, L.; Stern, C. L.; Marks, T. J. Organometallics 2000, 19, 3332-

3337. (b) Li, L.; Marks, T. J. Organometallics 1998, 17, 3996-4003. (c)

Chen, Y.-X.; Stern, C. L.; Yang, S.; Marks, T. J. J. Am. Chem. Soc. 1996,

118, 12 451-12 452.6. (d) also see ref 2c,d. (e) For a recent chelating borane

review, see: Piers, W. E.; Irvine, G. J.; Williams, V. C. Eur. J. Inorg.

Chem. 2000, 2131-2142.

(6) (a) Chien, J. C. W.; Tsai, W.-M.; Rausch, M. D. J. Am. Chem. Soc. 1991,

113, 8570-8571. (b) Yang, X.; Stern, C. L.; Marks, T. J. Organometallics

1991, 10, 840-842. (c) Ewen, J. A.; Elder, M. J. Eur. Pat. Appl. 426637

1991; Chem. Abstr. 1991, 115, 136 987c, 136 988d.

-

that the strong coordinative characteristics of FAl(2-C₆F₅C₆F₄)₃

(B)^8a,blead to more tightly bound, stereochemically immobile

ion pairs, accounting both for the decrease in polymerization

activity and for the enhancement in stereoselectivity. A solvent

effect was also observed: in low-polarity toluene, polymeriza-

tion activity, product syndiotacticity, and product molecular

weight are sensitively dependent on counteranion identity. In

contrast, a “leveling effect” on product stereoregularity is

observed in polar 1,3-dichlorobenzene, i.e., the anion depen-

dence is strongly attenuated.

These findings and the questions raised by the apparent

significance of ion pairing in C_s-symmetric polymerization

systems motivate the present broader and more quantitative

investigation of solution-phase catalyst structure and dynamics,

and correlation of these results with polymerization activity,

chain transfer pathways, and tacticity/microstructure, as well

as detailed determination of ion pair structures in the solid state.

Ion pairing effects are found to manifest themselves differently

for different processes occurring during polymerization, allowing

nonsystematic effects on directly observable product polymer

properties (e.g., on syndiotacticity or average molecular weight)

to emerge from systematic effects on individual processes

(propagation, site epimerization, chain release, etc.).

(7) For related fluorinated tetraarylborates, see: (a) Kaul, F. A. R.; Puchta, G.

T.; Schneider, H.; Grosche, M.; Mihalios, D.; Herrmann, W. A. J.

Organomet. Chem. 2001, 621, 184-189. (b) also see refs 2g,h,k. (c) Jia,

L.; Yang, X.; Stern, C. L.; Marks, T. J. Organometallics 1997, 16, 842-

857. (d) Jia, L.; Yang, X.; Ishihara, A.; Marks, T. J. Organometallics 1995,

14, 3135-3137.

(8) (a) Chen, Y.-X.; Metz, M. V.; Li, L.; Stern, C. L.; Marks, T. J. J. Am.

Chem. Soc. 1998, 120, 6287-6305. (b) Chen, Y. X.; Stern, C. L.; Marks,

T. J. J. Am. Chem. Soc. 1997, 119, 2582-2583. (c) Elder, M. J.; Ewen, J.

A. Eur. Pat. Appl. EP 573, 403, 1993; Chem. Abstr. 1994, 121, 0207d. (d)

also see Ref. 2o.

Furthermore, recent reports of concentration-dependent ion

pair aggregation and anion exchange processes in zirconocenium

(9) (a) Razavi, A.; Thewalt, U. J. Organomet. Chem. 1993, 445, 111-114.

(b) Razavi, A.; Ferrara, J. J. Organomet. Chem. 1992, 435, 299-310.

(10) Chen, M.-C.; Marks, T. J. J. Am. Chem. Soc. 2001, 123, 11 803-11 804.

(11) (a) Resconi, L.; Cavallo, L.; Fait, A.; Piemontesi, F. in ref. 1c, pp 1253-

1345. (b) Coates, G. W. in ref 1c, pp 1223-1252. (c) Veghini, D.; Henling,

L. M.; Burkhardt, T. J.; Bercaw, J. E. J. Am. Chem. Soc. 1999, 121, 564-

573. (d) Ewen, J. A.; Jones, R. L.; Razavi, A.; Ferrara, J. D. J. Am. Chem.

Soc. 1988, 110, 6255-6256.

(12) (a) Beswick, C. L.; Marks, T. J. J. Am. Chem. Soc. 2000, 122, 10 358-

10 370. (b) Deck, P. A.; Beswick, C. L.; Marks, T. J. J. Am. Chem. Soc.

1998, 120, 1772-1784. (c) Luo, L.; Marks, T. J. in ref 1d, pp 97-106. (d)

also see refs 7c and 8a.

(13) (a) See ref 2a. (b) Busico, V.; Cipullo, R. Prog. Polym. Sci. 2001, 26,

443-533. (c) Farina, M.; Terragni, A. Makromol. Chem., Rapid Commun.

1993, 14, 791-798. (d) Such techniques are reviewed in ref 11a.

9

4606 J. AM. CHEM. SOC. VOL. 126, NO. 14, 2004

Counteranion Effects on Olefin Polymerization

A R T I C L E S

Scheme 1

MeB(C₆F₅)₃^-systems and their sensitivity to Li⁺MeB(C₆F₅)₃

polarity effects on ion pairing during polymerization, cocatalyst

effects in an even less polar solvent, octane, are compared to

previous results. In addition, the spectrum of cocatalysts studied

has been expanded to include MAO, to compare ion-pairing

effects in this broadly utilized cocatalyst to the previously

investigated species. Propylene polymerization catalyzed by

1+MAO is carried out over a range of temperatures and

propylene pressures, and in the various solvents, and these

-

addition¹⁴prompt questions about potential ion pair aggregation

effects on enchainment in this system. Thus, catalyst concentra-

tion effects are also examined here, as are possible influences

of counteranion exchange and added Li⁺MeB(C₆F₅)₃on

-

propylene polymerization. To further detail and elucidate solvent

(14) Beck, S.; Lieber, S.; Schaper, F.; Geyer, A.; Brintzinger, H. H. J. J. Am.

Chem. Soc. 2001, 123, 1483-1489.

9

J. AM. CHEM. SOC. VOL. 126, NO. 14, 2004 4607

A R T I C L E S

Chen et al.

results are compared to those obtained with the other cocatalysts.

In addition, X-ray crystallographic characterization of [Me₂C-

(Cp)(Flu)]ZrMe⁺MeB(C₆F₅)₃^-(7) and [Me₂C(Cp)(Flu)]ZrMe⁺

FAl(2-C₆F₅C₆F₄)₃^-(10), and determination of the solution phase

molecular dynamics of these complexes, lead to a description

of the molecular basis for catalytic activity and selectivity.

Compared to B(C₆F₅)₄^-, CH₃B(C₆F₅)₃^-, which is commonly

accepted as more strongly coordinating, exhibits an expected

lower polymerization activity,^1b,7c,7dbut surprisingly and without

precedent, lower polypropylene stereoregularity. Using a straight-

forward kinetic model, we provide here a rationalization of this

interesting, counterintuitive result and then generalize it to all

of the cocatalyst systems examined here.

Experimental Section

Materials and Methods. All manipulations of air-sensitive materials

were performed with rigorous exclusion of oxygen and moisture in

flamed Schlenk-type glassware on a dual-manifold Schlenk line or

interfaced to a high-vacuum line (10^-6Torr), or in an N₂-filled Vacuum

Atmospheres glovebox with a high capacity recirculator (<1 ppm O₂).

Argon (Matheson, prepurified), and propylene (Matheson, polymeri-

zation grade) were purified by passage through a supported MnO

oxygen-removal column and an activated Davison 4A molecular sieve

column. Hydrocarbon solvents (toluene and pentane) were distilled

under nitrogen from Na/K alloy/benzophenone ketyl. All solvents for

high-vacuum line manipulations were stored in vacuo over Na/K alloy

in Teflon-valved bulbs. Deuterated solvents were obtained from

Cambridge Isotope Laboratories (all g 99 atom %D), were freeze-

pump-thaw degassed, dried over Na/K alloy, and stored in re-sealable

flasks. Other nonhalogenated solvents were dried over Na/K alloy, and

halogenated solvents were distilled from CaH₂. Methylaluminoxane

(MAO, obtained as a toluene solution from Aldrich) was dried under

high vacuum for 24 h to remove excess volatile aluminum alkyls before

use. [Me₂C(Cp)(Flu)]ZrMe₂(1),⁹B(C₆F₅)₃(3),¹⁵B(2-C₆F₅C₆F₄)₃(4),^8a

Figure 1. High-pressure polymerization reaction system.

measured by DSC (DSC 2920, TA Instruments, Inc.) from the second

scan with a heating rate of 10 °C/min. GPC analyses of polymer samples

were performed at the Dow Chemical Co., Chemical Sciences Catalysis

Laboratory, Midland, Michigan, on a Waters Alliance GPCV 2000 high-

temperature instrument. A polystyrene/polypropylene universal calibra-

tion was carried out using polystyrene standards.

Propylene Polymerization Experiments. Ambient-pressure pro-

pylene polymerizations were carried out on a high vacuum line (10^-6

Torr) in 250 mL round-bottom three-neck Morton flasks equipped with

large magnetic stirring bars, and with rapid stirring (∼1000 rpm) to

minimize mass transfer,¹⁷and thermocouple probes to monitor exotherm

effects.^2cIn a typical experiment, dry toluene (50 mL) was vacuum

transferred into the flask from Na/K, presaturated under 1.0 atm of

rigorously purified propylene, and equilibrated at the desired reaction

temperature using an external water bath. The catalytically active species

was freshly generated in 2∼4 mL of dry toluene in the glovebox.

Control NMR experiments revealed quantitative activation of the

catalyst under these conditions (vide infra). The catalyst solution was

then quickly injected into the rapidly stirred flask using a gastight

syringe. The temperature of the reaction mixture during polymerization

was monitored in real time using a thermocouple thermometer

(OMEGA Type K). The temperature rise was invariably less than

3 °C during these polymerizations, and temperature was controlled by

occasional addition of ice to the external water bath. After a measured

time interval, the reaction was quenched by the addition of 10 mL 2%

acidified methanol. Another 300-400 mL methanol was then added

and the polymer was collected by filtration, washed with methanol,

and dried on the high vacuum line to a constant weight.

Ph₃C⁺B(C₆F₅)₄(5),⁶Ph₃C⁺FAl(2-C₆F₅C₆F₄)₃(6)^8awere prepared

-

according to literature procedures.

Physical and Analytical Measurements. NMR spectra were

recorded on Varian ^UNITYInova-500 (FT, 500 MHz, ¹H; 125 MHz, ¹³C),

1

^UNITYInova-400 (FT, 400 MHz, H; 100 MHz, ¹³C), Mercury-400 (FT

1

400 MHz, H; 100 MHz, ¹³C; 377 MHz, ¹⁹F) or Gemini-300 (FT 300

MHz, ¹H; 75 MHz, ¹³C; 282 MHz, ¹⁹F) instruments. Variable-

temperature measurements were carried out using the ^UNITYInova-400

instrument with a 5-mm inverse probe or 5-mm broadband probe.

Probehead temperature calibration was conducted using methanol and

ethylene glycol standard samples (Varian, Inc.). Chemical shifts for

¹H and ¹³C spectra were referenced using internal solvent resonances

and are reported relative to tetramethylsilane. ¹⁹F NMR spectra were

referenced to external CFCl₃. NMR experiments on air-sensitive samples

were conducted in Teflon valve-sealed NMR tubes (J. Young). For

¹³C NMR analyses of homopolymer microstructures, 300∼400 mg

polymer samples were dissolved in 4 mL C₂D₂Cl₄, heated with a heat

gun in a 10 mm NMR tube, and transferred to the NMR spectrometer

with the probehead pre-equilibrated at 125 °C. A 2.0 s acquisition time

was used with a pulse delay of 6.0 s. A total of 4000-6000 transients

were accumulated for each spectrum. Pentad signals were assigned

according to literature criteria.¹⁶Melting temperatures of polymers were

High-pressure polymerization experiments in toluene solutions were

carried out in a 350 mL heavy wall glass pressure reactor, (Chemglass

Co., maximum pressure, 10 atm) equipped with a septum port, a large

magnetic stirring bar (∼1000 rpm), and an internal thermocouple

(OMEGA Type K), and connected to a high-pressure manifold equipped

with a gas inlet, diaphragm capacitance pressure gauge (0-200 psig),

and gas outlet (Figure 1). CAUTION: All of these procedures should

be performed behind a blast shield. In a typical procedure, in

glovebox, the reactor was charged with dry toluene (50 mL) and the

apparatus was assembled, removed, and then connected to the high-

pressure manifold. Under rapid stirring, rigorously purified propylene

was pressurized into the flask to reach ∼5-6 atm over 5 min and then

(15) Massey, A. G.; Park, A. J. J. Organomet. Chem. 1964, 2, 245-250.

(16) (a) Busico, V.; Cipullo, R.; Monaco, G.; Vacatello, M. Macromolecules

1997, 30, 6251-6263. (b) Pellecchia, C.; Pappalardo, D.; D′Arco, M.;

Zambelli, A. Macromolecules 1996, 29, 1158. (c) Busico, V.; Cipullo, R.;

Corradini, P.; Landriani, L.; Vacatello, M.; Segre, A. L. Macromolecules

1995, 28, 1887. (d) Miyatake, T.; Miaunuma, K.; Kakugo, M. Macromol.

Symp. 1993, 66, 203. (e) Kakugo, M.; Miyatake, T.; Miaunuma, K. Stud.

Surf. Sci. Catal. 1990, 56, 517. (f) Longo, P.; Grassi, A. Makromol. Chem.

1990, 191, 2387. (g) Randall, J. C. J. Polym. Sci., Part B: Polym. Phys.

1975, 13, 889.

9

4608 J. AM. CHEM. SOC. VOL. 126, NO. 14, 2004

Counteranion Effects on Olefin Polymerization

A R T I C L E S

slowly released to 1.0 atm over 5 min. This fill and release process

was repeated five times. The solution was then equilibrated at the

desired propylene pressure (1.0-5.0 atm) and reaction temperature

adjusted using an external water bath. Preparation of the catalytically

active species, temperature control, and reaction quenching were

performed as described for ambient-pressure reactions. Propylene

pressure was then released to 1.0 atm and polypropylene workup was

carried out using the procedure described above. Polymerization

experiments in 1,3-dichlorobenzene or octane solutions were carried

out as described above, but with addition of 50 mL dry 1,3-

dichlorobenzene or octane by cannula through the septum port. Ion

pair complexes were prepared and introduced as described above.

Microstructural Analysis of Polypropylene ¹³C NMR Spectra.

Polymer methyl resonances were assigned according to established

criteria,¹⁶and were analyzed at the pentad level. All polymer NMR

spectra were collected with identical temperature, solvent, instrument

field strength, and acquisition and processing parameters. Steric pentad

distributions were determined from direct integration of the following

regions (ppm): δ 21.91-21.7 (mmmm); 21.63-21.46 (mmmr); 21.43-

21.24 (rmmr); 21.16-20.94 (mmrr); 20.94-20.74 (xrmx); 20.74-20.58

(rmrm); 20.58-19.74 (rrrr + rrrm + mrrm). Pentad distributions were

modeled using the syndiospecific Bernoullian model outlined in Table

16 of ref 11a (p. 1316), having probability parameters P_mand P_mmof

formation for m and mm stereodefects, respectively. These probabilities

were determined by successive nonlinear least-squares minimization

of the function

of complex 9 results in a red-brown oily residue, and the solution

gradually turns brown-green. In a reaction of 1 with 5 at a 4-fold higher

concentration, the red-brown oily residue forms immediately, and

generates mixtures of unidentified complexes. Reagents 1 (2.0 mg, 0.005

mmol) and 6 (Ph₃C⁺FAl(2-C₆F₅C₆F₄)₃^-, 6.5 mg, 0.0050 mmol) were

combined and complete reaction of 1 with rapid formation of [Me₂C-

(Cp)(Flu)]ZrMe⁺FAl(2-C₆F₅C₆F₄)₃^-(10) and Ph₃CCH₃was observed

(NMR data are presented below).

In Situ Generation of Ion Pairs 7-10 for Polymerization Studies.

[Me₂C(Cp)(Flu)]ZrMe₂(1) and the required cocatalyst in a 1:1 ratio

(3, 5, 6) or a 0.5:1 ratio (4) were loaded in the glovebox into a vial

equipped with a septum, and 2-4 mL of toluene was added. The

mixture was shaken vigorously at room temperature for 10 min (3, 4,

6) or 2 min (5) before use. Total amounts used were chosen/refined as

required for temperature control and are reported herein (see Tables 7,

8, and 12-14).

-

Synthesis of [Me₂C(Cp)(Flu)]ZrMe⁺MeB(C₆F₅)₃(7). In the

glovebox, [Me₂C(Cp)(Flu)]ZrMe₂(1, 97.5 mg, 0.250 mmol), B(C₆F₅)₃

(3, 128 mg, 0.250 mmol), and 50 mL toluene were loaded into a 100-

mL reaction flask having a filter frit and stirred for 2 h at room

temperature. The solvent was next reduced in vacuo to 10 mL, and 50

mL pentane was condensed into the flask. The resulting suspension

was filtered, and the collected solid was washed with 5 mL of pentane

and dried under vacuum to afford 174 mg of the title complex; yield,

1

77%. H NMR peak assignments are determined from combined 1-D

and 2-D NMR techniques, and are as follows (labeling outlined in C):

¹H NMR (C₇D₈, 23 °C): δ 7.65 (d, J_H-H) 8.5 Hz, 1 H, H_A), 7.60

(d, J_H-H) 8.5 Hz, 1 H, H_A′), 7.12 (d, J_H-H) 6.3 Hz, 1 H, H_B), 7.05

(d, J_H-H) 8.2 Hz, 1 H, H_D), 6.92 (d, J_H-H) 8.3 Hz, 1 H, H_D′), 6.74

(t, J_H-H) 7.7 Hz, 1 H, H_C), 6.63 (t, J_H-H) 7.1 Hz, 1 H, H_C′), 6.41

(t, J_H-H) 8.2 Hz, 1 H, H_B′), 5.93 (d, J_H-H) 3.0 Hz, 1 H, H_F′), 5.55

(d, J_H-H) 3.0 Hz, 1 H, H_F), 5.20 (d, J_H-H) 3.0 Hz, 1 H, H_E), 4.45

(d, J_H-H) 3.0 Hz, 1 H, H_E′), 1.50 (s, 3 H, Me_R), 1.46 (s, 3 H, Me_R′),

-0.53 (s, br, 3 H, Me_B), -0.92 (s, 3 H, Me_M). ¹⁹F NMR (C₇D₈,

w

(I_exp- I_calcd)²(1 + I_exp

)

∑

wR²)

(3)

2

I_exp

∑

where I_expand I_calcdare experimental and calculated integral values

(normalized to ΣI_exp) 1) with weighting factor w ) 25 for all regions,

with the exception of the rrrr + rrrm + mrrm integral, for which

w ) 0. This weighting scheme increases the contribution to wR²of

stronger signals (having greater S/N ratios), while ensuring that the

rrrr + rrrm + mrrm integral, which is substantially larger than the

3

23 °C): δ -133.39 (d, J_F-F) 22.60 Hz, 6 F, o-F), -159.60 (t,

3

³J_F-F) 20.6 Hz, 3 F, p-F), -164.62 (t, J_F-F) 18.3 Hz, 6 F, m-F).

Anal. Calcd. for C₄₁H₂₄BF₁₅Zr: C, 54.49; H, 2.68. Found: C, 54.37;

H, 2.84.

rest, does not dominate the refinement. Agreement factors calculated

2,13c

according to the standard method, R²) Σ(I_exp- I_calcd)²/Σ(I_exp

)

are

less than 0.001 in all but two cases, with 0.0022 the highest value. I_exp

,

I

_calcd, and weighting multipliers (1 + I_exp)^w, for each experiment, along

with P_m, P_mm, wR², and R²for each set, are given in the Supporting

Information.

Reaction of [Me₂C(Cp)(Flu)]ZrMe₂with B(C₆F₅)₃, B(2-C₆F₅C₆F₄)₃,

Ph₃C⁺B(C₆F₅)₄^-, or Ph₃C⁺FAl(2-C₆F₅C₆F₄)₃^-. [Me₂C(Cp)(Flu)]-

ZrMe₂(1) and cocatalysts (3-6) were loaded into a J. Young NMR

tube and 0.5 mL of toluene-d₈was transferred in via pipet. Each sample

was then shaken vigorously and removed directly to the NMR

spectrometer. Reagents 1 (3.9 mg, 10 µmol) and 3 (B(C₆F₅)₃, 5.1 mg,

10 µmol) were combined, and complete reaction with rapid formation

-

of [Me₂C(Cp)(Flu)]ZrMe⁺MeB(C₆F₅)₃(7) was observed (NMR data

are presented below). Reagents 1 (3.9 mg, 10 µmol) and 4 (B(2-

C₆F₅C₆F₄)₃, 4.8 mg, 5.0 µmol) were combined, and complete reaction

with rapid formation of {([Me₂C(Cp)(Flu)]ZrMe)₂(µ-Me)}⁺MeB(2-

C₆F₅C₆F₄)₃(8) was observed. H and ¹⁹F NMR data for 8 are given

in a previous report.^8aReagents 1 (1.0 mg, 2.5 µmol) and 5

(Ph₃C⁺B(C₆F₅)₄^-, 2.3 mg, 2.5 µmol) were combined, and complete

reaction with rapid formation of [Me₂C(Cp)(Flu)]ZrMe⁺B(C₆F₅)₄^-(9)

and Ph₃CCH₃was observed. The C₆H₄signals of the fluorenyl region

could not be assigned completely due to overlap of the signals with

the solvent. ¹H NMR for 9 (C₇D₈, 23 °C): δ 7.8 (d, 1 H, C₆H₄), 7.6 (d,

1 H, C₆H₄), 5.742 (m, 1 H, C₅H₄), 4.824 (m, 1 H, C₅H₄), 4.435 (m, 1

H,C₅H₄), 3.813 (m, 1 H, C₅H₄), 1.667 (s, 3 H, CMe₂), 1.477 (s, 3 H,

CMe₂), -1.142 (s, 3 H, Zr-CH₃). ¹⁹F NMR (C₇D₈, 23 °C): δ -132.06

(m, o-F), -132.49 (m, o-F), -162.77 (t, ³J_F-F) 21.5 Hz, p-F), -163.0

(m, p-F), -166.72 (m, m-F), -166.96 (m, m-F). Prolonged standing

Synthesis of [Me₂C(Cp)(Flu)]ZrMe⁺FAl(2-C₆F₅C₆F₄)₃^-(10). In

the glovebox, [Me₂C(Cp)(Flu)]ZrMe₂(1, 97.5 mg, 0.250 mmol), Ph₃C⁺-

FAl(2-C₆F₅C₆F₄)₃^-(6, 328 mg, 0.250 mmol), and 100 mL toluene were

loaded into a 250 mL reaction flask having a filter frit, and stirred for

2 h at room temperature. The solvent was next reduced in vacuo to 10

mL, and 100 mL pentane was condensed into the flask. The resulting

suspension was filtered, and the collected solid was washed with 20

mL of pentane and dried under vacuum to afford 280 mg of the title

complex; yield, 82%. As measured from ¹H spectra, a pair of

diastereomers is evident in a 1.6:1 ratio at 23 °C. Assignment of the

¹H NMR spectrum was accomplished with combined NOE, EXSY,

and COSY techniques; atom labeling is described in C. Major and minor

-

1

9

J. AM. CHEM. SOC. VOL. 126, NO. 14, 2004 4609

A R T I C L E S

Chen et al.

Table 1. Summary of the Crystal Structure Data for Complexes 7

and 10^a

diastereomers are differentiated with upper- and lower-case subscripts,

respectively. Certain of the C₆H₄signals of the fluorenyl region could

not be clearly assigned due to overlap between the signals of the two

complex

7

10

1

isomers. H NMR (C₇D₈, 23 °C) for major diastereomer: δ 7.99 (d,

formula

formula weight

C₄₈H₃₂BF₁₅Zr

995.77

C₅₈H₂₁Al F₂₈Zr

1367.95

J_H-H) 8.0 Hz, 1 H, H_A′), 7.90 (dm, J_H-H) 4.0 Hz, 1 H, H_A), 7.20

(m, 2 H, H_B′and H_D), 7.00 (1 H, H_D′), 6.78 (m, 2 H, H_Cand H_C′), 6.20

(s, 1 H, H_F′), 6.08 (t, J_H-H) 8.0 Hz, 1 H, H_B), 5.44 (m, 1 H, H_E), 4.85

(s, 1 H, H_F), 4.61 (m, 1 H, H_E′), 1.61 (s, 3 H, Me_R), 1.44 (s, 3 H,

crystal color, habit

red, plate

red, block

crystal dimensions (mm) 0.284 × 0.178 × 0.044 0.194 × 0.174 × 0.166

crystal system

space group

a, Å

b, Å

c, Å

monoclinic

P2₁/c

monoclinic

P2₁/c

Me_R′), -1.03 (s, 3 H, Me_M). Minor diastereomer: δ 7.65 (d, J_H-H

)

13.5302 (18)

26.815 (4)

12.4684 (16)

116.673 (2)

4042.3 (9)

4

1.636

0.378

0.91577-0.98342

35 835

16.4939 (10)

19.6187 (12)

16.9722 (10)

112.4710 (10)

5075.0 (5)

4

1.790

0.380

0.925 39-0.952 86

46 570

8.4 Hz, 1 H, H_a′), 7.55 (dm, J_H-H) 8.0 Hz, 1 H, H_a), 7.20 (m, 1 H,

H_d), 7.09 (1 H, H_b′), 7.00 (2 H, H_cand H_d′), 6.74 (m, 1 H, H_c′), 6.32 (s,

1 H, Hf ′), 5.99 (t, J_H-H) 8.0 Hz, 1 H, H_b), 5.44 (m, 1 H, H_e), 5.00 (s,

1 H, H_f), 4.58 (m, 1 H, H_e′), 1.65 (s, 3 H, Me_r), 1.50 (s, 3 H, Me_r′),

-1.07 (s, 3 H, Me_m). ¹⁹F NMR (C₇D₈, 23 °C) for major diastereomer:

δ -113.62 (s, br, 3F), -133.90 (m, 3F), -134.60 (s, br, Al-F),

â, deg

V, Å³

Z

d (calc), g/cm³

µ, mm^-1

T_min-T_max

measured reflections

-138.04 (m, 3F), -139.24 (t, ³J_F-F) 21.5 Hz, 3F), 153.27 (t, ³J_F-F

)

19.8 Hz, 6F), 154.87 (m, 3F), 161.38 (m, 3F), 163.03 (t, ³J_F-F) 21.2

Hz, 3F). Minor diastereomer: δ 116.01 (s, br, 3F), -132.42 (s, br,

independent reflections 9692

12346

3

reflections > 2σ (I)

4357

8033

0.0705

0.0503

0.1374

0.981

796

Al-F), -133.90 (m, 3F), -138.68 (m, 3F), -139.55 (t, J_F-F) 18.9

R_int

0.1409

0.0548

0.1344

0.880

601

3

Hz, 3F), 153.52 (t, J_F-F) 21.2 Hz, 3F), 153.68 (t, J_F-F) 23.7 Hz,

3F), 153.89 (m, 3F), 161.22 (dd, J_F-F) 21.2, 7.6 Hz, 3F), 162.84 (t,

³J_F-F) 23.7 Hz, 3F). Anal. Calcd for C₅₈H₂₁AlF₂₈Zr: C, 50.92; H,

1.55. Found: C, 50.64; H, 1.73.

R[F²> 2σ (F²)]

wR(F²)

S

no. of parameters

X-ray Crystal Structure Determinations of [Me₂C(Cp)(Flu)]Zr-

^aCCD area detector diffractometer; æ and ω scans; temperature for data

collection, 153 (2) K; Mo KR radiation; λ ) 0.710 73 Å.

Me⁺MeB(C₆F₅)₃(7) and [Me₂C(Cp)(Flu)]ZrMe⁺FAl(2-C₆F₅C₆-

-

F₄)₃^-(10). Crystals of the title complexes suitable for X-ray diffraction

were obtained by slow diffusion of pentane into toluene solutions at

0 °C. Crystals were selected and mounted under Infineum V8512 oil,

and held under a nitrogen cold-stream at 153(2) K for data collection.

Diffraction data were obtained using a Bruker SMART 1000 CCD area

detector diffractometer with a fine-focus, sealed tube Mo KR radiation

source (λ ) 0.71073 Å) and graphite monochromator. For both 7 and

10, the initial crystal structure solution was obtained via Patterson

synthesis, refined through successive least-squares cycles, and subjected

to a face-indexed absorption correction. The refinements were carried

to convergence, with hydrogen atoms placed in idealized positions and

refined isotropically with fixed U_equnder standard riding model

constraints, with the following exception: in complex 7, hydrogen

atoms H₃C-B were refined isotropically with group thermal, H-C

distance, and H-H distance parameters. Crystal data collection and

refinement parameters are summarized in Table 1 and can be found in

the Crystallographic Information File (CIF, see Supporting Information).

t₂dimension and combined Gaussian weighting and 1 Hz line-

broadening in the t₁dimension, unless otherwise noted. Rate constant

calculations are described in the discussion.¹⁹For 7 (τ_m) 600 ms,

20.5 °C), quadrupolar relaxation of the ¹¹B- and ¹⁰B-coupled Me_B

protons precludes accurate determination of rates for exchange involving

these resonances using this technique. For 10, exchange rates between

resonances H_Aand H_aand between H_A′and H_a′were averaged to

determine anion racemization rates at 87.5 °C (τ_m) 185 ms) and 117.5

°C (τ_m) 40 ms). Anion racemization rate constants calculated from

EXSY data are in good agreement with values obtained from line shape

analysis (vide infra). At 127 °C, τ_mwas optimized to determine, or to

place a higher limit on, the rate of ion pair reorganization (τ_m) 800ms).

In this case, 14 000 real points in the t₂dimension and 256 points in

the t₁dimension were collected, and the data were processed with no

zero-filling and 1 Hz line broadening in t₂, and with linear prediction

to 512 points, zero-filling to 8192 points, 1 Hz line broadening, and

0.036 s Gaussian weighting in t₁. With 10, cross-peaks corresponding

to ion pair reorganization are absent at all temperatures measured,

invariantly with mixing time and data processing parameters. An upper

limit for ion pair reorganization is established as described in the

Discussion Section.

2D EXSY NMR Studies of Ion Pair Reorganization/Symmetri-

zation in 7 and 10. Toluene-d₈solutions of pure [Me₂C(Cp)(Flu)]-

ZrMe⁺MeB(C₆F₅)₃(7, 4.0 mg, 5.5 µM) or [Me₂C(Cp)(Flu)]ZrMe⁺

-

FAl(2-C₆F₅C₆F₄)₃^-(10, 7.0 mg, 6.4 µM) were prepared in the glovebox,

and filtered directly into J. Young NMR tubes. Spectra were collected

using the NOESY pulse sequence,¹⁸with acquisition parameters

optimized to resolve peaks of interest. Mixing times τ_m) 40-800

DNMR Studies of Ion Pair Reorganization/Symmetrization in 7

and 10 in Toluene-d₈. Pure [Me₂C(Cp)(Flu)]ZrMe⁺MeB(C₆F₅)₃^-(7,

ms, were chosen to minimize the error in calculated exchange rates,

8.0 mg, 8.8 µmol) or [Me₂C(Cp)(Flu)]ZrMe⁺FAl(2-C₆F₅C₆F₄)₃(10,

-

-1

according to τ_m) (T₁+ k_AB+ k_BA)

^-1, where k_ABand k_BAare

7.0 mg, 5.1 µmol) were loaded in the glovebox into capped vials, and

0.80 mL of a stock solution of CH₃Si(C₆H₅)₃(internal standard, 11

mM for 7 and 6.4 mM for 10) in toluene-d₈was transferred into each

vial. The resultant solutions were filtered and transferred into J. Young

NMR tubes. Temperatures were varied over the range 0-92.5 °C for

7, and over 23-132.5 °C for 10. Prior to each acquisition, the NMR

probehead was preequilibrated at the desired temperature for 10 min.

Each acquisition consisted of 65536 points spanning 4360 Hz (resolution

0.067 Hz), and 4908 Hz (resolution 0.075 Hz), for 7 and 10,

respectively. The raw data were zero-filled to 2 × np. Unweighted

transforms for both 7 and 10 were phased carefully and subjected to

reference deconvolution on the methyl resonance of triphenylmethyl-

silane as the internal line shape standard using the Hilbert algorithm,²⁰

estimates of the AfB and BfA exchange rate constants, respectively.¹⁹

Data were zero-filled to 2 × np and 4 × ni in the t₂and t₁dimensions,

respectively, and apodized using appropriate Gaussian weighting in the

(17) At 20 °C, rate of C₃H₆absorption is estimated 0.029 mol/min in toluene at

1.0 atm of C₃H₆, and propylene mass transfer effects (mass transport

coefficient) in the (2-PhInd)₂ZrCl₂/ MAO system in toluene (100 mL) are

observed to be insensitive to the presence of up to 4 g of isotactic PP with

a maximum stirring speed (1460 rpm), See: Lin, S.; Tagge, C. D.;

Waymouth, R. M.; Nele, M.; Collins, S.; Pinto, J. C. J. Am. Chem. Soc.

2000, 122, 11 275-11 285. In the present study, the most active ion pair

system, 5 (9) at 40 °C and 1.0 atm of C₃H₆, has a maximum rate of

propylene consumption of ∼0.015 mol/min (0.77/42*(75/60)), which should

be lower at 20 °C because lower activity is observed at lower temperatures.

Thus, propylene mass transfer effects should be negligible for all catalysts

under the present conditions.

(18) Macura, S.; Ernst, R. R. Mol. Phys. 1980, 41, 95-117.

(19) Perrin, C. L.; Dwyer, T. J. Chem. ReV. 1990. 90, 935-967, and also see

ref 14.

(20) Rutledge, D. N., Ed. Signal Treatment and Signal Analysis in NMR; Elsevier

Science: New York, 2003; Ch. 16.

9

4610 J. AM. CHEM. SOC. VOL. 126, NO. 14, 2004

Counteranion Effects on Olefin Polymerization

A R T I C L E S

Scheme 2. X^-, X′^-represent alternate stereochemical

100 mL toluene solutions in the high-pressure reaction vessel as

described above. The catalytically active species was freshly generated

in 2-10 mL of dry toluene using [Me₂C(Cp)(Flu)]ZrMe₂samples

(1, 1.0 mg, 2.5 µmol; 2.0 mg, 5.1 µmol; 4.0 mg, 10 µmol; 8.0 mg,

20 µmol; 16.0 mg, 40.8 µmol; 32.0 mg, 81.6 µmol) activated with

1.0 equivalents of B(C₆F₅)₃(3).

configurations at Al^a

Propylene Polymerization Catalyzed by 1 + 3 with Added Li⁺

MeB(C₆F₅)₃^-. Polymerization experiments were carried out in 100 mL

toluene solutions in the high-pressure reaction vessel described above.

-

Li⁺MeB(C₆F₅)₃was prepared in situ by mixing a 1:1 molar ratio of

dry LiMe powder and B(C₆F₅)₃in toluene, and the mixture was shaken

vigorously at room temperature for 30 min before use. ¹H NMR (C₇D₈,

23 °C): δ 0.79 (s, 3 H, Me). ¹⁹F NMR (C₇D₈, 23 °C): δ -136.84

(d, br, J_F-F) 23.0 Hz, 6 F), -159.39 (t, J_F-F) 19.6 Hz, 3 F), -163.03

(t, J_F-F) 19.6 Hz, 6 F). The catalytically active species was freshly

generated in 2∼4 mL toluene, as described above. The mixture was

-

then combined with the corresponding Li⁺MeB(C₆F₅)₃solution in

toluene, shaken vigorously at room temperature for 3 min, and then

injected immediately into the polymerization reactor.

Propylene Polymerization Catalyzed by 1 + MAO (2). Polym-

erization experiments were carried out in the high-pressure reaction

vessel described above. In a typical polymerization, [Me₂C(Cp)(Flu)]-

ZrMe₂(1, 3.9 mg) and MAO (2, 40 mg) were loaded in the glovebox

into a septum-capped vial, to which 4 mL of toluene was added. The

mixture was shaken vigorously at room temperature for 20 min before

use. In the polymerization reaction flask, MAO (2, 80 mg) was loaded

with 50 mL of toluene. The solution was then equilibrated at the desired

polymerization temperature and pressure as described above.

^aThe symbols used to represent these configurations are intended to

convey a qualitative picture of anion racemization (see Figure 4).

along with baseline and drift corrections, such that the final standard

peak width was 1.500 Hz in all spectra. Modeling of the ²⁹Si satellites

of the reference signal was included in the reference deconvolution

(²J_Si-H) 6.633 Hz). Application of reference deconvolution was found

to significantly improve variances, both for borane migration and for

ion pair reorganization, compared to line fits generated from use of

Results and Discussion

the approximation for half-height signal widths, W_signal) W_real

+

∆W_natural+ ∆W_exchange. For 7, broadening of the proton signals of the

diastereotopic i-Pr methyl groups (Me_R, Me_R′) and of the zirconocenium

methyl group (Me_M) were monitored over the temperature range, 57.8-

92.3 °C. Rate constants at each temperature were calculated from the

half-height widths of these signals (measured using the VNMR

command, dres) as compared to their widths in the slow-exchange limit

(0 °C).²¹Values and confidence intervals for ∆H^‡, ∆S^‡, and ∆G^‡were

determined from linear regression analysis of ln(k/T) vs 1/T, and are

reported at the 90% confidence level (Table 6).

For complex 10, spectra were recorded over the temperature range

78.5-132.5 °C, and referenced to a spectrum collected at 23 °C.

Complete line shape analysis (CLSA)²²of these spectra converged for

an exchange protocol including only diastereomer interconversion via

anion racemization (Scheme 2), but failed to converge at all temper-

atures when a rate parameter for ion pair reorganization was included.

Coalescence of the i-Pr bridge methyl signals at δ 1.61 (Me_R) and δ

1.65 (Me_r) of the two diastereomers was observed at 127.5 °C.

DNMR Studies of Ion Pair Reorganization/Symmetrization in

8, 9. Ion pairs 8 and 9 were prepared in situ from [Me₂C(Cp)(Flu)]-

ZrMe₂(1, 3.9 mg, 10 µmol for 8, or 1.0 mg, 2.5 µmol for 9) and B(2-

C₆F₅C₆F₄)₃(4, 4.8 mg, 5.0 µmol) or Ph₃C⁺B(C₆F₅)₄^-(2.3 mg, 2.5 µmol)

with 0.5 mL of o-xylene-d₁₀. Decomposition of 8 begins at ∼80 °C.

Complex 8 also decomposes rapidly in toluene at 115 °C on a time

scale of ∼10 min to give a deep blue-purple precipitate. Immediate

decomposition of 9 was detected at 80 °C and broadening of the

diastereotopic methyl signals on the i-Pr bridge could not be clearly

observed.

The following analysis couples complementary methods for

studying ion pairing effects in polymerization systems, and is

presented in three parts. First, a description of the preparation

and characterization of the in situ generated and isolated active

catalytic species is presented, with detailed structural and

solution/dynamic characterization of 7 and 10, with a discussion

extending these results to polymerization behavior. Second, a

detailed examination is presented of the effects of varying the

cocatalyst/counteranion on polymerization dynamics and product

polymer characteristics, as functions of temperature, monomer

concentration, and solvent. Using these results, a comparative

kinetic treatment is derived which provides a self-consistent

model for the effects of ion pairing/counteranion identity on

polymerization behavior and product polymer attributes. Finally,

we present an examination of putative catalyst and counteranion

concentration effects on syndioselection and other product

polymer characteristics.

This discussion focuses on the importance of ion-pairing

dynamics. The cation-anion interaction is recognized to have

both electrostatic and covalent components;^23,24thus the poten-

tial barrier to ion pair reorganization in the isolated catalyst

systems discussed below, or analogous site epimerization

processes operative during polymerization, may have both

electrostatic and covalent/coordinative components. Although

there are doubtless differences in the relative magnitudes of the

coordinative/covalent and electrostatic contributions to the ion

pair reorganization barrier among the various ion pair complexes

studied here, we do not distinguish between these components

Concentration Dependence Study of Propylene Polymerization

Catalyzed by 1 + 3. Polymerization experiments were carried out in

(21) (a) Sandstrom, J. Dynamic NMR Spectroscopy; Academic Press: New York,

1982; pp 77-92, and also see ref 4a. (b) Here k is the rate constant in s^-1

;

∆W ) W₂, - W₁, where W₂is line width at half-height of the exchange

broadened peak in Hz, and W₁is the line width in the absence of exchange

(the no-exchange limit, 0 °C for 7 and 23 °C for 10). The corresponding

free energies of activation can also be derived using ∆G^‡) - RT[ln(k/T)

+ ln(h/k)].

(23) Strauss, S. H. Chem. ReV. 1993, 93, 927-942.

(24) (a) Lanza, G.; Fragala, I. L.; Marks, T. J. Organometallics 2002, 21, 5594-

5612. (b) Lanza, G.; Fragala, I. L.; Marks, T. J. Organometallics 2001,

20, 4006-4017. (c) Lanza, G.; Fragala, I. L.; Marks, T. J. J. Am. Chem.

Soc. 2000, 122, 12 764-12 777.

(22) Budzelaar, P. H. M. gNMR V. 4.1.0; Adept Scientific plc, 1999

9

J. AM. CHEM. SOC. VOL. 126, NO. 14, 2004 4611

A R T I C L E S

Chen et al.

using the present experimental results, and refer to kinetic

inertness of the ion pair toward reorganization as “ion pairing

strength,” or “coordinative tendency.”

Interactions between the cationic and anionic portions of ion-

pair complexes 7-10 can be evaluated qualitatively from

1

ambient temperature 1-D H and ¹⁹F NMR spectroscopy. The

¹H NMR spectra of these compounds are straightforward, with

four distinct resonances at δ 4-6 ppm for the C₅H₄ring and

two methyl signals at δ 1.5-2.0 ppm for the ansa-isopropyl

group, in accord with the dissymmetry generated by the ion

pairing. The Zr-CH₃proton signals in 7, 9, and 10 invariably

appear at ∼δ -1 ppm; in contrast, the Zr-CH₃signals of 8

appear at -1.21 and -1.27 (for major and minor diastereomers,

respectively), and the µ-Me (bridging Zr-CH₃-Zr) proton

resonance of 8 appears in the region typical of this chemical

environment (δ -3.33 ppm). As previously shown, metalloce-

I. Zirconocenium Cations Generated via Reaction of

[Me₂C(Cp)(Flu)]ZrMe₂with Cocatalysts 3-6. The substantial

body of available structural and spectroscopic data on complexes

7-10 permits qualitative and quantitative evaluation of cation-

anion interactions. These interactions exhibit diverse structural/

dynamic behavior, which is quantified using X-ray diffraction

and dynamic NMR spectroscopy. The purpose of the following

discussion is to highlight key structural and kinetic features of

these systems, and to set the stage for correlation with

polymerization characteristics.

-

nium cation-MeB(2-C₆F₅C₆F₄)₃interactions are considerably

A. Synthesis and Spectroscopy. Under rigorously anhydrous/

anaerobic conditions, [Me₂C(Cp)(Flu)]ZrMe₂(1) undergoes

reaction with B(C₆F₅)₃(3), B(2-C₆F₅C₆F₄)₃(4),^8aPh₃C⁺B(C₆F₅)₄^-

weaker than those involving MeB(C₆F₅)₃^-, and the equilibrium

solution structure of 8 argues that neutral [Me₂C(Cp)(Flu)]ZrMe₂

(5), and Ph₃C⁺FAl(2-C₆F₅C₆F₄)₃(6)^8ato generate the corre-

-

has a greater affinity for the cation than does MeB(2-C₆F₅C₆-

- 5a,8a

F₄)₃

.

Thus, the relative coordinative tendency of these

sponding ion pairs (7-10; eq 4) and to afford highly active

olefin polymerization catalysts.²⁵Except for 9, these ion pairs

can be isolated and characterized by standard 1-D and 2-D ¹H/

¹⁹F NMR, and analytical techniques (see Experimental Section

for details); 7 and 10 have been further characterized by single-

crystal X-ray diffraction (vide infra). The reaction of cocatalyst

methyl fluoroarylborate anions versus neutral [Me₂C(Cp)(Flu)]-

ZrMe₂with respect to the [Me₂C(Cp)(Flu)]ZrMe⁺cation follows

-

the order MeB(C₆F₅)₃> [Me₂C(Cp)(Flu)]ZrMe₂> MeB(2-

-

C₆F₅C₆F₄)₃

.

In a prior communication, we observed that ion pairing

interactions between metallocenium cations and fluoroaryl

counteranions significantly influence fluoroaryl ¹⁹F NMR

chemical shifts.^8aThus, the ¹⁹F spectrum of complex 9 shows

no substantial chemical shift differences from that of Ph₃C⁺

B(C₆F₅)₃(3) or Ph₃C⁺FAl(2-C₆F₅C₆F₄)₃(6) with 1 cleanly

-

generates the monomeric ion pairs 7 and 10 in good isolated

yield. In contrast, B(2-C₆F₅C₆F₄)₃(4) preferentially yields

cationic µ-Me bridged dinuclear complex 8 as diastereomers

in a ratio of 1.8:1 (D and E, depicted below), even with

stoichiometric excesses of reagent 4 and long reaction times.^8a

-

B(C₆F₅)₄(5) at room temperature, suggesting that B(C₆F₅)₄

coordination to Zr⁺is weak and labile. Conversely, chemical

shift evidence for strong cation-anion interactions in 10 in

solution is readily detected in the ¹⁹F NMR spectra. Although

The reaction of cocatalyst Ph₃C⁺B(C₆F₅)₄(5) with 1 affords

-

1

ion pair 9 as suggested by H NMR (along with 1.0 stoichio-

-

Ph₃C⁺FAl(2-C₆F₅C₆F₄)₃(6) exhibits seven ¹⁹F signals (1:1:

metric equivalents of Ph₃CCH₃), with 9 being the least stable

of the present four ion pairs. Attempts to isolate or crystallize

this complex have been unsuccessful, and have resulted in dark

oily residues and a yellow-green solution (similar behavior is

1:1:1:2:2), 10 exhibits nine fluoroaryl signals, indicative of

restricted internal C₆F₅rotation but free anion rotation about

the Zr-F-Al linkage. The existence of 10 as diastereomers in

toluene-d₈solution, together with other structural and dynamic

data, demonstrates that the mutual o-perfluorobiphenyl group

orientations impart chirality to the Al center in solution, as seen

in the solid state (vide infra). The ¹⁹F NMR spectrum of 10

exhibits a characteristic broad F-Al resonance at δ -132.2

ppm, which, compared to the F-Al chemical shift of Ph₃C⁺

-

observed for most known group 4 metallocenium B(C₆F₅)₄

complexes).^7c

-

FAl(2-C₆F₅C₆F₄)₃(δ -175.60 ppm), demonstrates a strong

M⁺‚‚‚F-Al- interaction, and is consistent with a (time-

averaged) preferred orientation of the fluoroaluminate ion with

respect to the cation. Diffraction data clearly confirm the

-

coordination of the FAl(2-C₆F₅C₆F₄)₃anion via the Zr-F-

(25) NMR-scale reaction of 1 with MAO is not amenable to interpretation, thus

will not be discussed here.

Al bridge in the solid-state structure of 10.

9

4612 J. AM. CHEM. SOC. VOL. 126, NO. 14, 2004

Counteranion Effects on Olefin Polymerization

A R T I C L E S

Table 2. Selected Bond Distances (Å) and Angles (deg) for Complex 7

bond distances (Å)

B-C1B

C7-Zr

C18-Zr

C22-Zr

B-C29

1.652(7)

2.615(4)

2.429(4)

2.433(4)

1.651(7)

Zr-C1B

C8-Zr

C19-Zr

C14-C15

B-C23

2.521(4)

C1-Zr

2.248(4)

2.524(4)

2.521(4)

1.512(6)

1.366(6)

C2-Zr

2.480(4)

2.414(4)

2.515(4)

1.650(7)

2.652(4)

2.446(4)

1.553(6)

1.669(7)

C13-Zr

C20-Zr

C15-C18

C24-F2

C14-Zr

C21-Zr

B-C35

bond angles (deg)

B-C1B-Zr

165.5(3)

C1-Zr-C1B

C1B-B-C29

C35-B-C29

C14-Zr-C18

94.15(17)

C1B-B-C23

108.2(4)

107.1(4)

99.5(3)

C1B-B-C35

C35-B-C23

C16-C15-C17

107.7(4)

112.8(4)

106.2(4)

110.1(4)

111.0(3)

57.76(15)

C29-B-C23

C14-C15-C18

covalent character of the Zr-Me_Binteraction. In comparison

-

with reported analogous zirconocenium MeB(C₆F₅)₃ion pair

crystal structures (F-J),^27a-ethe present result affords the

shortest B-Me_B(0.024Å shorter than the average of F-I)^27fand

Zr-Me_Bbond distances observed to date (0.047 Å shorter than

the average), and the largest mean C(C₆F₅)-B-Me_Bangle (1.8°

larger than the average; Table 3), suggesting that the covalent

character of the action-anion interaction, whereas evident, is

least in the present case.

The observed Cp(centroid)-Zr-Cp(centroid) angle (bite

angle) for 7 (118.6°), as compared with F (127.0°),^27amay be

correlated with closer proximity of the bridging methyl carbon

and counteranion boron atoms to the metal in structure 7.

However, the steric bulk of the ancillary ligand structure and

difference in backbone composition in F also possibly contribute

to the observed differences in Zr-Me_B-B(C₆F₅)₃geometry.

Figure 2. Perspective ORTEP drawing of the molecular structure of the

complex [Me₂C(Cp)(Flu)]ZrMe⁺MeB(C₆F₅)₃^-(7). Thermal ellipsoids are

drawn at the 30% probability level.

The crystal structure of 10 shows the [Me₂C(Cp)(Flu)]ZrMe⁺

moiety in close contact with sterically congested FAl(2-

-

C₆F₅C₆F₄)₃through a Zr-F-Al bridge (Figure 3). Important

B. X-ray Crystal Structures of [Me₂C(Cp)(Flu)]ZrMe⁺

bond distances and bond angles of 10 are summarized in Table

4. For 10, the Zr-F-Al (162.21(10)°) and F-Zr-Me

(92.65(10)°) bond angles, as well as the Zr-Me (2.245(3)Å)

and Al-F (1.7858(17)Å) bond distances are reminiscent of those

MeB(C₆F₅)₃(7), and [Me₂C(Cp)(Flu)]ZrMe⁺FAl(2-C₆F₅-

-

C₆F₄)₃(10). Attempts were made during the course of this

study to grow single-crystal samples of complexes 7-10, and

crystals of more stable ion pairs 7 and 10 suitable for X-ray

analysis were obtained.²⁶The structure of 7 shows the [Me₂C-

(Cp)(Flu)]ZrMe⁺cation in contact with the counteranion through

in [rac-Me₂Si(Ind)₂ZrMe]⁺FAl(2-C₆F₅C₆F₄)₃(166.5(8)°,

-

90.8(6)°, 2.24(2)Å, 1.81(1)Å, respectively).^8aThe anion in 10

adopts a pseudotetrahedral geometry, with the C₆F₅-C₆F₄

torsion angles substantially divergent from 90° (72.3° on

average; ranging from 70.6° to 79.0°). In comparison with trityl

salt 6, which shows no cation-anion coordinative interaction

in the solid state,^8a10 exhibits a much longer Al-F bond

distance (1.786(2)Å vs 1.682(5)Å) and much smaller average

of the three F-Al-C₁₂F₉bond angles (103.0(1)° vs 107.7(3)°),

demonstrating that the impact of the zirconocenium cation on

the structure of the fluoroaluminate anion is large in comparison

to the cation influence on the anion structure in 7.

-

the MeB(C₆F₅)₃methyl group (Figure 2). Important bond

distances and angles for 7 are summarized in Table 2. The Zr-

Me_B-B bridge is nearly linear (bond angle 165.5(3)°). This

interaction has been shown by ab initio calculations to be

predominantly electrostatic in nature.²⁴The Me_B-Zr-Me_Mbond

angle is 94.15(17)°, with the Zr-Me_Mdistance (2.248(4)Å),

significantly shorter than the Zr-Me_Bdistance (2.521(4)Å). In

-

comparison with the noncoordinating Me_BB(C₆F₅)₃anion in

previously reported structure {([Me₂C(Cp)(Flu)]Zr(C₆F₅))₂-

- 8a

(µ-F)}⁺MeB(C₆F₅)₃

,

the slightly longer Me_B-B bond

Direct comparison of the cation structures in complexes 7

and 10 shows a subtle relationship between counteranion identity

and Zr environment. In 10, the larger Cp(centroid)-Cp(flu,

centroid) distance (3.796(11)Å vs 3.763(14)Å), greater metal-

distance (1.652(7)Å vs 1.64(2)Å) and smaller mean C(C₆F₅)-

B-Me_Bangle (108.7(4)° vs 111.4(9)°) show the effect on anion

structure in 7 due to cation-anion interaction. This observed

coordination-induced lengthening of the B-Me_Bbond and

flattening of the B(C₆F₅)₃substructure (compared to uncoordi-

nated MeB(C₆F₅)₃^-) possibly reflect the degree to which the

cation and anion share the Me_Bmoiety, hence the degree of

(27) (a) F, [Me₂Si(Cp′)₂]ZrMe⁺MeB(C₆F₅)₃^-; Cp′ ) C₅H₂(Me)(t-Bu), ref 14.

(b) G, (Cp)₂ZrMe⁺MeB(C₆F₅)₃Guzei, I. A.; Stockland, R. A.; Jordan,

-

,

R. F. Acta Crystallogr., Sect. C (Cryst. Str. Comm.) 2000, C56, 635-636.

(c) H, [Me₄C₂(Cp)₂]ZrMe⁺MeB(C₆F₅)₃^-, Beck, S.; Prosenc, M. H.;

Brintzinger, H. H.; Goretzki, R.; Herfert, N.; Fink, G. J. Mol. Catal. A:

Chem. 1996, 111, 67-79. (d) I, [(1,3-C₅H₃R₂)₂]ZrMe⁺MeB(C₆F₅)₃

;

-

(26) When a toluene solution of complex 7 was left standing at room temperature

for two weeks, crystals of decomposition product {[Me₂C(Flu)(Cp)Zr-

R ) SiMe₃, ref. 4a, also see: Bochmann, M.; Lancaster, S. J.; Hursthouse,

M. B.; Malik, K. M. A. Organometallics 1994, 13, 2235-2243. (e) J, [(1,2-

C₅H₃Me₂)₂]ZrMe⁺MeB(C₆F₅)₃^-, ref. 4a. (f) Average of the five complexes,

F-J.

-

(C₆F₅)]₂(µ-F)}⁺MeB(C₆F₅)₃were obtained, as reported previously, see

ref 8a.

9

J. AM. CHEM. SOC. VOL. 126, NO. 14, 2004 4613

A R T I C L E S

Chen et al.

-

Table 3. Comparison of Selected Bond Distances (Å) and Angles (deg) in Complex 7 to Those of Analogous Zirconocenium-MeB(C₆F₅)₃

Ion Pairs²⁷

^aAverage of structures F-J. ^bMean B-C(C₆F₅) distance. ^cMean Me_M-B-C(C₆F₅) angle. ^dBite angle.

indicate a stronger coordinative interaction with the bulkier but

-

more strongly donating FAl(2-C₆F₅C₆F₄)₃anion (Table 5).

Note that in 10, π-π stacking is also observed among the C₁₂F₉

groups of the anion, where the C₆F₄rings bound to the Al center

engage in stacking with the end C₆F₅rings on adjacent C₁₂F₉

groups. The sense of the corkscrew motif described by the C₁₂F₉

groups determines the stereochemical configuration at the Al

center. In contrast, this π-π stacking interaction is conspicu-

ously absent in the solid-state structure of trityl salt 6. This,

together with the observed interconversion of diastereomers of

10 in solution (vide infra), suggests a subtle reciprocation

between ancillary ligand architecture and structural dynamics.^8a

C. Solution Dynamics of Ion Pair Reorganization/Sym-

metrization in [Me₂C(Cp)(Flu)]ZrMe⁺MeB(C₆F₅)₃^-(7) and

-

[Me₂C(Cp)(Flu)]ZrMe⁺FAl(2-C₆F₅C₆F₄)₃(10). Exchange

processes available to ion pairs 7 and 10 in the absence of olefin

can be correlated with polymerization behavior of these catalyst

systems. The principal structural reorganization process of

Figure 3. Perspective ORTEP drawing of the molecular structure of the

complex [Me₂C(Cp)(Flu)]ZrMe⁺FAl(2-C₆F₅C₆F₄)₃(10). Thermal el-

lipsoids are drawn at the 30% probability level. The terminal C₆F₅groups

not only twist out of coplanarity with connected C₆F₄fragments but also

exhibit π-π interactions with the C₆F₄groups on adjacent C₁₂F₉ligands.

interest in each of these systems is migration of the anionic

portion of the catalyst-cocatalyst system from one side of the

zirconocenium-methyl metal center to the other (ion pair

reorganization). This process mirrors the site epimerization of

the zirconocenium-polymeryl-anion ensemble thought to occur

during polymerization and to give rise to product m stereodefects

in the absence of synchronous propylene enchainment (eq 2;

Scheme 1B). In the absence of olefin, both 7 and 10 undergo

background exchange processes as well, and ion pair reorga-

nization must be studied in the context of all extant reorganiza-

tion processes.

-

ligand distances (Zr-Cp(centroid), 2.176(3)Å vs 2.157(4)Å;

Zr-C(i-Pr bridging), 3.124(3)Å vs 3.109(4)Å), and greater

C(bridgehead, Cp)-C(i-Pr bridging) bond distance (1.530(5)Å

vs 1.512(6)Å) in 10, reveal: (a) that the Zr center is displaced

slightly out of the [Me₂C(Cp)(Flu)]^2-ligand pocket in 10 as

compared to 7, and (b) that the [Me₂C(Cp)(Flu)]^2-ligand is

pried open by FAl(2-C₆F₅C₆F₄)₃^-. The above observations

9

4614 J. AM. CHEM. SOC. VOL. 126, NO. 14, 2004

Counteranion Effects on Olefin Polymerization

A R T I C L E S

Table 4. Selected Bond Distances (Å) and Angles (deg) for Complex 10

bond distances (Å)

Al-F1

1.7858(17)

2.621(3)

2.444(3)

2.460(3)

2.012(3)

F1-Zr

2.1165(15)

C1-Zr

2.245(3)

2.498(3)

2.530(3)

1.530(5)

1.497(4)

C2-Zr

2.531(3)

2.419(3)

2.538(3)

2.014(3)

1.369(3)

C7-Zr

C8-Zr

2.594(3)

2.451(3)

1.554(4)

2.025(3)

C13-Zr

C20-Zr

C15-C18

C28-C29

C14-Zr

C21-Zr

Al-C23

C24-F2

C18-Zr

C22-Zr

Al-C35

C19-Zr

C14-C15

Al-C47

bond angles (deg)

Al-F1-Zr

162.21(10)

F1-Zr-C1

92.65(10)

105.32(10)

113.98(12)

57.89(10)

F1-Al-C23

C35-Al-C23

C18-C15-C14

C23-C28-C29

103.52(10)

116.00(12)

99.5(2)

F1-Al-C35

C35-Al-C47

C16-C15-C17

100.14(10)

115.34(12)

106.5(3)

F1-Al-C47

C23-Al-C47

C14-Zr-C18

123.2(3)

Table 5. Comparative Bond Distances (Å) and Angles (deg) for

Complexes 7 and 10

Table 6. NMR-Derived Rate Constants and Free Energies of

Activation for Solution Dynamic Processes of Complexes 7 and 10

in Toluene-d₈

j

a

b

Zr−C

Zr−C_Cp

Zr−C

C_Cp−Zr−C

C −C

C −C −C

14 15 18

Flu

15

Flu

15

14

15

18

a

d

q

cocat.

(cat.)

T

(°C)

k

k₁

∆G ^q

(kcal/mol)

k

∆G

reorg

total

1

reorg

(s^-1

7

2.221(4) 2.157(4) 3.109(4) 118.6(2) 1.553(4) 1.512(6) 99.5(3)

(s^-1

)

(s^-1

)

(kcal/mol)

10 2.221(3) 2.176(3) 3.124(3) 119.0(1) 1.554(4) 1.530(5) 99.5(2)

3(7)^e

77.5

92.5

87.5

6.0

18.6

1.1

2.8

8.4

18.3

15.0

n.d.ⁱ

5.2^b

19.6(10)

19.4(10)

21.2(6)

20.5

21.4(6)

20.8

0.8(4)

2.2(4)

∼0^g

20.6(36)

20.8(36)

n.d.

>24.8

3(7)^e

16.4^b

1.1^c

^aCentroid of C₁₃H₈ligand. ^bCentroid of C₅H₄ligand.

6(10)^e

6(10)^f

6(10)^h

6(10)^f

6(10)^e

6(10)^f

87.5

2.8^c

∼0^g

1

The H EXSY spectrum of ion pair 7 at 23 °C shows NOE

contact between protons Me_Band H_F(see atom labeling scheme

C in Experimental Section), with sufficient intensity to indicate

that the time-averaged solution structure of 7 is a dissymmetric

contact ion pair, with a preferred orientation of the anion with

respect to the cation.²⁸This spectrum reveals exchange between

Me_Band Me_Msignals (methyl-methide exchange), arising from

borane migration (dissociation of B(C₆F₅)₃from Me_Band

subsequent transfer to Me_M) and also permutation of diaste-

reotopic Me_Rand Me_R′resonances and exchange between

corresponding fluorenyl and cyclopentadienyl ring proton pairs

(ligand side-side exchange, arising from both borane migration

and ion pair reorganization, eq 5). Relative rates of borane

migration and ion pair reorganization are sensitively dependent

on metal identity and ligand architecture, as shown in previous

studies of archetypal Group 4 metallocene dimethyl precatalysts

117.5

127.5

8.4^c

∼0^g

18.3^c

15.0^c

n.d.ⁱ

∼0^g

21.5(6)

∼0^g

n.d.ⁱ

<0.25

^ak_total) k₁+ k_reorg

.

^bFor 7, k₁) rate constant for B(C₆F₅)₃migration.

d

^cFor 10, k₁) rate constant for anion racemization. k_reorg) rate constant

of ion pair reorganization. ^eTaken from line-broadening analysis. ^fFrom

2D-EXSY NMR. ^gAssumed based on EXSY results at 127.5 °C. ^hProjected

values from line-broadening analysis. ⁱAnion racemization saturation

regime, see discussion. ^jConfidence intervals presented at the 90%

confidence level.

broadening in the Me_Rand Me_R′resonances, and the rate of

borane migration, determined from broadening of Me_M.³⁰

Kinetic results are summarized in Table 6. Confidence intervals

for ∆H^‡and ∆S^‡are determined using standard error values

from linear regression analysis on the data used to generate the

Eyring plot.

4a,12,29

activated with B(C₆F₅)_3.

Exchange-broadening observed in variable-temperature 1-D

¹H spectra of 10 arises from two discrete processes: ion pair

reorganization, and interconversion between stereochemical

configurations at the chiral Al center (anion racemization, see

Scheme 2 for both processes). Both processes interconvert major

and minor diastereomers, however ion pair reorganization does

so with ligand side-side exchange, whereas anion racemization

does not. Thus EXSY can be used to differentiate between them.

EXSY spectra collected at 23 °C, 87.5 °C, and 117.5 °C (τ_m)

1200 ms, 185 ms, 40 ms, respectively) exhibit cross-peaks

corresponding to interconversion between diastereomers, but no

cross-peak intensity indicative of ligand side exchange. This

observation motivated EXSY data collection at 127.5 °C (τ_m)

800 ms) specifically to examine the possibility of very slow

ion pair reorganization. This spectrum also reveals no apparent

side-side exchange cross-peak intensity (Figure 4), and is

analyzed to establish a lower limit for the ion pair reorganization

1-D ¹H NMR spectral data collected for 7 over a 40°

temperature range afford kinetic parameters for both of these

processes. Broadening of the i-Pr methyl (Me_Rand Me_R′) and

methide resonances (Me_M) can be used to determine the rates

of both processes using the standard modified Bloch two-site

exchange line-broadening formalism, k ) π(∆W).^21bAt a given

temperature, the ion pair reorganization rate is taken as the

difference between total side-side exchange rate, taken from

(30) Temperature-dependent quadrupolar broadening of the Me_Bresonance

precludes measurement of exchange broadening on this signal, which is

instead assumed to be equal to that of the Me_Mresonance. The detection

limit for broadening is determined by the digital resolution (0.067 Hz,

corresponding to k ) 0.21 s^-1), thus the site exchange rate at 23 °C

(k ≈ 0.2 s^-1) as determined from EXSY data demonstrates that 0 °C is a

suitable temperature to take as the zero-exchange limit for the purposes of

line-shape analysis.

(28) Zuccaccia, C.; Stahl, N. G.; Macchioni, A.; Chen, M.-C.; Roberts, J. A.;

Marks, T. J. J. Am. Chem. Soc. 2004, 126, 144891464.

(29) Line broadening is found to be independent of concentration over an 8-fold

range for 7, arguing that an intramolecular exchange process is prevalent.

9

J. AM. CHEM. SOC. VOL. 126, NO. 14, 2004 4615

A R T I C L E S

Chen et al.

S_d,iwith intensities I_d,iand cross-peak intensity limit I_n, where

Φ_d,iis the measured signal-to-noise ratio of signal S_d,iover the

(noise) region where a corresponding cross-peak is sought, and

Φ_nis the arbitrary signal-to-noise ratio limit below which real

intensity might be mistaken for background noise. Not all

observed diagonal peaks need be employed, but for each pair

of cross-peaks sought, both corresponding diagonal peaks are

measured, pairing each cross-peak region with a diagonal peak,

as dictated by the F2 dimension direction, when the exchange

interconverts species that are present in unequal concentrations.

The above formulation effectively averages the noise intensity

for the q regions measured and is valid inasmuch as the noise

intensity is constant across the noise regions used to measure

the Φ_d,i's.

In the present case, diagonal peaks A′ and a give intensities

I_A′and I_awith signal-to-noise ratios Φ_A′(over region aA′) and

Φ_a(over A′a), respectively, and q ) 2. We extend the above

general formulation to the present system using a factor D

representing distribution of cross-peak intensity due to back-

ground saturation-regime exchange. Anion racemization dis-

tributes expected total cross-peak intensity such that the summed

intensity of the inter-diastereomer cross-peaks is one-half of

the total expected cross-peak intensity arising from ion pair

reorganization, thus D ) 1/2. We set the arbitrary but reasonable

criterion, that a cross-peak having intensity I_cross< 2 I_n, may

be extant but indistinguishable from background noise, thus

setting Φ_n) 2. Supposing such a peak exists, r_llis then

Figure 4. EXSY spectrum (in toluene-d₈) of complex 10, 127.5 °C, τ_m

)

800 ms. Diagonal peaks, lower-left to upper-right, correspond to resonances

H_A′, H_A, H_a′, and H_a, respectively. Spectra a. and b. are F2 slices passing

through the points of greatest intensity in resonances H_aand H_A′,

respectively.

q

φ

∑

rate at this temperature. In Figure 4, Aa and aA, or collectively

[Aa], refer to the pair of H_A-H_aexchange peaks (see atom

labeling scheme C in the Experimental Section), and I_Aa) I_aA

refers to the cross-peak intensity at these positions.³¹Cross-

peaks at [Aa], and at [A′a′] arise exclusively from anion

racemization. Intensity at [Aa′], [A′a], [AA′], and [aa′] is

expected from ion pair reorganization + rapid anion racemiza-

tion, with only the former process permuting ligand sides. Any

cross-peak intensity from ion pair reorganization is distributed

over these eight locations by the background anion racemization

(for which the present conditions represent the saturation regime,

with k ) 28 s^-1at 127.5 °C from dynamic NMR).

d,i

φ_A'+ φ_a

i

r > r_ll≈ D

)

(8)

qφ_n

8

The signal-to-noise ratios Φ_A′and Φ_a(33.4 and 64.4, respec-

tively, see Figure 4) are determined from F₂slices passing

through the highest points in A′ and a, setting r > r_ll) 12.2

and thereby giving a higher limit, k < 0.25 s^-1at 127.5 °C for

ion pair reorganization, including a correction for implicit NOE

intensity.³²

Ion pair complexes formed by activation of metallocene

precatalysts with Ph₃C⁺B(C₆F₅)₄are found to be inisolable

-

For two-site exchange, a suitable model for the proposed ion

pair reorganization under saturation conditions for anion race-

mization, rates is given by eq 6.¹⁹

- 7c

and very unstable, excepting Cp₂*ThMe⁺B(C₆F₅)₄

.

For

complex 9, determining the rate of a putative dynamic reorga-

nization/symmetrization process analogous to that observed with

7 was unsuccessful, with extensive decomposition occurring at

much lower temperatures (∼50 °C) compared to complexes 7

and 10. Similarly, decomposition of complex 8 is observed upon

heating (above ∼80 °C) and results in an insoluble, blue-purple

product in toluene. It will be seen from evidence derived from

polymerization results, that the site epimerization rate constant

for 9 is 11.1(11) s^-1at 60 °C, the highest value for all systems

studied, suggesting that the ion pairing interaction is weakest

for B(C₆F₅)₄^-as compared to the other systems (see discussion

below). The structure of the active species corresponding to 8

I

∑

1

r + 1

diag

k ) ln

with r )

(6)

(

)

τ_mr - 1

I

cross

A lower limit for r, r_ll, can be established using extant diagonal

peak intensities and a suitable higher limit for the corresponding

cross-peak intensities I_cross. What follows is a general method

for estimating r_llfor systems undergoing two-site exchange

without distribution of cross-peak intensity by background,

saturation-regime processes for some q extant diagonal signals

q

(32) RMS signal-to-noise ratios for specific signals over specific noise regions

are obtained from the VNMR command, dsn. Scant NOE intensity detected

in the EXSY spectrum of 10 at 127.5 °C at locations [A′B′], [a′b′], [A′b′],

and [B′a′] (distributed by rapid anion racemization) is measured, and used

together with the crystal data of 10 to estimate the expected NOE intensity

at A′a and aA′ (accompanied by intensities at [Aa′], [A′A], and [a′a]) and

increases the higher limit for the ion pair reorganization rate constant by

0.05 s^-1, assuming total cancellation of NOE and exchange cross-peak

intensity at these positions.

I

φ n

φ

d,i

∑

i

d,i

i

r > r_ll≈

)

(7)

qI_n

qφ_nn

qφ_n

(31) Signal overlap precludes use of other sets of signals for this determination.

9

4616 J. AM. CHEM. SOC. VOL. 126, NO. 14, 2004

Counteranion Effects on Olefin Polymerization

A R T I C L E S

during polymerization is unknown, however results discussed

below allow an estimation of k ) 10.9(10) s^-1at 60 °C for a

putative site epimerization in this species.

microstructure, and how these may reflect the coordinative

component of anion interaction with the cationic metal center.^7c

Several trends are immediately evident in the data. Product

polydispersities are consistent with well-defined single-site

processes and are rather temperature- and anion-insensitive.

Polymer production rates, however, are highly anion-sensitives

the intrinsic steric and electronic characteristics of the anions

appearing to have a major influence on monomer activation and

enchainment. The most strongly (FAl(2-C₆F₅C₆F₄)₃^-) and

weakly (MeB(2-C₆F₅C₆F₄)₃^-, B(C₆F₅)₄^-) coordinating anions

generally exhibit the lowest and highest polymerization rates,

respectively (Table 7). Not surprisingly,^1,11product molecular

weights fall with rising reaction temperature, in all cases (Figure

5). FAl(2-C₆F₅C₆F₄)₃^-affords the highest M_wproduct polymer

at all temperatures (vide infra for pressure effects). Most

interesting, however, is the striking pattern in polypropylene

stereodefect probabilities³⁵(P_m, generally attributed to site

epimerization,³⁶Scheme 1B, and P_mm, from propylene enan-

tiofacial misinsertion or chain epimerization, Schemes 1D, 1E,

respectively) as a function of anion and temperature (Figure

6). It can be seen that the FAl(2-C₆F₅C₆F₄)₃^--based catalyst

exhibits far higher syndiospecificity, with far lower m and

somewhat lower mm stereodefect production. All systems exhibit

a not unprecedented erosion in syndioselectivity with increasing

temperature,¹¹likely due to acceleration of m steric dyad

production vs enchainment, least prevalent in the FAl(2-

C₆F₅C₆F₄)₃^--based catalyst. The NMR-derived ion pair reor-

ganization/symmetrization kinetic results and the comparatively

low polymerization activity temperature dependence for 10 argue

that tighter ion pairing raises the activation energy for site

epimerization. Interestingly, mm stereodefects are far less

temperature-sensitive for all catalysts, with the FAl(2-C₆F₅C₆F₄)₃^-

The assembled NMR-derived kinetic data (Table 6) indicate

a fundamental and substantial difference in the lability of

MeB(C₆F₅)₃^-and FAl(2-C₆F₅C₆F₄)₃^-as counteranions for the

zirconocenium fragment.³³Indeed, the lower limit derived for

the barrier to ion pair reorganization in 10 is conservative, and

considering that this interaction may be even stronger, it is

remarkable that 10 produces polymer at all, inviting speculation

on the pathway for monomer enchainment. Multiple pathways

for insertion have been postulated in computational studies

where the catalyst-cocatalyst interaction is included,²⁴and the

collection of systems presented here may serve to differentiate

among these possibilities: specifically, an enchainment pathway

with concerted anion displacement may be favored in the FAl-

-

(2-C₆F₅C₆F₄)₃system, in contradistinction to the B(C₆F₅)₄

system, for example. Also, considering that monomer enchain-

ment is in general impeded by stronger ion pairing/increasing

counteranion coordinative tendencies,^12cand that in similar

systems ion pairing strength has been shown to diminish with

increasing zirconocenium alkyl steric bulk,^12ait is possible that

this latter differential effect may diminish more rapidly with

more strongly binding counteranions.

Independent of mechanistic considerations, differences in

counteranion coordinative ability manifest themselves measur-

ably in the rate constants for propylene insertion relative to those

of competing processes believed to occur during polymerization.

It is the goal of the following sections to examine these effects.

II. Catalytic Propylene Polymerization Mediated by

Complexes 7-10. It will be seen that substantial counteranion

effects are evident in the polymerization characteristics and

polypropylene microstructures obtained using the present cata-

lyst systems. The following sections examine cocatalyst, tem-

perature, monomer concentration, and solvent polarity effects

on stereodefect production, polymerization activity, and termi-

nation/chain-transfer kinetics. These effects represent an inter-

play of structural, kinetic, and thermodynamic influences, among

which the dominant factor is argued to be the lability of the

catalyst cation-anion interaction.

A. Counteranion and Temperature Effects on Propylene

Polymerization. Under rigorously anhydrous/anaerobic condi-

tions, complex 1 was activated with MAO (2) or perfluoroaryl

cocatalysts 3-6 to generate catalytically active ion pairs in situ.³⁴

Polymerizations were carried out under 1.0 atm propylene

pressure in toluene solution over the temperature range of -10°

to 60 °C using conditions minimizing mass transfer¹⁷and

exotherm effects (see the Experimental Section for details);^2c,10

product isolation and characterization utilized standard tech-

niques.¹¹The results of these propylene polymerization experi-

ments are summarized in Table 7. The data are analyzed with

a view toward discerning cocatalyst-dependent effects on

polymerization activity, molecular weight characteristics, and

-

catalyst again slightly superior. In contrast, the MeB(C₆F₅)₃

catalyst exhibits the lowest syndiotacticity with greatest increase

of m and mm stereodefect production with rising polymerization

temperature (vide infra for detailed explanation). In general, as

the temperature is increased, polymerization activities increase,

except near 60 °C, where activities decrease for all ion pairs.

Not only lower ion pair thermal stability, but also decreased

propylene solubility at higher temperatures doubtless contributes

to the lower activity observed in all systems at 60 °C (in toluene,

[propylene] ) 0.36 M at 60 °C vs 0.83 M at 25 °C, at 1.0 atm

system pressure).³⁷In addition, the B(C₆F₅)₄^--derived catalyst

exhibits the most significant erosion in performance, likely

reflecting the poor thermal stability of this complex as noted

above. In comparison to the FAl(2-C₆F₅C₆F₄)₃^--based polym-

erization system, lower product syndiotacticities but higher

polymerization activities are observed in the 1 + MAO system.

In agreement with previous polymerization studies using [Me₂C-

(35) For 1 + MAO, syndiotacticity falls with increasing polymerization

temperature, ref 11, while for C₁-symmetric catalysts, isotacticity sometimes

increases with increasing polymerization temperature: (a) Kleinschmidt.

R.; Reffke, M.; Fink, G. Macromol. Rapid Commun. 1999, 20, 284-288.

(b) Grisi, F.; Longo, P.; Zambelli, A.; Ewen, J. A. J. Mol. Catal. A: Chem.

1999, 140, 225-233. (c) For an example of a C₂-symmetric catalyst

propylene polymerization temperature dependent study, see: Resconi, L.;

Piemontesi, F.; Camurati, I.; Sudmeijer, O.; Nifant′ev, I. E.; Ivchenko, P.

V.; Kuz′mina, G. K. J. Am. Chem. Soc. 1998, 120, 2308-2321.

(36) See ref 11. Other proposed processes giving rise to m stereodefects are

discussed below. See ref 2c.

(37) An empirical model for the calculation of the solution-phase composition

of propylene in toluene and isododecane under relavent conditions is

presented in (a) Dariva, C.; Lovisi, H.; Santa Mariac, L. C.; Coutinho, F.

M. B.; Oliveira, J. V.; Pinto, J. C. Can. J. Chem. Eng. 2003, 81, 147-152.

(b) also see ref 17.

(33) Dynamic NMR experiments with 7 and 10 1,2-dichlorobenzene-d₄as solvent

reveal that in this more polar medium, the barrier to ion pair reorganization

in 7 is lowered, whereas with 10 such a process is still undetectable.

(34) For experiments using MAO (2) as cocatalyst, an Al:Zr ratio of 60: 1 is

employed, to improve comparability with results collected using molecular

cocatalysts. Control experiments in which the Al:Zr ratio is varied across

a 30-fold range show no significant dependence of the pentad distribution

on this ratio. These results are presented in Table 2 of the Supporting

Information.

9

J. AM. CHEM. SOC. VOL. 126, NO. 14, 2004 4617

A R T I C L E S

Chen et al.

Table 7. Propylene Polymerization Results for the Reactions Mediated by 1+ Indicated Cocatalysts under 1.0 atm of Propylene over the

Temperature Range from -10 to +60 °C^a

k_p^e

T_m

(°C)

M

w

P_m

P_mm

r ^h

(%)

rrrr

c

d

f

g

i

cocat.

(cat.)

exp.

no.

T

(°C)

cat.

(µmol)

[C H ]^b

(M)

time

(s)

yield

(g)

v_p,apparent

(M*s^-1

p

3

6

)

(M^-1*s^-1

)

(kg*mol^-1

)

P.D.I.^f

(%)

2

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

-10

0

10

2.83

1.87

1.31

0.83

0.56

0.36

2.83

1.87

1.31

0.83

0.56

0.36

2.83

1.87

1.31

0.83

0.56

0.36

2.83

1.87

1.31

0.83

0.56

0.36

2.83

1.87

1.31

0.83

0.56

0.36

1800

1200

600

1200

1800

3600

1800

600

1800

720

600

300

600

300

180

120

180

0.77

0.75

0.73

2.20

1.80

1.24

0.50

1.14

4.68

4.40

0.94

2.48

0.88

0.62

0.69

2.92

3.71

1.74

1.18

0.74

0.42

0.73

0.77

1.27

0.85

0.54

1.58

5.00

0.51

0.25

0.188

0.278

0.535

1.61

0.36

1.6

6.6

156.0

151.0

148.9

141.5

129.5

N.O.^k

147.0

140.4

127.3

101.4

69.0

138

139

125

124

1.88

1.85

1.83

1.88

1.87

2.16

1.75

1.56

1.87

1.81

1.92

2.38

1.83

1.92

1.85

1.78

1.82

1.95

1.93

1.98

1.95

1.96

1.82

1.86

2.04

1.96

1.85

2.09

1.95

0.3

0.4

0.5

1.3

2.7

8.0

1.3

2.0

3.1

6.7

12

20

0.5

0.7

1.0

2.4

5.3

0.7

0.9

1.4

1.5

2.8

1.4

1.6

1.5

2.0

2.5

4.3

1.3

1.4

1.7

1.9

2.6

3.2

1.1

1.3

1.5

1.9

2.3

3.2

0.9

1.0

1.2

1.5

2.0

2.7

98.2

97.8

97.6

96.1

94.3

87.6

95.8

94.7

93.9

89.7

83.9

74.7

97.3

96.6

95.8

94.1

90.6

82.0

97.0

96.5

95.6

93.8

89.9

82.4

97.8

97.6

97.1

96.1

94.1

89.8

95.3

94.1

93.6

88.7

83.1

63.8

88.6

85.2

81.8

69.0

53.1

34.6

92.9

90.8

88.5

82.7

72.4

49.5

92.1

90.7

88.3

82.4

70.8

50.8

94.3

93.9

92.6

89.5

84.2

70.8

5.0

3.3

2.5

20

10

25

40

60

-10

0

10

25

40

60

-10

0

10

25

40

60

-10

0

10

25

40

60

-10

0

32

1.32

51

80.8

0.459

0.122

0.279

0.572

1.08

0.689

0.606

0.538

0.455

1.01

4.28

5.44

1.28

1.73

1.81

1.54

1.78

4.52

27^j

36.7

79.8

78.8

3(7)^l

4(8)

0.12

0.40

1.2

126

3.5

79.0

41.6

11.9

6.6

20

4.5^j

1.4

N.O.^k

150.9

147.4

143.2

130.3

108.2

N.O.^k

151.5

147.6

143.5

130.7

110.3

N.O.^k

156.5

154.5

151.2

145.7

136.0

N.O.^k

7.6

5.0

10

6.2

10

1.5

1.3

2.6

10

20

201

2.6

8.4

28

168

132

101

82.8

53.1

229

180

153

112

82.6

55.8

290

84

19^j

13

5(9)^l

6(10)

22

40

49

90

0.4

0.8

1.2

2.4

5.1

75

600

167

0.932

0.035

0.066

0.155

0.489

0.062

0.061

14^j

13

10 800

3600

4500

3600

1800

0.030

0.10

0.32

1.6

0.3

0.5

0.9

1.7

5.3

242

204

147

104

10

25

40

60

0.30^j

0.45^j

66.5

^aIn 54 mL of toluene. ^bSee ref 37a. ^cAfter workup (see Experimental Section). ^dAs calculated from polymerization yield. ^eAs calculated from V_p,apparent

.

^fDetermined from GPC analysis relative to polystyrene standards; polydispersity index ) M_w/M_n. ^gDetermined from polymer ¹³C NMR pentad analysis.

i

^hFractional dyad content, r ) (Σ_x,yxmry + Σ_x,yxrmy + 2Σ_x,yxrry)/2, with x, y ∈ {r, m}. Calculated rrrr signal integral; see Supporting Information for

experimental and calculated pentad distributions. ^jNot used for estimating k_pat 60 °C. ^kNot observed. ^lThese data supplemented with additional results for

estimation of k_pat 60 °C (see Supporting Information).

Figure 5. Product molecular weight (M_w) data for polypropylenes produced

by 1+ the indicated cocatalysts over the temperature range from -10° to

+60 °C under 1.0 atm of propylene.

(Cp)(Flu)]ZrCl₂+ MAO (% rrrr ) 93.1 at 10 °C),^35acompa-

rable rrrr pentad contents (93.6%, Table 7, entry 3) are obtained

in the present work. The temperature dependence of derived

kinetic parameters will be addressed below.

Figure 6. A. Syndiotacticity (%rrrr) data and calculated m and mm

stereodefect production probabilities (relative to insertion, P_mand P_mm

respectively, discussed below) for polypropylenes produced by 1+ indicated

cocatalysts under 1.0 atm of propylene over the temperature range from

-10° to +60 °C.

B. Monomer Concentration Effects. Polymerization series

in which [propylene] is systematically varied reveal anion

dependences that are subtle compared to the anion sensitivity

of the temperature effects described above. These experiments

were carried out with T ) 60 °C, to maximize signal-to-noise

ratios for dilute pentad signals. The mechanistic consequences

of increasing [propylene] can be ascribed to increased rates of

bimolecular reactions such as insertion or enantiofacial misin-

sertion vs those of competing unimolecular processes such as

site epimerization and â-hydrogen elimination to Zr (resulting

in either chain epimerization or termination and reasonably

assumed to be zero-order in monomer). Generally, any observed

[propylene] effect on a measurable polymer feature can be

interpreted as arising from a combination of processes having

9

4618 J. AM. CHEM. SOC. VOL. 126, NO. 14, 2004

Counteranion Effects on Olefin Polymerization

A R T I C L E S

Table 8. Propylene Polymerization Results for the Reactions Mediated by 1+ Indicated Cocatalysts at 60 °C over the Pressure Range of

1-5 atm of Propylene^a

k_p^e

T_m

(°C)

M

w

P_m

P_mm

r^h

(%)

rrrr

c

d

f

g

i

cocat.

(cat.)

exp.

no.

P

(atm)

cat.

(µmol)

[C H ]^b

(M)

time

(s)

yield

(g)

v_p,apparent

(M*s^-1

3

6

)

(M^-1*s^-1

)

(kg*mol^-1

)

P.D.I.^f

(%)

2

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

1.0

2.0

3.0

4.0

5.0

1.0

2.0

3.0

4.0

5.0

1.0

2.0

3.0

4.0

5.0

1.0

2.0

3.0

4.0

5.0

1.0

2.0

3.0

4.0

5.0

2.5

1.5

20

6.6

5.9

5.1

10

2.9

2.5

2.9

2.5

10

5.1

2.6

20

0.36

0.76

1.18

1.61

2.05

0.36

0.76

1.18

1.61

2.05

0.36

0.76

1.18

1.61

2.05

0.36

0.76

1.18

1.61

2.05

0.36

0.76

1.18

1.61

2.05

1200

240

120

180

120

1800

1200

600

1200

600

1.24

0.92

1.03

1.88

2.19

2.48

1.18

2.41

2.82

1.52

1.74

1.31

1.88

1.19

1.25

1.27

1.21

1.01

1.62

1.56

0.25

2.09

3.54

6

0.454

1.69

3.79

4.60

8.03

27

80

N.O.^j

113.8

122.2

125.4

131.6

N.O.^j

99.0

106.0

109.6

N.O.^j

105.8

106.4

109.7

N.O.^j

119.4

124.5

130.0

127.2

36.7

48.8

56.7

63.2

71.2

11.9

19.0

25.0

29.5

33.9

53.1

53.5

55.8

56.6

58.6

55.8

57.1

59.3

61.2

63.2

66.5

68.6

71.2

73.3

70.8

2.16

1.87

1.84

1.81

1.80

2.38

2.03

2.55

2.57

1.89

1.82

1.80

1.89

1.86

1.81

1.82

2.23

1.97

1.82

1.68

1.95

2.04

1.88

1.87

1.86

8.0

5.0

3.8

2.9

2.3

2.8

2.3

2.6

2.0

2.1

4.3

2.9

3.3

3.0

3.2

3.0

3.1

2.7

3.2

2.9

3.2

2.6

2.7

2.3

2.8

2.3

2.5

87.0

90.8

91.1

93.2

93.6

74.6

79.8

81.7

83.9

84.9

81.7

86.2

87.5

89.9

90.7

82.4

86.1

87.9

90.0

89.2

89.6

92.4

91.6

93.6

93.0

62.7

72.9

75.2

80.2

82.2

33.6

43.9

49.5

54.5

57.1

48.9

60.2

64.6

70.5

73.5

50.5

60.3

65.9

70.8

70.2

70.4

78.1

77.5

82.1

80.8

116

103

141

4.5

4.6

6.9

5.9

2.9

19

12

13

10

4.8

14

6.2

6.5

7.7

5.8

0.45

0.91

0.99

0.82

0.94

3(7)

4(8)

5(9)

6(10)

0.606

0.433

0.884

1.03

0.560

1.28

0.480

0.689

0.873

0.458

0.932

0.444

0.370

0.594

0.572

0.061

0.511

1.30

20

16

13

11

9.9

13

8.6

6.5

5.2

4.8

1200

600

13

1200

1800

1200

1800

8.8

6.8

5.4

4.3

5.3

3.2

2.9

2.1

2.2

40

60

20

1.47

0.714

2.92

^aIn 54 mL of toluene. ^bSee ref 37a. ^cAfter workup (see Experimental Section). ^dAs calculated from polymerization yield. ^eAs calculated from V_p,apparent

.

^fDetermined from GPC analysis relative to polystyrene standards; polydispersity index ) M_w/M_n. ^gDetermined from polymer ¹³C NMR pentad analysis.

^hfractional dyad content, r ) (Σ_x,yxmry + Σ_x,yxrmy + 2Σ_x,yxrry)/2, with x, y ∈ {r, m}. ⁱCalculated rrrr signal integral; see Supporting Information for

experimental and calculated pentad distributions. ^jNot observed.

proposed rate laws that differ in their [propylene] depen-

dence.^11,38This approach is used for analyzing product molecular

weight and the abundance of m and mm stereodefects, always

against the background of chain propagation, reasonably as-

sumed to be first-order in monomer.³⁹The present work reveals

that these effects are particularly sensitive to counteranion

identity.

Limited, but anion-dependent increases in product molecular

weights (Table 8) are observed with increasing monomer

pressure, arguing that [monomer]-dependent termination pro-

cesses are significant.^11,38The sensitivity of M_nto propylene

-

pressure change is markedly higher in the MeB(C₆F₅)₃

polymerization system (7), suggesting that the ratio of rates for

unimolecular termination (υ_t1) vs [monomer]-dependent termi-

nation (υ_t2,propylene) is higher in this case.¹⁷This is illustrated in

Figure 7: assuming negligible chain transfer involving species

other than propylene, the slope and intercept from a linear fit

Figure 7. Product 1/P_nat 60 °C plotted vs 1/[propylene] for 1+ indicated

cocatalysts (Table 8).

of 1/P_nvs. 1/[propylene] (P_nis the number-average degree of

polymerization; see eq 9) are equal to k_t1/k_p, and k_t2,propyene/k_p

(38) (a) For C_s-symmetric catalysts, lower propylene concentrations correlate

with lower product molecular weights and tacticities (mostly m stereo-

defects), ref 13c. (b) In contrast, declining isotacticity with increasing

monomer concentration is observed in C₁-symmetric catalysts: Kukral, J.;

Lehmus, P.; Feifel, T.; Troll, C.; Rieger, B. Organometallics 2000, 19,

3767-3775. (c) For C₂-symmetric catalyst propylene concentration studies,

see ref 17, and also: (d) Busico, V.; Cipullo, R.; Cutillo, F.; Vacatello, M.

Macromolecules 2002, 35, 349-354, (e) Busico, V.; Brita, D.; Caporaso,

L.; Cipullo, R.; Vacatello, M. Macromolecules 1997, 30, 3971-3977, and

(f) Resconi, L.; Fait, A.; Piemontesi, F.; Colonesi, M. Macromolecules 1995,

28, 6667-6676.

(39) Examination of insertion rate vs propylene concentration from Table 8

reveals an approximately linear correlation in several systems, thus insertion

is assumed to be first-order in monomer in the present model. This was

also observed for the polymerization of 1-hexene catalyzed by [rac-(C₂H₄-

(1-indenyl)₂)ZrMe] [MeB(C₆F₅)₃] over the temperature range of -10 °C

to 50 °C, see ref 40. However, for C₂-symmetric catalyst propylene studies,

there is debate in the literature as to the exact order of monomer in

production of isotactic polypropylene. See refs 17, 38d-f. Ref 38d contains

a model reconciling observed apparent propagation [propylene] dependences

using a rigorous approach that holds propagation to be first-order in

monomer.

ν_t2,propylene+ ν_t1k_t2,propylenek_t1

1

P_n

1

)

+

(9)

(

)

ν_p

k_p

[propylene]

respectively (V_pbeing the rate of polymerization, assumed to

be first-order in [propylene]).¹⁷The other catalyst-cocatalyst

systems studied here, including MAO (2, for which chain

transfer to aluminum alkyls cannot be ruled out), are indistin-

guishable in this respect, in particular exhibiting a general

suppression of unimolecular termination (Figure 7). However,

the variance in GPC-determined M_wvalues propagates to

substantial uncertainties in the quantitation of k_t1/k_p, and k_t2,prp

/

k_p.^11c1H NMR end-group analysis of the product polymers

reveals that in these systems, 2,1-misinsertion followed by

â-hydrogen transfer to Zr is not significant (less than 10%),⁴⁰

9

J. AM. CHEM. SOC. VOL. 126, NO. 14, 2004 4619

A R T I C L E S

Chen et al.

also arguing that chain transfer to monomer is the preferred

bimolecular termination route, in these systems.

Increased monomer concentrations are accompanied by

increases in syndiotacticity in all systems studied (Figure 6).

Statistical techniques for modeling polymer ¹³C NMR pentad

distributions have been described for C_s-symmetric catalyst

systems, in particular for simultaneous estimation of probabilities

(relative to propagation) for events that produce m steric dyads

and mm steric triads in the product polymer.¹³These models

have the advantage of accounting for steric pentads (or any

n-ads) containing multiple stereodefects ('shared pentads' e.g.

mrmr, or mmrm). We apply here a standard statistical model^13c,d

to extract the probability P_mmof mm-generating processes and

P_mof m-generating processes, that takes into account their

contributions to shared pentad intensity. This model is based

on the assumption of perfect enantiomorphic site control, as

has been justified in several previous examples for this class of

catalysts.¹³Experimental and calculated pentad distributions for

both pressure- and temperature-dependence polymerization

series appear in the Supporting Information.

Figure 8. Product P_mat 60 °C plotted vs 1/[propylene] for 1+ indicated

cocatalysts (Table 8).

Table 9. Slopes, Intercepts, and Rate Ratios (%) of m

Stereodefects Originating from Site Epimerization vs “Back-Side”

Misinsertion Obtained from P_mvs 1/[Propylene] Plots for

Polymerizations Mediated by 1+ Indicated Cocatalysts under

1.0-5.0 atm Propylene at 60 °C^a

cocat.

(cat)

k_se/k_p^b

(slope)

k

_bsa/k_p^b

(intercept)

Catalyst site epimerization, having the proposed rate law,

V_se) k_se[catalyst], leads to the formation of m steric dyads as

stereodefects in the product polymer when followed by “normal”

chain-migratory insertion, but is not necessarily the most

significant factor in degradation of syndiotacticity at the

temperature maintained for this set of experiments (60 °C). In

another possible scenario, â-hydrogen transfer to the catalyst

metal center is followed by re-si interconversion of the resulting

π-macroolefin complex, and reinsertion (chain epimerization,

Scheme 1E). Concomitant stereoinversion at the metal (ion-

pair reorganization, Scheme 1E, pathway i) also generates an

m stereodefect. This process would then have the same stereo-

sequence and rate law as site epimerization. It is possible that

in certain systems, ion-pair reorganization of the π-macroolefin

complex proceeds rapidly compared to the reinsertion step (the

macroolefin being 1,1′-disubstituted). Chain epimerization

without stereoinversion of the metal generates an mm stereo-

defect (Scheme 1E, pathway ii).^11a,11c,41As Busico et al. have

recently observed,^2am stereodefects can in principle also arise

from insertion without chain migration (“back-side attack,”

opposite the anion, rather than same-side attack). In fact, any

1,2-insertion in which no net stereochemical inversion of the

catalyst occurs, if followed by a “normal” chain-migratory

insertion, will give rise to an isolated m stereodefect, either as

an mrr or rmr tetrad, depending on the enantiofacial orientation

of the back-side misinserted monomer (Scheme 1D). If we

assume V_bsa) k_bsa[catalyst][propylene] for such a process, then

the m stereodefect probability can be expressed as in eq 10.

c

v_se/v_p^c

v_sc/v_sc+ v_bsa

2

0.0243(19)

0.0442(62)

0.034(29)

0.0374(37)

0.0141(12)

0.0146(28)

0.0863(90)

0.034(43)

0.0319(55)

0.0144(18)

0.067

0.12

0.1

0.093

0.039

0.82

0.58

0.76

0.73

3(7)

4(8)

5(9)

6(10)

^aConfidence intervals presented at the 90% confidence level. ^bk_p, k_se,

and k_bsaare as defined in eq 10. ^cAt 60 °C, 1.0 atm, [propylene] ) 0.364

M, see ref 37a for conversion. V_se/V_p) (k_se/k_p) *(1/[propylene]); V_se/(V_se

+

V

_bsa) represents the fraction of m stereodefects attributable to site epimer-

ization at 60 °C, 1.0 atm.

[propylene]. The present results indicate that k_bsais detectably

nonzero at 60 °C with MeB(C₆F₅)₃as the anion (7, in

-

agreement with ref 2a), and indeed for all activators treated in

the present report. The observed anion ordering in k_se/k_pis FAl-

(2-C₆F₅C₆F₄)₃^-(counteranion in 10) < MeMAO^-(2) < MeB-

(2-C₆F₅C₆F₄)₃(8) < B(C₆F₅)₄(9) < MeB(C₆F₅)₃(7). This

ordering does not track ion pairing strength, but there is no

reason to expect it should:^2a,bit is possible that ion pairing

dynamics and counteranion structure/electronics influence the

propagation and site epimerization processes to different degrees.

The anion ordering in k_bsa/k_p, 10 ∼ 2 < 9 ∼ 8 < 7, also fails

to adhere to a specific trend.

-

This analysis employs the assumptions that backside attack

and site epimerization occur according to the above rate laws

and are the only processes giving rise to m stereodefects.

Independent evidence supporting a backside reaction pathway

is desired. Also, structure/function relationships are meaningful

only when determined for elementary processes; the product

polymer features analyzed herein are each derivative phenom-

ena. Estimates for k_pusing available activity data are required

to extract approximate values for k_seand k_bsa; the results of this

analysis, along with a discussion of its inherent limitations, are

presented below.

ν_bsa+ ν_sek_bsak_se

1

P_m)

)

+

(10)

(

)

ν_p

k_p

[propylene]

A linear fit of P_mvs 1/[propylene] then gives estimates for k_bsa

/

k_pand k_se/k_p(as intercept and slope, respectively, eq 10; see

Figure 8, Table 9). In this and subsequent models, V_p, the rate

of polymerization, represents the sum of rates for chain

migratory insertion, misinsertion via backside attack, and

enantiofacial misinsertion, all assumed to be first order in

The decrease in m stereodefect abundance with increasing

monomer concentration is found here to be greater than that of

the mm stereodefect abundance, this latter decline being largest

-

with MeB(C₆F₅)₃system but undetectable for the FAl(2-

(40) Liu, Z.; Somsook, E.; White, C. B.; Rosaaen, K. A.; Landis, C. R. J. Am.

Chem. Soc. 2001, 123, 11 193-11 207.

(41) Sillars, D. R.; Landis, C. R. J. Am. Chem. Soc. 2003, 125, 9894-9895.

C₆F₅C₆F₄)₃^--derived catalyst 10. Rate constants for chain

epimerization (V_ce) k_ce[catalyst]) and enantiofacial misinsertion

9

4620 J. AM. CHEM. SOC. VOL. 126, NO. 14, 2004

Counteranion Effects on Olefin Polymerization

A R T I C L E S

systematically underrepresented owing to reversible and ir-

reversible catalyst deactivation and catalyst induction, each of

which is likely both cocatalyst- and temperature-dependent.⁴³

Correcting for such effects would have the effect of inflating

k_p,apparent, thus increasing the estimates for elementary rate

constants k_se, k_bsa, k_ce, and k_emfrom the relative quantities derived

using eqs 10 and 11. However, reaction of catalysts with

adventitious stoichiometric poisons can be reasonably assumed

to proceed to completion in all cases. Thus the depression of

V_p,apparentdue to unintentional contamination should be statistical

and independent of catalyst and temperature, and can be safely

ignored when comparing k_p,apparentamong the present catalyst

systems, with the exception of 1 + 2, wherein large quantities

of excess alkylaluminoxanes introduce a further systematic

uncertainty. Catalyst thermal decomposition is clearly both

temperature- and catalyst-dependent (and possibly also [catalyst]-

dependent), and will thus introduce systematic errors into

estimates of the elementary rate constants k_se, k_bsa, k_ce, and k_em

for each system. However, we observe that greater polymeri-

zation rates (V_p,apparent, see Table 7) are generally associated with

greater thermal instability, thus suggesting a greater underrep-

resentation of these elementary rate constants with systems that

are more active.⁴⁴It has been observed by Landis et al. that the

fraction of catalytically active states in isospecific 1-hexene

polymerizations is moderately higher with B(C₆F₅)₃as activator

than with [PhNH(Me₂)]⁺B(C₆F₅)₄^-, suggesting that the actual

propagation rate constant is indeed higher in the latter case.

Thus, underrepresentation of rate constants k_se, k_bsa, k_ce, and k_em

Figure 9. Product P_mmat 60 °C plotted vs 1/[propylene] for 1+ indicated

cocatalysts (Table 8).

Table 10. Slopes, Intercepts, and Rate Ratios (%) of mm

Stereodefects Originating from Chain Epimerization vs

Enantiofacial Misinsertion Obtained from P_mmvs 1/[Propylene]

Plots for Polymerizations Mediated by 1+ Indicated Cocatalysts

under 1.0-5.0 atm Propylene at 60 °C^c

cocat.

(cat)

k_ce/k_p^b

(slope)

k

_em/k_p^b

(intercept)

c

v_ce/v_p^c

v_ce/v_ce+ v_em

2

0.0026(13)

0.0057(18)

0.0009(11)

0.0023(14)

0.0009(13)

0.0204(20)

0.026(26)

0.0291(17)

0.0261(20)

0.0239(19)

0.0071

0.016

0.0063

0.0025

0.26

0.38

0.19

0.078

0.094

3(7)

4(8)

5(9)

6(10)

^aConfidence intervals presented at the 90% confidence level. ^bk_p, k_ce,

and k_emare as defined in eq 11. ^cAt 60 °C, 1.0 atm, [propylene] ) 0.364

M, see ref 37a for conversion. V_ce/V_p) (k_ce/k_p) * (1/[propylene]); V_ce/

(V_ce+ V_em) represents the fraction of mm stereodefects attributable to site

epimerization at 60°C, 1.0 atm.

-

can reasonably be expected to be larger with B(C₆F₅)₄than

with MeB(C₆F₅)₃^-in the present case. Also, in both cases, the

fraction of active catalysts at any given time is approximately

the same as the fraction of catalysts that were active at some

time, suggesting that the formation of dormant states on the

time scale of their experiments (0.01-1000 s) is not significant.^43b

Subsequent direct NMR observation of catalyst polymeryl

species has shown that accumulation of dormant states is

insignificant.^43dAssuming this holds in the present systems, the

above findings on active site count and activity provide the

following condition for comparability: in comparing two

catalysts of different activity, if the more active catalyst gives

the larger apparent value for some elementary rate constant,

the difference in this value for the two catalysts is thus

underrepresented, and the ordering is reliably given by the data.

The ordering can be established, in any comparison for which

this condition is satisfied. Estimates for k_p, k_se, k_bsa, k_ce, and k_em

among the present series of catalysts are summarized in Table

11.⁴⁵Values for k_pat 60 °C were determined by extrapolation

of activity Arrhenius plots established from the temperature-

(V_em) k_em[catalyst][propylene]) vs propagation can be calcu-

lated from P_mmfor each catalyst (eq 11, with ν_pas defined

above), by analyzing the plots of P_mmvs 1/[propylene] for each

system (Figure 9, Table 10). The present collected results

ν_em+ ν_cek_emk_ce

1

P_mm

)

+

(11)

(

)

ν_p

k_p

[propylene]

suggest that enantiofacial misinsertion is the prevailing pro-

cess for the generation of mm stereodefects in all systems

studied here. Similar evaluation of the data for (1,2-SiMe₂)₂-

{C₅H_2-4-R}{C₅H-3,5-(CHMe₂)₂}ZrCl₂(R ) H, CHMe₂, SiMe₃)

activated by MAO at 24 °C suggests that in this case,^11cthe

relative contribution of enantiofacial misinsertion is small. This

distinction is likely due to metallocene structural differences,

and possibly also to differences in reaction temperature.

C. Anion Mobility during Propylene Polymerization

Probed in Situ by Enchainment Syndioselection. Considering

the large observed counteranion effects on M_w, m, and mm

stereodefects, it is important to inquire into their origin;⁴²as

indicated above, counteranion effects, or any effects, are best

analyzed against the rates for individual processes. However,

use of polymerization analytical yields for determination of k_p

values, necessary for determination of k_se, k_bsa, k_ce, and k_em, can

be relied on only in certain cases. Propagation rates are

(43) Whereas the thermal instability of group IV metallocenium salts of

-

B(C₆F₅)₄is well-known (refs 1b,2,7c) previous work (ref 7c) indicates

that B(C₆F₅)₄^--based systems are to some degree stabilized in the presence

of olefin, and extant literature finds catalyst activity during polymerization

to be more or less constant: (a) Wester, T. S.; Johnsen, H.; Kittilsen, P.;

Rytter, E. Makromol. Chem. Phys. 1998, 199, 1989-2004. Negligible

catalyst deactivation is observed in 1-hexene polymerizations using [rac-

C₂H₄(indenyl)₂ZrMe][MeB(C₆F₅)₃], see: (b) Liu, Z.; Somsook, E.; Landis,

C. R. J. Am. Chem. Soc. 2001, 123, 2915-2916. For relevant ethylene

homopolymerization results using mononuclear and binuclear constrained

geometry catalysts see: (c) Abramo, G. P.; Li, L.; Marks, T. J. J. Am.

Chem. Soc. 2002, 124, 13 966-13 967, and also see ref 2c. (d) Landis, C.

R.; Rosaaen, K. A.; Sillars, D. R. J. Am. Chem. Soc. 2003, 125, 1710-

1711.

(42) For general kinetic models, see: (a) Nele, M.; Mohammed, M.; Xin, S.;

Collins, S.; Dias, M. L.; Pinto, J. C. Macromolecules 2001, 34, 3830-

3841. (b) Grisi, F.; Longo, P.; Zambelli, A.; Ewen, J. A. J. Mol. Catal. A:

Chem. 1999, 140, 225. (c) Ewen, J. A. J. Mol. Catal. A: Chem. 1998,

128, 103.

(44) For example, in NMR study of the freshly generated ion pairs, we observe

systems 8 and 9 to undergo rapid decomposition (with 9 faster than 8),

while 7 decomposes slowly and 10 exhibits comparatively high thermal

stability.

9

J. AM. CHEM. SOC. VOL. 126, NO. 14, 2004 4621

A R T I C L E S

Chen et al.

Table 11. Estimated Absolute Rate Constants for Propagation,

Site Epimerization, “Back-Side” Misinsertion, Chain Epimerization,

and Enantiofacial Misinsertion Obtained from Activities, P_mvs

1/[Propylene] Plots, and P_mmvs 1/[Propylene] Plots for

Polymerizations Mediated by 1+ Indicated Cocatalysts under

1.0-5.0 atm Propylene at 60 °C^a

coordinating metrics developed on the basis of dynamic NMR-

derived ion pair reorganization barriers and anion displacement

equilibria.^{1b,4a,7c,8a,12}Surprisingly, the observed trend in k_bsa(L

mol^-1s^-1) also tracks ion pairing strength: FAl(2-C₆F₅C₆F₄)₃

-

(10, 0.486(65)) < MeB(C₆F₅)₃^-(7, 3.61(38)) < B(C₆F₅)₄^-(9,

c

d

k_p^b

k_se

(s^-1

k_bsa

(M^-1*s^-1

k_ce

(s^-1

k_em

(M^-1*s^-1

)

9.5(16)) ≈ MeB(2-C₆F₅C₆F₄)₃(8, 10.9(14) s^-1). If this trend

-

cocat.

(cat)

(M^-1*s^-1

)

accurately reflects selectivity for same-side approach of the

olefin with respect to the counteranion, it seems counterintuitive

from the standpoint of steric factors: certainly a more intimately

bound anion would be more likely to suppress same-side attack.

However, it is possible that a strongly coordinating anion

stabilizes the distribution of positive charge at the catalyst metal

center such that same side attack (as with “normal” chain-

migratory insertion) is favored.²⁸Considering that anion ster-

eochemical mobility, catalyst stability, and reactivity all seem

to depend strongly on the coordinative contribution to the

cation-anion interaction, it seems likely that charge distribution

and stabilization of the cationic moiety is highly anion-

dependent, and that the anion is intimately involved in the

insertion reaction (and in competing processes). The formal

charge of the cationic olefin π-adduct fragment (assuming one

exists) is the same as the isolated catalyst, and it is likely that

such species also persist as contact ion pairs in solution.²⁸

Interaction between the anion and a π-adduct may indeed

inVolVe the olefin (inasmuch as the anion’s charge is itself

localized), the anion possibly assisting in olefin activation.²⁴

This would explain the remarkable fact that the fluoroaluminate

system yields polymer at all, given the remarkable kinetic

inertness of the anion as a ligand in this system. Further evidence

of ion pairing influences can be seen in the trend in enantiofacial

misinsertion (k_em, L mol^-1s^-1), with 10 (0.807(74)) < 7 (1.09-

(11)) , 9 (7.75(62)) < 8 (9.36(65)). This trend in particular

suggests a complex dependence on both electronic and steric

factors.

Chain epimerization is observable in all systems with the

notable exception of catalyst system 10 (k_ce) 0.03(44),

indistinguishable from zero, as with 8, having k_ce) 0.29(37),

a rather large confidence interval; see Table 11) and is also

suppressed, but to a lesser extent, with 7 (k_ce) 0.239(75)).⁴⁶

Conversely, the weakly coordinating B(C₆F₅)₄^-exhibits a chain

epimerization rate constant of 0.69(42), ∼20× greater than that

of the FAl(2-C₆F₅C₆F₄)₃^-anion (although indistinguishable from

10 at the 90% confidence level). The diminished chain epimer-

ization channel observed with 10, if well-represented by the

present data, is possibly due to suppression of â-hydrogen

transfer to Zr (Scheme 1E). A tightly bound counteranion would

a priori be expected to destabilize the 4-center transition state

generally considered necessary for â-hydrogen transfer (termi-

nation via â-hydrogen transfer is undetectable with 10, but also

with 8 and 9).^11a,41This possibility is not inconsistent with (but

by no means conclusively demonstrates) the ideas of: (i) inti-

mate involvement of the counteranion in polymerization events,

and (ii) multiple available, counteranion-differentiated pathways

for monomer activation and enchainment.

2

402(44)

41.8(8)

321(12)

297(4)

9.8(13)

1.85(26)

10.9(10)

11.1(11)

0.474(47)

5.9(13)

3.61(38)

10.9(14)

9.5(16)

1.04(55)

0.239(75) 1.09(11)

0.29(37)

0.69(42)

8.2(12)

3(7)

4(8)

5(9)

6(10)

9.36(65)

7.75(62)

0.807(74)

33.7(16)

0.486(65) 0.03(44)

^aConfidence intervals presented at the 90% confidence level;k_p, k_se, k_bsa

,

k_ce, and k_emas defined in eqs 10 and 11. ^bk_pvalues are extrapolated from

plots of ln(k_p,apparent) vs. 1/T (see text). ^cFrom plots using eq 10. ^dFrom

plots using eq 11.

Figure 10. Plots of -ln(k_p) vs. 1/(polymerization temperature) for 1+

indicated cocatalysts under 1.0 atm propylene over the temperature range

from -10° to +60 °C in toluene (Table 7; k_pvalues corrected for [propylene]

temperature dependence).

dependence data (Figure 10), excluding experiments in which

k_pis obviously underrepresented by k_p,apparent(see Table 7).

Confidence intervals for these projected k_pvalues are given at

the 90% confidence level, as determined from linear regression

analysis of ln(k_p,apparent) vs 1/T for each system. Confidence

intervals for k_se/k_pand k_bsa/k_pwere determined from linear

regression analysis of P_mvs 1/[propylene], and for k_ce/k_pand

k_em/k_p, from P_mmvs 1/[propylene]; these are also given at the

90% confidence level. Confidence intervals for k_se, k_bsa, k_ce, and

k_emare derived from those of the parent quantities. Due to the

presence of large excesses of aluminoxanes in reactions using

2 as cocatalyst, and the large confidence interval associated with

k_pfor this system (402(44) s^-1) these data are excluded from

the present discussion. Also, lack of detailed information on

the structure(s) of the catalyst system obtained using this

cocatalyst impedes interpretation of rate data from a mechanistic

standpoint; data for 2 are presented for consideration in Table

11.

The trends in propagation and site epimerization rates both

track ion pairing strength. For k_p(L mol^-1s^-1), the observed

anion ordering is as follows: FAl(2-C₆F₅C₆F₄)₃^-(10, 33.7(16))

< MeB(C₆F₅)₃^-(7, 41.8(8)) , B(C₆F₅)₄^-(9, 297(4)) < MeB-

The above observed trends mirror ion pairing strength quite

well, and since these orderings arise from estimates of k_p, which

(2-C₆F₅C₆F₄)₃(8, 321(12)). For k_se(s^-1), the ordering is the

-

same: 10 (0.474(47)) < 7 (1.85(26)) , 8 (10.9(10)) ≈ 9

(46) Absolute rates for chain epimerization and enantiofacial misinsertion for

(11.1(11)). These general trends are consistent with the anion

each catalyst at each pressure can be roughly gauged using eq 11 (P_mm

)

and insertion rates. The chain epimerization rates follow the ordering:

3 > 5 > 2 > 4 > 6. It is likely in any event that the various steps involved

in chain epimerization are subject to anion-dependent steric and electronic

influences in a complex manner.

(45) See Table 1 in Supporting Information for additional polymerization results

used together with those in Table 7 to generate rate constants k_pfor Table

11.

9

4622 J. AM. CHEM. SOC. VOL. 126, NO. 14, 2004

Counteranion Effects on Olefin Polymerization

A R T I C L E S

Table 12. Propylene Polymerization Results for the Reactions Mediated by 1+ Indicated Cocatalysts in 1,3-Dichlorobenzene^aor Octane^a

under 1.0 atm of Propylene at 25 °C

exp.

no.

cocat.

(cat.)

cat.

(µmol)

time

(s)

yield^b

(g)

activity^c

(×10⁶)

T_m

(°C)

M ^d

rmmr^e

(%)

mmrr^e

(%)

rrmr^e

(%)

rrrr^e

(%)

w

solvent^a

(kg*mol^-1

)

P.D.I.^d

1,3-dichloro-benzene

1

2

3

4

5

6

7

8

9

2

2.5

10

3.8

8.0

10

3.6

20

5.0

10

120

360

900

120

900

0.56

2.67

0.82

1.21

2.4

0.83

3.16

2.09

2.45

2.47

6.5

N.O.^f

113.8

122.2

125.4

131.6

N.O.^f

110

97.6

93.3

104

127

147

49.7

104

139

154

1.85

1.76

2.24

1.92

1.78

1.86

2.01

1.82

1.83

1.89

1.6

2.3

2.4

2.3

1.3

1.6

1.9

1.5

1.4

4.3

4.6

4.4

4.3

2.8

3.4

4.0

3.2

2.8

18

17

3.4

10

6.5

4.9

1.9

49

50

85

69

77

82

89

3(7)

4(8)

5(9)

6(10)

2

3(7)

4(8)

5(9)

6(10)

2.67

0.86

4.53

0.96

1.4

0.063

1.256

0.49

0.148

octane

600

9000

1200

1800

1500

10

40

^a50 mL, 4 mL toluene injected with catalyst solution. ^bAfter workup (see Experimental Section). ^cUnits: g polymer/(mol cat.*atm*h). ^dDetermined

from GPC analysis relative to polystyrene standards; polydispersity index ) M_w/M_n. ^ePentad integrals from polymer ¹³C NMR. ^fNot observed.

we also observe to track ion pairing strength, it is worthwhile

to entertain the possibility that errors in the k_pestimates, rather

than systematic chemical structure/function relationships, domi-

nate the analysis. The present findings, if valid, do shed light

on some interesting observations¹⁰that have also received

-

attention in recent literature:^2a,bfor example, B(C₆F₅)₄(9) is

found to produce a higher-syndiotacticity polymer than

MeB(C₆F₅)₃^-(7) at all temperatures, even though it is thought

to be more weakly bound to the cation. At [propylene] ) 0.36

M (1 atm. system pressure, T ) 60 °C), the present results give

P_m,7/P_m,9ranging from 1.5 to 2.8, as determined directly from

NMR analysis of the product polymers from these two catalysts.

The aforementioned observations are consistent with a scenario

in which P_m,7/P_m,9< V_p,9/V_p,7(see eq 10 above). Thus at 25 °C,

for example, relative insertion rates V_p,9/V_p,7> 2.8 would give

V_m,9> V_m,7, as expected (P_m× V_p) V_m) V_bsa+ V_se). In such

a scenario, m stereodefect generation proceeds more rapidly in

9 than 7, but by a smaller margin than propagation. In fact,

activity measurements give an estimated V_p,9/V_p,7) 25.9(79),

at 25 °C, suggesting that this is indeed the case. A similar

argument explains comparative polymer M_wvalues: whereas

attenuated â-hydrogen elimination with MeB(C₆F₅)₃^-and FAl-

(2-C₆F₅C₆F₄)₃^-does contribute to increased polymer M_wvalues,

B(C₆F₅)₄^-produces a higher M_wproduct due to its much greater

propagation rate.

Figure 11. A. log(polymerization activity), B. Polypropylene M_w, C. rrrr

pentad intensity (%), and D. xmrx pentad intensity (%) data for polypro-

pylenes produced by 1+ indicated cocatalysts under 1.0 atm of propylene

at 25 °C in octane, toluene, and 1,3-dichlorobenzene solutions.

ion pair complexes in octane is a tenable explanation for the

reduced productivity.^43a

E. Catalyst Concentration and Added Li⁺MeB(C₆F₅)₃

-

Effects on Syndiospecific Polymerizations Mediated by 7.

Increased zirconocenium ion pair concentrations and the addition

of Li⁺MeB(C₆F₅)₃(11) have recently been reported to

-

accelerate anion exchange/catalyst symmetrization processes (cf.,

eq 2).¹⁴An ion quadruple (K) or higher aggregate was proposed

to be the key intermediate in such acceleration. To test the

possible effects of putative ion pair aggregation and the

introduction of lithium counteranion salts on the present

polymerization system, experiments in which the concentration

of 7 was varied, and experiments examining the effect of added

D. Solvent Effects on Polymerization Stereocontrol. In

nonpolar hydrocarbon solvents such as toluene (ꢀ ) 2.15), which

are commonly used for olefin polymerization, strong cation-

anion interactions¹²are doubtless an important modulator of

reactivity in the present class of catalysts.²

To probe the effects of polar solvation-induced ion pair

weakening on enchainment stereochemistry, polymerizations

were also carried out in more polar 1,3-dichlorobenzene (ꢀ )

5.04; Table 12). The net result is dramatic compression in the

dispersion of polymerization rates and collapse of rrrr, m, and

mm stereosequence percentages to the experimentally indistin-

guishable values of 50%, 17.5%, and 4%, respectively, for all

cocatalysts studied, indicating that polar solvents significantly

weaken ion pairing effects on stereocontrol in this system

(Figure 11).⁴⁷In contrast, polymerizations in less polar octane

(ꢀ ) 2.08; Table 12) evidence trends similar to those in toluene,

but with more dramatic decreases in polymerization activity and

slightly lower to negligible changes in product syndiotacticities.

Although stronger ion pairing effects in octane vs toluene can

be used to explain the observed decrease in activity in 9 and

10 (95% and 25% respectively), the lower solubility of these

Li⁺MeB(C₆F₅)₃on propylene polymerizations catalyzed by

-

1 + 3, were carried out. A priori, structures such as K might

be expected to exhibit enhanced degrees of stereochemical

mobility, which would consequently erode product syndiotac-

ticity. Indeed, evidence for¹⁴and against^12,14increased catalyst

(47) (a) Propylene solubilities at 1 atm can be expected to vary somewhat with

solvent. However, with 1,3-dichlorobenzene as solvent we observe that

whereas activity (hence insertion rate) increases, so also does the rate of

xmrx steric pentad formation, relatiVe to insertion. On the basis of what

we have demonstrated here from the propylene concentration study, we

would expect the opposite, were the effect exclusively a concentration effect.

(b) Herfert, N.; Fink, G. Makromol. Chem. 1992, 193, 773-778. (c)

Coevoet, D.; Cramail, H.; Deffieux, A. Makromol. Chem. Phys. 1999, 200,

1208-1214. (d) For solvent effects in C₂-symmetric catalyst system, Forlini,

F.; Tritto, I.; Locatelli, P.; Sacchi, M. C.; Piemontesi, F. Makromol. Chem.

Phys. 2000, 201, 401-408. (d) For CGC catalyst systems, see: Klein-

schmidt, R.; Griebenow, Y.; Fink, G. J. Mol. Catal. A-Chem. 2000, 157,

83-90.

9

J. AM. CHEM. SOC. VOL. 126, NO. 14, 2004 4623

A R T I C L E S

Chen et al.

Table 13. Concentration Effects on Propylene Polymerization Results for the Reactions Mediated by 1+ B(C₆F₅)₃(3) in Toluene under 1.0

atm of Propylene at 25 °C^a

exp.

no.

cat.

(µmol)

time

(s)

yield^b

(g)

activity^c

(×10⁶)

T_m

(°C)

M ^d

rmmr^e

(%)

mmrr^e

(%)

rrmr^e

(%)

rrrr^e

(%)

w

(kg*mol^-1

)

P.D.I.^d

1

2

3

4

5

6

800

400

200

100

50

600

900

2400

3600

7200

4.92

2.30

5.90

2.59

0.86

2.10

0.37

0.46

0.44

0.39

0.17

0.42

102.5

101.4

102.5

101.0

101.3

56.9

66.4

79.4

77.5

76.0

81.9

1.66

1.7

1.81

1.85

1.78

1.83

1.6

1.5

1.6

1.5

1.7

3.3

3.1

3.2

11.0

10.5

10.6

10.5

10.8

10.9

68.1

69.5

69.4

69.6

69.2

68.7

25

^aIn 104 mL toluene. ^bAfter workup (see Experimental Section). ^cUnits: g polymer/(mol cat.*atm*h). ^dDetermined from GPC analysis relative to polystyrene

standards; polydispersity index ) M_w/M_n. ^ePentad integrals from ¹³C NMR.

-

Table 14. Propylene Polymerization Results for the Reactions Mediated by 1 + 3 with Addition of Li⁺MeB(C₆F₅)₃(11) in Toluene under

1.0 atm of Propylene at 25 °C^a

exp.

no.

cat.

(µmol)

added 11

(µmol)

time

(s)

yield^b

(g)

activity^c

(×10⁶)

T_m

(°C)

M ^d

rmmr^e

(%)

mmrr^e

(%)

rrmr^e

(%)

rrrr^e

(%)

w

(kg*mol^-1

)

P.D.I.^d

1

2

3

4

5

6

20

0

20

40

60

80

2400

1800

1200

5.90

2.21

4.02

3.45

4.43

4.90

0.44

0.17

0.40

0.52

0.67

0.74

101.4

109.5

108.5

114.5

112.2

114.6

79.0

84.6

83.6

86.6

84.5

93.0

1.81

1.78

1.84

1.8

1.82

1.84

1.5

1.7

1.8

1.5

1.6

1.5

3.1

3.3

2.9

3

10.6

9.2

9.4

8

8.9

7.8

69.4

72

70.9

75.6

73.2

76

100

2.9

^aIn 104 mL toluene. ^bAfter workup (see Experimental). ^cUnits: g polymer/(mol cat.*atm*h). ^dDetermined from GPC analysis relative to polystyrene

standards; polydispersity index ) M_w/M_n. ^ePentad integrals from ¹³C NMR.

present conditions, either the ion exchange rate (if exchange

occurs at all) is dictated by a slow dissociative step, and is thus

invariant with [Li⁺MeB(C₆F₅)₃^-], or that exchange occurs

associatively but without stereoinversion, i.e., through same-

side attack (compare eqs 12a and 12b), as proposed by

Brintzinger et al.⁴⁹As in the concentration-dependence studies

of 7, catalyst activity and product syndiotacticity are modestly

to negligibly invariant to added Li⁺MeB(C₆F₅)₃^-, while product

site epimerization rates (eq 2) with increasing catalyst concen-

M_wmay be more sensitive.

tration in the absence of olefin has been presented in the

literature. However, in the present experiments with 7, catalyst

activity and product syndiotacticity were found to be essentially

concentration-invariant over a 32-fold catalyst concentration

range, while product M_wdata decline only modestly at the

highest concentrations (Table 13). These results argue that under

typical polymerization conditions, where [catalyst] ) 25-800

µM (in contrast with the aggregation experiments, in which

[catalyst] ) 2-20 mM),¹⁴inter-ion pair exchange via aggrega-

tion is not an important factor influencing activity or enchain-

ment stereochemistry. In work presented elsewhere,^28,48ion-

pair aggregation is shown by cryoscopy and pulsed field gradient

spin-echo NMR spectroscopy to be insignificant in benzene

Conclusions

A series of stable, structurally well-characterized, highly

reactive C_s-symmetric zirconocenium ion pair propylene po-

lymerization catalysts has been studied with regard to the

molecular and ion pair structure and structural dynamics, both

in the solid state and in solution. Ion-pairing differences are

evaluated on the basis of detailed spectroscopic/crystallographic

characterization and ion pair reorganization/symmetrization

kinetics, and reveal strongly anion-dependent correlations with

product polypropylene molecular weight and microstructural

features. A distinctive signature of the catalyst-cocatalyst

interaction emerges: polymerization activity, polymer micro-

structure, and molecular weight, in particular the relative rates

of termination pathways and stereodefect-generating side reac-

or toluene solutions for a broad range of single-site metalloce-

-

nium MeB(C₆F₅)₃or B(C₆F₅)₄ion pairs, even at concentra-

tions substantially higher than employed here. Taken together,

these findings argue that formation of ion quadruples (K) or

higher-order aggregates is unlikely to be of importance in the

present zirconocene-based catalyst systems for single-site R-ole-

fin polymerization.

Experiments examining the effects of added Li⁺MeB(C₆F₅)₃

-

(11) on propylene polymerizations catalyzed by 1 + 3 reveal

at most a minor increase in product M_w, syndiotacticity, or

melting point with increased 11 concentrations over a broad

range (Table 14). This observation suggests that under the

(48) (a) Stahl, N. G.; Zuccaccia, C.; Jensen, T. R.; Marks, T. J. J. Am. Chem.

Soc. 2003, 125, 5256-5257. (b) Stahl, N. G.; Marks, T. J.; Macchioni, A.;

Zuccaccia, C. Presented in part at the 222nd ACS National Meeting,

Chicago, IL, August 2001, Abstract INORG 407.

-

(49) Li⁺MeB(C₆F₅)₃is not assumed to be a solvent-separated or dissociated

ion pair.

9

4624 J. AM. CHEM. SOC. VOL. 126, NO. 14, 2004

Counteranion Effects on Olefin Polymerization

A R T I C L E S

tions relative to syndiospecific propylene enchainment, are all

highly sensitive to the sterics and energetics of cocatalyst

binding.

the ion pairing interactions, all else being equal. However, due

to its tendency to form [Zr-(µ-Me)-Zr]⁺structures, the latter

system is more complex. It is evident, however, that the

-

Comparison of solid-state structures demonstrates that the

B(C₆F₅)₄counteranion is the most weakly bound, exhibiting

-

tightly bound FAl(2-C₆F₅C₆F₄)₃counteranion actually draws

the most rapid production of m and mm stereodefects, and by a

greater margin, the most rapid chain propagation. Consistently

across the present series, polymerization activity decreases and

the rate of catalyst site epimerization decreases, as the ion

pairing strength is increased. Estimation of absolute rates for

propagation, site epimerization, backside misinsertion, enan-

tiofacial misinsertion, and chain epimerization for each of these

cation-anion systems provides a complete and self-consistent

explanation of the relative syndiotacticities of product polymers

from each system. Spectroscopic, theoretical,^24,51and polym-

erization studies argue that polar solvents significantly weaken

ion pairing, and in accord with a picture in which ion pairing

modulates syndiospecific enchainment, we find here that dif-

ferential anion effects on propagation rates diminish and those

on stereodefects completely Vanish in a more polar solvent.

While the present results serve to elucidate the importance of

catalyst-cocatalyst interactions in the production of syndiotactic

polypropylene,⁵²they also reveal a unique new feature of the

the [Me₂C(Cp)(Flu)]ZrMe⁺Zr center slightly further out of the

ligand pocket than does MeB(C₆F₅)₃^-. Reciprocal effects on

anion structure are larger: differences in flattening of the Group

13 atom geometry and lengthening of the Group 13 atom bond

to the bridging moiety with coordination of the anion, demon-

strate strong differences in the cation-anion interaction in 7

and 10. These counteranion differences are further manifested

in the rate constants of dynamic unimolecular reorganization

processes in isolated [Me₂C(Cp)(fluorenyl)]ZrMe⁺X^-ion pairs;

from NMR kinetic analysis, we find that the fluoroaluminate

ion pair has a far higher barrier to reorganization: ∆G^‡> 24.8

-

kcal/mol vs 21.3(36) kcal/mol for MeB(C₆F₅)₃at 127.5 °C.

Ion pair reorganization is herein assumed to be kinetically

accessible to the metallocene Group 4 olefin polymerization

catalysts as a class, but possibly requires the presence of olefin

in the fluoroaluminate system (10). This is consistent with the

commonly accepted chain-swinging model developed in con-

junction with metallocene polymerization catalysts for stereo-

regular propylene polymerization.^11,50

The above observations and conclusions help reconcile the

present accumulated evidence for appreciable counteranion/

cocatalyst influences on product polymer features with the

complex manifold of processes proposed to be kinetically

accessible during polymerization. Stereodefect frequencies and

molecular weights examined as a function of polymerization

temperature and monomer concentration across the entire

cocatalyst series allow quantitation of anion effects on a

-

FAl(2-C₆F₅C₆F₄)₃counteranion. This remarkable species,

showing the greatest affinity for the cationic zirconocenium

fragment, exhibits the highest, least temperature-dependent

syndioselectivity. This unprecedented, cocatalyst-derived sta-

bilization suggests completely new strategies for selectivity

enhancement in single-site polymerization processes.

Acknowledgment. Financial support by DOE (DE-FG02-

86ER1351) is gratefully acknowledged. M.-C. C. thanks Dow

Chemical for a postdoctoral fellowship and Dr. P. Nickias of

Dow for GPC measurements. We also thank Dr. L. Li, Dr. H.

Ahn, Dr. C. Zuccaccia, Dr. T. R. Jensen, and Mr. N. G. Stahl

for helpful discussions.

collection of processes, and by extension, comparison of FAl-

-

(2-C₆F₅C₆F₄)₃^-, MeB(C₆F₅)₃^-, B(C₆F₅)₄^-, MeB(2-C₆F₅C₆F₄)₃

and MAO. Comparing FAl(2-C₆F₅C₆F₄)₃and MeB(C₆F₅)₃

,

-

Supporting Information Available: Complete X-ray experi-

mental details and tables of bond lengths, angles, and positional

parameters for the crystal structures of 7 and 10, additional

polymerization results as an extension to results in Table 11

above, and a full-page, low peak-threshold EXSY spectrum for

10 at 127.5 °C, τ_m) 800 ms (15 pages, print/PDF). This

material is available free of charge via the Internet at

http://pubs.acs.org.

we find that the latter exhibits a greater proclivity toward

unimolecular reorganization/symmetrization and termination

processes than does the former, suggesting not only that the

difference in ion pairing strength persists during polymerization,

but that the fluoroaluminate anion suppresses chain transfer more

strongly than does the methylborate. The catalytic activities and

-

polymer syndiotacticities exhibited by MeB(2-C₆F₅C₆F₄)₃lie

-

between those of MeB(C₆F₅)₃and B(C₆F₅)₄^-, in accord with

previous evidence suggesting that, as compared with the former,

the bulkier ancillary structure of MeB(2-C₆F₅C₆F₄)₃reduces

JA036288K

-

(51) (a) Beswick, C. L.; Marks, T. J. Organometallics 1999, 18, 2410-2412.

(b) Deck, P. A.; Beswick, C. L.; Marks, T. J. J. Am. Chem. Soc. 1998,

120, 12 167-12 167. (c) and also see ref. 12.

(52) Chen, M.-C.; Roberts, J. A. S.; Marks, T. J. Organometallics 2004, 23,

932-935.

(50) (a) Cossee, P. Tetrahedron Lett. 1960, 17, 12. (b) Cossee, P. Tetrahedron

Lett. 1960, 17, 17. (c) Arlman, E. J.; Cossee, P. J. Catal. 1964, 3, 99.

Cossee, P. J. Catal. 1964, 3, 80. Cossee, P.

9

J. AM. CHEM. SOC. VOL. 126, NO. 14, 2004 4625

Article Doi

Marked Counteranion Effects on Single-Site Olefin Polymerization Processes. Correlations of Ion Pair Structure and Dynamics with Polymerization Activity, Chain Transfer, and Syndioselectivity

DOI: 10.1021/ja036288k

Source and publish data:

Authors:

Article abstract of DOI:10.1021/ja036288k

Full text of DOI:10.1021/ja036288k

Products guided by the article

R&D Labs maybe for 206852-73-5

Relevant to this article

Hot Product