Highly electrophilic olefin polymerization catalysts. Quantitative reaction coordinates for fluoroarylborane/alumoxane methide abstraction and ion-pair reorganization in group 4 metallocene and 'constrained geometry' catalysts

Article abstract of DOI:10.1021/ja972912f

Reaction enthalpies of group 4 metallocenes having the general formula L₂M(CH₃)₂ (L = Cp, 1,2-Me₂Cp, Me₅Cp; L₂ = Me₂Si(Me₄Cp)((t)BuN); M = Ti, Zr, and Hf) with the strong organo-Lewis acid B(C₆F₅)₃ were measured using batch titration calorimetry in toluene. Methide abstraction to form the corresponding L₂MCH₃⁺CH₃B(C₆F₅)₃^- contact ion pairs is highly exothermic in all cases. Exothermicity increases with increasing Cp methyl substitution: for M = Zr, ΔH = -23.1(3), -24.3(4), and -36.7(5) kcal mol^-1 for L = Cp, Me₂Cp, and Me₅Cp, respectively for M = Hf and L = 1,2-Me₂Cp, ΔH = -20.8(5) kcal mol^-1. 'Constrained geometry' complexes (L₂ = Me₂Si(Me₄Cp)((t)BuN)) exhibit similar exothermicities, with ΔH = -22.6(2), -23.9(4), and -19.3(6) kcal mol^-1 for M = Ti, Zr, and Hf, respectively. In contrast, analogous reactions with methylalumoxane (M:Al = 1:50) are less exothermic, with ΔH = -10.9(3) and -8.9(4) kcal mol^-1 for L = 1,2-Me₂Cp and M = Zr and Hf, respectively. Under identical conditions, (1,2-Me₂Cp)₂M-(CH₃)₂ (M = Zr, Hf) complexes also undergo methide abstraction with the less Lewis-acidic triarylboranes (C₆F₅)2BAr (AT = 3,5-C₆H₃F₂, Ph, and 3,5-C₆H₃Me₂); however, conversions to the corresponding (Me₂-Cp)₂MCH₃⁺ CH₃B(C₆F₅)₂Ar^- ion pairs, are incomplete. Variable-temperature NMR measurements yield thermodynamic parameters for partial methide abstraction by these less Lewis-acidic boranes. For Ar = 3,5-C₆H₃F₂, ΔH = -18.7(7) and -15.2(8) kcal mol^-1 with ΔS = -42(2) and -35(3) e.u:; for Ar = Ph, ΔH = -14.8(8) and -13.3(6) kcal mol^-1 with ΔS = -31(2) and -39(2) e.u.; for Ar = 3,5-C₆H₃Me₂, ΔH = -10.8(6) and -12.7(5) with ΔS = -19(2) and -36(4) e.u., in each case for M = Zr and Hf, respectively. Dynamic NMR analyses reveal that the activation barriers for methide abstraction from the neutral metallocene dialkyls are small and relatively insensitive to the borane identity (AH = 2-6 kcal mol^-1) while ion-pair separation/recombination processes are greatly facilitated by polar solvents. Ethylene polymerization activities for eight (Me₂Cp)₂MCH₃⁺CH₃B(C₆F₅)₂Ar^- complexes measured in toluene solution (25°C, 1 atm) follow a trend in metal (Zr > Hf)as well as a substantial trend in triarylborane (Ar = C₆F₅ > 3,5-C₆H₃F₂ > Ph ~3,5-C₆H₃-Me₂). Polymerization activities correlate roughly with MCH₃^{+ 13}C NMR chemical shifts and enthalpies of methide abstraction.

Full text of DOI:10.1021/ja972912f

1772
J. Am. Chem. Soc. 1998, 120, 1772-1784
Highly Electrophilic Olefin Polymerization Catalysts. Quantitative
Reaction Coordinates for Fluoroarylborane/Alumoxane Methide
Abstraction and Ion-Pair Reorganization in Group 4 Metallocene and
“Constrained Geometry” Catalysts
Paul A. Deck, Colin L. Beswick, and Tobin J. Marks*
Contribution from the Department of Chemistry, Northwestern UniVersity, EVanston, Illinois 60608-3113
ReceiVed August 18, 1997
Abstract: Reaction enthalpies of group 4 metallocenes having the general formula L2M(CH3)2 (L ) Cp, 1,2-
Me2Cp, Me5Cp; L2 ) Me2Si(Me4Cp)(tBuN); M ) Ti, Zr, and Hf) with the strong organo-Lewis acid B(C6F5)3
were measured using batch titration calorimetry in toluene. Methide abstraction to form the corresponding
-
L2MCH3+CH3B(C6F5)3 contact ion pairs is highly exothermic in all cases. Exothermicity increases with
increasing Cp methyl substitution: for M ) Zr, ∆H ) -23.1(3), -24.3(4), and -36.7(5) kcal mol-1 for L )
Cp, Me2Cp, and Me5Cp, respectively. For M ) Hf and L ) 1,2-Me2Cp, ∆H ) -20.8(5) kcal mol-1
.
“Constrained geometry” complexes (L2 ) Me2Si(Me4Cp)(tBuN)) exhibit similar exothermicities, with ∆H )
-22.6(2), -23.9(4), and -19.3(6) kcal mol-1 for M ) Ti, Zr, and Hf, respectively. In contrast, analogous
reactions with methylalumoxane (M:Al ) 1:50) are less exothermic, with ∆H ) -10.9(3) and -8.9(4) kcal
mol-1 for L ) 1,2-Me2Cp and M ) Zr and Hf, respectively. Under identical conditions, (1,2-Me2Cp)2M-
(CH3)2 (M ) Zr, Hf) complexes also undergo methide abstraction with the less Lewis-acidic triarylboranes
(C6F5)2BAr (Ar ) 3,5-C6H3F2, Ph, and 3,5-C6H3Me2); however, conversions to the corresponding (Me2-
Cp)2MCH3 CH3B(C6F5)2Ar- ion pairs are incomplete. Variable-temperature NMR measurements yield
+
thermodynamic parameters for partial methide abstraction by these less Lewis-acidic boranes. For Ar ) 3,5-
C6H3F2, ∆H ) -18.7(7) and -15.2(8) kcal mol-1 with ∆S ) -42(2) and -35(3) e.u.; for Ar ) Ph, ∆H )
-14.8(8) and -13.3(6) kcal mol-1 with ∆S ) -31(2) and -39(2) e.u.; for Ar ) 3,5-C6H3Me2, ∆H ) -10.8-
(6) and -12.7(5) with ∆S ) -19(2) and -36(4) e.u., in each case for M ) Zr and Hf, respectively. Dynamic
NMR analyses reveal that the actiVation barriers for methide abstraction from the neutral metallocene dialkyls
are small and relatively insensitive to the borane identity (∆Hq ) 2-6 kcal mol-1) while ion-pair separation/
recombination processes are greatly facilitated by polar solvents. Ethylene polymerization activities for eight
(Me2Cp)2MCH3 CH3B(C6F5)2Ar- complexes measured in toluene solution (25 °C, 1 atm) follow a trend in
+
metal (Zr > Hf) as well as a substantial trend in triarylborane (Ar ) C6F5 > 3,5-C6H3F2 > Ph ∼ 3,5-C6H3-
Me2). Polymerization activities correlate roughly with MCH3+ 13C NMR chemical shifts and enthalpies of
methide abstraction.
Introduction
number of these species have now been structurally
characterized.2a,3
Transformations attending alkyl or hydride anion abstraction
from group 4 complexes by strong Brønsted or Lewis acids are
currently of great interest owing to their central significance in
the activation and function of homogeneous “single-site”
Ziegler-Natta catalysts1 based on metallocene1 and quasi-
metallocene1,2 frameworks. Synthetic studies have shown that
with carefully selected metallocenes, Lewis acidic reagents, and
reaction conditions, this abstraction process yields highly
electrophilic, mononuclear ion pairs (e.g., A), which exhibit
unprecedented olefin polymerization activity and selectivity. A
In reality, “simple” functional models for homogeneous
Ziegler-Natta catalysts remain scarce. Surprisingly few Lewis
(2) For representative recent group 4 work involving nonmetallocene
skeletons, see: (a) Baumann, R.; Davis, W. M.; Schrock, R. R. J. Am. Chem.
Soc. 1997, 119, 3830-3831. (b) Scollard, J. D.; McConville, D. H. J. Am.
Chem. Soc. 1996, 118, 10008-10009. (c) Gillis, D. J.; Quyoum, R.; Tudoret,
M.-J.; Wang, Q.; Jeremic, D.; Roszak, A. W.; Baird, M. C. Organometallics
1996, 15, 3600-3605. (d) Mack, H.; Eisen, M. S. J. Organomet. Chem.
1996, 525, 81-7. (e) Shsh, S. A. A.; Dorn, H.; Voigt, A.; Roesky, H. W.;
Parisini, E.; Schmidt, H.-G.; Noltemeyer, M. Organometallics 1996, 15,
3176-81. (f) Horton, A. D.; de With, J.; van der Linden, A. J.; van de
Weg, H. Organometallics 1996, 15, 2672-2674. (g) Cloke, F. G. N.;
Geldbach, T. J.; Hitchcock, P. B., Love, J. B. J. Organomet. Chem. 1996,
506, 343. (h) Van der Linden, A.; Schaverien, C. J.; Meijboon, N.; Ganter,
C.; Orpen, A. G. J. Am. Chem. Soc. 1995, 117, 3008-21. (i) Pellecchia,
C.; Grassi, A.; Zambelli, A. J. Mol. Catal. 1994, 82, 57-65. (j) Uhrhammer,
R.; Black, D. G.; Gardner, T. G.; Olsen, T. D.; Jordan, R. F. J. Am. Chem.
Soc. 1993, 115, 8493-4.
(1) For recent reviews, see: (a) Kaminsky, W.; Arndt, M. AdV. Polym.
Sci. 1997, 127, 144-187. (b) Bochmann, M. J. Chem. Soc., Dalton Trans.
1996, 255-270. (c) Grubbs, R. H.; Coates, G. W. Acc. Chem. Res. 1996,
29, 85-93. (d) Brintzinger, H. H.; Fischer, D.; Mu¨lhaupt, R.; Rieger, B.;
Waymouth, R. M. Angew. Chem., Int. Ed. Engl. 1995, 34, 1143-1170. (e)
Guram, A. S.; Jordan, R. F. In ComprehensiVe Organometallic Chemistry;
Abel, E. W., Stone, F. G. A., Wilkinson, G., Eds.; Elsevier: Oxford, 1995;
Vol II, Chapter 12. (f) Soga, K.; Terano, M., Eds. Catalyst Design for Tailor-
Made Polyolefins; Elsevier: Tokyo, 1994. (g) Mo¨hring, P. C.; Coville, N.
J. J. Organomet. Chem. 1994, 479, 1-29. (h) Marks, T. J. Acc. Chem. Res.
1992, 25, 57-65.
S0002-7863(97)02912-0 CCC: $15.00 © 1998 American Chemical Society
Published on Web 02/14/1998

Electrophilic Olefin Polymerization Catalysts
J. Am. Chem. Soc., Vol. 120, No. 8, 1998 1773
Figure 1. Schematic enthalpic reaction coordinates for metallocenium ion pair reorganization processes: (A) ion pair separation/recombination
(ips), (B) Lewis acid methide abstraction/recombination (dr).
acids react with metallocenes to cleanly afford soluble com-
plexes with desirable catalytic activities, and still fewer afford
rigorously characterized active species. Effective cocatalysts
include Ph3C+ (with appropriate weakly coordinating counter-
anions),4 perfluoroarylboranes,4,5 and MAO (methylalumox-
ane).1,6 The most important common features appear to be: (1)
high native acidity of the electron-deficient centers and (2) lack
of labile nucleophilic substituents that might serve as catalyst
poisons. For example, simple trihaloboranes (BF3, BCl3)
irreversibly transfer F- or Cl- to the metal center affording
inactive metallocene halides,1 while alkylaluminum halides lead
to M(µ-Cl)Al structures, which have generally proven chal-
lenging to characterize and exhibit only modest catalytic
activity.7 Even in the absence of halogens and other donors,
“base-free” cationic metallocenes can form µ-Me dinuclear
species (B)1,3d,f,4a which likely stabilize the highly electrophilic
metal fragments.3d,4a,8 In contrast to these qualitative observa-
tions, quantitative aspects of this chemistry remain virtually
undocumented.
Complex temperature-dependent chemical/structural equilibria
are ubiquitous in metallocenium chemistry. The simplest of
these is the reVersible abstraction of alkide or hydride ligands
by a Lewis acid (A, eq 1; R ) alkyl, H). With A ) Ph3C+,
(3) (a) Sun, Y.; Spence, R. E. V. H.; Piers, W. E.; Parvez, M.; Yap, G.
P. A. J. Am. Chem. Soc. 1997, 119, 5132-5143 and references therein. (b)
Chen, Y.-X.; Stern, C. L.; Marks, T. J. J. Am. Chem. Soc. 1997, 119, 2582-
2583. (c) Wang, Q.; Gillis, D. J.; Quyoum, R.; Jeremic, D.; Tudoret, M.-J.;
Baird, M. C.; J. Organomet. Chem. 1997, 527, 7-14. (d) Chen, Y.-X.;
Stern, C. L.; Yang, S. T.; Marks, T. J. J. Am. Chem. Soc. 1996, 118, 12451-
2. (e) Yang, X.; Stern, C. L.; Marks, T. J. J. Am. Chem. Soc. 1994, 116,
10015-31. (f) Bochmann, M.; Lancaster, S. J.; Hursthouse, M. B.; Malik,
K. M. A. Organometallics 1994, 13, 2235-43. (g) Pellechia, C.; Grassi,
A.; Immirzi, A. J. Am. Chem. Soc. 1993, 115, 1160-1162. (h) Yang, X.;
King, W. A.; Sabat, M.; Marks, T. J. Organometallics 1993, 12, 4254-8.
(i) Yang, X.; Stern, C. L.; Marks, T. J. Angew. Chem., Int. Ed. Engl. 1992,
31, 1375. (j) Horton, A. D.; Orpen, A. G. Organometallics 1991, 10, 3910-
18. (k) Yang, X.; Stern, C. L.; Marks, T. J. J. Am. Chem. Soc. 1991, 113,
3623-5.
(4) (a) Jia, L.; Yang, X.; Stern, C. L.; Marks, T. J. Organometallics 1997,
16, 842-857 and references therein. (b) Xie, Z.; Manning, J.; Reed, R.
W.; Mathur, R.; Boyd, P. D. W.; Benesi, A.; Reed, C. A. J. Am. Chem.
Soc. 1996, 118, 2922-2928 and references therein. (c) Jia, L.; Yang, X.;
Ishihara, A.; Marks, T. J. Organometallics 1995, 14, 5-7. (d) Jia, L.; Yang,
X.; Stern, C. L.; Marks, T. J. Organometallics 1994, 13, 3755-3757. (e)
Strauss, S. H. Chem. ReV. 1993, 93, 927-942. (f) Chien, J. C. W.; Tsai,
W.-M.; Rausch, M. D. J. Am. Chem. Soc. 1991, 113, 8570-8571. (g) Yang,
X.; Stern, C. L.; Marks, T. J. Organometallics 1991, 10, 840-842. (h) Ewen,
J. A.; Elder, M. J. Eur. Pat. Appl. 1991, 426637; Chem. Abstr 1991, 115,
136987c, 136988d. (i) Hlatky, G. G.; Upton, D. J.; Turner, H. W. U S Pat.
Appl. 1990, 459921; Chem. Abstr. 1991, 115, 256897v.
(5) (a) Wu, Z.; Jordan, R. F.; Petersen, J. L. J. Am. Chem. Soc. 1995,
117, 5867-5868. (b) Pellecchia, C.; Pappalardo, D.; Oliva, L.; Zambelli,
A. J. Am. Chem. Soc. 1995, 117, 6593-6594. (c) Temme, B.; Erker, G.;
Karl, J.; Luftmann, H.; Fro¨hlich, R.; Kotila, S. Angew. Chem., Int. Ed. Engl.
1995, 34, 1755-1757. (d) Bochmann, M.; Lancaster, S. J.; Hursthouse, M.
B.; Malik K. M. A. Organometallics 1994, 13, 2235-2243. (e) Gillis, D.
J.; Tudoret, M.-J.; Baird, M. C. J. Am. Chem. Soc. 1993, 115, 2543-2545.
(f) Ewen, J. A.; Elder, M. J. Chem. Abstr. 1991, 115, 136998g.
(6) (a) For a recent review of the properties of MAO, see: Reddy, S. S.;
Sivaram, S. Prog. Polym. Sci. 1995, 20, 309-367. (b) Harlan, C. J.; Bott,
S. G.; Barron, A. R. J. Am. Chem. Soc. 1995, 117, 6465-6474. (c) Sishta,
C.; Hathorn, R.; Marks, T. J. J. Am. Chem. Soc. 1992, 114, 1112-4. (d)
Pasynkiewicz, S. Polyhedron 1990, 9, 429-453.
L2MR2 + A a L2MR+RA-
(1)
this equilibrium likely lies far to the right, while with A ) MAO
and B(C6F5)3, those factors affecting the position of this
equilibrium are presently unclear, however the relative acidities
of A and the metallocenium cation, as well as the energetics of
contact ion-pairing, must certainly be chief among them. While
studies of ligand steric and electronic effects,1 solvent effects,9
and anion effects3b,d,4,10 offer important incremental improve-
ment in understanding these ion pairs, the kinetic and thermo-
dynamic constraints under which neutral dialkylmetallocenes
are converted to metallocenium species remain largely unquan-
tified. Such information is crucial to understanding catalyst
stability, activity, and stereoselectivity.
(7) (a) Eisch, J. J.; Pombrik; S. I.; Zheng, G.-X. Organometallics 1993,
12, 3856-3863. (b) Kaminsky, W.; Engehausen, R.; Zounis, K.; Spaleck,
W.; Rohrmann, J. Makromol. Chem. 1992, 193, 1643-51. (c) Mo¨hring,
P.; Coville, N. J. J. Mol. Catal. 1992, 77, 41-50.
(8) (a) Chen, Y.-X.; Marks, T. J. Organometallics 1997, 16, 3649-3657.
(b) Beck, S.; Brintzinger, H.-H.; Goretzki, R.; Herfert, N.; Fink, G. J. Mol.
Catal. 1989, 111, 67.
(9) (a) Chien, J. C. W.; Song, W.; Rausch, M. D. J. Polym. Sci., Part A:
Polym. Chem. 1994, 32, 2387-93. (b) Herfert, N.; Fink, G. Makromol.
Chem. 1992, 193, 773-8.
(10) (a) Shiomura, T.; Asanuma, T.; Inoue, N. Macromol. Rapid
Commun. 1996, 17, 9-14. (b) Eisch, J. J.; Pombrik, S. I.; Gurtzgen, S.;
Rieger, R.; Vzick, W. In ref 1e, pp 221-235. (c) Giardello, M. A.; Eisen,
M. S.; Stern, C. L.; Marks, T. J. J. Am. Chem. Soc. 1995, 117, 12114-
12129. (d) Vizzini, J. C.; Chien, J. C. W.; Gaddam, N. B.; Newmark, R. A.
J. Polym. Sci., Part A.: Polym. Chem. 1994, 32, 2049-56.

1774 J. Am. Chem. Soc., Vol. 120, No. 8, 1998
Deck et al.
Isotec and was freshly vacuum-transferred from Na/K. Dimethylcy-
clopentadiene was synthesized by the method of Skattebøl18 and
converted to the corresponding lithium salt by treatment with n-BuLi
in hexanes.
In the present contribution, we quantify metallocenium ion-
pair formation/reorganization thermodynamics and kinetics as
represented in Figure 1 for a series of variously substituted Ti,
Zr, and Hf metallocene and quasi-metallocene dialkyls with a
series of triarylborane and methylalumoxane organo-Lewis
acids.11 We show that enthalpies and entropies of methide
abstraction (eq 1; R ) CH3) can be determined by titration
NMR Measurements. Routine characterization data were obtained
using Varian GEM-300, VXRS-300, and Unity 400-Plus instruments.
Variable-temperature 1H NMR experiments for line-shape analysis and
equilbrium constant determinations were carried out on the U400+
instrument using a 5-mm inverse detection probe calibrated with
ethylene glycol and methanol temperature standards. Samples were
contained in Teflon-valved NMR tubes which were loaded in the
glovebox with scrupulously dry, deoxygenated solvent distilled on the
high vacuum line. Equilibrium studies were carried out at low
concentrations (each precisely known, but nominally 1-4 mM) in order
that at least one integrable signal for each important species could be
independently measured at four or five temperatures (10 °C intervals)
in the slow-exchange regime. Pulse delays of 5 s were found to be
adequate for obtaining reliable integration of 1H NMR signals.
Exchange-broadening phenomena were found to be independent of
concentration over a 20-fold range. EXSY experiments were carried
out according to the methods of Perrin and Dwyer using nominal
analytic concentrations of 10 mM.12
1
calorimetry or by H NMR equilibration measurements, while
kinetic aspects can be assessed by dynamic NMR line shape
analysis or by 2D EXSY techniques.12 Furthermore, because
triarylboranes can serve as precursors to discrete, well-defined
metallocenium catalysts, we have synthesized and characterized
three new arylbis(pentafluorophenyl)boranes as well as the
corresponding ion-paired adducts with (1,2-Me2Cp)2M(CH3)2
(M ) Zr, Hf) to obtain complete thermodynamic and kinetic
methide abstraction profiles as a function of borane Lewis
acidity. The effects of metallocene ligation and triarylborane
substitution on the thermodynamic and kinetic parameters of
methide abstraction are quantified. The electrophilicity of the
metallocenium species is additionally examined by measuring
metal-bound 13CH3 NMR parameters. Finally, possible cor-
relations with ethylene polymerization activities are examined.
Due to the limited solubility of a number of the metallocene-borane
adducts in toluene at temperatures in the regime of slow exchange,
metallocene methyl ligands for 13C chemical shift measurements were
isotopically enriched in order to enhance 13C NMR spectral sensitivity.
Alkylation of the corresponding dichlorides was carried out using
toluene suspensions of fully enriched 13CH3Li prepared from 13CH3I
(Cambridge Isotope Laboratories) and lithium in ether followed by
recrystallization of the crude product from pentane at -30 °C. Mixtures
of isotopically enriched metallocene (nominally 9 mmol) and a slight
stoichiometric excess of triarylborane in toluene-d8 were examined at
-60 °C using the Varian VXRS-300 instrument. Chemical shifts (δC)
and coupling constants (1JCH) for the slightly exchanged-broadened
metal methyl (M-CH3) and quadrupolar-broadened boron-methyl
(M‚‚‚CH3-B) signals are provided in Table 6.
Reaction Calorimetry. Enthaplies of reaction between metallocenes
and Lewis acids were determined at 25.0 °C in toluene solution.
B(C6F5)3 and MAO solutions were prepared by vacuum-transferring
sufficient toluene from Na/K onto the pure solutes to prepare 100 mL
(nominal) of solution, while solutions of organometallic complexes
having precisely known concentrations were prepared from Na/K-dried
toluene in the glovebox. The organometallic titrant was metered into
the stirred excess Lewis acid solution19 using a calibrated buret
controlled by a clutched synchronous motor, and temperature changes
were recorded using a precision thermistor amplified by a Wheatstone
bridge interfaced to an analog recorder. The calorimeter heat capacity
was determined by monitoring temperature changes when heat was
introduced by a calibrated resistive heater. Each run consisted of several
(10-20) sequential titrations carried out for each reaction; average
reported deviations are from the mean between these. Overall
instrument calibrations using established methods20 ruled out significant
sources of systematic error. Other details of the techniques and
Experimental Section
General Considerations. All organometallic complexes were
manipulated under rigorously anaerobic conditions using a Vacuum
Atmospheres nitrogen-filled glovebox with an efficient recirculator (e1
ppm O2) or in glassware sealed with Viton O-rings and interfaced to a
high-vacuum manifold bearing valved inlets for argon purified by
passage through a supported MnO bed.3e,4a Other reactions were carried
out using standard Schlenk techniques. Solvents were freshly distilled
from appropriate drying agents.3e,4a Elemental analyses were performed
by Oneida Research Services (Whitesboro, NY).
Starting Materials and Reagents. The preparation of the pen-
tafluorophenyllithium used in several of the following procedures
followed a published method.13 Reactions of trihaloboranes with
arylsilanes were modifications (solvent, temperature) of published
procedures.14 Methylalumoxane (MAO, obtained as a toluene solution
from Aldrich) was dried under high vacuum for 15 h to remove excess
volatile aluminum alkyls and to produce a solid which could be easily
weighed.6c,15 Tris(pentafluorophenyl)borane (5a) was a gift from the
Dow Chemical Co. (Freeport, TX). It was purified by crystallization
from pentane at -30 °C followed by high vacuum sublimation (80
°C/10-5 Torr). Purity was verified by 19F NMR. The complexes Cp2-
Zr(CH3)2,16a (Me5Cp)2Zr(CH3)2,16b and (1,2-Me2Cp)2Zr(CH3)216c and the
constrained-geometry complexes Me2Si(Me4Cp)(tBuN)M(CH3)2 (M
) Ti, Zr, Hf)17 were prepared by published methods. PhBCl2 and
1-bromo-3,5-dimethylbenzene (Aldrich) and 1-bromo-3,5-difluoroben-
zene (PCR) were used as received. Toluene-d8 was purchased from
(11) Some aspects of this work have been previously communicated:
Deck, P. A.; Marks, T. J. J. Am. Chem. Soc. 1995, 117, 6128-6129.
(12) Perrin, C. L.; Dwyer, T. J. Chem. ReV. 1990, 90, 935-937.
(13) (a) Massey, A. G.; Park, A. J. J. Organomet. Chem. 1964, 2, 245-
250. (b) Massey, A. G.; Park, A. J. J. Organomet. Chem. 1966, 5, 218-
225.
(14) (a) Kaufmann, D. Chem. Ber. 1987, 120, 853-854. (b) Soundarara-
jan, R.; Matteson, D. S. J. Org. Chem. 1990, 55, 2274-2275.
(15) Resconi, L.; Bossi, S.; Abis, L. Macromolecules 1990, 23, 4489-
4491.
(18) (a) Holm, K. H.; Skattebøl, L. Acta Chim. Scand., Ser. B 1984, 38,
783-94. (b) Brun, R.; Grace, D. S.; Holm, K. H.; Skattebøl, L. Acta Chim.
Scand., Ser. B 1986, 40, 21-25.
(19) 1H NMR experiments indicated that the reactions of the metallocene
dimethyl complexes with excess B(C6F5)3 quantitatively form mononuclear
+
metallocenium species of the formula L2MCH3 CH3B(C6F5)3-. Further
(16) (a) Samuel, E.; Rausch, M. D. J. Am. Chem. Soc. 1973, 95, 6263-
6237. (b) Manriquez, J. M.; McAlister, D. R.; Sanner, R. D.; Bercaw, J. E.
J. Am. Chem. Soc. 1978, 100, 2716-2724. (c) Smith, G. M. Ph.D. Thesis,
Northwestern University, Evanston, IL, 1987.
(17) (a) Stevens, J. C. In Studies in Surface Science and Catalysis,
Hightower, J. W., Delglass, W. N., Iglesia, E., Bell, A. T., Eds.; Elsevier:
Amsterdam, 1996, Vol. 101, pp 11-20, and references therein. (b) Canich,
J. M.; Hlatky, G. G.; Turner, H. W. PCT Appl. WO 92-00333, 1992;
Canich, J. M. Eur. Pat. Appl. EP 420 436-A1, 1991 (Exxon Chemical Co.).
(c) Stevens, J. C.; Timmers, F. J.; Wilson, D. R.; Schmidt, G. F.; Nickias,
P. N.; Rosen, R. K.; Knight, G. W.; Lai, S. Eur. Pat. Appl. EP 416 815-
A2, 1991 (Dow Chemical Co.).
reactions were not observed within 1 h at 25 °C. With the less Lewis acidic
triarylboranes described in the text, the methide abstraction process was
incomplete; under these conditions, reaction enthalpies are not conveniently
determined by titration calorimetry.
(20) Hill, J. O.; O¨ jelund, G.; Wadso¨, I. J. Chem. Thermodyn. 1969, 1,
111-116.
(21) (a) King, W. A.; Di Bella, S.; Lanza, G.; Khan, K.; Duncalf, D. J.;
Cloke, F. G. N.; Fragala´, I. L.; Marks, T. J. J. Am. Chem. Soc. 1996, 118,
627-635. (b) King, W. A.; Marks, T. J. Inorg. Chim. Acta 1995, 229, 343-
354. (c) Nolan, S. P.; Stern, D.; Marks, T. J. J. Am. Chem. Soc. 1989, 111,
7844-7854. (d) Schock, L. E.; Marks, T. J. J. Am. Chem. Soc. 1988, 110,
7701-7715.

Electrophilic Olefin Polymerization Catalysts
J. Am. Chem. Soc., Vol. 120, No. 8, 1998 1775
3
4
apparatus for carrying out the calorimetric measurements have been
described elsewhere.21 Data analysis followed standard published
methods.22
(C6D6): δ 162.8 (dd, JCF ) 252 Hz, JCF ) 11 Hz, CF), 140.8 (br s,
3 4 3
CB), 119.6 (dd, JCF ) 18 Hz, JCF ) 7 Hz, CH), 110.2 (t, JCF ) 25
Hz, CH).
Ethylene Polymerization Experiments.3e,4a For each polymeriza-
tion, a precise amount of metallocenium catalyst (7a-d or 8a-d,
nominally 7 mmol) was dissolved in 2.0 mL of Na/K-dried toluene in
the glovebox. The catalyst solution was quickly injected with a gastight
syringe into a rapidly stirred, ethylene-saturated toluene solution (50
mL) immersed in a room-temperature water bath. The mixture was
maintained under 1 atm of ethylene (passed through an activated MnO/
SiO2 column). The temperatures of selected runs were monitored in
situ as a function of time using an inserted digital thermometer/
thermocouple (OMEGA Engineering, model HH21). For the most
active catalysts studied, exotherms were never greater than 8 °C over
the reaction periods employed (typically 90 s). After a measured time
interval, the reaction was quenched by injecting 10 mL of methanol
into the mixture. Reaction times were dictated by the activity of the
catalyst (as short as 30 s for Table 6, entry 1; as long as 30 min for
Table 6, entry 8). The resulting polymeric solid was collected by
filtration. Evaporation of the filtrate afforded an insignificant amount
of residue. Several runs with each catalyst were performed to derive
standard uncertainties in the activities. These data are collected in Table
6.
3,5-Difluorophenylbis(pentafluorophenyl)borane (5b). A solution
of dibromo(3,5-difluorophenyl)borane (2.1 g, 7.4 mmol) in pentane (20
mL) was added by cannula to a stirred pentane suspension of
pentafluorophenyllithium (prepared from C6F5Br and n-BuLi in rapidly
stirred pentane at - 78 °C). The resulting mixture was stirred for 12
h while slowly warming from - 78 to 25 °C. A fine white precipitate
of LiBr was then separated by filtration. Evaporation of the filtrate
and recrystallization of the residue from 40 mL of pentane at -30 °C
afforded the crude triarylborane in two crops of off-white, feathery
crystals. Sublimation (90 °C, 1 × 10-5 Torr) afforded a colorless solid
(1.0 g, 2.2 mmol, 30%); 1H NMR (toluene-d8): δ 7.12 (m, 2 H), 6.62
3
4
(m, tt, JHF ) 8.7 Hz, JHH ) 2.4 Hz, 1 H); 13C NMR (CD2Cl2): δ
1
3
1
163.4 (CF, dd, JCF ) 250 Hz, JCF ) 11 Hz), 147.6 (CF, d, JCF
)
242 Hz), 144.3 (CF, d, 1JCF ) 252 Hz), 138.0 (CF, d, 1JCF ) 254 Hz),
120.5 (CH, dd, 2JCF ) 23 Hz, 4JCF ) 9 Hz), 113.8 (C, m), 111.2 (CH,
t, 2JCF ) 25 Hz) (the signal for the boron-attached carbon of the C6H3F2
ring was not observed); 19F (C6D6): δ -109.8 (t, 3JFH ) 7.1 Hz, 2 F),
-129.8 (complex d, JFF ) 23 Hz, 4 H), -146.3 (t, JFF ) 21 Hz, 2
F), -161.0 (complex t, JFF ) 21 Hz, 4 F). Anal. Calcd for C18H3-
3
3
3
BF12: C, 47.20; H, 0.66. Found: C, 46.92; H, 0.59; and C, 46.99; H,
0.56.
Representative Propylene Polymerization Experiment. An O-
ring-sealed Pyrex reactor fitted with a magnetic stirring bar, a well-
sealed septum inlet, and a vacuum line interface, was flamed out under
high vacuum and then charged with 30 mmol of dimethylmetallocene
precatalyst (6a) in the glovebox. Toluene (50 mL) was vacuum-
transferred into the reactor from Na/K, and the reactor was back-filled
with an atmosphere of propylene (passed through a MnO/SiO2 column)
and equilibrated at 25 °C. A solution of phenylbis(pentafluorophenyl)-
borane (5c, 45 mmol in 2 mL of toluene) was then injected in a single
portion using a gastight syringe. The mixture was rapidly stirred at
room temperature for 1.0 h under an atmosphere of propylene. The
reaction was then quenched with methanol (1 mL), and the mixture
was washed with aqueous sodium bicarbonate. A portion of the organic
layer was subjected to GC/MS analysis to determine the production of
dimers and lower oligomers, but none were detected. The remaining
toluene solution was evaporated and dried under high vacuum to leave
only a small residue (<50 mg) of a dark liquid, which had an
Phenylbis(pentafluorophenyl)borane (5c). To a suspension of
C6F5Li in pentane at - 78 °C, prepared from C6F5Br (34 g, 0.14 mol)
and n-butyllithium (90 mL, 1.6 M in hexanes, 0.14 mol) in rapidly
stirred pentane at - 78 °C, was added dichlorophenylborane (10 g,
0.063 mol) using a syringe. The mixture was stirred while warming
to 25 °C over a period of 12 h and then filtered to remove LiCl. The
filtrate was evaporated, and the resulting tan residue was recrystallized
from 25 mL pentane to afford an off-white solid, which was sublimed
at 100 °C (1 × 10-5 Torr) to afford the desired triarylborane as a
colorless solid (8.5 g, 0.020 mol, 32%): 1H NMR (toluene-d8): δ 7.78
(d, 3J ) 7.4 Hz, 2 H), d 7.49 (tt, 3J ) 6.9 Hz, 4J ) 2.1 Hz, 1 H), 7.36
3
1
(t, J ) 7.1 Hz); 13C NMR (toluene-d8): δ 146.7 (CF, d, JCF ) 247
1
Hz), 143.2 (CF, d, JCF ) 256 Hz), 139.6 (CH), 138.7 (C, br), 137.7
(CF, d, 1JCF ) 253 Hz), 136.8 (CH), 128.6 (CH), 114.1 (C, br m); 19
F
3
3
NMR (C6D6): δ -130.6, (d, JFF ) 26 Hz, 4 F), -149.3 (t, JFF ) -
21 Hz, 2 F), -161.9 (complex t, 3JFF ) 22 Hz, 4 F). Anal. Calcd for
C18H5BF10: C, 51.23; H, 1.19. Found: C, 51.11; H, 1.01; and C, 51.03;
H, 1.00.
1
appearance and H NMR spectrum consistent with atactic polypropy-
Trimethyl(3,5-dimethylphenyl)silane (3d).23b To a solution of 3,5-
dimethyl-1-bromobenzene (28 g, 0.15 mol) in diethyl ether was added
n-BuLi (80 mL, 2.5 M in hexanes, 0.20 mol); LiBr began to precipitate
immediately. The mixture was stirred under reflux for 1 h and then
cooled to 25 °C. Chlorotrimethylsilane (27 g, 0.25 mol) was added
via syringe, resulting in the precipitation of LiCl, and the solution was
stirred under reflux for an additional 1 h. The solution was then cooled
to 25 °C, and the solvents were removed under reduced pressure.
Aqueous workup afforded a pale yellow oil, which was distilled under
reduced pressure (88-90 °C, 11 Torr) to afford a colorless oil (24 g,
0.13 mol, 87%) of trimethyl(3,5-dimethylphenyl)silane. 1H NMR
(CDCl3): δ 7.19 (s, 2 H), 7.04 (s, 1 H), 2.38 (s, 6 H), 0.31 (s, 9 H);
13C NMR (CDCl3): δ 140.2 (CCH3), 137.0 (CSi), 131.0 (CH), 130.5
(CH), 21.4 (ArCH3), 1.1 (CH3Si).
lene.
Trimethyl(3,5-difluorophenyl)silane (3b).23a To a solution of
1-bromo-3,5-difluorobenzene (19 g, 0.10 mol) in ether (300 mL)
maintained at -78 °C was added n-butyllithium (45 mL, 2.5 M in
hexanes, 0.11 mol), and the resulting mixture was stirred at -78 °C
for 2 h. Chlorotrimethylsilane (16 g, 0.15 mol) was then added in one
portion. The resulting mixture was warmed to 25 °C with stirring.
Aqueous workup and distillation of the crude oil afforded trimethyl-
(3,5-difluorophenyl)silane as a colorless liquid (14 g, 0.075 mol,
1
75%): bp 69-72 °C (20 Torr); H NMR (CDCl3): δ 6.99 (m, 2 H),
6.76 (tt, JHF ) 8.6 Hz, JHH ) 2.4 Hz, 1 H), 0.28 (s, 9 H); 13C NMR
3
4
(CDCl3): δ 162.8 (CF, dd, 1JCF ) 262 Hz, 3JCF ) 14 Hz), 115.3 (CSi,
3
2
2
d, JCF ) 9 Hz), 115.3, (CH, JCF ) 22 Hz), 104.0 (CH, t, JCF ) 25
Hz), 1.4 (CH3, s).
Dichloro(3,5-dimethylphenyl)borane (4d). Trimethyl(3,5-dimeth-
ylphenyl)silane (18 g, 0.10 mol) was added to a solution of boron
trichloride (120 mL, 1.0 M in dichloromethane, 0.12 mol) maintained
at - 78 °C in a 200-mL Schlenk flask. A dry ice condenser was fitted,
and the mixture was warmed to 25 °C and then stirred under gentle
reflux for 2 h. The solution was next cooled to 25 °C, and the volatile
components were transferred to a trap under vacuum. The resulting
oil was distilled under reduced pressure (94-97 °C at 9 mmHg) to
afford 15 g (0.080 mol, 80%) of dichloro(3,5-dimethylphenyl)borane
as a colorless, exceedingly moisture-sensitive liquid: 1H NMR (C6D6):
δ 7.79 (s, 2 H), 6.91 (s, 1 H), 2.03 (s, 6 H); 13C NMR (C6D6): δ 137.6
(CCH3), 136.9 (CH), 134.6 (CH), 134.0 (CB), 21.2 (ArCH3).
3,5-Dimethylphenylbis(pentafluorophenyl)borane (5d). A solu-
tion of dichloro(3,5-dimethylphenyl)borane (9.4 g, 0.050 mol) in
pentane (50 mL) was added via cannula to a stirred suspension of
pentafluorophenyllithium (prepared from C6F5Br and n-BuLi in rapidly
Dibromo(3,5-difluorophenyl)borane (4b). A solution of trimethyl-
(3,5-difluorophenyl)silane (4.35 g, 23.4 mmol) and boron tribromide
(13.0 g, 50.0 mmol) in heptane (50 mL) was stirred under reflux for
24 h. After cooling the reaction mixture, the volatile components were
removed under vacuum to leave a dark residue, which contained about
10% of the starting silane according to 19F NMR analysis. Two
crystallizations from pentane afforded 2.06 g (7.25 mmol, 31%) of
colorless, highly deliquescent crystals of the desired dibromoarylbo-
rane: 1H NMR (C6D6): δ 7.36 (m, 1 H), 6.54 (m, 2 H); 13C NMR
(22) Hansen, L. D.; Lewis, E. A.; Eatough, D. J. In Analytical Solution
Calorimetry; Grime, K. J., Ed.; Chemical Analysis Monograph 79; Wiley:
New York, 1985; Chapter 3.
(23) (a) Effenberger, F.; Haebich, D. Liebigs Ann. Chem. 1979, 6, 842-
867. (b) Laquerre, M.; Dunogues, J.; Calas, R., Duffalt, N. J. Organomet.
Chem. 1976, 112, 49-59.

1776 J. Am. Chem. Soc., Vol. 120, No. 8, 1998
Deck et al.
stirred pentane at -78 °C). The resulting mixture was stirred for 6 h
while slowly warming from -78 to 25 °C. A fine white precipitate of
LiCl was separated by filtration, and the filtrate was evaporated under
vacuum. The residue was recrystallized from 50 mL of pentane at
-30 °C to afford 8.1 g (0.018 mol, 36%) of the desired borane. A
portion (2.0 g) of the crude borane was sublimed (100 °C, 1 × 10-5
Torr) to afford 1.75 g of a highly moisture-sensitive colorless solid:
1H NMR (C6D6): δ 7.35 (s, 2 H), 6.90 (s, 1 H), 2.01 (s, 6 H); 13C
Table 1. Synthetic Yields and Analytical Data for
(1,2-Me2Cp)2MCH3 CH3BAr(C6F5)2 Complexes
+
-
analytical data
found
calcd
entry
M
Ar
% yield % C % H % C % H
7b
7c
7d
8b
8c
8d
Zr 3,5-C6H3F2
Zr C6H5
Zr 3,5-C6H3(CH3)2
Hf 3,5-C6H3F2
Hf C6H5
83
85
72
95
86
66
53.34 3.55 53.54 4.28
55.97 4.01 56.02 3.91
57.07 4.39 57.11 4.16
47.88 3.19 48.06 3.48
49.99 3.58 49.55 3.50
51.18 3.94 51.11 3.92
1
NMR (toluene-d8): δ 146.7 (CF, d, JCF ) 244 Hz), 143.2 (CF, d,
1
1JCF ) 256 Hz), 139.2 (C), 138.2 (C), 137.8 (CF, d, JCF ) 251 Hz),
137.5 (CH), 137.4 (CH), 114.2 (C, m), 21.0 (CH3); 19F NMR (C6D6):
δ -131.0 (d, 3JFF ) 24 Hz, 4 F), -149.6 (t, 3JFF ) 20 Hz, 2 F), -161.8
(complex t, 3JFF ) 21 Hz, 4 F). Anal. Calcd for C20H9BF10: C, 53.37;
H, 2.02. Found: C, 53.26; H, 1.93; and C, 53.49; H, 1.81.
Hf 3,5-C6H3(CH3)2
precipitation of the product, which was quickly filtered while cold and
dried under vacuum. Yields of 7b-d (M ) Zr) and 8b-d (M ) Hf)
are collected in Table 1 along with combustion analytical data. Product
yields may be significantly reduced if the mixture is filtered at room
temperature or if the precipitates are washed with pentane. 1H NMR
data for the new metallocenium complexes are collected in Table 2.
Bis(1,2-dimethylcyclopentadienyl)zirconium(IV)methyl methyltris(pen-
tafluorophenyl)borate (7a) was prepared by the literature procedure.3e
Bis(1,2-dimethylcyclopentadienyl)hafnium(IV) Dichloride.
A
solution of HfCl4(THF)2 (15 g, 32 mmol) and lithium 1,2-dimethyl-
cyclopentadienide (7.0 g, 70 mmol) in tetrahydrofuran (500 mL) was
stirred at 25 °C for 2 d. The solvent was then removed under vacuum,
and the resulting solid was extracted with chloroform. Concentration
of the chloroform solution and recrystallization at -30 °C afforded
the desired complex (8.4 g, 19 mmol, 60%) in three crops of colorless
needles: 1H NMR (CDCl3): δ 6.01 (t, 3J ) 3.4 Hz, 2 H), 5.95 (d, 3J )
3.5 Hz, 4 H), 2.14 (s, 12 H); 13C NMR (1H coupled): δ 126.7 (s, CCH3),
Results
113.8 (d, 1JCH ) 171 Hz, CH), 105.3 (dt, 1JCH ) 172 Hz, CH, 2JCH
)
In this section, the synthesis and characterization of several
families of metallocenium ion pairs formed from triarylboranes
of varying Lewis acidity are described. Then, thermodynamic
parameters obtained by solution calorimetry and 1H NMR
equilibrium measurements are reported for methide abstraction
from the various metallocenes and quasi-metallocenes by a series
of organo-Lewis acids. These thermochemical data are subse-
1
6 Hz), 13.4 (q, JCH ) 125 Hz, CH3). Anal. Calcd for C14H18Cl2Hf:
C, 38.59; H, 4.16. Found: C 38.87; H, 3.99; and C, 38.74; H, 3.96.
Bis(1,2-dimethylcyclopentadienyl)hafnium(IV) Dimethyl (6b). A
slurry of bis(1,2-dimethylcyclopentadienyl)hafnium(IV) dichloride (4.0
g, 9.2 mmol) and methyllithium (800 mg, 0.037 mmol) in toluene was
stirred at 25 °C for 3 d. The mixture was then filtered through Celite
to remove lithium chloride and unreacted methyllithium. Evaporation
of the colorless filtrate afforded a white solid (3.0 g, 7.8 mmol, 85%),
which was shown to be >95% pure by NMR. Recrystallization from
toluene gave an analytically pure sample as colorless needles: 1H NMR
(C6D6): δ 5.37 (d, 3J ) 3.2 Hz, 4 H), 5.18 (t, 3J ) 3.2 Hz, 2 H), 1.92
(s, 6 H), -0.48 (s, 3 H); 13C NMR (proton-coupled, C6D6): δ 120.6 (s,
CCH3) 109.4 (d, 1JCH ) 169 Hz, CH), 102.9 (d, 1JCH ) 171 Hz, CH),
1
quently combined with kinetic parameters obtained from H
NMR line shape analyses to derive enthalpic and entropic
reaction coordinates for methide abstraction/transfer and ion-
pair reorganization processes. Trends in the effects of ancillary
ligation and abstraction reagent Lewis acidity on methide
transfer thermodynamic and kinetic parameters are then dis-
cussed. Finally, these data are correlated with NMR 13C
chemical shift and coupling constant data as well as with
ethylene polymerization activities.
1
1
37.0 (q, JCH ) 114 Hz, ZrCH3), 13.0 (q, JCH ) 126 Hz, CpCH3).
Anal. Calcd for C16H24Hf: C, 48.67; H, 6.13. Found: C, 48.84; H,
5.95; and C, 48.53; H, 5.95.
Bis(1,2-dimethylcyclopentadienyl)hafnium(IV)methyl Methyltris-
(pentafluorophenyl)borate (8a). A swivel-frit reaction apparatus was
charged with bis(1,2-dimethylcyclopentadienyl)hafnium(IV) dimethyl
(160 mg, 0.40 mmol) and tris(pentafluorophenyl)borane (210 mg, 0.41
mmol) in the glovebox. The reactor was connected to the vacuum
line, and pentane (25 mL) was condensed onto the solids at -78 °C
from Na/K. The mixture was warmed with stirring to 25 °C, stirred
for 2 h, and then filtered. The precipitate was washed with pentane to
+
Synthesis and Characterization of (1,2-Me2Cp)2MCH3
-
CH3BAr(C6F5)2 Complexes. To examine the effects of
varying triarylborane Lewis acidity on the metallocene catalyst
activation process, a series of metallocenium salts was prepared
in which a methide group is transferred from the group 4 metal
center to a triarylborane in which two of the aryl groups are
pentafluorophenyl and the third is varied. This overall approach
was chosen for the efficiency with which metallocenium borates
can be synthesized and rigorously characterized as discreet,
monometallic contact ion pairs, and in which ligands and Cp
substituents provide convenient spectroscopic probes in 1H NMR
spectral line-shape analytical and integration measurements. The
triarylborane substituents were chosen in order to incrementally
modulate the Lewis acidity of the triarylborane as compared to
B(C6F5)3; the substantially less Lewis acidic BPh3 exhibits no
detectable reaction with the present metallocene dimethyls at
room temperature.3e With the metallocene starting materials
already in hand, triarylboranes were synthesized via organosilane
intermediates as outlined in Scheme 1. Metal-halogen ex-
change of bromoarenes (1) with n-butyllithium afforded inter-
mediate lithioarenes (2), which were quenched with chlorotri-
methylsilane to afford the corresponding arylsilanes (3). The
arylsilanes were then treated with trihaloboranes4a,14 to afford
intermediate dihaloarylboranes (4), which were isolated either
as a waxy solid (4b) or as an exceedingly moisture-sensitive
liquid (4d). Treatment of the crude dihaloarylboranes (4b and
4d) and the commercially available phenyl derivative (4c) with
pentafluorophenyllithium in pentane afforded the respective
1
afford a colorless solid (330 mg, 0.37 mmol, 92%); H NMR (C6D6):
3
4
3
δ 5.37 (t, J ) J ) 3.0 Hz, 2 H), 5.22 (t, J ) 3.0 Hz, 2 H), 4.72 (t,
3J ) J ) 2.7 Hz, 2 H), 1.55 (s, 6 H), 1.26 (s, 6 H), 0.29 (br s, 3 H),
4
-0.10 (s, 3 H); 13C NMR (C6D6): δ 148.6 (d, JCF ) 244 Hz, CF),
1
1
1
139.5 (d, JCF ) 253 Hz, CF), 137.6 (d, JCF ) 242 Hz, CF). 114.5
(C), 110.9 (CH), 105.6 (CH), ring CCH3 signals not observed (obscured
by solvent), C6F5 ipso signal not observed, 42.8 (ZrCH3), 21.5 (br,
BCH3), 12.5 (CpCH3), 12.3 (CpCH3); 19F NMR (C6D6): δ -133.3 (d,
3J ) 23 Hz, 6 F), -159.2 (t, 3J ) 21 Hz, 3 F), -164.4 (t, 3J ) 21 Hz,
6 F). Anal. Calcd for C34H24BF15Hf: C, 45.03; H, 2.67. Found: C,
45.37; H, 2.49; and C, 45.31; H, 2.56.
General Synthetic Procedure for Bis(1,2-dimethylcyclopentadi-
enyl)metal(IV)methyl (Metal ) Zirconium or Hafnium) Methyl-
(aryl)bis(pentafluorophenyl)borates. A swivel-frit reaction apparatus,
fitted with two 50-mL (or 100-mL) flasks and a vacuum-line interface,
was charged in the glovebox with bis(1,2-dimethylcyclopentadienyl)-
metal(IV) dimethyl (metal ) zirconium (6a) or hafnium (6b), nominally
0.1 to 0.5 mmol and a slight (10%) excess of the triarylborane (5b-
d). The frit assembly was then mounted on the high vacuum line, and
pentane (10-25 mL) was transferred from Na/K onto the solids at -78
°C with stirring. The mixture was then warmed to 25 °C and stirred
for an additional 1 h, to afford either a clear solution or a suspension
of fine precipitate. Cooling the mixture to -78 °C induced complete

Electrophilic Olefin Polymerization Catalysts
J. Am. Chem. Soc., Vol. 120, No. 8, 1998 1777
Table 2. 1H NMR Data for (1,2-Me2Cp)2MCH3 CH3BAr(C6F5)2 Complexes
+
-
entry
M
Zr
Ar
1H NMR data in toluene-d8 (chemical shifts, ppm)
7b
3,5-C6H3(CH3)2
7.11 (sa), 6.82 (s, 1 H), 5.31 (t,b 2 H), 5.11 (t, 3J ) 3.0 Hz, 2 H), 4.67 (t,b 2 H), 1.49 (s, 6 H),
1.22 (s, 6 H), 0.02 (s, 3 H), -0.07 (br s, 3 H)
7c
7d
8b
8c
8d
Zr
Zr
Hf
Hf
Hf
C6H5
7.42 (d, 3J ) 6.8 Hz, 2 H), 7.38 (t, 3J ) 7.0 Hz, 2 H), 7.23 (t, 3J ) 6.9 Hz, 1 H), 5.35 (t,b 2 H),
5.10 (t, 3J ) 3.0 Hz, 2 H), 4.61 (t,b 2 H), 1.51 (s, 6 H), 1.20 (s, 6 H), 0.04 (s, 3 H), -0.07 (br s, 3 H)
7.00 (ma), 6.56 (t, 3JHF ) 8 Hz, 1 H), 5.36 (t,b 2 H), 5.12 (t, 3J ) 3.0 Hz, 2 H), 4.64 (t,b 2 H),
1.53 (s, 6 H), 1.23 (s, 6 H), 0.07 (s, 3 H), -0.06 (br s, 3 H)
3,5-C6H3F2
3,5-C6H3(CH3)2
C6H5
7.1 (sa), 6.79 (s, 1 H), 5.23 (t,b 2 H), 5.15 (t, 3J ) 2.7 Hz, 2 H), 4.55 (t,b 2 H), 1.52 (s, 6 H),
1.21 (s, 6 H), 0.15 (br s, 3 H), -0.11 (s, 3 H)
7.46 (d, 3J ) 7.2 Hz, 2 H), 7.39 (t, 3J ) 7.4 Hz, 2 H), 7.27 (t, 3J ) 7.2 Hz, 1 H), 5.17 (t,b 2 H),
5.01 (t, 3J ) 2.7 Hz, 2 H), 4.48 (t,b 2 H), 1.52 (s, 6 H), 1.22 (s, 6 H), 0.18 (br s, 3 H), -0.13 (s, 3 H)
7.02 (ma), 6.58 (t, 3JHF ) 9 Hz, 1 H), 5.26 (t,b 2 H), 5.14 (t, 3J ) 2.5 Hz, 2 H), 4.60 (t,b 2 H),
1.55 (s, 6 H), 1.23 (s, 6 H), 0.11 (br s, 3 H), -0.14 (s, 3 H)
3,5-C6H3F2
a Signal partially obscured by solvent residual proton resonances. b Exchange-broadening prevented accurate determination of coupling constant.
Table 3. Enthalpies of Methide Abstraction from Group 4
Metallocenes and Related Compounds by Organo-Lewis Acids
B(C6F5)3 and MAO According to L2M(CH3)2 + A a L2MCH3
CH3A-
Scheme 1
+
a
Lewis
acid (A)
-∆Hdr
entry
complex
(kcal mol-1
)
1
2
3
4
5
6
7
8
9
Me2Si(Me4Cp)(tBuN)Ti(CH3)2
Cp2Zr(CH3)2
B(C6F5)3
B(C6F5)3
B(C6F5)3
B(C6F5)3
B(C6F5)3
B(C6F5)3
-22.6 (2)
-23.1 (3)
-24.3 (4)
-36.7 (5)
-23.9 (4)
-20.8 (5)
-19.3 (6)
-10.9 (3)
-8.9 (4)
(1,2-Me2Cp)2Zr(CH3)2
(Me5Cp)2Zr(CH3)2
Me2Si(Me4Cp)(tBuN)Zr(CH3)2
(1,2-Me2Cp)2Hf(CH3)2
Me2Si(Me2Cp)(tBuN)Hf(CH3)2 B(C6F5)3
(1,2-Me2Cp)2Zr(CH3)2
(1,2-Me2Cp)2Hf(CH3)2
MAO
MAO
a See Figure 1 for definition. Errors are average deviations from
the mean for two experiments, each of which comprises approximately
20 sequential batch titrations.
Scheme 2
of the titration.24 Calorimetric data are presented in Table 3,
in which reported uncertainties correspond to 95% confidence
limits.
In contrast to the metallocene dimethyl behavior with
B(C6F5)3 and MAO, analogous 1H NMR studies using the less
acidic boranes 5b-d reveal incomplete methide abstraction from
metallocenes at concentrations and temperatures comparable to
the calorimetric conditions. We were unable to detect signals
in the 1H NMR spectra assignable to ligand-bridged or dinuclear
µ-CH3 complexes.3d,f Rather, the spectra are consistent with a
dynamic equilibrium between the neutral metallocene, borane,
and the corresponding metallocenium salt (eq 2), as determined
by comparisons with published chemical shift data for analogous
compounds.1,3
triarylboranes (5) in moderate yields. Metallocenium methyl-
triarylborates (7, M ) Zr and 8, M ) Hf) were then prepared
by reacting the triarylboranes (5) with metallocene dimethyl
complexes (6) in pentane, and cooling the mixtures to -78 °C
to precipitate the desired products (Scheme 2); synthetic yields
and analytical data are provided in Table 1, and NMR data in
Table 2.
(1,2-Me2Cp)2M(CH3)2 + ArB(C6F5)2 a
-
(1,2-Me2Cp)2MCH3+ CH3BAr(C6F5)2 (2)
Thermodynamics of Metallocene Methide Abstraction by
Organo-Lewis Acids. Enthalpies of reaction for the forward
reactions shown in eq 1 where A is B(C6F5)3 or a representative
MAO sample (see Experimental Section for details) and R )
CH3 were measured by incrementally titrating approximately
0.04 M metallocene solutions in toluene into Lewis acid-
toluene solutions (total nominal B:M ratios of 10:1 and Al:M
These observations present a unique opportunity to measure
the thermodynamic aspects of eq 2 as a quantitative assessment
of the effectiveness of differently substituted organo-Lewis acids
in creating “cation-like” metallocene species. When metallo-
cenium salts 7 and 8 are dissolved in toluene-d8, integration of
well-resolved 1H NMR spectral signals (slow-exchange regime)
allows the equilibrium constants for eq 2 to be evaluated
1
ratios of 50:1 were used). Under these conditions, H NMR
studies indicate that the reactions with B(C6F5)3 are rapid, clean,
and complete. We have previously shown by solid-state
CPMAS 13C NMR that methylalumoxane (MAO) Al:M ratios
of ∼12:1 are adequate to effect complete conversion to
metallocenium species.6c Since the metallocene solutions are
titrated into the MAO or borane solutions in these experiments,
initial Al:M and B:M ratios are rather large (∼1000:1 Al:M
for MAO) and decline through the course of the titration. In
no case is the derived heat of reaction dependent upon the stage
(24) In principle, “breaks” in the MAO-zirconocene thermochemical
titration curves might indicate the formation of discrete species of discrete
stoichiometry. Such evidence was not observed in these and other
thermochemical measurements to be discussed elsewhere (Luo, L.; Marks,
T. J. Manuscript in preparation).
(25) (a) Diogo, H. R.; de Alencar Simoni, J.; Minas de Piedade, M. E.;
Dias, A. R.; Martinho-Simoes, J. A. J. Am. Chem. Soc. 1993, 115, 2764-
2774. (b) Martinho Simoes, J. A.; Beauchamp, J. L. Chem. ReV. 1990, 90,
629-688 and references therein.

1778 J. Am. Chem. Soc., Vol. 120, No. 8, 1998
Deck et al.
mol-1. Abstraction from Ti (Table 3, entry 1) proceeds with
an exothermicity between that of Zr (entry 5) and Hf (entry 6)
for identical ancillary ligation. Second, increased methyl
substitution of the bis-cyclopentadienyl ancillary ligation results
in consistent but nonlinear changes in enthalpies of reaction
for eq 1. Table 3 entries 2 and 3 evidence a modest increase in
exothermicity when four methyl substituents are appended to
the Cp ancillary ligands, while (Me5Cp)2ZrMe2 evidences an
increase of 12 kcal mol-1. A priori, it might be predicted that
Cp electronic substituent effects might have an approximately
linear influence on cation stabilization,26 and, therefore, the
present data suggest that steric effects may also contribute to
the observed large increase in the magnitude of ∆H observed
for (Me5Cp)2ZrMe2 (see Discussion for further comments). A
third noteworthy observation is that methide abstraction by
MAO is considerably less exothermic than the corresponding
abstraction by B(C6F5)3 (Table 3: (entries 3 vs 8, 6 vs 9). The
observed difference would appear to argue for a higher relative
Lewis acidity of B(C6F5)3 versus MAO. Finally, methide
abstraction enthalpies involving “constrained-geometry” com-
plexes (Table 3, entries 5 and 7)17,27 are similar to those observed
for the conventional metallocenes (entries 3 and 6).
The data in Table 4 reveal several additional noteworthy
trends in methide abstraction thermochemical parameters as a
function of metal (Zr or Hf) and triarylborane substituents. First,
large but orderly decreases in abstraction exothermicity are
observed for methide transfer to triarylboranes having the
formula ArB(C6F5)2 for the progression Ar ) C6F5 f 3,5-
C6H3F2 f C6H5 f 3,5-C6H3(CH3)2. This result shows that
variations in the substituents on a single aryl group in these
organoborane Lewis acids have a profound effect on the ability
to generate electrophilic “cation-like” species. The progression
in Keq (estimated at 298 K) is consistent with this conclusion,
while the significantly negative values of ∆Seq are consistent
with the associative nature of the ion-pair formation process.
Also noteworthy is the roughly 9 kcal mol-1 difference in ∆Hdr
for Ar ) C6F5 vs Ar ) C6H5 (entries 1 vs 3 and entries 5 vs 7).
In qualitative agreement with these results, HF/3-21G-level
quantum chemical calculations show B(C6F5)3 to be more acidic
than B(C6H5)(C6F5)2 by about 17.8 kcal mol-1 with respect to
methide as the base.28
Figure 2. Van’t Hoff plots for metallocenium ion pair formation/
dissociation (eq 2).
according to eq 3, where [N], [B], and [C] are the concentrations
[C]
[C]
[N]2
Keq
)
)
(3)
[N][B]
of neutral metallocene, free borane, and metallocenium cation,
respectively. The equilibrium constants defined in eq 3 were
determined for all six metallocenium salts (7b-d, 8b-d) at
four or more temperatures over a 30 °C temperature range, and
least-squares analyses of the resulting van’t Hoff plots for M
) Zr and M ) Hf are given in Figure 2. ∆Hdr and ∆Sdr
parameters (Figure 1) are readily obtained from the slope and
intercept of each plot, respectively. The ∆Hdr values so obtained
may be compared directly with the calorimetric -∆Hdr values
shown for methide abstraction in Table 3, eq 1; these data are
collected in Table 4. The van’t Hoff plots also allow estimation
of Keq at a single temperature (298 K), and these values are
also compared in Table 4.
Ion-Pair Dissociation/Reorganization Kinetics. As is ap-
parent in Figure 1, the construction of quantitative enthalpic
surfaces for methide abstraction requires not only thermody-
namic but also kinetic information in order to fix the relative
positions of enthalpic minima and maxima, respectively, and
ultimately to understand the driving force for cationic catalyst
formation as well as the ion-pairing dynamics likely operative
in stereoregulation processes. In this regard, the molecular
+
structure of the (1,2-Me2Cp)2MCH3 cation offers unique
The data in Tables 3 and 4 show that the abstraction of
methide from a metallocene using ArB(C6F5)2 or MAO reagents
is rather exothermic, even though enthalpies of homolytic bond
disruption for M-CH3 (M ) Zr or Hf) in these systems are
substantial.21d,25 The strengths of methide-Lewis acid bonds
formed in the abstraction process and the ability of ring
substituents to stabilize the cationic charge are doubtless
dominant (but highly variable) contributors to the observed
exothermicities (see Discussion section).
From the data in Tables 3 and 4, several noteworthy trends
may readily be inferred from the enthalpies of reaction corre-
sponding to variation in ancillary ligation (L2), metal (M), and
Lewis acid (A). First, methide abstraction from Hf is less
exothermic than from Zr (Table 3: entry 3 vs 6, entry 5 vs 7;
Table 4: entry 2 vs 6; entry 3 vs 7; entry 4 vs 8) by 2-4 kcal
dynamic NMR spectroscopic probes of ion-pair reorganization
dynamics and how those dynamics depend upon M, Lewis acid
abstractor, and solvent. Cation-anion dissociation-recombina-
tion processes which invert the symmetry of the dissymmetric
ion pair structure (Scheme 3, kips) permute diastereotopic Cp-
Me and ring C-H groups. Processes which involve B-CH3
(26) Gassman, P. G.; Winter, C. H. J. Am. Chem. Soc. 1988, 110, 6130-
35 and references therein.
(27) (a) Woo, T. K.; Margl, P. M.; Lohrenz, J. C. W.; Blochl, P. E.;
Ziegler, T. J. Am. Chem. Soc. 1996, 118, 13021-13030. (b) Carpenetti, D.
W.; Kloppenbrug, L.; Kupec, J. T.; Petersen, J. L. Organometallics 1996,
15, 1572-1581. (c) Devore, D. D.; Timmers, F. J.; Hasha, D. L.; Rosen,
R. K.; Marks, T. J.; Deck, P. A.; Stern, C. L. Organometallics 1995, 14,
3132-3134.
(28) LaPointe, R. E.; Stevens, J. C.; Nickias, P. N.; McAdon, M. H.
Eur. Pat. Appl. 92305730.1, 1992.

Electrophilic Olefin Polymerization Catalysts
J. Am. Chem. Soc., Vol. 120, No. 8, 1998 1779
Table 4. Thermochemical and Kinetic Data for Methide Abstraction from (Me2Cp)2M(CH3)2 Complexes in Toluene Solution as a Function of
-
Triarylborane As Depicted in (1,2-Me2Cp)2M(CH3)2 + ArB(C6F5)2 a (1,2-Me2Cp)2MCH3+CH3BAr(C6F5)2
thermodynamic data from Van’t Hoff
Analysis
kinetic data from Eyring
analysisd
-∆Hdr
(kcal/mol)
∆Hq
dr
entry
M
Ar
ion pair
-∆Sdr (eu)
Keq at 298 Kc
(kcal/mol)
∆Sqdr (eu)
kdr (s-1) (298 K)
1
2
3
4
5
6
7
8
Zr
Zr
Zr
Zr
Hf
Hf
Hf
Hf
C6F5
3,5-C6H3F2
C6H5
3,5-C6H3(CH3)2
C6F5
3,5-C6H3F2
C6F5
7a
7b
7c
7d
8a
8b
8c
8d
-24.2(5)a
-18.7(7)
-14.8(8)
-10.8(6)
-21.4(3)a
-15.2(8)
-13.3(6)
-12.7(5)
3 × 108 b
3 × 104
1 × 104
2 × 103
5 × 107 b
3 × 103
1 × 102
3 × 101
27(2)
23(2)
19(2)
17(2)
22(1)
16(1)
15(1)
15(1)
22(3)
20(5)
14(2)
9(4)
16(1)
4(4)
0.003
2.2
55
110
1.3
160
4600
9800
-42(2)
-31(2)
-19(2)
-35(3)
-39(2)
-36(4)
10(3)
10(3)
3,5-C6H3(CH3)2
a Obtained by titration calorimetry. Keq too large to determine by the NMR equilibration methods described in the text. Values were estimated
assuming ∆Sdr ) 42 eu (M ) Zr) and ∆Sdr ) 35 eu (M ) HF), as these are the experimental ∆Sdr values for Zr ) 3,5-C6H3F2. These values should
be considered rough estimates (within two orders of magnitude), since Keq is quite sensitive to ∆Sdr, in which there is substantial variability.
c Defined by eqs 1 and 2 in the text. d Activation parameters for the reverse of eq 1 (see Figure 1) in the text determined by NMR line-shape
analysis.
b
Scheme 3
dissociation and subsequent recombination (kdr) also permute
Cp-Me and ring C-H groups but additionally permute both
B-CH3 and M-CH3 substituents, all at identical rates. In-
tramolecular versus intermolecular processes can be probed by
examining the concentration dependence of the site permutation
processes as well as the influence of added BAr3.
The series 7a-d and 8a-d were studied by variable-
temperature NMR techniques in toluene solution (see Experi-
mental Section for details), and rates of Cp-Me (kips) and
M-Me/B-Me (kdr) exchange were determined using standard
Figure 3. Eyring plots for the kinetics of metallocenium ion pair
dissociation/recombination (Figure 1B).
only exception is complex 7a, the (1,2-Me2Cp)MCH3+ ion pair
with the most strongly bound borane (B(C6F5)3 ) as judged from
the calorimetry data (Table 3). In this case, ion-pair sym-
metrization (ion-pair separation-recombination, the scenario of
Figure 1A) is more rapid than borane dissociation/recombination
at all temperatures where line-broadening measurements are
possible (e.g., kips )30 s-1 and kdr ) 3 s-1 at 298 K). In regard
to the mechanism of the process associated with kdr, the
following lines of evidence argue that it is dissociative in
character. First, line shapes are insensitive to the addition of a
10-fold molar excess of borane, arguing against an associative
SE-2-like interchange process. Second, the large positive
entropies of activation observed for kdr (Table 4) are consistent
with a dissociative process. Third, to exclude the possibility
that kdr represents purely intramolecular repositioning of borane
between M-CH3 groups (eq 4), 19F EXSY experiments on
modified Bloch equation line-shape analysis techniques.29 In
several cases, dilution experiments revealed the line-broadening
effects to be independent of concentration over a 20-fold range.
The rates of site exchange were furthermore examined over 30-
50 °C temperature ranges, and least-squares fits to Eyring plots
(Figure 3) yielded the activation parameters set out in Table 4.
Several observations are noteworthy. Mechanistically, the rate
of ion-pair symmetrization in toluene is undetectably small in
comparison to borane dissociation-recombination (the scenario
of Figure 1B; the rate of Cp-Me exchange is identical with
that of M-Me/B-Me exchange within experimental error). The
(29) (a) Sandstrøm, J. Dynamic NMR Spectroscopy; Academic Press:
New York, 1982; pp 77-92. (b) Ham, N. S.; Mole, T. Prog. Nucl. Magn.
Reson. Spectrosc. 1969, 4, 91-192. (c) Kaplan, J. I.; Fraenkel, G. NMR of
Chemically Exhanging Systems; Academic Press: New York, 1980; pp 71-
128.

1780 J. Am. Chem. Soc., Vol. 120, No. 8, 1998
Deck et al.
+
An alternative probe of M-CH3 electrophilicity is the
magnitude of 1JC-H, with a relevant example being the correla-
1
tion of JC-H (Ti-CH3) in a series of L2Ti(Cl)CH3 complexes
with increasing electron-withdrawing character of the L ) Cp-
type ligand substituents.31 Qualitatively, the effect is associated
with CH3 rehybridization leading to greater C 2s character in
the C-H bonds; for the above Ti series, the full variation of
JC-H coupling constants is e 3.5 Hz. In the present 7a-d and
8a-d series, any consistent variation in JC-H(M-CH3) with
borane identity/Lewis acidity/∆H (complex formation) is below
the experimental uncertainty in the -60 °C measurement. The
same is true for the JC-H(B-CH3) parameters, where quadru-
polar broadening further increases the uncertainty in the
measurement.
Polymerization Studies. Ethylene and propylene polymer-
ization experiments with complexes 7a-d and 8a-d were
carried out at 25 °C using methodology designed to minimize
mass-transfer effects.3e,4a With the exception of complex 7a,
propylene polymerization activity under these conditions was
too low to warrant further investigation. Ethylene polymeri-
zation activity results are presented in Table 6 along with
estimations (from Keq data in Table 4) of the percentage of the
metallocene methyl complex in the cation-like form under
typical polymerization conditions in toluene at 25 °C. It can
be seen for M ) Zr that there is a qualitative correlation between
δ(Zr-13CH3) and ethylene polymerization activity. For entries
1-4, an approximate correction for the percentage of metal-
locene in the cation-like form suggests that 7a is intrinsically
more reactive. In the case of M ) Hf, there is again a qualitative
correlation between δ(Hf-13CH3) and ethylene polymerization
activity. However, factoring in the percentages of metallocene
in the cation-like form renders the correlation less montonic.
Further comparisons will be made in the following section.
7a with 1.0 equiv of added B(C6F5)3 were carried out. The
EXSY experiments reveal that B-Me/M-Me and bound
B(C6F5)3/free B(C6F5)3 exchange rates are indistinguishable.
Since solvent is known to have a significant effect on
metallocenium polymerization activity and stereoregulation,9 the
-
dynamics of (1,2-Me2Cp)2MMe+ MeB(C6F5)4 ion-pair reor-
ganization were also examined as a function of solvent polarity/
1
coordinating ability using H NMR line-shape analysis tech-
niques. Data are compiled in Table 5. For M ) Zr, the only
case where the rate of ion-pair symmetrization (Scheme 3, kips)
is detectable in toluene (ꢀ ) 2.37) solution, the effect of C6D5-
Cl (ꢀ ) 5.71) and 1,2-C6D4Cl2 (ꢀ ) 9.93) solvation is a Very
large enhancement (∼2000x) in the rate of ion-pair symmetriza-
tion (kips) and a more modest increase in the rate of borane
dissociation/recombination (kdr). In the case of M ) Hf, where
kips is not detectable in toluene solution, dissolution in more
polar solvents has several effects. In the case of C6F6 (ꢀ )
2.03) and CCl2FCF2Cl2 (ꢀ ) 2.41), small increases in kdr are
detected; however, ion-pair symmetrization remains relatively
slow. However, in the case of higher dielectric solvents such
as C6D5Cl, 1,2-C6D4Cl2, and CD2Cl2 (ꢀ ) 9.08), kips becomes
nearly as large as in the case of M ) Zr. At the first level of
analysis, solvent effects on the ion-pair separation process can
be related to more facile separation of charged species in higher
dielectric media. Any depression in rates of borane dissociation
(cf. Table 5, entries 7-9) could then be ascribed to stabilization
of the charge-separated ground-state versus the neutral dissocia-
tion products (cf. Figure 1). The possibility that the charge-
separated metallocene cation might be stabilized by complex
formation with the halocarbon solvent30 (e.g., C) was investi-
Discussion
The present study provides the first quantitative information
on the energetics of forming, dissociating, and reorganizing
organo-Lewis acid-derived metallocenium and quasi-metallo-
cenium ion pairs (A) active to varying degrees in homogeneous
catalytic olefin polymerization. In the discussion which follows,
we focus first on the chemical origins of the measured
thermodynamic parameters, then on the reaction coordinates for
ion-pair formation, dissociation, and rearrangement, and finally,
on possible correlations between physicochemical observables
and olefin polymerization activity.
Thermochemistry of Ion-Pair Formation. The magnitudes
of the thermodynamic parameters describing eqs 1 and 2, Figure
1, and reported in Tables 3 and 4, represent a subtle interplay
of several molecular parameters as illustrated by the approximate
thermodynamic cycle of Figure 4. It can be seen that metal/
ligand sensitivity of the measured methide abstraction enthalpy
(-∆Hdr) is sensitive to the M-CH3 homolytic bond dissociation
enthalpy, the L2MCH3 ionization potential, and the ion-pair
binding energetics, the latter of which likely contain nonelec-
trostatic components (e.g., covalent and repulsive nonbonded
interactions) judging from crystal structure data3e and qualitative
orderings of anion coordinating tendencies.4a,8 The electron
affinity of the methyl radical is invariant (180 kcal/mol),32 while
the borane methide affinity will be invariant for a metallocene
series with constant BAr3 (but is presently unknown). Making
the pragmatic assumption that the effects of solvation by
gated by a variety of low-temperature NMR experiments (e.g.,
incremental halocarbon addition to the metallocenium salt in
toluene-d8 at low temperatures). In no case was a halocarbon
complex detected. The ∆Sq parameters in Table 5 suggest that
the transition states for both borane dissociation and ion-pair
separation are more highly organized in the higher dielectric
solvents, possibly reflecting reorganization of the solvation
sphere.
Metallocenium Methyltriarylborate NMR Parameters. In
an effort to further probe (1,2-Me2-Cp)2MMe+ MeBAr(C6F5)2
-
electronic structure in the 7a-d and 8a-d series, δ(M-13CH3)
and 1JC-H(M-CH3 and B-CH3) values were measured. Within
a homologous series, the former parameter should provide a
qualitative measure of metallocenium electron deficiency/
electrophilicity.1,3d,e,4a,b From the data in Tables 4 and 6, it can
be seen that δ (M-13CH3) scales qualitatively with the electron-
withdrawing tendencies of the Ar substituents in the ArB(C6F5)2
series. Further correlations will be examined in the Discussion
section.
(30) (a) Peng, T.-S.; Winter, C. H.; Gladysz, J. A. Inorg. Chem. 1994,
33, 2534-2542 and references therein. (b) Igau, A.; Gladysz, J. A.
Organometallics 1991, 10, 2327-2334. (c) Colsman, M. R.; Newbound,
T. D.; Marshall, L. J.; Noirot, M. D.; Miller, M. M.; Wulfsberg, G. P.;
Frye, J. S.; Anderson, O. P.; Strauss, S. H. J. Am. Chem. Soc., 1990, 112,
2349-2362 and references therein. (d) Kulawiec, R. J.; Crabtree, R. H.
Coord. Chem. ReV, 1990, 99, 89-115 and references therein.
(31) Finch, W. C.; Anslyn, E. V.; Grubbs, R. H. J. Am. Chem. Soc. 1988,
110, 2406-2413.
(32) Huheey, J. E.; Keiter, E. A.; Keiter, R. L. Inorganic Chemistry, 4th
ed.; Harper Collins: New York, 1993; pp 40-43.

Electrophilic Olefin Polymerization Catalysts
J. Am. Chem. Soc., Vol. 120, No. 8, 1998 1781
-
Table 5. Kinetic Data for Ion-Pair Reorganization Processes in (1,2-Me2Cp)2MCH3+CH3B(C6F5)3 Complexes as a Function of Solvent and
Solvent Dielectric Constant
c
c
c
c
∆Hq
∆Sq
∆Hq
∆Sq
ips
dr
ips
dr
entry
M
solvent (ꢀ)a
kipsb (10-3/s)
kdrb (10-3/s)
(kcal/mol)
(eu)
(kcal/mol)
(eu)
1
2
3
4
5
6
7
8
9
Zr
Zr
Zr
Hf
Hf
Hf
Hf
Hf
Hf
toluene-d8 (2.37)
C6D5Cl (5.71)
1,2-C6D4Cl2 (9.93)
toluene-d8 (2.37)
C6F6 (2.03)
CCl2FCF2Cl (2.41)
C6D5Cl (5.71)
1,2-C6D4Cl2 (9.93)
CD2Cl2 (9.08)
30 (10)
60 000 (20 000)
70 000 (20 000)
d
3 (2)
20 (8)
24 (1)
11 (2)
12 (2)
17 (2)
-15 (8)
-10 (4)
27 (2)
19 (1)
22 (3)
0 (2)
<1
1300 (400)
5600 (200)
8000 (4000)
600 (300)
5 (2)
22 (1)
16 (1)
15 (2)
20 (1)
23 (1)
16 (1)
-2 (4)
-2 (4)
9 (3)
d
d
15 000 (9 000)
9 000 (4 000)
20 000 (10 000)
13 (4)
12 (3)
11 (1)
-9 (1)
-5 (8)
-16 (2)
7 (1)
<1
a Dielectric constant from CRC Handbook of Chemistry and Physics, 77th ed.; CRC Press: New York, 1996; p G-161. b Rate constant at 298
K derived from least-squares fitting of Eyring plot. c Eyring parameters derived from line-shape analysis. d Rate too slow to determine; kips << kdr.
Table 6. M-13CH3 Chemical Shifts (δC), M-13CH3 and B-13CH3 JCH Coupling Constants, and Ethylene Polymerization Activities for
1
+
-
(1,2-Me2Cp)2MCH3 CH3BAr(C6F5)2 Complexes in Toluene Solution
δC(MCH3)a
((0.02 ppm)
1JCH (MCH3)a
((0.5 Hz)
1JCH (BCH3)a
((1.0 Hz)
%
entry
catalyst
M
Ar
cationb
activityc
1
2
3
4
5
6
7
8
7a
7b
7c
7d
8a
8b
8c
8d
Zr
Zr
Zr
Zr
Hf
Hf
Hf
Hf
C6F5
3,6-C6H3F2
C6H5
3,5-C6H3(CH3)2
C6F5
3,5-C6H3F2
C6H5
3,5-C6H3(CH3)2
44.07
43.85
43.17
43.12
42.33
42.23
41.71
41.71
120.8
120.2
120.7
120.8
117.5
117.1
117.1
117.6
116.6
116.3
115.8
115.6
116.4
115.3
115.3
115.4
>99
62
5.2(4) × 106
5.0(4) × 106
1.3(11) × 105
3.4(15) × 105
1.3(10) × 105
5.8(12) × 104
6.5(28) × 103
1.6(30) × 103
44
19
>99
24
2
<1
a NMR parameters measured at -60 °C. b Mole percentage in the cation-like form, estimated from the equilibrium constants in Table 4 for a
toluene solution initially 0.140 mM in catalyst at 25 °C. c Activities measured in toluene solution at 25 °C and expressed in (g polyethylene) (mol
catalyst)-1 (atm C2H4)-1 h-1. Uncertainties in parentheses are standard deviations from several independent runs.
Figure 4. Qualitative thermodynamic cycle for metallocenium ion pair
formation from the neutral precursors.
hydrocarbon solvents are relatively small and constant for a
homologous series,33 then the difference in methide abstraction
+
Figure 5. Experimental enthalpic profiles for (1,2-Me2Cp)2MCH3
-
CH3B(C6F5)3- ion pair formation and reorganization for M ) Zr or Hf
in toluene-d8 and chlorobenzene-d5 solutions.
enthalpies for two different metallocene dimethyls, L2MMe2 and
L2MMe2′, can be expressed as in eq 5. In the case of Cp2-
ZrMe2 versus (Me5Cp)2ZrMe2 versus Cp2HfMe2, as an example,
D(Cp2Zr(Me)-Me) and D((Me5Cp)2Zr(Me)-Me) are virtually
identical at 67.2(1.0) and 67.0(1.0) kcal/mol,21d respectively,
∆∆H ) [D(M-Me) - D(M-Me′)] + (IP - IP′) -
(∆Hips - ∆Hips′) (5)

1782 J. Am. Chem. Soc., Vol. 120, No. 8, 1998
Deck et al.
-
-
Figure 6. Relation of ∆Hq for (1,2-Me2Cp)2MCH3+CH3B(C6F5)3
/
Figure 7. Relation of ∆Sq for (1,2-Me2Cp)2MCH3+CH3B(C6F5)3
/
dr
dr
CH3B(C6F5)2Ar- dissociation to neutral reactants and the corresponding
enthalpy of the same process for M ) Zr or Hf. The boranes are those
of Table 4.
CH3B(C6F5)2Ar- dissociation to neutral reactants and the corresponding
entropy of the same process for M ) Zr or Hf. The boranes are those
of Table 4.
d8) and Me2Si(Me4Cp)(tBuN)ZrCH3+ CH3B(C6F5)3- (∆Hq
)
ips
while D((Me5Cp)2Hf(Me)-Me) is probably slightly larger (72.0-
(3.7) kcal/mol).21d While few ionization potentials have been
recorded for trivalent group 4 metallocenes (for 1,3-tBu2Cp)2ZrI,
IP ) 5.72 eV),34 gas-phase photoelectron spectroscopic (PES)
data for Cp2ZrCl2 (IP ) 8.60 eV)35 and (Me5Cp)2ZrCl2 (IP )
7.55 eV)36 indicate that ring methylation should have a
substantial effect on the energetic ordering in Figure 4 (∆IP )
24.2 kcal/mol for this tetravalent pair). PES data for Cp2HfCl2
(IP ) 8.89 eV)35 indicate that group 4 metal identity will have
a smaller effect (∆IP ) 6.7 kcal/mol).35 Although ∆Hips data
have not been measured, they would appear to be reasonably
19(1) kcal/mol in toluene-d8).37 On the other hand, a greater
dispersion in ∆Hips and ∆Hq values is likely in anions with
ips
greater coordinative tendencies (e.g., tris(2,2′,2′′-nonafluorobi-
phenyl)fluoroaluminate, D).3b
approximated by ∆Hq (Table 5; i.e., assuming ∆Hq for ion-
ips
pair recombination is small) and seem unlikely to vary greatly
for constant CH3BAr3- from one L2MCH3+ CH3BAr3- system
to another as judged by dynamic NMR results for (1,2-Me2-
Cp)2ZrCH3+ CH3B(C6F5)3- (∆Hqips ) 24(1) kcal/mol in toluene-
From the above analysis, the results in Tables 3 and 4 are
not difficult to understand. For the types of group 4 metal-
locenes and quasi-metallocenes studied here, ancillary ligation
as expressed in cation stabilizing ability (e.g., IP) appears to
have the greatest influence on ∆Hdr for constant borane, while
D(M-CH3) and ∆Hips are somewhat less important. It is likely
that differences between Zr and Hf reside in a combination of
∆IP and ∆D(M-CH3) effects.
Reaction Coordinates for Ion-Pair Formation and Reor-
ganization. The topographies of the schematic potential
surfaces such as in Figure 1 describe how rapidly the metallo-
cenium catalyst is formed from the neutral precursors, how
stable it is thermodynamically with respect to the neutral
precursors, how rapidly it reverts to the neutral precursors, and
how labile the ion pairing is with respect to separation of cation
and anion. The nature of the ion pairing is known to influence
catalytic activity, stereoselectivity, and chain transfer charac-
teristics, while the migration of the anion in concert with
reorientation of the M+-polymer vector may facilitate stereo-
(33) (a) This is most reasonable within homologous L2MCH3+ systems
where L ) a cyclopentadienyl ligand. For less coordinatively saturated/
+
sterically encumbered structures, such as Me2Si(Me4Cp)(tBuN)MCH3
,
arene complexes have been detected in cases of minimally coordinating
anions.4a,8 (b) Solvation energies in nonpolar solvents have been estimated
to be on the order of 15-30 kcal/mol.33c They should differ negligibly in
a homologous series.21 (c) Richardson, D. E.; Alameddin, N. G.; Ryan, M.
F.; Hayes, T.; Eyler, J. R.; Siedle, A. R. J. Am. Chem. Soc. 1996, 118,
11244-11253 and references therein. (d) Corrections in gas-phase homolytic
bond dissociation enthalpies for nonpolar solvents should be minor:
Castelhano, A. L.; Griller, D. J. Am. Chem. Soc. 1982, 104, 3655-3659.
(e) Differences in heats of sublimation should be minor in a homologous
series.33f,g (f) Connor, J. A.; Zafarani-Moattar, M. T.; Bickerton, J.; Saied,
N. I.; Suradi, S.; Carson, R.; Takhin, G. A.; Skinner, H. A. Organometallics
1982, 1, 1166-1174 and references therein. (g) Yoneda, G.; Lin, S.-M.;
Wang, L.-P.; Blake, D. M. J. Am. Chem. Soc. 1981, 103, 5768-5771.
(34) King, W. A.; di Bella, S.; Gulino, A.; Lanza, G.; Fragala´, I.; Marks,
T. J. Manuscript in preparation.
(35) (a) Green, J. C. Struct. Bonding (Berlin) 1981, 43, 37-112. (b)
Condorelli, G.; Fragala´, I.; Centineo, A.; Tondello, E. J. Organomet. Chem.
1975, 87, 311-315.
(36) Ciliberto, E.; Condorelli, G.; Fagan, P. J.; Manriquez, J. M.; Fragala´,
I.; Marks, T. J. J. Am. Chem. Soc. 1981, 103, 4755-4759.
(37) Fu, P.-F.; Wilson, D. C.; Rudolf, P. R.; Lanza, G.; Fragala´, I.; Marks,
T. J. Manuscript in preparation.

Electrophilic Olefin Polymerization Catalysts
J. Am. Chem. Soc., Vol. 120, No. 8, 1998 1783
Figure 8. Relationship between the enthalpy of ion pair formation
from the neutral precursors and the M-13CH3 chemical shift for a series
of (1,2-Me2Cp)2MCH3+CH3B(C6F5)3-/CH3B(C6F5)2Ar- complexes. The
boranes are those of Table 4.
-
Figure 9. Relationship of (1,2-Me2Cp)2MCH3+CH3B(C6F5)3
/
CH3B(C6F5)2Ar- ethylene polymerization activity to the enthalpy of
ion pair dissociation to neutrals. The boranes are those of Table 4.
reason is not immediately obvious since a more polar solvent
might be expected to stabilize the charged ion pair versus the
neutral constituents. One contributing explanation may be a
Lewis acid-base interaction between the borane and chloro-
carbon solvent (e.g., E). The Lewis acidity of B(C6F5)3 is
selection/olefin activation/facial selection processes.1,3b,9,10,37 The
present thermodynamic and kinetic data allow construction of
quantitative experimental enthalpic reaction coordinates describ-
ing the energetics of the above processes. Figure 5 and Table
+
5 show that in toluene solution, (1,2-Me2Cp)2ZrCH3
-
CH3B(C6F5)3 undergoes symmetrization more rapidly than
borane dissociation/recombination while the opposite scenario
+
holds for (1,2-Me2Cp)2HfCH3 CH3B(C6F5)3-. Both reaction
coordinates indicate that the enthalpic barrier to ion-pair
estimated to lie between that of BCl3 and BBr3.13 However, in
the M ) Hf series where there is a larger database and where
the magnitudes of both ∆Hdr and ∆Hqdr are somewhat smaller,
there is no evidence that polar solvents enhance the rate of
dissociation to neutrals. Importantly, the data all indicate that
formation from the neutral reactants is very small; i.e., ∆Hq
dr
- ∆Hdr ) a few kcal/mol at most. The same trend holds for
the ion-pair complexes of the less Lewis acidic boranes in
toluene solution (Table 4). In regard to dissociation of the ion
pairs to neutrals, Figures 6 and 7 illustrate approximately linear,
Brønsted/Hammett-like correlations39 between the enthalpies and
entropies of ion-pair dissociation into neutrals and the corre-
sponding enthalpies and entropies of activation for the same
processes. This demonstrates rather classical and well-behaved
transmission of substituent effects in the methide abstraction
process.
+
the repositioning rate of the CH3B(C6F5)-3 anion and M-CH3
vector, from one side of the L-M-L plane to the other,
increases dramatically in polar solvents.
Ion-Pair Thermodynamics, NMR Spectroscopic Param-
eters, and Polymerization Activity. To the extent that δc(M-
CH3) for a closely related, homologous series measures the
electrophilicity/electron deficiency of the metallocenium metal
center as modulated by the metallocenium ion-pair components,
it can be seen in Figure 8 that there is a roughly linear
relationship between the enthalpy of ion-pair formation and the
Figure 5 also depicts the effects of solvent polarity on the
+
-
(1,2-Me2Cp)2MCH3 CH3B(C6F5)3 ion-pair symmetrization,
formation, and dissociation profiles. For both M ) Zr and Hf,
polar solvents substantially reduce the kinetic barrier to ion-
pair separation, with rate enhancements reminiscent of more
classical systems.40 This is readily rationalized on the basis of
solvation and/or complexation (C) effects. In the case of M )
Zr, the barrier to dissociation to neutrals is also reduced. The
13C NMR downfield shift of the M-CH3 groups in the (1,2-
+
Me2Cp)2MCH3+ CH3B(C6F5)3-/CH3B(C6F5)2-Ar series. These
results strongly argue that fluoroarylborane Lewis acidity
modulates the electrophilicity/electron deficiency of the met-
allocenium cation.
(38) Chien, J. C. W.; Iwamoto, Y.; Rausch, M. D.; Wedler, W.; Winter,
H. H. Macromolecules 1997, 30, 3447-3458 and references therein.
(39) (a) Moore, J. W.; Pearson, R. G. Kinetics and Mechanism, 3rd ed.;
John Wiley: New York, 1981; pp 353-363 and references therein. (b)
Maskill, H. The Physical Basis of Organic Chemistry; Oxford University
Press: Oxford, 1985; Chapter 10.3-10.5. (c) Isaacs, N. S. Physical Organic
Chemistry, 2nd ed; Longman: Essex, 1995, Chapter 5.
Correlations of metallocenium physical observables with
olefin polymerization activities are complicated by several
factors. First, many of the (1,2-Me2Cp)2MCH3+ CH3B(C6F5)2-
Ar complexes are not completely in the cationic form for 0.140
mM solutions in toluene at 25 °C (Table 6). Second, the active
forms of the catalysts in ethylene polymerization are either
(40) Reference 39c, Chapter 5.

1784 J. Am. Chem. Soc., Vol. 120, No. 8, 1998
Deck et al.
interactions1 expected in the former as well as differing cation-
anion interactions likely in the latter.3e Nevertheless, plots of
ethylene polymerization activity versus either ∆Hdr or δ(M-
CH3) for the (1,2-Me2Cp)2MCH3+CH3B(C6F5)3-/CH3B(C6F5)2-
Ar- series evidence a (surprising) positive correlation (Figures
9 and 10). Such phenomenological relationships should be
useful in targeting future classes of highly active cationic olefin
polymerization catalysts and certainly warrant additional scru-
tiny.
Conclusions
The thermodynamic propensity and kinetic facility of a series
of triarylboranes and methylalumoxane to abstract the methide
anion from group 4 metallocene and related complexes and to
form highly electrophilic cations was quantified for the first
time by titration calorimetry as well as by static and dynamic
NMR spectroscopy. The energetics and kinetics of cation
formation and structural reorganization are a sensitive function
of borane acidity, metal ancillary ligation, and solvent polarity.
Many of the trends can be interpreted in terms of the electron
withdrawing power of the borane substituents, the ability of the
metallocene ancillary ligands to stabilize the cationic charge,
and the homolytic M-CH3 bond dissociation enthalpies. Within
the (1,2-Me2Cp)2MCH3+CH3B(C6F5)3-/CH3B(C6F5)2Ar- series,
qualitative correlations exist between the enthalpies of methide
+
abstraction, the 13C chemical shifts of the resulting M-CH3
groups, and the ethylene polymerization activities.
-
Figure 10. Relationship of (1,2-Me2Cp)2MCH3+CH3B(C6F5)3
/
CH3B(C6F5)2Ar- ethylene polymerization activity to the MCH3 13C
+
Acknowledgment. This research was supported by the
Department of Energy (Grant # DE-FG02-86ER13511) and by
the Dow Chemical Co. P.A.D. thanks the National Science
Foundation for a postdoctoral fellowship. We thank Dr. Wayne
A. King for advice on calorimetry.
chemical shift. The boranes are those of Table 4.
polyalkyl (L2M(CH2CH2)nH+ X- ) or hydrido (L2MH+ X-
)
cations. The metallocenium steric and electronic environment
is therefore not identical with that in the L2MCH3 X-
+
precursors, with differing alkyl steric encumberance and agostic
JA972912F