NOE and PGSE NMR Spectroscopic Studies of Solution Structure and Aggregation in Metallocenium Ion-Pairs

Published on Web 01/20/2004

NOE and PGSE NMR Spectroscopic Studies of Solution

Structure and Aggregation in Metallocenium Ion-Pairs

Cristiano Zuccaccia,^†,‡Nicholas G. Stahl,^†Alceo Macchioni,*^,‡Ming-Chou Chen,^†

John A. Roberts,^†and Tobin J. Marks*^,†

Contribution from the Department of Chemistry, Northwestern UniVersity, EVanston, Illinois

60208-3113, USA, and Dipartimento di Chimica, UniVersita` di Perugia, 06123, Perugia, Italy

Received September 25, 2003; E-mail: t-marks@northwestern.edu; alceo@unipg.it

Abstract: The solution structures of the metallocenium homogeneous polymerization catalyst ion-pairs

[Cp₂ZrMe]⁺[MeB(C₆F₅)₃]^-(1), [(1,2-Me₂Cp)₂ZrMe]⁺[MeB(C₆F₅)₃]^-(2), [(Me₂SiCp₂)ZrMe]⁺[MeB(C₆F₅)₃]^-(3),

[Me₂C(Fluorenyl)(Cp)ZrMe]⁺[FPBA]^-(FPBA ) tris(2,2′,2′′-nonafluorobiphenyl)fluoroaluminate) (4), [rac-

Et(Indenyl)₂ZrMe]⁺[FPBA]^-(5), [(Me₅Cp)₂ThMe]⁺[B(C₆F₅)₄]^-(6), [(Me₂SiCp₂)Zr(Me)(THF)]⁺[MeB(C₆F₅)₃]^-

(7), [(Me₂SiCp₂)Zr(Me)(PPh₃)]⁺[MeB(C₆F₅)₃]^-(8), [(Me₂SiCp₂)Zr(Me)(THF)]⁺[B(C₆F₅)₄]^-(9), [(Me₂Si(Me₄-

Cp)(t-BuN)Zr(Me)(solvent)]⁺[B(C₆F₅)₄]^-(solvent ) benzene, toluene) (10), [(Cp₂ZrMe)₂(µ-Me)]⁺[MePBB]^-

(PBB ) tris(2,2′,2”-nonafluorobiphenyl)borane) (11), and [(Cp₂Zr)₂(µ-CH₂)(µ-Me)]⁺[MePBB]^-(12), having

the counteranion in the inner (1, 3, 4, 5, and 6) or outer (7, 8, 9, 10, 11, and 12) coordination sphere, have

been investigated for the first time in solvents with low relative permittivity such as benzene or toluene by

¹H NOESY and ¹H,¹⁹F HOESY NMR spectroscopy. It is found that the average interionic solution structures

of the inner sphere contact ion-pairs are similar to those in the solid state with the anion B-Me (1, 3) or Al-F

(5) vectors oriented toward the free zirconium coordination site. The HOESY spectrum of complex 6 is in

agreement with the reported solid-state structure. In contrast, in outer sphere contact ion-pairs 7, 8, 9, and

10, the anion is located far from the Zr-Me⁺moiety and much nearer to the Me₂Si bridge than in 3. The

interionic structure of 8 is concentration-dependent, and for concentrations greater than 2 mM, a loss of

structural localization is observed. PGSE NMR measurements as a function of concentration (0.1-5.0

mM) indicate that the tendency to form aggregates of nuclearity higher than simple ion-pairs is dependent

on whether the anion is in the inner or outer coordination sphere of the metallocenium cation. Complexes

2, 3, 4, 5, and 6 show no evidence of aggregation up to 5 mM (well above concentrations typically used

in catalysis) or at the limit of saturated solutions (complexes 3 and 6), while concentration-dependent

behavior is observed for complexes 7, 8, 10, and 11. These outer sphere ion-pairs begin to exhibit significant

evidence for ion-quadruples in solutions having concentrations greater than 0.5 mM with the tendency to

aggregate being a function of metal ligation and anion structure. Above 2 mM, compound 8 exists as higher

aggregates that are probably responsible for the loss of interionic structural specificity.

Introduction

polymers is also known to be strongly affected by ion-pairing,

while the nature of the influence depends on the chemical and

It is now well-known that ion-pairing phenomena play a

fundamental role in the performance of single-site group 4

metallocenium olefin polymerization catalysts. Numerous ex-

amples exist where cation-anion interactions substantially affect

catalyst activity, stability, polymerization selectivity,^1,2and the

microstructural properties of the resulting polyolefins.³In addi-

tion, dianionic cocatalysts have been shown to enforce spatial

confinement between two catalyst centers, allowing the produc-

tion of LLDPE (linear low-density polyethylene) from a single

ethylene feed.^3cThe stereochemistry of single-site-derived

structural properties of the cationic and anionic moieties. Thus,

the stereoregularity of polypropylene produced by C_s- or C₁-

symmetric catalysts is modulated via ion-pairing strength effects

on the relative rates of olefin insertion to catalyst stereochemical

inversion.^3a,d,eIn contrast, the stereorigid systems based on rac-

(2) For recent examples of catalyst-activator interplay see: (a) Metz, M. V.;

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

references therein. (b) Chen, Y.-X.; Kruper, W. J.; Roof, G.; Wilson, D.

R. J. Am. Chem. Soc. 2001, 123, 745. (c) Zhou, J.; Lancaster, S. J.; Walter,

D. A.; Beck, S.; Thornton-Pett, M.; Bochmann, M. J. Am. Chem. Soc. 2001,

123, 223. (d) Chase, P. A.; Piers, W. E.; Patrick, B. O. J. Am. Chem. Soc.

2000, 122, 12911. (e) Chen, Y.-X.; Marks, T. J. in ref 1b, p 1391. (f) Metz,

M. V.; Schwartz, D. J.; Stern, C. L.; Nickias, P. N.; Marks, T. J. Angew.

Chem., Int. Ed. 2000, 39, 1312.

(3) (a) Mohammed, M.; Nele, M.; Al-Humydi, A.; Xin, S.; Stapleton, R. A.;

Collins, S. J. Am. Chem. Soc. 2003, 125, 7930. (b) Li, L.; Metz, M. V.; Li,

H.; Chen, M.-C.; Marks, T. J.; Liable-Sands, L.; Rheingold, A. L. J. Am.

Chem. Soc. 2002, 124, 12725 and references therein. (c) Abramo, G. P.;

Li, L.; Marks, T. J. J. Am. Chem. Soc. 2002, 124, 13966. (d) Chen, M.-C.;

Marks, T. J. J. Am. Chem. Soc. 2001, 123, 11803. (e) Chen, M.-C.; Roberts,

J. A. S.; Marks, T. J. J. Am. Chem. Soc., in press. (d) Chen, M.-C.; Roberts,

J. A. S.; Marks, T. J. Organometallics, in press.

^†Northwestern University.

^‡Universita` di Perugia.

(1) For recent reviews of single-site olefin polymerization, see: (a) Gibson V.

C.; Spitzmesser, S. K. Chem. ReV. 2003, 103, 283 (b) Pedeutour, J.-N.;

Radhakrishnan, K.; Cramail, H.; Deffieux, A. Macromol. Rapid Commun.

2001, 22, 1095. (c) Gladysz, J. A., Ed. Chem. ReV. 2000, 100 (special issue

on “Frontiers in Metal-Catalyzed Polymerization”). (d) Marks, T. J.;

Stevens, J. C., Eds. Top. Catal. 1999, 15 and references therein. (e)

Britovsek, G. J. P.; Gibson, V. C.; Wass, D. F. Angew. Chem., Int. Ed.

Engl. 1999, 38, 428. (f) Kaminsky, W.; Arndt, M. AdV. Polym. Sci. 1997,

127, 144. (g) Bochmann, M. J. Chem. Soc., Dalton Trans. 1996, 255.

9

1448

J. AM. CHEM. SOC. 2004, 126, 1448-1464

Solution-Phase Interionic Structure and Aggregation

A R T I C L E S

C₂-symmetric ansa-metallocenes with homotopic sites are less

sensitive to the nature of the cocatalyst as manifested in the

microstructure of the resulting polymer.⁴Finally, as proposed

by Busico and co-workers,⁵strong anion association with the

“mobile” active cation in “oscillating metallocene catalysts”⁶

can play a role analogous to that of the covalent bridge in ansa-

metallocenes.

less is known about interionic structure during the olefin

enchainment and propagation processes.^11,12

In the past few years, it has been demonstrated that the

solution-phase interionic structures of transition metal-organic

ion-pairs can be successfully defined by combining information

derived from NOE (nuclear Overhauser effect)¹³and PGSE

(pulsed field gradient spin-echo)¹⁴NMR experiments. Semi-

quantitative or quantitative NOE experiments allow deduction

of the relative anion-cation positions and orientations (when

both moieties are unsymmetrical). PGSE experiments provide

an estimate of the average volume of the ionic species present

and, consequently, the aggregation tendency.

Although the phenomenological observation of ion-pairing

effects is well recognized, the detailed connections between the

catalyst/cocatalyst interplay and polymerization activity are still

not completely understood. For example, it is not clear whether

cocatalyst effects are exerted in the ground or transition state

of “L₂M(olefin)R⁺” species. Landis and co-workers⁷used

heavy-atom kinetic isotope effects to determine the nature of

the transition state in 1-hexene homopolymerization catalyzed

by rac-[C₂H₄(1-Indenyl)₂]ZrMe₂with various cocatalysts. They

demonstrated the following: (1) the transition state in which

alkene is committed to irreversible insertion into the growing

polymer does not change significantly as a function of cocata-

lyst, (2) the alkene binding to the metal center is reversible,

and (3) the dramatic effect of the cocatalyst structure on reaction

rate arises from counteranion effects on the alkene association

equilibrium constant and consequently, on the ground state.

Seemingly in contrast, Bochmann⁸and co-workers used quenched-

flow propylene polymerization kinetic studies to demonstrate

that, even in the presence of essentially identical numbers of

active sites, i.e. in conditions of similar ground-state energies,

the activation energetics of the chain growth are strongly

modulated by the counteranion properties, implying a model in

which the ion-pair is the actual propagating species. In the same

work, the authors also suggested that anion exchange within

higher-order ion-pair aggregates may provide a low-energy

pathway for monomer binding, so that ion-quadruples or even

higher aggregates may be the actual propagating species.

In this scenario it is clear that direct structural characterization

of the cation-anion interactions in solution (i.e., in the medium

in which these catalysts actually function) would greatly

facilitate the basic mechanistic understanding of these complex

systems. Although numerous X-ray diffraction,⁹NMR spectro-

scopic,¹⁰and theoretical¹¹studies have been carried out to

elucidate metallocenium ion-pair structure and dynamics, there

is a paucity of direct experimental solution-phase metrical

information concerning relative cation-anion positions. Even

With the principal aim of better understanding metallocenium

ion-pair structure-reactivity relationships, we have investigated

a wide range of metallocenium ions by modifying the metal

ligation as well as the structure of the weakly coordinating

counteranion. Both ion-pairs in which the counteranion occupies

one of the metal coordination sites (an inner sphere ion-pair,

ISIP) and ion-pairs in which the counteranion is displaced from

the first coordination sphere (an outer sphere ion-pair, OSIP)

have been investigated. In ISIPs, the cation-anion interaction

is primarily electrostatic in nature¹¹but still possesses some

coordinative character in that the anion occupies one of the

otherwise formally vacant coordination sites of the cationic metal

center. Recent calculations by Lanza and co-workers^11a,cand

by Ziegler and co-workers^11dsuggest a single concerted step

for olefin uptake and insertion. In this case an ISIP can be

considered to be one of the most experimentally accessible

models for the catalyst during turnover. OSIPs represent contact

ion-pairs in which the coordinatively saturated first coordination

sphere of the cation is no longer accessible to the anion, and as

a consequence, the anion is relegated to the second coordination

sphere, interacting with the cation through electrostatic and other

weak forces (H-bonding, π-π, CH-π, etc.) only. In the classical

two-step Cossee-type mechanism,¹⁵the outer sphere ion-pair

can be considered to be a model to simulate the structural

characteristics of the active site.

Preliminary results on the aggregation behavior in benzene

solution of a limited series of isolable metallocenium ISIPs by

a colligative method (freezing point depression) and by PGSE

NMR have been recently reported by some of us.¹⁶The results

are consistent with a structural model for homogeneous met-

allocenium Ziegler-Natta polymerization catalysts consisting

of a metal cation paired with a weakly coordinating anion to

form an intimate 1:1 ion-pair in which aggregation is not

(4) Resconi, L.; Cavallo, L.; Fait, A.; Piemontesi, F. in ref 1c, p 1253.

(5) (a) Busico, V.; Van Axel Castelli, V.; Aprea, P.; Cipullo, R.; Segre, A.;

Talarico, G.; Vacatello, M. J. Am. Chem. Soc. 2003, 125, 5451. (b) Busico,

V.; Cipullo, R.; Kretschmer, W. P.; Talarico, G.; Vacatello, M.; Van Axel

Castelli, V. Angew. Chem., Int. Ed. Engl. 2002, 41, 505.

1

detectable. Here we report a far more detailed H NOESY,

¹⁹F,¹H HOESY, and PGSE NMR investigation of a broad series

of both ISIP (Chart 1) and OSIP metallocenium salts differing

in metal, ancillary ligation, and counteranion in toluene-d₈and

(6) (a) Lin, S.; Waymouth, R. M. Acc. Chem. Res. 2002, 35, 765. (b) Coates,

G. W.; Waymouth, R. M. Science 1995, 267, 217.

(7) Landis, C. L.; Rosaaen, K. A.; Uddin, J. J. Am. Chem. Soc. 2002, 124,

12062.

(8) Song, F.; Cannon, R. D.; Bochmann, M. J. Am. Chem. Soc. 2003, 125,

7641.

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

(b) Vranka, K.; Ziegler, T. Organometallics 2001, 20, 905. (c) Lanza, G.;

Fragala`, I. L.; Marks, T. J. J. Am. Chem. Soc. 2000, 122, 12764. (d) Chan,

M. S. W.; Ziegler, T. Organometallics 2000, 19, 5182. (e) Fusco, R.; Longo,

L.; Proto, A.; Masi, F.; Garbasi, F. Macromol. Rapid Commun. 1998, 19,

257. (f) Fusco, R.; Longo, L.; Masi, F.; Garbasi, F. Macromol. Rapid

Commun. 1997, 18, 433.

(12) (a) Stoebenau, E. J.; Jordan R. F. J. Am. Chem. Soc. 2003, 125, 3222. (b)

Landis, C. R.; Rosaaen, K. A.; Sillars, D. R. J. Am. Chem. Soc. 2003, 125,

1710.

(13) Macchioni, A. Eur. J. Inorg. Chem. 2003, 195 and references therein.

(14) Binotti, B.; Macchioni, A.; Zuccaccia, C.; Zuccaccia, D. Comments Inorg.

Chem. 2002, 6, 417.

(15) Cossee, P. J. Catal. 1964, 3, 80.

(16) Stahl, N. G.; Zuccaccia, C.; Jensen, T. R.; Marks, T. J. J. Am. Chem. Soc.

2003, 125, 5256.

(9) For recent examples see: (a) Metz, M. V.; Schwartz, D. L.; Stern, C. L.;

Marks, T. J. Organometallics 2002, 21, 4159. (b) Liu, Z.; Somsook, E.;

Landis, C. L. J. Am. Chem. Soc. 2001, 123, 2915. (c) Chen, E. Y.-X.;

Kruper, W. J.; Roof, G.; Wilson, D. R. J. Am. Chem. Soc. 2001, 123, 745.

(d) Bazan, G. C.; Cotter, W. D.; Komon, Z. J. A.; Lee, R. A.; Lachicotte,

R. J. J. Am. Chem. Soc. 2000, 122, 1371. (e) Stephan, D. W.; Stewart, J.

C.; Gue´rin, F.; Spence, R. E. v. H.; Xu, W.; Harrison, D. G. Organometallics

1999, 18, 116. (f) Bochmann, M.; Lancaster, S. J.; Hursthouse, M. B.;

Malik, K. M. A. Organometallics 1994, 13, 2235.

(10) (a) Beswick, C. L.; Marks, T. J. J. Am. Chem. Soc. 2000, 122, 10358. (b)

Deck, P. A.; Beswick, C. L.; Marks, T. J. J. Am. Chem. Soc. 1998, 120,

1772. (c) Chen Y.-X.; Metz, M. V.; Li, L.; Stern, C. L.; Marks, T. J. J.

Am. Chem. Soc. 1998, 120, 6287. (d) Chen, Y.-X.; Stern, C. L.; Marks, T.

J. J. Am. Chem. Soc. 1997, 119, 2582.

9

J. AM. CHEM. SOC. VOL. 126, NO. 5, 2004 1449

A R T I C L E S

Zuccaccia et al.

Chart 1

[FPBA]^-(5),^10c[(Me₅Cp)₂ThMe]⁺[B(C₆F₅)₄]^-(6),²⁰[(Me₂Si(Me₄Cp)(t-

BuN)Zr(Me)(Solvent)]⁺[B(C₆F₅)₄]^-(S ) benzene, toluene) (10),²⁰and

[(Cp₂ZrMe)₂(µ-Me)]⁺[MePBB]^-(PBB ) tris(2,2′,2′′-nonafluorobiphe-

nyl)borane) (11)^10cwere prepared and purified according to literature

benzene-d₆solutions at room temperature over significant

concentration ranges. The results of the NMR studies on the

ISIP and OSIP complexes are presented in sequence.

21

procedures. Si(p-tolyl)₄was prepared according to the literature

Experimental Section

procedure and further purified by vacuum sublimation. Si[Si(CH₃)₃]₄

was purchased from Aldrich and used without further purification.

(Me₅Cp)₂ZrMe₂,²²Cp₂ZrMe₂,²³(Me₂Si(Me₄Cp)(t-BuN)ZrMe₂,²⁴

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,

interfaced to a high-vacuum line (10^-5Torr), or in a nitrogen-filled

MBraun glovebox with a high-capacity recirculator (<1 ppm O₂and

H₂O). All solvents were freeze-pump-thaw degassed on the high-

vacuum line, dried over Na/K alloy, and vacuum-transferred to a dry

storage tube having a PTFE valve. B(C₆F₅)₃was a gift from Dow

Chemical and was purified by recrystallization from pentane and

vacuum sublimation. PPh₃was purchased from Aldrich and purified

by vacuum sublimation. [Cp₂ZrMe]⁺[MeB(C₆F₅)₃]^-(1),¹⁷[(1,2-Me₂-

Cp)₂ZrMe]⁺[MeB(C₆F₅)₃]^-(2),^10b[(Me₂SiCp₂)ZrMe]⁺[MeB(C₆F₅)₃]^-

(3),¹⁸[Me₂C(Fluorenyl)(Cp)ZrMe]⁺[FPBA]^-(FPBA ) tris(2,2′,2′′-

nonafluorobiphenyl)fluoroaluminate) (4),¹⁹[rac-Et(Indenyl)₂ZrMe]⁺-

Me₂C(Fluorenyl)(Cp)ZrMe₂,²⁵

Me₂C(Fluorenyl)(Cp)Zr(CH₂Ph)₂,²⁶

(20) Jia, L.; Yang, X.; Stern, C. L.; Marks, T. J. Organometallics 1997, 16,

842.

(21) Lambert, J. B.; Zhao, J.; Stern, C. L.; J. Phys. Org. Chem. 1997, 10, 229.

(22) Manriquez, J. M.; McAlister, D. R.; Sanner, R. D.; Bercaw, J. E. J. Am.

Chem. Soc. 1978, 100, 2716.

(23) Samuel, E.; Rausch, M. D. J. Am. Chem. Soc. 1973, 95, 6263.

(24) (a) Stevens, J. C.; Timmers, F. J.; Wilson, D. R.; Schmidt, G. F.; Nickias,

P. N.; Rosen, R. K.; Knight, G. W.; Lai, S. Y. (Dow Chemical Co.).

Constrained Geometry Addition Polymerization Catalysts, Processes for

Their Preparation, Precursors Therefore, Methods of Use, and Novel

Polymers Formed Therewith. EP0416815, Mar 13, 1991. (b) Canich, J.

M.; Hlatky, G. G.; Turner, H. W. (Exxon Chemical Patents, Inc.).

Aluminum-Free Monocyclopentadienyl Metallocene Catalysts for Olefin

Polymerization. WO-9200333 A2, Jan 9, 1992. (c) Canich, J. A. M. (Exxon

Chemical Patents, Inc.). Olefin Polymerization Catalysts. EP-420436A1,

April 3, 1991.

(17) Yang, X.; Stern, C. L.; Marks, T. J. J. Am. Chem. Soc. 1994, 116, 10015.

(18) Beck, S.; Prosenc, M. H.; Brintzinger, H.-H.; Goretzki, R.; Herfert, N.;

Fink, G. J. Mol. Catal. A: Chem. 1996, 111, 67.

(25) (a) Razavi, A.; Thewalt, U. J. Organomet. Chem. 1993, 445, 111. (b) Razavi,

A.; Ferrara, J. J. Organomet. Chem. 1992, 435, 299.

(26) Bochmann, M.; Lancaster, S. J. Organometallics 1989, 8, 476.

(19) Chen, M.-C.; Marks, T. J. J. Am. Chem. Soc. 2001, 123, 11803.

9

1450 J. AM. CHEM. SOC. VOL. 126, NO. 5, 2004

Solution-Phase Interionic Structure and Aggregation

A R T I C L E S

27

3

and (Me₅Cp)TiMe₃were synthesized and purified according to

25 °C): δ -137.07 (br, 8F, o-F), -164.85 (t, J_FF) 20.9 Hz, 4F,

p-F), -166.30 (m, 8F, m-F).

literature procedures. [(Me₂SiCp₂)Zr(Me)(THF)]⁺[MeB(C₆F₅)₃]^-(7)²⁸

and [(Me₂SiCp₂)Zr(Me)(PPh₃)]⁺[MeB(C₆F₅)₃]^-(8)²⁹were prepared by

in situ generation according to literature procedures or by a scale-up

reaction followed by purification. Complete resonance assignments for

compounds 8, 9, and 12 are available in Supporting Information.

[(Cp₂Zr)₂(µ-CH₂)(µ-Me)]⁺[MePBB]^-(12). Compound 12 was

generated cleanly in situ by allowing a solution of 11 to stand at 25 °C

1

for 7 days in a J-Young NMR tube. H NMR (C₆D₆, 25 °C): δ 7.51

(s, 2H), 5.89 (s, 20H), -0.83 (br, 3H), -1.49 (s, 3H). ¹⁹F NMR (C₆D₆,

[(Me₂SiCp₂)Zr(Me)(THF)]⁺[MeB(C₆F₅)₃]^-(7). Modified Litera-

ture Synthesis. [(Me₂SiCp₂)ZrMe]⁺[MeB(C₆F₅)₃]^-(21.4 mg, 26.1

µmol) was loaded into a flip-frit apparatus, which was then interfaced

to the high-vacuum line. Dry toluene (approximately 25 mL) was

condensed in, under vacuum, in a dry ice/acetone bath, and THF [freshly

distilled from Na/K alloy (2.3 µL, 29 µmol)], was added via syringe.

The cold bath was next removed and the solution allowed to warm to

25 °C while stirring. The toluene was removed in vacuo. The resultant

residue was recooled, and benzene was condensed in. The mixture was

warmed to 25 °C and stirred, and the benzene was then removed in

vacuo to remove residual toluene. The residue was then triturated

overnight in 20 mL of pentane with vigorous stirring and subsequently

dried in vacuo. Finally, the solid product was rinsed with benzene (3

× 2 mL) to remove minor, highly soluble impurities. The final product

is very pure by ¹H NMR, and the spectrum is consistent with previous

reports.²⁸

In situ Preparation. In the glovebox, [(Me₂SiCp₂)ZrMe]⁺[MeB-

(C₆F₅)₃]^-(11.4 mg, 13.9 µmol) was loaded into a screw-top vial and

dissolved in approximately 5 mL of benzene-d₆. In a separate glovebox,

9.0 µL (14 µmol) of a THF solution (1.54 M in C₆D₆) was added via

gastight syringe. The solution was lightly shaken, and finely dispersed

oil was observed. The mixture was loaded into a J-Young NMR tube,

and the denser oily phase was allowed to settle to the bottom of the

tube. The sample was then used for measurements without further

purification.

[(Me₂SiCp₂)Zr(Me)(PPh₃)]⁺[MeB(C₆F₅)₃]^-(8). In situ Prepara-

tion. In the glovebox, (Me₂SiCp₂)ZrMe]⁺[MeB(C₆F₅)₃]^-(16.0 mg, 19.5

µmol) and PPh₃(5.3 mg, 19.6 µmol) were loaded into a screw-top

vial, and approximately 5 mL of benzene-d₆was added. The solution

was lightly shaken, and finely dispersed oil was observed. The mixture

was loaded into a J-Young NMR tube, and the denser oily phase was

allowed to settle to the bottom of the tube. The sample was then used

for measurements without further purification. ¹H NMR (C₆D₆, 25

°C): δ 7.07 (m, 9H), 6.93 (m, 6H), 6.64 (pseudo q, 2H), 6.23 (pseudo

q, 2H), 5.29 (pseudo q, 2H), 4.89 (pseudo q, 2H), 1.39 (br, 3H), 0.38

(s, 3H), 0.14 (s, 3H), -0.02 (s, 3H). ¹⁹F NMR (C₆D₆, 25 °C): δ

25 °C): δ -123.94 (broad d, 3F), -139.26 (d,³J_FF) 23.7 Hz, 3F),

-139.48 (dd,³J_FF) 22.7 Hz, J_FF) 8.9 Hz, 3F), -139.98 (d,³J_FF

)

4

24.1 Hz, 3F), -155.61 (t,³J_FF) 21.3 Hz, 3F), -159.60 (t,³J_FF) 22.7

Hz, 3F), -163.00 (t,³J_FF) 22.0 Hz, 3F), -163.45 (dt,³J_FF) 22.3 Hz,

⁴J_FF) 7.2 Hz, 3F), -164.16 (m, 3F).

NOE Measurements. All the NMR experiments were performed

using a Bruker Avance DRX 400 equipped with a direct QNP probe

and a z-gradient coil controlled by a Great 1/10 gradient unit or a Varian

UNITYInova 400 MHz NMR instrument equipped with an inverse

probe and a z-gradient coil controlled by a Performa III PFG unit. The

¹H NOESY³⁰NMR experiments were acquired using the standard three-

pulse sequence or by the PFG version as described by Wagner and

Berger.³¹One-dimensional GOESY experiments were carried out as

proposed by Shaka and co-workers.³²Two-dimensional ¹⁹F,¹H HOESY

NMR experiments were acquired by using the standard four-pulse

sequence or the modified version as proposed by Lix, Sonnichsen, and

Syches;³³in both cases, the fluorine spectra were acquired in the direct

dimension. A simple modification of the latter sequence (see Supporting

Information) was used to acquire proton-detected one-dimensional

¹H,¹⁹F HOESY spectra. The shape of the selective pulses was Gaussian,

and the duration was 8 ms, while the transmitter power was adjusted

to obtain 90° pulses on the fluorine channel. Only a single high-

frequency coil for both ¹H and ¹⁹F was used for the measurements

carried out with the Varian UNITYInova instrument.³⁴The number of

transients and the number of data points were chosen according to the

sample concentration and to the desired final digital resolution.

Qualitative or semiquantitative one- or two-dimensional Overhauser

spectra were acquired using a 1-2 s relaxation delay and 300-800

ms mixing times. The 2D experiment is preferred in the case of

compounds with many resonances in the fluorine spectrum, while 1D

¹H-detected HOESY spectra with selective excitation of ¹⁹F resonances

are preferred in the case of low-concentration samples.³⁵As a control,

the two techniques were shown to yield equivalent results in the case

of complex 3.

Quantitative Overhauser experiments for complex 1 were carried

out in toluene-d₈at 25 °C. For both the ¹H NOESY and ¹H,¹⁹F HOESY

NMR experiments, a relaxation delay of 25 s and a mixing time of 0.1

s were employed (initial rate approximation).³⁶Quantitative ¹H NOESY

3

-137.07 (brd, J_FF) 21.0 Hz, 6F, o-F), -164.85 (t, J_FF) 21.4 Hz,

3F, p-F), -167.15 (m, 6F, m-F).

[(Me₂SiCp₂)Zr(Me)(THF)]⁺[B(C₆F₅)₄]^-(9). (Me₂SiCp₂)ZrMe₂(50.0

and proton-detected one-dimensional H,¹⁹F HOESY experiments for

1

mg, 0.163 mmol) and Ph₃C⁺B(C₆F₅)₄(71.2 mg, 0.077 mmol) were

-

complexes 3 and 7 were carried out in benzene-d₆at 25 °C using the

pulse sequence shown in Figure S1. A relaxation delay of 6 s and a

mixing time of 0.150 s were employed (initial rate approximation).³⁶

loaded into a flip-frit apparatus and interfaced to the high-vacuum line.

Dry toluene (approximately 25 mL) was condensed in under vacuum

in a dry ice/acetone bath. The cold bath was removed and the mixture

allowed to warm slowly to 25 °C with stirring. After stirring at 25 °C

approximately 30 min, 0.27 mL of a toluene solution of THF was added

(0.30 M, 0.081 mmol THF). The toluene was then removed in vacuo.

The resulting residue was recooled, and pentane (approximately 25 mL)

was condensed in. The mixture was stirred for a few minutes, and the

pentane was filtered off to remove unreacted (Me₂SiCp₂)ZrMe₂and

Ph₃CCH₃. The solid product was then dried in vacuo (10^-6Torr).

Finally, the solid product was rinsed with benzene (3 × 2 mL) to

remove minor soluble impurities. ¹H NMR (C₆D₆, 25 °C): δ 6.47

PGSE Measurements. All PGSE experiments were performed on

the Varian UNITYInova 400 MHz NMR instrument. The standard

Stejskal-Tanner pulse sequence, as described by Pregosin and co-

(30) Jeener, J.; Meier, B. H.; Bachmann, P.; Ernst, R. R. J. Chem. Phys. 1979,

71, 4546.

(31) Wagner, R.; Berger, S. J. Magn. Reson. A 1996, 123, 119.

(32) (a) Stott, K.; Stonehouse, J.; Keeler, J.; Hwang, T. L.; Shaka, A. J. J. Am.

Chem. Soc. 1995, 117, 4199. (b) Stonehouse, J.; Adell, P.; Keeler, J.; Shaka,

A. J. J. Am. Chem. Soc. 1994, 116, 6037.

(33) Lix, B.; So¨nnichsen, F. D.; Sykes, B. D. J. Magn. Reson. A 1996, 121, 83.

(34) Urhin, D.; Beatty, E. J.; Barlow, P. N.; Ramage, R.; Sandor, P.; McSparron,

H.; Wilken, J.; Starkmann, B.; Young, D. W. Magn. Moments Online 1999,

10, 1.

(dt,³J_HH) 3.0 Hz, ⁴J_HH) 1.8 Hz, 2H), 5.84 (dt,³J_HH) 3.0 Hz, ⁴J_HH

1.8 Hz, 2H), 5.54 (pseudo q, 2H), 5.01 (pseudo q, 2H), 2.83 (m, 4H),

1.13 (m, 4H), 0.32 (s, 3H), 0.23 (s, 3H), -0.07 (s, 3H). ¹⁹F NMR (C₆D₆,

)

(35) The proton-detected 1D experiment is, as expected, more sensitive to

subtraction artifacts and sometimes proves to be incompletely efficient in

suppressing strong signals not involved in the interactions (e.g., the solvent

signal or Si(p-tolyl)₄signals). On the other hand, this represents only a

“cosmetic” effect on the final spectrum and does not introduce any

difficulties in data interpretation.

(27) Mena, M.; Royo, P.; Serrano, R.; Pellinghelli, M. A.; Tiripicchio, A.

Organometallics 1993, 12, 633.

(28) Schaper, F.; Geyer, A.; Brintzinger, H.-H. Organometallics 2002, 21, 473.

(29) Beck, S.; Prosenc, M. H.; Brintzinger H.-H. J. Mol. Catal. A: Chem. 1998,

128, 41.

(36) Neuhaus, D.; Williamson, M. The Nuclear OVerhauser Effect in Structural

and Conformational Analysis; Wiley-VCH: New York, 2000.

9

J. AM. CHEM. SOC. VOL. 126, NO. 5, 2004 1451

A R T I C L E S

Zuccaccia et al.

workers, was employed (∆ ) 54 ms, δ ) 3 ms).³⁷The nonstimulated

version was used because of its greater sensitivity. For this reason,

only singlet resonances were usable in the analysis. The gradient pulses

were rectangular, and their strength was varied during the course of

the experiment. Each spectrum was acquired using 16 K complex points,

while the number of transients acquired at each gradient level was

adjusted depending on the sample concentration (32 scans at the highest

concentrations, 1024 for the lowest concentration). The recycle delay

was adjusted for each sample by measuring the T₁value (using the

standard inversion recovery method) for all of the resonances of interest

and setting the recycle delay to 5 times the T₁value of the most slowly

relaxing resonance of interest. Signal intensities, I, were measured by

integration. Plots of ln(I/I₀) vs G², according to eq 1, were fitted using

a standard linear regression algorithm implemented in the Origin data

analysis software package, yielding a slope m (directly proportional to

the diffusion coefficient D), with a correlation coefficient consistently

greater than 0.999.

is a numerical factor that depends on the size and shape of the solute

and the hydrodynamic behavior of the solute-solvent system, and η

is the viscosity of the solution. Assuming that both solute and solvent

are spherical, the factor c is found to depend on the solute:solvent ratio

of radii according to eq 4:⁴¹

6

c )

(4)

3r_solv

1

r_solv

+

r_H

1 +

[

]

r_H

A semiempirical improvement of eq 4 has been derived:⁴²

6

c )

(5)

2.234

r_solv

1 + 0.695

(

)

[

]

r_H

By substituting eq 5 into eq 3, the following expression can be obtained:

I

I₀

δ

3

ln ) mG²in which m ) (γδ)²D ∆ -

(1)

(

)

2.234

r_solv

kT 1 + 0.695

(

)

[

]

r_H

The diffusion coefficients of four different molecules (standards),

namely Si(p-tolyl)₄, Si[Si(CH₃)₃]₄, (Me₅Cp)₂ZrMe₂, and Cp₂ZrMe₂, were

estimated according to eq 2 by measuring the m parameter for a sample

of HDO in D₂O obtained from Aldrich (%D ) 99.8).³⁸

D )

(6)

6πηr_H

It is known from the literature that, for relatively small molecules,

the van der Waals radius is the best approximation for the hydrodynamic

radius employed in the Stokes-Einstein equation.⁴³Introducing the

van der Waals radius of benzene-d₆(2.7 Å) in eq 6, the hydrodynamic

radii of Si(p-tolyl)₄, Si[Si(CH₃)₃]₄, (Me₅Cp)₂ZrMe₂, and Cp₂ZrMe₂were

estimated, and the hydrodynamic volumes (V_H) were calculated,

pragmatically assuming a spherical shape. Estimation of hydrodynamic

volumes for metallocenium ion-pairs by determination of diffusion

coefficients could in principle be carried out in the same way, provided

that errors originating from variability in gradient strength and

temperature could be gauged and that the metallocene concentrations

remained low. However, this approach is then not appropriate for

concentration-dependent studies. Measurement relative to an internal

reference of known diffusion coefficient will reduce errors due to

instrument instabilities as well as systematic errors due to changes in

solution viscosity.⁴⁴In principle, any internal reference with a molecular

volume similar to that of the species under investigation could be used.

We chose Si(p-tolyl)₄as an internal standard which combines high

chemical inertness toward metallocenium ion-pair complexes and a

suitable NMR spectrum with a sharp singlet at δ 2.10 ppm in benzene-

d₆that does not overlap with the resonances of the species under

investigation.

D_HDO

D_standard) m_standard

(2)

m_HDO

The measurements were carried out on a 10^-4M benzene-d₆solution

of the standards (in this condition it is a reasonable assumption that

the solution viscosity can be approximated by that of pure benzene-

d₆³⁹) at room temperature (∼21-23 °C) with the instrument temperature

control turned off and without spinning. Due to the temperature

dependence of the diffusion coefficient of HDO in D₂O, as well as the

temperature dependence of the solvent viscosity, the actual sample

temperature inside the probe must be measured. At the beginning and

at the end of each PGSE experiment on the standards, both neat

methanol and neat ethylene glycol were used to measure the temper-

ature. The actual temperature values were obtained using the standard

“tempcal(“m”)” and “tempcal(“e”)” commands implemented in the

VNMR 6.1C software package. At 22.0 ( 0.5 °C, the diffusion

coefficient of the four compounds were found to be: Si(p-tolyl)₄, D )

8.31 × 10^-10m²s^-1; Si[Si(CH₃)₃]₄, D ) 1.12 × 10^-9m²s^-1; (Me₅-

Cp)₂ZrMe₂, D ) 1.01 × 10^-9m²s^-1; Cp₂ZrMe₂, D ) 1.22 × 10^-9m²

s^-1

.

Samples were prepared in the glovebox using a 10^-3M stock solution

of Si(p-tolyl)₄in benzene-d₆. The samples were then transferred into a

PTFE-valved J-Young NMR tube and kept frozen at -78 °C before

use. The actual concentration of each sample was determined by

integration versus the Si(p-tolyl)₄internal standard. Variation of the

sample concentrations was achieved by diluting the most concentrated

sample with the same 10^-3M stock solution of Si(p-tolyl)₄in benzene-

d₆. In some cases, to reduce total experiment time, additional Si(p-

tolyl)₄was added to a given NMR sample. For each experiment, the

diffusion coefficient value obtained for a given compound at a given

concentration was corrected using the internal standard. In this way,

The diffusion data were treated as recently described by some of

us.⁴⁰According to the Stokes-Einstein model, the diffusion coefficient

D can be related to the hydrodynamic radius (r_H) of the diffusing particle

according to eq 3:

kT

cπηr_H

D )

(3)

in which k is the Boltzmann constant, T is the absolute temperature, c

(37) (a) Valentini, M.; Ru¨egger, H.; Pregosin, P. S. HelV. Chim. Acta 2001, 84,

2833. (b) Valentini, M.; Pregosin, P. S.; Ru¨egger, H. Organometallics 2000,

19, 2551. (c) Holz, M.; Weinga¨rtner, H. J. Magn. Reson. 1991, 92, 115.

(38) Mills, R. J. Phys. Chem. 1973, 77, 685. Data at different temperatures were

(41) (a) Espinosa, P. J.; de la Torre, J. G. J. Phys. Chem. 1987, 91, 3612. (b)

Gierer, A.; Wirtz, K. Z. Naturforsch. A 1953, 8, 522. (c) Wirtz, K. Z.

Naturforsch. A 1953, 8, 532.

(42) Chen, H.-C.; Chen, H.-S. J. Phys. Chem. 1984, 88, 5118.

(43) Edward, J. T. J. Chem. Educ. 1970, 47, 261.

(44) See for example: (a) Macchioni, A.; Romani, A.; Zuccaccia, C.; Gugliel-

metti G.; Querci C. Organometallics 2003, 22, 1526. (b) Babushkin, D.

E.; Brintzinger, H.-H. J. Am. Chem. Soc. 2002, 124, 12869. (c) Burini, A.;

Fackler, J. P. Jr.; Galassi, R.; Macchioni, A.; Omary, M. A.; Rawashdeh-

Omary, M. A.; Pietroni, B. R.; Sabatini, S.; Zuccaccia, C. J. Am. Chem.

Soc. 2002, 124, 4570. (d) Zuccaccia, C.; Bellachioma, G.; Cardaci G.;

Macchioni A. Organometallics 2000, 19, 4663.

estimated by interpolation of the data reported by Mills, giving D_HDO

)

1.748 × 10^-9m²s^-1at 22 °C.

(39) Viscosity of C₆D₆was estimated to be 0.672 cp at 22 °C by interpolation

of the data reported for C₆H₆(CRC Handbook of Chemistry and Physics,

51st ed.; Weast, R. C., Ed.; Chemical Rubber: Cleveland, 1971 and CRC

Handbook of Chemistry and Physics, 77th ed.; Lide, D. R., Ed.; CRC

Press: New York, 1996) and application of the correction proposed by

Mills (Mills, R. J. Phys. Chem. 1976, 80, 888.): η(C₆D₆) ) 1.063 × η-

(C₆H₆).

(40) Zuccaccia, D.; Sabatini, S.; Bellachioma, G.; Cardaci, G.; Clot, E.;

Macchioni, A. Inorg. Chem. 2003, 42, 5465.

9

1452 J. AM. CHEM. SOC. VOL. 126, NO. 5, 2004

Solution-Phase Interionic Structure and Aggregation

A R T I C L E S

Table 1. Diffusion Coefficients (D, 10^-10m²s^-1), Corrected Diffusion Coefficients^a(D*, 10^-10m²s^-1), Hydrodynamic Radii (r_H, Å),

Hydrodynamic Volumes (V_H, Å³), c Factor, and Aggregation Number (N) Values at Various Concentrations (units 10^-3M) for the Standards,

the Neutral Metallocenium Precursors, and the ISIPs^b

concn

D

TPTS

D*

r_H

c

V_H

N

Si(p-tolyl)₄(TPTS)

Si[Si(CH₃)₃]₄(TMSS)

(Me₅Cp)₂ZrMe₂

Cp₂ZrMe₂

0.10

9.75

2.46

8.74

2.00

7.34

15.67

9.90

4.80

0.80

2.50

5.00

1.00

8.60

2.80

1.37

0.68

5.00

2.80

1.15

3.50

2.05

8.31^c

11.2^c

10.1^c

12.2^c

11.8

9.44

9.92

7.69

9.63

11.1

6.86

6.64

6.62

7.08

6.92

6.72

6.53

6.55

6.58

6.81

5.34

5.76

6.20

6.00

5.40

4.66

3.80

4.06

3.63

3.64

4.13

3.94

4.70

4.11

3.47

5.12

5.19

5.39

4.92

5.00

5.05

5.00

4.97

5.76

5.72

5.70

5.72

5.0

4.5

4.7

4.4

4.7

4.6

5.0

4.7

4.3

5.1

5.2

5.1

5.3

424

232

280

200

202

295

256

435

291

175

562

586

656

499

524

539

524

514

800

784

776

784

1.20

0.86

1.11

0.95

1.01

0.86

0.96

0.89

0.97

0.99

1.03

0.96

1.11

1.03

1.01

0.98

0.97

0.99

0.97

0.96

0.97

(Me₅Cp)TiMe₃

8.20

7.94

7.77

8.02

7.11

7.79

7.66

8.02

7.63

7.61

7.40

7.27

7.21

7.24

7.42

7.03

7.54

8.08

7.82

7.07

12.0

Me₂C(Fluorenyl)(Cp)ZrMe₂

Me₂Si(Me₄Cp)(t-BuN)ZrMe₂

Me₂C(Fluorenyl)(Cp)Zr(CH₂Ph)₂

(Me₅Cp)₂ZrMe₂

9.89

10.6

8.23

9.97

12.9

Cp₂ZrMe₂

[Me₂Si(Me₄Cp)(t-BuN)TiMe]⁺[MeB(C₆F₅)₃]^-

[Me₂C(Fluorenyl)(Cp)ZrMe]⁺[MeB(C₆F₅)₃]^-

[(Me₅Cp)₂ThMe]⁺[B(C₆F₅)₄]^-(6)

[Cp₂ZrMe]⁺[MeB(C₆F₅)₃]^-(1)

[(Me₂Cp)₂ZrMe]⁺[MeB(C₆F₅)₃]^-(2)

7.32

7.20

6.85

7.71

7.56

7.46

7.55

7.56

7.63

6.31

6.35

6.38

6.37

6.35

[(Me₂SiCp)₂ZrMe]⁺[MeB(C₆F₅)₃]^-(3)

[Me₂C(Fluorenyl)(Cp)ZrMe]⁺[FPBA]^-(4)

[C₂H₄(Indenyl)₂ZrMe]⁺[FPBA]^-(5)

^aThe corrected D* values correspond to a hypothetical measurement carried out at 22 °C, in a solution containing the reported nominal concentration,

but having the viscosity of pure C₆D₆(see Experimental Section for details). ^bThe numbers in bold are represented graphically in Figure 1. ^cCarried out at

22 °C.

the corrected values correspond to a hypothetical experiment carried

out at 22.0 ( 0.5 °C in which the solution viscosity is approximated

by the bulk solvent viscosity. Hydrodynamic radii (r_H) were obtained

by application of eq 6, and hydrodynamic volumes (V_H) were calculated

assuming a spherical shape. Data are reported in terms of the ratio of

the apparent PGSE experimental volume at a given concentration to

the van der Waals volume of the 1:1 ion-pair calculated from X-ray

coordinates or molecular modeling. This ratio represents the aggregation

number (N) in a way similar to that defined by Pochapsky⁴⁵and is

useful when trends over a range of concentrations are to be compared.

Structural models for cations and anions were obtained from experi-

mental crystal structure data (Cambridge Structural Database) or by a

simple energy minimization using the Spartan software package when

a crystal structure was not available. In the case of the outer sphere

ion-pairs, the two ions were modeled separately. These models were

then used to estimate the van der Waals volumes using the software

package WebLab Viewer.

The measurement uncertainty in the diffusion data was estimated

by determining the standard deviation of the ratio m_species/m_standardfor a

benzene-d₆solution containing 2.8 × 10^-3M of [(Me₂SiCp₂)ZrMe]⁺-

[MeB(C₆F₅)₃]^-(3) and 1 × 10^-3M of Si(p-tolyl)₄across the series ∆

) 34, 54, 74, and 94 ms.³⁷Standard propagation of errors analysis

yielded a standard deviation of approximately 1.7% in the radius and

thus 5% in the volume. Any remaining deviations of the PGSE-derived

volumes from the computed van der Waals volumes of the ISIPs are

most likely due to deficiencies in the modified Stokes-Einstein model

and in the assumption that hard-sphere van der Waals volumes are a

good representation of the hydrodynamic volume of a given species.

Representative examples of the PGSE data acquisitions are reported

in the Supporting Information.

an excess of solid 9 and benzene-d₆to reflux and allowing it to stand

at room temperature for 3 h. A crystal of dimensions 0.540 mm ×

0.150 mm × 0.100 mm was 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:

space group P2₁/c, Z ) 4, a ) 16.037(3) Å, b ) 16.263(3) Å, c )

18.351(3) Å, â ) 104.515(3)°, V ) 4633.2(14) Å³, D_calc) 1.608 g/cm³,

F(000) ) 2240. Of 37809 measured reflections, 10806 were indepen-

dent and 6273 gave I > 2σ(I). The initial crystal structure solution

was obtained by direct methods and refined through successive least-

squares cycles, and a face-indexed absorption correction was applied:

µ ) 0.381 mm^-1, T_min) 0.87164, T_max) 0.96566. The refinement

was carried to convergence. Hydrogen atoms were placed in idealized

positions and refined isotropically with fixed U_equnder standard riding

model constraints: (∆/σ)_max) 0.022, (∆/σ)_mean) 0.001, ∆F_max) 1.488

e/Å³(adjacent to Zr), ∆F_min) -0.734 e/Å³, R[F²> 2σ(F²)] ) 5.92%,

wR(F²) ) 15.84%. Full crystal data collection and refinement param-

eters can be found in the Crystallographic Information File.

Results

PGSE and NOE NMR Investigations of Inner Sphere Ion-

Pairs (ISIPs). Table 1 summarizes the results of the PGSE

measurements carried out on inner sphere metallocenium ion-

pairs and on some neutral alkylmetallocenes with various ligand

frameworks. Hydrodynamic volumes are graphically depicted

in Figure 1 in comparison with the crystallographically or

computationally derived van der Waals volumes. There is very

good agreement between the experimental hydrodynamic vol-

umes and the van der Waals volumes, which indicates that all

of these complexes exist in solution predominantly, if not

exclusively, as discrete 1:1 ion-pairs. As far as the ISIPs are

concerned, concentration-dependent PGSE measurements were

performed on compounds 2, 3, 4, and 5. All show invariant

X-ray Crystal Structure Determination of [(Me₂SiCp₂)Zr(Me)-

(THF)]⁺[B(C₆F₅)₄]^-(9). Crystals of 9 suitable for X-ray diffraction

were obtained as colorless needles by heating an NMR tube containing

(45) (a) Mo, H.; Pochapsky, C. T. J. Phys. Chem. B. 1997, 101, 4485. (b)

Pochapsky, S. S.; Mo, H.; Pochapsky, C. T. J. Chem. Soc., Chem. Commun.

1995, 2513.

9

J. AM. CHEM. SOC. VOL. 126, NO. 5, 2004 1453

A R T I C L E S

Zuccaccia et al.

Figure 1. Plot of PGSE-derived hydrodynamic volumes (V_PGSE) versus

van der Waals volumes (V_vdW) computed for various metallocenium

compounds. The data points refer to the values reported in bold in Tables

1 and 5.

apparent molecular volumes over a wide range of concentrations.

Compound 3, for example, continues to behave as a single 1:1

ion-pair up to 8.0 mM.⁴⁶

The ¹H,¹⁹F HOESY spectrum of complex 1 evidences strong

NOE interactions between the o-F nuclei of the B(C₆F₅)₃moiety

and the µ-Me group as well as between B(C₆F₅)₃o-F nuclei

and protons on the Zr-Me group and Cp ligands (Cp )

cyclopentadienyl). No cation-anion NOE interactions are

observed for the B(C₆F₅)₃m-F and p-F fluorine nuclei. In

complex 3 the ansa-Me₂Si bridge inhibits free rotation of the

Cp ligands, allowing determination of the relative orientation

Figure 2. Section of the ¹⁹F,¹H HOESY spectrum (376.4 MHz, relaxation

delay ) 2 s, mixing time ) 800 ms, toluene-d₈, 298 K) of complex 3. The

F1 trace (indirect dimension) relative to the o-F resonance is reported on

the right. See Figure S7 for the corresponding proton-detected one-

1

dimensional H,¹⁹F HOESY experiments.

1

of the anion with respect to the cation. The H,¹⁹F HOESY

spectrum of complex 3 is shown in Figure 2, and a section of

1

the corresponding H NOESY spectrum is shown in Figure 3.

The intensity of the interactions between the o-F fluorine nuclei

on the anion and the Cp protons in the ¹H,¹⁹F HOESY spectrum

follows the order H2 > H1 > H3, and no cross-peaks are

detectable for the H4 proton. This indicates that the preferred

contact point for the anion is proximate to H2 (Figure 2). This

contention is confirmed by the strong homonuclear cross-peak

between the µ-Me group and the H2 proton in the ¹H NOESY

spectrum (Figure 3). As in the case of complex 1, no ¹⁹F-¹H

interionic interactions are detected for the m-F and p-F nuclei.

Complex 5 exists in solution as a 69:31 mixture of diaster-

eomers as a consequence of the chirality in both the [FPBA]^-

anion and in the cation. This observation alone suggests that

this complex is present in solution mainly as a tight ion-pair.

Several interionic proton-fluorine dipolar interactions can be

detected in the corresponding ¹H,¹⁹F HOESY spectrum (Figure

4).⁴⁷The Al-F, F6, F5, and o-F* fluorine atoms interact with

the protons on the cation, and in particular, for the major

diastereomer, the following contacts can be detected: Al-F with

H2, Zr-Me, H13, and H12/H3; F6 with H2, H1, Zr-Me, and

1

Figure 3. Section of the H NOESY spectrum (399.94 MHz, relaxation

delay ) 6 s, mixing time ) 150 ms, benzene-d₆, 298 K) of complex 3.

H12/H3; F5 with H3/H12, H10, and H11; and o-F* with H2,

H11, H1, and H12/H3. The ¹⁹F,¹H HOESY spectrum of

compound 6 (see Supporting Information) shows that both the

o-F and m-F fluorine atoms of the [B(C₆F₅)₄]^-anion exhibit

NOE interactions with both the Cp-Me and the Th-Me protons.

A very small interaction is also observed between the Cp-Me

group and the p-F nuclei.

The possibility of estimating the average internuclear dis-

tances in solution was also explored by acquiring quantitative

homo- and heteronuclear NOE spectra using the initial rate

approximation.³⁶The average interionic internuclear distances

derived for complexes 1 and 3 are compared in Tables 2 and 3.

PGSE and NOE NMR Investigations of Outer Sphere Ion-

Pairs (OSIPs). Upon addition of 1.0 equiv of THF or PPh₃to

complex 3, the methylborate anion is displaced from the metal

center to form OSIPs 7 and 8, respectively.^28,29The ¹H NOESY

(46) By dissolving purified complex 3 (see ref 16) at room temperature in

benzene-d₆, the highest concentration we were able to achieve was 2.8 mM.

The 8.6 mM solution was obtained by heating the NMR tube at the reflux

temperature of the solvent while some solid was still present on the bottom

of the tube; the initial NMR spectrum shows little evidence of decomposi-

tion. According to the data of Brintzinger and co-workers (Beck S.; Lieber,

S.; Schaper, F.; Geyer, A.; Brintzinger, H.-H. J. Am. Chem. Soc. 2001,

123, 1483), this same compound, formed in situ from the corresponding

dimethyl complex with a slight excess (1.1 equiv) of B(C₆F₅)₃, could be

maintained in C₆D₆solution at a concentration around 20 mM.

(47) The assignments of both ¹H and ¹⁹F resonances were made using

homonuclear COSY and NOESY experiments (see Supporting Information).

Due to the extensive spectral overlap in the ¹H spectrum of the major isomer,

the proton resonances were assigned using the more resolved resonances

of the minor isomer, taking advantage of the selective EXSY peaks due to

the rapid anion racemization (see ref 10c).

9

1454 J. AM. CHEM. SOC. VOL. 126, NO. 5, 2004

Solution-Phase Interionic Structure and Aggregation

A R T I C L E S

Table 2. Comparison between Experimental and Theoretical Average Internuclear Distances (Å) for Complex 1

c

calculated^a

experimental^b

calculated

experimental^c

r

_Zr-Me/Cp/ r

_o-F/B-Me/ r

1.00

0.69

0.64

0.81^d

0.99

0.69

0.73

r

3.8

4.8

5.2

4.1^d

3.8

4.8

4.6

B-Me/Cp

o-F/Zr-Me

o-F/Cp

B-Me/Cp

o-F/Zr-Me

o-F/Cp

^aCalculated from X-ray metrical parameters,⁴⁹considering all possible internuclear vectors, assuming that rotation of the methyl groups is faster than the

overall molecular correlation time (r^-3average).^{75 b}Calculated from the relative cross-peak integrals, taking into account the number of equivalent nuclei.⁷⁶

^cIn the case of H-H distances, the reference distance is r _Zr-Me/Cp(3.8 Å), while in the case of F-H, the reference distance is r _o-F/BMe(3.4 Å). Both are

calculated from the X-ray structure,⁴⁹considering all the possible internuclear vectors and assuming that the rotation of the methyl groups is more rapid than

the overall molecular correlation time (r^-3average). ^dCalculated from X-ray metrical parameters,⁴⁹considering all possible internuclear vectors and assuming

that rotation of the interionic average vector connecting the Cp protons and the o-F fluorine atoms is slower than the overall ion-pair correlation time (r^-6

average). In this case, it is likely that the difference between the internal motion and the overall relaxation time is less pronounced, so that neither model

reproduces the experimental data well.

Table 3. NOESY- and HOESY-Derived Internuclear Distances (Å)

for Complex 3

experimental

r

3.4^a

4.1^b

4.7^b

5.1^b

5.0^b

r

3.1^a

3.2^b

3.0^b

3.7^b

3.9^b

o-F/B-Me

o-F/H2

Zr-Me/H3

Me(A)/H1

Me(B)/H4

Zr-Me/H4

B-Me/H2

B-Me/H1

B-Me/H3

o-F/H1

o-F/H3

o-F/Me-Zr

^aCalculated from X-ray metrical parameters of complex 1, considering

all possible internuclear vectors and assuming that rotation of the methyl

groups is faster than the overall molecular correlation time (r^-3average).⁷⁵

^bCalculated from the relative cross-peak integrals, taking into account the

number of equivalent nuclei,⁷⁶using the r _o-F/B-Me(3.4 Å) as a reference

distance for the H-F distances, and the r

distances.

(3.1 Å) for the H-H

Zr-Me/H3

Table 4. NOESY- and HOESY-derived Interionic Heteronuclear

Distances (Å) for Complex 7

experimental

calculated^a

r

3.4^a

4.6^b

4.5^b

5.1^b

4.9^b

5.1^b

r

3.1^a

3.3^b

3.4^b

3.2^b

3.7^b

4.1^b

4.5^b

3.7^b

3.2^b

3.9^b

3.1

3.3

3.4

3.2

4.0

o-F/B-Me

o-F/H2

o-F/H1

o-F/H3

o-F/R

Zr-Me/H3

Me(A)/H1

Me(B)/H4

Zr-Me/H4

Zr-Me/R

B-Me/H1

_â/H2

_R/H1

_R/H2

_R/H3

o-F/â

4.6

4.1

3.2

4.4

^aCalculated from X-ray metrical parameters of the cationic portion of

complex 9, considering all possible internuclear vectors and assuming that

rotation of the methyl groups is faster than the overall molecular correlation

time (r^-3average).^{75 b}Calculated from the relative cross-peak integrals,

Figure 4. ¹⁹F,¹H HOESY spectrum (376.6 MHz, relaxation delay ) 1 s,

mixing time ) 800 ms, benzene-d₆, 298 K) of complex 5.

taking into account the number of equivalent nuclei,⁷⁶using the r

o-F/B-Me

(3.4 Å) as a reference distance for the H-F distances, and the r

Zr-Me/H3

spectrum of a 1.7 mM solution of 7 in benzene-d₆is illustrated

in Figure 5. Homonuclear interionic NOE interactions are

observed between the B-Me group and the H1 protons of the

[(Me₂SiCp₂)Zr(Me)(THF)]⁺cation. In addition, the interactions

of both the R and â THF protons and the Cp protons with the

(3.1 Å) for the H-H distances.

Cp and PPh₃protons interact with the anion o-F and m-F nuclei

(Figures 7 and S3). In contrast, the specificity of interionic

interactions in 7 appears unaffected or affected to a lesser extent

by increased concentration (Figure 6). The average interionic

internuclear distances measured for complex 7 are summarized

in Table 4.

The apparent PGSE-derived volumes of complexes 7 and 8

exhibit dramatic concentration dependence. The results are

reported in Table 5 and graphically depicted in Figure 8a. It

can be seen that with increasing concentration, both 7 and 8

exhibit similar tendencies to forms ion-quadruples or even higher

aggregates in the case of 8, with the principal difference being

that 8, probably due to the presence of an increased number of

phenyl substituents, can be maintained at higher concentrations

in benzene-d₆.

-

o-F(medium) and m-F(medium-weak) nuclei of the MeB(C₆F₅)₃

counteranion are clearly visible in the 1D ¹H,¹⁹F HOESY

spectrum (Figure 6). The interaction intensities for the Cp

protons follow the order H1 > H2 . H3 and, in contrast to

complex 3, the strongest contact is no longer with H2 but with

H1.

1

The 1D H,¹⁹F HOESY spectrum of a 1.1 mM benzene-d₆

solution of complex 8 reveals the same interionic interactions

with almost the same intensity trends as observed for 7 above

(Figure 7). Increasing the concentration of 8 leads to a loss of

specificity in anion-cation interactions; the 1D ¹H,¹⁹F HOESY

spectrum of a 2.4 mM solution of 8 shows that all of the cation

9

J. AM. CHEM. SOC. VOL. 126, NO. 5, 2004 1455

A R T I C L E S

Zuccaccia et al.

1

Figure 6. Proton (bottom) and 1D H,¹⁹F HOESY spectra (middle: o-F

and m-F irradiations; top: o-F irradiation, 1.0 mM) of complex 7 (399.94

MHz, relaxation delay ) 2 s, mixing time ) 800 ms, 1.7 mM in benzene-

d₆, 298 K).

Figure 5. (a) Section of the ¹H NOESY spectrum (399.94 MHz, relaxation

delay ) 6 s, mixing time ) 150 ms, benzene-d₆, 298 K) of a 1.7 mM

solution of complex 7. As evident from the F2 traces, the apparent H1/H3

and H2/H4 cross-peaks are dispersion peaks due to subtraction imperfections.

(b) F2 trace (direct dimension) corresponding to the H1 resonance of the

2D spectrum. (c) 1D-GOESY experiment (H1 irradiation, relaxation delay

) 6 s, mixing time ) 150 ms, benzene-d₆, 298 K) of a ca. 0.6 mM solution

of complex 7.

Compound 9 has very low solubility in benzene-d₆at room

temperature. Its apparent volume (Figure 8b) in a saturated

solution (ca. 0.4 mM) is consistent with the presence of 1:1

ion-pairs. There is no evidence for aggregate formation. Higher

concentrations can be reached by heating the NMR tube

containing solid 9 and benzene-d₆to 60 °C. After approximately

30 min at 25 °C, complex 9 begins to separate from this

supersaturated solution as crystals. Even if a quantitative PGSE

investigation cannot be carried out under these conditions, a

1

Figure 7. Proton (bottom) and 1D H,¹⁹F HOESY spectra (middle: o-F

irradiation; top: o-F irradiation, 1.1 mM) of compound 8 (399.94 MHz,

relaxation delay ) 2 s, mixing time ) 800 ms, 2.4 mM in benzene-d₆, 298

K).

quickly executed experiment indicates the formation of ag-

gregates (N ) apparent aggregation number ) 1.7 at 1.5 mM).

9

1456 J. AM. CHEM. SOC. VOL. 126, NO. 5, 2004

Solution-Phase Interionic Structure and Aggregation

A R T I C L E S

Table 5. Diffusion Coefficients (D, 10^-10m²s^-1), Corrected Diffusion Coefficients^a(D*, 10^-10m²s^-1), Hydrodynamic Radii (r_H, Å),

Hydrodynamic Volumes (V_H, Å³), Factor c, and Aggregation Number (N) Values at Different Concentrations (Concentration, 10^-3M) for the

OSIP^a

concn

D

TPTS

D*

r_H

c

V_H

N

[(Me₂SiCp)₂Zr(Me)(THF)]⁺[MeB(C₆F₅)₃]^-(7)

2.20

1.70

1.60

1.20

0.99

0.90

0.70

0.09

3.10

2.35

1.40

1.05

0.91

0.86

0.45

0.12

1.48

0.82

0.38

0.64

0.42

0.24

0.12

5.20

3.30

2.00

0.97

0.55

0.11

5.17

5.76

4.87

5.76

6.01

5.93

6.20

6.74

4.11

4.29

4.93

5.06

4.96

5.36

5.80

5.85

5.15

5.88

6.37

5.04

5.38

5.64

5.73

4.50

4.22

4.58

4.85

5.02

5.15

7.45

7.98

6.67

7.55

7.59

7.42

7.36

7.71

7.20

7.21

7.53

7.44

7.18

7.67

7.50

7.65

7.74

7.70

7.86

7.20

7.19

7.28

7.30

7.35

6.79

7.00

7.06

7.01

5.76

6.00

6.07

6.34

6.57

6.65

7.00

7.27

4.74

4.95

5.44

5.66

5.73

5.81

6.43

6.36

5.53

6.34

6.73

5.81

6.22

6.43

6.53

5.09

5.17

5.44

5.76

5.91

6.11

6.19

5.99

5.94

5.73

5.57

5.52

5.30

5.15

7.30

7.03

6.50

6.29

6.22

6.15

5.67

5.72

6.40

5.73

5.47

6.15

5.82

5.66

5.60

6.87

6.78

6.49

6.19

6.07

5.90

5.4

5.3

5.2

5.6

5.5

5.4

5.3

5.5

5.3

5.2

5.4

5.3

5.5

5.4

993

900

878

788

724

705

624

572

1630

1455

1150

1042

1008

974

764

784

1098

788

686

974

826

760

736

1358

1306

1145

993

937

860

1.72

1.56

1.52

1.37

1.26

1.22

1.08

0.99

2.19

1.95

1.54

1.40

1.35

1.31

1.02

1.05

1.68

1.20

1.05

1.33

1.13

1.04

1.01

1.59

1.53

1.34

1.16

1.09

1.01

[(Me₂SiCp)₂Zr(Me)(PPh₃)]⁺[MeB(C₆F₅)₃]^-(8)

[(Me₂SiCp)₂Zr(Me)(THF)]⁺[B(C₆F₅)₄]^-(9)

[Me₂Si(Me₄Cp)(t-BuN)Zr(Me)(C₆D₆)]⁺[B(C₆F₅)₄]^-(10)

[Cp₂ZrMe)₂(µ-Me)]⁺[MePBB]^-(11)

^aThe numbers in bold are represented graphically in Figure 1. ^aThe corrected D* value correspond to a hypothetical measurement carried out at 22 °C,

in a solution containing the reported nominal concentration, but having the viscosity of pure C₆D₆(see Experimental Section for details).

Figure 8. PGSE-derived aggregation number (N) as a function of concentration for inner- and outer-sphere metallocenium complexes. The two arrows in

frame b indicate measurements that refer to a supersaturated solution. 0 [(Me₂SiCp₂)Zr(Me)(PPh₃)]⁺[MeB(C₆F₅)₃]^-(8); 9 [(Me₂SiCp₂)Zr(Me)(THF)]⁺[MeB(C₆F₅)₃]^-

(7); b [(1,2-Me₂Cp)₂ZrMe]⁺[MeB(C₆F₅)₃]^-(2); O [(Me₂SiCp₂)ZrMe]⁺[MeB(C₆F₅)₃]^-(3); 1 [(Me₂Si(Me₄Cp)(t-BuN)Zr(Me)(benzene-d₆)]⁺[B(C₆F₅)₄]^-(10);

3 [(Me₂SiCp₂)Zr(Me)(THF)]⁺[B(C₆F₅)₄]^-(9); × [(Me₅Cp)₂ThMe]⁺[B(C₆F₅)₄]^-(6); [ [(Cp₂ZrMe)₂(µ-Me)]⁺[MePBB]^-(11); ] [Me₂C(Fluorenyl)(Cp)-

ZrMe]⁺[FPBA]^-(4); 2 [rac-Et(Indenyl)₂ZrMe]⁺[FPBA]^-(5).

The low solubility of 9 in benzene-d₆precludes investigation

of the interionic structure, but taking advantage of the thermal

stability, the 1D ¹H,¹⁹F HOESY spectrum was recorded in

toluene-d₈at 50 °C (concentration ca. 0.8 mM, o-F irradiation).

The S/N ratio of this spectrum is not optimal, but as in complex

7, interactions are detected between the o-F nuclei and the THF

protons as well as with the H1 and H2 resonances. No

interaction is observed between the o-F nuclei and the Zr-Me

group.

and 8 (Figure 8b). In addition, the interionic structure of complex

10 was investigated in toluene-d₈by irradiating the o-F

resonance of the [B(C₆F₅)₄]^-counteranion. The corresponding

1D ¹H,¹⁹F HOESY spectrum is presented in Figure 9; the order

of the interaction strength is Me(2) > Me(1) . Me(3) ≈ t-Bu

. Me(A), while no contact is observed for the Zr-Me group.

Complex 11, in which the low coordinating ability of the

[MePBB]^-anion^10censures the formation of an outer sphere

ion-pair, forms aggregates at increasing concentrations, albeit

with a reduced tendency compared to 7, 8, and 10 (Figure 8c).

Strong contacts are observed between the single Cp resonance

of 11 and all the fluorine nuclei belonging to the disubstituted

C₆F₄ring (in particular, the intensity of such interactions

decreases in the order F6 ≈ F5 > F4 > F3 ≈ o-F). Complex

11 slowly decomposes in benzene-d₆at room temperature, and

1

Previous 1D H NMR results²⁰as well as our own HOESY

data (vide infra) indicate that complex 10 must be formulated

as a solvent adduct in solution. This compound is therefore

considered to be an outer sphere ion-pair similar to 7 and 8

with the aromatic solvent acting as a weak ligand. Accordingly,

its aggregation behavior is very similar to that of complexes 7

9

J. AM. CHEM. SOC. VOL. 126, NO. 5, 2004 1457

A R T I C L E S

Zuccaccia et al.

or anionic moiety, suggesting that this is a general phenomenon

for group 4 metallocenium inner sphere olefin polymerization

catalyst ion-pairs. All the data (NOE, PGSE, and cryoscopy¹⁶)

are consistent with a picture in which these systems exist in

solution as 1:1 ion pairs. It is also evident that this behavior

seems to be independent of concentration over the range 0.5-

1

20 mM.¹⁶In accordance with these observations, the H,¹⁹F

HOESY spectra of complexes 1, 3, and 5 indicate that a well-

defined relative orientation predominates in solution consistent

with the minimal tendency of ISIPs to aggregate.

In complexes 1 and 3, the anion binds to the cation through

a µ-Me group, and the phenyl rings are directed away from the

cationic metal center. As expected, if for complex 1 the same

relative cation-anion orientation is present in solution as found

in the solid-state crystal structure,⁴⁹only the B(C₆F₅)₃fluorine

nuclei at the ortho positions should be sufficiently proximate

to the cation to undergo significant dipolar relaxation with the

Cp protons, as is observed. Although a lack of NOE interaction

does not always indicate a large internuclear distance,⁵⁰two

lines of reasoning lead us to suggest that the spectra shown in

Figures 2 and 3 and in Supporting Information closely reflect

the average structures in solution. First, changes in the cation-

anion interionic orientation in other ion-pairs, also associated

with aggregation, lead to the observation of interionic cross-

peaks for the fluorine nuclei in the meta position (vide infra);

second, the quantitative analysis (initial rate approximation³⁶)

1

of the H NOESY and H,¹⁹F HOESY spectra in the case of

complex 1 agree quite well with the reported crystal structure

data (Table 2).

Figure 9. Proton (bottom) and 1D ¹H,¹⁹F HOESY spectra (o-F irradiation,

top) of compound 10 (399.94 MHz, relaxation delay ) 2 s, mixing time )

800 ms, 0.8 mM in toluene-d₈, 298 K).

The approach used for interionic distance estimation here is

based on the physically reasonable assumption that any internal

motion is more rapid than the overall correlation time of the

ion-pairs; in this case, the NOE-sensitive average internuclear

distances can be estimated from the X-ray single crystal

conformation in r^-3space.³⁶As can be seen in Tables 2 and 3,

in which the reported distances should be considered to have

uncertainties of ca. 10%,⁵¹this assumption appears to be valid

when a terminal CH₃group is considered. This is most likely

due to the fact that rapid terminal CH₃group internal rotation

dominates the local value of the correlation time for any

internuclear vector connecting this group with other groups.

after several days a mixture of 11 and 12 is obtained, with

concomitant methane formation (eq 7, see Supporting Informa-

tion).⁴⁸

Using one of these average distances (e.g., the r

in

Zr-Me/Cp

the case of complex 1 and the r _Zr-Me/H3in the case of complex

3) all the other distances involving the terminal CH₃groups

are nicely reproduced. On the other hand, this simple assumption

may be less valid when the average distance to be investigated

does not involve freely rotating CH₃groups. In the case of the

o-F/Cp average distance in 1, for example, neither the r^-6nor

the r^-3average for the single X-ray conformer accurately

reproduces the experimental data (Table 2).⁵²

The interionic structure of complex 5 is consistent, as

expected, with the solid-state structure of the analogous

compound [rac-Me₂Si(Indenyl)₂ZrMe]⁺[FPBA]^-,^10cin that the

FPBA^-counteranion pairs with the cation via a strong Zr-F-

Even if it is slow, this elimination reaction is surprisingly

clean at room temperature. In complex 12 the Cp protons

(around 5.9 ppm) exhibit the same contacts as in 11, but the

inspection of the ¹⁹F,¹H HOESY spectrum reveals that the

µ-methyl group now interacts with the fluorine nuclei in

positions F6 and F5, while no interactions are detectable for

the µ-CH₂group (see Supporting Information).

Discussion

Inner Sphere Ion-Pairs. None of the ISIPs examined in this

study exhibit detectable aggregation, irrespective of the cationic

(49) Guzei, I. A.; Stockland, R. A.; Jordan, R. F. Acta Crystallogr., Sect. C

2000, 56, 635.

(50) In general the extent of NOE enhancement is not dependent on the

internuclear distance alone. Factors such as the correlation time and the T₁

can prevent enhancement to an extent indistinguishable from background

noise.

(48) For similar decomposition pathways see: (a) Brownie, J. H.; Baird, M.

C.; Zakharov, L. N.; Rheingold, A. L. Organometallics 2003, 22, 33. (b)

Zhang, S.; Piers, W. E. Organometallics 2001, 20, 2088. (c) Bochmann,

M.; Cuenca, T.; Hardy, D. T. J. Organomet. Chem. 1994, 484, C10.

(51) For a discussion of the effects of internal motion on NOE-derived average

distances, see ref 36, Chapter 5.

9

1458 J. AM. CHEM. SOC. VOL. 126, NO. 5, 2004

Solution-Phase Interionic Structure and Aggregation

A R T I C L E S

Al interaction as indicated by a strong cross-peak with the

adjacent Zr-Me group (Figure 4; see Supporting Information

for additional details). However, the presence of two diastere-

omers with some overlapping signals in the ¹H spectrum makes

it difficult to describe in detail the network of the interionic

dipolar interactions. In addition to the Al-F group, the F6, F5,

and o-F* nuclei can interact with different protons on the metal

cation (the ¹⁹F resonances are assigned by means of standard

¹⁹F COSY and ¹⁹F NOESY experiments), and the general trend

indicates that the F6 interaction is, not surprisingly, the strongest

(Figure 4). Interestingly, there is a strong homonuclear NOE

interaction between the F6 and the o-F* atoms on the counter-

anion. This contact is not expected to come from within the

same biphenyl moiety as confirmed by a similar ¹⁹F NOESY

experiment on the simple [Ph₃C]⁺[FPBA]^-salt. Rather, this

interaction is likely indicative of π-stacking between one C₆F₄

ring and a C₆F₅ring of an adjacent biphenyl similar to that

observed in the solid state.^3e

unsaturated cations. These species, containing very weakly

coordinating counteranions, are likely to exist as solvent-

separated ion-pairs in solvents with relatively high permittivities

such as bromobenzene (ꢀ_r²⁹³) 5.45), chlorobenzene (ꢀ_r²⁹³

)

5.69), and methylene chloride (ꢀ_r²⁹⁸) 8.93). However, indirect

evidence and classical calculations based on the theory of

Fuoss⁵⁴indicate that these complexes should behave as intimate

ion-pairs in the relatively low-permittivity solvents typically used

in single-site polymerization reactions (i.e., benzene, toluene,

and saturated hydrocarbon solvents).⁵⁵Accordingly, recent work

by Landis,^12bBochmann,⁸Waymouth,⁵⁶and Busico,⁵as well

as recent results from our laboratory,^{2e,3b-d,10,19}conclusively

demonstrates that the ion-pairs are the effective propagation

species and that the anion cannot be considered as a mere

spectator during the enchainment process. In fact, in the classical

scenario of a two-step Cossee-type mechanism,¹⁵consisting of

a series of equilibria in which reversible alkene association is

followed by alkene insertion into the polymeryl σ-bond, the

present OSIPs plausibly model one component of this equilib-

rium and consequently, together with the zirconocenium-

polymeryl anion ISIP, the resting state of the catalyst. It has

already been shown that the first step in the formation of

catalytically active species in metallocene-mediated polymeri-

zation of simple olefins is likely to be a monomer association/

dissociation preequilibrium involving the electron-deficient

metallocenium center.⁵⁷On the other hand, nonchelated alkyl-

alkene cationic group 4 d⁰complexes have not been directly

observed so far. In addition, as asserted by Busico, “for a

monomer molecule to insert, it is assumed that the anion must

be partly displaced, but to where exactly is hard to say.”^5aWith

the aim of better understanding the cation-anion interplay after

generation of the putative catalytically active species, we applied

the combined NOE and PGSE techniques to the ion-pairs formed

via anion displacement by a Lewis base. Relatively strong Lewis

bases have been used in several instances to stabilize cationic

zirconocene complexes,^29,58while weaker Lewis bases have been

employed to study the equilibria and kinetics of anion displace-

ment reactions.²⁸The use of moderately strong Lewis bases (i.e.,

The results of these homo- and heteronuclear NOE investiga-

tions on inner sphere intimate ion-pairs such as complexes 1-5

1

are in agreement with previous investigations using simple H

and/or ¹⁹F NMR spectroscopy and arguing from chemical shift

displacement accompanying coordination. For example, the

chemical shift of the bridging fluorine is an excellent indicator

of M‚‚‚F-Al coordination,^10cwhile changes in the m-F vs p-F

chemical shift difference reflect coordination of the [RB(C₆F₅)₃]^-

anions.^18,29,53These data can be interpreted in a straightforward

way: the changes in chemical shifts reflect the strength of the

anion-cation coordinative interaction and are in general related

to an interplay of steric and electronic constraints at both cation

and anion.^2e,10c,12bIn addition, there are now a number of X-ray

diffraction studies for this class of compounds from which

detailed metrical parameters can be analyzed and compared.⁹

Theoretical calculations at the ab initio level indicate that the

cation-anion interaction in these systems is primarily electro-

static in nature,^11a,cbut the residual coordinative ability of the

anion is sufficient to enforce a localized anion/cation geometry.

It is therefore likely that the X-ray-derived solid-state structures

of ISIPs are a reasonable approximation of the solution-state

structures in low-polarity solvents. This conclusion is also in

good agreement with the “gas-phase” and solvated ground-state

geometries computed in theoretical studies.¹¹

(54) Fuoss, R. M. J. Am. Chem. Soc. 1958, 80, 5059.

(55) Computational studies (see refs 11a and 30) show that multiple geometries

are energetically accessible, and the cation-anion interactions in these kinds

of ion-pairs are poorly localized. On the other hand, it has been proposed

that NOE is sensitive in distinguishing between conformations differing

by only a few kJ/mol (see ref 52d).

Outer Sphere Ion-Pairs. Far more difficult is the solution

structural characterization of species in which the anions are

not coordinated to/strongly interacting with, the formally

(56) Wilmes, G. M.; Polse, J. L.; Waymouth, R. M. Macromolecules 2002, 35,

6766.

(57) (a) Dalmann, M.; Erker, G.; Bergander, K. J. Am. Chem. Soc. 2000, 122,

7986. (b) Karl, J.; Dalmann, M.; Erker, G.; Bergander, K. J. Am. Chem.

Soc. 1998, 120, 5643. (c) Galakhov, M. V.; Heinz, G.; Royo, P. Chem.

Commun. 1998, 1, 17. (d) Wu, Z.; Jordan, R. F.; Petersen, J. L. J. Am.

Chem. Soc. 1995, 117, 5867. (e) Casey, C. P.; Hallenbeck, S. L.; Pollok,

D. W.; Landis, C. R. J. Am. Chem. Soc. 1995, 117, 9770.

(52) Experimental methods to estimate the actual value of the local correlation

time rely on measurement of the dipolar contribution to the ¹³C relaxation

time (see, for example: (a) Gaemers, S.; van Slageren, J.; O’Connor, C.

M.; Vos, J. G.; Hage, R.; Elsevier, C. J. Organometallics 1999, 18, 5238.

(b) Bu¨hl, M.; Hopp, J.; von Philipsborn, W.; Beck, S.; Prosenc, M. H.;

Rief, U.; Brintzinger, H.-H Organometallics 1996, 15, 778. (c) Abragam,

A. The Principles of Nuclear Magnetism; Clarendon Press: Oxford, 1961)

or on estimation of the homo- or heteronuclear NOE response at different

temperatures (see, for example: (d) Macchioni, A.; Magistrato, A.; Orabona,

I.; Ruffo, F.; Ro¨thlisberger, U.; Zuccaccia C. New J. Chem. 2003, 27,

455. (e) Zuccaccia, C.; Bellachioma, G.; Cardaci, G.; Macchioni, A. J.

Am. Chem. Soc. 2001, 123, 11020). A more in-depth analysis could, in

principle, be achieved using the two-dimensional conformer population

analysis algorithm proposed by Landis and co-workers (see, for example:

(f) Casey, C. P.; Hallenbeck, S. L.; Wright, J. M.; Landis, C. R. J. Am.

Chem. Soc. 1997, 119, 9680. (g) Landis, C. R.; Luck, L.; Wright, J. M. J.

Magn. Reson., Ser. B 1995, 109, 44. (h) Landis, C.; Allured, V. S. J. Am.

Chem. Soc. 1991, 113, 9493), but this is beyond the scope of the present

work.

(58) (a) Carpentier, J.-F.; Maryin, V. P.; Lucy, J.; Jordan, R. F. J. Am. Chem.

Soc. 2001, 123, 898. (b) Ringelberg, S. N.; Meetsma, A.; Hessen, B.;

Teuben, J. H. J. Am. Chem. Soc. 1999, 121, 6082. (c) Witte, P. T.; Meetsma,

A.; Hessen, B.; Budzeelar, P. H. M. J. Am. Chem. Soc. 1997, 119, 10561.

(d) Alelyunas, Y. W.; Baezinger, N. C.; Bradley, P. K.; Jordan, R. F.

Organometallics 1994, 13, 148. (e) Alelyunas, Y. W.; Guo, Z.; LaPointe,

R. E.; Jordan, R. F. Organometallics 1993, 12, 544. (f) Eshuis, J. J. W.;

Tan, Y. Y.; Meetsma, A.; Teuben, J. H. Organometallics 1992, 11, 362.

(g) Burowsky, S. L.; Jordan, R. F.; Hinch, G. D. Organometallics 1991,

10, 1268. (h) Alelyunas, Y. W.; Jordan, R. F.; Echols, S. F.; Borkowsky,

S. L.; Bradley, P. K. Organometallics 1991, 10, 1406. (i) Jordan, R. F.;

Bradley, P. K.; Baenzinger, N. C.; LaPointe, R. E. J. Am. Chem. Soc. 1990,

112, 1289. (j) Eshuis, J. J. W.; Tan, Y. Y.; Teuben, J. H. J. Mol. Catal.

1990, 62, 277. (k) Jordan, R. F.; Guram, A. S. Organometallics 1990, 9,

2116. (l) Jordan, R. F.; LaPointe, R. E.; Bradley, P. K.; Baenziger, N.

Organometallics 1989, 8, 2892. (m) Taube, R.; Krukowa, L. J. Organomet.

Chem. 1988, 347, C9. (n) Jordan, R. F.; Bajgur, C. S.; Dasher, W. E.

Organometallics 1987, 6, 1041. (o) Jordan, R. F.; Bajgur, C. S.; Willett,

R.; Scott, B. J. Am. Chem. Soc. 1986, 108, 7410.

(53) (a) Hayes, P. G.; Welch, G. C.; Emslie, D. J. H.; Noack, C. L.; Piers, W.

E.; Parvez, M. Organometallics 2003, 22, 1577. (b) Horton, A. D.; de With,

J. Organometallics 1997, 16, 5424.

9

J. AM. CHEM. SOC. VOL. 126, NO. 5, 2004 1459

A R T I C L E S

Zuccaccia et al.

THF, PPh₃) is preferred for the present model complex

investigation, in that the anion displacement equilibrium lies

far to the right (toward the Lewis base adduct) and affords stable

compounds suitable for long-duration NMR experiments such

as ¹⁹F ,¹H HOESY. In fact, formation of ion-pairs 7 and 8

proceeds quickly and cleanly after addition of one equivalent

of THF or PPh₃, respectively, to complex 3 to afford new ion-

pairs in which the anion is relegated to the second coordination

sphere. Brintzinger and co-workers have shown that reaction

of some ion-pairs with various Lewis bases proceeds with large

negative values of ∆S°, providing convincing indirect evidence

that the anion remains associated with the cation in benzene

solution.²⁸The present observation of interionic dipolar interac-

tions in the ¹H NOESY and ¹⁹F,¹H HOESY spectra of 7 and 8

directly indicate that intimate ion-pairs are formed, the solution

interionic structures of which are investigated here for the first

time.

in Table 4. The average distance between the B-Me group on

the anion and the H1 protons on the cation is estimated to be

4.1 Å, while the average distance between the same cation nuclei

and the o-F fluorine on the anion is estimated to be 4.5 Å. Thus,

in contrast to parent compound 3 (Table 3), the anion B-Me

and the o-F groups in the Lewis base complexes are ca. 1.1

and ca. 0.4 Å more distant, respectively, from the closest cation

proton (H1 in 7 and H2 in 3).

In the case of complex 7 (1.7 mM), the dipolar interionic

contact between the B-Me group and cation H1 protons is

detectable in the ¹H NOESY spectrum (Figure 5).⁵⁹In addition,

specific cation-anion interactions were also observed in the

corresponding ¹⁹F,¹H HOESY spectrum (Figure 6). Taken

together, these data prove directly that the displacement of

the anion by the THF generates an intimate OSIP in which

the anion is preferentially localized on the THF side of the

cation, shifted slightly toward the Me(A) group and further away

from the Zr-Me group as indicated by the increased interaction

with H1. The absence of interactions between the B-Me group

and R⁶⁰and â THF and Me(A) protons seems to indicate a

favored anion orientation in which the B-Me moiety points away

from the metal center. In contrast to precursor complex 3,

interionic contacts in complex 7 are detected not only for the

o-F, but also for the m-F fluorine nuclei (Figure 6). Interestingly,

the specificity in the interionic interactions is not affected

by an increase in concentration (from 1.0 to 1.7 mM) even

though PGSE data indicate formation of aggregates (vide infra).

In an effort to determine if the B-Me/H1 contact could be

due to aggregation phenomena, ¹H NOESY and a 1D GOESY³²

(H1 irradiation, Figure 5c) experiments were attempted at ca.

0.6 mM, where PGSE measurements indicate the presence of

mainly 1:1 ion-pairs (N ≈ 1). Unfortunately, the S/N ratio is

insufficient to conclude unambiguously that the observed B-Me/

H1 cross-peak arises from aggregation.⁶¹In light of the parallel

PGSE results it was decided to limit the quantitative analysis

to compound 7 and to continue to investigate all the other

systems from a semiquantitative point of view (i.e., the spectra

were not rigorously recorded in the initial rate approximation

regime).

These data strongly suggest that in complex 7 the anion is

energetically freer to assume various orientations with respect

to the cation, as expected in the absence of localized coordinative

interactions. On the other hand, the relative anion-cation posi-

tion is well defined, with the anion localized primarily on the

side of the cation to which THF is coordinated and shifted

toward the backside of the metallocenium cation. Computation-

ally optimized interionic structures for analogous metalloce-

nium-olefin adducts (namely, [Cp₂ZrMe(C₂H₄)]⁺[MeB(C₆F₅)₃]^-)

are in excellent qualitative agreement with the experimental data

reported here.⁶²In addition, using the data reported in Table 4

and the Zr-H1 distance (3 Å) from the X-ray structure of

complex 9 (see below), we can roughly estimate, in the most

extreme situation of a linear B-H1-Zr alignment, that the

Zr-B distance is ca. 7.2-7.3 Å. The corresponding DFT-

derived Zr-B distance calculated for the compound [Cp₂ZrMe-

(C₂H₄)]⁺[MeB(C₆F₅)₃]^-is 6.7 Å,⁶²while the corresponding ab

initio-derived distance is 7.4 Å in the related complex [H₂Si-

(C₅H₄)(t-BuN)Ti(Me)(C₂H₄)]⁺[MeB(C₆F₅)₃]^-.^11a

In contrast to complex 7, the interionic structure of complex

8 at 2.4 mM is substantially less localized than that at 1.1 mM

(Figure 7). This behavior can be understood by the results of

detailed concentration-dependent PGSE NMR investigations for

both ion-pairs 8 (0.1-3.1 mM) and 7 (0.1-2.2 mM) in that a

concentration dependence of the observed interionic interactions

might be a consequence of aggregation. As noted previously in

the literature for similar compounds (and in the present

Experimental Section),^18,20,29OSIPs 7 and 8, synthesized in situ,

begin to separate as finely dispersed oily phases at higher

concentrations, generating new sets of resonances with different

chemical shifts (probably due to the changes in the local

diamagnetic susceptibility in the new nonmonodisperse phase⁶³).

The highest concentrations used in this work, determined by

integration with respect to an internal standard, refer to the

1

Quantitative H-¹H (Figure 5) and H,¹⁹F (see Supporting

Information) analyses for a 1.7 mM solution of 7, using an

approach similar to that discussed above, afford the data reported

(59) Note that these NOESY spectra were recorded with a 6 s relaxation delay,

150 ms mixing time, and 32 scans per increment. If the parameters routinely

used for small molecule NOESY experiments (namely 1s relaxation delay,

800 ms mixing time, and 16 scans per increment) are employed, the B-Me/

H1 cross-peak is hardly visible. This is likely a consequence of the rapid

B-Me relaxation time in combination with the expectation of very little

NOE enhancement.

(60) The cross-peak corresponding to the F2 R resonance and to the F1 B-Me

resonance visible in Figure 5 is most probably an artifact in that the

corresponding mirror image peak across the diagonal is not observed.

(61) The 2D-NOESY experiment (Figure 5, a and b) and the 1D-GOESY

experiment (Figure 5c) were acquired in the initial rate approximation.

(62) Nifant’ev, I. E.; Ustynyuk, L. Y.; Laikov, D. N. Organometallics 2001,

20, 5375.

(63) Preliminary NMR results obtained for [(1,2-Me₂Cp)₂ZrMe(PPh₃)]⁺-

[MeB(C₆F₅)₃]^-indicate a benzene:metallocenium ion-pair mole ratio of

ca. 20:1 within the oily phase that separates completely from benzene

solution.

9

1460 J. AM. CHEM. SOC. VOL. 126, NO. 5, 2004

Solution-Phase Interionic Structure and Aggregation

A R T I C L E S

important for ISIPs,¹⁶and we now extend this work to OSIPs.

The results reported here for a wide-spectrum ion-pairs support

the following general conclusion: ISIPs do not aggregate to a

detectable extent in low-polarity hydrocarbon solvents, whereas

OSIPs can aggregate significantly at concentrations above

approximately 5 × 10^-4M. Furthermore, the extent to which

OSIPs aggregate is a function of the cation ligand framework

as well as the structure of the counteranion.

The behavior of complex 9 deserves additional comments.

The X-ray analysis in the solid-state (see below) and the

solution-phase NMR data are consistent with a formulation of

compound 9 as discrete cationic and anionic fragments (an outer

sphere ion-pair). Unfortunately, 9 is sparingly soluble (saturated

solution ca. 0.4 mM), and consequently, its aggregation tendency

cannot be fully explored. Some indication of aggregation can

be observed when an excess of 9 is heated in refluxing benzene-

d₆for about 1.0 min (Figure 8c). In this case, a metastable

supersaturated solution can be prepared, but as expected, crystals

form over a period of a few hours, rendering the measurement

of the diffusion coefficient necessarily imprecise (i.e., the actual

concentration cannot be defined, and signal intensities decrease

because of the concomitant precipitation). As for the interionic

structure, the results obtained for complex 9 are very similar to

those for complex 7: the anion is localized on the side of the

cation closer to the H1 protons. The low intensity of the signals

Figure 10. Proposed average solution structures of ion-pair (a) and ion-

quadruple (b) for the outer sphere ion-pairs 7 and 8. These structures are

consistent with both the NOE and PGSE data. The asterisks indicate the

carbon atoms to which the silicon bridge is connected.

portion of the ion-pairs still homogeneously dissolved in the

benzene-d₆solution and must be considered to be the concentra-

tion of a saturated solution.

As can be seen from Figure 8a, complexes 7 and 8 exhibit a

markedly higher tendency to aggregate than parent complex 3.

It is quite surprising that the interionic structure of complex 7

retains a high degree of localization while the corresponding

aggregation number indicates a substantial presence of ion-

quadruples at 1.7 mM concentration (N ) 1.56, Table 5). A

possible explanation for this behavior is that the average solution

structure of the ion-quadruple formed by 7 at higher concentra-

tions is the one depicted in Figure 10.⁶⁴By reasonably assuming

that the ion-quadruple structures of 7 and 8 are the same, we

suggest that the observed lower level of specificity in the anion-

cation interaction for a 2.4 mM solution of 8 (N ) 1.95, Table

5) must be attributed to the partial formation of ion-hextuples

that cannot form, starting from ion-quadruples, without leading

to a loss of structural localization.⁶⁵The fact that complexes 7

and 8 exhibit a completely different solution-structural behavior

than previously investigated ion-pairs 1-6 does suggest that

there is a pervasive modification of properties on transition from

inner- to outer sphere metallocenium ion pairs. To further test

this hypothesis, concentration-dependent PGSE measurements

were extended to compounds 2, 3, 4, and 5 (ISIPs) and to

complexes 9, 10, and 11 (OSIPs).

1

in the 1D H,¹⁹F HOESY spectrum recorded at 50 °C (o-F

irradiation, see Supporting Information) suggests weaker ion-

pairing in 9 vs that in parent compound 7, but this could also

be due to reduced cross-relaxation efficiency due to a decrease

in the correlation time.

Complex 10 exhibits a behavior very similar to that of

complexes 7 and 8 although with somewhat greater tendency

for aggregation. Interestingly, 10 is reasonably stable in toluene-

d₈but appears to decompose more rapidly in benzene-d₆at

concentrations greater than ca. 0.6 mM. The 1D ¹H,¹⁹F HOESY

spectrum (o-F irradiation, Figure 9) was recorded at 0.8 mM in

toluene-d₈to take advantage of the higher sample stability.⁶⁷

In accordance with the formulation of this compound and in

accordance with the aforementioned results on complexes 7

and 8, the interionic structure of 10 exhibits pronounced

localization, with the anion predominantly residing on the

side of the cation opposite to the Zr-Me, as confirmed by

observation of a small interaction with the Me(A) group and

the absence of any interaction with the Zr-Me moiety. The

observation that Me(2) interacts with the anion more strongly

than the t-Bu group, and the lack of the Zr-Me interaction,

are also in agreement with the presence of a solvent molecule

in the formally vacant coordination site. A structure in which

the [B(C₆F₅)₄]^-anion is in the second coordination sphere of

the Zr center is also consistent when these data are compared

with the ¹⁹F,¹H HOESY data for the Th-containing ion-pair

6. For the latter, the ¹⁹F,¹H HOESY results confirm the

interaction between the anion and the Th-Me group, as expected

Beck et al. initially communicated evidence that aggregation

may be an important consideration in group 4 metallocenium

ion-pairs.⁶⁶We have recently shown that aggregation is not

(64) That structures similar to that sketched in Figure 10 for ion-quadruples are

in fact accessible is supported by the crystal-structure packing in similar

compounds: (a) Gibson, V. C.; Humphries, M. J.; Tellmann, K. P.; Wass,

D. F.; White, A. J. P.; Williams, D. J. Chem. Commun. 2001, 2252. (b)

Chernega, A.; Cook, J.; Green, M. L.; Labella, L.; Simpson, S. J.; Souter,

J.; Stephens, A. H. H. J. Chem. Soc., Dalton Trans. 1997, 3225. (c)

Amorose, D. M.; Lee, R. A.; Petersen, J. L. Organometallics 1991, 10,

2191.

(65) Another interpretation that cannot be ruled out is that the average solution

interionic structures in ion-quadruples formed by complexes 7 and 8 are

substantially different. In addition, it could be argued that the observed

lower level of localization arises from an increased rate of site epimeriza-

tion that exchanges the magnetically nonequivalent protons (H1/H4,

H2/H3, Me(A)/Me(B)). We are confident that the observed nonspecific

interactions reflect the actual quasi-static structure in solution for two

principal reasons: (1) Reported rates for site epimerization in complexes

similar to complex 8 are on the order of 0.1 s^-1, (i.e., too slow to satisfy

the condition of fast exchange on the relaxation time scale (see ref

36) necessary to produce the spectrum reported in Figure 7). (2) In

(66) Beck, S.; Geyer, A.; Brintzinger, H.-H. Chem. Commun. 1999, 24,

2477.

(67) The higher stability of compound 10 in toluene could be due in principle

to either (1) a higher intrinsic stability due to coordination of toluene instead

of benzene (see, for example: Hayes, P. G.; Piers, W. E.; Paervez, M. J.

Am. Chem. Soc. 2003, 125, 5622) or (2) a lower level of aggregation that

impedes possible bimolecular decomposition pathways (see, for example:

Li, L.; Hung, M.; Xue, Z. J. Am. Chem. Soc. 1995, 117, 12746) or, more

likely, (3) a combination of 1 and 2.

1

accordance with point 1, H NOESY spectra of complex 8 recorded with

the same mixing time do not exhibit evidence for additional NOE

enhancement (H1 with H3 and H2 with H4) due to H1/H4 and H2/H3

exchanges.

9

J. AM. CHEM. SOC. VOL. 126, NO. 5, 2004 1461

A R T I C L E S

Zuccaccia et al.

Figure 11. X-ray-derived stick representations of the [FPBA]^-and [MePBB]^-anions, taken fron refs 10c and 10d. (a) “Coordinated” [FPBA]^-viewed

along the F-Al bond in complex [(Me₂Si(Me₄Cp)(t-BuN)Zr(Me)]⁺[FPBA]^-, (b) “Free” [FPBA]^-viewed along the F-Al bond in compound [Ph₃C]⁺[FPBA]^-.

(c) [MePBB]^-viewed along the CH₃-B bond in compound [((1,2-Me₂Cp)₂ZrMe)₂(µ-Me)]⁺[MePBB]^-. The cationic moieties have been omitted for the

purpose of clarity.

if the solid-state structure is primarily retained in solution, adding

evidence for the formulation of 6 as an inner sphere ion-pair.⁶⁸

nonequivalent o-F atoms interacts strongly with the B-Me group

but not with the cation ligand protons, while the other o-F atom

is the only one on the C₆F₅fragment that exhibits interionic

interaction with the cation. The additional observation that the

relative strength of interionic interactions between Cp protons

and the anion C₆F₄fluorine nuclei follows the order F6 ≈ F5

> F4 > F3 conclusively demonstrates that in complex 11 the

B-Me vector of the [MePBB]^-anion is directed away from the

metal center in a manner similar to that found in the solid state

for the analogous [((1,2-Me₂Cp)₂ZrMe)₂(µ-Me)]⁺[MePBB]^-

complex.^10c

The observation of a different interionic structure on passing

from 11 to 12 reflects the large changes in the electronic charge

distribution on the cationic moiety and simultaneously illustrates

that the HOESY experiment is very sensitive to detecting these

changes. In fact, the anion is localized on the µ-CH₃side of

the metallocene distal to the µ-CH₂group, and as in the case of

11, the B-Me vector of the [MePBB]^-anion is directed away

from the metal center. Note that in the case of 12 or related

complexes, even with X-ray structural data in hand, it might

be difficult to determine the exact location of the anion with

respect to the cation due to the known difficulties in resolving

disorders frequently associated with similar bridging ligands.^48a,b

Crystal Structure of [(Me₂SiCp₂)Zr(Me)(THF)]⁺[B(C₆F₅)₄]^-

(9). In complex 9, the coordination environment around the Zr

atom consists of the silicon-bridged bis-η⁵-cyclopentadienyl

ligand, a methyl ligand, and an η¹-oxygen-bound THF ligand,

affording the expected pseudotetrahedral metallocenium geom-

etry at Zr (Figure 12). The Zr-CH₃distance is 2.259(5) Å, the

Zr-O distance is 2.210(3) Å, and the CH₃-Zr-O angle is

96.8°. While the Zr-CH₃distance and the CH₃-Zr-O angle

are very similar to the corresponding values observed for

[Cp₂ZrMe-

The interionic structure of binuclear complex 11, due to the

simplicity of the metal cation, allows us to focus on the weak

coordinative properties of the [MePBB]^-anion.^2e,10cThe anion

is localized on the side of the cation proximate to the Cp ligands

as demonstrated by the absence of any hydrogen-fluorine

interaction with either the µ-methyl or the terminal methyl

groups. In addition, proton-proton interactions are not observed

1

in the H NOESY spectrum. To elucidate the relative anion

orientation with respect to the cation, the structure of the

[MePBB]^-anion must be taken into account. Unlike [FPBA]^-,

[MePBB]^-does not coordinate to the Zr center,⁶⁹and after

methide abstraction by the PBB Lewis acid cocatalyst from the

Zr dimethyl precursor, the geometry of the anion is similar to

that of the “free” [FPBA]^-anion with the C₆F₅rings of the

biphenyl groups shielding the B-Me group (Figure 11). Con-

sequently, the B-Me group cannot interact with the cationic

protons, even if directed toward them. However, the relative

cation-anion orientations can be probed in solution by means

of ¹⁹F,¹H HOESY spectroscopy.⁷⁰In fact, one of the two

(THF)]⁺[BPh₄]^-by Jordan and co-workers (2.256(10) Å and

97.4°),^58othe Zr-O bond distance and the THF arrangement

are distinctly different. That is, the Zr-O bond distance is ca.

0.1 Å longer (2.210(3) Å vs (2.122(14) Å), and the torsion

angles for the two C_methyl-Zr-O_THF-C_R,THFlinkages are 36.3°

and -153.7°, which places the THF molecule roughly 31.3°

(68) Another possibility for 6 is to assume it forms an OSIP in which there is

a lower degree of specificity in the anion-cation interaction. While this

hypothesis is difficult to substantiate due to the NMR equivalence of the

Cp-methyl resonances, it seems very unlikely in the context of the behavior

of all the other ion-pairs investigated in this work.

(69) The same non-coordinative characteristics are observed in the case of

[FPBB]^-(FPBB ) tris(2,2′,2′′-nonafluorobiphenyl)fluoroborate, (unpub-

lished results from this laboratory), suggesting that the central atom (B

versus Al) plays a fundamental role.

(70) This should be true also in the case of “free” [FPBA]^-, but the rapid

relaxation of the Al-F resonance precludes further investigations by ¹⁹F

NOESY experiments.

9

1462 J. AM. CHEM. SOC. VOL. 126, NO. 5, 2004

Solution-Phase Interionic Structure and Aggregation

A R T I C L E S

out of the plane described by atoms C_methyl, Zr, and O_THF. This

result is in contrast to the nearly perpendicular arrangement

(77.7°) reported for [Cp₂ZrMe(THF)]⁺[BPh₄]^-.^58oHowever,

THF orientations similar to those in 9 are observed in

[(Me₅Cp)₂Zr(CH₂SiMe₃)(THF)]⁺[BPh₄]^-by Petersen and co-

workers^64cand for [(Me₅Cp)₂TiMe(THF)]⁺[BPh₄]^-by Boch-

mann and co-workers.⁷¹A rationale for the different THF

orientations in these compounds vs that in [Cp₂ZrMe(THF)]⁺-

[BPh₄]^-has been presented.^64cThe steric hindrance exerted by

the pentamethyl-substituted cyclopentadienyl rings prevents the

THF from achieving the perpendicular orientation necessary for

good overlap between the O π-donor orbital primarily of p

character and the vacant Cp₂Zr hybridized d orbital of a₁

symmetry,⁷²resulting in a weaker interaction and longer Zr-O

distance. In the present case, the Zr-O distance of 2.210(3) Å

is closer to that of [(Me₅Cp)₂Zr(CH₂SiMe₃)(THF)]⁺[BPh₄]^-

(2.243(3) Å).^64cThe same steric argument appears sufficient to

explain the THF orientation/bonding found in 9. On one hand,

complex 9 presents an even more open coordination sphere

(Cp_centroid-Zr-Cp_centroidangle ) 126.54°) at the Zr vs

[Cp₂ZrMe(THF)]⁺[BPh₄]^-(Cp_centroid-Zr-Cp_centroidangle )

129.59°), suggesting there is more space to accommodate the

THF ligand in the perpendicular orientation. On the other hand,

the bridging Me₂Si group prevents free Cp ring rotation, and

the minimal steric hindrance provided by the hydrogens on the

Cp ring â position (H2 and H10) cannot be released until the

THF is rotated at least partially toward the electronically less

favorable “parallel” orientation.

Figure 12. (a) ORTEP drawing of the cationic part of complex

[(Me₂SiCp₂)Zr(Me)(THF)]⁺[B(C₆F₅)₄]^-‚C₆H₆(9) with thermal ellipsoids

plot (50% probability for all non-hydrogen atoms). (b) PLUTO drawing

showing the position of the nearest-neighbor anions and solvent molecules

to a given cation in the crystal of 9. The number reported next to each

anion, represented here by the B(C_ipso

) unit, is the Zr-B distance in Å.

4

Hydrogen atoms are omitted for clarity. Both drawings were created using

the ORTEP-3 for Windows; Farrugia, L. J. J. Appl. Crystallogr. 1997, 30,

565.

in the solid state of solvent molecules occupying the solution

anion-preferred face of the cation suggests another possible,

intriguing explanation for the observed weak NOE intensity

arising from cation-anion interactions in solution: it is possible

that the solvent preference for this face of the cation persists in

solution, i.e., the solution ion-pair is solvent-separated. Cor-

relation between the average interionic solution structure and

the most stable ion-pairing found in the solid state is non-trivial,

requiring daunting calculations of electrostatic and the “steric”

contributions to the total energy for each ionic pair found in

the solid state.⁷³

Regarding preferential ion-pairing in the solid state, of the

three nearest-neighbor anions of a given cation in the crystal of

9, none are positioned such that they would produce the

observed NOE interactions in solution, where the preferred anion

position is proximate (syn) to the THF ligand (see above). In

particular, the B atom of the nearest anion (Zr-B ) 7.021 Å)

is located midway between the CH₃and the THF ligands,

displaced ca. 26° from the plane described by C_methyl, Zr, O_THF

(Figure 12). The B atom of the second-nearest anion (Zr-B )

7.254 Å) is positioned approximately anti to the THF ligand

and lies ca. 35.7° out of the same plane but in the opposite

direction. In addition, solvent molecules (benzene) closest to a

given Zr atom are positioned approximately anti to the CH₃

ligand. While it will require further investigations, the presence

Conclusions

The results reported here show, for the first time, that

application of complementary NOE and PGSE methodologies

(73) (a) Gruet, K.; Clot, E.; Eisenstein, O.; Lee, D. H.; Patel, B.; Macchioni,

A.; Crabtree, R. H. New J. Chem. 2003, 27, 80. (b) Macchioni, A.;

Zuccaccia, C.; Clot, E.; Gruet, K.; Crabtree, R. H. Organometallics 2001,

20, 2367. (c) Macchioni, A.; Bellachioma, G.; Cardaci, G.; Cruciani, G.;

Foresti, E.; Sabatino, P.; Zuccaccia, C. Organometallics 1998, 17, 5549.

(71) Bochmann, M.; Jagger, A. J.; Wilson, L. M.; Hursthouse, M. B.; Motevalli,

M. Polyhedron 1989, 8, 1838.

(72) Lauher, J. W.; Hoffmann, R. J. Am. Chem. Soc. 1976, 98, 1729.

9

J. AM. CHEM. SOC. VOL. 126, NO. 5, 2004 1463

A R T I C L E S

Zuccaccia et al.

to group 4 metallocenium polymerization catalyst ion-pairs

affords unambiguous information concerning solution-phase

interionic structure and aggregation. Both features are found to

be dependent on whether the counteranion is in the inner or

outer coordination sphere. In the case of ISIPs, the derived

interionic solution structures are in excellent agreement with

those in the solid state, and there is no detectable ion-pair

aggregation. In other words, the residual coordinative character

of the cation-anion interaction is sufficently strong to enforce

a relatively low degree of ion-pair mobility, and thus the solid-

state, structurally frozen regime represents a good approximation

of the relative anion-cation orientation in solution. The same

residual coordinative interaction also tends to reduce the

separated ion character of these systems, which accounts for

the minimal aggregation observed in benzene solution, even at

fairly high concentrations (10-20 mM).

In the case of outer sphere ion-pairs, the anion is localized

proximate to the cation in a position different from that in the

ISIPs, while the lack of any residual coordinative interaction

allows the anion to explore a much greater range of orientations.

In particular, in the cases examined, the anion is preferentially

localized on the side of the cation closer to the coordinated

Lewis base, slightly shifted toward the backside of the metal-

locenium cation, and farther away from the metal-methyl

group. The absence of any apparent residual coordinative

interaction also accounts for: (1) the observation that the

diffraction-derived solid-state arrangements are unlikely to be

descriptive of the solution structure in that both lattice and

solvation energies must play a significant role in determining

the lowest-energy configuration, (2) the expected higher polarity

of these species dramatically increases their tendency to form

aggregates higher than 1:1 ion-pairs.⁷⁴Noteworthy here also

are the results of quantitative NOE investigations in solution,

which are in excellent agreement with the computationally

optimized structures^11a,62for similar compounds in which the

Lewis base consists of an ethylene molecule. This result is

encouraging in that it implies that substitution of the olefinic

substrate by a different Lewis base, incapable of undergoing

subsequent insertion, affords a good working model at least as

far as solution-phase interionic structure is concerned. Further

direct experimental information on cation-anion structural

interactions should be accessible and should be extendable to

systems in which anion effects have already been established.

Another important issue that is currently attracting much

catalytic mechanistic interest is whether zirconocenium ion-pair

aggregates are relevant in olefin polymerization enchainment

processes.⁸The results reported here indicate that the tendency

to form aggregates is dramatically increased when the coun-

teranion is displaced from the first coordination sphere but also

confirm that only at relatively elevated concentrations (>0.5

mM) does the concentration of ion-quadruples (ca. 10%) become

appreciable in nonpolar aromatic solutions. Hence, the increased

tendency to form aggregates observed for ion-pairs where the

counteranion is in the second coordination sphere seems thus

far insufficient to consider ion-quadruples relevant under typical

catalytic conditions.

Acknowledgment. This research was supported by the U.S.

Department of Energy (DE-FG 02-86 ER13511) and by the

MIUR (Rome, Italy, Programma di Rilevante Interesse Nazio-

nale, Cofinanziamento 2002-2003). We thank Dr. D. Zuccaccia

for helpful discussions. M.-C.C. and C.Z. thank the Dow

Chemical Co. and the Italian CNR, respectively, for postdoctoral

research fellowships.

Supporting Information Available: 1D HOESY pulse se-

quence schematic; NMR spectra of compounds 1, 3, 5, 6, 7, 8,

9, 11, and 12; and representative plots of PGSE data (PDF).

Crystallographic data of compound 9 in CIF format. This

material is available free of charge via the Internet at

http://pubs.acs.org.

JA0387296

(75) When the motion is more rapid than molecular tumbling, the effective

distance “sensed by the NOE” is r_effectiveg (1/N ∑^N_µ)1r^-_IS³_,µ^1/3. On the other

)

hand, when the motion is slower than overall molecular tumbling, the

corresponding effective distance is: r_effectiveg (1/N ∑^N_µ)1r_I^-_S⁶_,µ

)

with the

1/6

(74) Further investigations of aggregation as a function of temperature are

required to better discriminate between predominantly enthalpically or

entropically driven processes. For example, it has been reported that the

enthalpy of aggregation of tetrabutylammonium chloride in chloroform is

very small and that aggregation is a response to crowding in solution rather

than to a favorable change in enthalpy (see ref 45b).

index µ indicating the different conformations assumed by the spin system.

See: (a) Yip, P. F.; Case, D. A.; Hoch, J. C.; Poulsen, F. M.; Redfield, C.,

Eds. Computational Aspect of the Study of Biological Macromolecules by

Nuclear Magnetic Resonance Spectroscopy; Plenum Press: New York,

1991, pp 317-330. (b) Tropp, J. J. Chem. Phys. 1980, 72, 6035.

(76) Macura, S.; Ernst, R. R. Mol. Phys. 1980, 41, 95.

9

1464 J. AM. CHEM. SOC. VOL. 126, NO. 5, 2004

Article Doi

NOE and PGSE NMR Spectroscopic Studies of Solution Structure and Aggregation in Metallocenium Ion-Pairs

DOI: 10.1021/ja0387296

Source and publish data:

Authors:

Article abstract of DOI:10.1021/ja0387296

Full text of DOI:10.1021/ja0387296

Products guided by the article

R&D Labs maybe for 671821-21-9

Relevant to this article

Hot Product