Displacement of H3CB(C6F5)3- anions from zirconocene methyl cations by neutral ligand molecules: Equilibria, kinetics, and mechanisms

Article abstract of DOI:10.1021/om010671l

The displacement of the MeB(C₆F₅)₃^- anion from seven different zirconocene methyl cations by neutral Lewis bases, such as dimethylaniline, benzyldimethylamine, and diⁿbutyl ether, was investigated by 1D and 2D NMR spectroscopy. Equilibrium constants for reactions with diⁿbutyl ether change by factors of less than 5 between the zirconocene contact ion pairs studied, despite substantial steric differences. Rate constants of this displacement reaction, however, change by a factor of more than 10⁵ between Me₂Si(C₅H₄)₂ZrMe^-MeB(C ₆F₅)₃^-, the most open complex, and rac-Me₂Si(2-Me-BzInd)₂ZrMe^-MeB(C₆F ₅)₃^-, the most highly substituted species studied. Kinetic and stereochemical data indicate that Lewis base-anion exchange proceeds by way of an associative mechanism, which occurs without side change of the zirconium-bound methyl group. DFT calculations support an associative substitution mechanism and propose five-coordinated reaction intermediates with the Lewis base coordinated to the central coordination site.

Full text of DOI:10.1021/om010671l

Organometallics 2002, 21, 473-483
473
-
Disp la cem en t of H3CB(C6F 5)3 An ion s fr om Zir con ocen e
Meth yl Ca tion s by Neu tr a l Liga n d Molecu les: Equ ilibr ia ,
Kin etics, a n d Mech a n ism s
Frank Schaper, Armin Geyer, and Hans H. Brintzinger*
Fachbereich Chemie, Universita¨t Konstanz, D-78457 Konstanz, Germany
Received J uly 26, 2001
The displacement of the MeB(C6F5)3- anion from seven different zirconocene methyl cations
by neutral Lewis bases, such as dimethylaniline, benzyldimethylamine, and dinbutyl ether,
was investigated by 1D and 2D NMR spectroscopy. Equilibrium constants for reactions with
dinbutyl ether change by factors of less than 5 between the zirconocene contact ion pairs
studied, despite substantial steric differences. Rate constants of this displacement reaction,
-
however, change by a factor of more than 105 between Me2Si(C5H4)2ZrMe+MeB(C6F5)3 , the
most “open” complex, and rac-Me2Si(2-Me-BzInd)2ZrMe+MeB(C6F5)3 , the most highly
-
substituted species studied. Kinetic and stereochemical data indicate that Lewis base-anion
exchange proceeds by way of an associative mechanism, which occurs without side change
of the zirconium-bound methyl group. DFT calculations support an associative substitution
mechanism and propose five-coordinated reaction intermediates with the Lewis base
coordinated to the central coordination site.
In tr od u ction
In a recent study on the side exchange of borate
anions in ansa-zirconocene methyl cations, we have
obtained evidence that this process occurs through
associative mechanisms, probably via ion quadruples or
higher ion aggregates present in these systems,9 and
that anion-free, solvated zirconocene methyl cations are
unlikely to be relevant intermediates for this exchange
process. Formation of an olefin-containing reaction
complex might thus likewise occur by way of an associa-
tive displacement of the anion rather than via olefin
coordination to a solvated cation.
The reaction paths through which active species arise
when zirconocene precatalyst complexes are combined
with suitable activators are still incompletely under-
stood. Particularly unsatisfactory, despite recent advan-
ces,1-3 is our understanding of the elementary reaction
steps induced by methylalumoxane (MAO) as an activa-
tor. For reaction systems generated from zirconocene
dimethyl complexes by perfluorophenyl borane or borate
activators,4,5 well-defined ion pairs have been shown to
contain metallocene alkyl cations in contact with
H3CB(C6F5)3-, B(C6F5)4-, or other related anions.6
Even here it is not clear, however, which course of
events generates the catalyst proper: Dissociation of the
weakly coordinating anion is frequently assumed to
release a zirconocene alkyl cation,7 which would then
enter into a catalytic cycle by repeated olefin uptake and
insertion. The active catalyst is seen here as a solvated
alkyl zirconocene cation, while the nature of the anion
determines only the actual concentration of this active
species. In line with this view, molecular modeling
studies of olefin insertion, chain termination, and iso-
merization mechanisms have mainly dealt with naked
cationic intermediates in the gas phase.8
A number of recent density functional or ab initio
calculations have dealt with reactions between ethene
and alkylmetallocene contact ion pairs.10-14 So far,
however, a consistent picture has emerged neither for
(7) J ordan, R. F. Adv. Organomet. Chem. 1991, 32, 325. Bochmann,
M. J . Chem. Soc., Dalton Trans. 1996, 255.
(8) Cf.: (a) Kawamura-Kuribayashi, H.; Koga, N.; Morokuma, K.
J . Am. Chem. Soc. 1992, 114, 8687. (b) Woo, T. K.; Ziegler, T.
Organometallics 1994, 13, 432. (c) Meier, R. J .; van Dormaele, G. H.
J .; Iarlori, S.; Buda, F. J . Am. Chem. Soc. 1994, 116, 7274. (d) Prosenc,
M.-H.; Schaper, F.; Brintzinger, H. H. In Metalorganic Catalysts for
Synthesis and Polymerization; Kaminsky, W., Eds.; Springer-Verlag:
Berlin, 1999; p 223. For recent reviews see: (e) Resconi, L.; Cavallo,
L.; Fait, A.; Piemontesi, F. Chem. Rev. 2000, 100, 1253. (f) Rappe´, A.
K.; Skiff, W. M.; Casewit, C. J . Chem. Rev. 2000, 100, 1435.
(9) (a) Beck, S.; Lieber, S.; Schaper, F.; Geyer, A.; Brintzinger, H.
H. J . Am. Chem. Soc. 2001, 123, 1483. (b) Beck, S.; Geyer, A.;
Brintzinger, H. H. J . Chem. Soc., Chem. Commun. 1999, 2477.
(10) (a) Fusco, R.; Longo, L.; Masi, F.; Garbassi, F. Macromol. Rapid
Comm. 1997, 18, 433. (b) Fusco, R.; Longo, L.; Proto, A.; Masi, F.;
Garbassi, F. Macromol. Rapid Commun. 1998, 19, 257.
(11) Bernardi, F.; Bottoni, A.; Miscione, G. P. Organometallics 1998,
17, 16.
* Corresponding author. E-mail: Hans.Brintzinger@uni-konstanz.de.
(1) Babushkin, D. E.; Semikolenova, N. V.; Zakharov, V. A.; Talsi,
E. P. Macromol. Chem. Phys. 2000, 201, 558.
(2) Wieser, U.; Brintzinger, H. H. In Organometallic Catalysts and
Olefin Polymerization; Blom, R., Arild, F., Rytter, E., Tilset, M.,
Ystenes, M., Eds.; Springer: Heidelberg 2001; p 3.
(3) Coevoet, D.; Cramail, H.; Deffieux, A. Macromol. Chem. Phys.
1998, 199, 1451.
(12) Lanza, G.; Fragala`, I. L.; Marks, T. J . J . Am. Chem. Soc. 1998,
120, 8257.
(4) Yang, X.; Stern, C. C.; Marks, T. J . J . Am. Chem. Soc. 1991,
113, 3623. Ewen, J . A.; Elder, M. J . Makromol. Chem., Macromol.
Symp. 1993, 66, 179.
(5) Turner, H. W.; Hlatky, G. G. Eur. Pat. Appl. 0,277,003, 1988.
Hlatky, G. G.; Upton, D. J .; Turner, H. W. (Exxon Chemical Co.) PCT
Int. Appl. 91/09882, 1991. Ewen, J . A.; Elder, M. J . Eur. Patent Appl.
0,427,697, 1991.
(13) (a) Chan, M. S. W.; Ziegler, T. Organometallics 2000, 19, 5182.
(b) Chan, M. S. W.; Vanka, K.; Pye, C. C.; Ziegler, T. Organometallics
1999, 18, 4624. (c) Vanka, K.; Chan, M. S. W.; Pye, C. C.; Ziegler, T.
Organometallics 2000, 19, 1841. (d) Vanka, K.; Ziegler, T. Organome-
tallics 2001, 20, 905.
(14) Nifant’ev, I. E.; Ustynyuk, L. Y.; Laikov, D. N. In Organome-
tallic Catalysts and Olefin Polymerization; Blom, R., Arild, F., Rytter,
E., Tilset, M., Ystenes, M., Eds.; Springer: Heidelberg, 2001; p 72.
(6) Chen, E. Y.-X.; Marks, T. J . Chem. Rev. 2000, 100, 1391.
10.1021/om010671l CCC: $22.00 © 2002 American Chemical Society
Publication on Web 01/05/2002

474 Organometallics, Vol. 21, No. 3, 2002
Schaper et al.
Sch em e 1
the course of this reaction, i.e., the position of olefin
attack, nor for its energetics.12-14
rane. Thermodynamics and mechanisms of these ligand
substitutions were characterized for various zirconocene
complexes (1a -7a , Scheme 1) with the aim to extrapo-
late useful information concerning the uptake of olefin
substrates by these ion pairs.
Direct evidence in this regard is hard to come by:
Even for zirconocene cations carrying intramolecularly
tethered olefin ligands,15-17 no observations on anion-
olefin exchange mechanisms have been reported so far.
Intermolecular olefin-anion exchange in zirconocene
alkyl cations, on the other hand, is even less likely to
become directly observable, due to the high reactivity
expected for olefin-containing zirconocene alkyl cations.
We have thus set out to study the characteristics of this
type of reaction by using an inert Lewis base instead of
an olefin substrate.
Strong Lewis bases have been used in several in-
stances to stabilize cationic zirconocene complexes.18-21
Our aim in this study, however, was to find more weakly
coordinating Lewis bases, which would allow us to
determine equilibria and kinetics of anion displacement
reactions for a series of different zirconocene complexes.
It has been reported, for example, that dimethylaniline
(DMA) is partially coordinated to zirconocene methyl
cations when dimethylanilinium borate salts are reacted
with zirconocene dimethyl complexes.22 We have thus
chosen to study the reactions of weakly binding ligand
molecules such as DMA with zirconocene contact ion
pairs, generated in situ from the appropriate dimethyl
complex by reaction with tris(pentafluorophenyl) bo-
Exp er im en ta l P a r t
(C5H5)2ZrCl2,23 Me4C2(C5H4)2ZrCl2,24 Me2Si(C5H4)2ZrCl2,25
(C9H7)2ZrCl2,26 rac-Me2Si(C9H6)2ZrCl2,27 and B(C6F5)3 were
28
synthesized as described in the literature. rac-Me2Si(2-
Me-C9H5)2ZrCl2 and rac-Me2Si(2-Me-C13H7)2ZrCl2 were do-
nated by BASF AG, Ludwigshafen. The dimethyl complexes
were obtained by reaction of the corresponding dichloride
complexes with methylmagnesium chloride as described in ref
29. All operations were carried out using drybox or Schlenk
techniques. Solvents were distilled from sodium; Lewis bases
were distilled from sodium or passed through a column of
molecular sieve (4 Å). NMR solvents were dried and stored
over molecular sieves under nitrogen atmosphere. 1D NMR
spectra were recorded at 250 MHz in benzene-d6 at 300 K with
C6D5H as internal standard (δ ) 7.15 ppm). 2D NMR spectra
were recorded at 600 MHz. The temperatures in NMR experi-
ments were determined using methanol or ethylene glycol
standards as chemical thermometers.
Gen er a l P r oced u r e for th e NMR Exp er im en ts. Stock
solutions of the corresponding dimethyl complexes, of B(C6F5)3,
and of the respective Lewis base in C6D6 were prepared and
stored over molecular sieves (4 Å) under nitrogen atmosphere.
Borane (1.05-1.1 equiv) was added to the zirconocene dimethyl
complex and the slightly yellow solution diluted to the desired
concentration, in most cases 2 mM. The desired amount of
Lewis base was added. In general coordination to the metal
(15) (a) Wu, Z.; J ordan, R. F.; Petersen, J . L. J . Am. Chem. Soc.
1995, 117, 5867. (b) Carpentier, J .-F.; Wu, Z.; Lee, C. W.; Stro¨mberg,
S.; Christopher, J . N.; J ordan, R. F. J . Am. Chem. Soc. 2000, 122, 7750.
(16) Galakhov, M. V.; Heinz, G.; Royo, P. J . Chem. Soc., Chem.
Commun. 1998, 17.
(17) Casey, C. P.; Carpenetti, D. W., II; Sakurai, H. J . Am. Chem.
Soc. 1999, 121, 9483. Casey, C. P.; Carpenetti, D. W., II. Organome-
tallics 2000, 19, 3970.
(18) (a) J ordan, R. F.; Bajgur, C. S.; Willett, R.; Scott, B. J . Am.
Chem. Soc. 1986, 108, 7410. (b) J ordan, R. F.; LaPointe, R. E.; Bradley,
P. K.; Baenziger, N. Organometallics 1989, 8, 2892. (c) Burowsky, S.
L.; J ordan, R. F.; Hinch, G. D. Organometallics 1991, 10, 1268. (d)
J ordan, R. F.; Bradley, P. K.; Baenzinger, N. C.; LaPointe, R. E. J .
Am. Chem. Soc. 1990, 112, 1289. (e) J ordan, R. F.; Bajgur, C. S.;
Dasher, W. E. Organometallics 1987, 6, 1041. (f) Alelyunas, Y. W.; Guo,
Z.; LaPointe, R. E.; J ordan, R. F. Organometallics 1993, 12, 544. (g)
J ordan, R. F.; Guram, A. S. Organometallics 1990, 9, 2116. (h)
Alelyunas, Y. W.; J ordan, R. F.; Echols, S. F.; Borkowsky, S. L.;
Bradley, P. K. Organometallics 1991, 10, 1406.
(19) Taube, R.; Krukowa, L. J . Organomet. Chem. 1988, 347, C9.
(20) (a) Eshuis, J . J . W.; Tan, Y. Y.; Teuben, J . H. J . Mol. Catal.
1990, 62, 277. (b) Eshuis, J . J . W.; Tan, Y. Y.; Meetsma, A.; Teuben, J .
H. Organometallics 1992, 11, 362. (c) Ringelberg, S. N.; Meetsma, A.;
Hessen, B.; Teuben, J . H. J . Am. Chem. Soc. 1999, 121, 6082.
(21) Beck, S.; Prosenc, M.-H.; Brintzinger, H. H. J . Mol. Catal. A:
Chem. 1998, 128, 41.
1
center was accompanied by decolorization of the solution. H
NMR data of the Lewis-base adducts obtained are given in
the Supporting Information.
Com p u ta tion a l Deta ils. Semiempirical calculations were
carried out with the program Spartan 4.1.30 For DFT calcula-
(23) Heyn, B.; Hippler, B.; Kreisel, G.; Schreer, H.; Walther, D.
Anorganische Synthesechemie-ein Integriertes Praktikum; Springer-
Verlag: Berlin, 1986; p 84.
(24) Schwemlein, H.; Brintzinger, H. H. J . Organomet. Chem. 1983,
254, 69. Hu¨sgen, N. Dissertation, University of Konstanz, C. Mu¨ller:
Konstanz, 1998.
(25) Klouras, N.; Ko¨pf, H. Monatsh. Chem. 1981, 112, 887. Bajgur,
C. S.; Tikannen, W. R.; Petersen, J . L. Inorg. Chem. 1985, 24, 2539.
(26) Samuel, E.; Rausch, M. D. J . Organomet. Chem. 1965, 4, 156.
Rausch, M. D.; Moriarty, K. J .; Atwood, J . L.; Hunter, W. J . Organomet.
Chem. 1987, 327, 39.
(27) Diamond, G. M.; Rodewald, S.; J ordan, R. F. Organometallics
1995, 14, 5.
(28) Pohlmann, J . L. W.; Brinckmann, F. E.; Tesi, G.; Donadio, R.
E. Z. Naturforsch. B 1965, 20b, 1. Pohlmann, J . L. W.; Brinckmann Z.
Naturforsch. B 1965, 20b, 5.
(22) Horton, A. D.; Orpen, A. G. Organometallics 1991, 10, 3910.
Hlatky, G. G.; Eckman, R. R.; Turner, H. W. Organometallics 1992,
11, 1413.
(29) Beck, S. Dissertation, University of Konstanz, Hartung-Gorre-
Verlag: Konstanz, 1998.
(30) Spartan 4.1; Wavefunction Inc.: Irvine, 1996.

-
Displacement of H3CB(C6F5)3 Anions
Organometallics, Vol. 21, No. 3, 2002 475
tions, TURBOMOLE31 was used for the calculation of energies
and gradients and OPTIMIZE32 for geometry relaxation. In
all cases the gradient-corrected nonlocal functionals of Becke33
and Perdew34 (BP86) were used. These had been found to yield
reaction energies for ethene insertion into a Ti-Me bond close
to those obtained by CCSD(T) methods.35 For structure
determinations, SVP (split valence with polarization) basis sets
were used, while final energies were obtained with TZVP
(triple-ú split valence with polarization) basis sets for all atoms
except zirconium,36 to which the ecp-28-mwbopt basis set was
applied in both cases.37 Reaction energies obtained with a
TZVP basis are more positive by 0-15 kJ /mol than those
Sch em e 2
-
obtained with a SVP basis in most cases; only for B(C6F5)4
Ta ble 1. Equ ilibr iu m Con sta n ts a n d
anions were reaction energies more negative by approximately
10 kJ /mol with the bigger basis set (see Supporting Informa-
tion). Corresponding auxiliary basis sets were used for the J ij
Coulombic term.38 For MeB(CF3)3- anions all stationary points
were checked by frequency calculations for the absence of
negative eigenvalues and the presence of only one for transi-
tion states. For both anions, MeB(C6F5)3- or B(C6F5)4-, geom-
etry optimizations were carried out until the internal gradient
was well below 3 × 10-4 au. Reaction energies in toluene
solution were calculated from single-point energies using the
conductor-like screening model (COSMO)39 for the gas-phase
Th er m od yn a m ic P a r a m eter for th e Disp la cem en t
-
of MeB(C6F 5)3 fr om Com p lexes 1a -3a by DMAa
contact ion
pair
K
∆Hø
∆Sø
b
[L‚mol-1
]
[kJ ‚mol-1
]
[J ‚mol-1‚K-1
]
1a
2a
3a
70
160
270
-53(3)c
-148(9)c
-47(4)d
-120(13)d
been reported for anion displacement from (C5H5)2-
Zr(CH3)+‚‚‚H3CB(C6F5)3 by trialkylphosphines.21
-
optimized geometries, with ꢀ ) 2.4 CV-1 m-1, rH ) 1.3, rC
)
[ZrMe(Lb) +A-]
[ZrMe+ A-]‚[Lb]
2.0, rB ) 1.08, rF ) 1.72, rO ) 1.72, rSi ) 2.52, and rZr ) 2.28
Å. The use of gas-phase-optimized geometries, instead of
geometries optimized by use of COSMO, was shown for several
examples to have only minor effects (1-6 kJ /mol) on the
absolute energies.
K )
(1)
ZrMe+A- ) 1a -7a , Lb ) Lewis base
Equilibrium constants and thermodynamic param-
eters of these reactions are shown in Table 1. Values of
∆SQ ) -150 and -120 J ‚(mol‚K)-1 support the view that
the anion remains boundspresumably electrostaticallys
to the DMA-containing cation in 1b-3b, even though
it is no longer coordinated to the Zr center. The ring-
bridged DMA complexes 2b and 3b have somewhat
higher stabilities than the parent complex 1b. This
appears to be due to entropic contributions, as indicated
by the less negative values of both ∆HQ and ∆SQ for
the formation of 2b. Since cyclopentadienyl rotation is
equally quenched in 2a as in 2b by the Me4C2 bridging
unit, the ∆SQ value of -120 J ‚(mol‚K)-1 for the forma-
tion of 2b is likely to reflect just the associative
character of this process. Rotation of the cyclopentadi-
enyl rings, on the other hand, is probably more strongly
hindered in 1b than in 1a ; this appears to contribute
additional -30 J ‚(mol‚K)-1 to ∆SQ for the formation of
1b.
For the indenyl complexes 4a -7a no reaction was
observed even with a 10-20-fold excess of DMA. To
investigate the effects of bridging, substitution, and
annelation of the cyclopentadienyl rings on these anion
displacement reactions, we have searched for Lewis
bases other than DMA, which allow a comparison of all
the zirconocene complexes in the series 1a -7a . These
contact ion pairs were thus reacted with a number of
aliphatic and aromatic, cyclic and noncyclic ethers,
thioethers, phosphines, and amines.
Resu lts a n d Discu ssion
1. An ion Disp la cem en t Equ ilibr ia . In a series of
NMR experiments, dimethylaniline was reacted with
the contact ion pairs 1a -7a , which were obtained by
addition of 1.1 equiv of B(C6F5)3 to the appropriate
zirconocene dimethyl complex in 1-10 mM solutions in
benzene-d6.40 The bridged and unbridged cyclopentadi-
enyl complexes 1a -3a react partially with DMA to yield
the DMA adducts 1b-3b. At constant [DMA]:[Zr] ratios,
formation of the DMA adducts is favored at higher
zirconocene concentrations. This concentration depen-
dence, which is in accord with the equilibrium relation
represented in eq 1, indicates that the displaced anion
remains associated with the DMA-containing cation in
the form of some outer-sphere complex in the nonpolar
benzene solutions and that formation of separated,
solvated ions does not occur to any significant extent
(Scheme 2).41 The formation of associated ion pairs,
rather than separately solvated ions, has previously
(31) Eichhorn, K.; Treutler, O.; O¨ hm, H.; Ha¨ser, M.; Ahlrichs, R.
Chem. Phys. Lett. 1995, 242, 652.
(32) Baker, J . OPTIMIZE, PQS, Fayetteville.
(33) Becke, A. D. J . Chem. Phys. 1986, 84, 4524. Becke, A. D. J .
Chem. Phys. 1988, 88, 1053. Becke, A. D. Phys. Rev. A 1988, 38, 3098.
(34) Perdew, J . P. Phys. Rev. B 1986, 33, 8822. Perdew, J . P. Phys.
Rev. B 1986, 34, 7406.
(35) J ensen, V. R.; Børve, K. J . J . Comput. Chem. 1998, 19, 947.
(36) Scha¨fer, A.; Horn, H.; Ahlrichs, R. J . Chem. Phys. 1992, 97,
2571.
(37) Ahlrichs, R.; Ba¨r, M.; Ha¨ser, M.; Horn, H.; Ko¨lmel, C. Chem.
Phys. Lett. 1989, 162, 165. Ha¨ser, M.; Ahlrichs, R. J . Comput. Chem.
1989, 10, 104.
(38) Eichkorn, K.; Weigand, F.; Treutler, O.; Ahlrichs, R. Theor.
Chem. Acc. 1997, 97, 119.
(39) Klamt, A.; Schu¨rmann, G. J . Chem. Soc., Perkin Trans. 1993,
2, 799.
(40) (a) A slight excess of borane was used to avoid the formation of
dinuclear species. (b) Beck, S.; Prosenc, M.-H.; Brintzinger, H. H.;
Goretzki, R.; Herfert, N.; Fink, G. J . Mol. Catal. 1996, 111, 67, and
references therein.
(41) Slight changes of the equilibrium constants according to eq 1
with the overall zirconocene concentration, indicated by the experi-
mental data, would require an exponent for [ZrMe+A-] greater than
that for [ZrMe(Lb)+A-] by ca. 0.2. This is in line with the experimen-
tally observed increased tendency of outer-sphere complexes to form
ion quadruples or higher aggregates (see refs 9b, 29). Since these effects
are small compared to the experimental error and are not distinguish-
able from ionic strength effects, no quantitative analysis was per-
formed.

476 Organometallics, Vol. 21, No. 3, 2002
Ta ble 2. Rea ction s of th e Con ta ct Ion P a ir s 1a -7a w ith Differ en t Lew is Ba sesa
Schaper et al.
Lewis base
(C6H5)3PdCH2
reactions with 1a -3a b
reactions with 4a -7a b
complete coordination, unidentified
side products, stable Lewis base
boran adductc
(CH2)5CdCH2
NMe3
no coordinationd
complete coordination
complete coordination
Me2N(CH2C6H5)
NMe2(SiMe3)
complete, fast coordination
partial coordination, K≈30-100 L‚mol-1
unidentified side products
partial coordination, CH-activation of
NMe as a side reaction
nearly complete, slow coordination
,
NMe2(C6H5)
no coordinationd
1,8-diazabicyclo[5.4.0]undec-7-ene
(DBU)
2-Me-1-pyrroline
2,6-lutidine
pyridazine
(C2H5)2O
complete coordination, some side products
partial coordination, overlapping peaks
prevent quantitative evaluation
complete coordination
complete coordination
fast CH-activation reactione
complete coordination
complete coordination
partial coordination
fast CH-activation reactione
(C4H9)2O
partial coordination
(CH2)5O
(CH2)3(CH)2O
complete coordination
broadened peaks prevent quantitative
evaluation
complete coordination
broadened peaks prevent quantitative
evaluation
(CH2)2(CH)2O
broadened peaks prevent quantitative
evaluation
(CH)4O
no coordinationd
P(OMe3)
unidentified side products
complete coordination
complete coordination
very low solubility and broadened peaks
prevent quantitative evaluationf
no coordinationd
Me2P(C6H5)
P(CH2C6H5)3
(CH2)4S
partial coordination
partial coordination
(CH)4S
Typical conditions: in benzene-d6 at 300 K, [Zr] ) 1-5 mM, [Zr]:[Lewis base] ) 2:1-1:20. bNot all combinations were investigated
a
(see Supporting Information). Cf. Do¨ring, S.; Erker, G.; Fro¨hlich, R.; Meyer, O.; Bergander, K. Organometallics 1998, 17, 2183. dAt 20-
fold excess of Lewis base. Cf. ref 18g. Cf. ref 20a.
c
e
f
Ta ble 3. Estim a tes for Equ ilibr iu m Con sta n ts for
th e Coor d in a tion of Differ en t Lew is Ba ses to
Com p lexes 1a -7a
Ta ble 4. Equ ilibr iu m a n d Ra te Con sta n ts for th e
Disp la cem en t of MeB(C6F 5)3 fr om Com p lexes 1a
-
a n d 3a -7a by Dibu tyl Eth er
K ) [ZrMe(Lb)+A-]/
1a a
3a a
4a b
5a c
6a b
7a b
Lewis base
[ZrMe+A-]‚[Lb]
K [L‚mol-1
]
2800
-20
270
60
4900
-21
4000
53
2000
-19
2
1500
-18
8
800
-17
0.1
1100
-17
0.03
82
d
PMe3, NMe3, Et2O, C4H8S, C4H8O
PMe2(C6H5), PBn3, Me2NBn
nBu2O
K . 1 mM-1
K > 1 mM-1
K ≈ 1 mM-1
K < 1 mM-1
K , 1 mM-1
∆GL
300 Κ
k [L‚mol-1‚s-1
]
d,e
∆Gq
72
68
79f
300 Κ
Me2N(C6H5), Me2N(SiMe3)
a
At 300 K in benzene-d6 solution. k determined from 2D EXSY
C4H4O, C4H4S, olefina
spectra. bk determined from conventional 1H spectra. k estimated
c
a
Estimated from DFT calculations.
from rate constants at higher and lower temperatures. dIn kJ /mol.
eDetermined from the Eyring equation ∆Gq ) -RT ln(hk/(cQkBT)).
f∆Hq ) 36(2) kJ /mol, ∆Sq ) -142(7) J /(mol‚K).
For most of the Lewis bases studied, determination
of equilibrium constants for the whole range of com-
plexes 1a -7a was prevented either by complete lack of
reactivity or by fast, quantitative reactions, accompa-
nied in some cases by side reactions (Table 2). Partial
coordination is achieved by increasing the steric bulk
of the Lewis base. Substituted phosphines such as
tribenzylphosphine or dimethylphenylphosphine gave
the desired equilibrium with the indenyl complexes
4a -7a , but coordinated completely to the unsubstituted
cyclopentadienyl complexes 1a -3a .
Qualitative estimates of coordination constants for
different Lewis bases vis a` vis zirconocene contact ion
pairs 1a -7a are given in Table 3. As expected, the
coordination ability decreases with increasing steric
bulk and decreasing basicity of the Lewis base. Of the
Lewis bases studied here, only dinbutyl ether (DBE) was
found to react with cyclopentadienyl complexes 1a -3a
as well as with the indenyl complexes 4a -7a under
partial formation of complexes which are reasonably
stable against side reactions at room temperature. This
ligand was therefore chosen for a comparative study of
anion displacement from complexes 1a -7a .
The unsubstituted dimethylsilyl-bridged complex 3a
has the highest tendency to coordinate this Lewis base,
followed by the parent zirconocene 1a (Table 4, Figure
1). Annelation of an aromatic ring reduces the coordina-
tion constant by a factor of 2-3. Bridging of the indenyl
ligands has surprisingly small effects on the stability
of the DBE adducts 4c and 5c. A methyl group in the
2-position of the indenyl ring diminishes the coordina-
tion constant by a factor of ca. 2, while annelation of
another benzene ring has only minimal effects on the
equilibrium constant. In the series 1a -7a , equilibrium
constants differ only by factors less than 5, correspond-
ing to ∆∆GQ e 4 kJ /mol, despite substantial steric
differences among these complexes.
2. Kin etics of th e An ion Disp la cem en t Rea ction .
Reaction of benzyldimethylamine (BDMA) with the
bridged 2-methylindenyl complexes 6a and 7a was
essentially quantitative but slow enough to be moni-
tored via standard 1H NMR spectroscopy. Rate con-
stants were determined under pseudo-first-order con-
ditions with a relatively large excess of the amine.
Alternatively, the amine was used in approximately

-
Displacement of H3CB(C6F5)3 Anions
Organometallics, Vol. 21, No. 3, 2002 477
Ta ble 6. Ra te Con sta n ts for th e Disp la cem en t of
-
MeB(C6F 5)3 by DMAa
contact ion pair
Zr:DMA k [L‚mol-1‚s-1
]
Me4C2(C5H4)2ZrMe+‚MeB(C6F5)3-, 2a b
Me2Si(C5H4)2ZrMe+‚MeB(C6F5)3-, 3a c
Me2Si(C5H4)2ZrMe+‚MeB(C6F5)3-, 3a c
1:1
1:1
1:7
30 ( 6
21 ( 4
22 ( 3
a
Determined from 2D-EXSY spectra in toluene-d8 at 0 °C with
an overall zirconocene concentration of 1 mmol/L. Errors are
determined from exchange rates observed for different proton
c
pairs. bMixing time 500 or 800 ms. Mixing time 500 ms.
cyclopentadienyl protons as well as for the Zr- and Si-
bound methyl groups (Figure 2). In this manner, pseudo-
first-order rate constants were obtained by the method
described by Perrin and Dwyer (see Supporting Infor-
mation).45 The reaction rate was found to depend
linearly on the concentrations of zirconocene and of
Lewis base (Table 6), in agreement with the rate law
found for the anion substitution by benzyldimethyl-
amine. Similar rate constants of 30 and 22 L/(mol‚s) are
found for the bridged cyclopentadienyl complexes 2a and
3a , respectively. Due to the lack of reaction of DMA with
complexes 4a -7a , determination of anion displacement
kinetics was not possible for these indenyl complexes.
Displacement of the anion by dinbutyl ether (DBE)
from complexes 1a and 3a was studied by 2D NMR
spectroscopy at 300 K. Rate constants in benzene-d6
were determined from 2D NOESY spectra and analyzed
according to the rate law in eq 2 (Table 4). For the
indenyl complexes 4a , 6a , and 7a , anion-base exchange
was too slow to be measured by 2D exchange spectros-
copy. In these cases it was possible, however, to monitor
the reactions with DBE by conventional one-dimen-
sional NMR spectroscopy at 300 K. Again, a first-order
dependence on the concentrations of the zirconocene
complex and of the Lewis base was found in all cases.
The reaction of DBE with 5a at room temperature was
too slow for 2D NMR spectroscopy, but too fast for
conventional 1D NMR measurements. To estimate the
rate constant of this reaction at 300 K, the exchange
rate at 320 K was determined via EXSY spectroscopy
(partial decomposition of the complex already occurred
at this temperature) and at lower temperatures (240-
250 K) in toluene solution by use of 1D NMR techniques.
The rate constants at 300 K (estimated in the case of
5a ) are displayed in Figure 1 and Table 4. Annelation
of a benzene ring to 3a decreases the rate constant for
complex 5a by a factor of about 500. Introduction of a
methyl group in the 2-position slows the reaction down
by another 2 orders of magnitude. The anion displace-
ment reaction thus appears to be particularly sensitive
toward substituents in R-position of a bridged cyclopen-
tadienyl ring, more so than toward substituents in the
â-position. Further annelation of a benzene ring to the
indenyl moiety has only minor effects on the reaction
rates of 7a vs 6a , in accordance with the results
obtained with benzyldimethylamine as Lewis base. The
Me2Si-bridged complexes 3a and 5a react faster by
approximately 1 order of magnitude than their un-
bridged analogues 1a and 4a . This is most likely due to
the more open ligand framework in the bridged com-
plexes and possibly also to a hindrance of ring rotation
in the unbridged complexes upon formation of a crowded
F igu r e 1. Equilibrium (0,9) and rate constants (O,b) for
the substitution of MeB(C6F5)3 by dibutyl ether (DBE)
from unbridged complexes 1a and 4a (open symbols) and
from Me2Si-bridged complexes 3a , 5a -7a (filled symbols).
-
Ta ble 5. Ra te Con sta n ts a n d Activa tion
P a r a m eter s for th e Disp la cem en t of MeB(C6F 5)3
-
fr om 6a a n d 7a by BDMA in Ben zen e-d 6
k
∆Hq
∆Sq
complex
T [K]
[L‚(mol‚s)-1
]
[kJ ‚mol-1
]
[J ‚mol-1‚K-1
]
6a
6a
6a
7a
7a
7a
300
328
340
300
327
348
0.094
1.2
2.1
0.059
0.065
2.9
64(4)
67(2)
-50(11)
-43(7)
stoichiometric amounts, and the resulting concentration
profiles were analyzed with a second-order rate law. In
either case, the data obtained are in excellent agreement
with the rate law expressed in eq 2, i.e., with a reaction
first-order in the concentration of zirconocene as well
as of Lewis base.
v ) k[Zr][BDMA]
(2)
Annelation of a further ring to the indenyl moiety
causes only minor changes of reaction rates from 6a to
7a (Table 5). Complexes 1a -4a , on the other hand,
reacted with approximately stoichiometric amounts of
BDMA quantitatively in less than 5 min at room tem-
perature.42 Determination of rate constants for these
reactions was thus not possible under the conditions
used.43
With DMA as a Lewis base, displacement of the
methylborate anion from the cyclopentadienyl com-
plexes 2a and 3a was observable by means of 2D
NOESY NMR spectra at 0 °C in toluene-d8 solution.44
Exchange signals were detected and integrated for the
(42) No coordination of BDMA was observed for the sterically
crowded permethyl complex (C5Me5)2ZrMe+MeB(C6F5)3
.
-
(43) Complexes 1a -7a and especially their base adducts tend to
form aggregated species at elevated zirconocene concentrations (see
ref 29). To avoid this in our kinetic experiments, zirconocene concen-
trations were limited to ca. 1-4 mmol/L at room temperature. Cooling
was found to enhance the formation of the more aggregated forms,
and at concentrations low enough to avoid this problem it was not
possible to integrate NMR spectra with reasonable accuracy.
(44) DMA adducts 1b-3b undergo CH activation of N-methyl groups
within a period of 0.5-2 h at room temperature to yield a mixture of
two products, probably either with the anion or the nitrogen donor
coordinated to the metal. Similar CH activations have been reported
by J ordan and co-workers for other Lewis bases (see ref 18 g). To avoid
these side reactions, 2D NMR spectra were recorded at 0 °C. At this
temperature, a mixing time of 500 µs was necessary to observe the
anion displacement.
(45) Perrin, C. L.; Dwyer, T. J . Chem. Rev. 1990, 90, 935.

478 Organometallics, Vol. 21, No. 3, 2002
Schaper et al.
F igu r e 3. Exchange signals for anion exchange between
3a and 3b and for side migration of the Zr-methyl group
in 3a in 2D NOESY spectra of a reaction mixture as
described in Figure 3.
F igu r e 2. Exchange processes observed in 2D NOESY
spectra for anion-DMA exchange in a reaction mixture
containing 3a , 3b, and dimethylaniline (DMA).
change were made, when a reaction mixture of 3a , DBE,
and 3c was studied by 2D NMR. This stereoselectivity
provides further clues with regard to the mechanism of
the anion displacement reaction, to be discussed below.
transition state. Activation parameters of ∆Hq ) 36(2)
kJ /mol and ∆Sq ) -142(7) J /(mol‚K) were determined
for the reaction of 6a with DBE in the temperature
range 300-340 K. The significantly higher activation
entropy compared to the reaction with BDMA indicates
that entropic effects, most likely a decrease in the
rotational freedom of the n-butyl chains, might cause
the relatively slow reaction with DBE.
While the equilibrium constants for Lewis-base ad-
duct formation are affected by factors of 5 or less by any
modification of the metallocene ligand framework, the
rates of the anion displacement vary by more than 5
orders of magnitude: Complex 3a reacts 120 000 times
faster with DBE than complex 7a . Possible origins of
these unexpected rate differences, which correspond to
activation energy differences of ∆∆Gq ≈ 30 kJ /mol, will
be discussed below.
3. Ster eoch em istr y of th e An ion Disp la cem en t
Rea ction . Assignment of the proton positions in contact
ion pairs 2a and 3a and in their DMA adducts 2b and
3b by use of NOE techniques shows that the anion
displacement reaction is stereospecific in the sense that
incoming DMA occupies the same coordination site that
the anion has left. We are unable to detect any exchange
signals, e.g., between protons on the methyl side of 3a
and protons on the DMA side of 3b, which would
indicate that the methyl group moves from one side of
the complex to the other during the substitution process
(Figure 2).
Another process observed in these reaction systems
is the exchange of the anions of complexes 2a and 3a
with the anions of their respective Lewis-base adducts
2b and 3b (Figure 3). Under the conditions applied, this
exchange was nearly saturated. We can thus estimate
only a lower boundary of the observed rate constant,
kobs > 2-6 s-1. This corresponds to rate constants of k
> 5000 L‚(mol‚s)-1 if we assume that the reaction is
first-order both in the contact ion pair and the DMA-
containing ion pair. Much lower pseudo-first-order rate
constants of kobs ≈ 0.1 s-1 for site migration of the
methyl group9a,46,47 are indicated by small exchange
signals between the methyl and the anion sides of 2a ,
2b, 3a , or 3b. Exchange of a coordinated anion against
one bound in an outer-sphere complex thus proceeds
with the same stereospecifity as that of a coordinated
anion against a Lewis base. The same observations
concerning anion-Lewis base and anion-anion ex-
4. Mech a n ism of th e An ion Su bstitu tion Rea c-
tion . The rate law represented in eq 2, which holds for
all of the anion displacement reactions studied here, can
arise in principle from either one of two mechanisms:
(i) Either the anion is simply substituted by the Lewis
base by way of an SN2-type, associative mechanism
involving a transition state with a five-coordinated Zr
center, or (ii) the anion becomes reversibly detached
from the Zr center to form an outer-sphere ion pair. In
this case, coordination of the Lewis base to the vacant
coordination site would have to be rate-determining
(Scheme 3). The latter alternative is quite unlikely,
however: Ab initio calculations have indicated that even
the uptake of a weakly basic olefin ligand by a methyl
zirconocene cation is free of any significant enthalpic
barriers.8a-c
Rather than by a reversible anion dissociation, the
anion substitution reactions studied here are thus most
likely initiated by an associative attack of the Lewis
base at the Zr center. This is supported by negative
activation entropies of ∆Sq ) -50 to -40 J ‚mol-1‚K-1
,
obtained for the reactions of 6a and 7a with BDMA at
temperatures of 300-350 K (Table 5). The stereochem-
ical course of the anion displacement process indicates,
furthermore, that the five-coordinated zirconocene spe-
cies, which functions as a transition state or intermedi-
ate in this process, retains the methyl ligand in its
initial coordination site and has the anion and the
incoming ligandseither Lewis base or excess anionsin
adjacent coordination sites.
To check, at least qualitatively, on the accessibility
of such a species, we have modeled the reaction of a
somewhat simplified species H2Si(C5H4)2ZrMe+‚
Me(C6F5)3- (8a ) with trimethylamine by a semiempirical
method (PM3-tm). Along the reaction path, five-coordi-
nated Zr centers are indeed found as intermediates
(8d -8f, Figure 4). Each of these is lower in energy by
200-250 kJ /mol than an outer-sphere ion pair with a
vacant coordination site (8g) plus free NMe3. This
significant value supportssdespite the low accuracy
level of the method usedsan associative mechanism of
the ligand exchange process studied.
Among the five-coordinated intermediates, 8d is more
stable by approximately 40 kJ /mol than 8e or 8f. This
energy ordering agrees with the observed stereochem-
istry of the ligand exchange: The intermediate of lowest
energy carries the incoming Lewis base in the middle
position such that ligand exchange occurs with retention
(46) Siedle, A. R. Proc. Met. Con. 1993; Catalyst Consultants Inc.:
Spring House, PA, 1993; p 351. Siedle, A. R.; Newmark, R. A. J .
Organomet. Chem. 1995, 497, 119.
(47) Deck, P. A.; Marks, T. J . J . Am. Chem. Soc. 1995, 117, 6128.
Chen, Y. X.; Stern, C. L.; Yang, S. T.; Marks, T. J . J . Am. Chem. Soc.
1996, 118, 12451.

-
Displacement of H3CB(C6F5)3 Anions
Organometallics, Vol. 21, No. 3, 2002 479
Sch em e 3
ometries are calculated by DFT methods (Table 7):
Zr-C distances differ by less than 2 pm, coordination
angles by less than 2 deg.50 Energy profiles for elonga-
tion of the Zr-(µ-Me) distance (anion dissociation) for
both complexes are similar within deviations of less
than 10% (see Supporting Information).
Anion displacement was first modeled with dimethyl
ether as a Lewis base. For reaction of 9a with Me2O,
an energy minimum is found for a five-coordinated
-
intermediate
H2Si(C5H4)2ZrMe(OMe2)+MeB(CF3)3
,
9b , with the Lewis base positioned between the
methyl ligand and the anion (Figure 5), while lateral
attack of the Lewis base does not yield stable intermedi-
ates.51 Zr-C and Zr-O distances in 9b, which are
similar to experimental values reported for tetracoor-
dinated cationic zirconocene-alkyl-ether complexes,18a,b,52
confirm the strong interaction between the cationic
zirconium center and the Lewis base in the five-
coordinated intermediate. The ether ligand is oriented
perpendicular to the equatorial Me-Zr-O plane. Simi-
lar coordination of THF to (C5H5)2ZrMe+ was found by
diffractometry and explained by participation of oxygen
p-orbitals in the zirconium-oxygen bond.18a
Dissociation of the anion leads to the outer-sphere ion
pair 9c, where the ether ligand is now rotated into the
equatorial plane. The anion is without bonding contact
to the Zr center.53 Accordingly, the latter adopts a
tetrahedral coordination geometry, with Zr-C and
Zr-O distances slightly shorter than those in interme-
diate 9b.
F igu r e 4. Geometries and energies (relative to 8d )
calculated for intermediates in the substitution of the anion
from 8a by NMe3, calculated by semiempirical methods
(PM3-tm). Hydrogen atoms are omitted for clarity (with
the exception of Zr-bound methyl groups).
of stereochemistry at the metal center. Due to the
uncertain accuracy of the semiempirical method em-
ployed, we abstain from detailed characterizations of
probable transition states.48 The geometries and ener-
gies of these can nevertheless be assumed to resemble
those of the intermediates discussed above.
5. DF T Ca lcu la tion s on An ion Disp la cem en t by
Olefin s. To extrapolate the insights gained with regard
to anion displacement reactions by Lewis bases also
toward corresponding reactions with an olefin substrate,
the course of these reactions was further investigated
by means of DFT calculations.
Formation of the five-coordinated intermediate 9b
from 9a and free dimethyl ether is found to be exother-
mic with a coordination energy of -21 kJ /mol. Subse-
(50) Different orientations are found for the ligands in the equatorial
plane. While the methyl groups of 8a are symmetrically coordinated
in the equatorial plane with comparable C-Zr-Si angles, the Zr(Me)-
(anion) fragment in 9a is rotated in the equatorial plane by ca. 15°
relative to the Me2Si bridge. (cf. Burger, P.; Diebold, J .; Gutmann, S.;
Hund, H.-U.; Brintzinger, H. H. Organometallics 1992, 11, 1319.
Brintzinger, H. H.; Prosenc, M.-H.; Schaper, F.; Weeber, A.; Wieser,
U. J . Mol. Struct. 1999, 485-486, 409.)
(51) When the Zr-O distance is fixed to 2.6 Å, Lewis base attack
from the methyl side or the anion side leads to structures that are 40
and 20 kJ /mol, respectively, higher in energy than the corresponding
five-coordinated complex with the Lewis base in the middle position.
(52) (a) Alelyunas, Y. W.; Baenziger, N. C.; Bradley, P. K.; J ordan,
R. F. Organometallics 1994, 13, 148. (b) Bijpost, E. A.; Zuideveld, M.
A.; Meetsma, A.; Teuben, J . H. J . Organomet. Chem. 1998, 551, 159.
(c) Amorose, D. M.; Lee, R. A.; Petersen, J . L. Organometallics 1991,
10, 2191.
To limit the CPU time required for this more accurate
method,49 complex structures were simplified as fol-
lows: For the zirconocene contact ion pair, structure 9a
was used, which corresponds to the experimentally
studied complex 3a , with Me2Si replaced by H2Si and
-
MeB(C6F5)3 by MeB(CF3)3-. For contact ion pairs
H2Si(C5H4)2ZrMe+MeB(C6F5)3-, 8a , and H2Si(C5H4)2-
ZrMe+MeB(CF3)3-, 9a , closely similar coordination ge-
(53) One fluorine atom of the anion is found at a distance of 3.5 Å
from the metal center; this is presumably responsible for the relatively
large Me-Zr-O angle of 108°, compared to other tetrahedral oxygen-
coordinated cations and to the anion-free complex 9d .
(48) Børve, K. J .; J ensen, V. R.; Karlsen, T.; Støvneng, J . A.; Swang,
O. J . Mol. Model. 1997, 3, 193.
(49) Ziegler, T. Chem. Rev. 1991, 91, 651.

480 Organometallics, Vol. 21, No. 3, 2002
Ta ble 7. Selected Bon d Dista n cesa (Å) a n d An gles (d eg) Ca lcu la ted for Com p lexes 8a a n d 9a -d
8a 9a 9b 9c
9d c
2.26 2.27 2.29 2.26 2.26
Schaper et al.
b
Zr-Cterm
Zr-Cµ (Zr-H)b
Zr-B
2.45 (2.25)
4.12
2.24, 2.25
2.45 (2.19)
4.06
2.23, 2.23
3.70 (2.77)
5.04
2.24, 2.25
2.28
4.80
5.89
2.23, 2.26
2.24
65
Zr-ZCp
Zr-O
2.23, 2.23
2.28
b
Cterm-Zr-Cµ
96
97
152
b
ZCp-Zr-ZCp
126
127
125
83
126
108
126
99
b
Cterm-Zr-O
a
b
c
In Å. Cterm: methyl ligand bonded to Zr, Cµ: bridging methyl or trifluormethyl group bonded to B. Anion-free cationic complex.
F igu r e 5. Geometries and energies of the five-coordinated
intermediate and the outer-sphere reaction product of
anion substitution from 9a by Me2O calculated by DFT
methods.
Ta ble 8. Rela tive En er giesa Ca lcu la ted for th e
F ive-Coor d in a ted In ter m ed ia tes 8b-12b a n d for
Ou ter -Sp h er e Rea ction P r od u cts 8c-12c Occu r in g
in th e Rea ction s of 8a , 9a , or 12a w ith Dim eth yl
Eth er or Olefin s
F igu r e 6. Geometries of intermediates and products of
anion substitution from 9a by ethene (10b,c) or propene
(11b,c) calculated by DFT methods.
intermediates
products
8-12b
8-12c
Lewis
base
anion
Etoluene Egas phase Etoluene Egas phase
asymmetry in the olefin bonding becomes more pro-
nounced, especially for propene, most likely due to a
buildup of positive charge on the â-C atom.
-
9
Me2O MeB(CF3)3-
-27
23
22
40
5
-21
30
29
39
18
-38
11
19
52
11
-32
21
28
58
18
11 C3H7 MeB(CF3)3-
10 C2H4 MeB(CF3)3
8
12 C2H4 B(C6F5)4
For the propene-coordinated outer-sphere ion pair
11c, it was possible to calculate the geometry without
any restraints. In the case of the ethene complex 10c,
however, the distance between the metal center and the
nearest fluorine atom of the anion has to be fixed to the
value obtained for 11c.55 As expected, olefins act as
much weaker Lewis bases than Me2O. Starting from the
contact ion pair 9a and ethene, formation of an outer-
sphere ion pair with Zr-bound ethene is calculated to
require 19 kJ /mol; this is more endothermic by 57 kJ /
mol than for the analogous reaction with Me2O (Table
8). The reaction energy for formation of the outer-sphere
ion pair 11c from 9a and propene is 11 kJ /mol. A higher
Lewis basicity of propene, presumably by stabilization
of positive charge on the â-C atom by the methyl
substituent, is in agreement with coordination energies
quoted by other authors, which are more exothermic for
propene by ca. 10 kJ /mol.8e
-
C2H4 MeB(C6F5)3
-
a
In kJ /mol, relative to noninteracting 8-12a + Lewis base.
quent reorganization to the outer-sphere complex 9c is
further downhill by -11 kJ /mol (Table 8, Figure 5).
When single-point energies of complexes 9a -9c are
calculated in toluene solvent rather than in vacuo,
complexes 9b and 9c are found to be stabilized by 6 kJ /
mol relative to the situation in the gas phase. Formation
of the outer-sphere ion pair is expected to be further
favored by entropy contributions. Reaction entropies
calculated from the harmonic frequencies at 298 K in
the gas phase are less negative by ca. 60 J /(mol‚K) for
the formation of 9c than for 9b. Accordingly, outer-
sphere ion pairs of type 9c are expected to be the
dominant products in solution.
For anion substitution by ethene or propene, five-
coordinated intermediates with the olefin in the middle
position (10b and 11b) are again found to arise from
9a (Figure 6). Bond distances and angles at the zirco-
nium center are similar to those obtained for 9b (Table
9). The ethene and propene ligands are coordinated
perpendicular to the equatorial Si-Zr-Me plane of the
complex with unequal Zr-C distances.54 In the tetra-
coordinated outer-sphere ion pairs, 10c and 11c, the
olefin is rotated into the equatorial plane and the
(54) Bonding distances calculated for coordinated propene in 11b
(Zr-C1 ) 2.67 Å and Zr-C2 ) 3.04 Å) are comparable to those observed
in crystal structures of tetrahedral zirconocene complexes containing
a
chelated O-CHMe-(CH2)2-CHdCH2 ligand (see ref 15b). As
reported in these cases as well as in previous calculations on cationic
alkyl zirconocene complexes (see refs 8a,b,d), the olefin is found to be
unsymmetrically bound.
(55) Without restraints, coordination of a trifluoromethyl group of
-
the hypothetical MeB(CF3)3 anion to the metal center is followed by
dissociation of the olefin ligand or by its insertion into the zirconium-
methyl bond; finally fluoride is abstracted from the anion.

-
Displacement of H3CB(C6F5)3 Anions
Organometallics, Vol. 21, No. 3, 2002 481
Ta ble 9. Selected Bon d Dista n cesa (Å) a n d An gles (d eg) Ca lcu la ted for Olefin Com p lexes w ith Str u ctu r es
10b-e a n d 11b-c
10b
2.31
3.41 (2.40)
4.72
2.23, 2.23
2.81, 2.83
141
10c
10e
11b
2.30
3.65 (2.66)
4.90
2.23, 2.23
2.67, 3.04
145
11c
2.25
7.4
5.95
2.22, 2.25
2.57, 3.11
b
Zr-Cterm
2.24
2.34
2.90 (2.32)
4.43
2.24, 2.24
3.01, 3.01
140
Zr-Cµ (Zr-H)b
Zr-B
7.4
5.91
2.22, 2.24
2.79, 2.82
Zr-ZCp
Zr-olefinb
b
Cterm-Zr-Cµ
b
ZCp-Zr-ZCp
126
177
126
139
126
163
126
173
126
150
Si-Zr-C1(olefin)
a
b
In Å. Cterm: methyl ligand bonded to Zr, Cµ: bridging methyl group bonded to B, carbon atoms of the olefin given in the order C1, C2.
F igu r e 7. Geometries of intermediates and products of anion substitution from 8a and 12a by ethene calculated by DFT
methods.
Formation of the five-coordinated complexes 10b and
11b, on the other hand, requires 22-23 kJ /mol, both
for ethene and propene. This similarity between the two
olefins is unexpected in view of their different Lewis
basicity. The more symmetrical bonding of the olefins
in these five-coordinated complexes would indicate a less
pronounced buildup of positive charge on the â-carbon.
In addition, steric interactions, which are more unfavor-
able in the case of propene, might counteract its
electronic advantages. The transition state, 10e, for the
formation of the five-coordinated ethene complex 10b
has a geometry that is rather similar to that of the latter
(see Table 9). It is located at an energy of 32 kJ /mol
above that of 9a and free ethene, i.e., 10 kJ /mol above
that of 10b.
While these reaction energies would indicate that five-
coordinated intermediates occur in notable concentra-
tions, formation of outer-sphere ion pairs is likely to be
favored again by reaction entropies (vide supra). Forma-
tion of the five-coordinated intermediate and of the
outer-sphere complex requires 35-41 kJ /mol less for
B(C6F5)4- than for MeB(C6F5)3- as the counterion.57 This
is in agreement with results from NMR experiments:
While DBE reacts with rac-Me2Si(2-Me-BenzInd)2ZrMe+-
MeB(C6F5)3-, 7a , only partly within about 1 h, an
immediate and complete reaction was observed when
B(C6F5)4- instead of MeB(C6F5)3- was displaced by DBE.
The increased rate and extent of B(C6F5)4- displacement
might explain the generally higher polymerization activ-
ity of zirconocene ion pairs with the tetrakis(penta-
fluorophenyl)borate anion.
To study the influence that the anion might have on
the substitution process, five-coordinated intermediates
and outer-sphere reaction products were calculated for
-
(56) B(C6F5)4 coordination is possible either through two fluorine
atoms in the ortho and meta or in the meta and para posititions of the
phenyl ring. Since the latter results in an increased zirconium-boron
distance, the resulting structures are higher in energy by 47 kJ /mol.
In both cases, coordination with the B(C6F5)3 fragment pointing to the
outside of the complex, i.e., with the ortho-fluorine in the lateral or
the para-fluorine in the central coordination site, is more stable by
1-2 kJ /mol (see Supporting Information).
the reactions of ethene with the methyltris(pentafluo-
rophenyl)borate ion pair H2Si(C5H4)2ZrMe+MeB(C6F5)3
,
-
8a , and with the tetrakis(pentafluorophenyl)borate
complex H2Si(C5H4)2ZrMe+B(C6F5)4-, 12a (Figure 7).56
Rearrangement of the five-coordinated intermediates to
the outer-sphere ion pairs, i.e., from 8b to 8c and 12b
to 12c, is calculated to be endothermic with reaction
energies of 12 and 6 kJ /mol, respectively (Table 8).
(57) In accord with this energy difference, we find that coordination
-
of the MeB(C6F5)3 anion in the contact ion pair 8a via the methyl
group is more stable by 31 kJ /mol than coordination via two of the
fluorine atoms (see Supporting Information).

482 Organometallics, Vol. 21, No. 3, 2002
Schaper et al.
Ta ble 10. Rea ction En er giesa (in k J /m ol) Ca lcu la ted for th e Disp la cem en t of MeB(C6F 5)3 a n d B(C6F 5)4
An ion s by Eth en e
-
-
H2Si(C5H4)2ZrMe+
Cp2ZrMe+
Cp2ZrEt+
Cp2ZrBu+
ref 13a
localb
(Me2C5H3)2ZrMe+
this work
ref 21 ref 13b
ref 14
ref 13d
this work
b
b
b
b
b,c
anion
nonlocalb
nonlocal
(44)
local
34
nonlocal
local
local
nonlocalb,d nonlocald,e
-
MeB(C6F5)3
B(C6F5)4
52 (58)
11 (18)
(9/-4)f
(38)
59
-65
71 (91)
-
(-19/-40) f
-59 (-41)
10 (31) 9 (32)
a
Single-point energies in toluene solution with gas-phase-optimized geometries (energies in gas phase are denoted with parentheses).
bFunctional used for determination of the molecular geometry. Geometries taken from ref 13d. dGeometries taken from ref 13d and
reoptimized with the gradient-corrected BP functional. Alternative geometry for the contact ion pair. Olefin complex without/with â-agostic
interaction of the ethyl group.
c
e
f
In any case, we find that anion displacement proceeds
by way of a five-coordinated intermediate with ethene
or propene in the central coordination position. While
alternative reaction pathways have been proposed in
several recent DFT studies,12,13 our results are fully in
agreement with a thorough study by Nifant’ev et al.,14
in which transition states for various pathways of anion
substitution were compared for the first time.
Con clu sion s
From the experimental and theoretical results ob-
tained above, we can derive the following conclusions
concerning elementary reactions in zirconocene-based
catalyst systems for the polymerization of olefins.
The unexpectedly small rate constants found for the
substitution of the methylborate anion by Lewis bases,
e.g., in the case of complex 7a , indicate that the initial
displacement of the anion from the metal centers will
result in observable induction periods in polymerization
experiments, especially for propene in conjunction with
demanding zirconocene substituents, for which unfavor-
able steric interactions are likely to interfere with the
formation of five-coordinated intermediates. Stopped-
flow experiments at 40 °C with the catalyst rac-Me2Si-
(2-Me-4-Ph-Ind)2ZrCl2/MAO have indeed revealed an
induction period of about 0.3 s for the growth of polymer
chains with propene as monomer, while none was
observed for ethene.59
In the outer-sphere ion pair arising from the initial
substitution reaction, subsequent olefin insertions are
likely to proceed by way of a chain migratory insertion
mechanism analogous to that calculated for “free” alkyl
zirconocene cations.8 Coordination of an olefin and chain
growth are then in constant competition with re-
coordination of the anion. The rate with which the model
ligand DBE is substituted by the methylborate anion,
k-1 ) k/K, can be obtained from the experimental data
in Table 4 to be in the range 3 × 10-5 to 0.8 s-1 for
complexes 1a -7a . Even though olefins, as weaker Lewis
bases,60 are probably replaced more rapidly than dibutyl
ether, it appears unlikely that substitution of the olefin
by the methylborate anion (and even more so by less
Reaction energies obtained here for exchange of
MeB(C6F5)3- by ethene in the gas phase and in toluene
solution (52 and 58 kJ /mol, respectively) agree well with
values of 34-59 kJ /mol previously calculated13 or
obtained by combining DFT methods and NMR results
(Table 10).21 Much more negative reaction energies were
-
obtained by Vanka et al. for exchange of B(C6F5)4 by
ethene.13c,d Reoptimization of the reported molecular
geometries13c,d with a nonlocal functional, however, lead
to substantial changes in the geometry of the contact
ion pair and to more endothermic reaction energies for
the formation of the outer-sphere complex, which are
in close agreement with our results.
For anion substitution in Cp2ZrEt+A-, exothermic
reaction energies have been calculated by Nifant’ev et
al. from geometries obtained with nonlocal functionals
(Table 10).14 Activation and reaction energies obtained
by these authors for the formation of five-coordinated
intermediates are also lower by 20-30 kJ /mol than the
values obtained here. These differences can be attrib-
uted at least in part to the stabilization of the elec-
tron-deficient olefin complex by a â-agostic interaction
of the ethyl group. Our values of 11 and 52 kJ /mol calcu-
-
lated for the displacement of B(C6F5)4- or MeB(C6F5)3
from H2Si(C5H4)2ZrMe+A- by ethene are thus likely to
represent an upper limit for the respective reaction
energies.
-
coordinating cocatalyst anions such as B(C6F5)4 or
X-MAO-) can compete with the insertion reaction, for
which rates of approximately 10-100 s-1 can be esti-
mated from overall polymerization activities.
In summary, our DFT calculations support the pic-
ture of the anion displacement by a Lewis base de-
rived from experimental data, most notably the associa-
tive character of this ligand exchange reaction as well
as its stereochemistry. With olefin as a Lewis base, the
reaction follows the same patterns. Although our re-
sults lead us to assume that anion displacement by
ethene or propene is likely to be slower than that by
Me2O, we refrain from quantitative estimates of the
rates with which these olefins might displace the
anion from alkyl zirconocene ion pairs, due to the ill-
defined effects of reaction entropies and solvent in-
teractions. Nevertheless, it is to be expected that the
relative rates of anion displacement by an olefin follow
a pattern similar to that observed for DBE in the series
1a -7a .58
It might be argued that re-coordination of the anion
would occur to a coordinatively unsaturated insertion
product after each insertion step, instead of by an olefin
displacement. Such a scenario would require that the
anion be displaced again by an olefin before the next
(58) From molecular mechanics calculations it can be estimated that
-
displacement of MeB(C6F5)3 follows a similar trend with regard to
steric interactions in the series 1a -7a for DBE as well as for ethene
and propene (see Supporting Information).
(59) Busico, V.; Cipullo, R.; Esposito, V. Macromol. Rapid Commun.
1999, 20, 116.
(60) Catalyst activity in MAO-activated polymerizations using the
dichloride analogue of 7a is completely quenched by addition of DBE
in concentrations of 10-4-10-3 mol/L, i.e., lower by several orders of
magnitude than that of propene (Schaper, F.; Kirsten, R.; Brintzinger,
H. H. Unpublished results).

-
Displacement of H3CB(C6F5)3 Anions
Organometallics, Vol. 21, No. 3, 2002 483
in these polymerizations differ by a factor of ca. 100 and
correlate qualitatively with the estimated rate constants
of anion re-coordination (Figure 8).
These considerations make it likely that practically
important polymer properties are connected with rela-
tive rates of ligand substitution in metallocene catalyst
systems. It would therefore appear highly desirable to
gain more direct access, by improved experimental or
theoretical methods, to the rates with which olefins
displace different anions orsin the case of MAO-
activated catalystsstrimethylaluminum units from co-
ordinative contact with the Zr centers of different
catalysts.1,2,62
Ack n ow led gm en t. We thank Ms. Anke Friemel
(University of Konstanz) for excellent technical as-
sistance with two-dimensional NMR measurements and
Prof. I. E. Nifant’ev for sharing unpublished results with
us. Financial support by BMBF, by BASF AG, and by a
fellowship of the Landesgraduiertenfo¨rderung Baden-
Wu¨rttemberg is gratefully acknowledged.
F igu r e 8. Correlation of estimated relative molecular
weights in propene and ethene polymerizations catalyzed
by the dichloride analogues of 1a -7a /MAO relative to 3a /
MAO with the rate constant of anion re-coordination,
estimated for the substitution of dibutyl ether (DBE) by
-
MeB(C6F5)3
.
insertion can occur. Our results indicate, however, that
such an anion displacement would be much too slow in
comparison with average rates of polymer chain growth,
at least for more highly substituted complexes such as
6a and 7a . This allows us to conclude that coordination
of an olefin to the coordinatively unsaturated insertion
product is faster than coordination of the anion by
reorganization of the outer-sphere ion pair.
1
Su p p or tin g In for m a tion Ava ila ble: Tables of proton H
NMR data for reactions of 1a -7a with Lewis bases. DFT
calculations regarding effects of the ligand framework. Mo-
lecular mechanics calculations of the substitution reaction.
Details on the determination of kinetic constants in the 2D
NMR experiments and on the determination of relative mo-
lecular weights. Energy profiles of anion dissociation from
complexes 8a and 9a . Tables of absolute energies and of atomic
coordinates for complexes calculated by DFT methods. This
material is available free of charge via the Internet at
http://pubs.acs.org.
Lifetimes of approximately 100 s, however, for which
a polymer chain remains attached to a zirconium center,
might be long enough for re-coordination of the anion
to occur. A possible consequence of the formation of a
zirconocene polymer contact ion pair is the termination
of chain growth. Such an anion-initiated chain termina-
tion might explain shorter chain lengths found in some
polymerizations using the methylborate anion MeB-
OM010671L
(61) (a) Karol, F. J .; Kao, S.-C.; Wasserman, E. P.; Brady, R. C. New
J . Chem. 1997, 21, 797. (b) Suhm, J .; Schneider, M. J .; Mu¨lhaupt, R.
J . Mol. Catal. A: Chem. 1998, 128, 215. (c) Spaleck, W.; Kuber, F.;
Winter, A.; Rohrmann, J .; Bachmann, B.; Antberg, M.; Dolle, V.;
Paulus, E. P. Organometallics 1994, 13, 954. (d) Stehling, U.; Diebold,
J .; Kirsten, R.; Ro¨ll, W.; Brintzinger, H. H.; J u¨ngling, S.; Mu¨lhaupt,
R.; Langhauser, F. Organometallics 1994, 13, 964. (e) Herfert, N.
Dissertation, MPI Mu¨lheim, Universita¨t Du¨sseldorf, 1992.
-
-
(C6F5)3 compared to the less coordinating B(C6F5)4
anion29,47b and relative molecular weights obtained in
MAO-activated ethene and propene polymerizations
using the dichloride analogues of 1a-7a.61 Chain lengths
(62) Bochmann, M.; Lancaster, S. J . Angew. Chem. 1994, 106, 1715.