Analysis of metallocene-methylalumoxane methide transfer processes in solution using a 19F NMR spectroscopic probe

Article abstract of DOI:10.1021/om030171n

The metallocenes (C₆F₅C₅H ₄)CpZrMe₂ (1, Cp = C₅H₅) and (C ₆F₅C₅H₄)₂MMe₂ (2a, M = Zr; 2b, M = Hf) were treated with B(C₆F₅) ₃ and, separately, with commercial methylalumoxane (MAO, depleted of excess Me₃Al). The ensuing methide transfer reactions were followed by ¹⁹F NMR. Spectra of the activated species obtained using MAO and excess B(C₆F₅)₃ were strikingly similar, suggesting that similar cation-like species are formed using either organo-Lewis acid, and the MAO-activated species were tentatively formulated as (C₆F₅C₅H₄CpZrMe+ Me[AlMeO]_n^- and as (C₆F₅C ₅H₄)₂MMe⁺ Me[AlMeO]_n ^- (M = Zr, Hf). The relative amount of L₂MMe₂ and [L₂MMe]⁺ species observed in solution was measured as a function of [MAO] and as a function of [M] at constant [MAO]. Results are interpreted in terms of a canonical model in which MAO contains relatively few highly active Lewis acidic sites.

Full text of DOI:10.1021/om030171n

3558
Organometallics 2003, 22, 3558-3565
An a lysis of Meta llocen e-Meth yla lu m oxa n e Meth id e
Tr a n sfer P r ocesses in Solu tion Usin g a 19F NMR
Sp ectr oscop ic P r obe
Eric J . Hawrelak and Paul A. Deck*
Department of Chemistry, Virginia Tech, Blacksburg, Virginia 24061-0212
Received March 7, 2003
The metallocenes (C6F5C5H4)CpZrMe2 (1, Cp ) C5H5) and (C6F5C5H4)2MMe2 (2a , M ) Zr;
2b, M ) Hf) were treated with B(C6F5)3 and, separately, with commercial methylalumoxane
(MAO, depleted of excess Me3Al). The ensuing methide transfer reactions were followed by
19F NMR. Spectra of the “activated” species obtained using MAO and excess B(C6F5)3 were
strikingly similar, suggesting that similar “cation-like” species are formed using either
organo-Lewis acid, and the MAO-activated species were tentatively formulated as (C6F5C5H4)-
-
-
CpZrMe+ Me[AlMeO]n and as (C6F5C5H4)2MMe+ Me[AlMeO]n (M ) Zr, Hf). The relative
amount of L2MMe2 and [L2MMe]+ species observed in solution was measured as a function
of [MAO] and as a function of [M] at constant [MAO]. Results are interpreted in terms of a
canonical model in which MAO contains relatively few highly active Lewis acidic sites.
In tr od u ction
may be either “active” or “dormant”.11-14 The structure
of the alumoxane may also aggregate or somehow
change upon contact with the metallocene, revealing
“latent” Lewis acidic sites.15,16 Excess Me3Al present in
commercial MAO solution may also play important roles
besides adventitious contaminant scavenging, possibly
in stabilizing the active catalyst by forming Zr-Al
heterobimetallic complexes.17-22 Finally, most catalytic
reactions use metallocene dichlorides, whereas much of
the model chemistry has revolved around metallocene
dimethyls. The various alkylation, ligand abstraction,
and cation stabilization roles of Me3Al and MAO are
difficult to uncouple.
Activation of group 4 metallocene olefin polymeriza-
tion catalysts using methylalumoxane (MAO) is often
described as a Lewis acid-base equilibrium (eqs 1 and
2, L ) η5-C5H5 or similar ligand; M ) Ti, Zr, or Hf).1
Using 13C labeling techniques, Siedle showed that
methyl groups exchange among MAO and Zr sites,
suggesting that the methide abstraction process (eq 1)
is reversible.2 However, if eq 1 represents a true
equilibrium, one would expect to find a constant reaction
quotient (Q ≡ Keq). Marks and co-workers observed such
behavior using well-defined organo-Lewis acids (tri-
arylboranes) as MAO models,3 and they determined the
corresponding equilibrium constants using 1H NMR
analysis in the slow exchange regime.
Important quantitative aspects of metallocene activa-
tion by MAO also remain poorly understood. Some
(9) Barron, A. R. In Preparation, Properties, and Technology of
Metallocene-Based Polyolefins; Sheirs, J ., Kaminsky, W., Eds.; Wiley:
New York, 1999; p 33.
(10) Bryant, P. L.; Harwell, C. R.; Mrse, A. A.; Emery, E. F.; Gan,
Z. H.; Caldwell, T.; Reyes, A. P.; Kuhns, P.; Hoyt, D. W.; Simeral, L.
S.; Hall, R. W.; Butler, L. G. J . Am. Chem. Soc. 2001, 123, 12009-
12017.
-
L2MMe2 + [AlMeO]n h L2MMe+Me[AlMeO]n (1)
M
A
C
[C]
Q )
(2)
[M][A]
(11) Babushkin, D. E.; Brintzinger, H. H. J . Am. Chem. Soc. 2002,
124, 12869-12873.
(12) Zurek, E.; Ziegler, T. Organometallics 2002, 21, 83-92.
(13) Zhitao, X.; Vanka, K.; Firman, T.; Michalak, A.; Zurek, E.; Zhu,
C.; Ziegler, T. Organometallics 2002, 21, 2444-2453.
(14) Xu, Z.; Vanka, K.; Firman, T.; Michalak, A.; Zurek, E.; Zhu,
C.; Ziegler, T. Organometallics 2002, 21, 2444-2453.
(15) Harlan, C. J .; Bott, S. G.; Barron, A. R. J . Am. Chem. Soc. 1995,
117, 6465-6474.
Metallocene activation with MAO is apparently more
complicated.4 Commercial alumoxanes are ill-character-
ized mixtures of organoaluminum complexes with vary-
ing Lewis acidities.5-10 Experimental and theoretical
studies suggest that particular alumoxane compositions
(16) Koide, Y.; Bott, S. G.; Barron, A. R. Organometallics 1996, 15,
5514-5518.
(1) Chen, E. Y. X.; Marks, T. J . Chem. Rev. 2000, 100, 1391-1434.
(2) Siedle, A. R.; Lamanna, W. M.; Newmark, R. A.; Schroepfer, J .
N. J . Mol. Catal. A Chem. 1998, 128, 257-271.
(3) Deck, P. A.; Beswick, C. L.; Marks, T. J . J . Am. Chem. Soc. 1998,
120, 1772-1784.
(4) Pe´doutour, J . N.; Radhakrishnan, K.; Cramail, H.; Deffieux, A.
Macromol. Rapid Commun. 2001, 22, 1095-1123.
(5) Sinn, H.; Kaminsky, W.; Vollmer, H. J .; Woldt, R. Angew. Chem.,
Int. Ed. Engl. 1980, 19, 390-392.
(6) Sinn, H. Macromol. Symp. 1995, 97, 27-52.
(7) Imhoff, D. W.; Simeral, L. S.; Blevins, D. R.; Beard, W. R. ACS
Symp. Ser. 2000, 749, 177-191.
(8) Watanabi, M.; McMahon, C. N.; Harlan, C. J .; Barron, A. R.
Organometallics 2001, 20, 460-467.
(17) Bochmann, M.; Lancaster, S. J . Angew. Chem., Int. Ed. Engl.
1994, 33, 1634-1637.
(18) Tritto, I.; Li, S. X.; Sacchi, M. C.; Locatelli, P.; Zannoni, G.
Macromolecules 1995, 28, 5358-5362.
(19) Tritto, I.; Sacchi, M. C.; Locatelli, P.; Li, S. X. Macromol. Chem.
Phys. 1996, 197, 1537-1544.
(20) Eilertsen, J . L.; Rytter, E.; Ystenes, M. Vib. Spectrosc. 2000,
24, 257-264.
(21) Rytter, E.; Stovneng, J . A.; Eilertsen, J . L.; Ystenes, M.
Organometallics 2001, 20, 4466-4468.
(22) Babushkin, D. E.; Semikolenova, N. V.; Panchenko, V. N.;
Sobolev, A. P.; Zakharov, V. A.; Talsi, E. P. Macromol. Chem. Phys.
1997, 198, 3845-3854.
10.1021/om030171n CCC: $25.00 © 2003 American Chemical Society
Publication on Web 07/24/2003

Metallocene-Methylalumoxane Methide Transfer
Organometallics, Vol. 22, No. 17, 2003 3559
useful qualitative trends have been established by
examining metallocene/MAO mixtures in solution using
NMR spectroscopy.2,18,23-25 Nuclei such as 13C, 91Zr, 27Al,
and 11B enjoy excellent dispersion but suffer from low
sensitivity,10,26 whereas the reverse is true of 1H.
Interference from alumoxanes at high Al:Zr ratios and
exchange broadening have presented problems in both
1H and 13C NMR studies of activated metallocenes.
Isotopic enrichment of methyl groups offers convenient
improvement in 13C NMR signal intensity, but scram-
bling of the methyl groups into unenriched MAO (e.g.,
commercial samples) presents an important source of
error in quantitative work.
Qualitatively, however, the dispersion of 13C is useful
in identifying several different species. Using CPMAS
NMR techniques, Marks and co-workers showed that
the residue remaining after evaporating the solvent
from metallocene/MAO solutions is entirely (i.e., [C]/[M]
> 10) activated using an Al:Zr ratio of only 12:1.27 This
result provided some of the earliest evidence that the
large Al:Zr ratios used to optimize empirical polymer-
ization activities were not actually required to generate
the active “cation-like” intermediates.
Tritto and co-workers analyzed mixtures of Cp2Zr-
(13CH3)2 with either MAO or B(C6F5)3 using solution 13C
NMR spectroscopy and assigned well-resolved signals
to [(Cp2ZrMe)2(µ-Me)]+ and [Cp2ZrMe]+.28,29 In both
cation-like species, the δ(13C) values for the Zr-Me and
Cp ligand were nearly the same whether the counter-
anion was MeB(C6F5)3 or MeMAO. They then examined
the distribution of cation-like species varying temper-
ature, concentration, and reactant stoichiometry (Al:Zr
ratio). In their study, the dimer [(Cp2ZrMe)2(µ-Me)]+
was present in significant and varying amounts, and
their discussion focused mainly on the effect of concen-
trations and conditions on monomer-dimer equilibria,
confirming earlier reports that dimer formation in-
creased with overall zirconium concentration.30
solution UV-vis spectroscopy.4,32 The intense ligand-
to-metal charge transfer (LMCT) bands enable reliable
relative concentration estimates at low metallocene
concentrations, even though the absorptions of different
species are not fully resolved. Although their conclusions
are qualitative (major species are assigned to broad Al:
Zr regimes), they observe complete conversion of rac-
[C2H4(Ind)2]ZrMe2 to a cationic species in toluene solu-
tion when Al:Zr > 150. Ma¨kela¨ and co-workers obtained
similar results with a series of 2-trialkylsiloxy-substi-
tuted bis(indenyl) and bis(tetrahydroindenyl) complexes
of zirconium.33 One particularly interesting aspect of the
UV-vis studies is their apparent sensitivity to the
presence of chloride, trimethylaluminum, and other
species that could be present in practical polymeriza-
tions. For example, Cramail and Deffieux observed
different species in the activation of L2ZrX2 type com-
plexes depending on X (Cl or Me),34 whereas Weiser and
Brintzinger observed a common species regardless of the
X ligands.35 Heterobimetallic species derived from Me3-
Al such as [Ind2Zr(µ-Me2)AlMe2]+, which may increase
catalyst lifetimes,17,21,36 are generally observed when
using MAO as the activator but sometimes not when
using other alkylalumoxanes.37 However, with only a
few exceptions the UV-vis technique has been limited
to conjugated systems such as indenyl complexes (or
M ) Ti38) to avoid interference from MAO in the region
280-330 nm.
Taken together, the complexity of existing spectro-
scopic and polymerization data convinces us that the
development of new approaches to the study of metal-
locene-alumoxane reactions is warranted. In particular,
quantitative data arising from analyses in solution are
still rather scarce. We now report39 that C6F5 substit-
uents on the metallocene ancillary ligands provide
strong yet well-resolved 19F NMR signals corresponding
to M and C (eq 1). We use the relative integrated
intensities of these signals to probe this activation model
(eq 1) as a function of MAO and metallocene concentra-
tions.
Babushkin and co-workers obtained comparable re-
1
sults using both 13C NMR and H NMR spectroscopy.31
On the basis of careful analysis of δH(C5H5) as a function
of Al:Zr ratio (holding [Al] constant to rule out medium
effects), they concluded that MAO may contain a
distribution of aluminum acids with varying competence
for methide abstraction and that different portions of
that distribution are sampled at different values of Al:
Zr.
Resu lts a n d Discu ssion
P r ecu r sor Syn th esis. We prepared the C6F5-sub-
stituted metallocenes 1, 2a (M ) Zr), and 2b (M ) Hf)
by standard methods (Scheme 1). Hydrocarbon suspen-
sions of CH3Li avoid substitution at the para positions
of the C6F5 groups, which can be a troublesome side-
reaction when ethereal solvents are used. Furthermore,
the methylation reactions and subsequent product isola-
tion must be carried out in the dark. Although the
Deffieux and Cramail followed the alkylation and
activation of rac-[C2H4(Ind)2]ZrX2 (X ) Cl, CH3) using
(23) Tritto, I.; Li, S. X.; Sacchi, M. C.; Zannoni, G. Macromolecules
1993, 26, 7111-7115.
(24) Nekhaeva, L. A.; Bondarenko, G. N.; Rykov, S. V.; Nekhaev,
A. I.; Krenstel, B. A.; Mar’in, V. P.; Vyshinskaya, L. I.; Khrapova, I.
M.; Polonski, A. V.; Korneev, N. N. J . Organomet. Chem. 1991, 406,
139-146.
(32) Pe´deutour, J . N.; Cramail, H.; Deffieux, A. J . Mol. Catal. A
Chem. 2001, 174, 81-87.
(33) Ma¨kela¨, N. I.; Knuuttila, H. R.; Linnolahti, M.; Pakkanen, T.
A.; Leskela¨, M. A. Macromolecules 2002, 35, 3395-3401.
(34) Pe´deutour, J . N.; Cramail, H.; Deffieux, A. J . Mol. Catal. A
Chem. 2001, 176, 87-94.
(35) Wieser, U.; Brintzinger, H. H. Organomet. Catal. Olefin Polym.
2001, 3-13.
(36) Reddy, S. S.; Shashidhar, G.; Sivaram, S. Macromolecules 1993,
26, 1180.
(37) Wang, Q.; Song, L.; Zhao, Y.; Feng, L. Macromol. Rapid Comm.
2001, 22, 1030-1034.
(38) Long, W. P.; Breslow, D. S. J . Am. Chem. Soc. 1960, 82, 1953-
(25) Zhang, Z. Y.; Duan, X. F.; Zheng, Y.; Wang, J .; Tu, G. Z.; Hong,
S. L. J . Appl. Polym. Sci. 2000, 77, 890-897.
(26) Komon, Z. J . A.; Rogers, J . S.; Bazan, G. C. Organometallics
2002, 21, 3189-3195.
(27) Sishta, C.; Hathorn, R. M.; Marks, T. J . J . Am. Chem. Soc. 1992,
114, 1112-1114.
(28) Tritto, I.; Donetti, R.; Sacchi, M. C.; Locatelli, P.; Zannoni, G.
Macromolecules 1997, 30, 1247-1252.
(29) Tritto, I.; Donetti, R.; Sacchi, M. C.; Locatelli, P.; Zannoni, G.
Macromolecules 1999, 32, 264-269.
1957.
(30) Beck, S.; Prosenc, M. H.; Brintzinger, H. H.; Goretzki, R.;
Herfert, N.; Fink, G. J . Mol. Catal. A Chem. 1996, 111, 67-79.
(31) Babushkin, D. E.; Semikolenova, N. V.; Zakharov, V. A.; Talsi,
E. P. Macromol. Chem. Phys. 2000, 201, 558-567.
(39) Preliminary data was presented orally: Deck, P. A.; Thornberry,
M. P.; Hawrelak, E. J . Book of Abstracts, 220th National Meeting of
the American Chemical Society, Washington, DC, Aug 20-24, 2000,
Abstract INOR 192.

3560 Organometallics, Vol. 22, No. 17, 2003
Hawrelak and Deck
spectrum of 3 (Figure 1b) shows the expected downfield
shifts in each of the three C6F5 signals as the metal-
locene becomes more electron-deficient. The conversion
of 2a /2b to 4a /4b (eq 4) is analogous. An isolated
analytical sample of 4a shows the expected equally
integrating sharp Zr-Me (0.19 ppm) and broadened
1
B-Me (0.34 ppm) signals in the H NMR spectrum and
the 2:3 ratio of C6F5 groups in the 19F NMR spectrum.40
1 + B(C6F5)3 h
[(C6F5C5H4)CpZrMe]+[MeB(C6F5)3]- (3) (3)
2a /2b + B(C6F5)3 h
[(C6F5C5H4)2MMe]+[MeB(C6F5)3]-
F igu r e 1. 19F NMR spectra in C6D6: (a) 1; (b) 1 + excess
B(C6F5)3; (c) 1 + MAO.
(4a , M ) Zr; 4b, M ) Hf) (4)
Interestingly, when metallocene 1 in C6D6 is treated
with about 0.5 equiv of B(C6F5)3, signals assigned to the
metallocene dimethyl (1) and the monomeric cation (3)
Sch em e 1
1
are observed in the H and 19F NMR spectra along with
signals assigned to a methyl-bridged dinuclear complex
(eq 5). The concentration ratio of the monometallic
contact ion pair 3 to the dinuclear species was 3:1 at
[Zr]total ) 30 mM. Starting with similar concentrations
of Cp2ZrMe2, Beck and Brintzinger found that treatment
with 0.5 equiv of B(C6F5)3 led to a mixture that was
much richer in the dinuclear cationic species.30 This
trend extends to the disubstituted metallocene (2a ).
When a 30 mM solution of 2a was treated with 0.5 equiv
of B(C6F5)3, only unreacted 2a and the monometallic
contact ion pair 4a were observed. C6F5 substituents
strongly attenuate the tendency to form these dinuclear
cationic species: the monometallic cation is more Lewis
acidic, but the neutral dimethyl complex is also less
Lewis basic, and on balance the reaction shown in eq 5
lies more toward the left.
L2ZrMe2 + [L2ZrMe]+[MeB(C6F5)3]- h
[L2ZrMe}2(µ-Me)]+[CH3B(C6F5)3]- (5)
When 1 is treated instead with MAO, the 19F chemical
shifts observed (Figure 1c) for the putative zirconoce-
nium aluminate (5) are nearly identical to those ob-
served in the “well-defined” monometallic species (3).
From this observation, we infer that 1 reacts with MAO
according to eq 6. We further infer that the cationic
portions of 3 and 5 are structurally analogous, within
the sensitivity of this 19F NMR probe. Similarly, 2a and
2b react with MAO to afford 6a and 6b, respectively
(eq 7), and the 19F NMR spectra of 4a /4b (cationic
portion) and 6a /6b are respectively coincident. Spectra
are provided in the Supporting Information.
products, when purified, are not especially photosensi-
tive, methylation reactions carried out under ambient
light slowly deposit intractable black solids and give
poor yields. The cause of this photodecomposition re-
mains unknown at present, although the possibility of
intramolecular CF activation in these complexes has
provided us a new line of inquiry.
1 + MAO h [(C6F5C5H4)CpZrMe]+[MeMAO]- (5)
Sp ectr oscop ic Assign m en t of Activa ted Meta l-
locen es. Figure 1a shows the 19F NMR spectrum of the
arylated zirconocene dimethyl complex (1). Upon treat-
ment with a slight excess of the well-defined activator
B(C6F5)3, 1 is completely converted (eq 3) to the corre-
sponding “cation-like” species 3.40-42 The 19F NMR
(6)
2a /2b + MAO h [(C6F5C5H4)2MMe]+[MeMAO]-
(6a , M ) Zr; 6b, M ) Hf) (7)
(41) Yang, X.; Stern, C. L.; Marks, T. J . J . Am. Chem. Soc. 1994,
116, 10015-10031.
(42) Siedle, A. R.; Newmark, R. A. J . Organomet. Chem. 1995, 497,
119-125.
(40) Yang, X. M.; Stern, C. L.; Marks, T. J . J . Am. Chem. Soc. 1991,
113, 3623-3625.

Metallocene-Methylalumoxane Methide Transfer
Organometallics, Vol. 22, No. 17, 2003 3561
Broadening in the 19F NMR spectra of 5, 6a , and 6b
is attributed to two phenomena. First, there is a
distribution of aluminate “counterion” structures, be-
cause MAO is, after all, a mixture of compounds. In
some spectra, features appear to emerge from these
broad bands. Potentially, reactions of 1 and 2a /2b with
well-defined aluminum activators could enable us to
further assign these bands; these studies are underway.
Importantly, the spectra do not have any signals cor-
responding to µ-Me-bridged dimeric cations. Well-known
fluxional processes can also contribute to line broaden-
ing.3,40,43 Such broadening can be seen in the signals
assigned to 1 in Figure 1c and is attributed to the
dynamic nature of the abstraction equilibrium (eq 6).
In an EXSY experiment,44 strong off-diagonal signals
correlated respective signals assigned to 1 and 5,
indicating chemical exchange in eq 6.
When a large excess of Me3Al was added to a solution
of the MAO-activated metallocenium complex (6a ), no
change was observed in the 19F NMR spectrum. If
Me3Al adducts represented only part of an equilibrating
mixture, then we would expect to see the band shape
assigned to 6a to change with [Me3Al].17,28,31,45 Never-
theless, our “Me3Al-depleted” MAO sample almost
certainly contains some residual Me3Al.7,46-48 Therefore
we concede that the possibility remains that Me3Al
adducts account for all of our activated species (5, 6)
and that the 19F chemical shift is completely insensitive
to the adduct formation. Addition of Me3Al to a solution
of the B(C6F5)3-activated model compound 4a resulted
in rapid decomposition, probably due to methyl-aryl
exchange between Al and B centers.49-51 Additional
study of the reactions of C6F5-substituted metallocenes
with simple aluminum alkyls and aluminum alkyl
halides is underway in our laboratories.
Con cen tr a tion Stu d ies. Titration of a toluene solu-
tion of either 1 or 2a /2b with MAO at 25 °C led to a
monotonic decrease in the intensity of the signals
assigned to 1 or 2a /2b and a corresponding increase in
the signals assigned to 5 and 6a /6b, respectively. In our
samples, the starting concentrations ([1]o [2]o, and
[MAO]o) were known, and the equilibrium ratios [1]:[5]
and [2]:[6] were obtained from integration of the well-
resolved para regions of the 19F NMR spectra. These
data enabled us to calculate a reaction quotient (Q)
according to eq 2. Importantly, if Q were constant with
respect to [metallocene]o, [MAO]o, and time, we could
infer that Q is an equilibrium constant.
F igu r e 2. Plot of Q (calculated by eq 2) vs [MAO]: (a)
1 + MAO; (b) 2a + MAO; (c) 2b + MAO.
findings. First, Q is within the range 0.1 < Q < 50.
Tritto and co-workers observed (using 13C NMR) a
reaction quotient of 4.0 for methide abstraction from
Cp2TiCH3Cl (eq 8) at -20 °C.52 Here it should be noted
that our calculation of [MAO] uses a formula weight of
58 g mol-1, corresponding to the arbitrary [AlMeO]
“repeat unit,” and then simply [MAO] ) [Al]. Actual
number-averaged molecular weights of MAO samples
are typically 400 or higher and are subject to complex
ligand disproportionation pathways as samples age.7,48
While [M]o was fixed at a nominal concentration of
10 mM, we determined Q at the widest range of MAO
concentrations that would still allow us to integrate
signals assigned to both forms of the metallocene (M
and C). The results (Figure 2) give rise to three general
Cp2TiMe2 + MAO‚AlMe2Cl f
[Cp2TiCl]+[MAO‚AlMe3]- (8)
(43) Siedle, A. R.; Newmark, R. A.; Lamanna, W. M.; Schroepfer, J .
N. Polyhedron 1990, 9, 301-308.
(44) Perrin, C. L.; Dwyer, T. J . Chem. Rev. 1990, 90, 935-967.
(45) Coevoet, D.; Cramail, H.; Deffieux, A. Macromol. Chem. Phys.
1998, 199, 1451-1457.
(46) Resconi, L.; Bossi, S. Macromolecules 1990, 23, 4489-4491.
(47) Barron, A. R. Organometallics 1995, 14, 3581-3583.
(48) Imhoff, D. W.; Simeral, L. S.; Sangokoya, S. A.; Peel, J . H.
Organometallics 1998, 17, 1941-1945.
(49) Klosin, J .; Roof, G. R.; Chen, E. Y. X.; Abboud, K. A. Organo-
metallics 2000, 19, 4684-4686.
Using 58 g mol-1 as the formula weight for MAO
allows direct comparison of Q to practical Al:Zr ratios.
Useful catalyst concentrations are micromolar or even
less (i.e., [C] ≈ [M]o ≈ 10-6 M). On the basis of typical
amounts of MAO reported to optimize catalytic activity,
a large “excess” of MAO (e.g., [A] ≈ 1000 [M]o ≈ 10-3
M) is required to keep the catalyst mostly in its active
form (e.g., [C]/[M] ≈ 10). These quantities lead to a very
rough estimate of about 104 for Q. In polymerization
(50) Lee, C. H.; Lee, S. J .; Park, J . W.; Kim, K. H.; Lee, B. Y.; Oh,
J . S. J . Mol. Catal. A Chem. 1998, 132, 231-239.
(51) Kim, J . S.; Wojcinski, L. M.; Liu, S. S.; Sworen, J . C.; Sen, A.
J . Am. Chem. Soc. 2000, 122, 5668-5669.

3562 Organometallics, Vol. 22, No. 17, 2003
Hawrelak and Deck
reactions, especially when large volumes of solvent are
used, one can never be sure how much alumoxane is
consumed scavenging impurities or “reactivating” dor-
mant species. Some silica-supported metallocene cata-
lysts are reported to be fully activated at Al:Zr ratios
well below 100, and as we recalled earlier, Marks found
complete activation at Al:Zr < 15.53-55
The second general finding arising from Figure 2 is
that a comparison of the reaction quotients for 1 (Q ≈
3.5) and 2a (Q ≈ 1.3) at a fixed alumoxane concentration
([MAO] ≈ 0.2 M) is consistent with our prior observa-
tions that C6F5 substituents are strongly electron-
withdrawing,56-60 destablizing the “cation-like” form (C)
more in the case of 2a . Comparison of the reaction
quotients for 2a and 2b at a fixed alumoxane concentra-
tion ([MAO] ≈ 0.2 M) is consistent with Marks’s
observation that abstraction of methide from hafnocene
dimethyl is less exothermic than abstraction from
zirconocene dimethyl.3
Our third finding is that Q increases with increasing
[MAO] (Figure 2) for metallocenes 1 and 2a . MAO is
arguably more polar than the reaction solvent (toluene),
and aggregation of MAO around the ion-paired species
as a diffuse “solvent effect” would stabilize C relative
to M with increasing [MAO].61,62 However, we expected
a “solvent effect” to be more gradual and to influence
methide abstraction from the three metallocenes simi-
larly. However, Figure 2 shows that the Q varies
dramatically with [MAO] and that the effect of [MAO]
on Q is more pronounced for 1 (a factor of 10 over 0.1
M < [MAO] < 0.6 M, Figure 2a) than for 2a (a factor of
5 over 0.1 M < [MAO] < 1.3 M, Figure 2b). The value
of Q, within experimental error, is constant for 2b over
the range 0.1 < [MAO] < 3.0 M (Figure 2c). The spectra
obtained for the hafnium complex 2b showed low
concentrations of 6b, even when [MAO] was close to
saturation. There is significant error in the determina-
tion of Q for 2b. However, this result suggests that
rational metallocene design can allow us to probe the
limits of the methide-abstraction capability of MAO.
Methide abstraction from 2b is not discussed further.
We investigated the possibility that more than one
MAO “molecule” may be required to abstract a methyl
group from a metallocene or that MAO forms discreet
aggregates with the activated “cationic” form of the
metallocene (C). Babushkin and Brintzinger recently
characterized a metallocene-alumoxane complex using
F igu r e 3. Plot of Q (calculated by eq 2) vs [2a ].
pulsed-field-gradient NMR techniques to estimate hy-
drodynamic radii and concluded that each metallocene
is associated with an alumoxane species having, on
average, about 50 aluminum atoms. Still it is not known
if all of those aluminum atoms are part of a single
covalent framework or if several smaller alumoxane
fragments might be aggregated. We attempted to model
our data by introducing the number of MAO molecules
required as a parameter (eq 9). Varying the aggregation
number (0 < a < 2.0) made Q constant with [MAO] at
high [MAO] but not at low [MAO]. In an earlier version
of this paper, a reviewer pointed out that we might lose
a more significant fraction of [MAO] to reaction with
adventitious moisture at low [MAO]. However CH4
signals (δH ) 0.15 ppm) are not observed in the 1H NMR
spectra of these mixtures, suggesting that in situ
hydrolysis of MAO is negligible. We conclude that we
cannot rule out the possibility that multiple Al sites are
involved in the abstraction process, but that theory just
does not fit our data better than interacting with only
one reactive Al site at a time (a ) 0).
Cp′2MMe2 + (a +1)[AlMeO]n h
Cp′2MMe+Me[AlMeO]n-‚a [AlMeO]n (9)
To rule out medium effects, we next conducted a
separate series of experiments in which we held [MAO]o
at a fixed value (0.35 M) and varied the initial metal-
locene concentration, [2a ]o, from about 2-10 mM (lim-
ited by our ability to obtain reasonable integrals). Q was
calculated according to eq 2, and the results are plotted
in Figure 3. In concert with Babushkin’s finding,31 we
observed a decrease in Q with increasing [1]o. Moreover,
we found that the concentration of the “cation-like”
species, [C], is roughly constant for all experiments,
suggesting that even after the first aliquot of metal-
locene is added (Al:Zr ) 70:1), all of the “active”
aluminum sites are already titrated.
Tem p er a tu r e Effects. A solution of 2a (nominally
30 mM with Al:Zr ≈ 35) was analyzed using 19F NMR
spectroscopy at several temperatures ranging from 50
to -80 °C. Between 50 and -40 °C, the percent of
metallocene in the cation-like form (6a ) was constant
(within (10%). Below -40 °C, a decrease in the appar-
ent concentration of 6a was noted, but we disregarded
these data because significant gelation of the sample
had also occurred. Tritto and co-workers had reported
a dramatic decrease in the “proportion” of cationic
(52) Tritto, I.; Sacchi, M. C.; Sanxi, L. Macromol. Rapid Commun.
1994, 15, 217-223.
(53) Hlatky, G. G. Chem. Rev. 2000, 100, 1347-1376.
(54) Fink, G.; Steinmetz, B.; Zechlin, J .; Przybyla, C.; Tesche, B.
Chem. Rev. 2000, 100, 1377-1390.
(55) J ezequel, M.; Dufaud, V.; Ruiz-Garcia, M. J .; Carrillo-Her-
mosilla, F.; Neugebauer, U.; Niccolai, G. P.; Lefebvre, F.; Bayard, F.;
Corker, J .; Fiddy, S.; Evans, J .; Broyer, J . P.; Malinge, J .; Basset, J .
M. J . Am. Chem. Soc. 2001, 123, 3520-3540.
(56) Deck, P. A.; J ackson, W. F.; Fronczek, F. R. Organometallics
1996, 15, 5287-5291.
(57) Deck, P. A.; Fronczek, F. R. Organometallics 2000, 19, 327-
333.
(58) Thornberry, M. P.; Slebodnick, C.; Deck, P. A.; Fronczek, F. R.
Organometallics 2000, 19, 5352-5369.
(59) Thornberry, M. P.; Slebodnick, C.; Deck, P. A.; Fronczek, F. R.
Organometallics 2001, 20, 920-926.
(60) Maldanis, R. J .; Chien, J . C. W.; Rausch, M. D. J . Organomet.
Chem. 2000, 599, 107-111.
(61) Vizzini, J . C.; Chien, J . C. W.; Babu, G. N.; Newmark, R. A. J .
Polym. Sci. Polym. Chem. 1994, 32, 2049-2056.
(62) Deck, P. A.; Marks, T. J . J . Am. Chem. Soc. 1995, 117, 6128-
6129.

Metallocene-Methylalumoxane Methide Transfer
Organometallics, Vol. 22, No. 17, 2003 3563
products when Cp2ZrMe2 was combined with 10 equiv
of MAO and the temperature lowered from 25 °C to
-78 °C,28 corroborating an earlier study using a mixture
of Cp2Ti13CH3Cl and 10 equiv of MAO.23 Their total
aluminum concentrations were 0.7 M, whereas ours
were marginally higher (1.0 M). In their earlier report
they also noted that solutions of MAO are “quite viscous,
especially at very low temperatures”.
Ti(NMe2)4, presumably by intramolecular substitution
in putative LTi(NMe2)3 intermediates. A 19F NMR
spectrum of 4a at -80 °C showed rapid rotation of the
C6F5 groups, indicating that a Fortho‚‚‚Zr interaction, if
present, is not strong enough to impede that dynamic
process. We have also found that 1 and 2a are active
catalysts for homopolymerization of ethylene and copo-
lymerization of ethylene and 1-hexene, whereas strong
intramolecular Fortho‚‚‚Zr coordination might be expected
to exert a poisoning effect. The polymerization activity
is not as high as Cp2ZrCl2, but ordinary substituent
trends could account for that decrease.60,63-65 These
polymerization data will be reported elsewhere.
We now propose a model to account for our observa-
tions. First, MAO comprises aluminum sites having a
distribution of Lewis acidities, and the number of acidic
sites that are “competent” to abstract methide from the
metallocene is relatively small. Thus, when a fixed
amount of MAO reacts with a small amount of 1, there
are enough “competent” sites to generate an appreciable
fraction of cationic metallocene. However, additional 1
does not generate much additional cationic metallocene,
because the most reactive alumoxane sites are already
consumed (titrated). Thus, the apparent Q is described
by a function that is approximately reciprocal in the
metallocene concentration (Figure 3). The same concept
could also rationalize the data in Figure 2. At low
[MAO], there are only enough “competent” Lewis acidic
sites to activate a small fraction of the metallocene. As
[MAO] increases, the number of “competent” sites and
the concentration of cationic metallocene also increase
proportionally. The effect on eq 2 is to increase [C] in
proportion to [A], but the concomitant decrease in [M]
raises Q. From Figure 2 it is also clear that the fraction
of “competent” sites is metallocene-dependent, such that
only alumoxane sites that are more acidic than the
corresponding activated metallocenium species are “com-
petent”. The more Lewis basic metallocene 1 samples a
different fraction of acidic sites than does 2a . Metal-
locene 2b may be sufficiently electron-deficient to
engage in a more even-handed equilibrium with perhaps
only a single type of alumoxane Lewis acidic site,
rationalizing the nearly constant value of Q observed
for 2b at various MAO concentrations (Figure 2c). The
temperature-independence of Q is also consistent with
a model in which a small number of highly acidic sites
are “titrated” by the metallocene. The “equilibrium
constant” for those sites is so high that the change in
temperature does not have a measurable effect. Yet, the
activation must still be a dynamic equilibrium to have
observed off-diagonal signals in the EXSY spectrum. In
the latter two aspects, the behavior of B(C6F5)3 presents
a good model. We concede that our model is speculative
and that additional experimentation will be needed to
test it further. We are also presently working to
understand the complex reaction pathways that occur
when a metallocene dichloride is treated with methyl-
alumoxane. Preliminary results are qualitatively con-
sistent with the UV-vis spectroscopic studies cited
above.
Exp er im en ta l Section
Gen er a l Con sid er a tion s. All manipulations were carried
out using standard nitrogen-atmosphere techniques. MeLi was
purchased from Aldrich as a solution in diethyl ether and dried
under high vacuum (3 × 10-5 Torr) at 60 °C. For these
procedures, it is important to remove as much ether as
possible. MAO was received from Albemarle as a 10% toluene
solution, filtered, and dried under high vacuum (3 × 10-5 Torr)
for 15 h to remove “free” trimethylaluminum. The same sample
of MAO was used for all the experiments described here.
B(C6F5)3 was obtained as a gift from Albemarle, sublimed
(120 °C, 5 × 10-6 Torr), recrystallized from hexanes, and found
be 99+% pure according to 19F NMR spectroscopy. [(C6F5)-
C5H4]2ZrCl2 and Na[(C6F5)C5H4] were prepared according to
our own methods.56 CpZrCl3(DME) was prepared according to
the method of Lund and Livinghouse.66 NMR spectra were
recorded on a Varian U-400 instrument at 22 °C. 19F NMR
spectra are referenced to external C6F6 in CDCl3 at -163.0
ppm. Elemental microanalyses were performed by Oneida
Research Services (Whitesboro, NY) or Desert Analytics
(Tucson, AZ).
(P en ta flu or op h en yl)zir con ocen e Dich lor id e. A mixture
of CpZrCl3(dme) (1.50 g, 4.25 mmol), Na[(C6F5)C5H4] (1.10 g,
4.33 mmol), and toluene (400 mL) was stirred at 110 °C for
2 h. The hot mixture was filtered, and the precipitate was
rinsed with an additional 100 mL of hot toluene. The yellow
filtrate was cooled to 25 °C, and the resulting crystalline
product was collected on a filter, washed with hexanes, and
dried under vacuum to afford 1.47 g (3.21 mmol, 76%) of a
yellow crystalline solid. 1H NMR (C6D6): δ 6.45 (m, 2 H), 5.89
(s, 5 H), 5.75 (m, 2 H). 19F NMR (C6D6): δ -140.06 (m, 2 F),
-155.18 (t, 3J ) 22 Hz, 1 F), -163.14 (m, 2 F). Anal. Calcd for
C
16H9Cl2F5Zr: C, 41.93; H, 1.98. Found: C, 42.05; H, 1.89.
(P en ta flu or op h en yl)zir con ocen e Dim eth id e (1). A mix-
ture of [(C6F5)C5H4]CpZrCl2 (500 mg, 1.09 mmol), methyl-
lithium (480 mg, 22 mmol), and toluene (20 mL) was stirred
at 25 °C in the dark for 18 h. The solvent was evaporated,
and the residue was recrystallized from 20 mL of hexanes to
afford 391 mg (0.94 mmol, 86%) of colorless crystals. 1H NMR
(C6D6): δ 6.29 (m, 2 H), 5.70 (s, 5 H), 5.58 (m, 2 H), -0.27 (s,
6 H). 19F NMR (C6D6): δ -141.08 (m, 2 F), -157.59 (t, 3J )
22 Hz, 1 F), -163.36 (m, 2 F). Satisfactory elemental analysis
1
could not be obtained in repeated attempts. H and 19F NMR
spectra are included (Figures S1 and S2 in the Supporting
Information) as evidence of substantial bulk purity.
Str u ctu r a l Da ta . We have not yet been able to obtain
crystal structures of any of our B(C6F5)3-activated
metallocenium complexes. Chiefly we want to know if
there are any Fortho‚‚‚Zr interactions. A crystal structure
of 2b (to be reported elsewhere) shows no F‚‚‚Zr interac-
tions. However, in related work still underway, we have
found that 3-(pentafluorophenyl)indene and (penta-
fluorophenyl)cyclopentadiene undergo regioselective
ortho C-F substitution by NMe2 upon treatment with
1,1′-Bis(pen taflu or oph en yl)zir con ocen e Dim eth ide (2a).
A mixture of [(C6F5)C5H4]2ZrCl2 (1.00 g, 1.60 mmol), methyl-
(63) Mohring, P. C.; Coville, N. J . J . Organomet. Chem. 1994, 479,
1-29.
(64) Mohring, P. C.; Coville, N. J . J . Mol. Catal. A Chem. 1995, 96,
181-195.
(65) Mohring, P. C.; Vlachakis, N.; Grimmer, N. E.; Coville, N. J . J .
Organomet. Chem. 1994, 483, 159-166.
(66) Lund, E. C.; Livinghouse, T. Organometallics 1990, 9, 2426-
2427.

3564 Organometallics, Vol. 22, No. 17, 2003
Hawrelak and Deck
lithium (700 mg, 32 mmol), and toluene (35 mL) was stirred
at 25 °C in the dark for 18 h. The solvent was evaporated,
and the residue was recrystallized from 35 mL of hexanes to
afford 770 mg (1.32 mmol, 82%) of colorless crystals. 1H NMR
(C6D6): δ 6.32 (m, 4 H), 5.60 (m, 4 H), -0.39 (s, 6 H). 19F NMR
(C6D6): δ -140.92 (m, 4 F), -156.95 (t, 3J ) 24 Hz, 2 F),
-163.11 (m, 4 F). Satisfactory elemental analysis could not
be obtained in repeated attempts. 1H and 19F NMR spectra
are included (Figures S3 and S4 in the Supporting Informa-
tion) as evidence of substantial bulk purity.
1,1′-Bis(p en ta flu or op h en yl)h a fn ocen e Dich lor id e. A
mixture of HfCl4(THF)2 (1.50 g, 3.23 mmol), sodium (pen-
tafluorophenyl)cyclopentadienide (1.72 g, 6.78 mmol), and
toluene (400 mL) was stirrred at 110 °C for 2 h. The hot
mixture was filtered, and the filter was washed with 50 mL
of additional hot toluene. Cooling the yellow filtrate to 25 °C
afforded a crystalline product, which was collected on a filter,
washed with hexane, and dried under vcuum to afford 1.51 g
(2.12 mmol, 66%) of pure [(C6F5)C5H4]2HfCl2. 1H NMR (C6D6):
δ 6.41 (m, 4 H), 5.73 (m, 4 H). 19F NMR (C6D6): δ -140.13 (d,
4 F), -154.76 (t, 2 F), -162.93 (m, 4 F). Anal. Calcd for C22H8-
Cl2F10Hf: C, 37.13; H, 1.13. Found: C, 37.79; H, 1.13.
1,1′-Bis(p en ta flu or op h en yl)h a fn ocen e Dim eth id e (2b).
A mixture of [(C6F5)C5H4]2HfCl2 (1.00 g, 1.41 mmol), methyl-
lithium (620 mg, 28 mmol), and toluene (35 mL) was stirred
at 25 °C in the absence of light for 18 h. Toluene was
evaporated, and the residue was extracted with 35 mL of warm
hexane. The mixture was filtered, and the filtrate was
concentrated and cooled to afford 508 mg (0.76 mmol, 54%) of
pure 2b. 1H NMR (C6D6): δ 6.26 (m, 4 H), 5.53 (m, 4 H), -0.59
(s, 6 H). 19F NMR (C6D6): δ -141.04 (m, 4 F), -156.85 (tt, 2
F), -163.81 (m, 4 F). Anal. Calcd for C24H14F10Hf: C, 42.97;
H, 2.10. Found: C, 42.78; H, 2.05.
-161.0 (m, 4 F, meta-CF of cation), -164.2 (m, 3 F, para-CF
of MeBAr3 anion), -166.8 (m, 6 F, meta-CF of MeBAr3 anion).
On the basis of integration data we conclude that the remain-
ing signal, assigned to the ortho-CF of the cation, is located
at -141.1 ppm (4 F), coincident with the ortho-CF signal of 1.
To confirm the assignment of the anion associated with the
dinuclear cation in the 19F NMR spectrum, we repeated the
experiment reported by Beck et al.,30 in which a 35 mM
solution of Cp2ZrMe2 was treated with about 0.5 equiv of
B(C6F5)3, but 19F NMR data were not reported. Under these
conditions we find, as did Beck et al., that the monometallo-
cenium cation (with its associated anion) is the minor product,
whereas the major product is the “associated” dinuclear cation
[(Cp2ZrCH3)2(µ-CH3)2][CH3B(C6F5)3]. The three strongest sig-
nals in the 19F NMR spectrum were at -132.4, -164.2, and
-166.8 ppm.
Syn th esis of (C6F 5C5H4)2Zr Me+MeB(C6F 5)3 (4). The
-
following procedure was carried out with minimal exposure
to light. A solution of 2 (105 mg, 0.18 mmol) and B(C6F5)3
(96 mg, 0.19 mmol) in toluene (15 mL) was stirred for 5 min
at 25 °C and then evaporated. The residue was triturated with
10 mL of pentane, and the yellow precipitate was collected on
a filter and dried under vacuum to afford 160 mg (0.15 mmol,
81%) of a yellow solid. 1H NMR (toluene-d8): δ 6.37 (br s,
2 H), 5.98 (br s, 2 H), 5.61 (m, 4 H), 0.34 (br s, 3 H), 0.19 (s,
3 H). 19F NMR (toluene-d8): δ -135.2 (d, 6 F), -141.5 (d, 4 F),
-152.3 (t, 2 F), -159.1 (t, 3 F), -161.1 (m, 4 F), -164.8 (m,
6 F). Anal. Calcd for C42H14BF25Zr: C, 46.05; H, 1.29. Found:
C, 46.45; H, 0.90. Combining 2 and a slight excess of B(C6F5)3
in hexane solution at 25 °C also affords good yields of 4, which
is collected on a filter and dried under vacuum.
Titr a tion of Dim eth yl Com p lexes w ith MAO. All weigh-
ings were carried out using an analytical balance ((0.1 mg
resolution) in a nitrogen glovebox, using caution to minimize
exposure of the samples to ambient light. A typical experiment
was conducted as follows. A solution of 1 (3.99 × 10-2 M) in
C6D6 was prepared by dissolving 50 mg of 1 in 3.0 mL of C6D6
(needed for spectrometer frequency lock and magnetic field
shim adjustments). A solution of MAO (0.865 M) in toluene
was prepared by dissolving 251 mg of MAO in 5.0 mL of
toluene. To each of five J -Young NMR tubes was added 380
mg of the C6D6 solution of 1. Assuming that 1 forms an ideal
solution in C6D6 (F ) 0.95 g mL-1), the density of the C6D6
solution of 1 was 0.97. A 380 mg sample of this solution
corresponds to 0.400 mL; therefore, 0.016 mmol of 1 was added.
To each of the five NMR tubes, a different amount (e.g., 160,
407, 622, 800, and 950 mg) of the MAO solution in toluene
was added. Assuming that MAO forms an ideal solution in
toluene, the density of the toluene (F ) 0.865) solution of MAO
was 0.915. The added weights of this solution correspond to
0.18, 0.44, 0.68, 0.87, and 1.04 mL, or 0.16, 0.38, 0.59, 0.75,
and 0.87 mmol of MAO, respectively. Then, sufficient toluene
was added by weight to each tube to bring the total calculated
volume to 1.45 mL. The samples were shaken vigorously,
allowed to equilibrate for 5 min, and then subjected to 19F
NMR analysis. Longer equilibration times (up to 2 h in the
dark at 25 °C) did not change the spectra noticably. Stacked
plots showing the treatment of 2a /2b with B(C6F5)3 to generate
4a /4b and with MAO to generate 6a /6b are provided in the
Supporting Information (Figures S7, S8).
Solu t ion Ob ser va t ion of “Ca t ion -Lik e” Sp ecies (C6-
F 5C5H4)Cp Zr Me+MeB(C6F 5)3- (3) a n d (C6F 5C5H4)2Zr Me+-
-
MeB(C6F 5)3 (4). In a resealable (J -Young) NMR tube, a
solution of 1 and B(C6F5)3 (slightly more than 1 equiv) was
prepared in C6D6. NMR spectra showed that conversion to 3
was complete, with no remaining 1 but some unreacted
B(C6F5)3. Data for 3: 1H NMR (C6D6): δ 6.20 (br s, 1 H), 5.85
(br s, 1 H), 5.49 (m, 2 H), 5.48 (s, 5 H), 0.23 (broad,
unsymmetrical singlet, 6 H). 19F NMR (C6D6): δ -134.31 (m,
3
6 F), -140.82 (m, 2 F), -152.30 (t, J ) 21 Hz, 1 F), -158.89
(m, 3 F), -160.72 (m, 2 F), -164.45 (m, 6 F). The same
procedure was used to convert 2a to 4a . Data for 4a : 1H NMR
(C6D6): δ 6.32 (br s, 2 H), 5.95 (br s, 2 H), 5.57 (m, 4 H), 0.33
(br s, 3 H, BCH3), 0.18 (s, ZrCH3). 19F NMR (C6D6): δ -134.54
(m, 6 F), -140.54 (m, 4 F), -151.64 (t, 3J ) 21 Hz, 2 F),
-158.71 (m, 3 F), -160.53 (m, 4 F), -164.41 (m, 6 F). The
same procedure was used to convert 2b to 4b, except that
toluene-d8 was used, and the NMR spectra were collected at
-30 °C; otherwise the signals were too broad to assign. Data
for 4b: 1H NMR (toluene-d8): δ 6.20 (br s, 2 H), 5.78 (m, 2 H),
5.48 (m, 4 H, two coincident Cp-H signals), 0.64 (br s, 3 H,
B-CH3), -0.08 (br s, 3 H, Hf-CH3). 19F NMR (toluene-d8): δ
3
-134.50 (m, 6 F), -140.00 (m, 4 F), -151.72 (t, J ) 22 Hz, 2
F), -158.15 (m, 3 H), -160.31 (m, 4 F), -164.02 (m, 6 F).
Obser va tion of [{[C6F 5C5H4)Cp Zr Me}2(µ-Me)]+[CH3B-
(C6F 5)3]-. In a resealable J -Young NMR tube, a solution of 1
1
(29 mM) and B(C6F5)3 (14 mM) was prepared in C6D6. H and
19F NMR analyses showed unreacted 1 (50% of Zr), mono-
nuclear contact ion pair 3 (30% of Zr), and a new set of signals
Me3Al Ad d ition Exp er im en ts. In the glovebox, an NMR
tube was charged with 2a (11.9 mg, 20.3 µmol), B(C6F5)3 (12.5
mg, 24.4 µmol), C6D6 (∼1.0 mL), and Me3Al (∼0.5 mL of a 2.0
M solution in toluene, ∼1.0 mmol). NMR spectra (1H and 19F)
showed no signals that could be assigned to 2a , 4a , or B(C6F5)3.
1
accounting for 20% of the Zr present. In the H NMR spectrum
most of the signals were obscured by signals arising from 1
and 3. Only one of the Cp-H signals was well-resolved at 6.42
ppm (2 H). A signal at -1.00 ppm was assigned to the µ-CH3
group (br s, 3 H). On the basis of integration data we conclude
that the terminal CH3 signal coincides with the signal arising
from the ZrMe2 group of 1 at -0.27 ppm. Other data for the
dinuclear complex: 19F NMR (C6D6): δ -132.5 (m, 6 F, ortho-
CF of MeBAr3 anion), -153.1 (m, 2 F, para-CF of cation),
1
Instead, the H NMR spectrum showed four strong multiplets
at 6.60, 6.33, 5.74, and 5.46 ppm and a set of four weaker,
broadened signals at 6.30, 6.16, 5.60, and 5.56 ppm, in addition
to a strong singlet at 0.75 ppm, a weaker triplet at 0.12 ppm,
and Me3Al at -0.15 ppm. The 19F NMR spectrum showed
strong signals at -115.9, -122.6, -130.4, -140.5, -140.9,

Metallocene-Methylalumoxane Methide Transfer
Organometallics, Vol. 22, No. 17, 2003 3565
-152.9, -155.2, -155.5, -156.2, -161.7, -161.8, and -162.8
ppm in addition to a smaller number of weak signals. A second
NMR tube was charged with 2a (13.2 mg, 22.6 µmol), MAO
(80.2 mg, 1.38 mmol), and C6D6 (∼1 mL). 1H and 19F NMR
spectra were recorded, showing a mixture of 2a and 6a . An
identical sample was prepared, but this time a large excess of
Me3Al (∼1.0 mmol as a 2.0 M toluene solution, obtained from
Aldrich) was added. The 1H and 19F NMR spectra were
unchanged except for a large signal assigned to the added Me3-
Al in the 1H NMR spectrum.
Va r ia ble-Tem p er a tu r e Stu d y. Small portions of 2a (10.0
mg, 17 µmol) and MAO (35 mg, 610 µmol, Al:Zr ) 36) were
weighed into an NMR tube in the glovebox. The tube was
sealed with a valved, Teflon-sealed cap and interfaced to a
high-vacuum line. Toluene-d8 (roughly 0.6 mL) was condensed
into the sample. Upon shaking at room temperature, a clear
yellow solution was obtained. 19F NMR analysis revealed the
ratio of 2a :6a at the following temperatures: 42:58 at 50 °C,
47:53 at 25 °C, 48:52 at 0 °C, 47:53 at -20 °C, 46:54 at -40
°C, 39:61 at -60 °C, and 27:73 at -80 °C. Cooling the sample
to -78 °C for 5 min using a dry ice-2-propanol bath resulted
in faint turbidity and significant gelation of the sample.
to well-defined model compounds, and their relative
concentrations are obtained simply by spectral integra-
tion. Apparent reaction quotients obtained for methide
abstraction are much lower than one might anticipate
based on Al:Zr ratios and concentrations actually used
in olefin polymerization but at least qualitatively con-
sistent with other spectroscopic studies. The observed
effects of [MAO] on Q suggest that a simple equilibrium
model (eqs 1 and 2) does not describe the methide
abstraction by MAO at the concentrations used in this
study. Rather, MAO behaves as though only a small
fraction of highly reactive sites are present. Ongoing
studies in our laboratory are directed toward under-
standing metallocene activation in greater depth.
Ack n ow led gm en t. We thank W. F. J ackson for
preliminary experiments. This work was supported by
the Petroleum Research Fund (30738-G) and the Na-
tional Science Foundation (CHE-98-75446).
Su p p or tin g In for m a tion Ava ila ble: Experimental de-
tails, including tables of data obtained from the NMR titration
experiments and the equations used to determine Q from
measured quantities. This information is available free of
charge via the Internet at http://pubs.acs.org.
Con clu sion s
19F NMR spectroscopy is a useful probe of methide
abstraction from metallocenes by MAO. Metallocene
dimethides and “cationic” methylmetallocenium species
are readily distinguished using chemical shifts assigned
OM030171N