Living Ziegler-Natta polymerization by early transition metals: Synthesis and evaluation of cationic zirconium alkyl complexes bearing β-hydrogens as models for propagating centers

Article abstract of DOI:10.1021/ja057866v

The synthesis and characterization of a series of cationic zirconium and hafnium complexes with alkyl substituents bearings-hydrogens of general formula {(η-C₅Me₅)MR[N(Et)C(Me)N(t-Bu)]}[B(C₆F ₅)₄] [M = Zr; R = Et, n-Pr, i-Pr, i-Bu, i-Bu, and 2-ethylbutyl (5a-f) and M = Hf; R = i-Bu and t-Bu (6 and 7, respectively)] is described, including several isotopically labeled derivatives. The ability of these complexes to serve as model complexes for the living Ziegler-Natta polymerization of olefins that can be effected using the initiator 2a (R = Me in 5) has been addressed. The results obtained shed additional light on the steric and electronic factors that can contribute to the living character of a Ziegler-Natta polymerization based on an early transition metal initiator.

Full text of DOI:10.1021/ja057866v

Published on Web 02/16/2006
Living Ziegler-Natta Polymerization by Early Transition
Metals: Synthesis and Evaluation of Cationic Zirconium Alkyl
Complexes Bearing â-Hydrogens as Models for Propagating
Centers
Matthew B. Harney, Richard J. Keaton, James C. Fettinger, and Lawrence R. Sita*
Contribution from the Department of Chemistry and Biochemistry, UniVersity of Maryland,
College Park, Maryland 20742
Received November 19, 2005; E-mail: lsita@umd.edu
Abstract: The synthesis and characterization of a series of cationic zirconium and hafnium complexes
with alkyl substituents bearing â-hydrogens of general formula {(η5-C5Me5)MR[N(Et)C(Me)N(t-Bu)]}[B(C6F5)4]
[M ) Zr; R ) Et, n-Pr, i-Pr, n-Bu, i-Bu, and 2-ethylbutyl (5a-f) and M ) Hf; R ) i-Bu and t-Bu (6 and 7,
respectively)] is described, including several isotopically labeled derivatives. The ability of these complexes
to serve as model complexes for the living Ziegler-Natta polymerization of olefins that can be effected
using the initiator 2a (R ) Me in 5) has been addressed. The results obtained shed additional light on the
steric and electronic factors that can contribute to the living character of a Ziegler-Natta polymerization
based on an early transition metal initiator.
Introduction
electronic factors that are required to allow a propagating center
involving an early transition metal to continue to propagate
The development of well-defined transition metal complexes
that can serve as initiators for the living, and in some cases
stereospecific, polymerization of R-olefins is the latest chapter
in a decades-long success story that has covered the discovery,
mechanistic understanding, and commercialization of homoge-
neous “single-site” catalysts for polyolefin production through
the Ziegler-Natta process.1-11 However, in contrast to the
wealth of experimental and theoretical information now available
regarding the various mechanisms for chain termination and
chain release in nonliVing group 4 bis(cyclopentadienyl)metal-
based systems (also known as metallocenes), which are largely
dominated by mononuclear â-hydride elimination and â-hydride
transfer reactions,12 little is still known of the structural and
without such type of termination. Once this knowledge becomes
available, though, the design of second and third generation
initiators for (stereospecific) living Ziegler-Natta polymeriza-
tions that can operate at higher temperatures, and for a wider
range of monomers, should be feasible.
For group 4 metallocenes, cationic complexes bearing alkyl
groups with â-hydrogens, and typically either an ethyl or
isobutyl group, have long been considered as viable experimental
and theoretical models for the proposed polymer chain-bound
propagating center.12,13 Further, from these studies, it is now
widely accepted that observation of a strong intramolecular
(6) (a) Jayaratne, K. C.; Sita, L. R. J. Am. Chem. Soc. 2000, 122, 958-959.
(b) Jayaratne, K. C.; Keaton, R. J.; Henningsen, D. A.; Sita, L. R. J. Am.
Chem. Soc. 2000, 122, 10490-10491. (c) Keaton, R. J.; Jayaratne, K. C.;
Fettinger, J. C.; Sita, L. R. J. Am. Chem. Soc. 2000, 122, 12909-12910.
(d) Keaton, R. J.; Jayaratne, K. C.; Henningsen, D. A.; Koterwas, L. A.;
Sita, L. R. J. Am. Chem. Soc. 2001, 123, 6197-6198. (e) Jayaratne, K. C.;
Sita, L. R. J. Am. Chem. Soc. 2001, 123, 10754-10755. (f) Zhang, Y.;
Keaton, R. J.; Sita, L. R. J. Am. Chem. Soc. 2003, 125, 9062-9069. (g)
Zhang, Y.; Sita, L. R. J. Am. Chem. Soc. 2004, 126, 7776-7777. (h)
Kissounko, D. A.; Zhang, Y.; Harney, M. B.; Sita, L. R. AdV. Synth. Catal.
2005, 347, 426-432.
(1) For a review of living Ziegler-Natta polymerization, see: Coates, G. W.;
Hustad, P. D.; Reinartz, S. Angew. Chem., Int. Ed. 2002, 41, 2236-2257.
(2) For recent reviews of homogeneous Ziegler-Natta polymerization, see:
(a) Brintzinger, H. H.; Fischer, D.; Mu¨lhaupt, R.; Rieger, B.; Waymouth,
R. M. Angew. Chem., Int. Ed. Engl. 1995, 34, 1143-1170. (b) Coates, G.
W. Chem. ReV. 2000, 100, 1223-1252. (c) Gibson, V. C.; Spitzmesser, S.
K. Chem. ReV. 2003, 103, 283-315.
(3) (a) Scollard, J. D.; McConville, D. H. J. Am. Chem. Soc. 1996, 118, 10008-
10009. (b) Scollard, J. D.; McConville, D. H.; Rettig, S. J. Organometallics
1997, 16, 1810-1812. (c) Scollard, J. D.; McConville, D. H.; Vittal, J. J.;
Payne, N. C. J. Mol. Catal. A 1998, 128, 201-204.
(7) (a) Tshuva, E. Y.; Goldberg, I.; Kol, M. J. Am. Chem. Soc. 2000, 122,
10706-10707. (b) Tshuva, E. Y.; Goldberg, I.; Kol, M.; Goldschmidt, Z.
J. Chem. Soc., Chem. Commun. 2001, 2120-2121. (c) Segal, S.; Goldberg,
I.; Kol, M. Organometallics 2005, 24, 200-202.
(4) (a) Baumann, R.; Davis, W. M.; Schrock, R. R. J. Am. Chem. Soc. 1997,
119, 3830-3831. (b) Baumann, R.; Schrock, R. R. J. Organomet. Chem.
1998, 557, 69-75. (c) Schrock, R. R.; Baumann, R.; Reid, S. M.; Goodman,
J. T.; Stumpf, R.; Davis, W. M. Organometallics 1999, 18, 3649-3670.
(d) Mehrkhodavandi, P.; Bonitatebus, P. J.; Schrock, R. R. J. Am. Chem.
Soc. 2000, 122, 7841-7842. (e) Mehrkhodavandi, P.; Schrock, R. R. J.
Am. Chem. Soc. 2001, 123, 10746-10747. (f) Mehrkhodavandi, P.; Schrock,
R. R.; Bonitatebus, P. J., Jr. Organometallics 2002, 21, 5785-5798. (g)
Mehrkhodavandi, P.; Schrock, R. R.; Pryor, L. L. Organometallics 2003,
22, 4569-4583. (h) Schrock, R. R.; Adamchuk, J.; Ruhland, K.; Lopez,
L. P. H. Organometallics 2005, 24, 857-866.
(8) (a) Mitani, M.; Mohri, J.-I.; Yoshida, Y.; Saito, J.; Ishii, S.; Tsuru, K.;
Matsui, S.; Furuyama, R.; Nakano, T.; Tanaka, H.; Kojoh, S.-I.; Matsugi,
T.; Kashiwa, N.; Fujita, T. J. Am. Chem. Soc. 2002, 124, 3327-3336. (b)
Mitani, M.; Nakano, T.; Fujita, T. Chem. Eur. J. 2003, 9, 2396-2403. (c)
Suzuki, Y.; Terao, H.; Fujita, T. Bull. Chem. Soc. Jpn. 2003, 76, 1493-
1517. (d) Makio, H.; Fujita, T. Bull. Chem. Soc. Jpn. 2005, 78, 52-66.
(9) (a) Tian, J.; Coates, G. W. Angew. Chem., Int. Ed. 2000, 39, 3626-3629.
(b) Tian, J.; Hustad, P. D.; Coates, G. W. J. Am. Chem. Soc. 2001, 123,
5134-5135. (c) Mason, A. F.; Coates, G. W. J. Am. Chem. Soc. 2004,
126, 16326-16327.
(5) (a) Hagihara, H.; Shiono, T.; Ikeda, T. Macromolecules 1998, 31, 3184-
3188. (b) Hasan, T.; Ioku, A.; Nishii, K.; Shiono, T.; Ikeda, T. Macro-
molecules 2001, 34, 3142-3146. (c) Nishii, K.; Shiono, T.; Ikeda, T.
Macromol. Rapid Commun. 2004, 25, 1029-1032.
(10) (a) Busico, V.; Cipullo, R.; Friederichs, N.; Ronca, S.; Togrou, M.
Macromolecules 2003, 36, 3806-3808. (b) Busico, V.; Cipullo, R.;
Friederichs, N.; Ronca, S.; Talarico, G.; Togrou, M.; Wang, B. Macro-
molecules 2004, 37, 8201-8203.
9
3420
J. AM. CHEM. SOC. 2006, 128, 3420-3432
10.1021/ja057866v CCC: $33.50 © 2006 American Chemical Society

Living Ziegler−Natta Polymerization by Early Transition Metals
A R T I C L E S
Scheme 1
agostic interaction between the metal center and a hydrogen
atom in the â-position of an alkyl substituent is a strong predictor
for â-hydride elimination, or â-hydride transfer, according to
Scheme 1.14 Given this, it is reasonable to expect that, by
preventing the formation of â-hydrogen agostic interactions,
perhaps through steric control of the ligand environment about
the metal center, one should be able to design systems that are
living by shutting down these termination pathways, assuming,
of course, that no other termination pathways exist. Indeed,
Schrock and co-workers4 have presented such a steric argument
to rationalize the living character of polymerizations mediated
by their {[NON]ZrMe}+[B(C6F5)4]- ([NON]2- ) [t-Bu-d6-N-
o-C6H4)2O]2-) initiator, where â-hydride elimination involving
the polymer chain is seen to occur only very slowly at 0 °C.
Further, in another of their living systems, direct observation
living polymerization of ethene and propene at 50 °C. On the
basis of DFT calculations, these investigators attributed the
observed remarkable stability of the propagating center to an
attractive force between an ortho-fluorine of the N-pentafluo-
rophenyl group within the ligand and a â-hydrogen on the
growing polymer chain that they feel is strong enough to
mitigate formation of a â-hydrogen agostic interaction with the
metal center, thereby preventing both â-hydride elimination and
â-hydride transfer to monomer. Chan and co-workers15 have
recently provided solution NMR evidence for a H‚‚‚F interaction
between a fluorine atom of an ortho-CF3 phenyl substituent and
the R-hydrogen of a benzyl group within a neutral (phenoxy-
pyridine)zirconium benzyl complex, and they have proposed
that, while not being a direct model, this observation lends
support to the Fujita hypothesis. However, subsequent theoretical
investigations indicate that the effects of the ortho-fluorine atoms
in the Fujita system are, at best, primarily steric in nature and
serve to destabilize the transition state for â-hydride transfer to
monomer.16
In addition to termination, â-hydride elimination within a
propagating center appears to be capable of playing a critical
role in introducing stereoerrors within the polymer microstruc-
ture when propene is polymerized by group 4 metallocene-based
catalysts under conditions of low monomer concentration.17-20
As Scheme 2 reveals, these stereoerrors appear to be the result
of an overall “chain-end epimerization” process that proceeds
through two consecutive sequences of steps involving â-hydride
elimination, alkene rotation about the metal, and reinsertion.
Originally proposed by Busico and co-workers,17 this mecha-
nism has received considerable support through polymerization
studies involving isotopically labeled propene by groups of
investigators led by Brintzinger18 and Bercaw.19 Even then, there
are elements of this overall process that remain enigmatic,
including the requirement for formation of a tertiary alkyl
1
by H NMR spectroscopy of the cationic isobutyl initiators,
{[MesNpy]M(i-Bu)}+[B(C6F5)4]- (M ) Zr and Hf; [MesNpy]2-
) [H3CC(2-C5H4N)(CH2Nmesityl)2]2-), did not reveal any
evidence for a â-hydrogen agostic interaction between the metal
and the methine proton of the isobutyl group. In apparent
accordance, then, with expectations based on the absence of
such a â-hydrogen agostic interaction, both of these initiators
for M ) Zr and Hf were found to decompose only very slowly
at 0 °C in a strictly first-order fashion via assumed â-hydride
elimination, with t1/2 ) 40 min for M ) Zr and t1/2 ) 21 h for
M ) Hf.4g
Fujita and co-workers8 have recently presented a different
hypothesis to account for the surprising robustness of their
titanium bis(phenoxyimine)-based system, derived from activa-
tion of TiCl2{η2-1-[C(H)dN(C6F5)]-2-O-3-(t-Bu)C6H3}2 with
methylaluminoxane (MAO), that is capable of effecting the
(11) For late transition-metal living systems, see: (a) Schmidt, G. F.; Brookhart,
M. J. Am. Chem. Soc. 1985, 107, 1443-1444. (b) Brookhart, M.; Volpe,
A. F., Jr.; Lincoln, D. M.; Horvath, I. T.; Millar, J. M. J. Am. Chem. Soc.
1990, 112, 5634-5636. (c) Brookhart, M.; DeSimone, J. D.; Grant, B. E.;
Tanner, M. J. Macromolecules 1995, 28, 5378-5380. (d) Killian, C. M.;
Tempel, D. J.; Johnson, L. K.; Brookhart, M. J. Am. Chem. Soc. 1996,
118, 11664-11665. (e) Ittel, S. D.; Johnson, L. K.; Brookhart, M. Chem.
ReV. 2000, 100, 1169-1204. (f) Gottfried, A. C.; Brookhart, M. Macro-
molecules 2001, 34, 1140-1142. (g) Gottfried, A. C.; Brookhart, M.
Macromolecules 2003, 36, 3085-3100.
(15) Kui, S. C. F.; Zhu, N.; Chan, M. C. W. Angew. Chem., Int. Ed. 2003, 42,
1628-1632.
(16) (a) Talarico, G.; Busico, V.; Cavallo, L. Organometallics 2004, 23, 5989-
5993. (b) Vanka, K.; Xu, Z.; Ziegler, T. Organometallics 2005, 24, 419-
430.
(12) (a) Schneider, M. J.; Suhm, J.; Mu¨lhaupt, R.; Prosenc, M.-H.; Brintzinger,
H.-H. Macromolecules 1997, 30, 3164-3168. (b) Deng, L.; Ziegler, T.;
Woo, T. K.; Margl, P.; Fan, L. Organometallics 1998, 17, 3240-3253. (c)
Margl, P.; Deng, L.; Ziegler, T. J. Am. Chem. Soc. 1999, 121, 154-162.
(d) Rappe´, A. K.; Skiff, W. M.; Casewit, C. J. Chem. ReV. 2000, 100,
1435-1456. (e) Chirik, P. J.; Bercaw, J. E. Organometallics 2005, 24,
5407-5423 and references therein.
(13) Jordan, R. F. AdV. Organomet. Chem. 1991, 32, 325-387.
(14) (a) Brookhart, M.; Green, M. L. H.; Wong, L. L. Prog. Inorg. Chem. 1988,
36, 1-124. (b) Lohrenz, J. C. W.; Woo, T. K.; Fan, L.; Ziegler, T. J.
Organomet. Chem. 1995, 497, 91-104. (c) Cavallo, L.; Guerra, G.
Macromolecules 1996, 29, 2729-2737.
(17) (a) Busico, V.; Cipullo, R. J. Am. Chem. Soc. 1994, 116, 9329-9330. (b)
Busico, V.; Cipullo, R. J. Organomet. Chem. 1995, 497, 113-118.
(18) (a) Leclerc, M. K.; Brintzinger, H. H. J. Am. Chem. Soc. 1995, 117, 1651-
1652. (b) Leclerc, M. K.; Brintzinger, H. H. J. Am. Chem. Soc. 1996, 118,
9024-9032. (c) Busico, V.; Caporaso, L.; Cipullo, R.; Landriani, L.;
Angelini, G.; Margonelli, A.; Segre, A. L. J. Am. Chem. Soc. 1996, 118,
2105-2106. (d) Busico, V.; Brita, D.; Caporaso, L.; Cipullo, R.; Vacatello,
M. Macromolecules 1997, 30, 3971-3977.
(19) Yoder, J. C.; Bercaw, J. E. J. Am. Chem. Soc. 2002, 124, 2548-2555.
(20) For an alternative mechanism for chain-end epimerization, see: Resconi,
L.; Fait, A.; Piemontesi, F.; Colonnesi, M.; Rychlicki, H.; Zeigler, R.
Macromolecules 1995, 28, 6667-6676.
9
J. AM. CHEM. SOC. VOL. 128, NO. 10, 2006 3421

A R T I C L E S
Scheme 2
Harney et al.
Scheme 3
within a cationic zirconocene secondary alkyl complex at low
temperatures. Accordingly, these observations by us and the
Landis group are consistent with the view that it is not that one
necessarily needs to avoid â-hydride elimination in order to have
a living Ziegler-Natta polymerization within an early transition
metal-based system, but rather, it is essential that one avoids
chain-release after â-hydride elimination and prior to reinser-
tion.24
intermediate through a formal 2,1-reinsertion of the alkene, and
the need to still be able to mechanistically account for the
observation that switching of the coordinated alkene enantioface
can sometimes occur prior to reinsertion without dissociatively
losing the alkene to the sea of monomer that is present. Finally,
it must be mentioned that isomerizations involving â-hydride
eliminations, such as chain-end epimerization, do not appear
to be confined only to the polymerization of propene, as
extensive end-group analyses by Chien and co-workers21 have
provided evidence, obtained from metallocene-based 1-hexene
polymerizations, for the mechanistically related process of chain-
walking of the propagating center back along the polymer
backbone and within backbone substituents.
Apart from the few studies performed by the Schrock group,
there has been only one other experimental investigation of
â-hydride elimination or â-hydride-based isomerizations within
cationic early transition metal alkyl complexes that can serve
as models for the propagating centers within liVing Ziegler-
Natta polymerization systems. In this regard, we have previously
reported that dimethyl monocyclopentadienylzirconium aceta-
midinates of the general structure (η5-C5R5)ZrMe2[N(R1)C-
(Me)N(R2)] (1) can serve as precursors to the cationic species
{(η5-C5R5)ZrMe[N(R1)C(Me)N(R2)]}[B(C6F5)4] (2) that are
highly active initiators for the living polymerization of R-olefins,
and in the case of 2a, where R ) Me, R1 ) t-Bu, and R2 ) Et,
these polymerizations proceed in a stereospecific fashion as well,
according to Scheme 3.6 We have further used NMR
spectroscopy to observe that, after polymerization, when all
monomer has been consumed, the cationic zirconium end-groups
of the living polymer species can engage in chain-end-confined
chain-walking, albeit at a rate that is orders of magnitude slower
than propagation.22 It must also be noted that, while not directly
modeling a known living Ziegler-Natta polymerization system,
Landis and co-workers23 have similarly used NMR to directly
observe chain-end epimerization within cationic zirconocene
polypropenyl species and a s-Bu f n-Bu structural isomerization
During the course of our extensive mechanistic investigations
of living Ziegler-Natta polymerization by the class of initiators
encompassed by 2,6 we discovered that the readily available
mono- and dialkyl monocyclopentadienylzirconium acetamidi-
nates, 3 and 4, respectively, are remarkably stable in solution
and the solid state, even when the alkyl groups possess
â-hydrogens (see Chart 1).25 Indeed, this stability permitted the
first successful hydrozirconation of internal alkenes that provides
kinetically stable, zirconium-bound secondary (chiral) alkyl
groups that are not prone to undergo subsequent structural
isomerizations that are the hallmark of products derived from
zirconocene-based hydrozirconations.26,27 Furthermore, in re-
cently communicated preliminary studies,22 we made additional
use of this stability to develop well-defined cationic zirconium
complexes bearing alkyl substituents with â-hydrogens that
could serve as model compounds for probing chain-walking
within living polymers derived from 2. Now, in the present
work, we extend these studies to a wider range of mixed methyl,
alkyl complexes, i.e., R1 ) Me in 4, to further investigate the
structures and solution stabilities of the corresponding cationic
monoalkyl complexes 5 that can be prepared in high yield from
4 through chemoselective methyl group protonolysis employing
the borate, [PhNHMe2][B(C6F5)4]. Incorporation of isotopic
labels within n-propyl and isobutyl substituents of 5 through
the use of the aforementioned hydrozirconation methodology,
together with the synthesis and in-depth study of analogous
cationic hafnium complexes, including the isotopically labeled
isobutyl derivative 6 and the tert-butyl derivative 7, provides
additional insights regarding the nature of isomerizations
involving â-hydrogen eliminations within this class of group 4
compounds. Importantly, as will be presented, these investiga-
tions have produced some intriguing results related to the
question of what steric and electronic factors contribute to the
living character of polymerizations initiated by 2. In particular,
(24) Chain-walking is, in fact, highly competitive with propagation within the
living polymerization of R-olefins by the late transition metal initiators
described by Brookhart and co-workers.11
(21) Babu, G. N.; Newmark, R. A.; Chien, J. C. W. Macromolecules 1994, 27,
3383-3388.
(25) Keaton, R. J.; Koterwas, L. A.; Fettinger, J. C.; Sita, L. R. J. Am. Chem.
Soc. 2002, 124, 5932-5933.
(22) Harney, M. B.; Keaton, R. J.; Sita, L. R. J. Am. Chem. Soc. 2004, 126,
4536-4537
(26) Zhang, Y.; Keaton, R. J.; Sita, L. R. J. Am. Chem. Soc. 2003, 125, 8746-
(23) (a) Sillars, D. R.; Landis, C. R. J. Am. Chem. Soc. 2003, 125, 9894-9895.
(b) Landis, C. R.; Sillars, D. R.; Batterton, J. M. J. Am. Chem. Soc. 2004,
126, 8890-8891.
8747.
(27) Chirik, P. J.; Day, M. W.; Labinger, J. A.; Bercaw, J. E. J. Am. Chem.
Soc. 1999, 121, 10308-10317 and references therein.
9
3422 J. AM. CHEM. SOC. VOL. 128, NO. 10, 2006

Living Ziegler−Natta Polymerization by Early Transition Metals
A R T I C L E S
Chart 1
Scheme 4
displays the molecular structures of the isopropyl and tert-butyl
derivatives, 3c and 3g, respectively, along with those for the
bromo, alkyl analogues of 3e and 3f. Significantly, in all of the
structures studied so far, the alkyl substituent is always directed
toward the N-Et substituent of the acetamidinate fragment, as
depicted in Scheme 4. Furthermore, as illustrated by the structure
of the bromo analogue of 3f, the steric bulk of large alkyl
substituents appears to be accommodated by “tucking-in” part
of the alkyl fragment beneath the acetamidinate moiety. Surpris-
ingly, none of the structures display any unusual geometries as
a result of strong nonbonded steric interactions, and this includes
that of the tert-butyl derivative 3g. With respect to geometric
parameters, the only bond lengths of particular note are the
zirconium-nitrogen bonds, which will be discussed in more
detail shortly. Finally, in none of the crystal structures for chloro,
alkyl derivatives of 3 has there ever been observed any
interaction between the metal and the â-hydrogens of the alkyl
substituent.
In solution, 1H NMR spectroscopy reveals that all the chloro,
alkyl derivatives of 3 appear to exist as a single diastereomer,
which is assumed to be the same as that depicted in Scheme 4
and which is found in all the crystal structures obtained to date.
In addition, variable-temperature NMR provides evidence that
all of these compounds are configurationally stable on the NMR
time frame with respect to racemization through “amidinate ring-
flipping” 28 that occurs through metal-centered epimerization.
Finally, variable-temperature NMR was also used to verify that,
at elevated temperatures up to 80 °C, the isopropyl derivative
3c does not isomerize to the n-propyl derivative 3b, nor does
the tert-butyl derivative 3g isomerize to the isobutyl derivative
3e. This high degree of stability toward isomerization through
a presumed â-hydride elimination/reinsertion mechanism stands
in sharp contrast to the observations with zirconocene-based
secondary and tertiary chloro, alkyl complexes.27
(b) Synthesis of Mixed Methyl, Alkyl Complexes 4. As
Scheme 5 shows, methylation of 3 using methyllithium at -78
°C in Et2O provided the desired mixed methyl, alkyl complexes
4 in good to excellent isolated yields, depending upon their ease
of crystallization. The only case where this procedure failed
completely was in the attempted methylation of the tert-butyl
derivative 3g. In fact, all attempts to methylate 3g through a
variety of different reagents, e.g., MeLi, MeMgCl, ZnMe2, and
AlMe3, and under different conditions, failed to produce the
desired tert-butyl, methyl derivative of 4, presumably due to
the large degree of steric shielding of the metal center that the
tert-butyl group provides. Under forcing conditions (e.g.,
elevated temperatures or excess MeLi), only isomerization to
the isobutyl derivative 4e was observed, and curiously, switching
the order of reagents for the synthesis from the dichloride 8,
we present the conclusion that the presence of a strong
â-hydrogen agostic interaction does not necessarily go hand-
in-hand with a low barrier for â-hydride elimination/chain
release. Importantly, the results of this work also serve to help
delineate the extent to which the cationic alkyl complexes 5
can be used for accurately modeling the corresponding chemistry
of the true propagating species involved in living polymeriza-
tions initiated by 2a.
Results
(a) Synthesis of Mixed Halide, Alkyl Complexes 3. As
previously reported,25 monoalkylation of the dichloro starting
material 8 can be achieved in high yield using alkylmagnesium
chloride reagents, or alkyllithium reagents in the case of the
sterically demanding tert-butyl group, in diethyl ether (Et2O)
at low temperature according to Scheme 4. In the case of the
tert-butyl derivative 3g,25 it is imperative that much longer
reaction times (∼10 h) at -78 °C be used since the large amount
of tert-butyllithium still remaining at short reaction times serves
to isomerize the zirconium-bound tert-butyl group to an isobutyl
group upon warming the reaction mixture to room temperature
(cf. ∼100% yield of the isobutyl derivative 3e after 1 h vs a
9:1 ratio of 3g and 3e after 10 h). Fortunately, the small amount
of 3e that is still produced can be removed by fractional
crystallization to provide analytically pure 3g that is then
indefinitely stable in solution. Indeed, all the derivatives of 3
shown in Scheme 4 are stable in solution and in the solid state
at room temperature, thereby greatly facilitating their full
characterization by spectroscopic and analytical means.
As a side note, when alkylmagnesium bromide reagents are
used in Scheme 4, transhalogenation at zirconium occurs to
provide the monoalkylated product as a mixture of bromo and
chloro derivatives. Although this halide mixture can be used in
subsequent alkylation reactions, we have found it to be
inconvenient with respect to obtaining clean spectroscopic data
and unequivocal chemical analysis data. The one instance where
it has proven to be useful, however, is in obtaining crystal
structures through single-crystal X-ray analysis, where the bromo
derivatives are oftentimes more likely to produce high-quality
single crystals than the chloro derivatives.
Regarding crystal structures, a number have been obtained
for derivatives of 3, or as the corresponding bromide compound,
and some of these have been previously reported.22,25 Figure 1
(28) (a) Sita, L. R.; Babcock, J. R. Organometallics 1998, 17, 5228-5230. (b)
Koterwas, L A.; Fettinger, J. C.; Sita, L. R. Organometallics 1999, 18,
4183-4190.
9
J. AM. CHEM. SOC. VOL. 128, NO. 10, 2006 3423

A R T I C L E S
Harney et al.
Figure 1. Molecular structure (30% thermal ellipsoids) of (a) 3c, (b) 3g, (c) 3e (Br analogue), and (d) 3f (Br analogue) with partial atom labeling. Hydrogen
atoms have been removed for the sake of clarity.
Scheme 5
example, the Zr-N bond lengths of the chloro, isopropyl
compound 3c are 2.2419(11) Å for Zr(1)-N(2) and 2.2528-
(11) Å for Zr(1)-N(1), whereas the corresponding values in
the methyl, isopropyl derivative 4c are 2.2594(9) and 2.2808-
(9) Å, respectively. Although these Zr-N bond differences
between 3 and 4 would, at first glance, appear to be small, they
are likely the origin of the dramatic difference in solution
configurational stability displayed by the two classes of
i.e., methyllithium, followed by tert-butyllithium, also failed to
yield the desired mixed methyl, tert-butyl product.
compounds (vide infra).
In solution, 1H NMR spectra suggest that all derivatives of 4
are configurationally unstable with respect to amidinate ring-
flipping, and as depicted by Scheme 6, at room temperature,
they exist as a mixture of two interconverting diastereomers,
the relative ratio of which depends on the steric bulk of the
zirconium-bonded alkyl group. Indeed, for the isobutyl deriva-
tive 4e, it was recently determined, through the use of 2D 13C-
{1H} EXSY NMR, that metal-centered epimerization is a facile
process, even at -10 °C.6f It is not surprising, then, that this
dynamic process presumably allows for the major diastereomer
of 4 to be obtained as the single crystalline product in each
case. More importantly, it is most probable that the observed
difference in configurational stability between the chloro, alkyl
derivatives of 3 (stable) and the methyl, alkyl derivatives of 4
(unstable) is due to the relative electronic effects associated with
the chloro and alkyl substituents. In other words, given the
greater electronegativity of the chloro group, the zirconium-
nitrogen interactions in 3 are potentially stronger than in 4, and
these are manifested not only in shorter Zr-N bond distances
but also in a larger energy barrier for amidinate ring-flipping.
Finally, it can be noted that this difference in configurational
stability of 3 and 4 is more than just of cursory interest, in that
it can be used to direct stereocontrol during the polymerization
As with the mixed chloro, alkyl complexes 3, all the mixed
methyl, alkyl derivatives of 4 were found to be stable in solution
and in the solid state at room temperature, and all could be
recrystallized to an analytically pure state as determined by 1H
NMR spectroscopy. However, whereas all the derivatives of 3
gave satisfactory chemical analyses, the dialkyl compounds often
failed to provide similarly conclusive microanalysis data, even
after repeated attempts, presumably due to various decomposi-
tion pathways that are available at moderate temperatures.25,29
Accordingly, in those cases, single-crystal X-ray analysis was
performed to verify composition, and Figure 2 provides the
molecular structures obtained for some of the more interesting
examples. As can be seen, in keeping with the structures of
Figure 1, the structures for 4a,c-e all adopt a configuration
that places the alkyl substituent on the same side as the N-Et
group of the acetamidinate fragment, and once again, no
interactions of â-hydrogens with the metal center are observed
in the solid state. With respect to geometrical parameters, it is
important to note that between the chloro (bromo), alkyl
derivatives of 3 and the methyl, alkyl derivatives of 4, the latter
class uniformly has longer zirconium-nitrogen bonds. For
(29) Keaton, R. J.; Sita, L. R. Organometallics 2002, 21, 4315-4317.
9
3424 J. AM. CHEM. SOC. VOL. 128, NO. 10, 2006

Living Ziegler−Natta Polymerization by Early Transition Metals
A R T I C L E S
Figure 2. Molecular structure (30% thermal ellipsoids) of (a) 4a, (b) 4d, (c) 4c, and (d) 4e with partial atom labeling. Hydrogen atoms have been removed
for the sake of clarity.
Scheme 6
Scheme 8
Scheme 7
with 13C in the R-position, and finally, (4) doubly labeled 3e***
that has 13C in the R-position and D in the â-position. In the
case of 3b*, it is important to note that hydrozirconation does
not proceed in a highly stereoselective fashion, and therefore,
a ∼3:1 mixture of diastereomers (with respect to the relative
configuration of the metal and â-carbon centers) is produced.
As will be related shortly, this mixture of diastereomers for 3b*
has no consequences for the results obtained. It is also important
to note that variable-temperature 1H NMR was used to confirm
that no scrambling of the isotopic labels occurs in these
derivatives of 3b and 3e, even at elevated temperatures of up
to at least 80 °C. Finally, following the standard synthetic
protocol, the corresponding isotopically labeled methyl, alkyl
derivatives, 4b*, 4e*, 4e**, and 4e***, were obtained, and once
again, variable-temperature 1H NMR was used to confirm that
the isotopic labels remain intact within the alkyl substituents,
even at elevated temperatures.
(d) Synthesis of Cationic Complexes 5 via Chemoselective
Methyl Group Protonolysis. As depicted in Scheme 8, it was
discovered that a range of derivatives of 5 can be prepared in
virtually quantitative yield through chemoselective protonolysis
of the methyl group of 4 using 1 equiv of [PhNHMe2][B(C6F5)4]
when this reaction is conducted at -10 °C using chlorobenzene
as the solvent. To determine just how chemoselective this
process is, a 13C-methyl-labeled derivative of 4e was prepared
from 3e using 13CH3Li (99% 13C) as the alkylating reagent.
Protonolysis of this compound employing [PhNHMe2][B(C6F5)4]
of R-olefins by newly discovered chloride and methyl group
degenerative-transfer living Ziegler-Natta polymerizations
based on 1-3.6f,g
(c) Synthesis of Isotopically Labeled Derivatives of 3 and
4. To better investigate structural isomerizations involving
â-hydride elimination/reinsertion, it was desirable to have
isotopically labeled derivatives of 3 and 4 in which the isotopic
label could be placed at either an R- or â-position within an
alkyl substituent with known certainty. Fortunately, the newly
devised hydrozirconation chemistry shown in Scheme 7 pro-
vided a convenient route to several of these compounds.22,26
Thus, hydrogenolysis of the chloro, silyl complex 9 using either
H2 or D2 generates the chloro hydride (deuteride) zirconium
species, (η5-C5Me5)Zr(Cl)(X)[N(t-Bu)C(Me)N(Et)] (X ) H or
D), that in the presence of an excess of either propene or
[1-13C]2-methylpropene engages in hydrozirconation to provide
in high yield (1) the singly labeled n-propyl derivative 3b* that
has D in the â-position, (2) the isobutyl derivative 3e* that is
similarly singly labeled in the â-position, (3) singly labeled 3e**
9
J. AM. CHEM. SOC. VOL. 128, NO. 10, 2006 3425

A R T I C L E S
Harney et al.
to generate 5, in each case, only a single, configurationally stable
species for 5 is observed in solution. This configurational
stability with respect to metal-centered epimerization is in
keeping with the greater electron deficiency of the metal and,
hence, the shorter (stronger) Zr-N interactions with the
amidinate fragment (vide supra).
Given the â-hydrogen agostic interaction observed in the
solid-state structure of 5e, this compound served as a good
starting point to look for similar agostic interactions in solution.
Figure 4 presents the 1H NMR spectrum of the cation 5e
recorded in chlorobenzene-d5 at -10 °C, and the resonance of
particular note is that appearing at -0.27 ppm. Unequivocal
assignment of this resonance as being that of the â-hydrogen
1
of the isobutyl group was made through a 2D H-1H COSY
NMR experiment that also established that the two resonances
appearing at 0.77 and 1.59 ppm as a pair of doublet of doublets
(dd) are for the two diastereomeric R-hydrogens of this same
fragment. Although the significant upfield chemical shift for
the â-hydrogen is already suggestive of an agostic interaction
with the metal center,14 more compelling evidence was provided
by the value of 96 Hz for the associated 1J(13Câ-1Hâ) coupling
constant, which was conveniently obtained from a 2D J-resolved
13C-1H HSQC NMR spectrum. More specifically, the substan-
tial reduction in this coupling constant relative to a standard,
unperturbed 1J(13C-1H) value of 125 Hz signifies a weaker Câ-
Hâ interaction, as expected if the â-hydrogen is engaged in a
strong agostic interaction with the metal center.14 The 2D
J-resolved 13C-1H HSQC NMR method, however, also simul-
Figure 3. Molecular structure (30% thermal ellipsoids) of 5e with partial
atom labeling. The [B(C6F5)] anion and hydrogen atoms, except the
crystallographically located â-hydrogen, have been removed for the sake
of clarity.
was then carried out as described, and the resulting cationic
zirconium species was immediately used as an initiator for the
living polymerization of 1-hexene at -10 °C. GPC subsequently
confirmed that the molecular weight of the resulting isotactic
poly(1-hexene) was in agreement with quantitative generation
of a cationic initiator, and 13C NMR analysis of this polymeric
material confirmed the absence of any incorporation of the 13C-
methyl label into the end-groups, thereby confirming that only
the cationic isobutyl derivative 5e was initially produced by
the procedure of Scheme 8.6f
1
taneously provides J(13C-1H) values for the other protons of
the alkyl fragment, and in the case of 5e, those of interest are
for the R-hydrogens, with 1J(13CR-1HR) values of 128 and 138
Hz. Finally, a series of 1D NOE difference NMR spectra
confirmed that the solution structure of 5e is indeed the same
as that in the solid state, namely, that the â-hydrogen agostic
interaction occurs on the N-ethyl side of the amidinate, as shown
in Figure 4.
Although it has been exceedingly difficult to obtain single
crystals of derivatives of 5, in the case of the isobutyl derivative
5e, this feat could be reproducibly achieved at -30 °C by
layering pentane on top of a chlorobenzene solution, all within
an NMR tube (∼20 mg of 5e in 0.5 mL of chlorobenzene and
1 mL of pentane). Figure 3 presents the molecular structure of
5e as obtained by X-ray crystallography. Within this structure,
the borate counterion (not shown) is well displaced from the
zirconium cation, and there are no close Zr-F interactions
between the ion pair. Of immediate interest is the position of
the crystallographically located hydrogen atom, H(20A), that
puts it into a â-agostic interaction with the zirconium center
[cf. Zr(1)-H(20A) distance of 2.25(3) Å].30 Additional evidence
for this â-hydrogen agostic interaction is provided by the
observation that the â-carbon atom, C(20), is more pyramidal-
ized toward planarity than tetrahedral, as supported by the sum
of the bond angles for C(19)-C(20)-C(21), C(19)-C(20)-
C(22), and C(21)-C(20)-C(22), which is 337.7° (cf. 328° and
360° expected for idealized tetrahedral and planar geometries,
respectively). It is also interesting to note that the zirconium-
nitrogen bonds in the cation 5e are very short, at 2.1376(18)
and 2.1882(18) Å, and so it would appear that this geometric
parameter is indeed a good measure of the electron deficiency
of the metal center, as now established by the zirconium-
nitrogen bond length trend: 5e > 3e > 4e.6c
1
1
Table 1 provides solution J(13CR-1HR) and J(13Câ-1Hâ)
values for other derivatives of 5 that were obtained in a similar
fashion. Interestingly, except for the ethyl derivative 5a, for
which a â-hydrogen agostic interaction with the freely rotating
â-methyl group cannot be unequivocally established at -10 °C,
all the 1J(13Câ-1Hâ) values of Table 1 are supportive of a strong
â-hydrogen agostic interaction with the metal center. Further-
more, 1D NOE difference NMR confirms the solution structure
of all of these derivatives to be the same as that of 5e, in which
the â-hydrogen agostic interaction occurs on the N-ethyl side
of the amidinate fragment. It is also interesting to note that, in
the case of the isopropyl derivative 5c, only one of the
diastereomeric â-methyl groups is clearly engaged in a â-hy-
drogen agostic interaction, while the other is not.
(e) Synthesis of the Hafnium Complexes 6 and 7. We have
recently reported that the cationic methyl and isobutyl hafnium
analogues of 2a and 5e can also serve as initiators for the
isospecific living polymerization of 1-hexene, albeit with a rate
of propagation that is ∼60 times slower than that observed with
their second-row cousins.6h Given this, it was decided to expand
the range of the present studies to include a study of isomer-
ization within the isotopically labeled cationic hafnium isobutyl
derivative 6. Furthermore, considering all the previous failed
attempts to prepare a cationic tert-butyl derivative of 5, it was
In solution, all of the cations 5a-f are remarkably stable at
-10 °C, thus allowing their solution structures to be probed by
a variety of NMR techniques at this and other temperatures.
Importantly, although a mixture of diastereomers of 4 is used
(30) Jordan, R. F.; Bradley, P. K.; Baenziger, N. C.; LaPointe, R. E. J. Am.
Chem. Soc. 1990, 112, 1289-1291.
9
3426 J. AM. CHEM. SOC. VOL. 128, NO. 10, 2006

Living Ziegler−Natta Polymerization by Early Transition Metals
A R T I C L E S
Figure 4. 1H NMR (400 MHz, C6D5Cl, -10 °C) spectrum of 5e. The resonance marked with an asterisk is for the methyl groups of PhNMe2.
Table 1. 1J(13C-1H) Coupling Constants for the R Group in 5a
tert-butyllithium to provide the chloro, tert-butyl product 14
according to Scheme 10. Single-crystal X-ray analysis confirmed
the structure of 14, which is very similar to that previously
reported for the zirconium analogue 3g.25,32 However, attempts
to methylate 14 produced only the unlabeled derivative of the
methyl, isobutyl compound 12. Fortunately, reversing the order
of alkylation of 13 was successful. Thus, as Scheme 10 shows,
methylation of 13 using methylmagnesium chloride first pro-
vided the chloro, methyl compound 15, which was then reacted
with tert-butyllithium overnight at -55 °C to afford the elusive
methyl, tert-butyl complex 16. Finally, upon treatment of 16
with 1 equiv of [PhNHMe2][B(C6F5)4], the desired cationic tert-
butyl hafnium complex 7 was quantitatively generated, as
R
1J(13CR
−
1HR)b
1J(13Câ 1Hâ)b
−
Et (5a)
n-Pr (5b)
i-Pr (5c)
n-Bu (5d)
i-Bu (5e)
2-Et-Bu (5f)
140
144
135
133
123
110
126
121
n.o.
85
109
155
140
138
140
133
128
120
92
86
a Obtained from 2D 13C,1H HSQC NMR in chlorobenzene-d5 at -10
°C. b Two values denote that two distinct resonances for diastereotopic
protons were observed.
felt that better success might be achieved in efforts to synthesize
the corresponding cationic hafnium tert-butyl species 7.
To begin, the synthesis of doubly isotopically labeled 6 was
achieved in a fashion analogous to the synthetic protocol used
to prepare the zirconium analogue 5e***. Thus, as Scheme 9
reveals, hydrogenolysis of the chloro, silyl hafnium complex
10 under a D2 atmosphere resulted in the hydrohafnation31 of
[1-13C]2-methylpropene to provide the desired isotopically
labeled chloro, isobutyl hafnium compound 11, albeit in only
moderate yield as compared to the analogous hydrozirconation
reaction (cf. 70% yield for 3e*** vs 48% for 11). Methylation
of 11 using the standard procedure uneventfully provided the
isotopically labeled methyl, isobutyl hafnium compound 12.
Finally, treatment of 12 with 1 equiv of [PhNHMe2][B(C6F5)4]
in chlorobenzene at -10 °C provided the desired cationic species
1
determined by H NMR spectroscopy.
(f) Decomposition and Isomerization of Cationic Zirco-
nium and Hafnium Complexes. As mentioned previously, all
derivatives of cationic zirconium 5 appear to be extremely stable
at -10 °C, with no detectable amount of decomposition or
1
isomerization (where applicable) being observed by H NMR
after ∼18 h. At the higher temperature of 30 °C, however,
decomposition does ensue, and an initial screening qualitatively
determined a relative order of stability of 5 to be R ) Et >
n-Pr > n-Bu > i-Pr > i-Bu > 2-ethylbutyl. Furthermore, in
keeping with our past observations,3h,33 the main metal-
containing product isolated from these decompositions, and
oftentimes in high yield, is the dichloro dication species, {η5-
C5Me5)Zr(µ-Cl)[N(Et)C(Me)N(t-Bu)]}2[B(C6F5)4]2, shown in
Scheme 11, that presumably arises from chloride abstraction
from the chlorobenzene solvent and for which a crystal structure
has been obtained.34 Similar µ-chloro dicationic dizirconium
complexes have been previously reported to form in halogenated
1
6 in quantitative fashion as determined by H NMR. In this
respect, it is important to note that it was previously reported
that no eVidence for a â-hydrogen agostic interaction could be
found for unlabeled 6 in solution using 1D and 2D NMR
techniques.6h
Regarding the synthesis of 7, the first attempt involved
alkylation of the dichloro hafnium starting material 136h with
(32) Detailed information is provided in the Supporting Information.
(33) Zhang, Y.; Reeder, E. K.; Keaton, R. J.; Sita, L. R. Organometallics 2004,
23, 3512-3520.
(31) Erker, G.; Schlund, R.; Kru¨ger, C. Organometallics 1989, 8, 2349-2355.
9
J. AM. CHEM. SOC. VOL. 128, NO. 10, 2006 3427

A R T I C L E S
Scheme 9
Harney et al.
Scheme 10
decene) (LPD) sample at 0 °C, disappearance of the zirconium-
bonded R-carbon 13C resonance for this material at 82.3 ppm6e
was then followed by inverse-gated 13C NMR spectroscopy.
Once again, decomposition of this LPD was seen to occur in
first-order fashion at 0 °C, to provide a kd of 0.22 h-1 (t1/2
)
3.2 h), a Value that is more than 6 times greater than that of
the isobutyl deriVatiVe 5e.
Additional insight into the nature of the processes by which
5e decomposes was obtained by observing the course of
decomposition of the singly deuterium-labeled derivative 5e*
and the doubly isotopically labeled derivative 5e***. Thus,
through the use of 1H NMR, it was first determined that,
concurrent with decomposition, the original â-positioned deu-
terium label of 5e* begins to undergo scrambling within the
isobutyl group almost immediately upon warming from -10
to 0 °C, with this process reaching apparent equilibrium in
approximately 11 h. In the case of 5e***, when decomposition
and scrambling was followed by 13C NMR, 1:1:1 triplets for
two diastereotopic 13CH2D-labeled methyl groups of the isobutyl
substituent were observed, with no evidence being obtained for
the corresponding singlets that should arise from formation of
singly 13C-labeled methyl groups (see Figure 5). In other words,
the mechanism by which scrambling occurs always places a D
and 13C label on the same carbon atom, as shown in Scheme
12, and this is consistent with the multistep sequence of events
also shown in this scheme that is identical in nature to the
mechanism proposed by Busico and co-workers17 for chain-
end epimerization.
solvents starting from cationic zirconium methyl species.35 In
the present case, we believe the reactivity of the solvent with
any metal hydride intermediates that might be formed upon
â-hydride elimination in 5, thereby providing the dichloro
dication, may be precluding the ability to directly observe and
isolate these species.
On the basis of the preliminary screen, several derivatives of
5 were targeted for more in-depth analysis, with the subset being
comprised of 5b (R ) n-Pr), 5c (R ) i-Pr), 5e (R ) i-Bu), and
5f (R ) 2-ethylbutyl) (see Chart 2). Importantly, as presented
earlier, all of these cationic complexes appear to engage in a
strong â-hydrogen agostic interaction in solution as depicted.
At 0 °C, both the isobutyl and 2-ethylbutyl derivatives, 5e
and 5f, respectively, were observed to decompose in a first-
order fashion, while the n-propyl and isopropyl derivatives, 5b
and 5c, respectively, remained unchanged. Surprisingly, the
more sterically encumbered derivative, 5f, decomposed ap-
proximately twice as fast as the isobutyl derivative 5e (kd )
0.067 h-1; t1/2 ) 10.3 h vs 0.034 h-1; 20.4 h, respectively). To
put these absolute and relative rates of decomposition for 5e
and 5f into proper perspective with respect to the living polymer
species derived from 2a, for which 5e and 5d had been hoped
from the start to serve as viable models, the rate of decomposi-
tion for the living polymer derived from the polymerization of
1-decene was also determined at 0 °C at the same concentration
of cationic zirconium species. To achieve this, [1-13C]1-decene
(99% 13C) was first oligomerized in living fashion in chlo-
robenzene-d5 at -10 °C to a degree of polymerization (DP) of
∼15 by judiciously setting the ratio of the monomer concentra-
tion to that of the initiator 2a (DP ) [M]/[I] for living
polymerizations). After equilibration of the living poly(1-13C-
A few other comments regarding the isomerization/decom-
position of 5e*** are in order. First, as Figure 5 reveals, in
addition to the 2-methylpropene product expected from â-hy-
drogen elimination, formation of 2-methylpropane (isobutane)
was also observed to occur both initially and continuously
throughout decomposition. Initially, [1-13C,2-D]2-methylpropane
is predominately produced from the presumed protonolysis of
the unscrambled isobutyl fragment, but near the end of
decomposition/isomerization, increasing amounts of [1-13C,1-
D]2-methylpropane are generated from protonolysis of the
isomerized isobutyl group in 5e***, as observed in Figure 5.
Schrock and co-workers4g previously reported formation of
2-methylpropane during decomposition of their isobutyl hafnium
initiator, but they attributed this product to the reaction of the
initiator with dimethylaniline, which was present in their system
as in ours, being the byproduct of using [PhNHMe2][B(C6F5)4]
to initially form a cationic species through protonolysis. In the
present study, replacing this borate with [Ph3C][B(C6F5)4], which
can also be used to chemoselectively generate 5 from 4 through
methide abstraction, unfortunately did not obviate production
of 2-methylpropane during decomposition of 5e, nor did any
other attempts to remove obvious proton sources from the
equation (e.g., use of silylated glassware). Given the strict first-
order decomposition of 5e at different concentrations, we are
accordingly led to believe that formation of the saturated
(34) Full characterization and crystallographic analysis will be presented
elsewhere as part of a structural study of the series {(η5-C5Me5)Zr(µ-X)-
[N(Et)C(Me)N(t-Bu)]}2[B(C6F5)4]2 (X ) F, Cl, Br, Me).
(35) (a) Gomez, R.; Green, M. L. H.; Haggit, J. L. J. Chem. Soc., Dalton Trans.
1996, 939-946. (b) Vollmerhaus, R.; Rahim, M.; Tomaszewski, R.; Xin,
S.; Taylor, N. J.; Collins, S. Organometallics 2000, 19, 2161-2169.
9
3428 J. AM. CHEM. SOC. VOL. 128, NO. 10, 2006

Living Ziegler−Natta Polymerization by Early Transition Metals
A R T I C L E S
Figure 5. 13C{1H} NMR (100 MHz, C6D5Cl, 0 °C) of 5e*** after 15 h at 0 °C. 0, diastereotopic doubly labeled methyl groups (13CH2D) of the isobutyl
moiety; b, DC(CH3)2(13CH3); O, HC(CH3)2(13CH2D); 1, H2CdC(CH3)(13CH2D).
Scheme 11
Chart 2
Scheme 12
hydrocarbon coproduct of decomposition of this species is due
to a competitive first-order intramolecular process. As shown
in Scheme 13, the most likely decomposition path leading to
2-methylpropane from 5e involves intramolecular H-abstraction
from one of the methyl groups of the cyclopentadienyl ligand
to produce the cationic “tuck-in” metal species 17, which can
reasonably be assumed to be highly reactive. Indeed, much
precedent exists in the literature for a transformation such as
that shown.36 Efforts are currently underway to unequivocally
substantiate this decomposition pathway, as it clearly has
possible relevance to attempts to utilize the initiators 2 at
temperatures well above that which has previously been shown
9
J. AM. CHEM. SOC. VOL. 128, NO. 10, 2006 3429

A R T I C L E S
Scheme 13
Harney et al.
Chart 3
for a similar â-hydrogen agostic interaction being observed by
NMR techniques for the LPB. In addition to these chain-walking
studies with living polymeryl species, further efforts were made
to determine whether such species can engage in chain-end
epimerization, which in the case of living polymers derived from
higher R-olefins (i.e, other than propene) must occur through a
process involving enantiofacial switching of the coordinated
olefin within an alkene hydride intermediate. Thus, using the
isobutyl compound 5e as an initiator in order to place an isobutyl
group at one end of the polymer chain, isotactic LPD (DP ≈
15) was once again produced at -10 °C and allowed to gestate
at this temperature for 18 h, during which time aliquots were
periodically taken and acid-quenched to produce a 13C-labeled
methyl end-group where the metal had been attached. As it has
previously been determined by us that two distinct 13C NMR
resonances exist for the two possible diastereomers I and II
shown in Chart 3, if chain-end epimerization of the LPD were
occurring over time, then the 13C NMR spectra for the aliquots
should show the appearance and growth of a new 13C-labeled
methyl end-group resonance for the epimerized diastereomer
II over time. In practice, all aliquots provided 13C NMR spectra
with only a single 13C-labeled methyl end-group for diastereomer
I, even after 18 h of gestation. Accordingly, chain-end epimer-
ization through enantiofacial switching does not appear to be
possible for LPD, whereas chain-walking is. Finally, it must
be mentioned that warming a sample of LPD, initially prepared
at -10 °C, to room temperature results in the nearly quantitative
formation of a vinyl end-group through â-hydride elimination.
Although studies are still underway to quantify the amount of
saturated end-groups that might be produced through competitive
cyclopentadienyl ligand H-abstraction, at the present it appears
that this decomposition path does not compete with â-hydride
elimination in LPD as effectively as it does in the model
compounds 5e and 5f.
Since the tert-butyl derivative 5h has not yet yielded to
synthesis, it has not been possible to investigate the stability of
this species relative to that of the isobutyl cation 5e. However,
with both the n-propyl and isopropyl derivatives, 5b and 5c,
respectively, in hand, a similar investigation was conducted with
them. As stated previously, both 5b and 5c remained unchanged
over extended periods of time in chlorobenzene solution at 0
°C, a temperature at which both 5e and 5f undergo decomposi-
tion by â-hydride elimination and intramolecular H-abstraction.
At 5 °C, however, the isopropyl complex 5c was observed to
isomerize fully to the n-propyl derivative 5b in a first-order
fashion (at 5 °C, t1/2 ) 68 h; at 10 °C, t1/2 ) 33 h), with only
a trace amount of decomposition being observed. To determine
if this isomerization process is reversible, decomposition of the
deuterium-labeled n-propyl derivative 5b* at more elevated
temperatures was studied. Thus, at 20 °C, while 5b* was observe
to decompose in first-order fashion, the deuterium label in the
â-position of the n-propyl substituent did not scramble, and this
remained true at even higher temperatures of up to 50 °C.
Interestingly, at no temperature was propene observed, and thus,
to provide clear living character in the polymerizations of higher
R-olefins (e.g., -10 °C in the case of 1-hexene).
A second point regarding the isomerization/decomposition
of 5e is that, assuming â-hydride elimination to be the rate-
determining step for isomerization, 1,2-reinsertion to regenerate
unscrambled 5e*** and isomerization of the isobutyl group
involving a tert-butyl intermediate through a 2,1-reinsertion
according to Scheme 12 must both be very facile processes that
effectively compete extremely well with irreversible chain-
release from alkene hydride intermediates. Indeed, scrambling
of isotopic labels in 5e*** appears immediately by 13C NMR
at 0 °C at a point when minimal loss of compound (decomposi-
tion) has yet occurred. No NMR spectroscopic evidence for such
a cationic tert-butyl intermediate (i.e., the hitherto unattained
5g), however, has yet been obtained. More disappointingly, all
attempts to identify a reagent that could be used to chemically
trap the presumed but unobserved-as-of-yet alkene hydride
intermediates in order to eliminate the reversibility of the
â-hydride elimination process and isotopic scrambling, and
thereby permit kinetic analyses to obtain clear-cut thermody-
namic parameters, including a kinetic isotope effect, have failed
to produce a viable candidate as of yet.
In the case of the 2-ethylbutyl derivative 5f, a large-scale
(100 mg) decomposition at 0 °C provided a volatile hydrocarbon
product that contained not only the expected 2-ethyl-1-butene
from direct â-hydride elimination but also small amounts of
3-methyl-1-pentene and (E/Z)-3-methyl-2-pentene that more
detailed spectroscopic studies have revealed most likely arise
from chain-walking through a series of â-hydride elimination/
alkene rotation/reinsertion sequences that are identical in nature
to those of chain-end epimerization. As before, a significant
amount of the saturated hydrocarbon, 2-ethylbutane, was also
obtained through presumed competitive intramolecular H-
abstraction from a methyl group of the cyclopentadienyl ligand
according to Scheme 13.
As previously mentioned, end-group-confined chain-walking
has been directly observed by us in living cationic zirconium
poly(1-butenyl) (LPB) (DP ≈ 15) at -10 °C,22 a temperature
at which the isobutyl derivative 5e*** does not engage in isotope
label scrambling, and at which both 5e and the 2-ethylbutyl
derivative 5f are indefinitely stable. Thus, the conclusion reached
from these observations is that living polymers derived from
the polymerization of higher R-olefins (i.e., 1-butene, 1-hexene,
etc.) possess a lower barrier to â-hydride elimination, and hence
chain-walking, relative to the model compounds 5e and 5f,
despite evidence for strong â-hydrogen agostic interactions in
solution for the latter two in contrast to no evidence as of yet
(36) (a) Schock, L. E.; Brock, C. P.; Marks, T. J. Organometallics 1987, 6,
232-241. (b) Bulls, A. R.; Schaefer, W. P.; Serfas, M.; Bercaw, J. E.
Organometallics 1987, 6, 1219-1226. (c) Carney, M. J.; Walsh, P. J.;
Bergman, R. G. J. Am. Chem. Soc. 1990, 112, 6426-6428. (d) Horton, A.
D. Organometallics 1992, 11, 3271-3275. (e) Sun, Y.; Spence, R. E. v.
H.; Piers, W. E.; Parvez, M.; Yap, G. P. A. J. Am. Chem. Soc. 1997, 119,
5132-5143.
9
3430 J. AM. CHEM. SOC. VOL. 128, NO. 10, 2006

Living Ziegler−Natta Polymerization by Early Transition Metals
A R T I C L E S
Scheme 14
likely proceeds through H-abstraction from a methyl group of
the cyclopentadienyl ligand of 5 to liberate an alkane. We do
not yet know, however, how the steric and electronic factors of
the alkyl/polymeryl substituent in the cationic zirconium
complexes may either act in concert to promote this path as the
dominant mode of decomposition in some cases or, alternatively,
render it noncompetitive with â-hydride elimination in others.
For instance, with living polymeryl species derived from the
polymerization of higher R-olefins with 2a, intramolecular
H-abstraction from the cyclopentadienyl ligand does not appear
to be as competitive with â-hydride elimination, possibly as a
result of steric inhibition by the bulky polymeryl chain and
backbone substituents. On the other hand, as we have yet to
observe ethene, propene, or 1-butene in the decomposition of
5a, 5b, and 5d, it is possible that, for these linear n-alkane
substituents, intramolecular H-abstraction is now the dominant,
if not single, mode of decomposition. We are presently engaged
in the synthesis of 5 and living polymeryl species from 2a in
which deuterium labeling in the cyclopentadienyl methyl groups
should allow us both to unequivocally establish this mode of
decomposition and to quantify to what extent it occurs for the
living polymeryl species at higher temperatures.
it is possible that intramolecular H-abstraction from the cyclo-
pentadienyl ligand is now the dominant decomposition path
for 5b.
(g) Decomposition and Isomerization of Hafnium Deriva-
tives. Upon quantitative generation through protonolysis, the
cationic tert-butyl hafnium complex 7 was found to be unstable,
isomerizing to the unlabeled isobutyl derivative of 6 at tem-
peratures as low as -35 °C. At -10 °C, quantitative isomer-
ization of 7 to 6 according to Scheme 14 was complete within
1 h, and through an Eyring analysis (five temperatures between
-35 and 0 °C), the activation parameters for this process were
determined to be ∆Hq ) 16.4 ( 0.5 kcal mol1 and ∆Sq ) -3.0
( 1.0 eu.
Regarding decomposition of the unlabeled isobutyl deriva-
tive of 6, relative to the zirconium analogue 5e, the former
compound was surprisingly found to be more unstable than the
latter, and in fact, it slowly decomposes even at -35 °C. Once
again, the metal-containing product of decomposition is the
dichloro dicationic species, {(η5-C5Me5)Hf(µ-Cl)[N(Et)C(Me)N-
(t-Bu)]}2[B(C6F5)4]2, as previously reported.6h At 0 °C, the rate
Given the uncertainty that still exists regarding the nature
and extent to which the competitive H-abstraction reaction is
contributing to the decomposition of 5, quantitative kinetic
analyses of â-hydride elimination for the different derivatives
are not yet possible. However, general trends have been revealed
by the present studies, and they suggest the relative importance
of different steric and electronic effects on this process within
living polymerizations initiated by 2a. Perhaps of most surpris-
ing note was documentation of the relative stability of 5e > 5f
> LPD at 0 °C. All three species possess a similarly configured
â-substituted primary alkyl substituent, and therefore, from an
electronic perspective, they would all appear rather similar. The
observed decrease in stability on going from 5e to 5f to LPD,
however, does track with an increase in the size of both this
â-substituent and the entire alkyl substituent itself. There are
two possible ways in which a steric factor can play a key role
in determining stability within the series. First is the more
obvious fact that, after initial â-hydride elimination, an increase
in nonbonded interactions between the metal-ligand sphere and
the newly generated alkene fragment should promote irreversible
chain release of the latter. More problematic is the apparent
result that, while 5e and 5f can engage in a strong â-hydrogen
agostic interaction in solution, LPD cannot due to a similar
increase in nonbonded interactions between the polymeryl
substituent and the ligand sphere. At first glance, the association
of a strong â-hydrogen agostic interaction with an increase in
stability toward â-hydride elimination would appear contradic-
tory. However, as first proposed by Bercaw and co-workers,37
the ability to form a â-hydrogen agostic interaction with the
highly electrophilic metal center might actually provide for
ground-state stabilization. A significant barrier to â-hydride
elimination would then still be in keeping with a late transition
state in which substantial positive charge has developed on the
â-carbon.12c,37 Observation that LPB can engage in chain-
walking at -10 °C, a temperature at which 5e*** does not
engage in isotopic label scrambling, further supports a lower
of decomposition of 6 was measured to be 0.055 h-1 (t1/2
)
12.6 h), which can be compared to 0.034 h-1 (t1/2 ) 20.4 h) for
5e (vide supra). More interestingly, the â-hydrogen resonance
for unlabeled 6 appears as a well-defined multiplet at 1.55 ppm.
Thus, unlike the isobutyl zirconium complex 5e, the hafnium
analogue does not appear to possess a strong â-hydrogen agostic
interaction in solution. Further, in contrast to the zirconium
analogue 5e***, the doubly labeled hafnium isobutyl derivative
6 did not show any evidence of isotopic label scrambling at 0
°C. Finally, analysis of the decomposition products at this
temperature revealed that the quantities of 2-methylpropene and
2-methylpropane obtained were similar to those of the decom-
position products of 5e***, and accordingly, it appears that
intramolecular H-abstraction from the cyclopentadienyl ligand
in the isobutyl hafnium compound 6 is a competitive decom-
position path, as it is for the zirconium series.
Discussion
The results obtained regarding the modes of decomposition
for the cationic zirconium and hafnium alkyl complexes 5-7
allow us to perhaps paint a clearer picture of the steric and
electronic factors that are important for establishing the living
character of Ziegler-Natta polymerizations based on the class
of initiators encompassed by 2. This new insight will undoubt-
edly aid in efforts to design the next generation of initiators
that are capable of operating at much higher temperatures. At
the same time, however, when compared with results obtained
from direct observation of the living polymeryl species them-
selves, distinct limitations have also been exposed regarding
the ability of the “model” compounds to actually serve as
reliable models.
To begin, the present studies have served to uncover a
competitive intramolecular decomposition pathway that most
(37) Burger, B. J.; Thompson, M. E.; Cotter, W. D.; Bercaw, J. E. J. Am. Chem.
Soc. 1990, 112, 1566-1577.
9
J. AM. CHEM. SOC. VOL. 128, NO. 10, 2006 3431

A R T I C L E S
Harney et al.
barrier to â-hydride elimination for the living polymeryl species
relative to that for the isobutyl model complex. It must be borne
in mind, however, that the exact mechanism by which these
structural isomerizations occur is still not well understood, and
rather than proceeding through discrete alkene hydride inter-
mediates arising from â-hydride elimination, it is possible that
“metal-assisted” 1,2-hydride shifts are involved that occur in
concerted fashion similar to that observed for other metal-
assisted 1,n-hydride shifts and metal-assisted â-hydride transfer
to monomer mechanisms that are supported by recent theoretical
studies.14b,c,38 If so, the polymeryl species may actually be able
to access the transition state for one of these 1,2-hydride shifts
more readily in the absence of a strong â-hydrogen agostic
interaction.
Regarding the relative stabilities of the n-propyl and isopropyl
derivatives, 5b and 5c, respectively, it is not surprising that the
linear primary n-alkyl substituent is favored over the secondary
one.39,40 Although not yet quantified, it would also appear that
a lower barrier to â-hydride elimination is present in 5e and 5f
than in either 5b or 5c, even in the presence of strong
â-hydrogen agostic interactions in all four. This observation is
in keeping with a greater stabilization of a transition state with
substantial positive charge on the â-carbon that is provided by
the â-alkyl group on the alkyl substituent in the former pair.37
It also seems quite reasonable that, after â-hydride elimination
in 5e, strong nonbonded interactions between the 2-methylpro-
pene (isobutylene) fragment and the ligand sphere within an
alkene hydride complex promote chain-release, leading to
decomposition, whereas with 5c, these steric interactions are
greatly reduced with the coordinated propene within the alkene
hydride, thus providing for greater stability through reinsertion,
if indeed the 5c f 5b isomerization does proceed through formal
â-hydride elimination. A more seemingly inexplicable discovery
is that, while the n-propyl derivative 5b apparently does not
(cannot) isomerize to the secondary isopropyl derivative 5c, even
at temperatures as high as 50 °C, the isobutyl derivative 5e can
undergo facile isomerization at the low temperature of 0 °C
that presumably goes through intermediacy of what one would
imagine to be the more sterically encumbered tert-butyl complex
5h. In this regard, it is highly unfortunate that 5h has not yet
yielded to synthesis, as a crystal structure of this species should
be highly informative regarding the steric and electronic factors
that may serve to render this tertiary alkyl species more
accessible from 5e than the isopropyl complex 5c is from 5b.
In fact, theoretical calculations predict that such a cationic
zirconium tertiary alkyl complex actually possesses substantial
positive charge on the tertiary carbon center that is stabilized
by the three alkyl substituents.18
cousin. One possible explanation for why the hafnium species
does not appear to engage in a strong â-hydrogen agostic
interaction in solution, while 5e does, might be found in the
observation that bond lengths between ligand fragments and the
metal center within neutral dialkyl hafnium analogues of 4 are
significantly shorter than in the zirconium counterparts, presum-
ably due to the effects of lanthanide contraction.6h Thus, for
the unlabeled isobutyl derivative 6, an increase in steric
interactions may preclude formation of a ground-state stabilizing
â-hydrogen interaction, and solid-state structural studies are
continuing with the aim of either substantiating or disproving
this conjecture. Finally, it is interesting to note that the isobutyl
hafnium derivative 6 does not engage in isotopic label scram-
bling through structural isomerizations that formally involve the
tert-butyl derivative 7, and that this former species is clearly
more thermodynamically favored over the latter, which can
undergo isomerization with a modest barrier.
Conclusions
For as much success as has recently been reported regarding
the development of early transition metal-based initiators that
can effect the living Ziegler-Natta polymerization of ethene,
propene, and higher R-olefins, general knowledge of the steric
and electronic factors that contribute to rendering these systems
living is still lacking. The present extensive study of different
structural model complexes for the living system based on 2a
has served to delineate the extent to which these complexes
are actually useful for realistically modeling the living polymeryl
species. At one extreme, there is clear evidence to suggest that
new modes of decomposition, such as the postulated intramo-
lecular H-abstraction process, might be more relevant to only
the model complexes. On the other hand, the model complexes
have served to reveal some general trends that should prove
useful in either designing new initiators for high-temperature
living polymerization or making predictions regarding which
systems should be more intrinsically stable. With respect to this
latter point, results obtained with the linear primary alkane model
complexes suggest that these species are intrinsically more stable
toward â-hydride elimination than those possessing â-alkyl
substituents. When translated to practice, this conclusion then
supports the prediction that the living polymerization of ethene
by the class of initiators represented by 2 should be amenable
to much higher temperatures than the living polymerization of
either propene or higher R-olefins. Efforts to substantiate this
prediction are currently in progress.
Acknowledgment. Funding for this work was provided by
the NSF (CHE-0092493), for which we are grateful. We also
thank the ACS Division of Organic Chemistry for graduate
fellowships to both M.B.H. and R.J.K., sponsored by the Proctor
and Gamble Co.
Of final note and importance is the observation that, in
contrast to the results obtained by Schrock and co-workers, the
isobutyl hafnium analogue of 5e is less stable than its zirconium
Supporting Information Available: Experimental details,
including H NMR, of compounds 5a-d,f, 5b*, 5e*, 5e**,
5e***, 6, and 7, and details of crystallographic analyses of
compounds 3c,e-g, 4a,c-e, 5e, and 14 (PDF, CIF). This
material is available free of charge via the Internet at
http://pubs.acs.org.
(38) Yu, Z.-X.; Houk, K. N. Angew. Chem., Int. Ed. 2003, 42, 808-811.
(39) Coleman, J. P.; Hedgedus, L. S.; Norton, J. R.; Finke, R. G. Principles
and Applications of Organotransition Metal Chemistry; University Sci-
ence: Mill Valley, CA, 1987.
(40) It is interesting to note that the reverse trend appears true within model
complexes for the late transition metal living polymerization system studied
by Brookhart and co-workers, see: (a) Shultz, L. H.; Tempel, D. J.;
Brookhart, M. J. Am. Chem. Soc. 2001, 123, 11539-11555. (b) Leatherman,
M. D.; Svejda, S. A.; Johnson, L. K.; Brookhart, M. J. Am. Chem. Soc.
2003, 125, 3068-3081.
1
JA057866V
9
3432 J. AM. CHEM. SOC. VOL. 128, NO. 10, 2006