Zirconium sandwich complexes with η9 indenyl ligands: Well-defined precursors for zirconocene-mediated coupling reactions

Article abstract of DOI:10.1021/ja045072v

A family of isolable, well-defined bis-indenyl zirconium sandwich complexes, (η⁵-C₉H₅-1,3-R ₂)-(η⁹-C₉H₅-1,3-R ₂)Zr (R = silyl, alkyl), have been prepared by either alkane reductive elimination or alkali metal reduction of a suitable zirconium(IV) dihalide precursor. Crystallographic characterization of two of these derivatives, R = SiMe₂CMe₃ and CHMe₂, reveals unprecedented η⁹ coordination of one of the indenyl ligands. Variable-temperature and EXSY NMR studies establish that the η⁵ and η⁹ rings are rapidly interconverting in solution. The sandwich complexes serve as effective sources of low-valent zirconium reacting rapidly with both olefins and alkynes at ambient temperature. In contrast to bis-cyclopentadienyl chemistry, the olefin adducts of the bis-indenyl zirconium sandwiches undergo preferential C-H activation to yield the corresponding allyl hydride compounds, although reaction with excess olefin proceeds through the η²-olefin adduct, forming the corresponding zirconacyclopentane.

Full text of DOI:10.1021/ja045072v

Published on Web 12/03/2004
Zirconium Sandwich Complexes with η9 Indenyl Ligands:
Well-Defined Precursors for Zirconocene-Mediated
Coupling Reactions
Christopher A. Bradley,† Ivan Keresztes,† Emil Lobkovsky,† Victor G. Young,‡ and
Paul J. Chirik*,†
Contribution from the Department of Chemistry and Chemical Biology, Baker Laboratory,
Cornell UniVersity, Ithaca, New York 14853, and Department of Chemistry,
UniVersity of Minnesota, Minneapolis, Minnesota 55455
Received August 16, 2004; E-mail: pc92@cornell.edu
Abstract: A family of isolable, well-defined bis-indenyl zirconium sandwich complexes, (η5-C9H5-1,3-R2)-
(η9-C9H5-1,3-R2)Zr (R ) silyl, alkyl), have been prepared by either alkane reductive elimination or alkali
metal reduction of a suitable zirconium(IV) dihalide precursor. Crystallographic characterization of two of
these derivatives, R ) SiMe2CMe3 and CHMe2, reveals unprecedented η9 coordination of one of the indenyl
ligands. Variable-temperature and EXSY NMR studies establish that the η5 and η9 rings are rapidly
interconverting in solution. The sandwich complexes serve as effective sources of low-valent zirconium
reacting rapidly with both olefins and alkynes at ambient temperature. In contrast to bis-cyclopentadienyl
chemistry, the olefin adducts of the bis-indenyl zirconium sandwiches undergo preferential C-H activation
to yield the corresponding allyl hydride compounds, although reaction with excess olefin proceeds through
the η2-olefin adduct, forming the corresponding zirconacyclopentane.
Introduction
molecule is quite reactive, promoting activation of dinitrogen
to form [(η5-C5Me5)2Ti]2(µ2,η1,η1-N2) as well as participating
Transition metal sandwich compounds, more commonly
referred to as metallocenes, occupy a central role, both histori-
cally and practically, in organotransition metal chemistry.1
Specifically, group 4 metal sandwich complexes are of long-
standing interest given their indispensable role in organic
coupling reactions2 and small molecule activation processes.3
Despite this utility, there is a paucity of isolable or even
spectroscopically characterized low-valent group 4 transition
metal sandwich complexes. Attempts to prepare [(η5-C5H5)2-
Ti] by either direct synthesis4 or reduction of a suitable halide
precursor leads to the bridging titanium fulvalene hydride [(η5-
C5H5)Ti]2(µ,η5-C10H8)(µ-H)2, arising from activation of cyclo-
pentadienyl C-H bonds.5,6 However, the putative titanium
sandwich can be trapped using a sufficiently π-accepting ligand
such as a phosphine,7 alkyne,8 carbon monoxide,9 or dinitro-
gen.10 Introduction of methyl groups onto the cyclopentadienyl
ligand provides steric protection of the titanium center, allowing
observation of the titanium sandwich, [(η5-C5Me5)2Ti].11 This
in competitive cyclometalation of the cyclopentadienyl methyl
groups to yield (η5-C5Me5)(η5-C5Me4-η1-CH2)TiH.12 Likewise,
our laboratory has observed a transient, silylated titanium
sandwich (η5-C5H3-1,3-(SiMe3)2)2Ti that reacts rapidly with N2
to afford the end-on dinitrogen complex [(η5-C5H3-1,3-(SiMe3)2)2-
Ti]2(µ2,η1,η1-N2).13 This molecule serves as a useful synthon
for the preparation of monomeric imido complexes that par-
ticipate in nitrene group transfer processes.
Crystallographic characterization of an isolable titanium
sandwich (η5-C5Me4SiMe2CMe3)2Ti was recently achieved by
Lawless by reduction of the corresponding titanium(III) monochlo-
ride.14 Independently, Mach and co-workers prepared several
other related examples of the general form (η5-C5Me4R)2Ti
(R ) SiMe3, SiMe2Ph, SiMePh2, SiMe2CH2CH2Ph).15,16 In-situ
(7) (a) PF3: Sikora, D. J.; Rausch, M. D.; Rogers, R. D.; Atwood, J. L. J. Am.
Chem. Soc. 1981, 103, 1265. (b) PMe3: Kool, L. B.; Rausch, M. D.; Alt,
H. G.; Herberhold, M.; Thewalt, U.; Honold, B. J. Organomet. Chem. 1986,
310, 27. (c) Me2PCH2CH2PMe2: Girolami, G. S.; Wilkinson, G.; Thor-
tonpett, M.; Hursthouse, M. B. J. Chem. Soc., Dalton Trans. 1984, 2347.
(8) Rosenthal, U.; Burlakov, V. V.; Arndt, P.; Baumann, W.; Spannenberg,
A. Organometallics 2003, 22, 884.
† Cornell University.
‡ University of Minnesota.
(9) Sikora, D. J.; Macomber, D. W.; Rausch, M. D. AdV. Organomet. Chem.
(1) For leading reviews on metallocene chemistry, see: The Metallocenes:
Synthesis, ReactiVity and Applications; Togni, A., Halterman, R. L., Eds.;
Wiley-VCH: Weinheim, 1998.
1986, 25, 317.
(10) Berry, D. H.; Procopio, L. J.; Carroll, P. J. Organometallics 1988, 7, 570.
(11) Bercaw, J. E.; Brintzinger, H. H. J. Am. Chem. Soc. 1971, 93, 2045.
(12) Bercaw, J. E. J. Am. Chem. Soc. 1974, 96, 5087.
(2) (a) Negishi, E.-I.; Takahashi, T. Acc. Chem. Res. 1994, 27, 124. (b)
Rosenthal, U.; Pellny, P.-M.; Kirchbauer, F. G.; Burlakov, V. V. Acc. Chem.
Res. 2000, 33, 119.
(3) (a) Fryzuk, M. D. Chem. Rec. 2003, 3, 2. (b) Fryzuk, M. D.; Johnson, S.
A. Coord. Chem. ReV. 2000, 200, 379.
(4) Fischer, A. K.; Wilkinson, G. J. Inorg. Nucl. Chem. 1956, 2, 149.
(5) Brintzinger, H. H.; Bercaw, J. E. J. Am. Chem. Soc. 1970, 92, 6182.
(6) Troyanov, S. I.; Antropiusova, H.; Mach, K. J. Organomet. Chem. 1992,
427, 49.
(13) Hanna, T. E.; Keresztes, I.; Lobkovsky, E.; Bernskoetter, W. H.; Chirik,
P. J. Organometallics 2004, 23, 3448.
(14) Hitchcock, P. B.; Kerton, F.; Lawless, G. A. J. Am. Chem. Soc. 1998, 120,
10264.
(15) Lukesova´, L.; Hora´cek, M.; Stepnicka, P.; Fejfarova´, K.; Gyepes, R.;
Cisorova´, I.; Kubista, J.; Mach, K. J. Organomet. Chem. 2002, 663, 134.
(16) Hora´cek, M.; Kupfer, V.; Thewalt, U.; Stepnicka, P.; Pola´sek, M.; Mach,
K. Organometallics 1999, 18, 3572.
9
10.1021/ja045072v CCC: $27.50 © 2004 American Chemical Society
J. AM. CHEM. SOC. 2004, 126, 16937-16950
16937

A R T I C L E S
Bradley et al.
infrared spectroscopic studies recently conducted in our labora-
tory have shown that these molecules coordinate two equivalents
of N2 at low temperature, forming monomeric titanium bis-
dinitrogen complexes that are isoelectronic and isostructural with
more familiar dicarbonyl derivatives.17
from the corresponding bis-indenyl zirconocene alkyl hydrides.
The indenyl anion, C9H7-, is related to the ubiquitous cyclo-
pentadienyl ligand with the potential to coordinate to a transition
metal as a monoanionic, five-electron, L2X-type ligand.27
However, most indenyl complexes show distinct η3,η2 coordina-
tion, where two of the five metal-carbon bonds are significantly
longer than the other three.28 Replacement of cyclopentadienyl
with indenyl ligands often leads to enhanced or unique reactiv-
ity.29 This behavior, termed the “indenyl effect”, is most
frequently observed in associative substitution reactions, where
indenyl complexes undergo faster ligand exchange as a result
of the indenyl ring slippage from η5 to η3 coordination.30
Indenyl anions are common ancillary ligands in group 4
metallocene chemistry, owing to their role in stereospecific
olefin polymerization31 and catalytic bond-forming reactions.32
In some cases, enhanced catalytic activity is observed for the
indenyl complex as compared to the cyclopentadienyl com-
pound.33 We recently communicated the effect of indenyl ligand
substitution on the outcome of alkane reductive elimination from
group 4 metallocene alkyl hydride complexes. Specifically, we
discovered that the silylated bis-indenyl zirconocene isobutyl
hydride (η5-C9H5-1,3-(SiMe3)2)2Zr(CH2CHMe2)H undergoes
facile extrusion of isobutane at ambient temperature to yield
the zirconium sandwich complex (η5-C9H5-1,3-(SiMe3)2)-
(η9-C9H5-1,3-(SiMe3)2)Zr and crystallographically characterized
its η5,η6-bis(indenyl)zirconium-THF adduct.34 Here we describe
a detailed study of bis-indenyl zirconium complexes including
complete solution spectroscopic and solid-state characterization.
X-ray crystallographic studies on two examples have established
an unprecedented η9 indenyl-metal interaction that masks the
formally low-valent zirconium center. This family of molecules
promotes carbon-carbon bond-forming reactions with olefins
and alkynes, providing isolable, well-characterized species for
“zirconocene”-mediated coupling reactions.
Sandwich complexes of zirconium and hafnium are expected
to be more challenging to synthesize owing to the greater
thermodynamic potential for these metals to attain their highest
oxidation state. While a bis-cyclopentadienyl zirconium complex
[(η5-C5H5)2ZrII] has not been isolated or even experimentally
demonstrated, Negishi18 developed a practical route to a
functional equivalent of the molecule by addition of two
equivalents of n-BuLi to zirconocene dichloride at low tem-
perature. Detailed spectroscopic studies on the in-situ-generated
“zirconocene” reagent revealed a complex mixture of Zr(III)
and Zr(IV) compounds rather than a discrete zirconium(II)
sandwich.19 Nonetheless, this procedure is effective for the
coupling of numerous unsaturated organic molecules20 and has
found widespread utility in organic synthesis21 and materials
chemistry.22 Attempts to protect the zirconium center with
substituted cyclopentadienyl ligands produces more well-defined
chemistry, although observation of a sandwich has not been
realized. For example, reduction of (η5-C5Me5)2ZrCl2 with
sodium amalgam induces dinitrogen activation to form the end-
on dinitrogen complex [(η5-C5Me5)2Zr(η1-N2)]2(µ2,η1,η1-N2).23
Analogous chemistry was also observed with the corresponding
permethylhafnocene.24
Attempts to prepare zirconium sandwich complexes have also
resulted in intramolecular C-H activation of cyclopentadienyl
substituents. Reduction of (η5-C5Me4SiMe3)2ZrCl2 with Mg at
elevated temperature affords a mixture of products, one of which
has been identified as the cyclometalated zirconocene hydride,
(η5-C5Me4SiMe3)(η5-C5Me4-η1-SiMe2CH2)ZrH.25 This product
is postulated to arise from oxidative addition of a [SiMe3] C-H
bond to a transient zirconium(II) sandwich. Our laboratory has
been studying the chemistry of low-valent silylated bis-
cyclopentadienyl zirconium(IV) compounds and discovered that
alkane reductive elimination may be used to prepare the reactive
cyclometalated hydride (η5-C5H3-1,3-(SiMe3)2)(η5-C5H3-η1-1-
SiMe2CH2-3-SiMe3)ZrH.26 This complex, when exposed to 1
atm of dinitrogen, forms the side-on-bound dinitrogen complex
[(η5-C5H3-1,3-(SiMe3)2)2Zr]2(µ2,η2,η2-N2).26
Results and Discussion
Synthesis and Observation of Bis-indenyl Zirconium
Isobutyl Hydride Complexes. We recently reported both the
steric and electronic properties of a family of disubstituted
indenyl ligands by evaluation of the structural and spectroscopic
properties of the resulting zirconocene dichloride and dicarbonyl
compounds in addition to the related iron(II) sandwiches.35 In
agreement with the trend established in cyclopentadienyl
chemistry,26 silylated indenes impart more electrophilic zirco-
Provided these results, we became interested in studying both
the mechanism and products of alkane reductive elimination
(17) Hanna, T. E.; Lobkovsky, E.; Chirik, P. J. J. Am. Chem. Soc. 2004, 126,
14688.
(27) For a discussion of this nomenclature, see: Crabtree, R. H. The Organo-
metallic Chemistry of the Transition Metals, 2nd ed.; John Wiley: New
York, 1994; pp 24-38.
(18) Negishi, E.-I.; Cederbaum, F. E.; Takahashi, T. Tetrahedron Lett. 1986,
27, 2829.
(28) (a) Calhorda, M. J.; Verios, L. F. Comm. Inorg. Chem. 2001, 22, 375. (b)
Zargarian, D. Coord. Chem. ReV. 2002, 233-234, 157. (c) Trnka, T. M.;
Bonanno, J. B.; Bridgewater, B. M.; Parkin, G. Organometallics 2001, 20,
3255.
(19) Dioumaev, V. K.; Harrod, J. F. Organometallics 1997, 16, 1452.
(20) Watt, G. W.; Drummond, F. O. J. Am. Chem. Soc. 1970, 92, 826.
(21) For recent examples, see: (a) Chinkov, N.; Majumdar, S.; Marek, I. J.
Am. Chem. Soc. 2003, 125, 13258. (b) Urabe, H.; Suzuki, D.; Sasaki, M.;
Sato, F. J. Am. Chem. Soc. 2003, 125, 4036. (c) Paquette, L. A.; Kim, I.
H.; Cuniere, N. Org. Lett. 2003, 5, 221. (d) Takahashi, T.; Li, Y.; Ito, T.;
Xu, F.; Nakajima, K.; Liu, Y. J. Am. Chem. Soc. 2002, 124, 1144.
(22) (a) Johnson, S. A.; Liu, F.-Q.; Suh, M. C.; Zu¨rcher, S.; Haufe, M.; Mao,
S. S. H.; Tilley, T. D. J. Am. Chem. Soc. 2003, 125, 4199. (b) Lucht, B.
L.; Buretea, M. A.; Tilley, T. D. Organometallics 2000, 19, 3469. (c)
Takahashi, T.; Kotora, M.; Hara, R.; Xi, Z. Bull. Chem. Soc. Jpn. 1999,
72, 2591. (d) Lucht, B. L.; Mao, S. S. H.; Tilley, T. D. J. Am. Chem. Soc.
1998, 120, 4354.
(23) Manriquez, J. M.; Bercaw, J. E. J. Am. Chem. Soc. 1974, 96, 6229.
(24) Roddick, D. M.; Fryzuk, M. D.; Seidler, P. F.; Hillhouse, G. L.; Bercaw,
J. E. Organometallics 1985, 4, 97.
(25) Hora´cek, M.; Stepnicka, P.; Kubista, J.; Fejfarova, K.; Gyepes, R.; Mach,
K. Organometallics 2003, 22, 861.
(26) Pool, J. A.; Lobkovsky, E.; Chirik, P. J. J. Am. Chem. Soc. 2003, 125,
2241.
(29) (a) Calhorda, M. J.; Romao, C. C.; Verios, L. F. Chem. Eur. J. 2002, 8,
868. (b) Marder, R. B.; Roe, D. C.; Milstein, D. Organometallics 1988, 7,
1451.
(30) (a) Rerek, M. E.; Basolo, F. J. Am. Chem. Soc. 1984, 106, 5908. (b) Kakkar,
A. K.; Taylor, N. J.; Marder, T. J.; Shen, N.; Hallinan, N.; Basolo, F. Inorg.
Chim. Acta 1992, 198-200, 219. (c) Wescott, S. A.; Kakkar, A. K.; Stringer,
G.; Taylor, N. J.; Marder, T. B. J. Organomet. Chem. 1990, 394, 777.
(31) Coates, G. W. Chem. ReV. 2000, 100, 1223.
(32) Hoveyda, A. H.; Morken, J. P. In The Metallocenes: Synthesis, ReactiVity
and Applications; Togni, A., Haltermann, R. L., Eds.; Wiley-VCH:
Weinheim, 1998; Vol. 2, Chapter 8.
(33) Okuda, J.; Ko¨nig, P.; Rushkin, I. L.; Kang, H.-C.; Massa, W. J. Organomet.
Chem. 1995, 501, 37.
(34) Bradley, C. A.; Lobkovsky, E.; Chirik, P. J. J. Am. Chem. Soc. 2003, 125,
8810.
(35) Bradley, C. A.; Flores-Torres, S.; Lobkovsky, E.; Abrun˜a, H. D.; Chirik,
P. J. Organometallics 2004, 23, 5332.
9
16938 J. AM. CHEM. SOC. VOL. 126, NO. 51, 2004

Zirconium Sandwich Complexes
A R T I C L E S
tions from a resonance structure where the chloride acts as a
π-donor, coordinatively saturating the zirconium center.37
Treatment of the other zirconocene dichloride complexes with
isobutyllithium yielded analogous chemistry. Alkylation of the
mixture of rac and meso diastereomers of 4-Cl2 produced three
isomeric isobutyl hydride complexes in approximately equal
amounts. Two of the isomers arise from the two distinct meso-
zirconocene isobutyl hydrides, while the other derives from the
one possible isomer of the rac compound. In the case of 7-Cl2,
reaction with two equivalents of isobutyllithium afforded a 70:
30 mixture of isomeric isobutyl hydrides.
Each of the bis-indenyl zirconocene isobutyl hydrides un-
dergoes reductive elimination of isobutane at ambient temper-
ature (vide infra). Consistent with the electron-donating prop-
erties of alkyl substituents and cooperative steric effects,26 the
isopropyl-substituted compound, 5-(CH2CHMe2)H, undergoes
the slowest reductive elimination in the series, allowing
observation of the isobutyl hydride for several hours at 22 °C.
In contrast, the silylated zirconocene isobutyl hydride complexes
undergo rapid reductive elimination, most likely a result of
electronic destabilization of the Zr(IV) ground state.38 Despite
the complications arising from swift reductive elimination, most
of the zirconocene isobutyl hydrides have been characterized
in situ by NMR spectroscopy. Each alkyl hydride was generated
in near quantitative yield and exhibited either Cs or C1 symmetry,
depending on the indenyl substitution pattern. Moreover, each
spectrum contained diagnostic upfield resonances in the vicinity
of 0 ppm assigned to the diastereotopic protons on the methylene
carbon bound to the zirconium (see Experimental Section and
Supporting Information).
Figure 1. Zirconocene dichloride precursors and their shorthand designa-
tions.
nium centers than their alkylated counterparts. This family of
zirconocene dichloride complexes serves as the foundation for
the present study (Figure 1). As we noted in our prior report,35
the zirconocene dichloride 4-Cl2 was isolated as a near
equimolar mixture of rac and meso diastereomers that were not
separated for subsequent experiments. In this study, we also
included two new “mixed-ring” zirconocene dichlorides 8-Cl2
and 9-Cl2 (Figure 1).
Synthesis of Homoleptic Bis-indenyl Zirconium Sandwich
Complexes. As we previously communicated,34 reductive
elimination of isobutane from 1-(CH2CHMe2)H proceeds
rapidly at 22 °C and affords a burgundy complex, 1 (eq 2). A
Synthesis of the desired zirconocene isobutyl hydride com-
plexes was accomplished by addition of two equivalents of
isobutyllithium to the corresponding zirconocene dichloride (eq
1). Attempts to alkylate 1-Cl2 with two equivalents of tert-
combination of elemental analysis, solution molecular weight
determinations, and reactivity studies with HCl, CO and alkynes
established the empirical formula of the molecule as (C9H5-
1,3-(SiMe3)2)2Zr, consistent with the bis-indenyl zirconium
sandwich complex. However, both 1H and 13C NMR spectros-
copy indicated a molecule of Cs symmetry with inequivalent
indenyl ligands rather than the C2V-symmetric zirconium
sandwich with η5,η5-indenyl rings. While structural elucidation
of 1 was frustrated by poor crystal quality, addition of THF
afforded large red blocks suitable for X-ray diffraction. The
molecular structure established the identity of the complex as
(η5-C9H5-1,3-(SiMe3)2)(η6-C9H5-1,3-(SiMe3)2)Zr(THF) (1-THF),
where one of the indenyl ligands had undergone net haptotropic
rearrangement.39,40 Importantly, the NMR spectra of 1 and
butyllithium26,36 produced an intractable mixture of products.
While definitive mechanistic data has not been obtained, isobutyl
hydride formation with LiCH2CHMe2 most likely proceeds by
initial hydride transfer to the zirconium liberating isobutene,
followed by alkylation of the intermediate zirconocene hydrido
chloride. The olefin byproduct was detected and quantified by
1
a combination of H NMR spectroscopy and GC-MS experi-
ments. Initial alkylation to form a zirconocene isobutyl chloride
followed by â-hydrogen elimination is unlikely as preparation
of 1-(CH2CHMe2)Cl has been accomplished and subsequent
â-hydrogen elimination was not observed. These results are in
agreement with previous studies of bis-cyclopentadienyl zirco-
nium primary alkyl chloride complexes that are resistant to
â-hydrogen elimination, most likely due to important contribu-
(37) Chirik, P. J.; Day, M. W.; Labinger, J. A.; Bercaw. J. E. J. Am. Chem.
Soc. 1999, 121, 10308.
(38) Pool, J. A.; Lobkovsky, E.; Chirik, P. J. Organometallics 2003, 22, 2797.
(39) Albright, T. A.; Hofmann, P.; Hoffmann, R.; Lillya, C. P.; Dobosh, P. A.
J. Am. Chem. Soc. 1983, 105, 3396.
(40) Opruneko, Y. F.; Akhmedov, N. G.; Laikov, D. N.; Malyugina, S. G.;
Mstislavsky, V. I.; Roznyatovsky, V. A.; Ustynyuk, Y. A.; Utsynyuk, N.
A. J. Organomet. Chem. 1999, 583, 136.
(36) Pool, J. A.; Bradley, C. A.; Chirik, P. J. Organometallics 2002, 21, 1271.
9
J. AM. CHEM. SOC. VOL. 126, NO. 51, 2004 16939

A R T I C L E S
Bradley et al.
one of which is presented in Figure 3. The other molecule
(Supporting Information) is positionally disordered and was
successfully modeled with two independent molecules, with an
87:13 occupancy of the two sites. The disorder is such that a
180° rotation along the common coaxial direction of the two
indenyl ligands brings all atoms, except the zirconium and four
carbons, into complete overlap. A similar recrystallization
procedure was used to obtain single crystals of 5. The solid-
state structure of this sandwich complex is presented in Figure
4.
The solid-state structures of both 3 and 5 established the
identity of the molecules as the bis-indenyl zirconium sandwich
complexes where one of the rings is engaged in an η9 interaction
with the zirconium. Selected bond distances and angles for both
compounds are presented in Table 1. While the hapticity of the
indenyl ligands in 3 and 5 are similar, the ring orientations are
significantly different. In the more sterically hindered sandwich
3 the indenyl rings are staggered with a rotational angle of
177.8(4)°. This conformation contrasts the eclipsed geometry
observed in the corresponding zirconium dicarbonyl complex.35
In the solid-state structure of 5, the ligands adopt a near gauche
arrangement with a rotational angle of 84.6(2)°, in analogy to
the solid-state orientations of most of the bis-indenyl zirconium
dichloride precursors.35 In both cases, the substituents orient
such that the large groups point away from the interior of the
sandwich. For 3, the Si-tBu groups are directed above and below
the planes of the rings, while in 5, the isopropyl methine
hydrogens are pointed “in” toward the zirconium. Though the
conformations of these two sandwiches appear to minimize
transannular steric interactions of the indenyl substituents, the
role of crystal packing effects and kinetic preferences associated
with crystallization cannot be excluded. As expected for a metal
sandwich complex, the angle between the centroid of the η5
and the centroid to the midpoint of the two bridging carbons
on the η9 ligand approaches 180°. For 3, this angle deviates
only 0.8(2)° from linearity, while a more pronounced bend of
6.2(2)° is observed for 5.
As can be seen from the data compiled in Table 1, the
zirconium-carbon distances for the six-membered ring of the
η9 indenyl ligand in both 3 and 5 range between 2.255 and 2.520
Å, within the range typically ascribed to zirconium-carbon
bonds in π-complexed ligands. For comparison, the correspond-
ing distances to the six-membered ring on the η5 ligand range
between 3.416 and 4.185 Å, outside the range of a bonding
interaction. Coordination of the benzo portion of the η9 indenyl
ligand breaks the planarity of the ring and shortens the bonds
between the zirconium and the bridgehead carbon atoms. The
planes of the five- and six-membered rings on the η9 ligand are
buckled by 37.1(3)° and 37.6(1)° in 3 and 5, respectively.
Likewise, the centroids of the five- and six-membered rings form
angles of 60.7(6)° and 61.3(4)° with the zirconium center.
Relatively short Zr-C(7) and Zr-C(8) distances of 2.264(4)
and 2.270(3) Å are observed for 3, while comparable distances
Figure 2. Labeling scheme and isolated yield for the bis-indenyl zirconium
sandwich complexes.
1-THF were similar, leading to the conclusion that 1 was the
formally zwitterionic zirconium sandwich (η5-C9H5-1,3-(SiMe3)2)-
(η6-C9H5-1,3-(SiMe3)2)Zr. As will be described in a later section,
newly acquired X-ray crystallographic data on related zirconium
sandwiches has established unprecedented η9 hapticity of one
of the indenyl ligands, indicating that the structure of 1 is in
fact (η5-C9H5-1,3-(SiMe3)2)(η9-C9H5-1,3-(SiMe3)2)Zr.
Reductive elimination from the other homoleptic bis-indenyl
zirconocene isobutyl hydrides, 2-(CH2CHMe2)H, 3-(CH2-
CHMe2)H, rac/meso-4-(CH2CHMe2)H, and 5-(CH2CHMe2)-
H, produced similar products. In each case, the analytical and
NMR spectroscopic data (vide infra) are consistent with the
formation of the bis-indenyl zirconium sandwich complex
(Figure 2). Each compound was isolated as a burgundy solid in
moderate yield. Monitoring the isobutane reductive elimination
reactions in situ by NMR spectroscopy revealed that each
sandwich compound is formed in near quantitative yield,
demonstrating that the low isolated yields are a consequence
of the high lipophilicity of the molecules. The bis-indenyl
zirconium sandwich complexes can also be prepared by alkali
metal reduction of the dichloride precursor (eq 3). Although in
most cases alkane reductive elimination is the preferred method,
synthesis of 5 is best achieved by reduction of 5-Cl2 with sodium
amalgam due to a secondary reaction of the resulting sandwich
with the isobutene byproduct (vide infra).
Crystallographic Characterization of 3 and 5 and Elec-
tronic Structure. Two of the bis-indenyl zirconium sandwich
complexes have been characterized by X-ray crystallography.
Cooling a saturated pentane solution of 3 to -35 °C afforded
burgundy blocks suitable for diffraction. The compound crystal-
lizes with two independent molecules in the asymmetric unit,
9
16940 J. AM. CHEM. SOC. VOL. 126, NO. 51, 2004

Zirconium Sandwich Complexes
A R T I C L E S
Table 1. Comparison of the Bond Distances (Å) and Angles (deg)
of the η5 and η9 Indenyl Ligands in 3 and 5
η
5 indenyl
η
9 indenyl
3
5
3
5
Zr-C(1)
2.513(3) 2.485(2) Zr-C(9)
2.455(3) 2.517(2) Zr-C(10)
2.496(3) 2.536(2) Zr-C(6)
2.544(3) 2.539(2) Zr-C(7)
2.561(4) 2.528(2) Zr-C(8)
3.507(4) 3.416(2) Zr-C(18)
4.185(4) 4.093(2) Zr-C(17)
4.171(4) 4.113(2) Zr-C(16)
3.465(4) 3.452(2) Zr-C(15)
2.612(3) 2.574(2)
2.752(3) 2.784(2)
2.618(3) 2.626(2)
2.264(4) 2.267(2)
2.270(3) 2.255(2)
2.383(4) 2.334(2)
2.520(4) 2.487(2)
2.518(4) 2.501(2)
2.380(4) 2.386(2)
Zr-C(2)
Zr-C(3)
Zr-C(4)
Zr-C(5)
Zr-C(11)
Zr-C(12)
Zr-C(13)
Zr-C(14)
C(11)-C(12)
C(12)-C(13)
C(13)-C(14)
1.350(6) 1.369(2) C(18)-C(17) 1.402(6) 1.426(3)
1.419(5) 1.417(4) C(17)-C(16) 1.384(6) 1.395(3)
1.366(6) 1.367(3) C(16)-C(15) 1.406(6) 1.418(3)
rotational anglea 177.8(4)
84.6(2)
61.3(4)
37.6(1)
6.3(2)
cent-Zr-centb 60.7(6)
η9 fold anglec
37.1(3)
4.9(2)
buckle
Figure 3. Molecular structure of one of the independent molecules of 3 at
30% probability ellipsoids. Hydrogen atoms omitted for clarity.
a Angle formed between the plane defined by Zr, C(2), and the centroid
between C(4)-C(5) and the plane defined by Zr, C(7)-C(8) centroid, and
C(16)-C(17) centroid. b Angle formed between the zirconium and the
centroids of the five- and six-membered rings of the η9 indenyl ligands.
c Angle formed between the intersection of the normal to the planes of the
centroid of the five- and six-membered rings of the η9 ligands.
2- 42
nularly bridged analogue pentalene, C8H6
. On the basis of
this relationship, indenyl ligands would be expected to engage
in bonding to transition metals with hapticities ranging from η1
to η9. Transition metal-indenyl bonding is usually confined to
direct interactions with the five-membered ring, although it is
well known that this class of ligand frequently “slips” from η5
to η3 coordination as a result of a gain in aromaticity within
the benzo ring.30 Examples of η6-indene, C9H8, complexes are
known, but in these cases the cyclopentadienyl ligand is
protonated in analogy to well-known transition metal arene
complexes.43a,b In addition, examples of heterobimetallic η6,η5
complexes where indenyl acts as a bridging ligand have also
been reported.43c
Figure 4. Molecular structure of 5 at 30% probability ellipsoids. Hydrogen
atoms omitted for clarity.
Crystallographic characterization of both 3 and 5 provides
the first structural evidence for η9 metal-indenyl bonding.42b
For the zirconium sandwich complexes, this interaction affords
a formally 18-electron, coordinatively saturated metal center.
These molecules are conceptually related to the group 4
transition metal pentalene compounds initially described by
Jonas and co-workers.44 Both “mixed-ring” compounds of the
form (η5-C5H5)(η8-C8H6)MX (M ) Ti, Zr, Hf; X ) halide),
and homoleptic bis-pentalene complexes (η8-C8H6)2M have been
isolated. Solution NMR studies on the latter compounds
established molecules of D2d molecular symmetry, consistent
with η8 coordination of both pentalene ligands, giving the
appearance of “20-electron” compounds. Computational studies
are inconsistent with this bonding picture, demonstrating that
only 9 of the 10 ligand π-orbitals interact with the metal center,
of 2.267(2) and 2.255(2) Å are present in 5. For reference, more
typical zirconium-carbon bond distances in the range from
2.539(2) to 2.561(4) Å are observed for C(4) and C(5) in the
η5 ligand. Moreover, the coordinated six-membered rings are
slightly buckled by 4.9(2)° and 6.2(2)° in 3 and 5, respectively,
less than the buckling of 19° observed in 1-THF,34 suggesting
that the aromaticity of the benzo ring is preserved upon
coordination to zirconium. Similar observations have been made
by Basolo in compounds where an η5 indenyl ligand “slips” to
η3 hapticity.30b These results, coupled with the structures of 3
and 5, underscore the flexibility of the indenyl π-system to
accommodate the coordination requirements of the metal.
The carbon-carbon bond distances in the six-membered rings
also lengthen somewhat upon coordination to the zirconium.
For the η9 indenyl ring in 3, the C(17)-C(18) and C(15)-C(16)
distances are 1.402(6) and 1.406(6) Å, respectively, slightly
elongated from the corresponding C(11)-C(12) and C(13)-
C(14) bond lengths of 1.350(6) and 1.366(6) Å in the η5 ligand.
A similar effect is also observed in 5, where the C(17)-C(18)
and C(15)-C(16) distances are 1.426(3) and 1.418(3) Å,
respectively, while C(11)-C(12) and C(13)-C(14) are 1.369(2)
and 1.367(3) Å.
(41) Katz, T. J. J. Am. Chem. Soc. 1960, 82, 3784.
(42) (a) Katz, T. J.; Rosenberger, M. J. Am. Chem. Soc. 1962, 84, 865. (b) The
η9 indenyl interaction was initially predicted computationally by Verios
before the determination of the solid state structures of 3 and 5. Verios, L.
F., manuscript in preparation.
(43) (a) Fessenbecker, A.; Stephan, M.; Grimes, R. N.; Pritzkow, H.; Zenneck,
U.; Siebert, W. J. Am. Chem. Soc. 1991, 113, 3061. (b) Treichel, P. M.;
Johnson, J. W. J. Organomet. Chem. 1975, 88, 207. (c) Bonifaci, C.;
Ceccon, A.; Gambaro, A.; Manoli, F.; Mantovani, L.; Ganis, P.; Santi, S.;
Venzo, A. J. Organomet. Chem. 1998, 577, 97. (d) Ceccon, A.; Gambaro,
A.; Santi, S.; Valle, G.; Venzo, A. Chem. Commun. 1989, 51.
Because the indenyl anion is a 10 π-electron system, it is
(44) Jonas, K.; Korb, P.; Kollbach, G.; Gabor, B.; Mynott, R.; Angermund, K.;
Heineman, O.; Kru¨ger, C. Angew. Chem., Int. Ed. Engl. 1997, 36, 1714.
isoelectronic with cyclooctatetrenyl dianion41 and its transan-
9
J. AM. CHEM. SOC. VOL. 126, NO. 51, 2004 16941

A R T I C L E S
Bradley et al.
suggesting a more traditional 18-electron molecule.45 This
computational model has also been corroborated by photoelec-
tron spectroscopy.46
While the two ligand types are isoelectronic, the resulting
bis-indenyl zirconium sandwich complexes have several notable
differences from the homoleptic bis-pentalene compounds. Most
significantly, dissociation of the neutral benzo portion of the
η9 ligand would provide a formally 14-electron, η5,η5-“zir-
conocene”-like sandwich and potentially the rich chemistry of
such species.2a,33 In the ground state, the bis-indenyl zirconium
sandwiches can be formulated as either Zr(II) compounds, where
the coordinated benzo ring functions as a neutral four-electron
donor, or Zr(IV) species with an LX2 olefin-alkyl fragment.
This redox ambiguity is similar to the bonding scheme
encountered with bent zirconocene diene complexes47 and
titanium arene compounds.48 The metrical parameters of both
3 and 5 (Table 1) establish the importance of the Zr(IV)
canonical form (B), as the CdC bonds of the coordinated benzo
ring are slightly elongated with respect to the corresponding η5
ligand. The slight buckling of the coordinated six-membered
ring also supports the LX2 bonding description, where the
σ-interactions occur from the C15 and C18 atoms. Coordination
of the cyclopentadienyl portion of the indenyl ligand in 3 and
5 attenuates the unsaturation of the zirconium and lessens the
importance of the Zr(IV) canonical structures.
Figure 5. Representations of the HOMO and LUMO of 5 as calculated
using ADF (2003.01, TZ2P, ZORA).
for 5 are presented in Figure 5 as representative examples. More
detailed frontier molecular orbital diagrams for both molecules
are contained in the Supporting Information. The HOMO of 5
(Figure 5) is principally composed of a zirconium 1a1 orbital53
interacting with a linear combination of p-orbitals of the six
carbons on the benzo portion of the η9 ligand. The major lobe
of the zirconium orbital bonds with the p-orbitals on the benzo
carbons proximal to the cyclopentadienyl ring. This interaction
is augmented by the favorable overlap of the toroid with the
p-orbitals of bridgehead carbons as well as those from the benzo
carbons distal from the five-membered ring. The LUMO
(-1.679 eV) lies 35 kcal/mol above the HOMO and is
principally π-orbitals on the benzo portion of the η5 indene that
are not engaged in bonding with the zirconium. The large
HOMO-LUMO gap is consistent with the observation of a
closed shell, coordinatively saturated molecule.
Additional zirconium-ring interactions are observed in the
HOMO-1 (-4.747 eV), which principally contains an interaction
between a cloverleaf zirconium d-orbital and one of the
cyclopentadienyl carbons substituted with an isopropyl group.
A similar zirconium-ring interaction is observed in the
HOMO-2 (-4.970 eV) orbital that is augmented by additional
bonding to the η5 indenyl ring. These interactions, in conjunction
with HOMO-4 (-6.208 eV), account for the bonding of all nine
indenyl carbons with the zirconium. Graphical representations67
of these orbitals can be found in the Supporting Information.
NMR Characterization and Solution Dynamics of Homo-
leptic Bis-indenyl Zirconium Sandwich Complexes. The
solution structure and dynamics of the homoleptic bis-indenyl
The electronic structure of the zirconium sandwich complexes
has been studied in more detail with the aid of density functional
theory calculations as implemented by the ADF suite of
programs.49-51 Full molecule geometry optimization of 5
provided a structure in excellent agreement with the experi-
mentally determined structure.52 A similar computation was
conducted on the full molecule of 1, and the optimized geometry
is in good agreement with the solid-state structure of 3. The
HOMO and LUMO of 1 and 5 are similar, and selected orbitals
1
zirconium sandwich complexes, 1-5, were studied by H and
13C NMR spectroscopy in conjunction with two-dimensional
g-COSY, g-HSQC, and g-HMBC experiments. Complete assign-
ments for each of the zirconium sandwich complexes are
provided in the Experimental Section and Supporting Informa-
tion. As a representative example, both 1H and 13C spectra of 3
are presented in Figure 6. As previously reported for 1,34 both
(45) Costuas, K.; Saillard, J.-Y. Chem. Commun. 1998, 2047.
(46) Gleiter, R.; Bethke, S.; Okubo J.; Jonas, M. Organometallics 2001, 20,
4274.
(47) Erker, G.; Kruger, C.; Muller, G. AdV. Organomet. Chem. 1985, 24, 1.
(48) Ozerov, O. V.; Ladipo, F. T.; Patrick, B. O. J. Am. Chem. Soc. 1999, 121,
7941.
(49) te Velde, G.; Bickelahaupt, F. M.; van Gisbergen, S. J. A.; Fonseca Guerra,
C.; Barends, E. J.; Snijders, J. G.; Zielger, T. J. Comput. Chem. 2001, 22,
931.
(50) Guerra, C. F.; Snijders, J. G.; te Velde, G.; Baerend E. J. Theor. Chem.
Acc. 1998, 99, 391.
(53) Lauher, J. W.; Hoffmann, R. J. Am. Chem. Soc. 1976, 98, 1729.
(54) Marat, K. Spinworks; University of Manitoba: Manitoba, Canada, 2004.
(55) For a recent example, see: Jahr, H. C.; Nieger, M.; Do¨tz, K. H. Chem.
Commun. 2003, 2866.
(56) (a) Negishi, E.; Kondakov, D. Y. Chem. Soc. ReV. 1996, 26, 417. (b)
Negishi, E.; Takahashi, T. Bull. Chem. Soc. Jpn. 1998, 71, 755.
(57) Fujita, K.; Shinokubo, H.; Oshima, K. Angew. Chem., Int. Ed. 2003, 42,
2550.
(58) Fujita, K.; Yorimitsu, H.; Shinokubo, H.; Oshima, K. J. Am. Chem. Soc.
2004, 126, 6776.
(59) McDade, C.; Bercaw, J. E. J. Organomet. Chem. 1985, 279, 281.
(60) Swanson, D. R.; Rousset, C. J.; Negishi, E. J. Org. Chem. 1989, 54, 3521.
(61) Pangborn, A. B.; Giardello, M. A.; Grubbs, R. H.; Rosen, R. K.; Timmers,
F. J. Organometallics 1996, 15, 1518.
(62) Go´mez, J. C. C.; Lo´pez, F. J. S. MestRe-C, version 3.7.1.9.0; Universidade
de Santiago de Compostela: 2004; www.mestrec.com.
(51) Baerends, E. J.; Autschbach, J. A.; Berces, A.; Bo, C.; Boerrigter, P. M.;
Cavallo, L.; Chong, D. P.; Deng, L.; Dickson, R. M.; Ellis, D. E.; Fan, L.;
Fischer, T. H.; Guerra, C. F.; van Gisbergen, S. J.; Groeneveld, J. A.;
Gritsenko, O. V.; Gru¨ning, M.; Harris, F. E.; van den Hoek, P.; Jacobsen,
H.; van Kessel, G.; Koostra, F.; van Lenthe, E.; Osinga, V. P.; Patchkovshii,
S.; Philipsen, P. H.; Post, D.; Pye, C. C.; Ravenek, W.; Ros, P.; Schipper,
P. R.; Schreckenbach, G.; Snijders, J. G.; Sola, M.; Swart, M.; Swerhone,
D.; te Velde, G.; Vernooijs, P.; Versluis, L.; Visser, O.; van Wezenbeek,
E.; Wiesenekker, G.; Wolff, S.; Woo, T.; Ziegler, T. ADF 2002.03 SCM
Theoretical Chemistry; Vrije Universiteit: Amsterdam, The Netherlands,
http://www.scm.com/.
(52) A calculation on staggered 5 was also performed and provided results similar
to those for gauche 5.
9
16942 J. AM. CHEM. SOC. VOL. 126, NO. 51, 2004

Zirconium Sandwich Complexes
A R T I C L E S
Figure 7. Eyring plots for indenyl ring interconversion for 1-4.
the original spectra. Special care should be exercised in heating
1 as prolonged thermolysis induces approximately 40% conver-
sion to another unidentified compound that most likely derives
from cyclometalation of the [SiMe3] substituents. The observed
temperature-dependent NMR spectra are consistent with a
fluxional process involving interconversion of the η9 and η5
indenyl ligands on the NMR time scale (eq 4). Notably, the 1H
NMR spectrum of 5 exhibits the number of resonances for a
Cs-symmetric molecule inconsistent with the solid-state structure
which is C1-symmetric, suggesting that ring rotation occurs even
when the ring interconversion is frozen out.
Figure 6. 1H and 13C NMR spectrum of 3 in benzene-d6 at 23 °C.
spectra exhibit the number of peaks consistent with a molecule
of Cs symmetry with inequivalent indenyl rings. One of the most
notable features of the 1H NMR spectrum of 3 is the resonances
centered at 4.02 and 5.37 ppm. Both g-COSY and g-HMBC
experiments establish diagnostic couplings that identify these
peaks as benzo resonances on the η9 indenyl ring. Similar
features are observed in the 13C spectrum, where upfield-shifted
carbons appear at 75.32 and 95.32 ppm for the carbons in the
six-membered ring coordinated to the zirconium. In contrast,
η9 hapticity has little effect on the chemical shift of the
cyclopentadienyl hydrogen, as 1H NMR signals centered at 6.20
and 6.45 ppm are observed for the protons on the inequivalent
cyclopentadienyl rings.
The other homoleptic bis-indenyl zirconium sandwich com-
plexes display similar spectral features. However, sandwiches
1-5 are dynamic in solution, and cooling the sample below
ambient temperature is often required for observation of a static
molecule. Variable-temperature 1H NMR spectroscopy was used
to interrogate the dynamics of each zirconium sandwich in
toluene-d8 solution. As a representative example, a stack plot
of the spectra of 2 as a function of temperature is presented in
the Supporting Information. At temperatures below -40 °C, a
Cs-symmetric molecule is observed with distinct resonances for
the benzo and cyclopentadienyl hydrogens. In the static limit,
the benzo hydrogens of the η9-indenyl ring appear at diagnostic
values of 3.66 and 5.56 ppm. Diastereotopic silyl methyl groups
are also observed at this temperature. Warming the sample to 2
°C broadens both the benzo and cyclopentadienyl hydrogens
and results in coalescence of one set of silyl methyls. At 23
°C, the benzo and cyclopentadienyl hydrogens are nearly
broadened into the baseline and both sets of diastereotopic
methyl groups appear as singlets.
Rate constants for the ring interchange were measured by a
combination of line-shape analysis54,62 and EXSY NMR experi-
ments.63 The most reliable line-broadening rate constants were
obtained from modeling the silyl and [CMe3] indenyl substit-
uents rather than the cyclopentadienyl or benzo hydrogens. For
both 2 and 3, the observation of two sets of diastereotopic methyl
groups provided redundant rate constants that are in good
agreement. One exception is 5, where the peak overlap in the
low-temperature spectra prevents an accurate determination of
∆ν∞. For consistency, the rate constants for ring exchange of
several other sandwiches were also measured by EXSY and
found to be in excellent agreement with the values obtained
from line broadening. For example, the rate of ring intercon-
version for 1 determined by line broadening was 5(1) s-1
whereas a value of 6(2) s-1 was measured by EXSY.
Ring-exchange rate constants for sandwiches 1-4 were
determined over a modest range of temperatures, typically
between 35 and 50 °C, allowing construction of Eyring plots
(Figure 7) and determination of activation parameters (Table
2). For comparison, the relative rates of ring interchange at a
common temperature, 23 °C, are also included in Table 2. The
values of ∆S‡ range between 2 and 12 eu and are consistent
Similar variable-temperature behavior was observed for the
other zirconium sandwich complexes. In all cases, the changes
are reversible as repeated warming and cooling cycles regenerate
(63) (a) Perrin, C. L.; Dwyer, T. J. Chem. ReV. 1990, 90, 935. (b) Derose, E.
F.; Castillo, J.; Saulys, D.; Morrison, J. J. Magn. Reson. 1991, 93, 347. (c)
Zolnai, Z.; Juranic, N.; Vikic-Topic, D.; Macura, S. J. Chem. Inf. Comput.
Sci. 2000, 40, 611.
9
J. AM. CHEM. SOC. VOL. 126, NO. 51, 2004 16943

A R T I C L E S
Bradley et al.
Figure 8. Relative rates of ring interconversion for bis-indenyl zirconium compounds.
Table 2. Activation Parameters and Relative Rates of Ring
6.40 ppm do not couple to any other peaks, identifying them as
cyclopentadienyl hydrogens. The remaining four upfield reso-
nances are therefore assigned as the four inequivalent hydrogens
on the benzo portion of the η9-indenyl ring. Through-bond
coupling was established by a g-COSY experiment and identi-
fied the resonances at 5.17 and 5.45 ppm as the benzo hydrogens
proximal to the five-membered ring, whereas those centered at
3.82 and 4.07 ppm are distal. Similar analyses were used to
establish connectivity in 8 and 9. Spectral assignments for all
three molecules can be found in the Experimental Section and
Supporting Information.
Interconversion for Homoleptic Bis-indenyl Zirconium Sandwich
Complexes
compound
∆H (kcal/mol)
∆S (eu)
∆
G (kcal/mol)a
relative ratea
1
2
3
4
5
17.9(1)
14.5(3)
20.4(3)
15.1(3)
NAb
6(1)
16.3(1)
13.8(3)
16.9(3)
14.4(3)
19.9(5)c
1.2
80
1
40
6
2(1)
12(5)
2(1)
NA
a T ) 296 K. b Not determined due to overlapping resonances. c Deter-
mined from a single rate constant at 296 K.
with a unimolecular, dissociative process. From these results
the relative rates of η5 f η9 ring interconversion have been
measured as a function of indenyl substituent (Figure 8). A
discussion of the rates and possible mechanisms for ring
interconversion will be presented in a later section.
While through-bond coupling experiments established con-
nectivity, these data do not identify which ligand is engaged in
the η9 interaction. A series of NOESY experiments were
conducted on 7, 8, and 9 to determine the identity of the
preferred isomer. As a representative example, the assignment
of 9 will be presented in detail. Cross-peaks in the -40 °C
NOESY spectrum were observed between the [CMe3] resonance
(1.37 ppm) and one cyclopentadienyl (5.02 ppm), one isopropyl
methine (3.34 ppm), and three benzo hydrogens. Importantly,
one of the benzo resonances (7.68 ppm) is in the η5 indenyl
region, while the two others (4.04 and 5.37 ppm) are in the
Synthesis of Heteroleptic Bis-indenyl Zirconium Sandwich
Complexes. Introduction of two chemically distinct indenyl
ligands into the coordination sphere of the zirconium raises the
question of what (if any) preference would be observed for η9
coordination. To test this hypothesis, the synthesis of several
heteroleptic bis-indenyl zirconium complexes was explored.
Treatment of 6-Cl2 with two equivalents of LiCH2CHMe2
rapidly produced isobutene from hydride transfer and isobutane
from reductive elimination, along with the desired sandwich
1
complex, 6 (eq 5). The H NMR spectrum of 6 in benzene-d6
exhibits twice the number of resonances in equal ratios expected
from a single Cs-symmetric bis-indenyl zirconium sandwich.
Similar behavior is observed by 13C NMR spectroscopy.
Subsequent EXSY NMR experiments on the mixture of
compounds demonstrate facile exchange between the two
isomers (eq 5). Thus, for this sandwich there is no measurable
preference for either ring to adopt η9 coordination.
Similar synthetic procedures were used to prepare the other
heteroleptic sandwiches, 7, 8, and 9 (Figure 2). In all three cases,
only one isomer was detected by 1H and 13C NMR spectroscopy.
A series of two-dimensional NMR experiments were conducted
to assign the identity of each observed isomer and determine
the preference for η9 coordination. Establishment of connectivity
and differentiation between cyclopentadienyl and benzo peaks
was accomplished by a combination of g-COSY, g-HMBC, and
g-HSQC experiments. As a representative example, the benzo
and cyclopentadienyl region of the g-COSY spectrum of 7 is
shown in Figure 9. The two resonances centered at 6.09 and
Figure 9. Indenyl region of the g-COSY NMR spectrum of 7 in benzene-
d6 at 21 °C.
9
16944 J. AM. CHEM. SOC. VOL. 126, NO. 51, 2004

Zirconium Sandwich Complexes
A R T I C L E S
Figure 11. Relative rate of ring interconversion, carbonyl stretching
frequencies of the corresponding zirconocene dicarbonyl, and oxidation
potential of the bis-indenyl iron sandwiches.
Figure 10. Possible pathways for η5 f η9-indenyl interconversion of bis-
indenyl zirconium sandwich complexes.
diagnostic η9 portion of the spectrum. Because no cross-peak
was observed between the [CMe3] group and any distal
hydrogen on the η5 indenyl, this substituent must be on the η5
ring. Enhancements in the other two benzo hydrogens are a
result of transannular NOE effects on the η9 ligand. Compli-
mentary cross-peaks to the other benzo resonances were
observed with the [SiMe3] substituent. Similar cross-peaks
involving the [CHMe2] groups as well as other benzo resonances
were observed and allowed complete assignment of the 1H NMR
spectrum.
Performing the NOESY at 22 °C produced similar cross-
peaks, although the η9-η5 through-space interactions were
reduced. Significantly, weak but observable exchange peaks are
also observed with a minor isomer present in ∼1% of the
mixture. These data clearly establish the thermodynamic prefer-
ence for the η9 coordination of the isopropyl-substituted ligand,
although the minor isomer is kinetically accessible. Similar data
were collected on 7 and 8 and allowed assignment of the major
isomer in both cases. In all three heteroleptic sandwich
complexes, the more alkylated indenyl ring preferentially adopts
η9 coordination (Figure 2).
Observation of a single isomer of 7 as compared to the near
equimolar mixture observed for 6 highlights the influence of
indenyl substitution on the ground-state preferences for η9
coordination. Replacement of a single silyl group with an alkyl
substituent perturbs the equilibrium such that predominantly one
isomer is observed in solution. The synthesis of single isomers
of 8 and 9 also corroborates the finding that alkylated indenyl
ligands prefer η9 hapticity. Previous studies from our laboratory
have established that alkylated indenyl ligands are more electron
donating than their silylated counterparts.35 Preferential interac-
tion of the more electron-donating benzo ring suggests that
π-donation from the coordinated six-membered ring is a key
component of metal-η9 indenyl bonding, supporting a signifi-
cant contribution from the zirconium(II) diolefin adduct.
Possible Mechanisms for Ring Interchange. Both the homo-
and heteroleptic bis-indenyl zirconium sandwich complexes are
dynamic in solution. Examining the frontier molecular orbitals
of the sandwiches (Figure 5 and Supporting Information)
suggests that dissociation of the benzo portion of the η9 indenyl
ligand to afford a symmetric η5,η5 intermediate appears to be a
plausible pathway for ring interconversion (Figure 10, Path A).
Experimental support for this pathway is provided by the
reaction of 1 with CO, alkynes, and (dimethylamino)pyridine
to afford bent zirconocene complexes with familiar η5,η5
coordination of the indenyl ligands.34 Although a detailed kinetic
study has not been performed, it is likely that ligand addition
to the 18-electron, coordinatively saturated zirconium sandwich
proceeds via initial dissociation of the benzo ring to afford a
reactive η5,η5 zirconium sandwich intermediate.
Alternatively, a mechanism involving formally zwitterionic
η5,η6 sandwich intermediates can also be envisioned (Figure
10, Path B). Isolation and crystallographic characterization of
an η5,η6 zirconocene-THF complex34 suggests that dissociation
of the five-membered ring is viable. It is important to note,
however, that in this mechanism the η6 indenyl ligand must
undergo a haptotropic rearrangement in order to interconvert
the two rings. Computational studies by Albright et al.39
demonstrated the feasibility of such ring migrations, and several
such rearrangements have been observed experimentally, most
frequently in group 6 arene chemistry.55 The computational
studies also suggest that ring migration through η3-allylic
intermediates are energetically favorable as rearrangement
through the center of the polycyclic rings is symmetry forbid-
den.39 A concerted pathway that interchanges the two indenyl
rings also cannot be excluded.
Attempts to distinguish between these pathways experimen-
tally have not been successful. While modest substituent effects
have been observed for the rate of ring interconversion, no
obvious trend has emerged that favors one pathway. The rate
of fluxionality versus electronic parameters for purely silyl-
substituted zirconium sandwiches is presented in Figure 11. Both
carbonyl stretching frequencies on the corresponding zir-
conocene dicarbonyl compounds and electrochemistry on related
bis-indenyl iron sandwiches35 have established the relative
electrophilicity of each zirconocene as 3 < 1 < 2. The ordering
of the electronic properties of the silylated zirconium sandwiches
correlates with their electronic parameters, although the effects
are quite small. Our previous studies demonstrated that the
electronic properties of 1 and 3 are almost indistinguishable as
are the rates of ring interconversion.35 In contrast, 2 is much
more electrophilic, and as a consequence, the rate of the
fluxional process increases dramatically as compared to 1 and
3. It is interesting to note that the two most sterically hindered
complexes are at opposite ends of the series.
A possible explanation for this trend could be ground-state
destabilization of benzo ring coordination based on observations
that the most electron-donating indenyl ligand adopts η9
coordination. However, the relative rates of ring interconversion
for the alkylated zirconium sandwiches, 4 and 5, do not support
this reasoning. Taken as a whole, the relative rates of indenyl
ring interchange are slightly dependent on the substituents and
a composite of both ground- and transition-state effects that are
not readily delineated based on the data collected.
9
J. AM. CHEM. SOC. VOL. 126, NO. 51, 2004 16945

A R T I C L E S
Bradley et al.
Reaction of Bis-indenyl Zirconium Sandwiches with
Unsaturated Organic Molecules. Synthesis and characteriza-
tion of bis-indenyl zirconium sandwich complexes prompted
an exploration of their utility in traditional “zirconocene”-
promoted reactions with olefins and alkynes.3a Because a series
of isolable zirconium sandwiches has been prepared, there is a
unique opportunity to tune the stereoelectronic properties of the
reagent, perhaps for specifically tailored applications. As noted
in a previous section, preparation of the alkylated zirconium
sandwich 5 by treatment of 5-Cl2 with two equivalents of
isobutyllithium produced the bis-indenyl zirconium methyl allyl
hydride complex 5-(η3-CH2C(Me)CH2)H (eq 6). Isolation of
yellow 5-(η3-CH2C(Me)CH2)H using this method is best
achieved in a closed system so that the isobutene cannot escape.
This C-H activation side reaction was not observed during the
synthesis of any other zirconium sandwich, highlighting the
increased propensity of 5 to participate in formal oxidative
addition reactions. The origin of this effect is 2-fold. The
presence of alkylated indenyl ligands produces a relatively
electron-rich zirconium that favors C-H bond addition.35
Moreover, this sandwich is the most open member of the series,
thus favoring the approach and activation of the olefin.
Synthesis of 5-(η3-CH2C(Me)CH2)H has also been indepen-
dently accomplished by addition of excess isobutene to the
sandwich. Working with isolated compound revealed that the
C-H activation reaction is reversible; reductive elimination of
isobutene is observed in solution over the course of hours at
ambient temperature. Characterization of 5-(η3-CH2C(Me)-
CH2)H has been accomplished by NMR spectroscopy, elemental
analysis, and X-ray diffraction. The solid-state structure of the
molecule is presented in Figure 12. As we observed previously,34
oxidative addition of a C-H bond to the zirconium sandwich
restores the more commonly observed η5,η5 hapticity of the
indenyl ligands. In contrast to the gauche conformation observed
in the solid-state structure of 5, the ring orientation in 5-(η3-
CH2C(Me)CH2)H is staggered with a rotational angle of
179.4(3)°. This change in conformation most likely arises from
a gearing effect induced by the methyl group on the allyl ligand.
To maximize bonding to the π-allyl, the methyl group must
point directly toward one of the indenyl ligands. As a result,
the ring orients such that the methyl substituent is nestled
between the two isopropyl groups whose methine hydrogens
are directed toward the interior of the metallocene. The other
ring then staggers with respect to the first to minimize
transannular interactions.
Figure 12. (a) Molecular structure of 5-(η3-CH2C(Me)CH2)H at 30%
probability ellipsoids. Hydrogen atoms, except the zirconium hydride,
omitted for clarity. (b) Top view of the molecule with isopropyl substituents
and hydrogen atoms, except for the zirconium hydride, omitted for clarity.
Table 3. Selected Bond Distances for 5-(η3-CH2C(Me)CH2)H
bond distance (Å)
Zr(1)-C(31)
Zr(1)-C(32)
Zr(1)-C(33)
C(31)-C(33)
C(32)-C(33)
C(33)-C(34)
2.486(6)
2.407(5)
2.571(5)
1.366(8)
1.409(7)
1.517(7)
at temperatures below -38 °C, where the molecule is in the
static limit. In this regime, the spectra exhibit the number of
resonances expected for a C1-symmetric molecule. Assignment
of each hydrogen and carbon was accomplished through a
combination of g-COSY, g-HSQC, and g-HMBC experiments
and is reported in detail in the Experimental Section.
Given the importance of the transient zirconocene 1-butene
complex, (η5-C5H5)2Zr(η2-CH2dCHCH2CH3), in Negishi-type
coupling reactions,56 we explored the reactivity of the bis-
indenyl zirconium sandwiches with R-olefins. Careful addition
of one equivalent of 1-butene to 5 affords a yellow compound
identified as 5-(η3-CH2CHC(Me)H)H on the basis of multi-
nuclear NMR experiments (eq 7). This molecule can also be
prepared by addition of either cis- or trans-2-butene to the
sandwich. As expected, the internal alkenes react much slower
than the terminal olefin.
1
Characterization of 5-(η3-CH2C(Me)CH2)H by solution H
and 13C NMR spectroscopy revealed a dynamic molecule at
ambient temperature. Variable-temperature NMR studies are
consistent with η1,η3 allyl interconversion as well as hydride
transfer between the zirconium and the allyl fragments.58 The
full details of these experiments will be published elsewhere.
Identification of both 1H and 13C resonances has been achieved
As with 5-(η3-CH2C(Me)CH2)H, 5-(η3-CH2CHC(Me)H)-
H is dynamic in solution, requiring temperatures below -40
°C for observation of spectra in the static limit. A combination
of low-temperature g-HSQC and g-HMBC experiments was
9
16946 J. AM. CHEM. SOC. VOL. 126, NO. 51, 2004

Zirconium Sandwich Complexes
A R T I C L E S
Figure 13. Reactivity of 5 with olefins and alkynes.
used to assign the allylic carbons, while g-COSY data allowed
differentiation and assignment of the benzo, cyclopentadienyl,
and zirconium-hydride resonances. Importantly, 5-(η3-CH2CHC-
(Me)H)H is more resistant to reductive elimination of olefin
than 5-(η3-CH2C(Me)CH2)H, consistent with observations in
cyclopentadienyl chemistry where more substituted zirconocene
alkyl hydride complexes participate in more facile reductive
elimination processes.38
observed in cyclopentadienyl chemistry,60 the alkyl groups are
positioned at the â and â′ positions on the ring. In addition,
both the 1H and 13C NMR spectra of 5-Cy-1 are consistent with
a molecule of C2 symmetry, establishing the trans disposition
of the two ethyl substituents. Treatment of 5-Cy-1 with gaseous
hydrogen chloride afforded the zirconocene dichloride complex
5-Cl2 and 3,4-dimethylhexane. Identification of the alkane was
accomplished by comparison of NMR and mass spectral data
to an authentic sample. Reaction of the zirconacyclopentane with
N-bromosuccinimide (NBS) produced no reaction.
The bis-indenyl zirconium sandwich complexes also promote
the coupling of alkynes. Addition of an excess of either 2-butyne
or 4-octyne to 5 furnished the corresponding zirconacyclopen-
tadiene in near quantitative yield (Figure 13). Addition of HCl
produced 5-Cl2 and the corresponding diene, whereas reaction
with NBS afforded 5-Br2 and the dibromoalkene. Attempts to
observe the zirconium alkyne adduct by addition of one
equivalent of 2-butyne to 5 have been unsuccessful, producing
a near equimolar mixture of the zirconacyclopentadiene and the
starting sandwich. A stable alkyne adduct can be isolated upon
reaction with diphenylacetylene, although no coupling is
observed when this adduct is exposed to excess PhCtCPh or
2-butyne (Figure 13).
Isolation of 5-(η3-CH2CHC(Me)H)H contrasts the behavior
and stability of the parent zirconocene-1-butene complex.
Dioumaev and Harrod19 demonstrated that (η5-C5H5)2Zr(η2-
CH2dCHCH2CH3) is in equilibrium with the zirconocene allyl
hydride, although the equilibrium favors the olefin complex.
Recently, Oshima and co-workers were able to exploit this
dynamic phenomenon by performing tandem reductions of 1,4-
diketones by delivering both a hydride and an allylic fragment
from the zirconium.57,58 In 5-(η3-CH2CHC(Me)H)H the in-
creased electron density imparted by the alkylated indenyl
ligands favors oxidative addition of the allylic C-H bond.
Similar preferences have been observed in permethylzirconocene
chemistry where several allyl hydride complexes have been
isolated.59 Unfortunately attempts to prepare similar compounds
with zirconocenes with electron-withdrawing indenyls such as
1 and 2 have yielded a complex mixture of products.
Silylated zirconocene sandwiches participate in similar
chemistry depending on both the silyl group and alkyne. For 1,
addition of one equivalent of 2-butyne allows observation of
the adduct whereas addition of excess alkyne yields the
zirconacyclopentadiene (Figure 14). Similar reactivity is ob-
served with 2, where the alkyne adduct is observed upon
stoichiometric addition of alkyne and zirconocyclopentadiene
is obtained in the presence of excess 2-butyne. Treatment of
the latter compound with NBS afforded the dibromoalkene and
2-Br2 (Figure 14).
Despite observation of allylic activation in the ground state,
both 5 and 5-(η3-CH2CHC(Me)H)H participate in coupling
reactions that involve the intermediacy of the η2-olefin com-
pound. Addition of an excess of 1-butene to either 5 or 5-(η3-
CH2CHC(Me)H)H resulted in selective coupling of the alkenes
to afford the zirconacyclopentane, 5-Cy-1 (Figure 13). As
The bis-indenyl zirconium sandwich with the largest silyl
substituents, 3, displays unique reactivity that reflects the
increased steric bulk of the ligand array. Treatment with an
excess of 2-butyne affords the corresponding alkyne adduct
9
J. AM. CHEM. SOC. VOL. 126, NO. 51, 2004 16947

A R T I C L E S
Bradley et al.
1
MHz for H and 125.704 MHz for 13C. The temperature for the VT
experiments was calibrated using methanol and ethylene glycol
standards. Line-shape analyses were performed using DNMR as
implemented by Spinworks 2.2.54
Qualitative EXSY spectra were acquired with a sweep width of 4.2
kHz and a mixing time of 200 ms in phase-sensitive mode. For
quantitative rate studies, mixing times of 200, 50, 20, 10, 5, and 1 ms
were used. A total of 200 complex points were collected in the indirectly
detected dimension with 16 scans and 2k points per increment. The
resulting matrix was zero filled to 1k × 1k complex data points, and
Gaussian line-broadening functions were applied in both dimensions
prior to Fourier transform. Cross-peak integration was performed in
Mest-ReC,62 intensities were normalized, and rates were calculated using
literature methods.63
NOESY spectra were acquired with a sweep width of 4.8 kHz and
a mixing time of 200 ms in phase-sensitive mode. A total of 200
complex points were collected in the indirectly detected dimension with
16 scans and 2k points per increment. The resulting matrix was zero
filled to 1k × 1k complex data points, and Gaussian line-broadening
functions were applied in both dimensions prior to Fourier transform.
All 2D spectra were acquired at 22 °C unless otherwise noted. Details
about more routine two-dimensional experiments can be found in the
Supporting Information.
Figure 14. Reactivity of silylated bis-indenyl zirconium sandwich com-
plexes with alkynes.
3-(MeCtCMe). No additional chemistry is observed upon
addition of excess alkyne. Addition of diphenylacetylene to 3
produced no reaction.
Concluding Remarks
Calculations were performed with the Amsterdam Density Functional
Theory (ADF2003.01) suite of programs. Relativistic effects were
included using the zero-order regular approximation (ZORA). The
Vosko, Wilk, and Nusair (VWM) local density approximation,64
Becke’s exchange,65 and Perdew’s correlation66 (BP86) were used. The
cores of the atoms were frozen up to 1s for C and N and 2p for Zr and
Si. Uncontracted Slater-type orbitals (STOs) of triple-ú quality with
two polarizations were employed. This basis set is denoted TZ2P in
the ADF program. Each geometry optimization was carried out without
symmetry constraints. Orbital pictures were generated using MOLE-
KEL67 to process the ADF results.
Single crystals suitable for X-ray diffraction were coated with
polyisobutylene oil in a drybox and quickly transferred to the
goniometer head of a Siemens SMART CCD area detector system
equipped with a molybdenum X-ray tube (λ ) 0.710 73 Å). Preliminary
data revealed the crystal system. A hemisphere routine was used for
data collection and determination of lattice constants. The space group
was identified, and the data were processed using the Bruker SAINT
program and corrected for absorption using SADABS. The structures
were solved using direct methods (SHELXS) completed by subsequent
Fourier synthesis and refined by full-matrix least-squares procedures.
Elemental analyses were performed at Robertson Microlit Laboratories,
Inc., in Madison, NJ. Mass spectrometry was performed on a Hewlett-
Packard 5980A GC/MS equipped with a 5970 mass-selective detector
and a DB-1, 0.25 µm film, fused silica capillary column with an
injection temperature of 35 °C, a ramp rate of 10 °C, and a final column
temperature of 260 °C. The organic products were characterized by
comparison to literature reports.68
A series of isolable bis-indenyl zirconium sandwich com-
plexes has been prepared by either alkane reductive elimination
or reduction with sodium amalgam. Crystallographic charac-
terization of two examples establishes unprecedented η9 coor-
dination of one of the indenyl ligands, providing coordinatively
saturated molecules and a new binding mode for a ubiquitous
organometallic ligand. NMR spectroscopic studies revealed
dynamic molecules in solution, where the η5 and η9 rings are
interconverting on the time scale of the NMR experiment. In
analogy to “zirconocene” chemistry, the bis-indenyl zirconium
sandwiches react with olefins and alkynes to form zirconacycles
with restored η5,η5 hapticity. Isolation of the intermediate olefin
complexes reveals that the equilibrium lies toward the zir-
conocene allyl hydride arising from C-H activation of an allylic
hydrogen. The results of these studies have provided a family
of isolable “zirconocene” reagents that can be easily tuned by
judicious choice of indenyl substituents.
Experimental Section
General Considerations. All air- and moisture-sensitive manipula-
tions were carried out using standard vacuum line, Schlenk, or cannula
techniques or in a M. Braun inert atmosphere drybox containing an
atmosphere of purified nitrogen. Solvents for air- and moisture-sensitive
manipulations were initially dried and deoxygenated using literature
procedures.61 Benzene-d6 and toluene-d8 for NMR spectroscopy were
distilled from sodium metal under an atmosphere of argon and stored
over 4 Å molecular sieves or sodium metal. All calibrated gas bulb
additions used a 31.6 mL calibrated gas bulb. Preparation of all
zirconocene dichloride complexes,35 1,34 and 1-THF34 was accomplished
as described previously. Isobutylene and 2-butyne were purchased from
Aldrich and dried over molecular sieves before use. Diphenylacetylene
was purchased from Acros and dried on the high vacuum line followed
by recrystallization from dry pentane at -35 °C. 1-Butene and cis-
and trans-2-butene were purchased from Matheson and stored over
LiAlH4 before use.
Observation of (η5-C9H5-1,3-(SiMe2Ph)2)2Zr(CH2CHMe2)H (2-
(CH2CHMe2)H). A J. Young NMR tube was charged with 20 mg
(0.024 mmol) of 2 and 3 mg (0.047 mmol) of LiCH2CHMe2. Benzene-
d6 was then added by pipet, and the tube was sealed with a Teflon
1
valve. The tube was shaken, and the reaction was monitored by H
1
NMR spectroscopy over the course of 2 h. H NMR (benzene-d6): δ
) -1.65 (br s, 1H, CH2CH(CH3)2), 0.52 (s, 6H, SiMe2Ph), 0.54 (s,
6H, SiMe2Ph), 0.67 (s, 6H, SiMe2Ph), 0.69 (s, 6H, SiMe2Ph), 0.86 (d,
1H and 13C NMR spectra were recorded on a Varian Inova 400
Spectrometer operating at 399.779 (1H) and 110.524 MHz (13C). All
(64) Vosko, S. H.; Wilk, L.; Nusair, M. Can. J. Phys. 1990, 58, 1200.
(65) Becke, A. D. Phys. ReV. 1988, A38, 2398.
(66) Perdew, J. P. Phys. ReV. 1986, B33, 8822.
1
chemical shifts are reported relative to SiMe4 using H (residual) or
(67) Flukiger, P.; Luthi, H. P.; Portmann, S.; Weber, J. MOLEKEL 4.3; Swiss
Center for Scientific Computing: Manno, 2000; http://www.cscs.ch/
molekel.
13C NMR chemical shifts of the solvent as a secondary standard.
Variable-temperature and two-dimensional NMR experiments were
performed on a Varian Inova 500 Spectrometer operating at 499.920
(68) Takahashi, T.; Yanzhong, L. Titanium and Zirconium in Organic Synthesis;
Marek, I., Ed.; Wiley-VCH: Weinheim, 2002; Chapter 2.
9
16948 J. AM. CHEM. SOC. VOL. 126, NO. 51, 2004

Zirconium Sandwich Complexes
A R T I C L E S
8 Hz, 6H, CH2CH(CH3)2), 2.38 (br m, 1H, CH2CH(CH3)2), 6.23 (s,
1H, Zr-H), 6.52 (s, 2H, Cp), 7.08-7.18 (br m, 20H, Ph), 7.23 (m, 2H,
Ind), 7.39 (m, 2H, Ind), 7.46 (m, 2H, Ind), 7.62 (m, 2H, Ind). One
CH2CH(CH3)2 resonance not located.
pentane at -35 °C afforded 0.336 g (72%) of 5 as a burgundy solid.
Anal. Calcd for C30H38Zr: C, 73.56; H, 7.82. Found: C, 73.45; H,
1
8.08. H NMR (toluene-d8, -34 °C): δ ) 1.07 (d, 3 Hz, 6H, CH-
(CH3)2), 1.10 (d, 3 Hz, 6H, CH(CH3)2), 1.14 (d, 3 Hz, 6H, CH(CH3)2),
1.18 (d, 3 Hz, 6H, CH(CH3)2), 2.58 (sept, 3 Hz, 2H, CH(CH3)2), 2.99
(sept, 3 Hz, 2H, CH(CH3)2), 3.65 (br s, 2H, Ind), 4.84 (br s, 2H, Ind),
5.83 (s, 1H, Cp), 6.62 (s, 1H, Cp), 6.85 (br s, 2H, Ind), 7.17 (br s, 2H,
Ind). 13C NMR (toluene-d8, -34 °C): δ ) 22.26, 22.30, 26.34, 27.20,
27.62, 29.44 (CH(CH3)2), 69.41, 96.98 (Ind), 102.49, 115.31, 116.02,
120.98, 121.02, 121.27, 123.69, 127.85 (Cp/Ind).
Preparation of (η5-C9H5-1,3-(SiMe2Ph)2)(η9-C9H5-1,3-(SiMe2Ph)2)-
Zr (2). A 100 mL round-bottomed flask was charged with 0.800 g
(0.86 mmol) of 2-Cl2 and approximately 30 mL of toluene forming a
yellow solution. At ambient temperature, 0.121 g (1.89 mmol) of LiCH2-
CHMe2 was added and the reaction mixture was stirred for 1 h. The
resulting burgundy solution was filtered through a pad of Celite, and
the solvent was removed in vacuo, leaving a burgundy oil. Recrystal-
lization from pentane at -35 °C afforded 0.291 g (39%) of 2. Anal.
Calcd for C50H54Si4Zr: C, 69.95; H, 6.34. Found: C, 69.69; H, 5.94.
1H NMR (toluene-d8, -34 °C): δ ) 0.39 (s, 6 H, SiMe2Ph), 0.44 (s,
6 H, SiMe2Ph), 0.70 (br s, 6 H, SiMe2Ph), 0.74 (br s, 6 H, SiMe2Ph),
3.66 (br s, 2H, Ind), 5.56 (br s, 2H, Ind), 5.86 (s, 1H, Cp), 5.97 (s, 1H,
Cp), 6.62 (m, 2H, Ind) 7.13 (br s, 12H, Ph), 7.36 (br s, 8H, Ph), 7.48
(m, 2H, Ind). 13C NMR (toluene-d8, -34 °C) δ ) -0.25 (SiMe2Ph),
-0.09 (SiMe2Ph), 1.35 (SiMe2Ph), 1.55 (SiMe2Ph), 75.10, 95.02 (Ind),
98.25, 110.30, 123.90, 126.57 (Cp), 128.46, 128.68, 129.61, 131.86,
132.98, 134.51, 134.59, 134.85, 137.62, 138.07, 140.26, 140.44 (Ind/
Ph).
Preparation of (η5-C9H5-1,3-(SiMe3)2)(η9-C9H5-1,3-(SiMe2Ph)2))-
Zr (6). This molecule was synthesized using a procedure similar to
that described for 2 with 0.654 g (0.81 mmol) of 6-Cl2 and 0.104 g
(1.62 mmol) of LiCH2CHMe2, yielding 0.129 g (22%) of burgundy
solid 6 as an equimolar mixture of isomers. Anal. Calcd for C40H50-
Si4Zr: C, 65.42; H, 6.86. Found: C, 65.41; H, 7.06. 1H NMR (toluene-
d8, -34 °C): δ ) 0.22 (s, 18H, SiMe3), 0.33 (s, 18H, SiMe3), 0.42 (s,
6H, SiMe2Ph), 0.50 (s, 6H, SiMe2Ph), 0.66 (s, 6H, SiMe2Ph), 0.72 (s,
6H, SiMe2Ph), 3.82 (br s, 2H, Ind), 3.90 (br s, 2H, Ind), 5.35 (br s, 2H,
Ind), 5.41 (br s, 2H, Ind), 5.35 (s, 1H, Cp), 5.41 (s, 1H, Cp), 6.19 (s,
1H, Cp), 6.23 (s, 1H, Cp), 6.67 (m, 2H, Ind), 6.82 (m, 2H, Ind), 7.33
(m, 10H, Ph), 7.40 (m, 10H, Ph), 7.45 (m, 2H, Ind), 7.48 (m, 2H, Ind).
13C NMR (toluene-d8, -34 °C): δ ) -0.38 (SiMe2Ph), -0.14 (SiMe2-
Ph), 1.03 (SiMe2Ph), 1.46 (SiMe3), 2.32 (SiMe3), 74.88, 75.24, 95.03,
95.23 (Ind), 99.97, 102.98, 109.95, 112.45, 118.78, 120.29, 123.50,
123.72 (Cp), 126.85, 131.55, 131.80, 133.11, 133.47, 134.47, 134.62,
134.72, 134.89, 136.51, 138.52, 140.31, 140.35 (Ind/Ph). One SiMe2-
Ph and three Ind/Ph resonances not located.
Preparation of (η5-C9H5-1,3-(SiMe2tBu)2)(η9-C9H5-1,3-(SiMe2tBu)2)-
Zr (3). This molecule was synthesized using a procedure similar to
that described for 2 with 0.545 g (0.64 mmol) of 3-Cl2 and 0.086 g
(1.3 mmol) of LiCH2CHMe2, yielding 0.203 g (40%) of 3 as a burgundy
solid. Anal. Calcd for C42H70Si4Zr: C, 64.79; H, 9.06. Found: C, 64.45;
1
t
H, 8.90. H NMR (benzene-d6): δ ) 0.10 (s, 6H, SiMe2 Bu), 0.24 (s,
t
t
t
Preparation of (η5-C9H5-1,3-(SiMe3)2)(η9-C9H5-1-(SiMe3)-3-(CMe3))-
Zr (7). This molecule was prepared in an identical manner to 2 with
0.769 g (1.16 mmol) of 7-Cl2 and 0.149 g (2.33 mmol) of LiCH2-
CHMe2, yielding 0.282 g (41%) of a burgundy solid identified as 7.
Anal. Calcd for C31H46Si3Zr: C, 62.66; H, 7.80. Found: C, 62.16; H,
6H, SiMe2 Bu), 0.47 (s, 6H, SiMe2 Bu), 0.58 (s, 6H, SiMe2 Bu), 0.81
(s, 36H, SiMe2 Bu), 4.02 (s, 2H, Ind), 5.37 (s, 2H, Ind), 6.20 (s, 1H,
Cp), 6.45 (s, 1H, Cp), 6.77 (s, 2H, Ind), 7.52 (s, 2H, Ind). 13C NMR
(benzene-d6): δ ) -5.12 (2 SiMe2 Bu), -3.08 (SiMe2 Bu), -2.76
(SiMe2tBu), 17.42 (SiMe2CMe3), 19.03 (SiMe2CMe3), 26.46 (SiMe2CMe3),
26.99 (SiMe2CMe3), 75.32, 95.32 (Ind), 101.21, 109.23, 119.82, 122.81
(Cp), 126.76, 131.68, 132.87, 137.90 (Ind).
t
t
t
1
7.52. H NMR (benzene-d6): δ ) 0.23 (s, 9H, SiMe3), 0.28 (s, 9H,
SiMe3), 0.34 (s, 9H, SiMe3), 1.28 (s, 9H, CMe3), 3.82 (t, 3 Hz, 1H,
Ind), 4.07 (t, 3 Hz, 1H, Ind), 5.15 (d, 3 Hz, 1H, Ind), 5.46 (d, 3 Hz,
1H, Ind), 6.09 (s, 1H, Cp), 6.40 (s, 1H, Cp), 6.79 (br s, 2H, Ind), 7.46
(d, 5 Hz, 1H, Ind), 7.60 (d, 5 Hz, 1H, Ind). 13C NMR (benzene-d6): δ
) 1.08 (SiMe3), 1.29 (SiMe3), 2.20 (SiMe3), 33.73 (CMe3), 34.27
(CMe3), 71.81, 74.05, 93.10, 94.89 (Ind), 102.86, 105.96, 108.31,
119.19, 122.70, 122.78, 122.91, 126.16, 126.50, 126.85, 128.03, 131.18,
131.46, 136.69 (Cp/Ind).
Preparation of rac/meso-(η5-C9H5-1-(SiMe3)-3-(CMe3))(η9-C9H5-
1-(SiMe3)-3-(CMe3))Zr (4). This molecule was prepared in an identical
manner to 2 with 0.745 g (1.15 mmol) of rac/meso-4-Cl2 and 0.151 g
(2.36 mmol) of LiCH2CHMe2, yielding 0.140 g (21%) of a burgundy
solid identified as a near equimolar mixture of rac/meso-4. Anal. Calcd
for C32H46Si2Zr: C, 66.48; H, 8.03. Found: C, 65.98; H, 7.98. 1H NMR
(toluene-d8, -34 °C): δ ) 0.21 (s, 9H, SiMe3), 0.30 (s, 9H, SiMe3),
0.34 (s, 9H, SiMe3), 0.38 (s, 9H, SiMe3), 1.23 (s, 9H, CMe3), 1.29 (s,
9H, CMe3), 1.30 (s, 9H, CMe3), 1.34 (s, 9H, CMe3), 3.71 (t, 5 Hz, 1H,
Ind), 3.88 (t, 5 Hz, 1H, Ind), 3.94 (t, 5 Hz, 1H, Ind), 4.21 (t, 5 Hz, 1H,
Ind), 5.11 (m, 2H, Ind), 5.37 (d, 5 Hz, 1H, Ind), 5.49 (d, 5 Hz, 1H,
Ind), 5.98 (s, 1H, Cp), 6.03 (s, 1H, Cp), 6.05 (s, 1H, Cp), 6.23 (s, 1H,
Cp), 6.68-6.75 (br m, 3H, Ind), 6.79 (t, 10 Hz, 1H, Ind), 7.26 (d, 10
Hz, 1H, Ind), 7.37 (d, 10 Hz, 1H, Ind), 7.62 (d, 10 Hz, 1H, Ind), 7.69
(d, 10 Hz, 1H, Ind). 13C NMR (toluene-d8, -34 °C): δ ) 1.53 (SiMe3),
1.86 (SiMe3), 2.36 (SiMe3), 2.55 (SiMe3), 30.28 (CMe3), 33.14 (CMe3),
33.17 (CMe3), 33.48 (CMe3), 33.72 (CMe3), 34.29 (CMe3), 34.46
(CMe3), 34.90 (CMe3), 71.29, 72.66, 73.25, 75.19, 93.58, 93.86, 95.24,
95.95, 96.19, 106.77, 112.09, 112.61 (Cp), 122.07, 122.27, 122.62,
122.69, 122.90, 123.76, 123.95, 124.54, 124.86, 125.05, 125.85, 126.09,
126.27, 126.59, 126.69, 126.74, 127.07, 128.69, 129.62, 136.55, 137.02,
137.62 (Ind). Two Ind resonances not located.
Preparation of (η5-C9H5-1,3-(CHMe2)2)(η9-C9H5-1,3-(CHMe2)2)-
Zr (5). A 100 mL round-bottomed flask was charged with 26.29 g
(0.131 mmol) of mercury and approximately 10 mL of toluene. While
stirring, 0.132 g (5.72 mmol) of sodium metal was added and the
resulting amalgam was stirred for an additional 20 min. A toluene slurry
containing 0.535 g (0.954 mmol) of 5-Cl2 was added. The resulting
reaction mixture was stirred vigorously for 2 days, and the purple
solution was decanted from the amalgam and filtered through Celite.
Removal of the toluene in vacuo followed by recrystallization from
Preparation of (η5-C9H5-1,3-(SiMe3)2)2Zr(η5-CH2C(Me)CH2)H (5-
(η3-CH2C(Me)CH2)H). A 25 mL round-bottomed flask was charged
with 0.105 g (0.214 mmol) of 5 and dissolved in 5 mL of pentane in
a drybox. A 180° needle valve was attached, and the reaction was
degassed before 625 Torr (1.07 mmol) of isobutylene was added by a
calibrated gas bulb. The reaction was stirred overnight before solvent
was removed in vacuo. Extraction with pentane followed by filtration
through a 1 mL frit and solvent removal in vacuo gave a foamy red
solid. Recrystallization from pentane at -35 °C afforded 0.042 g (36%)
of 5-(η3-CH2C(Me)CH2)H as a red solid. Anal. Calcd for C42H70Si4-
1
Zr: C, 74.80; H, 8.49. Found: C, 74.36; H, 8.30. H NMR (toluene-
d8, -38 °C): δ ) -2.66 (s, 1H, CH2CMeCH2), 1.06 (d, 8 Hz, 3H,
CH(CH3)2), 1.09 (d, 8 Hz, 3H, CH(CH3)2), 1.21 (m, 6H, CH(CH3)2),
1.21 (s, 1H, CH2CMeCH2), 1.30 (d, 8 Hz, 3H, CH(CH3)2), 1.34 (d, 8
Hz, 3H, CH(CH3)2), 1.44 (s, 3H, CH2CMeCH2), 1.77 (d, 8 Hz, 3H,
CH(CH3)2), 2.18 (d, 8 Hz, 3H, CH(CH3)2), 2.48 (s, 1H, Zr-H), 2.76
(m, 1H, CH(CH3)2), 2.78 (s, 1H, Cp), 2.87 (m, 1H, CH(CH3)2), 2.90
(s, 1H, CH2CMeCH2), 3.20 (m, 1H, CH(CH3)2), 3.59 (s, 1H, Cp), 3.92
(s, 1H, CH2CMeCH2), 3.99 (m, 1H, CH(CH3)2), 6.57 (m, 1H, Ind),
6.65 (m, 1H, Ind), 6.87 (br s, 3H, Ind), 7.51 (m, 1H, Ind), 7.57 (br s,
2H, Ind). 13C NMR (toluene-d8, -38 °C): δ ) 22.59, 23.46, 24.25,
25.04, 25.89, 26.41, 26.84, 27.03 (CH(CH3)2), 28.21 (CH2CMeCH2),
28.73, 29.33, 29.67, 31.88, 35.00 (CH(CH3)2), 58.19 (CH2CMeCH2),
70.66 (CH2CMeCH2), 105.89 (Cp), 111.25 (Cp/Ind), 112.67 (Cp/Ind),
9
J. AM. CHEM. SOC. VOL. 126, NO. 51, 2004 16949

A R T I C L E S
Bradley et al.
112.81 (Cp), 115.65, 117.32, 118.91, 120.47, 121.69, 121.83, 121.98,
122.18, 122.58, 123.17, 134.10 (Cp/Ind). Three Cp/Ind resonances not
located.
120.50, 120.80, 121.39, 121.68, 121.79, 122.15, 122.77, 123.28, 123.51,
123.69, 125.06 (Cp/Ind).
Preparation of (η5-C9H5-1,3-(CHMe2)2)2Zr(η3-CH2CHC(Me)H)-
H (5-(η3-CH2C(Me)CH2)H). This molecule was prepared in an similar
manner as 5-(η5-CH2C(Me)CH2)H with 0.015 g (0.031 mmol) of 5
and 18 Torr (0.30 mmol) of 1-butene. 1H NMR (toluene-d8, -38 °C):
δ ) -1.01 (d, 10 Hz, 1H, CH2CHCHCH3), 1.18 (m, 1H, CH2-
CHCHCH3), 1.23 (d, 8 Hz, 3H, CH(CH3)2), 1.25 (d, 8 Hz, 3H, CH-
(CH3)2), 1.29 (d, 8 Hz, 3H, CH(CH3)2), 1.38 (d, 8 Hz, 3H, CH(CH3)2),
1.43 (d, 8 Hz, 3H, CH(CH3)2), 1.46 (d, 8 Hz, 3H, CH(CH3)2), 1.56 (d,
8 Hz, 3H, CH(CH3)2), 1.60 (d, 8 Hz, 3H, CH(CH3)2), 1.80 (d, 10 Hz,
1H, CH2CHCHCH3), 2.02 (br s, 1H, CH2CHCHCH3), 2.17 (br s, 3H,
CH2CHCHCH3), 3.07 (m, 1H, CH(CH3)2), 3.23 (m, 2H, CH(CH3)2 and
Zr-H), 3.45 (m, 2H, CH(CH3)2), 4.51 (s, 1H, Cp), 4.95 (s, 1H, Cp),
6.54 (m, 1H, Ind), 6.64 (m, 1H, Ind), 6.68 (m, 1H, Ind), 6.77 (m, 1H,
Ind), 6.98 (d, 12 Hz, 1H, Ind), 7.05 (d, 12 Hz, 1H, Ind), 7.20 (d, 12
Hz, 1H, Ind), 7.39 (d, 12 Hz, 1H, Ind). 13C NMR (toluene-d8, -38
°C): δ ) 13.67, 21.58, 22.90, 23.92, 25.23, 25.35, 25.63, 26.22, 26.40,
27.58, 28.28, 28.49, 29.00, 29.45, 30.16, 52.57 (CH(CH3)2/CH2-
CHCHCH3), 95.10, 105.05, 115.54, 116.81, 117.53, 118.03, 118.34,
Acknowledgment. We thank the Cornell University Depart-
ment of Chemistry and Chemical Biology, the donors of the
Petroleum Research Fund administered by the American
Chemical Society, and the National Science Foundation (pre-
doctoral fellowship to C.A.B. and CAREER Award to P.J.C.)
for generous financial support. P.J.C. is also a Cottrell Scholar
and acknowledges financial support from the Research Corpora-
tion.
Supporting Information Available: Additional experimental
procedures, NMR spectroscopic data, computational results, and
crystallographic data for 3, 5, and 5-(η3-CH2C(Me)CH2)H
including full atom labeling schemes and complete bond
distances and angles (PDF, CIF). This material is available free
of charge via the Internet at http://pubs.acs.org.
JA045072V
9
16950 J. AM. CHEM. SOC. VOL. 126, NO. 51, 2004