Carbon-oxygen bond cleavage with η9,η5- bis(indenyl)zirconium sandwich complexes

Article abstract of DOI:10.1021/ja065456g

Treatment of the η⁹,η⁵-bis(indenyl)zirconium sandwich complex, (η⁹-C₉H₅-1,3-(SiMe ₃)₂)(η⁵-C₉H₅-1,3- (SiMe₃)₂)Zr, with dialkyl ethers such as diethyl ether, CH₃OR (R = Et, ⁿBu, ^tBu), ⁿBu ₂O, or ⁱPr₂O resulted in facile C-O bond scission furnishing an η⁵,η⁵-bis(indenyl)zirconium alkoxy hydride complex and free olefin. In cases where ethylene is formed, trapping by the zirconocene sandwich yields a rare example of a crystallographically characterized, base-free η⁵, η⁵-bis(indenyl)zirconium ethylene complex. Observation of normal, primary kinetic isotope effects in combination with rate studies and the stability of various model compounds support a mechanism involving rate-determining C-H activation to yield an η⁵,η⁵- bis(indenyl)-zirconium alkyl hydride intermediate followed by rapid β-alkoxide elimination. For isolable η⁶,η⁵- bis(indenyl)-zirconium THF compounds, thermolysis at 85 °C also resulted in C-O bond cleavage to yield the corresponding zirconacycle. Both mechanistic and computational studies again support a pathway involving haptotropic rearrangement to η⁵,η⁵-bis(indenyl)zirconium intermediates that promote rate-determining C-H activation and ultimately C-O bond scission.

Full text of DOI:10.1021/ja065456g

Published on Web 12/08/2006
Carbon-Oxygen Bond Cleavage with
η9,η5-Bis(indenyl)zirconium Sandwich Complexes
Christopher A. Bradley,† Luis F. Veiros,‡ Doris Pun,† Emil Lobkovsky,†
Ivan Keresztes,† and Paul J. Chirik*,†
Contribution from the Department of Chemistry and Chemical Biology, Baker Laboratory,
Cornell UniVersity, Ithaca, New York 14853, and Centro de Quimica Estrutural, Complexo I,
Instituto Superior Te´cnico, AV. RoVisco Pais 1, 1049-001 Lisbon, Portugal
Received July 28, 2006; E-mail: pc92@cornell.edu
Abstract: Treatment of the η9,η5-bis(indenyl)zirconium sandwich complex, (η9-C9H5-1,3-(SiMe3)2)(η5-C9H5-
1,3-(SiMe3)2)Zr, with dialkyl ethers such as diethyl ether, CH3OR (R d Et, nBu, tBu), nBu2O, or iPr2O resulted
in facile CsO bond scission furnishing an η5,η5-bis(indenyl)zirconium alkoxy hydride complex and free
olefin. In cases where ethylene is formed, trapping by the zirconocene sandwich yields a rare example of
a crystallographically characterized, base-free η5,η5-bis(indenyl)zirconium ethylene complex. Observation
of normal, primary kinetic isotope effects in combination with rate studies and the stability of various model
compounds support a mechanism involving rate-determining CsH activation to yield an η5,η5-bis(indenyl)-
zirconium alkyl hydride intermediate followed by rapid â-alkoxide elimination. For isolable η6,η5-bis(indenyl)-
zirconium THF compounds, thermolysis at 85 °C also resulted in CsO bond cleavage to yield the
corresponding zirconacycle. Both mechanistic and computational studies again support a pathway involving
haptotropic rearrangement to η5,η5-bis(indenyl)zirconium intermediates that promote rate-determining Cs
H activation and ultimately CsO bond scission.
Introduction
are stabilized by the introduction of sterically demanding silyl
substituents (R d SiMe3, SiMe2 Bu, SiMe2Ph, etc.) on the
t
The rich chemistry of low oxidation state, reducing group 4
transition metal complexes continues to attract considerable
attention given the utility of these compounds to promote
numerous organic transformations1 and small molecule activa-
tion processes.2 Substituted bis(cyclopentadienyl) complexes of
titanium and zirconium (group 4 metallocenes) have been the
most extensively studied, owing to their commercial availability,
steric and electronic modularity, and convenient synthetic access
to low oxidation states by treatment with alkyl lithium reagents3
or alkali metal reductants.4
cyclopentadienyl rings. Subsequent studies from our laboratory
have demonstrated that exposure of many of these compounds
to dinitrogen at low temperature yields monomeric bent ti-
tanocene mono- and bis(dinitrogen) complexes, depending on
the specific cyclopentadienyl substituent.8 This chemistry
has recently been extended to the isolobal and isoelectronic
carbon monoxide derivatives, (η5-C5Me4R)2Ti(CO), and the
observation of mixed titanocene dinitrogen, carbonyl com-
pounds.9
Such an approach has not been successful for the preparation
of the corresponding bis(cyclopentadienyl)zirconium(II) sand-
wich compounds.10 The larger metallic atomic radius and greater
thermodynamic driving force to achieve the +4 oxidation state
make such targets more challenging to access. Our laboratory
has discovered that replacing the cyclopentadienyl anions with
related six π-electron, 1,3-disubstituted indenyl ligands has
allowed observation11 and ultimately complete spectroscopic and
structural characterization12 of the first bis(indenyl)zirconium
The first isolable examples5 of bis(cyclopentadienyl)titanium-
(II) sandwiches were independently reported by Lawless6 and
Mach.7 These compounds of the general form (η5-C5Me4R)2Ti
† Cornell University.
‡ Instituto Superior Te´cnico.
(1) For a recent review, see: Titanium and Zirconium in Organic Synthesis;
Marek, I., Ed.; Wiley-VCH: Weinheim, 2002.
(2) (a) Pool, J. A.; Chirik, P. J. Can. J. Chem. 2005, 83, 286. (b) Fryzuk, M.
D.; Johnson, S. A. Coord. Chem. ReV. 2000, 200, 379.
(3) (a) Negishi, E.-I.; Takahashi, T. Acc. Chem. Res. 1994, 27, 124. (b)
Rosenthal, U.; Pellny, P.-M.; Kirchbauer, F. G.; Burkalov, V. V. Acc. Chem.
Res. 2000, 33, 119.
(7) Lukesova´, L.; Hora´cek, M.; Stepnicka, P.; Fejfarova´, K.; Gyepes, R.;
Cisorova´, I.; Kubista, J.; Mach, K. J. Organomet. Chem. 2002, 663, 134.
(8) Hanna, T. E.; Lobkovsky, E.; Chirik, P. J. J. Am. Chem. Soc. 2004, 126,
14688.
(4) (a) Gambarotta, S.; Scott, J. Angew. Chem., Int. Ed. 2004, 43, 5298. (b)
Fryzuk, M. D. Chem. Rec. 2003, 3, 2.
(5) For observation of bis(cyclopentadienyl)titanium sandwich complexes in
solution see: (a) Bercaw, J. E.; Brintzinger, H. H. J. Am. Chem. Soc. 1971,
93, 2045. (b) Bercaw, J. E. J. Am. Chem. Soc. 1974, 96, 5087. (c) deWolf,
J. M.; Blaauw, R.; Meetsma, A.; Teuben, J. H.; Gyepes, R.; Varga, V.;
Mach, K.; Veldman, N.; Spek, A. L. Organometallics 1996, 15, 4977. (d)
Hanna, T. E.; Keresztes, I.; Lobkovsky, E.; Bernskoetter, W. H.; Chirik,
P. J. Organometallics 2004, 23, 3448.
(9) Hanna, T. E.; Lobkovsky, E.; Chirik, P. J. J. Am. Chem. Soc. 2006, 128,
6018.
(10) Hora´cek, M.; Stepnicka, P.; Kubista, J.; Fejfarova´, K.; Gyepes, R.; Mach,
K. Organometallics 2003, 22, 861.
(11) Bradley, C. A.; Lobkovsky, E.; Chirik, P. J. J. Am. Chem. Soc. 2003, 125,
8110.
(6) Hitchcock, P. B.; Kerton, F.; Lawless, G. A. J. Am. Chem. Soc. 1998, 120,
10264.
(12) Bradley, C. A.; Keresztes, I.; Lobkovsky, E.; Young, V. G.; Chirik, P. J.
J. Am. Chem. Soc. 2004, 126, 16937.
9
16600
J. AM. CHEM. SOC. 2006, 128, 16600-16612
10.1021/ja065456g CCC: $33.50 © 2006 American Chemical Society

η9
,
η
5-Bis(indenyl)Zr Sandwich Compds for C
−O Bond Cleavage
A R T I C L E S
Results and Discussion
Carbon-Oxygen Bond Cleavage in Dialkyl Ethers. Interest
in the reactivity of η9,η5-bis(indenyl)zirconium sandwich com-
pounds with dialkyl ethers stems from the observation of ligand-
induced η9 to η6 haptotropic rearrangement upon addition of
THF or DME (DME ) 1,2-dimethoxyethane).17 Computational
studies suggest that migration of the metal to the benzo rather
than cyclopentadienyl ring allows additional backbonding with
the six-membered ring thereby stabilizing the η6 over the η5
haptomer.18 In this study, the scope of this transformation was
explored with simple, commercially available dialkyl ethers.
Stirring either of the silyl-substituted η9,η5-bis(indenyl)-
zirconium sandwich complexes, 1 or 2, in diethyl ether at 23
°C resulted in cleavage of the CsO bond to form an equimolar
mixture of the zirconocene ethoxy hydrides, (η5-C9H5-1,3-
(SiMe2R)2)2Zr(OCH2CH3)H (R ) Me, 1-(OEt)H; R ) Ph,
2-(OEt)H), and the corresponding ethylene compounds, (η5-
C9H5-1,3-(SiMe2R)2)2Zr(η2-CH2dCH2) (R ) Me, 1-(η2-CH2d
CH2); R ) Ph, 2-(η2-CH2dCH2)) (eq 2).
Figure 1. Bonding description for η9,η5-bis(indenyl)zirconium sandwich
compounds.
sandwich complexes. As was computationally predicted,13 these
compounds do not adopt the familiar parallel ring structure
typically associated with bis(cyclopentadienyl)metal sandwich
complexes but rather contain one “η9” indenyl ring where the
carbocycle is buckled and all nine carbons of the ligand are
coordinated to the zirconium.14 A more detailed investigation
into the electronic structure of these compounds13 has established
a bonding description between two limiting canonical forms.
In one extreme, the η9,η5-bis(indenyl)zirconium sandwich can
be viewed as a Zr(II), d2 species with a neutral benzo ligand
while in the other as a Zr(IV), d0 complex where the benzo
ring is reduced by two electrons to form an LX2 fragment
(Figure 1).
In light of these results, the possibility for these molecules
to serve as sources of [η5-(C9H5-1,3-R2)2ZrII]15 by dissociation
of the benzo portion of the η9 indenyl ring has been an area of
considerable experimental and theoretical interest. Both line
broadening and EXSY NMR experiments have established that
the η9,η5-bis(indenyl)zirconium sandwich complexes undergo
q
facile (∆G296 ≈ 14-20 kcal/mol) ring interchange in solution,
most likely involving the desired [η5-(C9H5-1,3-R2)2ZrII] inter-
mediate (eq 1).12 These experiments have been supported by
For diethyl ether cleavage with 1, recrystallization of the
product mixture from pentane at -35 °C furnished crystals of
pure 1-(η2-CH2dCH2). Attempts to isolate the remaining
1-(OEt)H from the mother liquor have been unsuccessful due
to cocrystallization with 1-(η2-CH2dCH2). The identity of the
zirconocene ethylene compound was further established by
independent synthesis, as treatment of 1 with 1 equiv of CH2d
computational studies, where a dissociative mechanism generat-
ing an [η5-(C9H5-1,3-R2)2ZrII] (S ) 0) species was found to be
the lowest energy pathway.16
1
CH2 also afforded 1-(η2-CH2dCH2). The H and 13C NMR
spectra of 1-(η2-CH2dCH2) exhibit the number of peaks for a
C
2V symmetric bent zirconocene derivative with two equivalent
In this contribution, the reactivity of η9,η5-bis(indenyl)-
zirconium sandwich complexes toward dialkyl ethers and THF
is explored both experimentally and computationally. Carbon-
oxygen bond cleavage has been observed under mild conditions
in solution and was found to proceed through reducing η5,η5-
bis(indenyl)zirconium intermediates that promote initial C-H
activation followed by C-O bond scission.
η5,η5-coordinated indenyl rings. One notable feature of the
benzene-d6 1H NMR spectrum of 1-(η2-CH2dCH2) is the
observation of a broad singlet centered at 0.87 ppm for the
vinylic hydrogens of the coordinated olefin. Confirmation of
this assignment was provided by the preparation of 1-(η2-CD2d
CD2) from the cleavage of diethyl ether-d10 or by treatment of
1 with CD2dCD2.
Single crystals of 1-(η2-CH2dCH2) were obtained from a
concentrated pentane solution chilled to -35 °C. Two inde-
pendent, nearly identical molecules were found in the asym-
metric unit, one of which is shown in Figure 2. Selected metrical
parameters are reported in Table 1. Consistent with the solution
NMR data, the indenyl rings are η5 coordinated with a rotational
(13) Veiros, L. F. Chem.sEur. J. 2005, 11, 2505.
(14) For a review on indenyl hapticity, see: Calhorda, M. J.; Roma˜o, C. C.;
Veiros, L. F. Chem.sEur. J. 2002, 8, 868.
(15) For recent examples of accessing low-valent early transition metals by
dissociation of a ligand, see: (a) Korobkov, I.; Arunachalampillai, A.;
Gambarotta, S. Organometallics 2004, 23, 6248. (b) Korobkov, I.;
Gambarotta, S.; Yap, G. P. A. Angew. Chem., Int. Ed. 2003, 42, 814. (c)
Evans, W. J.; Davis, B. L. Chem. ReV. 2002, 102, 2119. (d) Diaconescu,
P.; Arnold, P. L.; Baker, T.; Mindiola, D. J.; Cummins, C. C. J. Am. Chem.
Soc. 2000, 12, 6108. (e) Tsai, Y.-C.; Johnson, M. A.; Mindiola, D. J.;
Cummins, C. C.; Klooster, W. T.; Koetzle, T. F. J. Am. Chem. Soc. 1999,
121, 10426.
(17) Bradley, C. A.; Kerestzes, I.; Lobkvosky, E.; Chirik, P. J. J. Am. Chem.
Soc. 2005, 127, 10291.
(18) Veiros, L. F. Organometallics 2006, 25, 4698.
(16) Veiros, L. F. Organometallics 2006, 25, 2266.
9
J. AM. CHEM. SOC. VOL. 128, NO. 51, 2006 16601

A R T I C L E S
Bradley et al.
Figure 2. Molecular structure of 1-(η2-CH2dCH2) (left) at 30% probability ellipsoids. Hydrogen atoms omitted for clarity. View of the core of the molecule
(right).
analysis, and degradation experiments. Similar to 1-(η2-CH2d
Table 1. Selected Bond Distances (Å) and Angles (deg) for
1-(η2-CH2dCH2)
CH2), 2-(η2-CH2dCH2) exhibited a singlet centered at 0.53 ppm
C(19)-C(20)
Zr(1)-C(19)
Zr(1)-C(20)
Ra
1.443(5)
2.285(3)
2.278(3)
63(3)
0.100(2), 0.1176(3)
3.2(4), 3.3(2)
116.8(3)
for the vinylic hydrogens of the coordinated ethylene, consistent
with zirconocyclopropane character. On a preparative scale,
recrystallization of the product mixture from pentane at -35
°C yielded pure 2-(OEt)H, opposite of the purification product
isolated from 1. Attempts to isolate pure 2-(η2-CH2dCH2) have
been unsuccessful due to contamination from 2-(OEt)H. For
both 1 and 2, attempts were made to detect an intermediate
η6,η5-bis(indenyl)zirconocene diethyl ether complex by 1H NMR
spectroscopy. However, addition of a large excess (>50 equiv)
of Et2O to a benzene-d6 solution of either 1 or 2 produced Cs
O cleavage products with no spectroscopic evidence for
intermediates.
Cleavage of the C-O bond in diethyl ether was also
accomplished with the more electron-rich, alkylated η9,η5-bis-
(indenyl)zirconium sandwich, (η9-C9H5-1,3-(CHMe2)2)(η5-C9H5-
1,3-(CHMe2)2)Zr (5).23 Stirring 5 in neat diethyl ether for 36 h
at 45 °C yielded the zirconocene ethoxy hydride (η5-C9H5-1,3-
(CHMe2)2)Zr(OEt)H (5-(OEt)H) and the metallacycle 5-(c-
C4H8), arising from coupling of 2 equiv of ethylene (eq 3). A
2:1 ratio of 5-(OEt)H to 5-(c-C4H8) was consistently observed.
Previous studies from our laboratory have demonstrated that
the less sterically congested sandwich compound promotes more
facile coupling of olefins and alkynes.12 Indeed, addition of
slip distortionsb
fold angles
rotational angle
a Angle defined by the normals of the C-H-H planes. b For definitions
of these parameters, see ref 24.
angle of 116.8(3)°. As expected for a bent zirconocene deriva-
tive, the ethylene ligand lies in the plane of the metallocene
wedge with a C(19)s(20) bond distance of 1.443(5) Å and an
R value of 63(3)°, consistent with the significant contribution
from the zirconacyclopropane descriptor. This view is also
corroborated by the 1H NMR chemical shift, where the vinylic
protons of the coordinated olefin are substantially shifted upfield
from the typical alkene region of the spectrum.
While implicated as intermediates in zirconocene-promoted
coupling reactions, isolable and crystallographically character-
ized base-free olefin complexes of group 4 metallocenes are
rare.19 Despite the absence of a supporting ligand, the carbon-
carbon bond distance in 1-(η2-CH2dCH2) is statistically
indistinguishable from the values of 1.451(5) and 1.455(9) Å
reported for the base-stabilized complexes, (η5-C9H7)2Zr(η2-
CH2dCH2)(THF) and (η5-C5H5)2Zr(η2-CH2dCH2)(THF), re-
spectively.20 These values are also in agreement with an ansa-
zirconocene ethylene hydride anion21 as well as the permethyltita-
nocene ethylene complex, (η5-C5Me5)2Ti(η2-CH2dCH2).22
Characterization of the diethyl ether cleavage products from
2 was also accomplished by NMR spectroscopy, combustion
(19) For an example of a crystallographically characterized base-free ansa-
zirconocene alkyne complex, see: Lefeber, C.; Baumann, W.; Tillack, A.;
Kempe, R.; Go¨rls, H.; Rosenthal, U. Organometallics 1996, 15, 3486.
(20) Fischer, R.; Walther, D.; Gebhardt, P.; Go¨rls, H. Organometallics 2000,
19, 2532.
(21) Lee, H.; Hascall, T.; Desrosiers, P. J.; Parkin, G. J. Am. Chem. Soc. 1998,
120, 5830.
(22) Cohen, S. A.; Auburn, P. R.; Bercaw, J. E. J. Am. Chem. Soc. 1983, 105,
1136.
9
16602 J. AM. CHEM. SOC. VOL. 128, NO. 51, 2006

η9
,
η
5-Bis(indenyl)Zr Sandwich Compds for C
−O Bond Cleavage
A R T I C L E S
Figure 3. Products of methyl alkyl ether cleavage with 1.
ethylene to 5 in benzene-d6 yielded only 5-(c-C4H8) independent
of the amount of olefin employed. Qualitatively, the rates of
diethyl ether cleavage with 1 and 2 were indistinguishable,
whereas those with 5 was slower.
(eq 4). The rest of the material was starting sandwich 1 and
excess dimethyl ether.
The observation of carbon-oxygen bond cleavage of diethyl
ether to yield well-defined zirconocene products prompted
further exploration into the scope of the reaction. The silylated
sandwich, 1, was used for the majority of subsequent studies
due to its relative ease of synthesis and high symmetry. In certain
instances, the chemistry of 5 was also examined due to its
smaller indenyl substituents and more electron-rich zirconium
center.24
Methyl alkyl ethers, MeOR (R ) Et, nBu, tBu), were initially
studied. Selective C-O bond cleavage of the ethers was
observed when large excesses, either neat solution or 50 equiv
in benzene-d6, were stirred with 1. In each case, the zirconocene
methoxy hydride compound, (η5-C9H5-1,3-(SiMe3)2)2Zr(OCH3)H
(1-(OMe)H), was observed along with a corresponding olefin
byproduct (Figure 3).
For MeOEt, the olefin product of CsO cleavage is ethylene
which is readily trapped by free sandwich compound to yield
1-(η2-CH2dCH2). In the case of methyl n-butyl ether cleavage,
a thermodynamic mixture of free cis- and trans-2-butene was
Diagnostic resonances17 for η6-indenyl coordination were
observed at 3.36 and 3.55 ppm assigned to a cyclopentadienyl
hydrogen on the η5 ring and a benzo hydrogen on the η6 ligand,
respectively. Over the course of minutes at 23 °C, a new Cs
symmetric product identified as the zirconocene alkyl hydride,
(η5-C9H5-1,3-(SiMe3)2)2Zr(CH2OCH3)H (1-(CH2OCH3)H), was
observed (eq 4). This compound arises from C-H activation
of an ether methyl group and exhibits cyclopentadienyl and
benzo hydrogens in the 6.50 to 7.50 ppm chemical shift region,
indicative of common η5,η5 indenyl hapticity. Exposure of the
product mixture to a vacuum resulted in the loss of dimethyl
ether, from either dissociation or reductive elimination, and
allowed isolation of the sandwich, 1. Warming the reaction
mixture to 85 °C for several days produced a complex mixture
of products, the majority of which contained 1, the correspond-
ing zirconocene cyclometalated hydride and 1-(CH2OCH3)H.
Importantly, no C-O bond cleavage was observed under these
conditions.
1
identified by H NMR spectroscopy. Previous work from our
laboratory has demonstrated that 1 isomerizes terminal olefins
to internal olefins, most likely through zirconocene crotyl
hydride intermediates.12 Once the 2-butenes are formed, the
olefins are sufficiently hindered such that coordination to
zirconium is not observed. This trend also holds for MeOtBu
(MTBE) cleavage, where 1 equiv of free isobutene was
observed.
Treatment of 1 with dimethyl ether produced a different
outcome. Monitoring the addition of 50 equiv of Me2O to a
1
benzene-d6 solution of 1 by H NMR spectroscopy revealed
approximately 10% conversion to the η6,η5-bis(indenyl)zirco-
nium dimethyl ether adduct (1-OMe2) after 15 min at 23 °C
(23) The labeling scheme used in this paper is adapted from ref 12.
(24) Bradley, C. A.; Flores-Torres, S.; Lobkovsky, E.; Abrun˜a, H. D.; Chirik,
P. J. Organometallics 2004, 23, 5332.
9
J. AM. CHEM. SOC. VOL. 128, NO. 51, 2006 16603

A R T I C L E S
Bradley et al.
Figure 4. Products of longer chain dialkyl ether cleavage with 1.
Table 2. Pseudo-First-Order Rate Constants Measured for Et2O
Cleavage with 1 at 45 °Ca
Observation of CsH activation with dimethyl ether prompted
more careful monitoring of the addition of other methyl alkyl
ethers to 1. Addition of 10 equiv of methyl ethyl ether or 50
equiv of methyl n-butyl ether to 1 in benzene-d6 allowed
observation of small quantities (10-30%) of the corresponding
zirconocene alkyl hydride compounds, (η5-C9H5-1,3-(SiMe3)2)2-
1
equiv of Et2O
kobs
×
103 (s-
)
10
25
50
75
1.1(2)
1.4(2)
2.2(3)
4.1(3)
n
Zr(CH2OR)H (R ) Et, 1-(CH2OEt)H; Bu, 1-(CH2OnBu)H),
a Each experiment was conducted with a 0.03 M solution of 1 in benzene-
over the course of minutes at 23 °C (eq 5).
d6.
Attempts to cleave (Me3Si)2O were unsuccessful; heating 1 to
95 °C in the presence of a large excess of ether produced no
Si-O cleavage products after 2 weeks.
To probe the mechanism of dialkyl ether cleavage, several
kinetic and isotopic labeling experiments were performed. Rate
constants for diethyl ether cleavage by 1 were measured at 45
°C in benzene-d6 solution by 1H NMR spectroscopy with Et2O
excesses ranging between 10 and 75 equiv. Under these
conditions, clean pseudo-first-order processes were obtained (see
Supporting Information). The observed rate constants from these
experiments are reported in Table 2. The nonlinear, nearly
sigmoidal dependence of the observed rate constant is most
likely a result of competing pathways involving diethyl ether
cleavage and coordination. One path is a productive mechanism
involving initial C-H bond cleavage ultimately leading to C-O
scission while the others are most likely due to ligand induced
haptotropic rearrangement to form η6,η5-bis(indenyl)zirconium
mono- and bis(ether) complexes. Previous studies have shown
that addition of 1 and 2 equiv of THF to 1 results in a similar
outcome where both processes are dependent on the THF
concentration.17
The kinetic isotope effect (kH/kD) for dialkyl ether cleavage
was also measured for both 1 and 2. Using Et2O as a
representative substrate, benzene-d6 solutions of 1 or 2 were
prepared with an equimolar mixture of diethyl ether and diethyl
ether-d10. A combination of 1H and 2H NMR spectroscopy was
used to confirm the formation of the expected products:
1-(OEt)H, 1-(OEt-d5)D, 1-(η2-CH2dCH2), and 1-(η2-CD2d
CD2). The kinetic isotope effects were measured by monitoring
the CsO bond cleavage as a function of time by integration of
the ZrsH resonance of 1-(OEt)H versus the indenyl ring
For both zirconocene alkyl hydride compounds, diagnostic
resonances upfield of 0 ppm were observed for the diastereotopic
hydrogens on the carbon directly bound to the zirconium. No
benzo or cyclopentadienyl peaks were present between 3 and 5
ppm that are indicative of η6 indenyl coordination. Additional
evidence for C-H activation was provided by degradation
experiments. Treatment of 1-(CH2OEt)H or 1-(CH2OnBu)H
with an excess of gaseous deuterium chloride afforded ROCH2D
and 1-Cl2. No 1H NMR spectroscopic evidence for C-H
activated products or η6,η5-bis(indenyl)zirconium ether com-
plexes was obtained upon treatment of 1 with a large excess of
methyl tert-butyl or diethyl ether.
Carbon-oxygen bond cleavage in longer chain dialkyl ethers
was also explored. Treatment of a benzene-d6 solution of 1 with
20 equiv of ethyl n-butyl ether produced an equimolar mixture
of (η5-C9H5-1,3-(SiMe3)2)2Zr(OnBu)H (1-(OnBu)H) and 1-(η2-
CH2dCH2) over the course of 3 days at 23 °C (Figure 4).
Thermolysis with di-n-butyl ether at 85 °C under similar
conditions furnished 1-(OnBu)H and a mixture of cis- and trans-
2-butene. Likewise, heating a benzene-d6 solution of 1 to 95
i
°C in the presence of 25 equiv of Pr2O yielded (η5-C9H5-1,3-
(SiMe3)2)2Zr(OiPr)H (1-(OiPr)H) after 5 days. The presumed
olefin byproduct, propene, was not detected by 1H NMR
spectroscopy.
9
16604 J. AM. CHEM. SOC. VOL. 128, NO. 51, 2006

η9
,
η
5-Bis(indenyl)Zr Sandwich Compds for C
−O Bond Cleavage
A R T I C L E S
Figure 5. Mechanistic possibilities for dialkyl ether cleavage with η9,η5-bis(indenyl)zirconium sandwich compounds.
protons for both isotopologs of the bis(indenyl)zirconium ethoxy
hydride(deuteride) compounds. Resonances for the two isoto-
pologs of the ethylene compounds could not be used as
competing isotopic exchange with the [SiMe3] complicated the
integration values. A similar control, whereby a benzene-d6
solution of 1-(OEt)H and 1-(OEt-d5)D was allowed to stand
at 23 °C for several days, produced no isotopic exchange. From
the average of four independent trials, a normal primary isotope
effect (kH/kD) of 5.0(5) was measured at 23 °C. A statistically
indistinguishable value of kH/kD of 4.7(4) was determined for 2
from seven independent trials. These data clearly indicate that
a CsH bond cleavage event precedes CsO bond scission either
in or prior to the rate-determining step.
Wolczanski27 and Parkin28 have provided a precedent for
γ-C-H activation in group 4 transition metal chemistry with
titanium amide and zirconium oxo complexes, respectively.
Other possibilities for C-H activation, especially those similar
to Path A, leading to C-O scission are possible from η6 indenyl
intermediates, but only one is shown for convenience. The final
mechanism, Path C, involves direct oxidative addition of the
C-O bond to form an intermediate zirconocene alkoxide alkyl
complex which then undergoes â-hydride elimination to form
ethylene (trapped by 1) and the observed zirconocene ethoxy
hydride compound.
In an attempt to definitively exclude the possibility of Path
C, the synthesis of a putative zirconocene alkoxy alkyl complex
was targeted. Alkylation of 1-(OMe)Cl, prepared from methyl
vinyl ether addition to 1-(H)Cl, with nBuLi yielded the
zirconocene methoxy butyl compound (η5-C9H5-1,3-(SiMe3)2)2-
Zr(OMe)nBu (1-(OMe)nBu), the requisite intermediate arising
from C-O oxidative addition from MeOnBu cleavage. Heating
a benzene-d6 solution of 1-(OMe)nBu to 65 °C for 2 days
produced no change in the 1H NMR spectrum. Recall that
cleavage of MeOnBu occurred at 23 °C in a similar time frame,
definitely excluding Path C as the mechanism of dialkyl ether
cleavage. The selectivity of the MeOR cleavage reactions is
also inconsistent with a direct C-O oxidative addition pathway.
Selective formation of 1-(OMe)H would not be expected for
the series of ethers (Figure 3).
At this point, it is useful to consider possible mechanisms
for dialkyl ether cleavage. Three general mechanistic proposals
are presented in Figure 5 using diethyl ether as a representative
substrate and 1 as the η9,η5-bis(indenyl)zirconium sandwich.
Path A involves C-H activation by 1, through a transient η5,η5-
bis(indenyl)zirconium sandwich (not shown), to yield a zir-
conocene alkyl hydride complex. Subsequent â-alkoxide elimi-
nation25,26 liberates free ethylene which is trapped by unreacted
1 to yield the observed zirconocene ethoxy hydride product.
C-H activation at the position adjacent to the oxygen is shown,
though activation of other positions along the chain are also
plausible. The second pathway involves η6,η5-bis(indenyl)-
zirconium intermediates and must be considered as these species
have been isolated from addition of THF or DME to 1.17 In
Path B, only one possibility involving η6 haptomers is il-
lustrated: γ-C-H activation promoted by the η6 indenyl ligand
followed by proton transfer and haptotropic rearrangement. Both
The feasibility of the C-H activation/â-alkoxide elimination
sequence found in Path A (and possibly in a variant of Path B)
was probed by a series of olefin insertion studies.29 The synthesis
of a stable bis(indenyl)zirconium dihydride complex, (η5-C9H5-
(27) Schaller, C. P.; Cummins, C. C.; Wolczanski, P. T. J. Am. Chem. Soc.
1996, 118, 591.
(25) Strazisar, S. A.; Wolczanski, P. T. J. Am. Chem. Soc. 2001, 123, 4728.
(26) For examples of â-chloride eliminations in group 4 metallocene chemistry,
see: Stockland, R. A.; Foley, S. R.; Jordan, R. F. J. Am. Chem. Soc. 2003,
125, 796.
(28) Howard, W. A.; Waters, M.; Parkin, G. J. Am. Chem. Soc. 1993, 115, 4917.
(29) For synthesis of similar compounds in cyclopentadienyl-based chemistry,
see: Buchwald, S. L.; Nielsen, R. B.; Dewan, J. C. Organometallics 1998,
7, 2324.
9
J. AM. CHEM. SOC. VOL. 128, NO. 51, 2006 16605

A R T I C L E S
Bradley et al.
Figure 6. Evaluation of â-alkoxide elimination with 5-H2.
1,3-(CHMe2)2)2ZrH2 (5-H2), was accomplished by oxidative
addition of H2 to 5 at 23 °C.30 Importantly, this procedure allows
isolation of the dihydride where the benzo rings remain intact
and do not undergo hydrogenation to the corresponding tet-
rahydroindenyl derivative. Ring hydrogenation was only ob-
served with continued exposure to H2 at 45 °C.
Because certain donors such as THF and DME are known to
induce haptotropic rearrangement to form isolable η6,η5-bis-
(indenyl)zirconium ligand derivatives, pathways that include
such haptomers must be considered. As mentioned previously,
the observation of η5,η5 haptomers arising from C-H activation
argues against the formation of such species. In addition, the
alkyl-substituted sandwich, 5, does not coordinate THF to form
the η6,η5-bis(indenyl)zirconium THF complex even in neat
THF.17 Initial spectroscopic17 and subsequent computational
studies16 demonstrate that a stronger metal-η9 indenyl interaction
in the ground state coupled with a weaker Zr-O bond in the
product, arising from a more electron-rich metal center,
decreases the propensity for ligand-induced haptotropic rear-
rangement. Because 5 is competent for diethyl ether cleavage
at rates only slightly slower than 1 and 2, it appears that η6-
haptomers are not a prerequisite for C-O bond scission. The
slower rate observed with 5 is most likely a ground state effect,
where a stronger η9 interaction increases the barrier to access
the intermediate η5,η5-bis(indenyl)zirconium species required
for C-H activation. The observed “leveling” in the rate
constants for diethyl ether cleavage with 1 suggests that at higher
ether concentrations formation of an unproductive η6,η5-bis-
(indenyl)zirconium ether compound competes with formation
of a productive η5,η5-bis(indenyl)zirconium intermediate that
is competent for rate-determining C-H activation followed by
fast â-alkoxide elimination.
The observed selectivity in EtOnBu cleavage is also readily
accommodated by mechanism A in Figure 5. Preferential Cs
H activation of the terminal methyl group of the ethyl chain
leads to rapid, irreversible â-alkoxide elimination to form the
observed products. The alternative pathway to form 1-(OEt)H
and 2-butenes would require either CsH activation of a more
sterically hindered methylene position along the butyl chain or
oxidative addition of the terminal methyl group followed by
chain running. Both of these processes are expected to have
higher barriers than those of the first pathway, and hence
1-(OnBu)H and 1-(η2-CH2dCH2) are formed exclusively.
Addition of 1 equiv of methyl vinyl ether to a benzene-d6
solution of 5-H2 at 23 °C immediately yielded 5-(OMe)H and
free ethylene (Figure 6). These products are a result of olefin
insertion into the Zr-H bond followed by rapid â-OMe
elimination; no resonances for the intermediate zirconocene alkyl
1
hydride were detected by H NMR spectroscopy. A similar
experiment was also performed with 1-hexenyl methyl ether.31
Addition of 1 equiv of the alkoxy-substituted olefin to a
benzene-d6 solution of 5-H2 at 23 °C yielded the zirconocene
alkyl hydride complex 5-(nHxOMe)H immediately upon mix-
ing. No chain running was observed until the solution was heated
to 65 °C, whereupon 5-(OMe)H and a mixture of 2-hexenes
was observed. These observations demonstrate that â-hydrogen
elimination from terminal zirconocene alkyls is possible32 and
continues until irreversible â-alkoxide elimination occurs. An
alternative pathway would involve alkane reductive elimination
to yield 5 followed by methylene C-H activation and â-meth-
oxide elimination.
Based on the experimental data, previous work with bis-
(indenyl)zirconium complexes and computational studies on
C-O activation for related THF compounds (vide infra), Path
A is the favored mechanism for dialkyl ether cleavage. Previ-
ously our laboratory has demonstrated that the η9,η5-bis-
(indenyl)zirconium sandwich, 1, promotes the C-H activation
of N,N-dimethylaminopyridine (DMAP), demonstrating that
such species are capable of C-H activation.11 In addition,
solution NMR12 and computational13,16 studies demonstrate that
dissociation of the η9 indenyl ring to form a transient η5,η5-
bis(indenyl)zirconium sandwich is energetically feasible at 23
°C. Moreover, C-H activation of the methyl groups of several
ethers to yield (η5-C9H5-1,3-(SiMe3)2)2Zr(CH2OR)H complexes
have been observed. These compounds are not on the reaction
coordinate for C-O bond cleavage, as â-alkoxide elimination
is not possible. Importantly, these compounds are models for
the C-H activation products of the longer alkyl chain. NMR
spectroscopy has clearly established η5 indenyl ligand coordina-
tion in the (η5-C9H5-1,3-(SiMe3)2)2Zr(CH2OR)H derivatives,
suggesting that productive C-H activation ultimately leading
to â-alkoxide elimination also proceeds through η5,η5-bis-
(indenyl)zirconium haptomers.
Carbon-Oxygen Bond Cleavage of THF. Although η5,η5-
bis(indenyl)zirconium intermediates are favored for dialkyl ether
cleavage, the possibility of C-O bond activation from isolable
η6,η5-bis(indenyl)zirconium ligand complexes was also explored.
Thermolysis of a benzene-d6 solution of (η6-C9H5-1,3-(SiMe3)2)-
(η5-C9H5-1,3-(SiMe3)2)Zr(THF) (1-THF) at 85 °C for several
days resulted in clean and quantitative formation of the ring-
opened product, (η5-C9H5-1,3-(SiMe3)2)2Zr(cyclo-OCH2(CH2)2-
CH2) (1-c-O(CH2)4) (eq 6). The zirconocene alkoxy alkyl
complex, 1-c-O(CH2)4, was characterized by multinuclear NMR
spectroscopy, combustion analysis, and degradation experiments.
Treatment with excess DCl produced 1-Cl2 and O-4-d2-n-
(30) Bradley, C. A.; Lobkovsky, E.; Keresztes, I.; Chirik, P. J. J. Am. Chem.
Soc. 2006, 128, 6454.
(31) (a) Brown, H. C.; Lynch, G. J. J. Org. Chem. 1981, 46, 531. (b) Newcomb,
M.; Filipkowski, M. A.; Johnson, C. C. Tetrahedron Lett. 1995, 36, 3643.
(32) Chirik, P. J.; Day, M. W.; Labinger, J. A.; Bercaw, J. E. J. Am. Chem.
Soc. 1999, 121, 10308.
1
2
butanol, identified by H and H NMR spectroscopy.
9
16606 J. AM. CHEM. SOC. VOL. 128, NO. 51, 2006

η9
,
η
5-Bis(indenyl)Zr Sandwich Compds for C
−O Bond Cleavage
A R T I C L E S
while the first-order rate constant for THF dissociation is 31
s-1 (eq 8). Comparing the latter value to the rate constants
measured for THF cleavage demonstrates that THF dissociation
to generate the η9,η5-bis(indenyl)zirconium sandwich is fast and
reversible relative to the C-O bond cleavage reaction.
Carbon-oxygen bond cleavage in THF was also explored
Kinetic isotope effects for THF scission were also measured.
Competition experiments were conducted, whereby benzene-
d6 solutions of 1 were heated to 92 °C with excesses of
equimolar mixtures of THF and THF-d8. The conversion was
with zirconocene sandwich complexes that do not form isolable
(or even observable) η6,η5-bis(indenyl)zirconium-THF deriva-
tives. Recall that the purely alkyl substituted 5 does not form
any detectable quantity of 5-THF, even in THF solution.17
Thermolysis of 5 at 85 °C in the presence of an excess of THF
resulted in complete conversion to (η5-C9H5-1,3-(CHMe2)2)2-
Zr(cyclo-OCH2(CH2)2CH2) (5-c-O(CH2)4) after 3 days (eq 7).
1
monitored as a function of time by integrating the H NMR
resonances of the indenyl ring protons relative to the metallo-
cycle. This procedure yielded a value of kH/kD of 3.5(3) at 92
°C as the average of five independent trials. The control
experiment where 1-c-O(CD2)4 was heated to 92 °C for 12 h
was conducted and produced no isotopic exchange into the
[SiMe3] substituents. As with diethyl ether cleavage, the
observation of a normal, primary kinetic isotope effect estab-
lishes a C-H bond-breaking step along the reaction coordinate
either in or before the rate-determining step.
The synthesis of potential intermediates from olefin insertion
with the bis(indenyl)zirconium dihydride 5-H2 was also ex-
plored. Treatment of 5-H2 with excess (>25 equiv) 2,5-
dihydrofuran at -78 °C followed by warming to ambient
temperature yielded the zirconocene alkyl hydride complex 5-(c-
OC4H7)H, where the zirconium has migrated to the carbon R
The ring opened product, 5-c-O(CH2)4, was characterized by
multinuclear NMR spectroscopy (1H, 13C), combustion analysis,
and X-ray diffraction. Representations of the solid-state structure
are presented in Figure 7. The overall molecular geometry of
the compound is typical for a bent zirconocene derivative with
two η5 indenyl rings with a rotational angle of only 11.6(3)°,
placing one benzo substituent over the metallocene wedge,
similar to the geometry observed in a related crotyl hydride
compound.12 The six-membered metallacycle adopts a chairlike
conformation with zirconium-carbon and zirconium-oxygen
bond distances of 2.3007(11) and 1.9180(8) Å, similar to those
in the related permethylated zirconocene complex, (η5-C5Me5)2-
Zr(cyclo-O(CH2)4), prepared by treatment of the zirconocene
dichloride with BrMgCH2(CH2)2CH2OMgBr.33
1
to the oxygen atom (eq 9). The H NMR spectrum of 5-(c-
Mechanistic Studies for THF Cleavage. As with diethyl
ether cleavage, kinetic studies were performed on the ring
opening of THF with 1. Experiments were performed at 85 °C
with excesses of THF ranging between 10 and 75 equiv. The
pseudo-first-order rate constants are presented in Table 3 and
were determined from monitoring the conversion of 1-THF to
1
1-c-O(CH2)4 by H NMR spectroscopy as a function of time.
The data compiled in Table 3 clearly establish that the rate
of C-O bond cleavage is independent of the THF concentration.
We have previously measured both the equilibrium and rate
constants for THF coordination to 1.17 In benzene-d6 at 23 °C,
the equilibrium constant for the addition of THF to 1 is 22 M-1
,
OC4H7)H exhibits the number of resonances consistent with a
near equimolar mixture of endo and exo isomers, where the
endo isomer is defined as the one in which the oxygen atom of
the ring is coordinated in the central position of the zirconocene
wedge. Catalytic isomerization of the bulk 2,5-dihydrofuran to
the 3,4 isomer occurred immediately upon addition to 5-H2.
Table 3. Observed Pseudo-First-Order Rate Constants for C-O
Bond Cleavage in THF with 1 at 85 °Ca
1
equiv of THF
kobs
×
103 (s-
)
10
50
75
1.8(1)
1.9(1)
1.8(1)
(33) Mashima, K.; Yamakawa, M.; Takaya, H. J. Chem. Soc., Dalton Trans.
1991, 2851.
a Each experiment conducted at 0.03 M 1 in benzene-d6.
9
J. AM. CHEM. SOC. VOL. 128, NO. 51, 2006 16607

A R T I C L E S
Bradley et al.
Figure 7. Molecular structure of 5-c-O(CH2)4 at 30% probability ellipsoids (left). Hydrogen atoms omitted for clarity. Top view of the molecule with
hydrogens and isopropyl methyl substituents removed for clarity (right).
Confirmation of the position of the zirconium atom on the ring
was accomplished by treatment with DCl, which yielded 5-Cl2
with deuterium exclusively in the position R to the oxygen.
Over the course of hours at 23 °C in a benzene-d6 solution,
5-(c-OC4H7)H gradually converts to the product of THF scission
5-c-O(CH2)4. Unfortunately attempts to gain more mechanistic
insight by treating 5-D2 with 2,5-dihydrofuran have been
unsuccessful as the isotopic labels are rapidly washed into the
excess olefin required for synthesis. Despite this limitation, these
results demonstrate that an observable C-H activated product
leads to C-O bond cleavage under mild conditions.
catalyst, [(COD)Ir(PPh3)(py)]PF6 (COD ) 1,5-cyclooctadiene,
py ) pyridine). Using this procedure, less than 5% deuterium
incorporation into the positions adjacent to the oxygen was
2
detected by H NMR spectroscopy. Other catalysts such as
(Ph3P)3RhCl produced higher levels of isotopic scrambling.
Addition of 3,4-d2-THF to a benzene-d6 solution of 1 afforded
the η6,η5-bis(indenyl)zirconium-THF-d2 complex, 1-THF-d2.
Two isomers (one with the deuterium atoms syn to the η6 ring,
the other with the deuteria anti) are most likely formed; however
1
they are not resolved by H NMR spectroscopy in part due to
the rapid dissociation and recoordination of the THF-d2 ligand.17
Thermolysis of 1-THF-d2 at 85 °C for 3 days resulted in ring
opening to yield the oxozirconacycle without scrambling of the
deuterium positions as judged by 2H and 13C NMR spectroscopy
(eq 11). The assignment of the 13C NMR spectrum (Figure 8)
Failure to add an excess of either 3,4- or 2,5-dihydrofuran to
5-H2 produced a different outcome. Under these conditions,
bimetallic insertion/ring opening to the bridging alkoxy alkyl
zirconocene hydride complex was observed (eq 10). Treatment
was accomplished by comparison of chemical shifts to known
bis(indenyl)zirconium compounds as well as with a g-HSQC
experiment. The appearance of two triplets in the {1H}13C
experiment for the carbons labeled “2” and “3” is consistent
with selective deuterium incorporation. Degradation of the
labeled ring opened product with excess HCl was consistent
with this observation yielding 1-Cl2 and 2,3-d2-butan-1-ol
exclusively (eq 11).
of this compound with 1 atm of dihydrogen at 45 °C yielded
the bis(tetrahydroindenyl)zirconocene dihydride, 5-THI-H2,30
and the alkoxy hydride, 5-(OnBu)H.
The C-O bond cleavage of 3,4-d2-THF with 1 was also
explored. This substrate was prepared in a straightforward
manner by D2 addition to 2,5-dihydrofuran using Crabtree’s
9
16608 J. AM. CHEM. SOC. VOL. 128, NO. 51, 2006

η9
,
η
5-Bis(indenyl)Zr Sandwich Compds for C
−O Bond Cleavage
A R T I C L E S
Figure 8. Partial {1H}13C NMR spectra of (a) 1-c-O(CH2)4 and (b) 1-c-O(CH2CH(D)CH(D)CH2) in benzene-d6.
The results of the isotopic labeling experiment with 3,4-THF-
d2 establish that chain running via reversible â-hydrogen
elimination to the 3 or 4 position of the THF ring, while
plausible from the model studies, is not a prerequisite for C-O
bond cleavage. If this were the case, deuterium scrambling into
the 1 and 4 positions (Figure 8) would be expected.
Computational Studies. The mechanism of both C-H and
C-O bond activation of THF by bis(indenyl)zirconium sand-
wich complexes was also investigated with DFT calculations34
using model complexes, whereby the 1,3-disubstituted indenyl
ligands were replaced with the parent indenyl anion, [C9H7]-.
The possibility of both bond activation reactions from the η6,η5-
bis(indenyl)zirconium THF complex, (η6-C9H7)(η5-C9H7)Zr-
(THF) (A) was initially computed. The free energy profile35
for these processes is presented in Figure 9.
The mechanism for C-H bond breaking from A (left side,
Figure 9) follows a single-step pathway. The reaction profile
corresponds to the oxidative addition of one C-H bond to the
zirconium, forming the zirconocene alkyl hydride complex, B,
where the η6,η5 hapticity of the indenyl rings is maintained. In
the transition state, TSAB, the scission of the C-H bond is just
started, while the formation of the Zr-H bond is nearly complete
with a Zr-H distance of 1.87 Å, only slightly longer than the
value of 1.82 Å found in the final product B. The Wiberg indices
(WI)36 also support this trend as the Zr-H has a value of 0.481
in TSAB and 0.776 in B; the C-H bond has a value of 0.383 in
TSAB and 0.037 in B. The Zr-O bond is maintained throughout
the course of the reaction as evidenced by the bond distances:
dZrsO ) 2.32, 2.34, and 2.30 Å, in A, TSAB, and B, respectively.
This interaction becomes progressively stronger along the
reaction coordinate (WIZrsO ) 0.239, 0.263, and 0.285, by the
same order). Perhaps most significantly, C-H activation from
A is endergonic with ∆G ) 8.0 kcal/mol and a rather high
activation energy of ∆Gq ) 26.7 kcal/mol.
The C-O bond breaking pathway from A also proceeds by
a single-step process. In this case, the product is the zirconacycle
(C) with η6 and η5 bound indenyl ligands. In this case, a
dissociative mechanism is clearly operative as the new Zr-C
bond has not started to form in the transition state (TSAC) as
dZrsC ) 3.44 Å. Importantly, C-O bond breaking in THF is
far advanced with an elongated bond length of 2.02 Å and
weaker WI (WICsO ) 0.355) than those of the reactant (dCsO
) 1.47 Å, WICsO ) 0.826). Although the reaction is computed
to be significantly exergonic, ∆G ) -34.9 kcal/mol, it involves
a high activation barrier of ∆Gq ) 22.1 kcal/mol.
The values obtained for both C-H and C-O bond cleavage
from A as well as the experimental results prompted an
(36) (a) Wiberg, K. B. Tetrahedron 1968, 24, 1083. (b) Wiberg indices are
electronic parameters related to the electron density in between atoms. They
can be obtained from a Natural Population Analysis and provide an
indication of the bond strength.
(34) Parr, R. G.; Yang, W. Density Functional Theory of Atoms and Molecules;
Oxford University Press: New York, 1989.
(35) All ∆G values were calculated at 298.15 K.
9
J. AM. CHEM. SOC. VOL. 128, NO. 51, 2006 16609

A R T I C L E S
Bradley et al.
Figure 9. Free energy profile for the mechanism of C-H (left) and C-O (right) bond activation, starting from (η6-C9H7)(η5-C9H7)Zr(THF) (A). The
minima and the transition states were optimized (B3LYP/VDZP), and the obtained structures are presented. The free energies (kcal/mol-1) are relative to A,
and the values in italics represent energy barriers. The most relevant atoms in each step are highlighted.
Figure 10. Free energy profile for the conversion of (η6-C9H7)(η5-C9H7)Zr(THF) (A) into (η5-C9H7)2Zr(THF) (G) by a dissociative mechanism. The minima
and the transition states were optimized (B3LYP/VDZP), and the obtained structures are presented. The free energies (kcal mol-1) are relative to A, and the
values in italics represent energy barriers. The most relevant atoms in each step are highlighted.
investigation into alternative mechanistic possibilities. In these
cases, the starting zirconium compound is the η5,η5-haptomer
of A, namely (η5-C9H7)2Zr(THF) (G). Therefore, the reaction
of converting A to G was initially studied (Figure 10). There
are two likely possibilities for this transformation. One is
intramolecular with a direct shift of the hapticity of the η6
indenyl to η5. The other possibility is dissociative, involving
initial loss of THF to generate the η9,η5-bis(indenyl)zirconium
sandwich, (η9-C9H7)(η5-C9H7)Zr (E), whereby readdition of
THF yields (η5-C9H7)2Zr(THF) (G).
The dissociative conversion of A to G occurs in four discrete
steps. The first two convert the η6,η5 complex, A, into the η9,η5
bis(indenyl)zirconium sandwich, (η9-C9H7)(η5-C9H7)Zr (E). The
first step is THF loss from A to produce (η9-C9H7)(η5-C9H7)Zr
(D), where the indenyl rings adopt an anti conformation.
Rotation of the indenyl ligands yields the gauche conformer,
E, which converts to gauche (η5-C9H7)2Zr (F) by benzo
dissociation. Finally, F coordinates THF through a transition
state (TSFG) with the two molecules loosely bound (dZrsO
)
4.07 Å). The overall activation free energy for this entire four-
step process is 12.0 kcal/mol, well below the value computed
(22.8 kcal/mol) for the intramolecular η9 to η5 shift13 and much
lower than those computed for C-H and C-O activation from
A (Figure 9).
The DFT computed free energy profile for THF cleavage from
the η5,η5-bis(indenyl)zirconocene THF adduct, G, is presented
in Figure 11. The left side of the figure corresponds to direct
C-O bond cleavage in the reactant, G, yielding the metallacycle
(η5-C9H7)2Zr(O-c-(CH2)4) (H) in a single step. Similar to the
mechanism from A, C-O bond cleavage is well advanced in
9
16610 J. AM. CHEM. SOC. VOL. 128, NO. 51, 2006

η9
,
η
5-Bis(indenyl)Zr Sandwich Compds for C
−O Bond Cleavage
A R T I C L E S
Figure 11. Free energy profile for the mechanism of C-H bond activation (right, two steps) and direct C-O bond breaking (left, one step), starting from
(η5-C9H7)2Zr(THF) (G). The minima and the transition states were optimized (B3LYP/VDZP), and the obtained structures are presented. The free energies
(kcal mol-1) are relative to (η6-C9H7)(η5-C9H7)Zr (A), and the values in italics represent energy barriers. The most relevant atoms in each step are highlighted.
the transition state (dCsO ) 1.91 Å, WICsO ) 0.422). This
process (G to H) is thermodynamically favorable with ∆G )
-51.0 kcal/mol and has an accessible activation free energy of
∆Gq ) 14.0 kcal/mol.37
kinetic isotope effect of 3.5(3) at 92 °C for THF versus THF-
d8 cleavage. Moreover, the initial dissociation of THF from the
η6,η5-bis(indenyl)zirconium THF compound to form the sand-
wich is followed by recoordination to yield the η5,η5-haptomer
prior to the rate-determining step and produces the overall zero-
order dependence on THF concentration. C-H bond cleavage
from the position adjacent to the oxygen without rearrangement
to the â-position is also in agreement with the cleavage of 3,4-
d2-THF.
It should be noted that previous computational studies13 have
identified the spin triplet form of (η5-C9H7)2Zr(THF) as being
more stable by 2.8 kcal/mol than the singlet complex, G.
Numerous attempts were made to find a plausible mechanism
on the triplet spin surface, but all cases studied involved
conversion to the singlet species, G, through the known13
crossing point, 3.2 kcal/mol above the triplet state itself. It should
be pointed out that the computational studies were conducted
on model compounds that lacked the sterically demanding 1,3-
substituents, and these are known to attenuate the singlet-triplet
gaps.16
The right side of Figure 11 corresponds to a two-step pathway
involving initial C-H activation to yield an intermediate η5,η5-
bis(indenyl)zirconium alkyl hydride complex (I) that undergoes
subsequent C-O bond cleavage to yield the final metallacycle
product, J. The difference between H and J is simply the
rotational conformation of the two indenyl ligands, and these
rotamers are expected to interconvert readily in solution.12,13
The first step in the path leading from G to J is analogous to
the corresponding C-H activation step from the η6,η5-haptomer,
A (Figure 9, left). In the transition state (TSGI), the Zr-H bond
is practically formed (dZrsH ) 1.89 Å, WIZrsH ) 0.418), and
the C-H bond is still considerable (dCsH ) 1.41 Å, WICsH
0.466). The conversion of I to J is concerted, and thus the Cs
O bond is weakened in the transition state (TSIJ) with dCsO
)
)
2.19 Å and WICsO ) 0.280 as compared to I (dCsO ) 1.49 Å,
WICsO ) 0.805). The incipient C-H bond is nearly formed in
TSIJ (dCsH ) 1.93 Å, WICsO ) 0.231), and both the carbon
and oxygen atoms remain coordinated throughout the conversion
from I to J (dZrsO ) 1.95-2.33 Å, dZrsC ) 2.12-2.31 Å. The
transformation of G to J is thermodynamically favored with
∆G ) -51.7 kcal/mol, and both steps have low activation free
energies (∆Gq ) 13.3 and 11.3 kcal/mol, respectively). Pathways
involving C-H activation R to the oxygen were most favorable
while those relying on activation of the â C-H bond were found
to be much higher in free energy.
The computational studies demonstrate that the most favorable
pathway for C-O bond cleavage in THF from A follows a
multistep mechanism involving conversion of the η6,η5-bis-
(indenyl)zirconium THF compound into the η5,η5-haptomer
followed by C-H activation and finally C-O bond breaking.
The rate-determining step for the entire transformation is the
C-H activation step (G f I, ∆Gq ) 13.3 kcal/mol) and is
consistent with the experimentally observed normal, primary
Concluding Remarks
This study has demonstrated that η9,η5-bis(indenyl)zirconium
sandwich complexes are effective for C-O bond cleavage in
dialkyl and cyclic ethers under mild conditions in solution to
yield well-defined organic and organometallic products. Prepa-
ration of model compounds, in combination with kinetic,
isotopic labeling, and computational studies, has established that
the C-O activation reaction involves initial conversion of the
η9,η5-bis(indenyl)zirconium sandwich to reducing η5,η5 inter-
mediates that promote C-H activation. In the dialkyl ether
examples, subsequent â-alkoxide elimination generates the
observed products. For isolable η6,η5-bis(indenyl)zirconium THF
compounds, dissociation of the oxygen donor followed by
haptotropic rearrangement yields the active species for rate-
determining C-H activation followed by C-O bond cleavage.
In these reactions, isotopic labeling studies have eliminated
competitive chain running during the course of THF opening.
(37) All energies are relative to the initial reactant, A.
9
J. AM. CHEM. SOC. VOL. 128, NO. 51, 2006 16611

A R T I C L E S
Bradley et al.
These studies highlight the potential of isolable η9,η5-bis-
(indenyl)zirconium sandwich compounds to mediate formal two-
electron reduction chemistry by access to η5,η5-intermediates
generated by facile benzo ring dissociation.
and 100 µL of a 0.031 M benzene-d6 solution containing sublimed
ferrocene. The appropriate amount (typically 5-25 µL) of dialkyl ether
was then added by a microliter syringe, and benzene-d6 was added to
bring the total volume to 0.5 mL. The tube was then quickly shaken
and frozen in liquid nitrogen. The tube was then transferred to a
preheated NMR probe at the appropriate temperature (45 or 85 °C),
and the disappearance of starting material and the appearance of product
were measured in an arrayed 1H NMR experiment. Typically the time
interval between data collection was 1-5 min depending upon the
reaction rate. Spectra were integrated versus the ferrocene standard.
Three runs at each dialkyl ether concentration were performed.
Typical Procedure for Kinetic Isotope Effect Determination. A
flame-dried J. Young NMR tube was charged with 300 µL of a 0.04
M benzene-d6 stock solution of 1 and 100 µL of a 0.031 M benzene-d6
solution containing sublimed ferrocene. 25 µL (>25 equiv) of an
equimolar solution of THF/THF-d8 or diethyl ether/diethyl ether-d10
was added to the tube. Dialkyl ether cleavage was then performed under
typical conditions. The solvent was removed in vacuo on a vacuum
line, and the samples were redissolved in 0.5 mL of benzene-d6.
Integration of the 1H NMR resonances of the resulting cleavage products
relative to ferrocene provided the kinetic isotope effects for the
reactions.
Experimental Section38
Diethyl Ether Cleavage with 1. A 25 mL round bottomed flask
was charged with 250 mg (0.41 mmol) of 1 and approximately 10 mL
of diethyl ether. The reaction mixture was stirred at ambient temperature
for 3 days. The solvent was removed in vacuo, and the resulting red
foam was extracted into pentane. Analysis of the products by NMR
spectroscopy revealed formation of a near equimolar mixture of 1-(η2-
CH2dCH2) and 1-(OEt)H. Analytically pure 1-(η2-CH2dCH2) can
be isolated by successive recrystallizations (typically two) from pentane
at -35 °C. However, attempts to obtain 1-(OEt)H free from 1-(η2-
CH2dCH2) have been unsuccessful.
Characterization of (η5-C9H5-1,3-(SiMe3)2)2Zr(η2-C2H4) 1-(η2-
CH2dCH2). Anal. Calcd for C32H50Si4Zr: C, 60.21; H, 7.90. Found:
C, 59.89; H, 7.89.1H NMR (benzene-d6): δ ) 0.12 (s, 36H, SiMe3),
0.87 (br s, 4H, C2H4), 6.67 (s, 2H, Cp), 7.07 (m, 4H, Benzo), 7.65 (m,
4H, Benzo).13 C NMR (benzene-d6): δ ) 1.34 (SiMe3), 94.64 (C2H4),
110.03, 124.97, 125.26, 127.01, 136.27 (Cp/Benzo).
Characterization of (η5-C9H5-1,3-(SiMe3)2)2Zr(OCH2CH3)H (1-
Acknowledgment. We thank the National Science Foundation
(CAREER Award to P.J.C. and predoctoral fellowship to
C.A.B.) for financial support. P.J.C. also acknowledges the
Research Corporation for a Cottrell Scholarship and the Packard
Foundation for a fellowship in science and engineering.
1
(OEt)H). H NMR (benzene-d6): δ ) 0.46 (s, 18H, SiMe3), 0.49 (s,
18H, SiMe3), 0.78 (t, 6 Hz, 3H, OCH2CH3,), 3.37 (q, 6 Hz, 2H, OCH2-
CH3), 5.16 (s, 2H, Cp), 5.23 (s, 1H, Zr-H), 6.74 (t, 8 Hz, 2H, Benzo),
6.81 (m, 2H, Benzo), 7.32 (d, 11 Hz, 2H, Benzo), 7.53 (d, 11 Hz, 2H,
Benzo). 13C NMR (benzene-d6): δ ) 1.76, 1.86 (SiMe3), 23.06
(OCH2CH3), 67.29 (OCH2CH3), 108.72, 123.46, 124.11, 124.93, 125.09,
125.25, 128.70, 133.83 (Cp/Benzo). One Cp/Benzo resonance not
located.
Supporting Information Available: Additional experimental
procedures, computational details, and spectral characterization.
Crystallographic data for 1-(η2-CH2dCH2) and 5-c-O(CH2)4
and all atomic coordinates for optimized species. This material
is available free of charge via the Internet at http://pubs.acs.org.
General Kinetics Procedure. A flame-dried J. Young NMR tube
was charged with 300 µL of a 0.05 M benzene-d6 stock solution of 1
(38) General considerations and additional experimental procedures are reported
in the Supporting Information.
JA065456G
9
16612 J. AM. CHEM. SOC. VOL. 128, NO. 51, 2006