C - D0 (D0 = π-donor, F) cleavage in H2C = CH(D0) by (Cp2ZrHCl)n: Mechanism, agostic fluorines, and a carbene of Zr(IV)

Article abstract of DOI:10.1021/ja0024340

Consistent with the C - O cleavage behavior of vinyl ethers, vinyl fluoride reacts with Cp₂ZrHCl to give Cp₂ZrFCl and C₂H₄ as primary products. DFT (B3PW91) calculations show this reaction to be highly exoenergetic ( - 55 kcal/mol), and reveal a σ-bond metathesis mechanism to be unfavorable compared to a Zr - H addition across the C = C bond, with regiochemistry placing F on C_β of the resulting fluoroethyl ligand. β-F elimination (onto Zr) then completes the reaction. There is no η²-olefin intermediate on the reaction path. DFT calculations seeking the energy and structure of the two carbenes Cp₂ZrHCl[CF(CH₃)] and Cp₂ZrFCl-[CH(CH₃)] are also described.

Full text of DOI:10.1021/ja0024340

J. Am. Chem. Soc. 2001, 123, 603-611
603
C-D0 (D0 ) π-donor, F) Cleavage in H2CdCH(D0) by (Cp2ZrHCl)n:
Mechanism, Agostic Fluorines, and a Carbene of Zr(IV)
Lori A. Watson, Dmitry V. Yandulov, and Kenneth G. Caulton*
Contribution from the Department of Chemistry, Indiana UniVersity, Bloomington, Indiana 47405-7102
ReceiVed July 5, 2000
Abstract: Consistent with the C-O cleavage behavior of vinyl ethers, vinyl fluoride reacts with Cp2ZrHCl to
give Cp2ZrFCl and C2H4 as primary products. DFT (B3PW91) calculations show this reaction to be highly
exoenergetic (-55 kcal/mol), and reveal a σ-bond metathesis mechanism to be unfavorable compared to a
Zr-H addition across the CdC bond, with regiochemistry placing F on Câ of the resulting fluoroethyl ligand.
â-F elimination (onto Zr) then completes the reaction. There is no η2-olefin intermediate on the reaction path.
DFT calculations seeking the energy and structure of the two carbenes Cp2ZrHCl[CF(CH3)] and Cp2ZrFCl-
[CH(CH3)] are also described.
Introduction
The molecular fragment RuHClL2 (L ) PiPr3), derived from
the dimer [RuH(µ-Cl)L2]2, has proven1 to be a π-rich species
which readily and selectively converts olefins containing
π-donor substituents (D0) into carbene ligands (eq 1). The
known molecules where a neutral carbene is bound to a d0
metal.3 This makes a species such as Cp2ZrHCl[CF(CH3)] or
Cp2ZrFCl[CH(CH3)] attractive from the structural point of view.
What is the bond order of the Zr/carbene bond and how will
the carbene p orbital acquire electron density? To put this into
perspective, a few d0 carbonyl complexes are known,4 for
example, Cp2Ti(CO)22+ and Cp2Zr(acetyl)(CO)+, and the bond-
ing in Cp*2Zr(H)2(CO) has been discussed4i,j with regard to its
surprisingly low νCO value.
Zirconium is termed “oxo-philic”, and is, more generally,
considered to form strong bonds with any light atom which bears
multiple lone pairs: amides, halides, and pseudohalides. It is
for this reason that olefins bearing such π-donor functionality
cannot be successfully polymerized by zirconium olefin po-
lymerization catalysts: the functionality is abstracted from
carbon, thereby poisoning the catalyst.
The early stoichiometric applications of Schwartz’ reagent,
Cp2ZrHCl, are consistent with the above principle: they
generally involved nonfunctionalized olefins and alkynes.5
Subsequent studies6 of Cp2ZrHCl reacting with vinyl ethers
H2CdCH(OR) confirmed the ability of this Lewis acidic
zirconium to “abstract” an OR group by C/O bond cleavage
(eq 2)
exothermic character of this transformation (an unusually easy
route to a carbene, and one which is 40-70 kcal/mol endo-
thermic for forming the free carbene) relies on π-electron
donation to the carbene carbon by both D0 and by the metal.
Taube showed2 long ago that the d6 configuration for Ru and
Os is very reducing, which is equivalent to π-donating (back-
donating).
Exploring the analogous chemistry for a d0 metal offers the
possibility of very different thermodynamic preferences. For
example, nonheteroatom stabilized carbenes (the so-called
Schrock carbenes) must rely wholly on metal back-donation to
occupy the carbene p orbital. This is why such carbenes
generally involve highly electropositive metals in low formal
oxidation states (considering the carbene as uncharged). Note
also that the neutral carbene formalism with much back-bonding
(A) is not different from the dianionic carbene formalism (B).
Cp2ZrHCl + H2C ) CH(OEt) f
Cp2Zr(OEt)Cl + H2C ) CH2 (2)
The present work is thus directed in part at exploring both
experimentally and computationally the accessibility to a carbene
ligand in an unprecedented environment: one with no d
electrons to contribute to the carbene (C). In fact, there are no
The released olefin was then rapidly hydrozirconated by
additional Cp2ZrHCl to give Cp2Zr(C2H5)Cl.
While examples of early transition metal complexes which
cleave the strong C-F bond7 are not as plentiful as those of
(1) (a) Coalter, J. N., III; Spivak, G. J.; Ge´rard, H.; Clot, E.; Davidson,
E. R.; Eisenstein, O.; Caulton, K. G. J. Am. Chem. Soc. 1998, 120, 9388.
(b) Coalter, J. N., III; Bollinger, J. C.; Huffman, J. C.; Werner-Zwanziger,
U.; Caulton, K. G.; Davidson, E. R.; Ge´rard, H.; Clot, E.; Eisenstein, O.
New J. Chem. 2000, 24, 9.
(3) A reviewer has commented that Arduengo carbenes are known to
coordinate to d0 metal centers. Since such carbenes involve much N f C
donation from the two heteroatoms on the carbene carbon, this behavior
towards d0 metals is predictable. Indeed, Arduengo carbenes are now proven
to be primarily σ-donors, not π-acceptors towards metals: Herrmann, W.
A., Kocher, C., Angew. Chem., Int. Ed. 1998, 37, 2490.
(2) Taube, H. Pure Appl. Chem. 1979, 51, 901.
10.1021/ja0024340 CCC: $20.00 © 2001 American Chemical Society
Published on Web 01/06/2001

604 J. Am. Chem. Soc., Vol. 123, No. 4, 2001
Watson et al.
Scheme 1
latter metals,8 there has been a recent report of [Cp2ZrH2]2
reacting with C6F6 to give Cp2Zr(C6F5)F as the major product.9
In contrast, reaction of Schwartz’s reagent with vinyl phos-
phines forms a nonagostic insertion product, which undergoes
no further reaction.10 Thus, the qualitative trend of zirconium
interacting much more strongly with first row “hard” bases and
rejecting other “softer” heteroatoms is again demonstrated.
Our interest here is in the mechanism of such C/heteroatom
cleavage by Zr(IV). The absence of d electrons makes this an
electrophilic, not an oxidative process, and indeed, it is for the
Cp2ZrXY unit that the σ-bond metathesis mechanism was
devised out of a necessity to avoid the more traditional oxidative
mechanisms.11 Our special focus here will be on C-F bond
scission, since this is a thermodynamically quite strong bond,
and one subject to intense recent study.12,8a The specific
organofluorine target is vinyl fluoride.
regioisomer (eq 4) showed it to be stable (i.e., no C/O bond
scission) under the conditions of eq 3.
We chose â-methoxy styrene as a vinyl ether in which steric
effects at the R-carbon might be anticipated to direct the
regiochemistry away from that of ethyl vinyl ether. This attempt
(eq 5) apparently fails since the products of this reaction
Results
Vinyl Ethers. The reaction of Cp2ZrHCl with ethyl vinyl
ether has been reported6 to yield Cp2ZrCl(OEt) and ethylene as
primary products. This was shown to proceed by the regio-
chemistry of eq 3 because independent synthesis of the opposite
(complete in C6D6 within 1.5 h at 25 °C) are Cp2ZrCl(OMe)
and the result of secondary reaction of liberated styrene with
Cp2ZrHCl: Cp2ZrCl(CH2CH2Ph) and Cp2(Cl)ZrCHPhCH3 (mi-
nor product). The former has equivalent Cp rings and an
AA′XX′ spin system pattern for the -CH2CH2- group. The
1H/1H coupling within this spin system was confirmed by a
COSY study, and the presence of two CH2 groups (contrast D
and E, which have only one CH2 group) was established by
(4) (a) Meyer, T. J. J. Am. Chem. Soc. 1976, 98, 6733. (b) Manriquez,
J. M.; McAlister, D. R.; Sanner, R. D.; Bercaw, J. E. J. Am. Chem. Soc.
1978, 100, 2716. (c) Antonelli, D. M.; Tjaden, E. B.; Stryker, J. M.
Organometallics 1994, 13, 763. (d) Guo, Z.; Swenson, D. C.; Guram, A.
S.; Jordan, R. F. Organometallics 1994, 13, 766. (e) Howard, W. A.; Parkin,
G.; Rheingold, A. Polyhedron 1995, 14, 25. (f) Procopio, L. J.; Carroll, P.
J.; Berry, D. H. Polyhedron 1995, 14, 45. (g) Calderazzo, F.; Pampaloni,
G.; Tripepi, G. Organometallics 1997, 16, 4943. (h) Pampaloni, G.; Tripepi,
G. J. Organomet. Chem. 2000, 593, 20. (i) Brintzinger, H. H. J. Organomet.
Chem. 1979, 171, 337. (j) Marsella, J. A.; Curtis, C. J.; Bercaw, J. E.;
Caulton, K. G. J. Am. Chem. Soc. 1980, 102, 7244.
(5) (a) Hart, D. W.; Schwartz, J. J. Am. Chem. Soc. 1974, 96, 8115. (b)
Labinger, J. A.; Hart, D. W.; Seibert, W. E., III; Schwartz, J. J. Am. Chem.
Soc. 1975, 97, 3851. (c) Hart, D. W.; Blackburn, T. F.; Schwartz, J. J. Am.
Chem. Soc. 1975, 97, 679. (d) Schwartz, J.; Labinger, J. A. Angew. Chem.,
Int. Ed. Engl. 1976, 15, 333. (e) Wipf, P.; John, H. Tetrahedron 1996, 52,
1253.
(6) Buchwald, S. L.; Nielson, R. B.; Dewan, J. C. Organometallics 1988,
7, 2324.
(7) (a) Burns, C. J.; Anderson, R. A. J. Chem. Soc., Chem. Commun.
1989, 136. (b) Watson, P. L.; Tulip, T. H.; Williams, I. Organometallics
1990, 9, 1999. (c) Weydert, M.; Anderson, R. A.; Bergman, R. G. J. Chem.
Soc. 1993, 115, 8837.
(8) (a) Kiplinger, J. L.; Richmond, T. G.; Osterburg, C. E. Chem ReV.
1994, 94, 373. (b) Murphy, E. F.; Murugavel, R.; Roesky, H. W. Chem.
ReV. 1997, 97, 3425.
(9) Edelbach, B. L.; Rahman, A. K. F.; Lachicotte, R. J.; Jones, W. D.
Organometallics 1999, 18, 3170.
HMQC and DEPT experiments. This same product was also
synthesized independently from styrene and Cp2ZrHCl. If the
reaction is executed in the stoichiometry 1Cp2ZrHCl:2(Ph)-
HCdCH(OMe), free styrene is in fact observed.
Similar results are obtained (Scheme 1) with cyclic vinyl
ethers (2,3-dihydrofuran or 3,4-dihydro-2H-pyran, for ex-
ample).13 Once again, the olefin inserts into the Zr-H bond
through an initial Zr-O agostic interaction, and then was
suggested to undergo C-O bond cleavage to form the alkoxides.
This affords the acyclic species Cp2Zr(Cl)O(CH2)xCHdCH2
which can further react with a second equivalent of Cp2ZrHCl
to form a dizirconated species.
(10) Zablocka, M.; Igau, A.; Majoral, J.-P.; Pietrusiewicz, K. M.
Organometallics 1993, 12, 603
(11) (a) Koga, N.; Morokuma, K. Chem. ReV. 1991, 91, 823. (b) Tolbert,
M. A.; Beauchamp, J. L. J. Am. Chem. Soc. 1984, 106, 8117. (c) Thompson,
M. E.; Baxter, S. M.; Bulls, A. R.; Burger, B. J.; Nolan, M. C.; Santarsiero,
B. D.; Schaefer, W. P.; Bercaw, J. E. J. Am. Chem. Soc. 1987, 109, 203.
(d) Christ, C. S.; Eyler, J. R.; Richardson, D. E. J. Am. Chem. Soc. 1990,
112, 596.
C/O bond cleavage is the shared feature of these reactions,
but it must not be interpreted as indicating the complete absence
(12) Burdeniuc, J.; Jedlicka, B.; Crabtree, R. H. Chem. Ber. Recl. 1997,
97, 3425.
(13) Ce´nac, N.; Zablocka, M.; Igau, A.; Commenyes, G.; Majoral, J. P.;
Skowronska, A. Organometallics 1996, 15, 1208.

C-D0 CleaVage in H2CdCH(D0) by (Cp2ZrHCl)n
J. Am. Chem. Soc., Vol. 123, No. 4, 2001 605
Scheme 2
Scheme 3
of Zr-H addition with the opposite regiochemistry (Scheme
2) since that might be reversible. For the later transition metal
RhIII, and where O f metal donation is not possible, it is
thermodynamically favored to have OR′ on CR rather than
(Câ).14 Such thermodynamic preference could certainly be
reversed when an O f metal bond is involved, and where ring
sizes differ. In the cases at hand, formation of the ultimately
preferred Zr-alkoxide linkage can only evolve from F, since
G has put the hydrogen on the “wrong” carbon for an olefinic
product. Indeed, from G, what might evolve is the carbene
species H.
Figure 1. Optimized geometries of reactants and products with selected
distances in Å.
the product mixture; however, a broad resonance at ap-
proximately the same chemical shift is observed. This resonance
might be broadened because of halide exchange with other
products in the reaction mixture.
No intermediates were observed in the course of this reaction.
In particular, no fluoroethyl complex or carbene species was
observed. The approximately 1.5-h reaction time is mass-transfer
limited, due to the extremely low solubility of Cp2ZrHCl. Thus,
kinetic studies will not depend on the molecular mechanism of
this reaction. We therefore undertook DFT calculations to
delineate the reaction mechanism (see below).
To generalize this reactivity of fluorine-substituted olefins,
CH2CF2 was reacted with Cp2ZrHCl; this reaction produces
CH2CHF, Cp2Zr(CH2CH3)Cl, and CH2CH2 (1H NMR evidence).
Again, the products formed indicate that the primary reaction
gives Cp2ZrClF and, in this case, vinyl fluoride, which can then
react with another equivalent of Cp2ZrHCl.
Calculated Reaction Energy. DFT calculations were carried
out on the overall reaction (Figure 1). Species 1 and 2 are very
similar in their arrangement of Cp ligands, differing significantly
only in the Zr-Cl bond distance. The Zr-Cl bond lengthens
by 0.016 Å from the reactant, Cp2ZrHCl, to the product, Cp2-
ZrClF. This is probably due to the increased donor ability of F
versus H, allowing slight lengthening of the Zr-Cl bond. The
Cp (centroid)-Zr-Cp-(centroid) angle changes very little
thoughout the reaction; the average of all calculated structures
reported in this study is 131.6°, with a standard deviation of
1.65°.
The overall reaction energy (with zero-point energy correction
included) was calculated to be -55.1 kcal/mol. This large
exothermic value is to be expected considering the strength of
the Zr-F bond formed.15,16
Mechanistic Questions. Because no intermediates were
observed for this reaction, DFT methods were used to evaluate
a number of mechanistic questions. DFT calculations on the
analogous reaction with C2H4 have been published.17 One can
Vinyl Fluoride. A slurry of Cp2ZrHCl in benzene reacts to
completion within 1.5 h with vinyl fluoride (100% excess). Cp2-
Zr(C2H5)Cl, which is the product of the secondary reaction of
Cp2ZrHCl with liberated ethylene is evident by Cp and ethyl
1H NMR signals. The continued production of this compound
diminishes with time, and free ethylene accumulates, due to
the depletion of Cp2ZrHCl. Three additional Cp resonances and
two 19F resonances grow during the reaction. Two of these show
1
somewhat broader Cp H NMR resonances, consistent with
unresolved coupling to 19F. These three Cp and two 19F
resonances are attributed to Cp2ZrClF and the F/Cl redistribution
products Cp2ZrCl2, Cp2ZrF2 and Cp2Zr(C2H5)F. Overall, these
results are explained by the reaction in Scheme 3.
Attempts to independently synthesize Cp2ZrF2 have been
successful. As Cp2ZrF2 is soluble in THF, it was hoped that
comparison of its chemical shift with the chemical shifts of the
products of the reaction of vinyl fluoride and Cp2ZrHCl in THF-
d8 could provide a basis for assigning at least the Cp resonance
of the supposed difluoride halide exchange product. In this
reaction in THF-d8, the analogous products were obtained, with
Cp2ZrCl(CH2CH3) and free ethylene both observed. One of the
Cp resonances at 6.37 matches well in shift and line width with
the Cp signal of Cp2ZrF2 at 6.38 ppm. The 19F shift of the
difluoride at -157.3 is not seen as a distinct peak in the 19F of
(15) Dias, A. R.; Martinho-Simoes, J. A. Polyhedron 1988, 7, 1531.
(16) Wipf, P.; Jahn, H. Tetrahedron 1996, 52, 12853.
(17) Endo, J.; Koga, N.; Morokuma, K. Organometallics 1993, 12, 2777.
(14) Mak, K. W.; Xue, F.; Mak, T. C. W.; Chan, K. S. J. Chem. Soc.,
Dalton Trans. 1999, 3333.

606 J. Am. Chem. Soc., Vol. 123, No. 4, 2001
Watson et al.
Figure 2. Possible mechanisms and approach directions.
Figure 4. “Outside” reaction pathway, with selected bond distances
(Å), angles (deg), and relative energies (with ZPE) in kcal/mol.
Scheme 4
Figure 3. σ-Bond metathesis transition states: “inside” (TSA) and
“outside” (TSB) with selected distances in Å.
The C-F bond is also stretched by 0.150 Å from its vinyl
fluoride length as it begins to interact with zirconium. There is
very little disruption of the C-C bond, however. It remains
essentially a double bond, stretched by only 0.016 Å from the
CH2CHF distance of 1.348 Å.
The energy of this transition state lies 21.1 kcal/mol above
the separated species. Using a typical18 ∆Sq of -27 cal deg-1
mol-1, this corresponds to a half-life (when [C2H3F] ) 1 M)
of 8 × 104 h at 25 °C. In reality, other pathways have a lower
energy barrier.
imagine two classes of mechanisms that would lead to the
observed experimental products (Figure 2). The olefin could
first insert into the Zr-H bond and then undergo â-F migration
and ethylene elimination. This mechanism corresponds to the
olefin entering with the πcc face first. Alternately, a four-centered
transition state corresponding to a σ-bond metathesis mechanism
could be formed. This would be the result of the F-C bond
encountering the zirconium center first. Both of these mecha-
nisms have been implicated in the reactions of early transition
metals, and both are allowed by the unsaturated metal center in
this case.
In each of these mechanisms, the olefin can approach from
two directions, between the hydride and chloride (“inside”) or
on the side opposite the chloride (“outside”); the third direction,
approach opposite hydride, will not lead to the needed hydride-
olefin interaction. There are also two possible regiochemistries
of addition; the first of these forms an initial â-fluoro alkyl
product, while insertion from the opposite direction will form
an ethyl product with the fluorine on the R-carbon. Although
only the â-fluoro interaction can lead to the observed products,
the energetic difference between the two addition regiochem-
istries leads to interesting conclusions and therefore will be
discussed later.
σ-Bond Metathesis. As zirconium(IV) has vacant d orbitals
available, a four-centered transition state is possible. Indeed,
σ-bond metathesis was first proposed to account for these same
types of reactions with d0 metalssparticularly zirconium.11
Additionally, such a concerted transition state has been proposed
as the mechanism in the C-F bond cleavage of a variety of
fluoroarenes by a similar system, [Cp2ZrH2]2.9
A transition state was found for the “inside” approach of CH2-
CHF, forming a four-centered transition state TSA (Figure 3).
This structure can best be described as one resulting from
electrophile-assisted nucleophilic attack. This is supported by
the obvious pyramidalization of the R-carbon (sum of angles
around the carbon is 343.5°) by the incoming hydride ligand
and the weakening of the C-F bond as it interacts with Zr. In
this transition state, the developing C-H bond is not very far
advanced, as is evident by a C-H bond distance of 1.319 Å
(compared to the C-H distance of 1.087 Å in free ethylene).
A similar geometry, but with a higher activation energy, is
obtained as a four-centered transition state from the “outside”
approach (TSB). The energy of this transition state lies 28.2
kcal/mol above the separated species and lies earlier still on
the reaction coordinate than the “inside” case. Bonds being
formed are further away from their product value. For example,
the Zr-F bond is longer (at 2.423 Å) than in the “inside”
transition state (2.346 Å) as is the developing C-H bond (1.418
vs 1.319 Å) (Figure 3).
Regardless of olefin approach trajectory, the experimental
evidence in this case argues against the σ-bond metathesis
mechanism. Namely, we did not observe C-F activation when
Cp2ZrHCl was reacted with a variety of fluoroarenes in a variety
of concentrations (Scheme 4, 1.5 equiv, or neat arene). If σ-bond
metathesis had been the operative mechanism, one would have
expected the formation of Cp2ZrClF and the corresponding
arene. However, one would expect no reaction with the arenes
if insertion was the predominant mechanism because insertion
is more difficult with aromatic systems than for an isolated vinyl
group due to loss of aromaticity. Thus, the fact that no reaction
with arene was seen even with extended reaction time (3 days
at room temperature) argues against a σ-bond metathesis
mechanism and for one that allows an initial insertion step.
Insertion Mechanisms. (a) “Outside” Attack. Attack of the
olefin face from the side opposite chloride is calculated to
encounter a +20.5 kcal/mol barrier to insertion (Figure 4). The
Zr-C distances (2.859 and 2.950 Å) at the TS13 indicate that
relatively little Zr/olefin bond formation has taken place. For
(18) Minas da Piedade, M. E.; Simo˜es, J. A. M. J. Organomet. Chem.
1996, 518, 167.

C-D0 CleaVage in H2CdCH(D0) by (Cp2ZrHCl)n
J. Am. Chem. Soc., Vol. 123, No. 4, 2001 607
Figure 5. “Inside” reaction pathway, with selected bond distances (Å),
angles (deg), and relative energies (with ZPE) in kcal/mol.
Figure 6. Space filling drawings of TS13 and TS14. Cl and F are
darkest.
comparison, Zr(IV)/C(olefin) distances of 2.68 and 2.89 Å have
been reported.19 There is no η2 intermediate found here; the
only η2-structure is this transition state. This is not surprising,
as Zr in this instance has no d-electrons to stabilize an olefin
complex by back-bonding. The high barrier is in contrast to
the low barrier one would expect for a normal Lewis acid/base
adduct (about 3 kcal/mol). While the higher calculated barrier
is partly due to the lack of such an η2-intermediate, the hydride
and chloride must move much closer together (from 100.2° in
Cp2ZrHCl to 74.0° in the transition state), providing some
destabilization from H/Cl repulsive forces.
From the initial insertion transition state, a â-H agostic
structure should be formed. While no minimum (i.e., intermedi-
ate) with this structure has been found, a product resulting from
rotation about the CH2F group, the â-F agostic structure (3), is
a minimum. This lies16.4 kcal/mol below the reactants, and
shows lengthening of the C-F bond and a close F-Zr contact
characteristic of an agostic structure.20 Additionally, the agostic
character is evident by a reduction in the Zr-CR-Câ angle to
98.0°. A C-C distance of 1.484 Å, much nearer a C-C single
bond length, implies that this structure is best described as a
â-F agostic ethyl complex. Supporting this is the Zr-Câ bond
distance, which at 2.995 Å is not within normal bonding distance
(2.2-2.4 Å).
(Figure 6). The olefin, instead of being sandwiched between
the two Cp rings as in the “outside” case, occupies a less
sterically crowded face of the zirconium where there is more
room between the Cp’s. Also, the Cl-Zr-H angle opens up to
140°, which eliminates any repulsion between the hydride and
chloride ligands that may have been present in the “outside”
case. Once again, there is no intermediate η2 olefin complex
(optimization of TS14 distorted toward the reactants resulted in
complete dissociation of H2CdCHF); an η2 adduct is a transition
state for insertion.
The next step on the calculated reaction pathway is a
transition state corresponding to CH2F rotation to form a
â-fluoroethyl intermediate (TS4). This transition state has an
energy of -18.5 kcal/mol below reactants. The â-hydrogen is
nearly coplanar with Zr, CR, and Câ; the dihedral angle is 353.9°.
The C-C bond distance indicates that this structure is very
similar to an η1 alkyl; the in-plane hydrogen appears not to
interact strongly with the Zr center (Câ-H distance of 1.099
vs 1.098 for the noncoplanar hydrogen).
In contrast to the â-fluoro agostic structure seen in the
“outside” reaction pathway (3), the â-fluoroethyl intermediate
in the “inside” path, 4, has an essentially nonbonding Zr-F
distance (3.281 Å); minor stretching of the C-F bond is seen
(1.399 in TS4, vs 1.419 Å in 4). The values, both taken from
sp3 carbons, can also be compared with a different conformation
where the CH2F group has no interaction at all with the metal
center, 5, (Figure 7); in this case, the weakly agostic â-fluoro-
ethyl C-F bond is still stretched (1.419 vs 1.401 Å for the
noninteracting case).
From this minimum, a transition state (TS32) is formed which
connects (3) to the final products. This transition state is
surprisingly low in energy, lying only about 3.4 kcal/mol above
the agostic intermediate.
Geometric parameters change in the expected manner through-
out the reaction coordinate. The C-C bond, which begins as
1.348 Å in CH2CHF lengthens to 1.355 Å at TS13. It lengthens
further to 1.484 Å at 3, consistent with the formation of an alkyl
fragment, before shortening to 1.404 Å in TS32, in preparation
for the elimination of ethylene. A similar trend is seen in the
continuous stretching of the C-F distance from 1.338 in TS13
to 1.470 in 3 and finally 1.768 Å in TS32. The Zr-F bond
shortens correspondingly from 2.416 in 3 to 1.938 Å in the final
Cp2ZrClF product (2).
(b) “Inside” Attack. Attack of the olefin between the hydride
and chloride ligands provides a much lower activation barrier
of +8.1 kcal/mol (Figure 5). This lower barrier to insertion
might be explained by less steric repulsion from the cyclopen-
tadienyl ligands when the olefin is in this “inside” position
As in the “outside” path, the energy rise from 4 to TS42 is
very small: 3.5 kcal/mol. The Zr-Câ distance is only slightly
longer in this “inside” transition state than in TS32 of the
“outside” path, at 3.034 vs 2.943 Å, respectively.
The geometric parameters again correlate with the expected
changes in structure along the reaction coordinate. For example,
the C-C bond distance goes from a double bond distance of
1.349 in TS14 to an η1-alkyl like 1.508 in 4. This bond then
shortens to 1.484 Å in TS42 in preparation for the elimination
of ethylene (CdC distance of 1.348 Å). The C-F bond
lengthens throughout the reaction path, going from 1.339 to
1.399 to 1.419 to 1.471 Å for TS14, TS4, 4, and TS42. The Zr-F
distance is also correspondingly shorter, decreasing dramatically
from 3.281 Å in 4 to 2.609 Å in TS42 in preparation for its
final value of 1.938 Å in 2. One other parameter which is
interesting to follow is the Zr-CR distance. It shortens consider-
ably from TS14 to 4 (2.728 to 2.297 Å), which is expected as a
(19) Wu, Z.; Jordan, R. F.; Petersen, J. L. J. Am. Chem. Soc. 1995, 117,
5867; Carpentier, J.-F.; Wu, Z.; Lee, C. W.; Stro¨mberg, S.; Christopher, J.
N.; Jordan, R. F. J. Am. Chem. Soc. 2000, 122, 7750.
(20) In agreement with literature precedent, we will refer to this as an
agostic interaction. See Brothers, P. J.; Roper, W. R. Chem. ReV. 1988, 88,
1293.

608 J. Am. Chem. Soc., Vol. 123, No. 4, 2001
Watson et al.
calculation methods differ (and so energies cannot be compared
exactly), this qualitative result confirms that a heteroatom on
the olefin does not fundamentally alter the mechanism.
Minima for the â-F Ethyl Intermediates. Perhaps one of
the most surprising results of this study was that â-fluoro agostic
interactions give very little or no stabilization to the metal
fragment (Figure 7). For example, 5 clearly has no interaction
of the fluorine with the metal center. The Zr-F distance is quite
long (4.538 Å) and there is no stretching of the C-F bond. 4
is the very weakly agostic â-fluoro intermediate seen in the
“inside” reaction pathway. The two are basically conformers
of each other, and differ only by 1.8 kcal/mol (well within the
expected <3 kcal/mol difference for conformers). On the other
hand, the -16.4 kcal/mol structure 3 (the â-F agostic intermedi-
ate from the “outside” pathway) is strongly agostic. In addition
to a short Zr-F distance of 2.416, the C-F bond is clearly
stretched (nearly 0.07 Å from the middle structure). However,
the agostic structure is nearly 5 kcal/mol higher in energy than
its nonagostic counterpart!
Since none of these structures show any agostic hydrogen
interactions, the energies are not influenced by such factors.
These structures, however, do illustrate that the inside LUMO
is preferred. That is, when fluorine does interact with the
zirconium center, it does so with overlap of the inside orbital.
The lower energy of the “inside” insertion transition state also
serves to confirm this.
Figure 7. Selected bond distances (in Å), angles (in degrees) of
â-fluoroethyl intermediates.
However, the fact that the Zr/F interaction leads to a less
stable conformation is counterintuitive. It not only contradicts
the generalization that Zr prefers to interact with first row
heteroatoms, it also goes against the intuitive response that a
16 e- complex is a Lewis acid. There may be several reasons
for this observed paradox. First, there is a marked increase in
the Cl-Zr-CR angle in 3 as compared to 5, as well as a marked
stretch in the Zr-CR bond (2.403 vs 2.297 Å). This distortion
from the optimized structure may cause a poor overlap between
the carbon and zirconium orbitals. Second, the C-F bond is
being stretched; this demands an energetic price, due to the
strength of the C-F bond. For example, lengthening the CF
bond in CFH3 by 0.07 Å is calculated to cost 1.5 kcal/mol. Third,
in contrast to 5, which is staggered around the C-C bond, the
fluoroethyl 3 is forced to be eclipsed. This, too, costs some
energy in steric repulsion. F/Cl repulsion and a reduction in
the Lewis acidity of Zr because of π-stabilization from Cl also
serve to destabilize 3.
Figure 8. Energetics of hydrozirconation of CH2CH2.17
structure similar to a simple η1 alkyl is formed. This distance
then lengthens again (to 2.322 Å in TS42) in preparation for
the elimination of ethylene.
General Conclusions about the Insertion Pathways. In both
the inside and outside insertion pathways, no π-olefin complex
exists as an intermediate at the chosen level of theory. Instead,
the only π-olefin complexes found corresponded to transition
states. A preference for the inside LUMO is seen in the
difference between the energies of both the initial insertion
transition states as well as in the â-fluoroethyl structures (see
below). From the small energy cost to go from the fluoroethyl
intermediate to the third transition state, we can also see that it
is relatively easy to migrate a â-fluorine to the Zr center, further
confirming the assertion that Zr interacts strongly with first row
heteroatoms.
Comparison to the Calculated Hydrozirconation of
CH2CH2. It was found17 that ethylene insertion into the Zr-H
bond of Cp2ZrHCl occurred between the chloride and hydride
ligands, as opposed to attacking from the side opposite to the
Cl ligand. The difference in activation energy between these
two transition states, 14.7 kcal/mol at the MP2 level of theory,
was said to result from the smaller distortion of the complex
required to access the transition state in the “inside” insertion
case (Figure 8).
Several other â-fluoroethyl intermediates, corresponding to
rotation around the C-C, Zr-C, or ring puckering (from the
strongly agostic species), were also found. In all cases, they
were essentially isoenergetic ((0.5 kcal/mol) with those struc-
tures reported here.
Interactions of R-Fluoro Species. We have also calculated
structures resulting from the “wrong” insertion regiochemistry,
species where fluorine is on CR. When examining these minima,
strong R-fluoro/Zr interactions were again seen to deliver
negligible stabilization, as in the â-fluoro case (Figure 9). For
example, 6 has little if any Zr-F interaction. Not only is the
distance fairly long, the Zr-CR-F bond angle is very close to
tetrahedral, at 108.8°. This also serves to show that the best Zr
acceptor orbital is not “outside.” Contrast this with 7, which
involves donation to the “inside” acceptor orbital of Zr. Not
only is the Zr-F distance considerably shorter, but two other
indicators of agostic interaction are present as well: namely,
the C-F bond stretch and the marked decrease in the Zr-CR-F
angle (to 72.0°). It is obvious, then, that 7 has a much greater
degree of interaction between the Zr and F. However, just as in
The mechanistic preference is similar to the one just discussed
for the reaction of Cp2ZrHCl with vinyl fluoride; in both cases,
the faster insertion occurred from “inside.” Although the

C-D0 CleaVage in H2CdCH(D0) by (Cp2ZrHCl)n
J. Am. Chem. Soc., Vol. 123, No. 4, 2001 609
Scheme 5
of lack of back-bonding, there is still a considerable thermo-
dynamic gain in forming this carbene species from the separated
carbene and zirconocene (Cp2ZrHCl and C(F)(CH3)) (Scheme
5). That is, the high energy of isomerizing vinyl fluoride to
free CF(CH3), a, 53.4 kcal/mol, costs more than the (consider-
able) binding energy of this carbene to Cp2ZrHCl, b, 32.3 kcal/
mol. The Zr/C distance, 2.306, is consistent with a single bond
length.
9 also has some unexpected structural parameters. The Zr-
Cl distance is much longer than in the starting Cp2ZrHCl
structure (2.723 vs 2.431 Å), and the Cl/Zr/CR angle is
exceptionally small (at 47.90°). This could perhaps be best
explained by incipient donation of a Cl lone pair into the carbene
p-orbital (I). It is apparently easier to contract the Cl/Zr/CR angle
than to pyramidalize the R-carbon (which remains essentially
planar with the sum of angles around CR in 9 equaling 358.0°)
to promote better π-accepting. The chloride is still sufficiently
far away (2.08 Å) such that no bond with CR is actually made;
Figure 9. R-Fluoroethyl reaction pathways, with selected bond
distances (Å), angles (deg), and relative energies in kcal/mol
however, the sum of van der Waals radii for Cl and a benzene
carbon is 3.55 Å.21 This further confirms that some interaction
is taking place between the carbene carbon and the chloride,
although the minimum energy structure is not at the “R-chloro
agostic” alkyl stage (J), and certainly not the fully inserted
species Cp2Zr(H)(CClF(CH3)) analogous to the inserted product
Cp2ZrF(CHCl(CH3)). This weak but non-negligible interaction
between cis ligands is analogous to the “cis-effect” in cis-MH-
(H2) species.22 When we optimized geometry beginning from
Cp2Zr(H)(CClF(CH3)) where CR had an “ordinary” sp3 geom-
etry, this again optimized to structure 9, and thus the C-Cl
bond is nearly cleaved.
Figure 10. “Carbene” geometries.
the â-F intermediates, there is no energetic stabilization derived
from this interaction. Any stability from the Zr-F interaction
is thus offset by the cost of increasing the C/Zr/Cl angle from
95.6° in 6 to 114.1° in 7. One other conformer (rotation about
the Zr-CR bond to place F out of the plane) was also found. It,
too, was essentially isoenergetic, having an energy of -20.1
kcal/mol.
Here, too, “outer” addition has a higher barrier than “inner”
addition (Figure 9). Once again, the preference for the inside
LUMO is seen. In both of these transition states, however, the
energies are 2-4 kcal/mol higher than the corresponding
energies for along the â-fluoro pathway that leads to the
observed products. The degree of bond formation for this
regiochemistry (taken from Zr-CR distances) is slightly less in
the “outside” insertion TS16, and slightly more in the case of
the “inside” TS17 than for the other regiochemistry.
Stability of Carbenes Resulting from R-Fluoroethyl Isomer-
izations. One could imagine an isomerization process that would
convert the R-fluoroethyl complexes into carbenes (either Cp2-
ZrClF[CH(CH3)] or Cp2ZrClH[CF(CH3)], shown in Figure 10).
A potential energy surface search shows that the non-π-
substituted carbene, Cp2ZrClF[CH(CH3)], is not a minimum.
Instead, from various initial geometries, the carbene inserts into
the Zr-Cl bond to form Cp2ZrF[CHCl(CH3)], 8. Placing the
carbene closer to the fluorine (making a smaller CR/Zr/F angle)
in the starting geometry for optimization still results in the
insertion into the Zr-Cl bond.
Conclusions
This study provides confirmation of several factors that have
been thought to be true concerning the reaction of Zr with
olefins, but also reveals a number of surprises about the reaction
of an early metal species with a hard fluoride functionality. As
expected, the overall reaction is highly exothermic, forms a
Zr-F bond, and fails to bind C2H4 to an unsaturated, but d0
product. Unlike the d6 example MHClL2 (M ) Ru or Os),
carbene products Cp2ZrClF[CH(CH3)] or Cp2ZrClH[CF(CH3)]
are thermodynamically unfavorable due to lack of back-bonding
from Zr(IV).
All of the surprises relate to the mechanism and the energies
and structures of the fluoroethyl intermediates and transition
states. The approach of either the slender F-C bond or the
bulkier face (C/C π system) of vinyl fluoride is significantly
easier in the “inside” orientation, as the “inside” unoccupied
However, the π-substituted carbene, Cp2ZrClH[CF(CH3)], 9,
is a minimum (Figure 10). While high in energy compared to
the separated Cp2ZrHCl and CH2CHF (+21.1 kcal/mol) because
(21) Bondi, A. J. J. Chem. Phys. 1964, 68, 441.
(22) Maseras, F.; Lledos, A.; Clot, E.; Eisenstein, O. Chem. ReV. 2000,
100, 601.

610 J. Am. Chem. Soc., Vol. 123, No. 4, 2001
Watson et al.
(m, 2H, Cp2(Cl)ZrCH2CH2Ph)sall CH2 protons exhibited AA′XX′
coupling; Cp2(Cl)ZrCHPhCH3 (minor product), δ 7.10 (m, 5H, Cp2-
(Cl)ZrCHPhCH3), 6.01 (s, 10H, Cp2(Cl)ZrCHPhCH3), 2.94 (q, 1H, Cp2-
(Cl)ZrCHPhCH3), 1.50 (d, J ) 10 Hz, 3H, Cp2(Cl)ZrCHPhCH3).
Reaction of Cp2ZrHCl with CH2CHF. Cp2ZrHCl (10.1 mg, 0.039
mmol) was slurried in approximately 0.5 mL of C6D6 and placed in an
NMR tube under argon. CH2CHF (1.502 equiv) was added via standard
gas line techniques. The previously clear supernatant turned a pale
yellow color and the reaction was nearly homogeneous after 1.5 h of
orbital is a better Lewis acid. The regiochemistry of addition
of Zr-H across the CH2dCHF bond is kinetically controlled
by a TS17 - TS14 ) 11.6 - 8.1 ) 3.5 kcal/mol difference in
activation energy, which calculates to a 0.3% predicted yield
for the R-fluoroethyl product.
What is especially nonintuitive is the weakness of both R-F
and â-F agostic interactions in the Cp2ZrCl(fluoroethyl) inter-
mediates, given the seemingly favorable hard acid/hard base
interaction. “Intimate ion pairing” by C-F f Zr donation has
become a major study area in catalyzed olefin polymerization.23
We have resorted to steric repulsion and the energy cost of
bending bond angles around Zr to rationalize these results, but
the outcome remains fundamentally surprising.24
Given that the σ-bond metathesis mechanism was initially
devised for d0 cases where an oxidative addition/reductive
elimination sequence is untenable, it is interesting that it is
inferior here to â-H addition/â-F elimination. Because there is
here an energetically less costly mechanism, the σ-bond
metathesis pathway fails to operate. The σ-bond metathesis
transition state can be described as electrophile assisted
(Zr(IV)) nucleophilic (H-) attack on the substituted vinyl carbon.
It is noteworthy that these transition states show only modest
disruption of the CdC double bonds, as judged by their length.
The absence of any η2-olefin intermediate, of either ethylene
or vinyl fluoride, is noteworthy as well. While this claim might
be viewed as subtle or even subjective, it can be made objective
by the criterion of Zr/C bond length; there is no intermediate
where anything approaching a typical Zr-C(olefin) bond
distance is present. This does not create a significant kinetic
barrier, however, since every TS except the first lies below the
energy of the separated reactants; η2-C2H4 binding, if it occurred,
would not accelerate product formation.
1
agitation. H NMR (400 MHz, C6D6): ethylene (δ 5.24(s)), Cp2Zr-
(CH2CH3)Cl26 δ 1.08 (q, J ) 7 Hz, 2H, Cp2Zr(CH2CH3)Cl), 1.44 (t, J
) 7 Hz, 3H, Cp2Zr(CH2CH3)Cl), 5.75 (s, 10H, Cp2Zr(CH2CH3)Cl).
Due to the insolubility of dihalide species, only excess vinyl fluoride
could be identified reliably in the 19F NMR spectra. Halide exchange
products were also obtained. Unreacted CH2CHF could be observed
by NMR even when reacted as 0.90 equivalents. With identical
conditions as above, Cp2ZrHCl and CH2CHF were reacted in d8-THF,
to better observe growth of dihalide products. Within 1 h, the solution
was homogeneous and slightly yellow in color. (1H NMR, 400 MHz):
Cp2ZrCH2CH3(Cl), δ 6.318 (s, 10H, Cp2ZrClCH2CH3), 1.314 (t, J )
7.6 Hz, 3H, Cp2ZrClCH2CH3), 0.970 (q, J ) 7.6 Hz, 2H, Cp2ZrClCH2-
CH3); CH2CH2, δ 5.360 (s); other Cp resonances at δ 6.490, 6.437,
6.367 (assigned to Cp2ZrF2 by comparison to an authentic sample),
6.229. (19F, 376.5 MHz): δ 28.378, 2.051, -94.948, broad resonance
at -140.
Reaction of Cp2ZrHCl with 1,1 CH2CF2. Cp2ZrHCl (10.0 mg,
0.039 mmol) was slurried in approximately 0.5 mL of C6D6 and placed
in an NMR tube under argon. CH2CF2 (1.57 equiv) was added via
standard gas line techniques. The slightly yellow solution appeared
mostly homogeneous within 1.5 h. Excess CH2CF2, CH2CHF, Cp2Zr-
1
(CH2CH3)Cl, and CH2CH2 were then visible in the H and 19F NMR
spectra.
Reaction of Cp2ZrHCl with C6H5F. Cp2ZrHCl (10.0 mg, 0.039
mmol) was slurried in approximately 0.5 mL of C6D6 and placed in an
NMR tube under argon. C6H5F (4.0 µL, 0.0426 mmol, 1.1 equiv) was
added via syringe. No change was observed by NMR after 3 days
tumbling at room temperature. The mixture was still heterogeneous
with a clear supernatant. Heating of the NMR tube at 60° for
approximately 1 h produced a homogeneous purple solution, from the
decomposition of Cp2ZrHCl.27 The 1H and 19F NMR of the C6H5F was
unchanged, and no C6H6 was produced. Identical results were obtained
when a large excess of C6H5F (1:1 C6H5F/C6D5CD3 as solvent) or a
catalytic amount (0.1 equiv) of C6H5F was used. No reaction occurred
as well when C6H6 or d12-cyclohexane were used as solvents.
Similar methods employed with 1,3 C6H4F2, 1,4 C6H4F2, 4-fluoro-
anisole, and C6F6 also showed no reaction with Cp2ZrHCl even with
extended reaction time. Heating caused the decomposition of Cp2ZrHCl.
Synthesis of Cp2ZrF2. This procedure is a slight modification of
Seyam’s preparation.28 Cp2ZrCl2 (1.6 g, 5.47 mmol) was dissolved in
150 mL of acetone (dried with MgSO4 and molecular sieves and
degassed prior to use). AgBF4 (2.189 g, 11.25 mmol, 2.06 equiv)
dissolved in 50 mL of acetone was added dropwise through an addition
funnel wrapped in aluminum foil over 30 min at 0 °C. A white solid
(AgCl) formed on contact. The reaction solution was stirred for an
additional 15 min at room temperature. The yellow supernatant was
removed via cannula and filtered. The volume was then reduced to
approximately 30 mL, and 60 mL of dry ether was layered on top.
Recrystallization at -40° afforded a very pale beige solid. Isolated yield,
68%. 1H NMR (d8-THF, 400 MHz): 6.38 (s, Cp2ZrF2). 19F NMR (d8-
THF, 376.5 MHz): -157.3 (br s).
Experimental Section
General. All manipulations were performed using standard Schlenk
line and glovebox techniques. Solvents were dried and distilled
following standard protocols and stored in airtight solvent bulbs under
argon. All reagents for which a synthesis is not given are commercially
available from Aldrich or Lancaster and were used as received with
no further purification. All NMR solvents were dried, vacuum-
1
transferred, and stored in an argon-filled glovebox. H, 19F, and 13C
NMR were recorded on a Varian Gemini 2000, or Inova 700
spectrometer. Chemical shifts are reported in ppm and referenced to
residual solvent peaks (1H and 13C) or external CF3COOH (neat), -78.5
ppm relative to CFCl3 (19F).
Reaction of Cp2ZrHCl with â-Methoxy Styrene. 10 mg (0.0392
mmol) of Cp2ZrHCl was slurried in approximately 1 mL C6D6 and
placed in an NMR tube. 5.5 µL (0.0410 mmol, 1.05 equiv) of (C6H5)-
CHCH(OCH3) was added via syringe, and the tube shaken to thoroughly
mix the reagents. Within 5 min, a noticeable yellow tint of the
supernatant was apparent; after 2 h, the yellow solution was entirely
homogeneous. Cp2Zr(OMe)Cl, Cp2(Cl)ZrCH2CH2Ph, and Cp2(Cl)-
1
ZrCHPhCH3 were observed by NMR. H NMR, COSY, HMQC, and
DEPT experiments were used to confirm the observed products;
Cp2(Cl)ZrCH2CH2Ph, and Cp2(Cl)ZrCHPhCH3 were synthesized in-
dependently by the reaction of PhCHCH2 with Cp2ZrHCl and compared
with literature results.25 1H NMR (300 MHz, C6D6): Cp2Zr(OMe)Cl,
δ 5.73 (s, 10 H, Cp2ZrClOMe), 3.64 (s, 3H, Cp2ZrClOMe); Cp2(Cl)Zr-
CH2CH2Ph (major product), δ 7.22 (m, 5H, Cp2(Cl)ZrCH2CH2Ph), 5.90
(s, 10H, Cp2ZrCH2CH2Ph), 2.86 (m, 2H, Cp2(Cl)ZrCH2CH2Ph), 1.31
Computational Details
All calculations were performed with the Gaussian 98 package29 at
the B3PW9130 level of theory. Basis sets used include LANL2DZ for
Zr and Cl, 6-31G for all Cp carbons and hydrogens,31 6-31G* for
fluorine and the fluoroethylene carbons, and 6-31G** for the hydride
(23) Chen, E. Y.; Marks, T. J. Chem. ReV. 2000, 100, 1391.
(24) A referee disagrees that this is surprising and feels it is caused by
transfer to Zr of “...considerable π charge from the Cp ligands...”.
(25) (a) Nelson, J. E.; Bercaw, J. E.; Labinger, J. A. Organometallics
1989, 8, 2484. (b) Erker, G.; Kropp, K.; Atwood, J. L.; Hunter, W. E.
Organometallics 1983, 2, 1555-1561. (c) Chirik, P. J.; Day, M. W.;
Labinger, J. A.; Bercaw, J. E. J. Am. Chem. Soc. 1999, 121, 10308.
(26) Gell, K. I.; Posin, B.; Schwartz, J.; Williams, G. J. Am. Chem. Soc.
1982, 104, 1846.
(27) Wailes, P. C.; Weigold, H. J. Organomet. Chem. 1970, 24, 405.
(28) Seyam, A.; Samha, H.; Hodali, H. Gazz. Chim. Ital. 1990, 120, 527.

C-D0 CleaVage in H2CdCH(D0) by (Cp2ZrHCl)n
J. Am. Chem. Soc., Vol. 123, No. 4, 2001 611
and fluoroethylene hydrogens. The basis set LANL2DZ is the Los
Alamos National Laboratory ECP plus a double-ú valence on Zr and
Cl.32 Because of the electronegativity of fluorine, and thus the probable
inclusion of some d-orbital character in its bonding orbitals, in all DFT
calculations discussed here polarization functions were included on all
non-Cp ligands. All calculations were performed with C1 symmetry,
and all stationary points were characterized as minima or transition
states by a frequency analysis, which was also used to compute zero-
point energy corrections without scaling. For the 4-centered, σ-bond
metathesis transition states, a grid method was employed to find the
TS. The C-H distance was frozen to values between 2.6 and 1.4, and
additional dihedrals corresponding to F-H-Zr-Cl and CR-H-
Zr-Cl were frozen to 0°. The structures were allowed to optimize,
and a transition-state calculation was started from the highest-energy
structure that had not dissociated. Further optimization of the obtained
structure (with three imaginary frequencies) resulted in a true transition
state. All transition states were confirmed to lie on the reaction pathway
by distorting the complex along the imaginary frequency and optimizing
the resulting structure.
(29) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb,
M. A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A., Jr.;
Stratmann, R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A.
D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi,
M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.;
Ochterski, J.; Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick,
D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.;
Ortiz, J. V.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi,
I.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.;
Peng, C. Y.; Nanayakkara, A.; Gonzalez, C.; Challacombe, M.; Gill, P. M.
W.; Johnson, B. G.; Chen, W.; Wong, M. W.; Andres, J. L.; Head-Gordon,
M.; Replogle, E. S.; Pople, J. A. Gaussian 98, revision A.7; Gaussian,
Inc.: Pittsburgh, PA, 1998.
(30) Becke, A. D. Phys. ReV. 1988, A38, 3098.; Becke, A. D. J. Chem.
Phys. 1993, 98, 1372; Becke, A. D. J. Chem. Phys. 1993, 98, 5648; Perdue,
J. P.; Wang, Y. Phys. ReV. B. 1991, 45, 13244.
(31) Hariharan, P. C.; Pople, J. A. Theor. Chim. Acta 1973, 28, 213.
(32) Hay, P. J.; Wadt, W. R., J. Chem. Phys. 1985, 82, 270; Wadt, W.
R.; Hay, P. J. J. Chem Phys. 1985, 82, 284; Hay, P. J.; Wadt, W. R. J.
Chem. Phys. 1985, 82, 299.
Acknowledgment. This work was supported by the donors
of the Petroleum Research Fund. L.A.W. gratefully acknowl-
edges the National Science Foundation for a graduate fellowship.
JA0024340