Zirconocene allyl complexes: Dynamics in solution, reaction with aluminum alkyls, B(C6F5)3-induced propene insertion, and density-functional calculations on possible formation and reaction pathways

Article abstract of DOI:10.1021/om9906058

The complex (C₅H₅)₂Zr(Me)(methallyl) (1) was prepared and studied as a model for Zr-bound allyl species which are likely to arise in zirconocene-based polymerization systems. 1 undergoes allyl-alkyl exchange with trimethyl aluminum (TMA) or with methyl alumoxane (MAO) at rates that are proportional to the Al concentration, but remain 1-2 orders of magnitude below those of typical olefin insertions. Its perfluorotriphenylborane adduct 2, i.e., the contact ion pair (C₅H₅)₂Zr(methallyl)⁺+H ₃CB(C₆F₅)₃^-, has been characterized with regard to the rearrangement dynamics of its allyl ligand. Propene reacts with 2, at rates which are substantially lower again than those of cationic Zr-alkyl species, under insertion between Zr and one of the allylic termini. Scrambling of deuterium from an allylic CD₂ terminus over several C atom positions next to the unsaturated chain end indicates rather extensive metal migration within an initial olefin insertion product. Density-functional calculations indicate that insertion of propene directly into an η³-coordinated Zr-allyl unit occurs with lower activation energy than insertion into an η¹-bound Zr-allyl species and that the lowest-energy pathway for the reactivation of a cationic zirconocene allyl species is its reconversion to the corresponding Zr-alkyl species by molecular hydrogen.

Full text of DOI:10.1021/om9906058

Organometallics 2000, 19, 377-387
377
Articles
Zir con ocen e Allyl Com p lexes: Dyn a m ics in Solu tion ,
Rea ction w ith Alu m in u m Alk yls, B(C6F 5)3-In d u ced
P r op en e In ser tion , a n d Den sity-F u n ction a l Ca lcu la tion s
on P ossible F or m a tion a n d Rea ction P a th w a ys
Susanna Lieber, Marc-Heinrich Prosenc,*,† and Hans-Herbert Brintzinger*
Fakulta¨t fu¨r Chemie, Universita¨t Konstanz, D-78457 Konstanz, Germany
Received J uly 30, 1999
The complex (C5H5)2Zr(Me)(methallyl) (1) was prepared and studied as a model for Zr-
bound allyl species which are likely to arise in zirconocene-based polymerization systems. 1
undergoes allyl-alkyl exchange with trimethyl aluminum (TMA) or with methyl alumoxane
(MAO) at rates that are proportional to the Al concentration, but remain 1-2 orders of
magnitude below those of typical olefin insertions. Its perfluorotriphenylborane adduct 2,
-
i.e., the contact ion pair (C5H5)2Zr(methallyl)+H3CB(C6F5)3 , has been characterized with
regard to the rearrangement dynamics of its allyl ligand. Propene reacts with 2, at rates
which are substantially lower again than those of cationic Zr-alkyl species, under insertion
between Zr and one of the allylic termini. Scrambling of deuterium from an allylic CD2
terminus over several C atom positions next to the unsaturated chain end indicates rather
extensive metal migration within an initial olefin insertion product. Density-functional
calculations indicate that insertion of propene directly into an η3-coordinated Zr-allyl unit
occurs with lower activation energy than insertion into an η1-bound Zr-allyl species and
that the lowest-energy pathway for the reactivation of a cationic zirconocene allyl species is
its reconversion to the corresponding Zr-alkyl species by molecular hydrogen.
In tr od u ction
nocene allyl complexes as deactivated species was first
observed by Schumann, Marks, and co-workers.4 In a
study on the catalyst system (C5Me5)2ZrMe(THT)+-
B(C6H5)4-,5 Teuben and co-workers noted that deuteri-
olysis of the deactivated catalyst, which gave a mixture
of propene-d1 and isobutene-d1, indicated the presence
of zirconocene-allyl complexes in this reaction system.
Zirconocene-based catalysts for the polymerization of
R-olefins have been widely studied with regard to the
production of homo- and copolymers with variable
molecular weights and degrees of tacticity.1 While most
of these catalyst systems are highly active initially, rates
of monomer uptake generally decrease to rather small
fractions of their initial values within a few hours or,
in some cases, even within a few minutes.2,3 At present,
it is not fully understood which mechanisms cause this
deactivation.
In a complex of the type (C5Me5)2Zr(alkyl)(alkene)+
a
hydrogen atom is apparently transferred from a CH3
substituent of the coordinated alkene to the Zr-bound
alkyl group, such that formation of a Zr-allyl cation is
accompanied by liberation of an alkane (Scheme 1).
Formation of zirconocene-allyl complexes through C-H
activation of isobutene in a zirconocene-alkyl cation
was also observed by Horton and by van der Heijden
and Hessen.10,11
Alternatively, a zirconocene-alkyl cation could un-
dergo â-H transfer to the metal and subsequent C-H
activation of a CH2R or CH3 substituent of the resulting
olefin ligand to give an allyl cation of the type (C5H5)2-
One mechanism that has been postulated in this
regard is the formation of less active or inactive zirco-
nium-allyl complexes.4-9 The formation of lantha-
† Present address: University of North Carolina at Chapel Hill,
Venable and Kenan Laboratories, B 421 CB #3290, Chapel Hill, NC
27599-3290.
(1) Brintzinger, H. H.; Fischer, D.; Mu¨lhaupt, R.; Rieger, B.;
Waymouth, R. Angew. Chem., Int. Ed. Engl. 1995, 34, 1143. Bochmann,
M. J . Chem. Soc., Dalton Trans. 1996, 255. Kaminsky, W. J . Chem.
Soc., Dalton Trans. 1998, 1413.
(2) Fischer, D.; Mu¨lhaupt, R. J . Organomet. Chem. 1991, 417, C7.
(3) Chien, J . C. W.; Sugimoto, R. J . J . Polym. Sci. 1991, 29, 459.
(4) J eske, G.; Lauke, H.; Schock, L. E.; Swepstone, P. N.; Schumann,
H.; Marks, T. J . J . Am. Chem. Soc. 1985, 107, 8103. J eske, G.; Lauke,
H.; Mauermann, H.; Swepstone, P. N.; Schumann, H.; Marks, T. J . J .
Am. Chem. Soc. 1985, 107, 8091.
(5) Eshuis, J . J . W.; Tan, Y. Y.; Meetsma, A.; Teuben, J . H.;
Renkema, J .; Evens, G. G. Organometallics 1992, 11, 362.
(6) Richardson, D. E.; Alameddin, G. N.; Ryan, M. F.; Hayes, T.;
Eyler, J . R.; Siedle, A. R. J . Am. Chem. Soc. 1996, 118, 11244.
(7) Pindando, G. J .; Thornton-Pett, M.; Bouwkamp, M.; Meetsma,
A.; Hessen, B.; Bochmann, M. Angew. Chem., Int. Ed. Engl. 1997, 37,
22358.
(8) Feichtinger, D.; Plattner, D. A.; Chen, P. J . Am. Chem. Soc. 1998,
120, 7125.
(9) Guyot, A.; Spitz, R.; Dassaud, J .-P.; Gomez, C. J . Mol. Catal.
1993, 82, 29.
(10) Horton, A. D. Organometallics 1996, 15, 2675.
(11) van der Heijden, H.; Hessen, B.; Orpen, A. G. J . Am. Chem.
Soc. 1998, 120, 1112.
10.1021/om9906058 CCC: $19.00 © 2000 American Chemical Society
Publication on Web 01/25/2000

378 Organometallics, Vol. 19, No. 4, 2000
Lieber et al.
Sch em e 1
Sch em e 2
Zr(allyl)+ together with dihydrogen (Scheme 2).12 For-
mation of dihydrogen gas in the course of zirconocene-
catalyzed olefin polymerizations was indeed observed
by Horton13 and by Karol and Wassermann14 and
proposed to be due to the formation of allyl complexes
in the catalyst systems studied. Direct observations on
the formation of Zr-methallyl complexes by C-H acti-
vation of isobutene under release of H2 were obtained
by Richardson and co-workers from gas-phase MS
studies.6 Recently, Resconi has discussed possible roles
of zirconocene-allyl complexes in stereoerror formation
by chain-end isomerizations,15 which are observed par-
ticularly at reduced monomer concentrations.16,17
The further fate of any such allyl species arising
during polymerization catalysis is not clear. Teuben and
co-workers assumed that a (C5Me5)2Zr(allyl)+ complex
is catalytically inactive.5 Guyot, Spitz, and co-workers
postulated that inactive Ti-allyl complexes formed by
Ti-based heterogeneous catalysts might be recon-
verted to active Ti-alkyl species by alkyl exchange
with aluminum alkyls.9 Erker and co-workers, on the
other hand, observed the insertion of R-olefins into a
zirconocene allyl-borate “betain” species, derived from
a Zr-butadiene complex by reaction with B(C6F5)3.18
It thus remains to be explored whethersand at which
ratessolefins can insert into Zr-allyl complexes which
might arise in a catalytic reaction system. Here we
report results of experimental and theoretical studies
on possible reaction pathways of zirconocene-allyl
model complexes.
Resu lts a n d Discu ssion
The dynamic behavior of Zr-allyl complexes and their
reactivity toward Lewis acids and olefins were studied
using the 2-methylallyl (methallyl) zirconocene complex
(C5H5)2ZrMe(C4H7), 1, as a model for an allyl complex
carrying a polymer alkyl chain. This complex was pre-
pared and isolated by reaction of excess methallylmag-
nesium chloride in THF with (C5H5)2ZrMeCl.19 Complex
1 is an air- and water-sensitive yellow solid; it can be
recrystallized from pentane in the form of small yellow
needles. Complex 1 is soluble and stable in ethers and
hydrocarbons but decomposes slowly in chlorinated
1
solvents. H NMR spectra of 1 in C6D6 at room temper-
ature give rise to only one resonance for all four allylic
protons. 1H and 13C chemical shifts of the terminal CH2
groups of the C4H7 ligand, 2.54 and 67 ppm, respec-
tively, and an IR absorption at 1520 cm-1 correspond
to an η3-π-bound, rather than to an η1-σ-bound allyl
group.20 The equivalence of all terminal allyl protons
even at -90 °C indicates a rapid haptotropic allyl
rearrangement21 with a low activation barrier of ∆Gq
< 30 kJ /mol.
1. Rea ction w ith Alu m in u m Alk yls. To study the
reactivity of complex 1 toward aluminum-based cocata-
lysts, 1 M solutions of trimethylaluminum (TMA) were
mixed with C6D6 solutions of complex 1 in a glovebox.
Reaction of complex 1 with 1 equiv of TMA (i.e., 0.5
equiv of Al2Me6) causes the bright yellow color of
complex 1 to diminish. New signals appearing in the
1H NMR spectra at 5.7 and -0.15 ppm were assigned
to (C5H5)2ZrMe2. A new methallyl species, presumably
Me2Al-C4H7, gives rise to a broad singlet at 2.92 ppm
and to a likewise broadened CH3 signal appearing at
1.76 ppm, rather than at 2.54 and 1.5 ppm, as observed
in complex 1.
(12) A similar reaction path for ethene polymerization was studied
by: Margl, P. M.; Woo, T. K.; Ziegler, T. Organometallics 1998, 17,
4997.
(13) Horton, A. D. Private communication, 1998.
(14) Karol, F. J .; Kao, S.-C.; Wassermann, E. P.; Brady, R. C. New
J . Chem. 1997, 21, 797. Wassermann, E. P. Abstracts of Papers, ACS,
081-poly, 1998; 216.
(15) Resconi, L.; Camurati, I.; Sudmeijer, O. Top. Catal. 1999, 7,
145.
(16) Busico, V.; Cipullo, R. J . Am. Chem. Soc. 1994, 116, 9329.
(17) (a) Leclerc, M. K.; Brintzinger, H. H. J . Am. Chem. Soc. 1995,
117, 1651. (b) Leclerc, M. K.; Brintzinger, H. H. J . Am. Chem. Soc.
1996, 118, 9024. (c) Prosenc, M. H.; Brintzinger, H. H. Organometallics
1997, 16, 3889.
(19) Wailes, P. C.; Weigold, H.; Bell, A. P. J . Organomet. Chem.
1971, 33, 181.
(20) Martin, H. A.; Lemaire, P. J .; J ellinek, F. J . Organomet. Chem.
1968, 14, 149. Schlosser, M.; Sta¨hle, M. Angew. Chem. 1980, 92, 497.
Wilke, G.; Bogdanovic, B.; Hardt, P.; Heimbach, P.; Keim, W.; Kro¨ner,
M.; Oberkirch, W.; Tanaka, K.; Steinru¨cke, E.; Walter, D.; Zimmer-
mann, H. Angew. Chem. 1966, 78, 157.
(18) Insertion of one propene/Zr into the “betaine” complex (C5H5)2-
Zr(η3-CH2CHCH2CH2B(C6F5)3) was reported by Erker and co-work-
ers: Temme, B.; Karl, J .; Erker, G. Chem. Eur. J . 1996, 2, 919. Karl,
J .; Erker, G.; Fro¨hlich, R. J . Am. Chem. Soc. 1997, 119, 11165. Karl,
J .; Erker, G. J . Mol. Catal. 1998, 128, 85.
(21) Krieger, J . K.; Deutch, J . M.; Whitesides, G. M. Inorg. Chem.
1973, 12, 1535. Erker, G.; Berg, K.; Angermund, K.; Kru¨ger, C.
Organometallics 1987, 6, 2620. Landis, C. R.; Cleveland, T.; Casey, C.
P. Inorg. Chem. 1995, 34, 1285; Abrahams, M. B.; Yoder, J . C.; Loeber.
C.; Day, M. W.; Bercaw, J . E. Organometallics 1999, 18, 1873.

Zirconocene Allyl Complexes
Sch em e 3
Organometallics, Vol. 19, No. 4, 2000 379
associative process, i.e., with an attack of AlMe3 at the
Zr-bound methallyl ligand.
Reaction of complex 1 with TMA-free MAO24 results
in the formation of (C5H5)2ZrMe2 and of a species with
a very broad resonance at ca. 3.1 ppm, presumably some
MAO-methallyl complex. At Zr:Al ratios of ca. 1:1, 1:2,
1:5, and 1:10, we observe 1 and (C5H5)2ZrMe2 in ratios
of 1.5:1, 1.2:1, 0.8:1, and 0.6:1, respectively. This indi-
cates that the equilibrium between complex 1 and
(C5H5)2ZrMe2 is established also in the presence of MAO
and that the equilibrium lies somewhat farther on the
side of (C5H5)2ZrMe2 and MAO-(C4H7), with an equi-
librium constant K ) 0.1 (mol/L)1/2. 2D-NOESY data for
the methyl exchange at a Zr:Al ratio of 1:10 give a lower
Ta ble 1. Rela tive Con cen tr a tion s of Com p lex 1
a n d (C5H5)2Zr Me2 a n d Equ ilibr iu m a n d Ra te
Con sta n ts for Th eir In ter con ver sion w ith Resp ect
to th e Zr /Al Ra tioa
limit of kobs ) 0.4 s-1 25
which is comparable to that
,
observed for (C5H5)2ZrMe(C4H7) and TMA.
RAl ) [Zr]/[Al]
In further experiments, complex 1 was equilibrated
with 10 equiv of TMA and 1 equiv of a variety of
zirconocene dimethyl complexes. In the presence of
(C5H4Me)2ZrMe2, (C5H4Me)2ZrMe(C4H7) and (C5H5)2-
ZrMe(C4H7) are formed in a ratio of 1:8. This indicates
that substitution of the C5-ring disfavors the formation
of an allyl complex. In accord with this, no allyl
exchange was detectable when (C5H5)2ZrMe(C4H7) was
analogously reacted with the more highly substituted
complexes Me2Si(2-Me-4-tBu-C5H2)2ZrMe2 or Me2Si(2-
Me-indenyl)2ZrMe2. In the presence of the more open
complexes Me4C2(C5H4)2ZrMe2, Me2Si(C5H4)2ZrMe2, or
Ph2C(C5H4)(fluorenyl)ZrMe2, on the other hand, the
respective methallyl complexes were found to dominate
in the equilibrium mixtures, in ratios of 4:1 to 6:1
relative to (C5H5)2ZrMe(C4H7), documenting again the
sensitivity of these allyl exchange equilibria with re-
spect to steric factors. All of these exchange reactions
occur only in the presence of TMA, as did those
described above for (C5H5)2ZrMe2 and (C5H5)2ZrMe-
(C4H7) alone.
2. Rea ction s of Com p lex 1 w ith B(C6F 5)3. Upon
addition of a slight excess (1.1 equiv) of tris(pentafluo-
rophenyl) borane, the color of a solution of complex 1 in
benzene changes from yellow to orange-red.26 This is
accompanied by a broadening of the singlet of the allylic
protons as well as a chemical shift of the Zr-bound CH3
group to -0.3 ppm. A chemical shift difference between
the m- and p-fluorine atoms of ∆δ ) 3 ppm indicates
the presence of a weakly bound methylborate anion in
an ion pair such as 2 (Scheme 4), which has been
observed before by Horton in a study of â-methyl
elimination from zirconocene neopentyl complexes.10
Tris(pentafluorophenyl)borane thus attacks (C5H5)2-
ZrMe(C4H7) at its methyl group rather than at its allyl
group, to yield a zirconocene-allyl cation with a coor-
dinated methylborate group, in contrast with the reac-
tion of TMA or MAO with (C5H5)2ZrMe(C4H7), which led
to exchange of the methallyl moiety.
1:1
1:3
1:5
1:10
0.71
Rint ) [1]/[(C5H5)2ZrMe2]b
1.95 1.38 1.0
equilibrium constantc Kex/(mol/L)1/2 0.07 0.06 0.07 0.08
exchange rated kobs/s-1
0.12 0.22 0.30 0.42
1.7 1.8 1.8 1.8
rate constante kex/103 (mol/L)-1 s-1
a
At a total zirconocene concentration of 0.08 mol/L in C6D6
b
solution at 300 K. Integral ratio of the C5H5 signals of complexes
1 and (C5H5)2ZrMe2. c According to Scheme 3. Determined by 1H-
d
NOESY spectra according to ref 22. e Derived from kex ) kobs
×
-1/2
2([Al2Me6]Kdiss
)
with Kdiss ) 2.6 × 10-8 mol/L (cf. ref 23).
This reaction between complex 1 and TMA occurred
instantaneously at room temperature but did not go to
completion. The final ratio between 1 and (C5H5)2ZrMe2
was ca. 2:1, while reaction with 1.5 equiv of Al2Me6 gave
1 and (C5H5)2ZrMe2 in a ratio of 1.4:1. For the equilib-
rium between these two compounds and excess Al2Me6
(cf. Scheme 3), the equilibrium constant Kex and appar-
ent exchange rates kobs were determined in C6D6 solu-
tions containing complex 1 and Al2Me6 in Zr:Al ratios
of 1:1, 1:3, 1:5, and 1:10, using two-dimensional NMR
techniques (cf. Table 1).
1H-NOESY spectra of these solutions indicate meth-
allyl-methyl exchange between the Zr centers of (C5H5)2-
ZrMe(C4H7) and (C5H5)2ZrMe2 and the Al centers of
TMA and the aluminum methallyl product. Methyl-
methallyl exchange between (C5H5)2ZrMe(C4H7) and
(C5H5)2ZrMe2 is entirely suppressed in the absence of
TMA. Rather than between (C5H5)2ZrMe(C4H7) and
(C5H5)2ZrMe2 directly, the exchange thus proceeds via
reversible exchange of methallyl and methyl ligands
between Zr and Al centers. The apparent exchange rate
constant kobs is approximately proportional to the square
root of the Al2Me6 concentration (cf. Table 1). This
indicates that reaction of monomeric AlMe3 with com-
plex 1 is rate determining for the exchange process.
Using an equilibrium constant Kdiss ) 2.6 × 10-8 mol/L
23
for the dissociation of Al2Me6 to monomeric AlMe3,
a
value of kex ) 1.7 × 103 L‚mol-1‚s-1 is determined for
the second-order rate constant of the exchange reaction
between AlMe3 and complex 1 (cf. Table 1).
At temperatures below -10 °C in toluene solution,
two singlets are observed for the allylic protons of
complex 2. Coalescence of these signals at 0 °C indicates
When this rate constant for methyl exchange between
complex 1 and AlMe3 is determined at varying temper-
atures in the range 0-40 °C, estimates of ∆Hq ) +34
( 1 kJ /mol and of ∆Sq ) -70 ( 2 J /(mol‚K) are obtained.
The negative activation entropy is in accord with an
(24) Tritto, I.; Donetti, R.; Sacchi, M. C.; Locatelli, P.; Zannoni, G.
Macromolecules 1997, 30, 1247.
(25) Exchange between (C5H5)2ZrMe(C4H7) and MAO is accompanied
by reaction of (C5H5)2ZrMe2 with MAO under formation of an oily
-
precipitate, presumably (C5H5)2ZrMe+MeMAO
(26) At zirconocene concentrations above
precipitates shortly after.
.
(22) Perrin, C. L.; Dwyer, T. J . Chem. Rev. 1990, 90, 935.
(23) Smith, M. B. J . Organomet. Chem. 1972, 46, 31.
1 mmol/L, a red oil

380 Organometallics, Vol. 19, No. 4, 2000
Lieber et al.
Sch em e 4
Sch em e 5
group, as indicated by evolution of isobutene.30 When
20
the complex (C5H5)2Zr(C4H7)2 was reacted with the
a free energy of activation of ∆Gq ) 54 kJ /mol for the
haptotropic allyl rearrangement, as compared to ∆Gq
< 30 kJ /mol in complex 1. This increased activation
barrier undoubtedly arises from an increased electron
deficiency at the Zr center in the ion pair 2, which
disfavors the η1-coordination of the allyl ligand required
for the exchange process.
When measured at -20 °C in bromobenzene, in which
it is soluble up to 50 mmol/L, complex 2 gives rise to
two separate C5H5 signals at 5.43 and 5.31 ppm and
two sharp CH2 signals at 2.97 and 2.45 ppm. Upon
warming, the C5H5 signals coalesce at 5 °C, indicating
a relatively facile rotation of the methallyl ligand.
Coalescence of the methallyl CH2 signals, on the
other hand, requires warming to 38 °C, from which we
estimate an allyl rearrangement barrier of ∆Gq ) 63
kJ /mol as compared to ∆Gq ) 54 kJ /mol in toluene.
The more polar solvent appears to cause an increased
charge separation in the ion pair 2 and, hence, an even
stronger preference for η3-coordination of the allyl
ligand.
N,N-dimethylanilinium salt, only one allyl group was
protonated, but the dimethylaniline produced is par-
tially coordinated to the Zr center, as documented by
proton NMR spectra.31
3. Rea ction of Ion P a ir 2 w ith P r op en e. Upon
addition of 1 equiv of propene to a solution of com-
plex 1 and borane at -60 °C, we observe formation of
short oligomeric chains, with large parts of ion pair 2
remaining unreacted, rather than insertion of one
propene molecule per Zr center. At room temperature,
oligomers are formed upon addition of ethene or pro-
pene. After reaction of 10-12 equiv of propene, some
unreacted 2 is still detectable by NMR. These observa-
tions indicate that the first propene insertion into the
Zr-allyl cation of ion pair 2 occurs significantly slower
than subsequent insertions into the ensuing cationic
polymeryl species.
In a sample prepared with a limited (ca. 50-fold) ex-
cess of propene, the oligomers obtained show only 2-pro-
penyl and n-propyl end groups. The absence of any de-
tectable fraction of isopropyl end groups indicates that
propene inserts only into the Zr-methallyl bond of com-
plex 2 and not into its Zr-CH3 bond. In an effort to clar-
ify the reaction mechanism of this insertion, in particu-
lar whether propene undergoes a (presumably R-agos-
tically assisted) insertion into the Zr-C σ-bond of an
η1-coordinated allyl ligand4,18 or an insertion directly
into the η3-coordinated Zr-allyl fragment (Scheme 6),
we have studied the distribution of D atoms in the end
groups of propene oligomers arising from the deu-
terated allyl complex (C5H5)2ZrMe(CH2C(CH3)CD2) in
the presence of B(C6F5)3 and a 10-12-fold excess of
propene.
Determination of the exchange rates by two-dimen-
1
sional H NMR methods in the temperature range -
20 to +25 °C gives activation parameters of ∆Hq ) +42
( 1 kJ /mol and ∆Sq ) - 67 ( 4 J /(mol‚K) for the allyl
rotation and of ∆Hq ) + 50 ( 4 kJ /mol and ∆Sq ) - 56
( 5 J /(mol‚K) for the haptotropic allyl rearrangement.
Isotope effects on the kinetics of this rearrangement
were studied with the methylborate ion pair derived
from the deuterated methallyl complex (C5H5)2ZrMe-
1
(CH2C(CH3)CD2) (3). Two-dimensional H NMR mea-
surements in bromobenzene solution at 25 °C gave an
exchange rate constant of kD ) 22.3 s-1 for the protons
at the CH2 terminus, whereas for the fully protonated
complex kH ) 16.0 s-1 was obtained for the same
process. The inverse kinetic isotope effect of kH/kD ) 0.72
documents that the frequencies of the C-H and C-D
bonds at the allylic termini increase in the exchange
transition state. This is indeed to be expected if the η3-
bound allyl ligand is converted to an η1-geometry (cf.
Scheme 5), since typical IR frequencies of νCH ) 3100
cm-1 and νCD ) 2250 cm-1, observed for CH2 and CD2
termini in uncomplexed R-olefins,27 are higher than
those of νCH ) 2922 cm-1 and νCD ) 2081 cm-1 observed
in complex 3.28,29
In-situ D-NMR spectra of the oligomers formed in
these reaction systems showed, to our surprise, sub-
stantial D-atom contents not only in the terminal and
penultimate methylene groups of the oligomers, which
derive from the deuterated allyl ligand, but also in
methylene groups inside the oligomer chains and espe-
cially in their methyl substituents (Figure 1).32 This
(27) Willi, A. V. Isotopeneffekte bei chemischen Reaktionen; G.
Thieme: Stuttgart, 1983; p 20.
(28) This excludes stabilization of the η1-coordinated intermediate
by an agostic interaction of the released CD2 terminus with the
zirconium center, for which a normal kinetic isotope effect of kH/kD
≈
1.3 is expected.
Attempts to generate cations with less strongly bound
anions by reaction of complex 1 with N,N-dimethyl-
anilinium tetrakis(pentafluorophenyl)borate or [Ph3C]+-
[B(C6F5)4]- were not successful. N,N-Dimethylanilinium
cation protonates both the methyl group and the allyl
(29) Prosenc, M. H.; J aniak, C.; Brintzinger, H. H. Organometallics
1992, 11, 4036.
(30) Similar observations were made by Hessen and van Hejden in
studies with related Hf complexes: Hessen, B.; van Hejden, H. J .
Organomet. Chem. 1997, 534, 237.
(31) Horton, A. D.; Orpen, A. G. Organometallics 1992, 11, 8.

Zirconocene Allyl Complexes
Organometallics, Vol. 19, No. 4, 2000 381
F igu r e 1. 2H NMR spectrum of the reaction product of [(C5H5)2Zr(C4H5D2)][MeB(C6F5)3] (4) with 10-12 equiv of propene.
Sch em e 6
observation indicates that D atoms are efficiently
scrambled between all positions of the methallyl end
group and at least one adjacent propene unit. Due to
this D-atom scrambling, we are not able to distinguish
between insertions into an η1-σ- and an η3-π-bound allyl
ligand. If the reaction is quenched with water im-
mediately (within 5 s) after the addition of propene, we
find substantial D-atom scrambling already; the degree
of scrambling increases, however, when the reaction
mixtures are kept for several hours without quenching.
At present it is not clear, therefore, to which degree this
D-atom scrambling occurs during the catalysis itself s
i.e., concomitantly with the initial insertion stepssor
by subsequent exposure of the olefinic oligomers to the
Lewis-acidic reaction system.
If zirconocene catalyst centers are indeed involved,
this rearrangement would probably be due to the
availability of five- and six-membered cyclic intermedi-
ates (cf. Scheme 7), which allow Zr centers to migrate
to either one of the adjacent methylene and methyl C
atoms in a manner reminescent of the D-label scram-
bling between methylene and methyl groups associated
with stereoerror formation by chain-end isomeriza-
tions.16,17
(32) In 13C NMR spectra of the oligomers, a multitude of unassign-
able signals are apparent, presumably due to the presence of widely
varying degrees of oligomerization and stereoregularity. Changes
observed in the spectra of the deuterated products were thus not
assignable either.

382 Organometallics, Vol. 19, No. 4, 2000
Lieber et al.
F igu r e 2. Energy profile and structures for the formation of a zirconium-allyl complex via C-H activation and subsequent
dihydrogen formation.
Sch em e 7
Com p u ta tion a l Stu d y
study the reaction of the resulting zirconocene allyl
cation with propene, conceivable reaction paths for
carbon-carbon bond formation were investigated by
studying the insertion of a propene ligand into a Zr-
σ-allyl or Zr-π-allyl bond.
4. Zir con iu m -Allyl F or m a tion . As the initial state
of the allyl formation reaction,6,10 we chose the cation
[(C5H5)2Zr-H(isobutene)]+, A-1. This cation can be
formed, with a calculated activation energy of 40 kJ /
mol,17c through â-hydrogen transfer to the metal from
a Zr-bound isobutyl ligand, which represents a reason-
able model for a Zr-bound polypropene chain. The
energy of cation A-1, which is calculated to be 24 kJ /
mol higher than the energy of the related isobutyl cation
A-0,17c was set to 0 kJ /mol as the initial state of the
allyl formation reaction.
Experimental indication for formation of zirconocene
allyl complexes and their reaction with propene raises
the question as to reaction sequences and relative
activation energies for these reactions in comparison to
activation energies for chain growth and chain epimer-
ization in homogeneous polymerization of propene with
zirconocenes.
To gain more detailed insights into formation of a
cationic zirconocene allyl complex during polymerization
and further reaction of this cation with propene, these
reactions were studied by density-functional methods.
We investigated the reaction path in Scheme 2 for
intramolecular C-H activation in the model system
[(C5H5)2Zr-H(isobutene)]+ with special interest in ac-
tivation energies for the formation of the dihydrogen
ligand and its replacement by a propene molecule. To
Geometry optimization of the cation A-1 revealed a

Zirconocene Allyl Complexes
Organometallics, Vol. 19, No. 4, 2000 383
Ta ble 2. Bon d Dista n ces a n d An gles for
Geom etr ies A-1 to A-3
below that of transition state A-2. In this cation A-3
the dihydrogen ligand is weakly bound to the metal
center by 25 kJ /mol and can be displaced by a propene
molecule or an anion such as MAO-X- or borate.
Dissociation of the dihydrogen ligand requires about the
same energy (25 kJ /mol) as the reverse reaction from
cation A-3 to A-1 (22 kJ /mol). It is thus likely that a
finite fraction of intermediate A-3 will react under loss
of dihydrogen especially in the presence of a large excess
of monomer.
Formation of the cationic zirconocene allyl complex
A-3 from a cationic zirconocene alkyl resting state
A-0 requires an overall activation barrier of 67 kJ /
mol and will thus normally be slow in comparison to
chain growth (ca. 25-30 kJ /mol37-42). At sufficiently low
olefin pressure, however, allyl formation could be com-
petitive.
5. In ser tion P a th w a ys. Displacement of the dihy-
drogen ligand in cation [(C5H5)2Zr-η3-methallyl(H2)]+,
A-3, by a propene molecule leads to intermediate B-1
(Figure 3). The geometry optimization of [(C5H5)2Zr-
methallyl(propene)]+ (B-1) reveals a loosely bound pro-
pene molecule and an η3-methallyl ligand attached to
the Zr center. The long Zr-propene bond distances of
d(Zr-C(5)) ) 441 pm and d(Zr-C(6)) ) 420 pm are
more indicative of an electrostatic interaction than of a
propene ligand π-bound to the metal. The binding
energy between the propene molecule and the [(C5H5)2-
Zr-methallyl]+ fragment is calculated to be -28 kJ /mol,
which is only 3 kJ /mol stronger than the bond energy
of the dihydrogen ligand. The geometry parameters of
the methallyl ligand (Table 3) remain unchanged com-
pared to complex A-3.
To assess the activation barrier for insertion of pro-
pene into a Zr-bound allyl group in the cation B-1, we
chose as the reaction coordinate the C(3)-C(5) distance
between the inner olefin terminus (C(5)) and the inner
methylene group of the allyl ligand (C(3)) (Figure 3,
right), with all other geometry parameters allowed to
relax.
A-1
A-2
A-3
Zr-C(1)
Zr-C(2)
Zr-C(3)
Zr-H(1)
Zr-H(2)
Zr-CEa
255
301
364
185
372
216/216
134
111
293
256
260
254
183
196
220/218
130
136
112
245
257
254
205
206
219/217
130
203
84
CE-Zr-CEa
C-H(2)
H(1)-H(2)
a
CE ) C5-ring centroid.
nearly Cs-symmetric conformation with Zr, H(1), C(1),
and C(2) lying in the plane that bisects the angle
between the C5-ring ligand planes. The isobutene ligand
is asymmetrically bound with Zr-C bond distances of
255 and 301 pm (Figure 2). Such asymmetrically bound
olefin molecules were previously reported in related
theoretical17c,33 and single-crystal structural studies.34
The hydride ligand H(1) is found at a distance of 185
pm from the Zr center: This is comparable with
reported experimental and theoretical data of zir-
conocene hydride cations.17c,33,35 Bond parameters cal-
culated for cation A-1 are listed in Table 2.
To model the intramolecular C-H activation, the
distance between the hydride ligand H(1) and one
hydrogen atom of an isobutyl methyl group, H(2), was
chosen as reaction coordinate (Figure 2). Decreasing the
H(1)-H(2) distance from 293 pm in A-1 to 112 pm leads
to the transition state A-2 (Figure 2). Geometry opti-
mization of cation A-2 revealed a relatively low energy
of 33 kJ /mol above A-1 and bond distances and angles
as listed in Table 2. The C(1)-C(2) bond vector is tilted
out of the C5H5-Zr-C5H5 bisector plane by 24° such
that one of the isobutene methyl groups (C(3)) appears
in the equatorial plane within bonding distance to the
metal center. The Zr-carbon bond distances (d(Zr-C(1))
) 256 pm and d(Zr-C(3)) ) 254 pm) differ slightly from
bond distances reported for a Zr-bound allyl ligand.36
The C(3)-H(2) bond is elongated to a value of 136 pm;
the Zr-H(2) distance of 196 pm for the bridging
hydrogen ligand is longer than the terminal Zr-H(1)
distance of 183 pm, which remains nearly unchanged
in comparison to initial state A-1.
The reaction profile calculated for optimized geom-
etries along the reaction coordinate is shown in Figure
3. Decreasing the C(3)-C(5) bond distance from ca. 295
pm to 260 pm leads, over a shallow plateau, to the
transition state B-2 at a C(3)-C(5) bond distance of 226
pm (Figure 3). In the optimized geometry of the transi-
tion state B-2 the metal-carbon bond distances of Zr-
C(5) ) 299 pm and Zr-C(6) ) 239 pm are shorter than
those in the initial state B-1. The methallyl ligand is
asymmetrically bound to the metal with bond distances
of d(Zr-C(1)) ) 248 pm and d(Zr-C(3)) ) 296 pm. The
coordination site thus opened at the metal is occupied
by the olefin ligand (Figure 3). The energy of transition
state B-2 is calculated to be 61 kJ /mol above that of the
initial state B-1. Further decrease of the bond distance
Further decrease of the H(1)-H(2) bond distance to
approximately 80 pm leads to a minimum A-3, the
cation [(C5H5)2Zr-η3-methallyl(H2)]+ (Figure 2). In this
final state A-3, the methallyl ligand is bound in an η3-
fashion to the Zr center with bond distances of d(Zr-
C(1)) ) 245 pm and d(Zr-C(3)) ) 254 pm, which are
slightly shorter than those in transition state A-2. The
geometry of the dihydrogen ligand shows a Zr-η2-H2
interaction with equal metal-hydrogen bond distances
(d(Zr-H(1)) ) 205 pm, d(Zr-H(2)) ) 206 pm); both
hydrogen atoms lie in the equatorial plane.
(37) Prosenc, M. H.; Schaper, F.; Brintzinger, H.-H. In Metalorganic
Catalysts for Synthesis and Polymerization,; Kaminsky, W., Ed.;
Springer-Verlag: Berlin, 1999; p 223.
(38) Margl, P.; Deng, L.; Ziegler, T. J . Am. Chem. Soc. 1998, 120,
5517.
The energy of the final state, A-3, is calculated to be
+11 kJ /mol above the initial state, A-1, and 22 kJ /mol
(33) Margl, P.; Deng, L.; Ziegler, T. Organometallics 1998, 17, 933.
(34) Wu, Z.; J ordan, R. F.; Petersen, J . L. J . Am. Chem. Soc. 1995,
117, 5867.
(35) Yang, X.; Stern, C. L.; Marks, T. J . Angew. Chem. 1992, 104,
1406-1408.
(36) Wielstra, Y.; Duchateau, R.; Gambarotta, S.; Bensimon, C.;
Gabe E. J . Organomet. Chem. 1991, 418, 183.
(39) Støvneng, J . A.; Rytter, E. J . Organomet. Chem. 1996, 519, 277.
(40) Lohrenz, J . C. W.; Woo, T. K.; Ziegler, T. J . Am. Chem. Soc,
1995, 117, 12893.
(41) Yoshida, T.; Koga, N.; Morokuma, K. Organometallics 1995,
14, 746.
(42) Yoshida, T.; Koga, N.; Morokuma, K. Organometallics 1996,
15, 766.

384 Organometallics, Vol. 19, No. 4, 2000
Lieber et al.
F igu r e 3. Energy profile and structures for the insertion of propene into a zirconium-allyl σ-bond (left) and into a
zirconium-allyl π-bond (right).
Ta ble 3. Bon d Dista n ces a n d An gles for
Geom etr ies B-1 to B-3
mol higher than that for insertion of a propene ligand
into a Zr-σ-ethyl bond in the cation [(C5H5)2Zr-ethyl-
(propene)]+.37,39,41,42
B-1
B-2
B-3
An alternative reaction path involving insertion of a
propene molecule into the Zr-C σ-bond of an η1-bound
allyl ligand was modeled by choosing the C(5)-C(1)
bond distance as the reaction coordinate.
Zr-C(1)
246
259
249
441
420
316
142
219/218
133
248
261
296
299
239
216
140
219/219
131
263
303
354
346
236
156
138
219/219
134
Zr-C(2)
Zr-C(3)
Zr-C(5)
Zr-C(6)
The optimized start geometry C-1 (Figure 3, left)
reveals that the η1-bound methallyl ligand is rotated
around its Zr-C(1) bond by 80°, with a Zr-C(1) distance
of 223 pm. The geometry of cation C-1 is similar to that
expected for a σ-bound alkyl group in related cationic
zirconocene complexes.37-42 In this geometry, the sub-
stituents of the methallyl ligand and the olefin molecule
point into an opening between the Zr-bound ligands to
reduce nonbonding repulsion between substituents. The
propene ligand is more strongly bound to the metal
center in cation C-1 than in cation B-1 as indicated by
shorter Zr-C(5) and Zr-C(6) bond distances (302 and
268 pm). The energy of species C-1 is calculated to be
54 kJ /mol above that of cation B-1. This is close to the
experimentally obtained activation energy for the ex-
change of syn- and anti-methylene protons of the allyl
ligand, which involves a hapticity change similar to that
described by the structural rearrangement from B-1 to
C-1.
The energy profile obtained from optimized geom-
etries along the reaction path for the insertion of a
propene ligand into the Zr-C(1) σ-bond is presented in
Figure 3. Starting from cation C-1, a decrease of the
Zr-C(1) bond distance leads over a plateau at 280 pm
to transition state C-2 at ca. 220 pm. The energy of this
transition state C-2 is calculated to be 32 kJ /mol above
that of cation C-1, i.e., 86 kJ /mol above that of cation
B-1. The geometry of transition state C-2 shows a strong
R-agostic interaction29,42 with a Zr-H bond distance of
C(3)-C(5)
C(1)-C(2)
Zr-CEa
CE-Zr-CEa
a
CE ) C5-ring centroid.
Ta ble 4. Bon d Dista n ces a n d An gles for
Geom etr ies C-1 to C-3
C-1
C-2
331
C-3
296
Zr-C(1)
223
334
Zr-C(2)
Zr-C(3)
Zr-C(5)
302
268
346
134
219/220
130
263
233
226
140
219/218
134
Zr-C(6)
221
146
154
217/217
135
C(1)-C(5)
C(1)-C(2)
Zr-CEa
CE-Zr-CEa
a
CE ) C5-ring centroid.
C(3)-C(5) leads to the final product B-3 at a C(3)-C(5)
bond distance of 156 pm. Geometry optimization of this
cation B-3 reveals the parameters listed in Table 3. The
former propene ligand is now σ-bonded with its CH2
terminus to the Zr center (d(Zr-C(6)) ) 236 pm). The
carbon atoms C(1) and C(2) of the allyl ligand form an
olefin terminus (d(C(1)-C(2)) ) 138 pm) and stabilize
the unsaturated Zr center as shown by bond distances
of d(Zr-C(1)) ) 263 pm and d(Zr-C(2)) ) 303 pm
(Figure 3). The activation barrier for the process B-1 to
B-3 is calculated to be 61 kJ /mol; it is thus ca. 35 kJ /

Zirconocene Allyl Complexes
Organometallics, Vol. 19, No. 4, 2000 385
214 pm and a C(1)-H bond distance of 114 pm and is
comparable to the geometries of transition states cal-
culated for the insertion of an olefin into a Zr-alkyl
σ-bond reported in the literature.37-42 Further decrease
of the C(1)-C(5) bond distance leads to final state C-3,
the geometry of which shows a strong γ-agostic interac-
tion. The energy of final state C-3 is calculated to be 31
kJ /mol below that of initial state C-1. This final state
C-3 can rearrangeswithout any significant activation
barriersto final state B-3 by replacing the γ-agostic
interaction by coordination of the olefin carbon atoms
(C(2),C(3)) to the metal center.
Overall activation barriers of +61 kJ /mol for insertion
of a propene ligand into a Zr-η3-methallyl bond repre-
sented by the sequence B-1 f B-2 f B-3 and of +86
kJ /mol for insertion into a Zr-η1-methallyl bond rep-
resented by the sequence B-1 f C-1 f C-2 f C-3 show
that insertion into a Zr-allyl bond is possible and is
more likely to take place into a Zr-allyl π-bond than
into a Zr-allyl σ-bond. Rearrangement of the allyl group
from initial state B-1 to C-1 is endothermic by 54 kJ /
mol and thus disfavors insertion of a propene ligand into
a Zr-methallyl σ-bond. Similar insertion mechanisms
of olefins into π-allyl ligands were previously reported
for insertion of butadiene into a Ni-allyl bond by
Taube44 and for ethylene insertion into Pd-allyl and
Zr-allyl bonds.12,45
of activity caused by addition of small amounts of
dihydrogen to zirconocene-based polymerization sys-
tems.46 In any case, formation of zirconocene-allyl
species through one of the pathways represented in
Schemes 1 and 2 is likely to cause serious reductions
in catalyst acitivities by tying up substantial fractions
of the catalyst.
Com p u ta tion a l Meth od s
For geometry optimizations of the complexes studied,
we used the gradient-corrected nonlocal density func-
tional according to Becke47 and Perdew48 (BP). For all
calculations the program DGAUSS (4.1)49 with the
medium integration grid option was used on a CRAY
J 916/6-1024. For the Zr atom we used an effective core
potential base set with double-ú quality in the valence
region.49 The DZVP21 split valence basis set was used
for all C and H atoms; for the J ij Coulombic term50
the A149 auxiliary basis was used for all atoms. All
stationary points were checked by frequency calcula-
tions for the absence of negative eigenvalues and the
presence of only one negative eigenvalue for each
transition state.
Exp er im en ta l Section
All reactions were performed under argon using Schlenk-
line techniques or under nitrogen in a glovebox. Solvents were
dried prior to use by refluxing over and distilling from sodium.
Deuterated solvents were dried over 4 Å molecular sieves.
(C5H5)2ZrClMe,19 (C5H5)2ZrMe2,50 (C5H5)2Zr(C4H7)2,20 (CH2C-
(CH3)CD2)MgBr,52 B(C6F5)3,53 Me2Si(2-Me-4-tBu-C5H2)2ZrMe2,54
Me2Si(2-Me-indenyl)2ZrMe2,54 C2Me4(C5H4)2ZrMe2,54 Me2Si-
Con clu sion
In conclusion, our results show that zirconocene-allyl
species arising during olefin polymerization catalysis
can be reactivated either by exchange with an alumi-
num alkyl cocatalyst orsin the presence of a borane
activatorsby olefin insertion into the Zr-allyl bond.
Allyl-alkyl exchange with trimethyl aluminum or with
MAO occurs, at moderate Al:Zr ratios, with rates on the
order of ca. 0.1-1 s-1, which are ca. 10-100 times
slower than typical olefin insertions but comparable to
the rates of chain transfer, e.g., to aluminum alkyls.
Formation of transmetalation products with Al-bound
methallyl end groups during polymerization catalysis
would escape recognition since the final hydrolysis
would convert them to normal 2-propenyl end groups.
Olefin insertion into a Zr-allyl species, which is
recognizable by the appearance of internal unsaturation
within the polymer chain,14,15 must also be slower than
olefin insertion into normal Zr-alkyl species by at least
one order of magnitude, as shown by the observation
that only a fraction of the borane-activated complex 2
reacts with added propene, which is then used up
through multiple insertions into the ensuing Zr-oligo-
mer units. DFT calculations illustrate that reactivation
of a Zr-bound allyl ligand through insertion of a propene
molecule requires a relatively high activation energy.
Reactivation by insertion of propene is thus likely to be
slower than that by reaction with dihydrogen, which
regenerates a highly reactive zirconocene-olefin-hy-
dride complex with an activation barrier of only ca. 20
kJ /mol. This provides an explanation for the increase
(C5H4)2ZrMe2,54 and Ph2C(C5H4)(fluorenyl)ZrMe2 were syn-
54
thesized according to literature procedures. A 25% solution of
MAO in toluene was donated by WITCO GmbH. [PhNMe2H]-
[B(C6F5)4] and [Ph3C][B(C6F5)4] were obtained from BASF AG.
All other chemicals were purchased from commercial suppliers
and used without further purification.
1H NMR spectra were recorded on Bruker AC 250 (250
MHz) or Bruker Avance 600 (600 MHz) instruments. 13C NMR
and 2H NMR spectra were recorded on a Bruker Avance 600
(600 MHz) instrument. 19F NMR spectra were recorded on a
Bruker 400 (400 MHz) instrument. Polymer spectra were
recorded on a Bruker AC 250 (250 MHz).
Syn th esis a n d Ch a r a cter iza tion of Com p lexes
(C5H5)2Zr Me(C4H7) (1). To a solution of (C5H5)2ZrMeCl (0.4
g, 2.3 mmol) in 15 mL of THF at 0 °C was added 5 mL of a
solution of (C4H7)MgCl in THF (c ) 0.5 mol/L, 2.5 mmol). The
mixture was allowed to warm to room temperature. After
about 1 h the color of the solution had turned from pale yellow
to orange. After stirring for another 2 h the solvent was
(46) Carvill, A.; Tritto, I.; Locatelli, P.; Sacchi, M. C. Macromolecules
1997, 30, 7056.
(47) Becke, A. J . Chem. Phys. 1986, 84, 4524. Becke, A. J . Chem.
Phys. 1988, 88, 1053. Becke, A. Phys. Rev. 1988, A38, 3098.
(48) Perdew, J . P. Phys. Rev. 1986, B33, 8822. Perdew, J . P. Phys.
Rev. 1986, B34, 7406.
(49) Andzelm, J . W. Density Functional Methods in Chemistry;
Labanowski, J . K., Andzelm, J . W., Eds.; Springer-Verlag: New York,
1991; p 155.
(50) Dunlap, B. I.; Connolly, J . W. D.; Sabin, J . R. J . Chem. Phys.
1979, 71, 3396.
(51) Samuel, E.; Rausch, M. D. J . Am. Chem. Soc. 1973, 95, 6263.
(52) Hill, E. A.; Boyd, W. A.; Desai, H.; Darki, A.; Bivens, L. J .
Organomet. Chem. 1996, 541, 1.
(53) Pohlmann, J . L. W.; Brinckman, F. E.; Tesi, G.; Donadio, R. E.
Z. Naturforsch. 1965, 20b, 1. Pohlmann, J . L. W.; Brinckman, F. E. Z.
Naturforsch. 1965, 20b, 5.
(43) Brookhart, M.; Green, M. L. H.; Wong, L.-L. Prog. Inorg. Chem.
1988, 36, 2.
(44) Tobisch, S.; Bogel, H.; Taube, R. Organometallics 1996, 15,
3563. Tobisch, S.; Bogel, H.; Taube, R. Organometallics 1998, 17, 1177.
(45) DiRenzo, G. M.; White, P. S.; Brookhart, M. J . Am. Chem. Soc.
1996, 118, 6225.
(54) Beck, S.; Prosenc, M. H.; Brintzinger, H. H.; Goretzki, R.;
Herfert, N.; Fink, G. J . Mol. Catal. A: Chem. 1996, 111, 67.

386 Organometallics, Vol. 19, No. 4, 2000
Lieber et al.
evaporated and the residue extracted with pentane. The
pentane solution was reduced in volume in vacuo and cooled
to -60 °C to yield 350 mg (53%) of (C5H5)2ZrMe(C4H7) (1) in
the form of yellow needles.
1H NMR (600 MHz, C6D6, 300 K): δ 5.29 (s, 10H, C5H5),
2.54 (s, 4H, CH2(CH3)CCH2), 1.48 (s, 4H, CH2(CH3)CCH2),
-0.25 (br, 3H, ZrCH3). 13C NMR (62.5 MHz, C6D6, 300 K): δ
108.8 (C5H5), 67.92 (CH2(CH3)CCH2), 30.2 (ZrCH3), 27.2 (CH2-
(CH3)CCH2).
bromobenzene gave signals assigned to complex [(C5H5)2Zr-
(C4H5D2)][MeB(C6F5)3] (4).
1H NMR (600 MHz, C6D5Br, 300 K): δ 5.47 (s, 10H, C5H5),
3.2 (s, 1H, CH2(CH3)CCD2), 2.51 (s, 1H, CH2(CH3)CCD2), 1.68
(s, 3H, CH2(CH3)CCD2), 1.21 (br, 3H, CH3B(C6F5)3). 19F NMR
(376 MHz, C6D5Br, 300 K): δ -132 (d, J F-F 22 Hz, o-F), -161.5
(t, J F-F 22 Hz, p-F), -166.1 (m, J F-F 22 Hz, m-F). 2H NMR
(61.4 MHz, C6H5Cl, 300 K): 2.55 (C4H5D2).
Rea ction of (C5H5)2Zr Me(C4H7) (1) w ith N,N-Dim eth -
yla n ilin iu m P er flu or otetr a p h en yl Bor a te. Reaction of
equimolar amounts of (C5H5)2ZrMe(C4H7) (c ) 2.5 mmol) and
N,N-dimethylanilinium tetrakis(pentafluorophenyl)borate in
bromobenzene at room temperature gave 1H NMR signals that
were assigned to (C5H5)2Zr(C4H7)+, (C5H5)2ZrMe+, N,N-dimeth-
ylaniline, isobutene, and methane in the following manner.
1H NMR (250 MHz, C6D6, 300 K): 6.4 (t, J H-H 7.5 Hz, m-H),
6.1 (t, J H-H 7.5 Hz, p-H), 5.9 (d, J H-H 7.5 Hz, o-H), 1.8
(coordinated aniline), 4.83, 4.75 (C5H5)2Zr(C4H7)+(N,N-dim-
ethylaniline), 2.41 (C5H5)2Zr(CH2C(CH3)CH2)+(N,N-dimethy-
laniline), 1.0 (C5H5)2Zr(CH2C(CH3)CH2)+(N,N-dimethylaniline),
4.73 (CH2C(CH3)2), 1.58 (CH2C(CH3)2), 0.14 CH4.
Rea ction of (C5H5)2Zr (C4H7)2 w ith N,N-Dim eth yla n i-
lin iu m P er flu or otetr a p h en yl Bor a te. Upon addition of 250
µL of borate solution (c ) 10 mmol/L) to 250 µL of a solution
of (C5H5)2Zr(C4H7)2 (c ) 10 mmol/L) in bromobenzene, an
instant color change from orange to pale yellow was observed,
and 1H NMR signals for isobutene, (C5H5)2Zr(C4H7)+(N,N-
dimethylaniline), and free N,N-dimethylaniline (DMA) were
detected in the reaction mixture.
1H NMR (250 MHz, C6D6, 300 K): 6.4 (t, J H-H 7.5 Hz, m-H),
6.1 (t, J H-H 7.5 Hz, p-H), 5.9 (d, J H-H 7.5 Hz, o-H), 1.8
(coordinated DMA), 4.83, 4.75 (C5H5)2Zr(C4H7)+(DMA), 2.41
(C5H5)2Zr(CH2C(CH3)CH2)+(DMA), 1.0 (C5H5)2Zr(CH2C(CH3)-
CH2)+ (DMA).
Rea ction of (C5H5)2Zr (C4H7)2 w ith Tr ityl P er flu or otet-
r a p h en yl Bor a te. Reaction of equimolar amounts of (C5H5)2-
Zr(C4H7)2 and trityl tetrakis(pentafluorophenyl)borate in bro-
mobenzene at 25 °C resulted in the appearance of several new
signals in the (C5H5) region, isobutene, and a green discolora-
tion, which indicates decomposition of the starting material.
Variation of solvent, concentration, or temperature did not
influence the reaction.
1H NMR (600 MHz, C6D5Br, 300 K): δ 5.47 (s, 10H, C5H5),
2.51 (s, 2H, CH2(CH3)CCH2), 3.2 (s, 2H, CH2(CH3)CCH2), 1.68
(s, 3H, CH2(CH3)CCH2), 1.21 (br, 3H, ZrCH3).
Anal. Found (Calcd): C, 61.56 (61.79); H, 7.18 (6.91).
(C5H5)2Zr Me(C4H5D2) (3). To a solution of 200 mg (0.74
mmol) of (C5H5)2ZrMeCl in 10 mL of THF at 0 °C was added
3 mL of a solution of (CH2C(CH3)CD2)MgBr in THF (c ) 0.25
mol/L). The mixture was allowed to warm to room tempera-
ture, during which the pale yellow solution turned orange.
After stirring for 2 h the solvent was exchanged for pentane
and the residue extracted three times. Partial removal of the
solvent and storing at -60 °C afforded (C5H5)2ZrMe(C4H5D2)
(3) in the form of small orange needles. Yield: 150 mg (0.51
mmol, 70%).
2H NMR (61.4 MHz, C6H6, 300 K): 2.42 (s, 2D). 1H NMR
(600 MHz, C6D6, 300 K): δ 5.29 (s, 10H, C5H5), 2.54 (s, 2H,
CH2(CH3)CCD2), 1.48 (s, 3H, CH2(CH3)CCH2), -0.25 (br, 3H,
ZrCH3).
R ea ct ion of (C5H 5)2Zr Me(C4H 7) (1) w it h Tr im et h yl-
a lu m in u m . Addition of 200 µL of a TMA solution (c ) 1 mol/
L) in C6D6 to 200 µL of complex 1 (c ) 0.1 mol/L) resulted in
an almost complete disappearance of the yellow color. New
NMR signals that appeared in the reaction mixture in addition
to those of TMA and 1 were assigned to (C5H5)2ZrMe2 and
(C4H7)Me5Al2. Ratios of (C5H5)2ZrMe(C4H7) to (C5H5)2ZrMe2 as
given in Table 1 were determined from the respective integrals
of the C5H5 signals.
1H NMR (600 MHz, C6D6, 300 K): δ_5.70 (s, 10H, C5H5-
ZrMe2), 5.29 (s, 10H, C5H5ZrMe(C4H7)), 2.92 (br, 4H, CH2-
(CH3)CCH2AlMe2), 2.54 (s, 4H, CH2(CH3)CCH2), 1.77(s, 3H,
CH2(CH3)CCH2AlMe2), 1.48 (s, 3H, CCH2(CH3)CCH2), -0.15
(s, 6H, Cp2Zr(CH3)2), -0.25 (br, 3H, Cp2ZrCH3(CH2(CH3)-
CCH2)), -0.42 (bbr, AlMe3 and CH2(CH3)CCH2AlMe2). 13C
NMR (150 MHz, C6D6, 300 K): δ 110.3 (C5H5ZrMe2), 108.8
(C5H5ZrMe(C4H7)), 51.5 (CH2(CH3)CCH2AlMe2), 67.9 (CH2-
(CH3)CCH2), 28.2 (CH2(CH3)CCH2AlMe2), 27.2 (CH2(CH3)-
CCH2), 30.2 (Cp2ZrCH3), -7.2 (AlMe3 and CH2(CH3)CCH2-
AlMe2).
R ea ct ion of [(C5H 5)2Zr (C4H 7)][MeB(C6F 5)3] (2) w it h
Olefin s. An NMR tube containing 0.5 mL of a 1 mM solution
of 2 in toluene, prepared as described above, was cooled to -60
°C. At this temperature 1 equiv of olefin was added via a
syringe. The resulting product was immediately measured
1
starting from -60 °C to room temperature. H NMR signals
Rea ction of (C5H5)2Zr Me(C4H7) (1) w ith B(C6F 5)3. Reac-
tion of 300 µL of a 0.1 mol/L solution of 1 and 300 µL of a
0.11 mol/L solution of B(C6F5)3 in C6D6 in an NMR tube at
room temperature gave NMR signals that were assigned to
complex [(C5H5)2Zr(C4H7)][MeB(C6F5)3] (2) as outlined be-
low. Formation of 2 was accompanied by precipitation of a red
oil, which we were not able to isolate without decomposition.
No precipitate occurred when orange complex 2 (0.05 mol/L)
was generated by an analogous reaction in bromobenzene
solution.
1H NMR (600 MHz, C6D6, 300 K): δ 5.25 (s, 10H, C5H5),
2.45 (s, 4H, CH2(CH3)CCH2), 1.4 (s, 3H, CH2(CH3)CCH2), -0.3
(br, 3H, CH3B(C6F5)3). 19F NMR (376 MHz, C6D6, 300 K): δ
132.1 (d, J F-F 22 Hz, o-F), 163.8 (t, J F-F 22 Hz, p-F), 166.4 (m,
J F-F 22 Hz, m-F).
1H NMR (600 MHz, C6D5Br, 300 K): δ 5.47 (s, 10H, C5H5),
3.2 (s, 2H, CH2(CH3)CCH2), 2.51 (s, 2H, CH2(CH3)CCH2),
1.68 (s, 3H, CH2(CH3)CCH2), 1.21 (br, 3H, CH3B(C6F5)3). 19F
NMR (376 MHz, C6D5Br, 300 K): δ -132.0 (d, J F-F 22 Hz,
o-F), -161.5 (t, J F-F 22 Hz, p-F), -166.1 (m, J F-F 22 Hz,
m-F).
at 4.7-4.5 and 2.0-0.9 ppm in a ratio of approximately 1:15,
assigned to short oligomers containing 4-5 monomer units,
and the signals of the unreacted complex 2 were detected.
Addition of a 50-fold excess of propene at room temperature
gave oligomeric chains, with 1H NMR signal integrals of the
olefinic end group and those for the remainder of the polymer
chain in a ratio of 1:80.
P olym er iza tion of P r op en e a n d Eth en e. A Schlenk
vessel was charged with toluene (50 mL) and triisobutylalu-
minum (0.5 mol). The mixture was thermostated at 40 °C and
saturated with the monomer to 1 bar. To start the polymeri-
zation, 5 mL of a 1 mM solution of 2 in toluene was added.
The reaction was stopped after 30 min by adding an acidic
methanol solution. The resulting solution was washed with
NaHCO3 solution and water, and the organic layer was
separated and dried with Na2SO4. Evaporation of the solvent
in vacuo gave an oil, which is soluble in C2D2Cl4 at room
temperature.
Rea ction of (C5H5)2Zr Me(C4H5D2) (3) w ith B(C6F 5)3 a n d
P r op en e. In an NMR tube 250 µL of a solution of 3 (0.025
mmol) in chlorobenzene was combined with 250 µL of a
solution of tris(pentafluorophenyl)borane (0.03 mmol) at room
temperature. To the resulting ion pair [(C5H5)2Zr(C4H5D2)]-
Rea ction of (C5H5)2Zr Me(C4H5D2) (3) w ith B(C6F 5)3. An
analogous reaction of (C5H5)2ZrMe(C4H5D2) with B(C6F5)3 in

Zirconocene Allyl Complexes
Organometallics, Vol. 19, No. 4, 2000 387
2
[MeB(C6F5)3] (4) was added 10 equiv of propene, and H NMR
spectra were recorded (cf. Figure 1). In another experiment,
10-12 equiv of propene was added to the ion pair 4; after 5-10
s the reaction mixture was quenched by adding 5 µL of H2O,
and 13C and 2H NMR spectra were recorded. In this case the
ratio of terminal D2CdC to the sum of the other D signals was
1:20.
gratefully acknowledged. We thank Witco GmbH for the
donation of MAO solutions. Financial support of this
work was provided by BASF AG, Bundesministerium
fu¨r Forschung und Technologie, and by funds of the
University of Konstanz.
Su p p or tin g In for m a tion Ava ila ble: Equations used for
the determination of exchange equilibrium constant Kex and
of rate constants kex for methyl/methallyl exchange and data
on the temperature dependence of kex are available free of
charge via the Internet at http://pubs.acs.org.
Ack n ow led gm en t. Helpful discussions with Profes-
sor Ulrich Steiner, University of Konstanz, the recording
of 2H and two-dimensional spectra by Dr. Armin Geyer
and Anke Friemel, and allocation of computing time by
the Computing Center of the University of Konstanz are
OM9906058