Zirconocene complexes with cyclopenta[l]phenanthrene ligands: syntheses, structural dynamics, and properties as olefin polymerization catalysts

Article abstract of DOI:10.1021/om000149j

(2-Methylcyclopenta[l]phenanthryl)₂ZrCl₂ and (2-phenylcyclopenta[l]phenanthryl)₂ZrCl₂ were synthesized and used, after activation with MAO, as catalysts for the polymerization of propene. The resulting polymers

Full text of DOI:10.1021/om000149j

Organometallics 2000, 19, 3597-3604
3597
Zir con ocen e Com p lexes w ith Cyclop en ta [l]p h en a n th r en e
Liga n d s: Syn th eses, Str u ctu r a l Dyn a m ics, a n d
P r op er ties a s Olefin P olym er iza tion Ca ta lysts
Nicole Schneider, Frank Schaper, Katrin Schmidt, Robin Kirsten,
Armin Geyer,* and Hans H. Brintzinger*
Fachbereich Chemie, Universita¨t Konstanz, Postfach 5560, D-78464 Konstanz, Germany
Received February 15, 2000
(2-Methylcyclopenta[l]phenanthryl)2ZrCl2 and (2-phenylcyclopenta[l]phenanthryl)2ZrCl2
were synthesized and used, after activation with MAO, as catalysts for the polymerization
of propene. The resulting polymers had low isotacticities (8-18% [mmmm]). Ligand rotation
barriers for the corresponding Zr-dibenzyl derivatives were determined by dynamic NMR
spectroscopy. Pathways and transition states for ligand rotation in this series of complexes
were studied by molecular-mechanics calculations and are discussed in relation to the
microstructure of the respective polymer products.
Sch em e 1
In tr od u ction
Thermoplastic elastomers with an isotactic-atactic
block structure were reported in 1995 by Hauptmann
and Waymouth to arise from the polymerization of
propene with (2-phenylindenyl)2ZrCl2 (1)/MAO.1 This
zirconocene catalyst is thought to interconvert between
two isomers with different stereospecificities by rotation
of the ligands, the rac-like chiral isomer producing
isotactic, crystalline blocks and the achiral meso-like
isomer being responsible for atactic, amorphous blocks
(Scheme 1). Under these premises, a nonrandom block
structure with distinct isotactic and atactic blocks will
form provided that the rate of olefin insertion is greater
than that of the mutual rotation of the ligand moieties.
A more recent comparison of polymers generated by the
unbridged complex 1 with reactor blends obtained with
mixtures of the rac- and meso-isomers of the corre-
sponding dimethylsilyl-bridged complex supports this
view.2
catalysts to characterize the factors that control the
formation of thermoplastic elastomeric polyolefin ma-
terials.
Since annelation of the indenyl ligands in bridged
zirconocene complexes with additional aromatic rings
was found to enhance catalyst activities and polymer
isotacticities,3,4 we have now prepared complexes similar
to 1 with further rings annelated to the indenyl ligands
and studied their properties as propene polymerization
Resu lts
P r epar ation an d Ch ar acter ization of Cyclopen ta-
[l]p h en a n th r en e Zir con iu m Com p lexes. Complexes
2 and 3 were synthesized by deprotonation of the
respective ligand precursors, 4 and 5,4 with n-butyl-
lithium and subsequent reaction with ZrCl4 in toluene
1
(1) (a) Coates, G. W.; Waymouth, R. M. Science 1995, 267, 217. (b)
Coates, G. W. Dissertation, Stanford University, 1994. (c) Hauptmann,
E.; Waymouth, R. M. J . Am. Chem. Soc. 1995, 117, 11586. (d)
Kravchenko, R.; Masood, A.; Waymouth, R. M. Organo-
metallics 1997, 16, 3635. (e) Maciejewski Petoff, J . L.; Bruce, M. D.;
Waymouth, R. M.; Masood, A.; Lal, T. K.; Quan, R. W.; Behrend, S. J .
Organometallics 1997, 16, 5909. (f) Tagge, C. D.; Kravchenko, R. L.;
Lal, T. K.; Waymouth, R. M. Organometallics 1999, 18, 380.
(2) Maciejewski Petoff, J . L.; Agoston, T.; Lal, T. K.; Waymouth, R.
M. J . Am. Chem. Soc. 1998, 120, 11316.
(3) (a) Spaleck, W.; Kuber, F.; Winter, A.; Rohrmann, J .; Bachmann,
B.; Antberg, M.; Dolle, V.; Paulus, E. F. Organometallics 1994, 13, 954.
(b) Stehling, U.; Diebold, J .; Kirsten, R.; Ro¨ll, W.; Brintzinger, H. H.;
J u¨ngling, S.; Mu¨lhaupt, R.; Langhauser, F. Organometallics 1994, 13,
964.
according to Scheme 2. Their H NMR spectra in CDCl3
solution (see Experimental Section) are consistent with
an apparent C2v symmetry, which indicates that ligand
rotation is fast on the 1H NMR time scale at room
temperature.
Crystals suitable for an X-ray structure analysis were
obtained from a concentrated solution of 3 in methylene
chloride. Unlike catalyst 1, the unit cell of which
contains both a rac-like and a meso-like isomer,1c
zirconocene 3 crystallizes, in space group P1h, as its rac-
like isomer only, with the phenyl substituents oriented
away from the chloride ligands. The phenyl rings are
bent by 4-5° out of the plane of the adjacent cyclopenta-
(4) (a) Schneider, N.; Huttenloch, M. E.; Stehling, U.; Kirsten, R.;
Schaper, F.; Brintzinger, H. H. Organometallics 1997, 16, 3413. (b)
Schneider, N.; Prosenc, M. H.; Brintzinger, H. H. J . Organomet. Chem.
1997, 7213.
10.1021/om000149j CCC: $19.00 © 2000 American Chemical Society
Publication on Web 08/05/2000

3598 Organometallics, Vol. 19, No. 18, 2000
Schneider et al.
Sch em e 2
Ta ble 1. Selected Dista n ces (Å) a n d An gles (d eg)
for Com p lexes 3 a n d 3B
Ta ble 2. MAO-Activa ted P r op en e P olym er iza tion ,
[Zr ] ) 55 µM, Al:Zr ) 1000:1
complex 3
complex 3B
activity
[105 gPP/
%
Zr-Cl(1)
Zr-Cl(2)
Zr-Cpc
2.410(2)
2.415(2)
2.241, 2.234
2.475(7)-2.598(6) Zr-CCp
Zr-C(47)
Zr-C(54)
Zr- Cpc
2.322(7)
2.263(7)
2.268, 2.259
2.500(7)-2.644(6)
catalyst T [°C] p [bar] (mol Zr‚h)] TOF [s-1
]
Mw
mmmm
a
1a
1
2
0
0
0
0
1.7
2
2
6.2
6.8
10.7
0.2
4.1
4.5
7.1
0.13
395 000
580 000
802 800
94 300
13.2
22
8
Zr-CCp
Cl(1)-Zr-Cl(2) 97.04(8)
Cpc,1-Zr-Cpc,2 131.9
C(47)-Zr-C(54) 95.5(3)
Cpc,1-Zr-Cpc,2
Cpc-Zr-C(47)
Cpc-Zr-C(54)
131.6
106.8, 101.7
107.5, 107.7
3
2
18
Cpc-Zr-Cl(1)
Cpc-Zr-Cl(2)
105.6, 105.6
104.8, 106.7
a
See ref 1.
a
Cpc is the centroid of the cyclopentadienyl ring.
F igu r e 2. Crystal structure of complex 3B (thermal
parameters are drawn at 50% probability; hydrogen atoms
are omitted for clarity).
with Zr-CH2-C6H5 angles of 126-133°, similar to
those reported for other benzyl zirconocene complexes
(Figure 2).5 In comparison with 3, the benzyl groups
appear to be more sterically demanding than chloride
ligands; they cause an elongation of the Zr-CCp dis-
tances and a more pronounced bending of the phen-
anthrene framework by 5-10° out of the cyclopenta-
dienyl plane (Table 1).
P r op en e P olym er iza tion s. Complexes 2 and 3 were
studied with regard to their properties as catalysts for
the polymerization of propene. The reaction systems
contained the respective complexes in a concentration
of 55 µM, together with MAO (Al:Zr ) 1000:1) in toluene
solution at 0 °C under a constant propene pressure of
2 bar (Table 2).
F igu r e 1. Crystal structure of complex 3 (thermal pa-
rameters are drawn at 50% probability; hydrogen atoms
are omitted for clarity).
dienyl rings; they are rotated out of this plane by 6-20°.
The phenanthrene moiety is likewise bent by 3-5° out
of the cyclopentadienyl plane (Figure 1, Table 1).
The corresponding dibenzyl complexes 2B and 3B
were obtained by treating the zirconocene dichloride
complexes 2 and 3 with benzylmagnesium chloride in
toluene. Crystals of 3B suitable for X-ray structural
analysis were obtained by slow evaporation of a benzene
solution.
3B crystallized, as yellow prisms, in the chiral
(5) J ordan, R. F.; LaPointe, R. E.; Baenzinger, N.; Hinch, G. H.
Organometallics 1990, 9, 1539.
space group P212121. The benzyl ligands are η1-bonded

Zirconocene Complexes as Polymerization Catalysts
Organometallics, Vol. 19, No. 18, 2000 3599
For comparison, complex 1 was resynthesized and
used under the same conditions. The activity of 3/MAO
was significantly less than that of 1/MAO. While the
molecular weight of the polymer obtained was also far
lower, the isotacticity was about the same as for 1/MAO.
Nevertheless, no melting transition was detected for the
polymer obtained with 3/MAO.6 Complex 2/MAO, on the
other hand, was more active than 1/MAO and generated
a polymer with much higher molecular weight. In this
case, however, the polymer was nearly atactic and thus
did not give rise to any melting transition either.
Possible reasons for the loss of thermoplastic proper-
ties of the polymer obtained with 2 and 3 could be either
a lower rate of propene insertion or a faster rotational
isomerization in comparison with 1. Both cases would
lead to a diminished isotactic block length and, hence,
to the absence of a melting transition. To ascertain to
which degree the polymer microstructure depends on
variations in the rates of ligand rotation, the latter were
investigated by variable-temperature NMR studies.
Dyn a m ic NMR In vestiga tion s on Liga n d Rota -
tion Ra tes. Room-temperature 1H NMR spectra of 2
and 3 in CDCl3 solution indicate an apparent C2v
symmetry, i.e., fast ligand rotations. Some signal broad-
ening from conformational exchange was detected for
complex 3, but even at a temperature of -75 °C the
internal rotation was not frozen out. Apparently, the
rotational activation barrier in 3 is too small for line-
shape analysis or EXSY spectroscopy to be feasible.
Instead, temperature-dependent NOE measurements
were used to determine the rotational barrier. Cross
relaxation rates detected in NOESY spectra depend both
on the proton-proton distances and on the rate of
molecular tumbling. In the temperature region of sign
inversion of the NOE, tumbling rates can be quantified
and internal energy barriers can be estimated.7 By
determining NOE intensities for 3 at seven tempera-
tures between 198 and 300 K, the activation energy of
the internal rotation was estimated to be EA ) 4.5 ( 2
kcal/mol in CDCl3 solution.8
F igu r e 3. Variable-temperature 1H NMR spectra of
complex 3B.
Ta ble 3. Liga n d Rota tion Ra tes a n d Activa tion
Ba r r ier s for Com p lexes 1B-3B
1Ba
2B
3B
T [K] k [s-1
]
∆Gq T [K] k [s-1
]
∆Gq T [K] k [s-1
]
∆Gq
203.1
213.7
222.3
232.8
243.4
255.0
264.9
274.7 1600 12.0
285.6 3300 12.0
292.1 6700 12.0
13 10.7 191.5
18 11.1 198.5
32 11.3 208.0
75 11.5 220.0
150 11.7
50
100
400
900
9.6 222.5
9.6 231.5
9.6 245.5
9.8 256.0
267.0
7.5 12.0
20
12.1
12.1
12.0
12.1
12.2
12.2
12.2
90
300
750
400 11.8
775 11.9
281.0
306.5 12000
314.5 22000
2000
a
Taken from ref 9.
While the room-temperature NMR spectra of 2B and
3B in toluene-d8 indicate, again, an apparent C2v
symmetry, i.e., fast rotation of the ring ligands, two
signal sets were detected for the cyclopenta[l]phenan-
threne protons at ca. 200 K (Figure 3). The NMR spectra
at this temperature correspond to the C2 symmetry of
the rac-like X-ray structure of 3B, where the two halves
of each phenanthrene ring have different chemical
environments and the protons at each benzyl methylene
group are diastereotopic. Signal separations of 56 Hz
for H1, 124 Hz for H2, 49 Hz for H3, 669 Hz for H4, and
615 Hz for H5 (cf. Figure 3) at 210 K led to well-
separated individual coalescence temperatures. Total
band-shape analysis10 was performed for 3B over a
range of more than 100 K, and ∆Gq was found to be
12.1 ( 0.2 kcal/mol at 0 °C (Table 3). The temperature
dependence indicates a slightly negative ∆Sq. The
diastereotopic methylene protons of the benzyl groups
split into an AB spin system below 230 K. An isomer-
ization barrier of 12.0 ( 0.5 kcal/mol was found for the
coupled two-site exchange of these protons, in good
agreement with the isomerization barrier of the cyclo-
penta[l]phenanthrene protons. We can thus conclude
that the racemization process of 3B is associated with
an internal energy barrier of 12.1 kcal/mol at 0 °C,
which is quite close to that of 12.0 kcal/mol reported
for complex 1B.9
Since a substantially higher isomerization barrier of
ca. 12 kcal/mol had been reported for the dibenzyl
derivative of 1 (1B) by NMR line-shape analysis at
temperatures of -72 to 20 °C,9 we have studied isomer-
ization rates also of the dibenzyl derivatives 2B and 3B
by NMR techniques.
(6) Mansel, S.; Pe´rez, E.; Benavente, R.; Peren˜a, J . M.; Bello, A.;
Ro¨ll, W.; Kirsten, R.; Beck, S.; Brintzinger, H. H. Macromol. Chem.
Phys. 1999, 200, 1292.
(7) Neuhaus, D.; Williamson, M. The Nuclear Overhauser Effect in
Structural and Conformational Analysis; VCH Publishers: New York,
1989.
(8) Measurements and calculations were performed as described
in: Poppe, L.; v. Halbeek, H. J . Am. Chem. Soc. 1992, 114, 1092. Cross
relaxation rates (σ) were calculated from the relative intensity of cross
(Ic) and diagonal (Id) signals in NOESY spectra with the mixing time
tm: σ ) -Ic/(Idtm); according to: Macura, S.; Farmer, B. T., II; Brown,
L. R. J . Magn. Reson. 1986, 70, 493. Overall and internal motions are
assumed to be thermally induced: tc ) A exp(Ea/RT). The NOE H1
f
H2 was chosen as representative for the overall tumbling rate of 3.
The cross relaxation becomes positive above 215 K in toluene, and the
activation energy of the overall molecular tumbling was determined
to be 1.8 kcal/mol. The NOE H5 f H6 has contributions from the H5
and H6 protons of the same phenanthrene ring and additionally it is
influenced by the short H5 to H6 distance (2.3 Å, estimated from the
X-ray structure) between the two phenanthrene rings of 3. This cross
relaxation rate is also influenced by the internal motion, and it becomes
positive above 225 K. We estimated the correlation time for the internal
dynamics according to the spectral density function of Lipari and Szabo
(Lipari, G.; Szabo, A. J . Am. Chem. Soc. 1982, 104, 4546), which
assumes two independent tumbling rates.
(9) Bruce, R. W.; Coates, G. W.; Hauptmann, E.; Waymouth, R. M.;
Ziller, J . W. J . Am. Chem. Soc. 1997, 119, 11174.
(10) Sandstro¨m, J . Dynamic NMR Spectroscopy; Academic Press:
New York, 1982.

3600 Organometallics, Vol. 19, No. 18, 2000
Schneider et al.
The methyl-substituted complex 2B shows signal
separations of only 58 Hz for H4 and 152 Hz for H5. Line-
shape analysis, in the temperature range of 191-220
K (Table 3), gave ∆Gq ) 9.6 ( 0.3 kcal/mol. To cor-
roborate this estimate, we recorded an EXSY spectrum
at 191.5 K with a very short mixing time of 10 ms. Since
no assumptions about T2 relaxation are necessary,
EXSY spectroscopy is in general associated with small
experimental errors, but was not feasible for complex
3B due to signal overlap. For 2B, an exchange rate of
55 s-1 determined from the intensity of the cross signals
between the H5 signals11 corresponds, again, to an
isomerization barrier of ∆Gq ) 9.5 ( 0.1 kcal/mol at this
temperature. A significantly lowered rotational barrier
thus results when the 2-phenyl substituent is replaced
by a 2-methyl group. As for complex 1B,9 meso-like
isomers of 2B and 3B are not detectable by NMR. The
question thus arises whether the interconversion be-
tween the two enantiomeric rac-like forms proceeds by
way of a meso-like intermediate.
F or ce-F ield Ca lcu la tion s. To characterize the reac-
tion paths of these isomerizations, we have studied this
ligand rotation process by a molecular-mechanics in-
vestigation. For this purpose, the Amber94 force field
was used to model the organic parts of the complexes.12
For the coordination geometry around the metal center,
we used optimized force-field parameters based on those
published by Doman et al.13a and Ho¨weler et al.13b (see
Supporting Information). Comparison of calculated and
experimentally determined structures for several sub-
stituted bisindenyl zirconium dichloride and dibenzyl
complexes (Figure 4)14 documents close agreement with
crystallographically determined bond parameters and,
in particular, with Cp-Cp torsion angles.15
In a first step, we investigated the isomerization of
the dichloride complexes 2 and 3 and, for comparison,
also that of complex 1, for which molecular-mechanics
calculations of ligand rotation have been published by
Cavallo et al.16 and Pietsch et al.17 Ligand rotation was
modeled by changes of the dihedral angles Φ1 and Φ2
in steps of 5° (Figure 5), followed by relaxation of all
other geometry parameters. The resulting energy sur-
face is shown in Figure 6; relative energies of stationary
points are shown in Table 4.
F igu r e 4. Observed and calculated structures for substi-
tuted bis(indenyl)zirconium complexes.
The relative energies obtained for the rotamers of
complex 1, while slightly higher than those obtained in
the previous force-field studies, show the same energy
ordering and similar geometries. The energy surface
contains four distinct minima and resembles that
obtained by Cavallo et al.16 The minima B and C (see
Figure 7) correspond to the rac-like and meso-like
rotamers found in the crystal structure of complex 1.
F igu r e 5. Definition of torsion angles Φ and θ.
As noted before,16,17 the minimum of lowest energy (A)
corresponds to another rac-like rotamer with a smaller
torsion angle θ of about 120° between the two ligand
moieties. The fourth minimum (D), which represents a
rotamer with the phenyl substituent of one ligand lying
between the two chloride ligands, is rather insignificant
due to its high energy.
For the interconversion between enantiomeric rac-
isomers (A and A′) two pathways are possible. The first
of these involves rotation of one ligand with the anne-
lated ring passing over the two chloride ligands (transi-
tion state E). The other one corresponds to the rotation
(11) Perrin, C. L.; Dwyer, T. J . Chem. Rev. 1990, 90, 935-967.
(12) Cornell, W. D.; Cieplak P.; Bayly C. I.; Gould I. R.; Merz K. M.,
J r.; Ferguson D. M.; Spellmeyer D. C.; Fox T.; Caldwell J . W.; Kollman
P. A. J . Am. Chem. Soc. 1995, 117, 5179.
(13) (a) Doman, T. N.; Hollis, T. K.; Bosnich, B. J . Am. Chem. Soc.
1995, 117, 1352. (b) Ho¨weler, U.; Mohr, R.; Knickmeier, M.; Erker, G.
Organometallics 1994, 13, 2380.
(14) Complexes 6 and 7 have not been used in the optimization of
the force field and can thus serve as a check on its accuracy.
(15) Deviations between calculated and crystallographically deter-
mined Cp-Cp torsion angles arise when strongly electron-withdrawing
substituents, such as fluorine or trifluoromethyl, are present in the
aromatic rings.
(16) Cavallo, L.; Guerra, G.; Corradini, P. Gazz. Chim. Ital. 1996,
126, 463.
(17) Pietsch, M. A.; Rappe´, A. K. J . Am. Chem. Soc. 1996, 118, 10908.

Zirconocene Complexes as Polymerization Catalysts
Organometallics, Vol. 19, No. 18, 2000 3601
F igu r e 8. Energy profile for the racemization of complex
1 along the isomerization paths indicated in Figure 6.
Ta ble 5. Rela tive En er gies Ca lcu la ted for th e
Dich lor id e Com p lexes 1-3 a n d for th e Diben zyl
Com p lexes 1B-3B [k ca l/m ol]
F igu r e 6. Energy surface of ligand rotations of complex
1 with isoenergetic lines at a distance of 1 kcal/mol
(redundant data for Φ > 180° given for clarity), with
alternative isomerization paths via transition states E
(dashed line) and F (solid line).
rotamer
1
3
2
1B
3B
2B
A (rac)
B (rac)
C (meso)
D
E
F
0
0
0.8
0
0.4
0.4
5.5
3.7
0
0
2.3
0
2.0
1.1
6.2
8.6
2.7
1.6
3.6
7.3
5.5
2.8
1.6
3.2
7.8
4.9
4.3
1.6
5.0
7.7
8.8
2.7
1.5
2.5
8.3
11.2
conclude that the activation barrier of the rac-rac
isomerization determined by NMR spectroscopy is iden-
tical with the barrier for the rac-meso isomerization,
proposed to be crucial for the mechanism of the polym-
erization reaction.
In all three complexes, rotation via transition state
F is favored. For complex 3, the activation barrier of
4.9 kcal/mol, calculated for the rotation over transition
state F , is in good agreement with the value of 4.5 ( 2
kcal/mol, determined by dynamical NMR. Among
complexes 1-3, the highest activation barrier of 5.5
kcal/mol is observed for complex 1, while a value of only
3.7 kcal/mol is obtained for the methyl-substituted
complex 2.
F igu r e 7. Geometries of minima and transition states on
the rotational energy surface of complex 1.
Ta ble 4. Rela tive En er gies a n d Tor sion An glesa
Ca lcu la ted for Com p lex 1
relative energies [kcal/mol]
rotamer this work ref 16 ref 17 this work ref 16 ref 17
torsion angles θa [deg]
Determination of the energy surfaces for the dibenzyl
complexes 1B-3B is complicated by the rotational
freedom of the benzyl ligands, which results in ad-
ditional minima and transition states. The values in
Table 5 represent the minima of lowest energy and the
lowest activation barriers for the interconversion be-
tween rac- and meso-like isomers. As for the dichlorides
1-3, racemization proceeds also for the dibenzyl com-
plexes 1B-3B by way of a transition state located
between a rac- and a meso-like isomer.
A (rac)
B (rac)
C (meso)
D
E
F
0
0
0
128
179
32
102
84
≈105 118
≈180 n.m.
2.7
1.6
3.6
7.3
5.5
1.4
1.0
n.m.
<10
<8
n.m.
0.6
n.m.
5
≈40
n.m.
n.m.
n.m.
23
n.m.
47
n.m.
n.m.
119
a
Definition of torsion angle θ see Figure 5; n.m.: not mentioned
or not calculated.
with the substituent in 2-position passing over the
chloride ligands (transition state F ).18
In contrast with the dichloride complexes 1-3, rota-
tion over transition state E is favored however for the
benzyl derivatives 1B-3B. The activation barriers thus
obtained follow similar trends as the experimental
values determined by dynamic NMR (Table 6): The
same values are found for complexes 1B and 3B, while
the value for 2B is smaller by 2 kcal/mol. The difference
of 4 kcal/mol between the calculated values of ∆Hq and
the measured values of ∆Gq can be attributed to entropic
factors. In the transition state for this rotation the
phenyl ring of one of the benzyl ligands is sandwiched
between both indenyl ligands (Figure 9). For complex
1B9 as for 3B slightly negative values of ∆Sq, derived
Analogous calculations for complexes 2 and 3 yield
similar energy surfaces, with results as summarized in
Table 5. For the three complexes calculated, the isomer-
ization between the two enantiomeric rac-like isomers
is possible only with the meso-like form as an inter-
mediate. On these isomerization paths the highest
activation barrier is always the one between the rac-
and meso-like rotamers, which differ by less than 2 kcal/
mol in their relative energies (Figure 8). We can thus
(18) In contrast with a previous report,17 our calculations show no
evidence of π-stacking interactions during ligand rotation. This is in
line with a recent experimental study on hydrogenated analogues of
complex 1.1d

3602 Organometallics, Vol. 19, No. 18, 2000
Schneider et al.
Based on experimental data for the dibenzyl com-
plexes 1B and 3B, and on calculations for 1 and 3, 1B
and 3B, and 1C and 3C, we have to assume similar
ligand rotation barriers for the 2-phenyl-substituted
catalysts 1/MAO and 3/MAO. The observation that
polymers produced with 3/MAO appear to contains
unlike the stereoblock polymers produced with 1/MAOs
only short isotactic sequences can be explained by the
unexpectedly low rate of olefin insertion in 3/MAO,
which is indicated by an activity 1-2 orders of magni-
tude lower than that of 1/MAO. For the 2-methyl-
substituted catalyst 2/MAO, on the other hand, which
has an even higher activity than 1/MAO, the substan-
tially reduced ligand rotation barrier is likely to cause
the lack of crystallizable isotactic stereoblocks in its
polypropene products.
F igu r e 9. Transition state E for the ligand rotation of
complex 1B.
Ta ble 6. Com p a r ison of Obser ved a n d Ca lcu la ted
Activa tion Ba r r ier s [k ca l/m ol]
While the relative rates of ligand rotation and the
polymerization activities estimated for catalysts 1-3/
MAO are thus qualitatively in line with the microstruc-
tures of their respective polypropene products, a dis-
turbing but probably significant quantitative discrepancy
remains to be explained:9 The ligand rotation barrier
of ∆Gq ≈ 5 kcal/mol, experimentally determined for the
dichloride complex 3 and calculated for this complex as
well as for the catalyst models 1C and 3C, corresponds
to rates of ligand rotation of ca. 108 s-1. These rotational
frequencies, and even the reduced ligand rotation rates
of 1000-1500 s-1 found for the dibenzyl complexes 1B
and 3B, exceed the average insertion rate of 0.1-10
olefins/s in these catalyst systems by several orders
of magnitude. This is not in line with the mechanism
proposed for stereoblock formation with these catalysts.
1B
3B
2B
∆Hq (calculated)
∆Gq (observed)
8
12
8
12
6
10
from the experimental data, are in agreement with a
loss of rotational freedom in the transition state.
To decide which rotational transition state (E as
observed for the dibenzyl derivatives or F as for the
dichlorides) is favored by the active species in polym-
erization, the latter was modeled for 1 and 3 in a
manner similar to that proposed by Pietsch et al. for
complex 1,17 i.e., by using a 2-methyl-1-butyl ligand to
model the growing polymer chain, while keeping the
second coordination site occupied by a methyl group.
Due to the reduced symmetry of these complexes and
the high number of possible rotamers of the 2-methyl-
1-butyl ligand, several additional minima and transition
states were found on the energy surfaces. Interconver-
sion between all minima proceeds most easily by rota-
tion via transition state F . For the phenyl-substituted
indenyl and cyclopenta[l]phenanthrene complexes, 1C
and 3C, we find rotational barriers of approximately
4-6 kcal/mol. Apparently, the transition to a meso-like
rotation intermediate occurs more easily in the presence
of one relatively large (2-methyl-1-butyl) and one small
(methyl) ligand than in the presence of the two medium-
sized benzyl ligands. With regard to the height of the
rotation barrier as well as to their rotation via transition
state F , these model complexes thus resemble more
closely their dichloride precursors than the dibenzyl
derivatives.
As an explanation for this apparent discrepancy, we
propose that olefin insertion rates during actual chain
growth are much higher than the apparent turnover
frequencies given in Table 2, which are derived from
time-averaged activities: Consideration of ligand dis-
placement equilibria and kinetics indicates that each
Zr-polymeryl cation is converted from the preponder-
ant, presumably inactive contact-ion pair to a reactive
outer-sphere cation-anion pair only for minute fractions
of its total lifetime.20 During these short periods,
uninterrupted chain growth would have to occur with
an actual rate much higher than that derived from time-
averaged activities.
Control of stereoblock polymer formation by ligand
rotation would then require that the periods of active
chain growth of a Zr-polymeryl species are shorter than
its average lifetime by a factor of at least 103 and
possibly of as much as 108, such that the actual rate of
olefin insertion during these growth periods exceeds
that of ligand rotation. While these relations appear
conceivable in light of energy differences of roughly 8
kcal/mol associated with anion displacement from alkyl-
metallocene cations,21 substantial further studies, e.g.,
with regard to the dependence of isotactic block lengths
Discu ssion
Our molecular-mechanics calculations allow us to
delineate in detail the reaction paths for the racemiza-
tion of a number of unbridged zirconocene dichloride and
dibenzyl complexes; they reproduce, furthermore, the
rotational barriers derived from dynamic NMR methods
with reasonable accuracy. The results of these calcula-
tions thus support the assumption that, for each of the
complexes studied, the experimentally determined ra-
cemization barrier corresponds directly to that of the
conversion between its rac-like and meso-like isomers,
i.e., for the isomerization process thought to be respon-
sible for the switch in stereoselectivity typical for these
catalyst systems.19
(19) The assumption that the rac-like isomers of these complexes
are isoselective is supported by the observation that the rac isomer of
the Me2Si-bridged analogue of 2 produces highly isotactic polypropene.4a
(20) (a) Fink, G.; Zoller, W. Makromol. Chem. 1981, 182, 3265, and
references therein. (b) Reichert, K. H. Angew. Makromol. Chem. 1981,
94, 1. (c) Beck, S.; Prosenc, M.-H.; Brintzinger, H. H. J . Mol. Catal. A:
Chem. 1998, 128, 41.
(21) Chan, M. S. W.; Vanka, K.; Pye, C. C.; Ziegler, T. Organo-
metallics 1999, 18, 4624.

Zirconocene Complexes as Polymerization Catalysts
Organometallics, Vol. 19, No. 18, 2000 3603
Ta ble 7. Cr ysta llogr a p h ic a n d Exp er im en ta l Da ta a
for Com p lexes 3 a n d 3B
on monomer concentrations,22 are clearly required to
corroborate this view.
3
3B
Exp er im en ta l Section
formula
cryst color and form yellow needle
C46H30Cl2Zr‚CH2Cl2 C60H44Zr
yellow orthogonal
prism
Gen er a l Com m en ts. Moisture- and air-sensitive com-
pounds were handled under an argon atmosphere using
Schlenk or drybox techniques. THF and toluene were distilled
from sodium, pentane was distilled from Na-K-alloy, and CH2-
Cl2 was distilled from CaH2. ZrCl4 (Fluka) was sublimed prior
to use. Cyclopenta[l]phenanthrenes 4 and 5 were synthesized
as described earlier.4 Benzylmagnesium chloride was synthe-
sized according to the literature.23 1H and 13C NMR spectra
were obtained with a Bruker DRX 600 spectrometer at a
proton resonance frequency of 600.13 MHz. Chemical shifts
are referenced to the solvent, and temperatures are determined
from the signal separation in a MeOH sample. 1H spin systems
were identified from TOCSY and DQF-COSY spectra, and 13C
resonances were assigned from inverse 13C-1H shift correla-
tions (HMQC).24 Mixing times in NOESY (EXSY) spectra were
chosen short enough to prevent spin diffusion. Exchange-
broadened spectra were simulated with the program DNMR-
SIM.25 ∆Gq values were calculated from the two-site exchange
rates with the Eyring equation.26
Gen er a l Meth od for Syn th eses of Bis(cyclop en ta [l]-
p h en a n t h r en e)zir con iu m Dich lor id es. A solution of 2.2
mmol of the respective cyclopenta[l]phenanthrene in 10 mL
of THF was treated with 1.44 mL of a 1.6 M solution of
n-butyllithium in hexane at -78 °C. After stirring overnight
the reaction mixture was evaporated to dryness, washed with
30 mL of pentane, and treated with a suspension of 415 mg of
zirconium tetrachloride (1.09 mmol) in 50 mL of toluene. The
mixture was stirred for 3 h at 90 °C, cooled to room temper-
ature, stirred overnight, and filtered. The yellow residue was
washed with 20 mL of pentane and THF.
cryst syst
space group
a [Å]
b [Å]
c [Å]
triclinic
P1h
orthorhombic
P212121
12.801(4)
17.842(5)
18.223(6)
90
10.218(5)
10.993(6)
16.430(9)
100.16(3)
97.03(3)
93.71(3)
2; 1796(2)
0.1 × 0.3 × 0.4
243; 1.534
0.640, 844
adaptive ω;
1.9-24.0
5299
R [deg]
â [deg]
90
90
γ [deg]
Z; V [Å3]
4; 4162(2)
0.3 × 0.3 × 0.4
243; 1.366
0.306, 1776
adaptive ω;
2.3-27.0
10 021
cryst size [mm]
T [K]; dcalcd [g/cm3]
µ [mm-1], F(000)
scan mode; θ range
[deg]
no. of reflns collcd
no. of ind reflns
obsd reflns
(I > 2σ(I))
4810
3920
8970
5004
solutionb
direct methods
470; 0.988
0.0627, 0.1514
1.020
direct methods
550; 1.037
0.0660, 0.1244
0.829
no. of params; GOF
R(F), Rw(F2)c
largest diff peak
[e Å3]
Flack x-param
-0.10(8)
a
Conditions: Siemens P4 four-circle diffractometer, Mo KR
radiation, graphite monochromator. All non-hydrogen atoms were
refined anisotropically. Hydrogens were refined on the calculated
positions with fixed isotropic U, using riding model techniques.
b
SHELXS-86, SHELXL-93 (G. Sheldrick, University of Go¨ttingen,
1990 and 1993). c Weighting scheme: 3: w-1 ) σ2(Fo2) + (0.0697P)2
+ 12.2662P. 3B: w-1 ) σ2(Fo2) + (0.0467P)2 + 2.3585P, with P )
(Fo + 2Fc2)/3.
2
Bis(2-m et h ylcyclop en t a [l]p h en a n t h r en e)zir con iu m
Dich lor id e (2). The yield was 390 mg (58%). 1H NMR
(600 MHz, CDCl3): δ 8.53-8.49 (m, 4 H), 8.94-7.97 (m, 4 H),
7.61-7.58 (m, 8 H), 6.64 (s, 4 H), 1.53 (s, 3 H). MS (EI): m/z
620 (28%, M+), 391 (100%, M+ - 4), 230 (85%, 4-H), 215 (28%,
4 - Me). Anal. Calcd for C36H26ZrCl2: C, 69.66; H, 4.22.
Found: C, 69.50; H, 4.52.
Bis(2-p h en ylcyclop en t a [l]p h en a n t h r en e)zir con iu m
Dich lor id e (3). The yield was 90 mg (11%). 1H NMR (600
MHz, CD2Cl2): δ 8.44 (d, J ) 8.2 Hz, 4 H1), 7.55 (t, J ) 8.2
Hz, 4 H2), 7.41 (t, J ) 7.8 Hz, 4 H3), 7.38 (d, J ) 7.8 Hz, 4 H4),
7.06 (s, 4 H5), 7.63-7.57 (m, 10 HPh). MS (EI): m/z 744
(12%, M+), 453 (41%, M+ - 5), 415 (18%, M+ - 5 - Cl), 291
(85%, 5 - H). Anal. Calcd for C46H30ZrCl2‚CH2Cl2: C, 68.03;
H, 3.89. Found: C, 68.14; H, 4.00.
Bis(2-m et h ylcyclop en t a [l]p h en a n t h r en e)zir con iu m
Diben zyl (2B). A 77 mg sample of 3 (0.12 mmol) in 10 mL of
toluene was treated with 0.23 mL of a 1.1 M solution of
benzylmagnesium chloride (0.25 mmol) in ether. The reaction
mixture was stirred for 2 days at room temperature, evapo-
rated to dryness, washed with 10 mL of pentane, and extracted
with 20 mL of toluene. The toluene extract was evaporated
to dryness to yield a yellow powder. 1H NMR (600 MHz,
C6D5CD3): δ 8.26 (d, J ) 7.5 Hz, 4 H1), 7.33-7.28 (m, 8 H2,3),
7.54 (d, J ) 7.4 Hz, 4 H4), 6.06 (s, 4 H5), 6.64 (d, J ) 7.4 Hz,
4 Ho-Ph), 7.07 (t, J ) 7.6 Hz, 4 Hm-Ph), 6.82 (t, J ) 7.2 Hz,
2 Hp-Ph), 1,33 (s, 6 HMe), 0,63 (s, 4 Hmethylene).
Bis(2-p h en ylcyclop en t a [l]p h en a n t h r en e)zir con iu m
Diben zyl (3B). A 20 mg sample of 4 (0.03 mmol) in 5 mL of
toluene was treated with 0.05 mL of a 1.1 M solution of
benzylmagnesium chloride (0.06 mmol) in ether. The reaction
mixture was kept for 2 days at room temperature and
evaporated to dryness to yield a yellow powder. 1H NMR (600
MHz, C6D5CD3): δ 8.08 (d, J ) 8.1 Hz, 4 H1), 7.21 (t, J ) 7.4
Hz, 4 H2), 7.07 (t, J ) 7.3 Hz, 4 H3), 6.6 (bs, 4 H4), 6.8 (bs, 4
H5), 7.50 (d, J ) 7.4 Hz, 4 Ho-Ph-Phe), 7.34 (t, J ) 7.4 Hz, 4
H
m-Ph-Phe), 7.26 (t, J ) 7.3 Hz, 2 Hp-Ph-Phe), 6.30 (d, J ) 7.5
Hz, 4 Ho-Ph-Bn), 6.98-6.96 (m, 4 Hm-Ph-Bn), 6.82 (t, J ) 7.3
Hz, 2 Hp-Ph-Bn), 0,52 (s, 4 Hmethylene).
P olym er iza tion s a n d P olym er An a lysis. Into a Bu¨chi 1
L autoclave (previously cleaned by stirring with a 0.5% solution
of AliBu3 in toluene and subsequently dried in vacuo at 50 °C)
were added, in the following order, 350 mL of toluene, 4.50 g
of methylaluminoxane MAO (5.21 mass % Al, molar mass ca.
800, Witco AG) in 10 mL of toluene, and 10 mL of a toluene
solution of the zirconocene complex (Table 2). After the reaction
mixture was stirred for 30 min at 0 °C under an argon
atmosphere of 1 bar, the autoclave was pressurized with a
propene pressure of 2 bar (i.e., to a total pressure of 3 bar).
During the polymerization reaction, the total pressure of 3 bar
was kept constant by further addition of propene. Finally, the
autoclave was vented and the reaction mixture drained into 1
L of methanol acidified with 10 mL of concentrated aqueous
HCl. The precipitated polymer was collected by filtration,
washed with methanol, and dried at 50 °C to constant weight.
Molecular weight distributions were determined by GPC,
melting points by DSC (Kunststoff-Labor, BASF AG). Polymer
13C NMR spectra were measured in C2D2Cl4 (δ 74.1 ppm) at
(22) Nele, M.; Collins, S.; Diaz, M. L.; Pinto, J . C.; Lin, S.;
Waymouth, R. M. Macromolecules, submitted.
(23) Organikum; Becker, H. G. O., et al. VEB Deutscher Verlag der
Wissenschaften: Berlin, 1977; p 623.
(24) General NMR procedures for biopolymer characterization
are described by: Kessler, H.; Schmitt, W. In Encyclopedia of
Nuclear Magnetic Resonance; Grant, D. M., Harris, R. K., Eds.;
Wiley: Chichester, 1996.
(25) DNMR-SIM (Ha¨gele, G., Heidelberg, Germany) is available via
http://www.uni-duesseldorf.de/WWW/MathNat/AC1/forschung/nmr/
dnmr-sim.htm. The program is based on DNMR 3 (Binsch, G.; Kessler,
H. Angew. Chem., Int. Ed. Engl. 1980, 19, 445).
(26) Kessler, H. Angew. Chem. 1970, 82, 237.

3604 Organometallics, Vol. 19, No. 18, 2000
Schneider et al.
120 °C on a 250 MHz spectrometer operated at 62.9 MHz and
analyzed by methods described in the literature.27,28
publication no. CSD 408120 and CSD 408121. Copies of
the data can be obtained, free of charge, on application to
CCDC, 12 Union Road, Cambridge CB2 1EZ, UK (fax: +44
1223 336033 or e-mail: deposit@ccdc.cam.ac.uk).
F or ce-F ield Ca lcu la tion s. Calculations were carried out
using the Hyperchem 5.0 program on a conventional PC
(Pentium 200) in the manner described above. In the case of
complexes 1B-3B, 1C, and 3C care was taken to find the
overall minimum for each point with respect to the rotational
freedom of the ligands. Minima on the obtained energy surface
were optimized without restraints. Transition states were
located as points of highest energy on rotation paths with one
torsion angle restrained with an accuracy of (1°.
Cr yst a l St r u ct u r e Det er m in a t ion s. Space group
determinations, diffraction data collection, and solution and
refinement of the structures were conducted as summarized
in Table 7. Crystallographic data (excluding structure factors)
for the structures in this paper have been deposited with the
Cambridge Crystallographic Data Centre as supplementary
Ack n ow led gm en t. We thank Prof. R. M. Waymouth
(Stanford University) for sharing unpublished results
with us, Monika Cavegn (University of Konstanz) for
two-dimensional NMR measurements, and Dr. Stefan
Mansel (CSIS, Madrid) and Dr. J u¨rgen Suhm (BASF
AG, Ludwigshafen) for DSC and GPC measurements.
Financial support by BMBF and by fellowships of the
Landesgraduiertenfo¨rderung Baden-Wu¨rttemberg is
gratefully acknowledged.
Su p p or tin g In for m a tion Ava ila ble: 2D plots of the
energy surfaces calculated, force-field parameters involving
metal coordination, and tables of the X-ray crystallographic
data for complexes 3 and 3B. This material is available free
of charge via the Internet at http://pubs.acs.org.
(27) Ewen, J . A. J . Am. Chem. Soc. 1984, 106, 6355. (b) Ewen, J . A.
In Catalytic Polymerizations of Olefins; Keji, T., Soga, K., Eds.;
Kohdansha Elsevier: Tokyo, 1986; p 271. (c) Cheng, H. N.; Ewen, J .
A. Makromol. Chem. 1989, 190, 1931.
(28) Busico, V.; Cipullo, R.; Corradini, P.; Landriani, L.; Vacatello,
M.; Segre, A. L. Macromolecules 1995, 28, 1887.
OM000149J