New C1-symmetric Ph2C-bridged multisubstituted ansa-zirconocenes for highly isospecific propylene polymerization: Synthetic approach via activated fulvenes

Article abstract of DOI:10.1021/om100289y

The synthesis of multisubstituted diphenylmethylene-bridged fluorenyl-cyclopentadienyl proligands Ph₂C(3,6-tBu ₂FluH)(3-R¹-5-R²C₅H₃) (Flu = fluorenyl; R¹ = tert-butyl, R²

Full text of DOI:10.1021/om100289y

Organometallics 2010, 29, 5073–5082 5073
DOI: 10.1021/om100289y
New C1-Symmetric Ph2C-Bridged Multisubstituted ansa-Zirconocenes
for Highly Isospecific Propylene Polymerization: Synthetic Approach via
Activated Fulvenes†
Evgeny Kirillov,*,‡,§ Nicolas Marquet,‡ Abbas Razavi, Vincenzo Belia, Frank Hampel,§
Thierry Roisnel,‡ John A. Gladysz,*,§,^ and Jean-Franc-ois Carpentier*,‡
‡
§
Universite de Rennes 1, Friedrich-Alexander Universitat Erlangen-Nurnberg, Total Petrochemicals,
ꢀ
€
€
Feluy, and ^Texas A&M University
Received April 9, 2010
The synthesis of multisubstituted diphenylmethylene-bridged fluorenyl-cyclopentadienyl proli-
gands Ph2C(3,6-tBu2FluH)(3-R1-5-R2C5H3) (Flu = fluorenyl; R1 = tert-butyl, R2 = H (2a); R1 =
tert-butyl, R2 = Me (2b); R1 = cumyl, R2 = Me (2c)) was developed using a nucleophilic addition
protocol based on regular and activated fulvenes. Two highly congested proligands (2b,c) were pre-
pared in a two-step procedure, starting first by addition of [3,6-tBu2Flu]þLi- onto 6,60-bis(p-
chlorophenyl)fulvenes, followed by a Pd-catalyzed reductive dechlorination. The X-ray crystal
structures of 2a and 2b CH2Cl2 were determined. These revealed a preorganized sandwichlike
3
positioning of the fluorenyl and cyclopentadienyl plane fragments, similar to that observed in the
metallocenes. The corresponding dichlorozirconium complexes {Ph2C(3,6-tBu2Flu)(3-R1-5-R2-
C5H2)}ZrCl2 (R1 = tert-butyl, R2 = H (3a); R1 = tert-butyl, R2 = Me (3b); R1 = cumyl, R2 = Me
(3c)) were prepared by salt metathesis and characterized by elemental analysis, NMR spectroscopy,
and X-ray crystallography (for 3a,c). These C1-symmetric zirconocenes exist as racemic mixtures of
two enantiomers, arising from planar chirality at the Cp ring. When activated with MAO, 3a,b
showed high activity in the polymerization of propylene (5020 and 3580 kg of iPP mol-1 h-1
,
respectively; toluene solution, 60 ꢀC), affording highly isotactic ([m]4 94.0%) polymers with mole-
cular weights in the range Mw = 30 000-175 000, which are similar to those obtained with the
corresponding CMe2-bridged catalysts. No activity was observed with the 3c/MAO system, which
was proposed to be due to a deactivation involving a cumyl phenyl ring. This was supported by DFT
computations on the cationic species [{Ph2C(3,6-tBu2Flu)(3-cumyl-5-Me-C5H2)}ZrMe]þ, which
revealed facile coordination of the phenyl ring to the metal center, followed by C-H(phenyl)
activation at the ortho position with concomitant elimination of methane; propylene insertion into
the resulting ortho-metalated species was found to be thermodynamically unfavorable.
Introduction
For instance, in the series of single-carbon-bridged systems,
Cs-symmetric precatalysts (Scheme 1; I), highly syndiotactic
polypropylene is commonly produced ([r]4 75-91%; liquid
propylene, MAO, Tpolym=40-80 ꢀC).2 Simple modification
of the ligand skeleton in the precatalyst, namely, installation
of a bulky substituent (tBu group) at the 3-Cp position that
imposes an overall C1 symmetry of the molecule (Scheme 1;
II and III), results in highly isospecific systems for polymer-
ization of propylene ([m]4 79-98%; liquid propylene, MAO,
Tpolym = 40-80 ꢀC).2b,3,4 All these stereoselective polymer-
ization catalyst systems operate via an enantiomorphic-site
control mechanism.
Among the great number of industrially relevant metallo-
cene-based polymerization catalysts, group 4 ansa-metallo-
cenes supported by fluorenyl-cyclopentadienyl ligands hold
a unique position. Their highly tunable ligand platform
allows the introduction of various substituents at different
positions of the cyclopentadienyl (Cp), fluorenyl (Flu), and
bridge moieties and, therefore, access to a class of catalysts
that features high catalytic activity, excellent control, and
remarkable stereospecificity in olefin (co)polymerization.1
† Part of the Dietmar Seyferth Festschrift. Dedicated to Prof. Dietmar
Seyferth in warm appreciation of his decades of valuable service to the
organometallic community.
(2) For leading references, see: (a) Ewen, J. A.; Jones, R. L.; Razavi,
A.; Ferrara, J. D. J. Am. Chem. Soc. 1988, 110, 6255. (b) Razavi, A.;
Bellia, V.; De Brauwer, Y.; Hortmann, K.; Peters, L.; Sirol, S.; Van Belle, S.;
Thewalt, U. Macromol. Chem. Phys. 2004, 205, 347.
(3) Razavi, A.; Bellia, V.; Baekelmans, D.; Slawinsky, M.; Sirol, S.;
Peters, L.; Thewalt, U. Kinet. Catal. 2006, 47, 257.
(4) (a) Wondimagegn, T.; Wang, D.; Razavi, A.; Ziegler, T. Organo-
metallics 2009, 28, 1383. (b) Tomasi, S.; Razavi, A.; Ziegler, T. Organo-
metallics 2009, 28, 2609.
*To whom correspondence should be addressed. E-mail: evgueni.
kirillov@univ-rennes1.fr (E.K.); gladysz@mail.chem.tamu.edu (J.A.
G.); jean-francois.carpentier@univ-rennes1.fr (J.-F.C.). Fax: (þ33)
223236939 (J.-F.C.). Tel: (þ33) 223235950 (J.-F.C.).
(1) For leading reviews, see: (a) Alt, H.; Samuel, E. Chem. Soc. Rev.
1998, 323. (b) Alt, H. Macromol. Chem. Symp. 2001, 173, 65. (c) Razavi, A.
Coord. Chem. Rev. 2006, 250, 155.
r
2010 American Chemical Society
Published on Web 06/07/2010
pubs.acs.org/Organometallics

5074 Organometallics, Vol. 29, No. 21, 2010
Kirillov et al.
Scheme 2. Synthesis of Fulvenes and Flu/Cp Ligands
Scheme 1
Another interesting feature of this class of catalysts is the
influence of the distal substituents R in the R2C bridge on the
catalytic activity and the molecular weight of the resulting
iPP. Thus, replacement of methyl groups with phenyl groups
in the R2C bridge enabled a ca. 4-fold increase of the molecular
weight of polypropylene obtained with the syndiospecific
system I; the same observation was made for the nonsub-
stituted fluorenyl system.5 For those syndiospecific systems,
it was proposed that such a replacement would result in an
increase of hapticity (from η1/η3 to η5) of the fluorenyl
moiety and a concomitant decrease of electrophilicity of
the metal center.6 Some Ph2C-bridged variations of the
3-substituted-Cp system II have been reported by Bercaw
et al. (2-adamantyl or norbornyl instead of tBu: hemiisotactic
PP, [m]4 14-28%; liquid propylene, MAO, Tpolym = 0-
20 ꢀC).7 In this case, a similar 4-fold increase of molecular
weights was observed upon going from Me2C- to Ph2C-
bridged systems. On the other hand, metallocene systems of
type III (i.e., 3,5-disubstituted Cp8) bearing such Ph2C
bridges, and their performance in propylene polymerization
(anticipated to not only be highly isospecific but also offer
high-molecular-weight polymers), have not been documen-
ted thus far.
Accordingly, we set out to synthesize Ph2C-bridged proli-
gands of the type III and ansa-zirconocene complexes de-
rived therefrom and assess the influence of the phenyl groups
on the production of iPP. As detailed here, the catalytic
performances of the zirconium complexes were, after their
activation, investigated in homogeneous polymerizations of
propylene and compared to those of the classic system based
on {Me2C(3,6-tBu2Flu)(3-tBu-5-MeC5H2)}ZrCl2 (III-Me2).
Interestingly, the syntheses of the highly sterically hindered
proligands and corresponding metallocenes proved to be
quite challenging, regarding significant modifications of
protocols used for analogous Me2C-bridged systems.
ever, that the success of this reaction strongly depends on the
nature of the substituents on the five-membered fulvene ring
and, even more importantly, of the terminal substituents at the
exomethylene group. For instance, 6,60-undecamethylene- anþd
6,60-tetradecamethylenefulvenes do not react with [Flu]-Li
(Flu- = C13H9- =fluorenyl anion), even in the presence of
HMPA.10 Indene-derived 6,60-diphenylbenzofulvene showed
no reactivity toward [Flu]-Liþ, as well.11 Also, 6,60-diphenyl-
2,5-dimethylfulvene and [Flu]-Liþ were found intact, whereas
6-phenyl-2,5-dimethylfulvene readily reacts with [Flu]-Liþ to
give (after hydrolytic workup) the desired product Ph(H)C-
(FluH)(2,5-Me2C5H3).12 Nonetheless, the successful synthesis
of several sterically hindered Ph2C(Flu#H)(3-RCpH) proli-
gands (Flu# =fluorenyl, 2,7-di-tert-butylfluorenyl, octame-
thyloctahydrodibenzofluorene and R= H, 2-adamantyl) was
reported by Bercaw.7,13
In the present investigation, a variety of fulvenes (1a14 and
1b-g) were synthesized in good yields using an inexpensive
and efficient methodology (Scheme 2).15 To prepare the
target proligands Ph2C(3,6-tBu2FluH)(3-R1-5-R2C5H3),
these fulvenes were subsequently reacted with the [3,6-
tBu2Flu]- anion (Scheme 2). It turned out that the relatively
less crowded fulvene 1a, which does not bear a methyl
substituent at the 5-position, reacts with [3,6-tBu2Flu]-Liþ
exclusively in Et2O at 90 ꢀC (in an autoclave) over several
days. The corresponding proligand was subsequently iso-
lated in low yield (21%).16 In striking contrast, the more
sterically encumbered fulvene 1b did not undergo nucleo-
philic addition of the fluorenyl anion under a variety of
Results and Discussion
Syntheses of Proligands. Nucleophilic addition of cyclo-
pentadienyl-type anions to fulvenes has proven to be the
most versatile and efficient route for the synthesis of the
monocarbon-bridged bis(cyclopentadienyl) ligands.9 A vari-
ety of such fluorenyl-cyclopentadienyl type ligands have
been prepared via this economical procedure using the
appropriate fulvene and fluorenyl anion. It appears, how-
(10) Alt, H.; Jung, M. J. Organomet. Chem. 1998, 568, 87.
(11) Alt, H.; Jung, M.; Kehr, J. J. Organomet. Chem. 1998, 562, 153.
(12) Won, Y. C.; Kwon, H. Y.; Lee, B. Y.; Park, Y.-W. J. Organomet.
Chem. 2003, 677, 133.
(13) Miller, S. A.; Bercaw, J. E. Organometallics 2004, 23, 1777.
(14) Miller, S. A.; Bercaw, J. E. Organometallics 2006, 25, 3576.
(15) (a) Kirillov, E.; Gladysz, J. A.; Razavi, A. (Friedrich-Alexander
(5) Razavi, A.; Peters, L.; Nafpliotis, L. J. Mol. Catal. A 1997, 115,
129.
(6) Razavi, A.; Atwood, J. L. J. Organomet. Chem. 1993, 459, 117.
(7) Miller, A.; Bercaw, J. E. Organometallics 2002, 21, 934.
(8) Those systems should be officially referred to as (1),2,4-trisub-
stituted. However, the (1),3,5 method of numbering the Cp ring in this
publication is adapted for historical reasons (researchers first prepared
the 3-tert-butyl-substituted complex and then complexes having an
additional substituent at the “5” position).
€
€
Universitat Erlangen-Nurnberg and Atofina), Eur. Pat. Appl. EP
05105162.1, 2005 (13/06/2005). (b) Carpentier, J.-F.; Kirillov, E.; Razavi,
ꢀ
A. (Universite de Rennes 1 and Total Petrochemicals) Eur. Pat. Appl. EP
(9) Day, J. H. Chem. Rev. 1953, 53, 167.
06121181.9, 2006 (25/09/2006).

Article
Scheme 3. Palladium-Catalyzed Reductive Dechlorination
of 2d,e
Organometallics, Vol. 29, No. 21, 2010 5075
conditions, some of them quite severe (solvents refluxing
THF, Et2O, nBu2O, toluene; addition of HMPA).17
These observations prompted us to strive for an indirect
approach, employing more reactive starting compounds.
Thus, the reaction of [3,6-tBu2Flu]-Liþ with “activated”
fulvenes 1d,e, which have slightly electron-withdrawing
p-chloro substituents on both phenyl rings, allowed access to
the corresponding ligand precursors 2d,e in good yields (60
and 80%, respectively). The latter compounds were further
subjected to a Pd-catalyzed reductive dechlorination using a
2:1 KOtBu/NaBPh4 mixture as the reducing agent18 and
Nolan’s catalyst (SIPr)Pd(C3H5)Cl19 to afford selectively the
desired compounds 2b,c (Scheme 3). The analogous reaction
of [3,6-tBu2Flu]-Liþ with the bis(p-fluorophenyl)fulvene 1f
afforded adduct 2f in significantly lower yield (7%).20 How-
ever, a similar reaction of [3,6-tBu2Flu]-Liþ with fulvene 1g,
which has two trifluoromethyl meta substituents on each
phenyl ring, gave a complex mixture of unidentified products.
Therefore, the use of chloro activating groups appears crucial
to achieve selective nucleophilic addition of the fluorenyl anion
and allow their subsequent selective removal, to eventually
yield the diphenylmethylene-bridged proligand.
Figure 1. Crystal structures of compounds 2a (a) and 2b CH2Cl2
3
(b). H atoms, except some of the five-membered rings, and
solvates are omitted for clarity; ellipsoids are drawn at the 50%
probability level.
All the proligands prepared were stable at room tempera-
ture in solution and in the solid state, and the structures were
authenticated by NMR spectroscopy, FAB-MS, and ele-
mental analysis. The steric congestion of 2a-f was evidenced
by NMR spectroscopy in CDCl3 or THF, which showed
fluxional behavior that was attributed to hindered rotation
of all substituents around the bridging quaternary carbon
atom (and possibly isomerization of CdC bonds within the
1
CpH ring as well). In fact, the H NMR spectra of these
compounds were remarkably broadened at ambient tempera-
ture. On the NMR time scale, sharp resonances are observed
at -40 ꢀC. However, the complexity of these spectra, arising
from the presence of several isomers and/or conformers
(and the consequent overlapping of resonances), made them
hardly informative (see the Supporting Information). Also,
the fast exchange regime could not be reached, as these
compounds decompose at 90-100 ꢀC in C2D2Cl4, giving
thus far unidentified materials (most likely arising from C-C
bond cleavage at the quaternary carbon atom).
(16) The solvent has a pronounced influence upon this reaction. For
example, when THF was used, no addition of the fluorenyl anion to
fulvene 1a took place, even at elevated temperatures and in the presence
of HMPA. When Et2O was used, the reaction proceeded; however,
significant heating was required (autoclave conditions).
(17) The preparation of proligand 2b in 71% yield by the direct
addition of [3,6-tBu2Flu]-Liþ to the nonactivated fulvene 1b has been
claimed following extended reaction times (refluxing diethyl ether, 13
days); however, several attempts to reproduce this experiment failed in
our hands. See: Ikenaga, S.; Okada, K.; Takayasu, H.; Inoue, N.;
Hirota, N.; Kaneyoshi, H.; Funaya, M.; Kawai, K.; Kawahara, N.;
Kojoh, S.; Kashiwa, N.; Mori, R. (Mitsui Chemicals Inc.) Eur. Pat.
Appl. 1614699, 2006; Chem. Abstr. 2006, 141, 332634.
To further characterize these constrained molecules, sin-
gle-crystal X-ray diffraction analyses were performed for
two of them, 2a and 2b (as the 2b CH2Cl2 solvate) (Figure 1).
(18) Considerably poorer results were obtained when KOtBu alone
was used as the reducing agent; in this case, many side products formed.
(19) Navarro, O.; Marion, N.; Oonishi, Y.; Kelly, R. A., III; Nolan,
S. P. J. Org. Chem. 2006, 71, 685.
3
Interestingly, in the solid state, both molecules were preor-
ganized in a sandwich-type structure with the Cp0H and
Flu0H groups arranged in a fashion similar to that expected
after coordination to a metal center. Furthermore, the con-
formers exhibit planar chirality, although both enantiomers
are found in the two unit cells. Striking consequences of the
sterically demanding substituents at the quaternary carbon
(20) The observation that 6,60-bis(arene)fulvenes can be “activated”
toward nucleophilic addition by introducing electron-withdrawing
groups X in the para positions of the arenes is in line with the effect of
such substituents described by their Hammett σp values. For example,
ligands 2b,c could not be prepared by the classical way starting from 1b,c
(X=H, σp=0.00), even under forcing reaction conditions. At the same
time, 2d,e were obtained in rather good yields (X = Cl, σp = 0.23),
whereas the yield of the fluorinated congener 2f was much lower (X=F,
σp =0.06).
atoms in 2a and 2b CH2Cl2 are the C(1)-C(6)-C(9) and
3

5076 Organometallics, Vol. 29, No. 21, 2010
Kirillov et al.
˚
Table 1. Selected Bond Distances (A) and Angles (deg) for Proligands 2a and 2b CH2Cl2, Complexes 3a,c, and Reference
1c
Molecule III-Me2
3
2a
2b CH2Cl2
3a
3c
III-Me2
3
Zr-Cl
2.425(2)
2.427(2)
2.440(7)
2.470(8)
2.602(7)
2.571(7)
2.468(8)
2.201(7)
2.422(7)
2.545(7)
2.704(7)
2.670(7)
2.515(7)
2.257(7)
118.07(5)
98.9(5)
2.4227(10)
2.4283(10)
2.442(4)
2.492(3)
2.630(4)
2.547(4)
2.457(4)
2.204(4)
2.426(3)
2.497(4)
2.666(4)
2.712(4)
2.588(4)
2.269(4)
118.03(4)
98.60(3)
104.00(4)
2.8020(6)
2.6390(11)a
2.427(3)
2.464(3)
2.611(4)
2.549(4)
2.461(4)
Zr-C(1)
Zr-C(2)
Zr-C(3)
Zr-C(4)
Zr-C(5)
Zr-CpCent
Zr-C(9)
2.432(4)
2.512(4)
2.653(4)
2.688(4)
2.611(4)
Zr-C(10)
Zr-C(11)
Zr-C(12)
Zr-C(13)
Zr-FluCent
CpCent-Zr-FluCent
C(1)-C(6)-C(9)
C(7)-C(6)-C(8)
118.5(3)
100.2(3)
105.3(3)
104.03(17)
103.28(17)
105.48(12)
103.03(13)
103.1(6)
a The crystal structure reported in ref 1c is for a mixed chloro-iodo zirconocene, and the Zr-Cl bond distances must therefore be interpreted
with caution.
Scheme 4. Synthesis of ansa-Zirconocenes 3a-c
basic sec-BuLi as the deprotonating reagent. A similar
approach afforded 3b as a reddish pink solid. However, this
complex was not entirely pure, as reflected by a few minor
broadened 1H NMR signals in addition to those expected for
3b (see the Supporting Information).22 All attempts to
further purify the material were unsuccessful. Nonetheless,
the formulations of all zirconium complexes were supported
by microanalyses (see the Experimental Section).
Selective deprotonation of compounds 2d,e by nBuLi, sec-
BuLi, tBuLi, and MeLi failed; rather, metalation/alkylation
sequences involving the chlorinated phenyl groups appeared
to occur.23 Also, when nBuLi/TMEDA, KOtBu/nBuLi in
Et2O, tBuLi in pentane, or LDA in refluxing THF were used,
only unidentified brownish solids were isolated after subse-
quent addition of ZrCl4. Attempts to isolate and characterize
dianionic derivatives of these ligands failed as well, appar-
ently reflecting their high instability.
C(7)-C(6)-C(8) bond angles (104.03(17), 103.28(17)ꢀ and
105.48(12), 103.03(13)ꢀ, respectively). Note, however, that
the C(1)-C(6)-C(9) bond angles remain expectedly much
larger (6-7ꢀ) than those observed in the corresponding
metallocenes (vide infra, Table 1).
Synthesis and Structure of Zirconocenes. In order to pre-
pare zirconocene dichlorides, standard salt metathesis reac-
tions between ZrCl4 and ligand dianions, generated in situ
in Et2O, were attempted (Scheme 4). Thus, zirconium com-
plex 3a was isolated in good yield as a characteristically pink,
microcrystalline material. Unexpectedly, similar one-pot
synthetic protocols to coordinate proligands 2b,c appeared
to fail. Instead, unidentified brownish amorphous solids
were obtained after in situ metalation of these proligands
and subsequent treatment with ZrCl4 in Et2O, followed by
the usual workup. In striking contrast to 2a, the more
sterically hindered molecules 2b,c, which incorporate an
additional methyl group at the 5-position of the Cp ligand,
do not undergo deprotonation at the sp3 carbon of the
fluorenyl moiety upon addition of MeLi or nBuLi (in Et2O
or pentane).21 Pure crystalline 3c was finally obtained from
the reaction conducted in toluene, using the more strongly
Single crystals of 3a,c suitable for X-ray diffraction studies
were grown from CH2Cl2/hexane (1:1 v/v) solutions at room
temperature. Their solid-state structures are depicted in
Figure 2, and selected geometrical parameters are given in
Table 1. The unit cells of both C1-symmetric complexes are
composed of pairs of enantiomers. These molecules exhibit
geometrical parameters essentially similar to those of zirco-
nocene dichlorides incorporating isopropylidene- and diphe-
nylmethylene-bridged {Cp/Flu} ligands.1c,7,13,24 In both
complexes, the coordination mode of the central five-mem-
bered ring of the fluorenyl ligand deviates slightly from η5
toward η3, as evidenced by the significant differences in the
˚
M-C(ring) distances (ca. 0.3 A between the shortest and the
longest bond lengths). The Cpcent-Zr-Flucent bite angles in
3a,c are almost identical (118.07(5) and 118.03(4)ꢀ, respec-
tively) and compare well with the corresponding value in the
isopropylidene-bridged metallocene III-Me2 (118.5(3)ꢀ). One
must keep in mind, however, that such Cp-Flu ansa-zirco-
nocenes may have very similar structural features in the solid
state but quite different chemical behavior (i.e., catalytic per-
formance in propylene polymerization, propensity of adjacent
(21) This tentative conclusion stems from the fact that, when ZrCl4
was added after the deprotonation step, the products isolated after
workup all featured H NMR spectra consistent with the formulation
1
{Ph2C(FluH)(Cp)}Zr(R)Cl2 (R=Me, Bu).
(22) Attempted syntheses of zirconocenes by reacting protio ligands
2b,c with Zr(CH2Ph)4 or Zr(NMe2)4 gave only products with Cp-
coordinated ligands but pendant FluH groups, i.e. {Ph2C(FluH)-
(Cp0)}ZrX3 with X=CH2Ph, NMe2, as revealed by NMR monitoring.
Similar alkane or amine elimination reactions conducted at elevated
temperatures resulted in substantial decomposition of the aforemen-
tioned intermediates.
(23) Bailey, W. F.; Luderer, M. R.; Jordan, K. P. J. Org. Chem. 2006,
71, 2825.
(24) Razavi, A.; Thewalt, U. J. Organomet. Chem. 2001, 621, 267 and
references cited therein.

Article
Organometallics, Vol. 29, No. 21, 2010 5077
(see the Experimental Section) was made by 2D (COSY,
1
HETCOR) experiments. Analysis of the H and 13C NMR
chemical shifts, in particular that of the C9-fluorenyl carbon,
which could be unambiguously assigned in the case of 3a,
was not conclusive regarding the actual hapticity of the
fluorenyl moiety in solution.26
Propylene Polymerization Catalysis. Metallocenes 3a-c,
in combination with MAO, were evaluated in the homo-
geneous polymerization of propylene (toluene, 5 bar of
constant pressure, 60 ꢀC). Each polymerization experiment
was repeated independently three times under the same
conditions, revealing good reproducibility in terms of activ-
ity (gas uptake) and productivity (polymer yield), as well as
physicochemical properties (Mw, Mn, Tm, isotacticity) of the
isolated polymer. For comparison purposes, the catalytic
performance of the system based on the reference metallo-
cene precatalyst III-Me2 under these conditions was deter-
mined as well. Selected polymerization results are summarized
in Table 2.
Analysis of these results reveals that systems based on
metallocenes 3a,b feature, under our experimental condi-
tions, superior activities (5020 and 3580 kg of iPP mol-1 h-1
,
respectively; entries 2 and 3) as compared to the already high
activity of the reference metallocene precatalyst III-Me2
(1710 kg of iPP mol-1 h-1; entry 1).27
Regarding the molecular weights of the polypropylenes
produced, systems based on III-Me2 and 3b (=III-Ph2)
yielded very similar products (compare entries 1 and 3).
These results are in contrast to the increased molecular
weights commonly observed when a Me2C bridge is replaced
by a Ph2C bridge. On the other hand, the molecular weight of
the polypropylene obtained from 3a is significantly lower,
about 20% of those obtained from III-Me2 and 3b (entry 2).
This trend is not unexpected and can be accounted for by the
absence of the 5-methyl substituent on the Cp ring of 3a, which
is known to be responsible for destabilizing a transition state
leading to chain transfer to monomer or β-H elimination.2b,3,4
The regioselectivities for primary insertion of the mono-
mer are nearly perfect for all metallocene precursors
(99.8-99.9%). In terms of stereocontrol, the new metallo-
cenes 3a,b also afforded polypropylenes featuring high pen-
tad [m]4 values, as determined by 13C NMR spectroscopy.
This also confirmed an operative enantiomorphic-site
stereocontrol mechanism (see the Supporting Information).
These isotacticities were quite similar to that for the polymer
obtained with the reference precatalyst III-Me2. This is also
reflected by the identical Tm values determined for these
polymers.
Remarkably, the system 3c/MAO appeared to be com-
pletely inactive under the conditions used (entry 4). The
latter observation may be accounted for by a deactivation
involving the phenyl ring of the cumyl group. Indeed, a
similar catalytic inactivity has been documented for the
metallocene precursor {Me2C(Flu)(3-(2-Ph-2-adamantyl)-
Cp)}ZrCl2.28
DFT Computations of a Possible Deactivation Pathway. To
gain a better insight into this deactivation phenomenon,
DFT computations were conducted starting from the putative
Figure 2. Crystal structures of complexes 3a (a) and 3c (b). H
atoms are omitted for clarity; ellipsoids are drawn at the 50%
probability level.
C6 rings of the fluorenyl moiety to be hydrogenated or not25)
in solution.
The 1H and 13C NMR spectra of 3a-c show a single set of
resonances, consistent with the C1 symmetry of these species.
1
Accurate assignment of the H and 13C NMR resonances
(25) For instance, the benzenoid rings of the fluorenyl ligand in
{Ph2C(Cp)(Flu)}ZrCl2 readily undergo hydrogenation while the iso-
structural Me2C-bridged analogue does not, even under forcing condi-
tions. This behavior has been rationalized by the η5 coordination mode
in the former case, while in the latter case, the prevalent coordination
mode is η3, with dynamic hapticity toward η1.6
(26) (a) Irwin, L. J.; Reibenspies, J. H.; Miller, S. A. Polyhedron 2005,
24, 1314. (b) Drago, D.; Pregosin, P. S.; Razavi, A. Organometallics 2000,
19, 1802.
(27) As a consequence, polymerizations with such systems must be
operated at very low metallocene concentration (10 μmol L-1) in order
to control the high exothermicity.
3

5078 Organometallics, Vol. 29, No. 21, 2010
Kirillov et al.
Table 2. Propylene Polymerizationa
productivity activity
Ccat
entry precatalyst (μmol L-1
amt of MAO time
(equiv)
Tm
[m]4 d 1,2-insertiond
c
(%)
c
)
(min) mp (g) (g/g of prectatalyst) (kg mol-1 h-1
)
(ꢀC)b 103Mw Mw/Mn (%)
1
2
3
4
III-Me2
3a
3b
10
10
10
200
5000
5000
5000
1000
30
30
30
30
1.39
3.46
2.42
0
1390
3460
2420
0
1710
5020
3580
0
152
152
152
166
30
175
2.4
2.1
2.3
95.2
93.9
94.0
>99.9
99.8
99.9
3c
a Polymerization conditions: 300 mL high pressure glass reactor; solvent toluene, 150 mL; P(propylene) = 5 bar; Tpol = 60 ꢀC. b Determined by DSC.
c Determined by GPC. d Determined by 13C NMR.
cationic methyl-zirconocene species obtained upon activa-
tion with MAO, namely [{Ph2C(3,6-tBu2Flu)(3-cumyl-5-
Me-C5H2)}ZrMe]þ (A; Scheme 5) (BP86 level, see the Sup-
porting Information for details). The objectives of these
nonexhaustive computations were to assess the energy pro-
files for two possible concurrent processes: i.e., (a) classical
coordination/insertion of propylene into the Zr-Me bond of
species A to afford the corresponding insertion product C
versus (b) intramolecular C-H activation in the pendant
phenyl group by the cationic metal center to form o-phenyl-
metalated species D (vide infra) and further reaction of the
latter species with propylene to yield species E. Although
metallocene species with all real substituents on the Cp and
fluorenyl rings were considered, a series of simplifications/
assumptions was made to reduce the calculation costs: (i) the
influence of the MAO counteranion was neglected and
only single cationic species in gas phase were computed;29,30
(ii) the methyl group (polymeryl chain) is located opposite to
the bulky cumyl substituent. In addition, when react-
ions with propylene were envisioned, it was considered that
(iii) propylene insertion proceeds in a primary fashion and
(iv) propylene coordinates to zirconium with its methyl
group positioned “head down” into the free space in the
central region of the fluorenyl ligand and the methylene
directed toward the less crowded quadrant (opposite to the
methyl group/polymeryl chain), as usually considered for
isospecific propylene polymerization mediated by C1-sym-
metric cyclopentadienyl-fluorenyl metallocene catalysts.
The propagation pathway was evaluated at the first inser-
tion stage: that is, conversion of π-complex A-si-C3H6
(Scheme 5) to isobutyl product C. This process appeared
favorable on both kinetic and thermodynamic grounds, with
a low calculated activation barrier of 6.9 kcal mol-1 for
and related species B-II (in which the hydrogen bound to the
interacting C(phenyl) points toward the Zr-Me group) was
found to be significantly exergonic (ΔEAfB = -10.2 and
-7.7 kcal.mol-1, respectively). There are precedents in the
literature for π coordination of pendant arene groups to
cationic group 4 metal centers (η1-η3 coordination mode
observed both in the solid state and in solution)32-34 similar
to that in computed species B-I/B-II (η1 coordination mode).
Exchange of the π-coordinated phenyl ring in B-I for a
propylene unit (leading to A-si-C3H6) is only slightly thermo-
dynamically unfavorable (ΔEA-si-C H6fB-I = þ3.8 kcal
3
mol-1), and that with B-II is even more accessible. The low
energy differences associated with these reversible processes
(B-I T A-si-C3H6 T B-II) are unlikely to account for the
observed complete inactivity toward propylene.
We therefore investigated subsequent chemical events. In
this respect, a likely phenomenon, mentioned above and
previously suggested by Bercaw et al.,28 is metalation (C-H
activation) of the cumyl phenyl ring at the ortho position and
concomitant alkane (methane) elimination. This chemical
event, which requires proximal C-H and Zr-Me groups as
found in B-II (and not in B-I), is computed to occur through
an accessible transition state (EqB-IIfD=13.9 kcal mol-1).35
If this reaction takes place in the absence of propylene, the
overall energy balance appears quite unfavorable (ΔEB-IIfD
=
þ6.4 kcal mol-1); however, in the presence of propylene,
a significant stabilization is calculated (ΔEB-IIfD
si-C3
H6 = -9.6 kcal mol-1), associated with the formation of
the new π complex D-re-C3H6.
Further propylene insertion in the resulting ortho-meta-
lated species D-re-C3H6 (to give species E) is computed to be
thermodynamically disfavored by 8.4 kcal mol-1. We also
checked that the product resulting from a second propylene
insertion31 and was significantly exergonic, -8.1 kcal mol-1
.
insertion is just further stabilized by only 2-3 kcal mol-1
.
On the other hand, coordination of the cumyl phenyl ring
to the zirconium center of the three-coordinated species A to
form eventually B-I (in which the hydrogen bound to the
interacting C(phenyl) points opposite to the Zr-Me group)
These results indicate that the formation of an ortho-
metalated species such as D-re-C3H6 is a likely process and
(32) Upon activation with [Ph3C]þ[B(C6F5)4]-, {PhMe2SiC5H4}-
TiMe3 undergoes coordination of the pendant phenyl ring onto the
cationic titanium center: (a) Deckers, P. J. W.; Van der Linden, A. J.;
Meetsma, A.; Hessen, B. Eur. J. Inorg. Chem. 2000, 929. (b) Deckers,
P. J. W.; Hessen, B.; Teuben, J. H. Angew. Chem., Int. Ed. 2001, 40, 2516.
(c) Deckers, P. J. W.; Hessen, B.; Teuben, J. H. Organometallics 2002, 21,
5122.
(28) No direct experimental evidence has thus far been obtained to
support this hypothesis. Attempts to generate catiþonic species starting
from the dimethylated version of 3c and [Ph3C] [B(C6F5)4]- led to
unidentified compounds, probably arising from rapid decomposition.
On the other hand, similar catalytic inactivity has been documented for
the metallocene precursor {Me2C(Flu)(3-(2-Ph-2-adamantyl)-Cp)}ZrCl2.14
(29) Previous QM/MM studies by Ziegler et al. on the polymerizat-
ion reactions have been conducted using the large size ion pair
[metallocenium]þ[MeB(C6F5)3]- as a model for the active species:
(a) Wang, D.; Tomasi, S.; Razavi, A.; Ziegler, T. Organometallics
2008, 27, 2861. (b) Wondimaggen, T.; Wang, D.; Razavi, A.; Ziegler, T.
Organometallics 2008, 27, 6434. See also ref 4.
(33) For computational studies of this process, see: (a) Blok, A. N. J.;
Budzelaar, P. H. M.; Gal, A. W. Organometallics 2003, 22, 2564. (b) De
Bruin, T. J. M.; Magna, L.; Raybaud, P.; Toulhoat, H. Organometallics 2003,
22, 3404. (c) Tobisch, S.; Ziegler, T. Organometallics 2003, 22, 5392.
(34) Experimental and theoretical studies on intramolecular C-H
activation involving the pendaþnt tolyl group in the cationic titanocene
[(Cp)(CpCMe2C6H4Me)TiMe] [MeB(C6F5)3]- were reported; see:
Sassmannshausen, J.; Baumgartner, J. Organometallics 2008, 27, 1996.
(35) DFT computations of the similar intramolecular C-H activa-
tion reaction of [(Cp)(CpCMe2C6H4Me)TiMe]þ[MeB(C6F5)3]-, yield-
ing [(Cp)(CpCMe2(σ-o-C6H3Me))Ti]þ[MeB(C6F5)3]- and methane,
gave the following parameters: ΔGq = 23.5 kcal mol-1 and ΔG = 3.1
(30) Theoretical investigations dealing with possible structures of
oligomeric MAO have been reported; see: Zurek, E.; Ziegler, T. Prog.
Polym. Sci. 2004, 29, 107.
(31) A 6.7 kcal mol-1 barrier for the (second) insertion of propylene
into the Zr-C bond was calculated for [{Me2C(3,6-tBu2Flu)(3-tBu-5-
Me-C5H2)}Zr(Polymeryl)]þ [MeB(C6F5)3]-.4a
kcal mol-1 34
.
3

Article
Organometallics, Vol. 29, No. 21, 2010 5079
Scheme 5. Energy Profiles Computed at the BP86 Level for Propylene Coordination/Insertion and Possible Decomposition Pathways for
Active Species Derived from 3ca
a Energies are given in kcal mol-1 relative to A.
that the latter species is stable, which is nonreactive toward
propylene.
ring. Other structural modifications of this class of metallo-
cenes to further improve catalytic performance and control
over iPP properties are underway and will be reported in due
course.
Conclusions
We have developed an effective synthetic approach toward
sterically encumbered Ph2C-bridged cyclopentadienyl-
fluorenyl-based ligand systems. A new series of C1-sym-
metric metallocenes was synthesized and successfully app-
lied to the highly isospecific polymerization of propylene.
Precursors 3a,b afford 2-3-fold more active systems than the
reference precatalyst {Me2C(3,6-tBu2Flu)(3-tBu-5-MeC5H2)}-
ZrCl2 (III-Me2) for the polymerization of propylene in
toluene. On the other hand, in striking contrast to both
systems I and II (Scheme 1), replacement of the bridge
methyl groups by phenyl groups in the isospecific precursor
III does not affect substantially the molecular weight of iPP.
The complete inactivity of the 3c/MAO system in propylene
polymerization could be rationalized by DFT calculations,
which indicate a concurrent decomposition pathway invol-
ving intramolecular C-H activation in the cumyl phenyl
Experimental Section
General Considerations. All manipulations were performed
under a purified argon atmosphere using standard Schlenk
techniques or in a glovebox. Solvents were distilled from Na/
benzophenone (THF, Et2O) and Na/K alloy (toluene, pentane)
under nitrogen, degassed thoroughly, and stored under nitrogen
prior to use. Deuterated solvents (benzene-d6, toluene-d8, THF-
d8; >99.5% D, Deutero GmbH and Euroisotop) were vacuum-
transferred from Na/K alloy into storage tubes. CDCl3, CD2Cl2,
and C2D2Cl4 were kept over calcium hydride and vacuum-trans-
ferred before use. The ligand precursors 3,6-di-tert-butylfluorene
and 6,60-diphenyl-3-tert-butyl-5-methylfulvene (1b) were gener-
ously provided by Total Petrochemicals. The precursors 3,6,60-
trimethylfulvene, 1-tert-butyl-3-methylcyclopentadiene (mixture
of isomers), and (1-methyl-3-tert-butylcyclopentadienyl)lith-
ium were prepared according to literature protocols36 and

5080 Organometallics, Vol. 29, No. 21, 2010
Kirillov et al.
characterized by 1H NMR spectroscopy. 1-tert-butylcyclopen-
tadiene (mixture of isomers) was prepared according to the
published procedure.37 The reference metallocene {Me2C(3,6-
tBu2-Flu)(3-tBu-5-MeC5H2)}ZrCl2 (III-Me2) was synthesized
as described in the patent literature.36 The catalyst (SIPr)Pd-
(C3H5)Cl was synthesized as reported.19 Other starting materi-
als were purchased from Acros, Strem, and Aldrich and used as
received. MAO (30 wt % solution in toluene, Albermale; con-
tains ca. 10 wt % of free AlMe3) was used as received.
n-butyllithium (6.67 mL of a 2.5 M solution in hexane, 16.7
mmol) at 0 ꢀC with stirring. After 2 h, a solution of 4,40-
dichlorobenzophenone (4.18 g, 16.66 mmol) in THF (50 mL)
was added dropwise with stirring. The mixture turned orange.
After 4 h, concentrated aqueous NH4Cl (50 mL) was added
slowly. This mixture was stirred overnight. The organic layer
was separated and dried over MgSO4, and all the volatiles were
removed in vacuo. The deep red residue was recrystallized from
1
hot methanol to give 1d (3.7 g, 10.02 mmol, 60%). H NMR
NMR spectra of complexes were recorded on Bruker
AC-200, AC-300, AM-400, and AM-500 spectrometers in
Teflon-valved NMR tubes at 25 ꢀC, unless otherwise indicated.
1H and 13C chemical shifts are reported in ppm vs SiMe4 (0.00),
as determined by reference to the residual solvent peaks. 19F
chemical shifts were determined by external reference to an
aqueous solution of NaBF4. The resonances of organometallic
complexes were assigned from 2D 1H-13C HMQC and HMBC
NMR experiments. Coupling constants are given in hertz.
Elemental analyses (C, H, N) were performed using a Flash
EA1112 CHNS Thermo Electron apparatus and are the aver-
age of two independent determinations. FAB-HRMS spectra
were recorded on a Micromass ZABSpecTOF high-resolution
MS/MS spectrometer. DSC measurements were performed on
a SETARAM Instrumentation DSC 131 differential scanning
calorimeter at a heating rate of 10 ꢀC/min; first and second runs
were recorded after cooling to 30 ꢀC, and the melting tempera-
tures reported in tables correspond to the second run. GPC
analyses of iPP samples were carried out in 1,2,4-trichloroben-
zene at 135 ꢀC in the research center of Total Petrochemicals in
Feluy, Belgium, using polystyrene standards for universal
calibration. 13C NMR analyses of iPP samples were run in
the research center of Total Petrochemicals in Feluy, Belgium,
on a AM-500 Bruker spectrometer using the following condi-
tions: solutions of ca. 200 mg of PP polymer in a trichloro-
benzene/C6D6 mixture at 135 ꢀC in 10 mm tubes, inverse gated
experiment, pulse angle 90ꢀ, delay 11 s, acquisition time 1.25 s,
number of scans 6000.
(CDCl3, 400 MHz, 25 ꢀC): δ 7.32 (m, 4H, Ph), 7.18 (m, 4H, Ph),
6.37 (s, 1H, CH), 5.66 (s, 1H, CH), 1.53 (s, 3H, CH3), 1.17 (s, 9H,
CCH3). Anal. Calcd for C23H22Cl2: C, 74.80; H, 7.00. Found: C,
74.85; H, 7.10.
6,60-Bis(4-chlorophenyl)-3-cumyl-5-methylfulvene (1e). Using
a protocol similar to that described above for 1d, compound
1e was obtained from 1-methyl-3-cumylcyclopentadiene (5.60 g,
28.3 mmol; mixture of isomers), n-butyllithium (11.30 mL of a
2.5 M solution in hexane, 28.3 mmol), and dichlorobenzophe-
none (6.00 g, 23.9 mmol). The crude product was recrystallized
from hot methanol to give deep red crystals of 1e (6.20 g,
14.4 mmol, 51%). 1H NMR (CDCl3, 200 MHz, 25 ꢀC): δ
7.41-7.10 (m, 8H, Ph), 6.02 (br m, 1H, CH), 5.85 (br m, 1H,
CH), 1.55 (s, 6H, CCH3), 1.46 (s, 3H, CH3). Anal. Calcd for
C23H22Cl2: C, 77.96; H, 5.61. Found: C, 78.11; H, 6.01.
6,60-Bis(4-fluorophenyl)-3-tert-butyl-5-methylfulvene (1f). Using
a protocol similar to that described above for 1d, compound 1f
was obtained from 1-methyl-3-tert-butylcyclopentadiene (2.29 g,
16.8 mmol; mixture of isomers), n-butyllithium (6.72 mL of a 2.5 M
solution in hexane, 16.8 mmol), and 4,40-difluorobenzophenone
(3.30 g, 15.1 mmol). The red residue was recrystallized from hot
methanol at 25 ꢀC to give 1f (0.74 g, 2.22 mmol, 15%). 1H NMR
(CDCl3, 300 MHz, 25 ꢀC): δ 7.17 (m, 4H, Ph), 7.03 (m, 4H, Ph),
6.32 (s, 1H, Cp), 5.63 (s, 1H, Cp), 1.47 (s, 3H, CH3), 1.13 (s, 9H,
CCH3). 19F{1H} NMR (CDCl3, 282 MHz, 25 ꢀC): δ -112.9 (s).
Anal. Calcd for C23H22F2: C, 82.11; H, 6.59. Found: C, 83.03; H,
7.34 (these data do not indicate analytical purity but are provided
to illustrate the best result obtained).
6,60-Bis(3,5-bis(trifluoromethyl)phenyl)-3-tert-butyl-5-methyl-
fulvene (1g). Using a protocol similar to that described above for
1d, compound 1g was obtained from 1-methyl-3-tert-butylcy-
clopentadiene (1.30 g, 9.54 mmol; mixture of isomers), n-butyl-
lithium (3.90 mL of a 2.5 M solution in hexane, 9.54 mmol),
and 3,30,5,50-tetrakis(trifluoromethyl)benzophenone (5.00 g,
9.57 mmol). The red residue was recrystallized from hot metha-
1-Cumyl-3-methylcyclopentadiene (Mixture of Isomers). To a
solution of 3,6,60-trimethylfulvene (27.70 g, 230.5 mmol) in THF
(500 mL) was added phenyllithium (120.0 mL of a 2.0 M solu-
tion in dibutyl ether, 240.0 mmol) at 0 ꢀC with stirring. After
20 h, the reaction mixture was hydrolyzed with a concentrated
solution of NH4Cl (150 mL). This mixture was stirred overnight.
The organic layer was separated, dried over MgSO4, and eva-
porated. The crude product was distilled under vacuum (80-
84 ꢀC/0.1 Torr) to give 1-cumyl-3-methylcyclopentadiene as a
mixture of isomers (40.63 g, 205.1 mmol, 89%). Anal. Calcd for
C15H18: C, 90.85; H, 9.15. Found: C, 91.13; H, 10.01.
1
nol at 25 ꢀC to give 1g (1.05 g, 1.83 mmol, 19%). H NMR
(CDCl3, 300 MHz, 25 ꢀC): δ 7.90 (d, 2H, Ph), 7.68 (d, 4H, Ph),
6.38 (s, 1H, Cp), 5.43 (s, 1H, Cp), 1.37 (s, 3H, CH3), 1.11 (s, 9H,
CCH3). 19F{1H} NMR (CDCl3, 282 MHz, 25 ꢀC): δ -62.51 (s,
6F), -62.56 (s, 6F). Anal. Calcd for C27H22F12: C, 56.45; H,
3.86. Found: C, 56.88; H, 4.08.
6,60-Diphenyl-3-tert-butylfulvene (1a). To a solution of tert-
butylcyclopentadiene (1.47 g, 12.0 mmol; mixture of isomers) in
diethyl ether (50 mL) was added n-butyllithium (4.8 mL of a
2.5 M solution in hexane, 12.0 mmol) at 0 ꢀC with stirring. After
2 h, a solution of benzophenone (2.20 g, 12.0 mmol) in diethyl
ether (30 mL) was added dropwise. The reaction mixture turned
orange. After 2 h, aqueous HCl (50 mL of a 20% solution) was
added slowly. The mixture was stirred overnight. The organic
layer was separated, washed with water and NaHCO3 solution,
and dried over MgSO4, and all the volatiles were removed in
vacuo. The orange residue was recrystallized from methanol at
-30 ꢀC to give 1a (1.72 g, 6.00 mmol, 50%). 1H NMR (CDCl3,
400 MHz, 25 ꢀC): δ 7.34 (m, 10H, Ph), 6.63 (dd, J=2.3, J=7.3,
1H, CH), 6.27 (dd, 2.3, J = 7.3, 1H, CH), 5.93 (t, J = 2.3, 1H,
CH), 1.18 (s, 9H, CCH3). Anal. Calcd for C22H22: C, 92.26; H,
7.74. Found: C, 92.84; H, 7.88.
3,6-Di-tert-butyl-9-[(3-tert-butylcyclopenta-1,4-dien-1-yl)di-
phenylmethyl]-9H-fluorene (2a). To a solution of 1a (1.68 g,
5.87 mmol) in Et2O (20 mL) was added at room temperature
a diethyl ether solution of (3,6-di-tert-butylfluorenyl)lithium
(30 mL) prepared from 3,6-di-tert-butylfluorene (1.63 g,
5.85 mmol) and n-butyllithium (2.35 mL of a 2.5 M solution
in hexane, 5.85 mmol) with stirring. The mixture was refluxed
for 7 days, quenched with saturated aqueous NH4Cl (50 mL),
and diluted with diethyl ether (45 mL). The organic layer was
separated, washed with water (2 ꢀ 150 mL), and dried over
CaCl2. The volatiles were removed in vacuo, and the residue was
dissolved in hot MeOH. The solution was cooled to -30 ꢀC, and
a white precipitate formed. The precipitate was collected by
filtration, washed with cold methanol (-30 ꢀC), and dried in
vacuo overnight to give 2a as a white powder (0.70 g, 1.24 mmol,
6,60-Bis(4-chlorophenyl)-3-tert-butyl-5-methylfulvene (1d). To
a solution of 1-methyl-3-tert-butylcyclopentadiene (2.27 g, 16.7
mmol; mixture of isomers) in THF (150 mL) was added
1
21%). H NMR (CDCl3, 300 MHz, 60 ꢀC): δ 7.6-6.8 (br m,
16H), 5.73 (br s, 1H, CH), 5.46 (s, 1H, 9-Flu), 5.40 (br s, 1H,
CH), 2.59 (br s, 2H, CH2), 1.33 (s, 18H, CCH3-Flu), 0.91 (s, 9H,
CCH3-Cp). Anal. Calcd for C43H48: C, 91.43; H, 8.57. Found:
C, 91.88; H, 8.97.
(36) Razavi, A. (Atofina) PCT Int. Appl. WO 00/49029, 1999.
(37) Moore, W. R.; King, B. J. J. Org. Chem. 1971, 36, 1882.

Article
3,6-Di-tert-butyl-9-{(4-tert-butyl-2-methylcyclopenta-1,4-dien-
Organometallics, Vol. 29, No. 21, 2010 5081
of 1f (0.74 g, 2.22 mmol) in Et2O (30 mL) was added at room
temperature a diethyl ether solution of (3,6-di-tert-butylfluor-
enyl)lithium (30 mL) prepared from 3,6-di-tert-butylfluorene
(0.62 g, 2.23 mmol) and n-butyllithium (0.89 mL of a 2.5 M
solution in hexane, 2.23 mmol) with stirring. The mixture was
refluxed for 6 days, quenched with saturated aqueous NH4Cl
(50 mL), and diluted with diethyl ether (50 mL). The organic
layer was separated, washed with water (2 ꢀ 200 mL), and dried
over CaCl2. The volatiles were removed in vacuo. The residue
was washed with MeOH and then with cold pentane (-30 ꢀC) on
a filter and dried in vacuo overnight to give 2f as a white powder
(0.10 g, 0.16 mmol, 7%). The 1H NMR spectrum of this
compound (90 ꢀC, CDCl3, Teflon-valved J. Young NMR tube)
featured extremely broadened resonances and was not informa-
tive. 19F{1H} NMR (CDCl3, 282 MHz, 90 ꢀC): δ -117.3 (br s).
MS-FAB (m-nitrobenzyl alcohol, m/z): 614.3 ([M - H]þ), 337.2
([(p-F-C6H4)2C(3-C4H9-5-CH3-C5H3)]þ), 277.3 ([2,6-(C4H9)2-
C13H6H]þ). Anal. Calcd for C44H48F2: C, 85.95; H, 7.87.
Found: C, 86.09; H, 8.25.
{Ph2C(3,6-tBu2-Flu)(3-tBu-C5H3)}ZrCl2 (3a). To a solution
of 2a (0.80 g, 1.42 mmol) in Et2O (30 mL) was added n-butyl-
lithium (1.10 mL of a 2.5 M solution in hexane, 2.84 mmol) at
0 ꢀC with stirring. After 12 h, anhydrous ZrCl4 (0.33 g, 1.43 mmol)
was added to the reaction flask in the glovebox. This resulting
pink solution was stirred at room temperature overnight. Then
the volatiles were evaporated in vacuo and hexane (ca. 40 mL)
was vacuum-transferred under reduced pressure. The mixture
was filtered and the solvent was removed from the filtrate to give
a pink powder. Hexane (5 mL) was added to the pink powder.
After 1 month at room temperature, a red microcrystalline
powder had formed, which was filtered and dried under vacuum
to give 3a (0.47 g, 0.65 mmol, 46%). 1H NMR (CD2Cl2, 300 MHz,
25 ꢀC): δ 8.16 (m, 2H, Flu), 7.99 (m, 2H, Ph), 7.89 (m, 2H, Ph),
7.47 (m, 2H, Ph), 7.34 (m, 2H, Ph), 7.07 (m, 2H, Flu), 6.35 (m,
2H, Flu), 6.19 (t, J=2.6, 1H, Cp), 5.74 (t, J=2.6, 1H, Cp), 5.58
(t, J=2.6, 1H, Cp), 1.44 (s, 9H, CCH3-Flu), 1.43 (s, 9H, CCH3-
Flu), 1.19 (s, 9H, CCH3-Flu). 13C NMR (CD2Cl2, 75 MHz,
25 ꢀC): δ 150.9, 146.1, 145.8, 145.7, 129.9, 129.6, 129.5, 129.4,
127.9, 127.8, 127.7, 127.6, 127.4, 127.2, 124.0, 123.5, 123.4,
120.7, 120.4, 120.0, 119.9, 115.4, 109.52, 105.84, 101.6, 77.9
(C9-Flu), 58.4 (Ph2C(Cp)(Flu)), 35.6 (CCH3), 33.6 (CCH3), 32.2
(CCH3), 32.1 (CCH3), 30.4 (CCH3). Anal. Calcd for
C43H46Cl2Zr: C, 71.24; H, 6.40. Found: C, 70.55; H, 7.18.
{Ph2C(3,6-tBu2-Flu)(3-tBu-5-Me-C5H2)ZrCl2 (3b). To a so-
lution of 2b (0.84 g, 14.5 mmol) in toluene (50 mL) was added
sec-butyllithium (2.22 mL of a 1.3 M solution in hexane, 14.5
mmol) at -30 ꢀC with stirring. The mixture was warmed to
room temperature. After 12 h, the flask was transferred to a
glovebox, and anhydrous ZrCl4 (0.34 g, 14.51 mmol) was added.
The deep red mixture was stirred at room temperature over-
night. The volatiles were evaporated in vacuo, and hexane (ca.
40 mL) was vacuum-transferred in under reduced pressure. The
mixture was filtered, and the solvent was removed form the
filtrate to give the crude 3b as a reddish pink powder (0.99 g, 13.4
mmol, 93%). Attempts to further purify this solid by recrystal-
lization failed. 1H NMR (CD2Cl2, 300 MHz, 25 ꢀC; crude
product): δ 8.10 (m, 2H), 7.89 (m, 3H), 7.70 (m, 1H), 7.45-
6.88 (m, 9H), 6.25 (d, J=9.2, 1H), 6.14 (d, J=3.0, 1H, Cp), 5.58
(d, J=3.0, 1H, Cp), 1.88 (s, 3H, CH3), 1.42 (s, 18H, CCH3), 1.09
(s, 9H, CCH3). Anal. Calcd for C44H48Cl2Zr: C, 71.51; H, 6.55.
Found: C, 72.15; H, 7.09.
1-yl)[bis(4-chlorophenyl)]methyl}-9H-fluorene (2d). To a solu-
tion of 1d (1.33 g, 3.60 mmol) in Et2O (30 mL) was added at
room temperature a diethyl ether solution of (3,6-di-tert-butyl-
fluorenyl)lithium (30 mL) prepared from 3,6-di-tert-butylfluor-
ene (1.00 g, 3.59 mmol) and n-butyllithium (1.44 mL of a 2.5 M
solution in hexane, 3.59 mmol) with stirring. The mixture was
refluxed for 5 days, quenched with saturated aqueous NH4Cl
(50 mL), and diluted with diethyl ether (50 mL). The organic
layer was separated, washed with water (2 ꢀ 200 mL), and dried
over CaCl2. The volatiles were removed in vacuo. The residue
was washed on a filter with MeOH and then with cold pentane
(-30 ꢀC) and then dried in vacuo overnight to give 2d as a white
powder (1.40 g, 2.16 mmol, 60%). 1H NMR (THF-d8, 300 MHz,
90 ꢀC): δ 7.53 (br s, 2H), 7.40-6.80 (br m, 14H), 6.20 (br s, 1H,
CH), 5.64 (s, 1H, 9-Flu), 2.78 (s, 2H, CH2, CH), 1.36 (s, 3H,
CH3), 1.32 (s, 18H, CCH3), 1.09 (s, 9H, CCH3). MS-FAB (m-
nitrobenzyl alcohol, m/z): 647.3 ([M - H]þ), 369.1 ([(p-Cl-
C6H4)2C(3-C4H9-5-CH3-C5H3)]þ), 277.3 ([2,6-(C4H9)2C13H6H]þ).
Anal. Calcd for C44H48Cl2: C, 81.58; H, 7.47. Found: C, 82.04; H,
7.55.
3,6-Di-tert-butyl-9-{(4-cumyl-2-methylcyclopenta-1,4-dien-1-
yl)[bis(4-chlorophenyl)]methyl}-9H-fluorene (2e). Using a proto-
col similar to that described above for 2d, compound 2e was
obtained from 3,6-di-tert-butylfluorene (4.00 g, 14.4 mmol),
n-butyllithium (5.80 mL of a 2.5 M solution in hexane, 14.4 mmol),
and 1e (6.20 g, 14.4 mmol). A similar workup gave 2e as an off-
1
white powder (8.15 g, 11.5 mmol, 80%). H NMR (CD2Cl2,
200 MHz, 25 ꢀC): δ 8.00-6.30 (br m, 19H), 5.70-5.20 (br m, 2H,
CH and 9-Flu), 3.00-2.00 (br m, 2H, CH2), 1.80-0.80 (br m,
27H, CCH3, CH3). MS-FAB (m-nitrobenzyl alcohol, m/z):
707.3 ([M - H]þ), 431.1 (p-Cl-C6H4)2C(3-C(CH3)2C6H6-5-
CH3-C5H3)]þ), 277.3 (([2,6-(C4H9)2C13H6H]þ). Anal. Calcd
for C49H50Cl2: C, 82.91; H, 7.10. Found: C, 83.01; H, 7.15.
3,6-Di-tert-butyl-9-{(4-tert-butyl-2-methylcyclopenta-1,4-dien-
1-yl)(diphenyl)methyl}-9H-fluorene (2b). In a Schlenk flask con-
taining 2d (2.00 g, 3.09 mmol) dissolved in dry THF (50 mL)
were added under an argon stream NaBPh4 (2.30 g, 6.72 mmol),
KOtBu (1.50 g, 13.4 mmol), (SIPr)Pd(C3H5)Cl (0.018 g,
0.030 mmol), and iPrOH (10 mL) with stirring. The mixture
was kept at 60 ꢀC for 3 h and then filtered through a silica pad.
The filtrate was evaporated, and the residue was recrystallized
from CH2Cl2/MeOH (ca. 1/1 v/v) to give colorless prisms of
2b CH2Cl2. The crystals were dissolved in toluene (3 mL). The
3
solution was evaporated and dried under vacuum overnight to
yield 2b (0.89 g, 1.54 mmol, 50%). 1H NMR (500 MHz, CD2Cl2,
-40 ꢀC; at least three CpH double bond isomers were detected):
δ 8.3-6.5 (m, 16H), 6.0-5.0 (m, 2H, CH and 9-Flu), 3.3-1.8 (m,
2H, CH2), 1.3-1.0 (m, 21H, CCH3 and CH3), 0.8-0.7 (m, 9H,
CCH3). MS-FAB (m-nitrobenzyl alcohol, m/z): 577.4 ([M -
H]þ), 301.2 ([(C6H5)2C(3-C4H9-5-CH3-C5H3)]þ), 277.3
([2,6-(C4H9)2C13H6H]þ). Anal. Calcd for C44H50: C, 91.28; H,
8.71. Found: C, 91.12; H, 9.23.
3,6-Di-tert-butyl-9-{(4-cumyl-2-methylcyclopenta-1,4-dien-1-
yl)diphenylmethyl}-9H-fluorene (2c). This ligand was prepared
using a procedure similar to that described above for 2b, starting
from 2e (6.30 g, 8.88 mmol), NaBPh4 (6.70 g, 19.6 mmol), KOtBu
(4.40 g, 39.2 mmol), (SIPr)Pd(C3H5)Cl (0.052 g, 0.090 mmol),
THF (100 mL), and iPrOH (20 mL). After recrystallization of
the crude material from MeOH and drying under vacuo, pure 2e
was isolated as an off-white powder (4.21 g, 6.57 mmol, 74%).
1H NMR (CD2Cl2, 200 MHz, 25 ꢀC): δ 8.10-6.40 (br m, 21H),
5.90-4.90 (br m, 2H, CH and 9-Flu), 3.00-1.90 (br m, 2H,
CH2), 1.80-0.60 (br m, 27H, CCH3 and CH3). MS-FAB (m-
nitrobenzyl alcohol, m/z): 639.4 ([M - H]þ), 363.2 (C6H5)2C(3-
C(CH3)2C6H6-5-CH3-C5H3)]þ), 277.3 ([2,6-(C4H9)2C13H6H]þ).
Anal. Calcd for C49H52: C, 91.82; H, 8.18. Found: C, 92.13;
H, 8.25.
{Ph2C(3,6-tBu2-Flu)(3-PhCMe2-5-Me-C5H2)}ZrCl2 (3c). Using
a protocol similar to that described above for 2b, compound 2d was
obtained from 2c (0.35 g, 0.55 mmol), sec-butyllithium (0.84 mL
of a 1.3 M solution in hexane, 1.10 mmol), and ZrCl4 (0.13 g,
0.55 mmol). This crude product was recrystallized from hexane
(3 mL) to give pure 3c as a pink powder (0.31 g, 0.39 mmol, 71%).
1H NMR (CD2Cl2, 300 MHz, 25 ꢀC): δ 8.16 (d, J=8.4, 2H), 8.04
(m, 2H), 7.99 (d, J=7.7, 1H), 7.88 (d, J=7.7, 1H), 7.60-7.27 (m,
8H), 7.25-7.12 (m, 4H), 7.11-7.00 (m, 2H), 6.28 (d, J=9.2, 1H),
3,6-Di-tert-butyl-9-{(4-tert-butyl-2-methylcyclopenta-1,4-dien-
1-yl)[bis(4-fluorophenyl)]methyl}-9H-fluorene (2f). To a solution

5082 Organometallics, Vol. 29, No. 21, 2010
Table 3. Summary of Crystal and Refinement Data for Compounds 2a, 2b CH2Cl2, and 3a,c
Kirillov et al.
3
2a
2b CH2Cl2
3
3a
3c
empirical formula
formula wt
temp, K
wavelength, A
cryst syst
C43H48
564.81
173(2)
0.710 73
monoclinic
P21/c
11.9603(3)
15.9036(7)
18.0774(7)
94.048(2)
3430.0(2)
4
C45H52Cl2
663.77
293(2)
0.710 73
triclinic
P1
12.2450(6)
12.8780(6)
15.5643(8)
73.819(3)
2048.59(17)
2
C43H46Cl2Zr
724.92
173(2)
0.710 73
triclinic
P1
12.6849(8)
20.6181(15)
21.5412(16)
77.198(5)
5480.9(7)
6
C49H50Cl2Zr
801.01
100(2)
0.710 73
monoclinic
Cc
22.104(4)
10.2250(16)
35.465(6)
92.746(9)
8006(2)
8
˚
space group
˚
a, A
˚
b, A
˚
c, A
β, deg
3
˚
V, A
Z
density (calcd), Mg/m3
abs coeff, mm-1
cryst size, mm3
no. of rflns collected
no. of indep rflns
max and min transmissn 0.9939 and 0.9939
no. of data/restraints/
params
1.094
0.061
1.076
0.186
1.318
0.476
1.329
0.442
0.10 ꢀ 0.10 ꢀ 0.10
0.45 ꢀ 0.18 ꢀ 0.14
0.15 ꢀ 0.10 ꢀ 0.10
0.4 ꢀ 0.18 ꢀ 0.09
13 897
7679 (R(int) = 0.0424)
40 262
95 753
58 396
9355 (R(int) = 0.0316)
0.974 and 0.920
9355/0/452
21 489 (R(int) = 0.2234)
0.9539 and 0.9320
21 489/0/1243
18 178 (R(int) = 0.0612)
0.961 and 0.830
18 178/2/956
7679/0/388
final R indices (I >2σ(I)) R1 = 0.0707, wR2 = 0.1859 R1 = 0.0648, wR2 = 0.2102 R1 = 0.0808, wR2 = 0.1818 R1 = 0.0490, wR2 = 0.0903
R indices (all data)
goodness of fit on F2
˚
largest diff peak, e A
R1 = 0.1313, wR2 = 0.2157 R1 = 0.0927, wR2 = 0.2273 R1 = 0.1890, wR2 = 0.2275 R1 = 0.0545, wR2 = 0.0926
1.063
1.175
0.906 and -0.344
1.119
0.353 and -0.507
1.13
0.603 and -1.088
-3
1.499 and -1.209
6.19 (d, J = 2.8, 1H, Cp), 5.90 (d, J = 2.8, 1H, Cp), 1.92 (s, 3H,
CH3), 1.76 (s, 3H, CH3), 1.48 (s, 18H, CCH3), 1.35 (s, 3H, CH3).
13C NMR (CD2Cl2, 75 MHz, 25 ꢀC; many signals in the aromatic
region overlapped; resonances from quaternary carbons were not
observed): δ 150.5, 149.7, 142.8, 128.7, 128.0, 126.9, 125.6, 124.0,
123.0, 122.7, 120.6, 120.5, 119.3, 119.2, 105.0, 104.3, 40.0
(PhC(CH3)2), 34.9 (CCH3), 32.8 (CCH3), 31.6 (CCH3), 31.5
(CCH3), 25.9 (PhC(CH3)2), 25.8 (PhC(CH3)2), 20.2 (CH3Cp).
Anal. Calcd for C49H50Cl2Zr: C, 73.47; H, 6.29. Found: C, 73.68;
H, 6.11.
Propylene Polymerization. Polymerizations were performed
in a 300 mL high-pressure glass reactor equipped with a
mechanical stirrer (Pelton turbine) and externally heated with
a double mantle with a circulating water bath. The reactor was
filled with toluene (80 to 150 mL) and MAO (1.5 mL of a 30 wt-%
solution in toluene) and pressurized at 5 atm of propylene (Air
Liquide, 99.99%). The reactor was thermally equilibrated at the
desired temperature for 30 min, the propylene pressure was
decreased to 1 atm, and a solution of the catalyst precursor in
toluene (ca. 2 mL) was added by syringe. The propylene pressure
was immediately increased to 5 atm (kept constant with a back
regulator) and the solution was stirred for the desired time
(typically 30 min). The temperature inside the reactor was
monitored using a thermocouple. The polymerization was
stopped by venting the vessel and quenching with a 10% HCl
solution in methanol (ca. 3 mL). The polymer was precipitated
in methanol (ca. 200 mL), and 35% aqueous HCl (ca. 1 mL) was
added to dissolve possible catalyst residues. The polymer was
collected by filtration, washed with methanol (ca. 200 mL), and
dried under vacuum overnight.
least-squares refinement based on F2 (programs SIR97 and
SHELXL-97).38 Many hydrogen atoms could be found from
the Fourier difference analysis. Carbon-bound hydrogen atoms
were placed at calculated positions and forced to ride on the
attached atom. The hydrogen atom contributions were calcu-
lated but not refined. All non-hydrogen atoms were refined with
anisotropic displacement parameters. Crystals of 2b were found
to contain lattice disordered solvent molecules, which could not
be sufficiently modeled in the refinement cycles. These molecules
were removed using the SQUEEZE procedure39 implemented in
the PLATON package.40 The locations of the largest peaks in
the final difference Fourier map calculation as well as the
magnitude of the residual electron densities were of no chemical
significance. Crystal data and details of data collection and
structure refinement for the different compounds are given in
Table 3. The principal crystallographic data (excluding struc-
ture factors) are available as Supporting Information, as CIF
files.
Acknowledgment. This work was financially supported
by Total Petrochemicals (grant to E.K., J.A.G., and N.
M.). J.-F.C. thanks the Institut Universitaire de France for
a Junior IUF fellowship (2005-2009).
Supporting Information Available: Figures giving representa-
tive 1H and 13C NMR spectra, CIF files giving crystallographic
data for 2a,b and 3a,c, and text and tables giving computational
details, including Cartesian coordinates of computed structures.
This material is available free of charge via the Internet at http://
pubs.acs.org.
Crystal Structure Determination of 2a,b and 3a,c. Diffraction
data were collected at 100 K using a Bruker APEX CCD
diffractometer with graphite-monochromated Mo KR radiation
(38) (a) Sheldrick, G. M. SHELXS-97, Program for the Determina-
€
€
tion of Crystal Structures; University of Gottingen, Gottingen, Germany,
1997. (b) Sheldrick, G. M. SHELXL-97, Program for the Refinement of
˚
(λ = 0.710 73 A). A combination of ω and φ scans was carried
€
€
Crystal Structures; University of Gottingen, Gottingen, Germany, 1997.
(c) Sheldrick, G. M. Acta Crystallogr. 2008, A64, 112.
out to obtain at least a unique data set. The crystal structures
were solved by direct methods; the remaining atoms were loc-
ated from difference Fourier synthesis followed by full-matrix
(39) Van der Sluis, P.; Spek, A. L. Acta Crystallogr. 1990, A46, 194.
(40) Spek, A. L. Acta Crystallogr. 1990, A46, C–34.