Neutral ansa-bis(fluorenyl)silane neodymium borohydrides: Synthesis, structural study and behaviour as catalysts in butadiene-ethylene copolymerisation

Article abstract of DOI:10.1039/c0nj00274g

The reaction of silylene-bridged bis(fluorenyl)dipotassium salts with neodymium tris(borohydride) afforded new neutral ansa-bis(fluorenyl)silane neodymium borohydrides: (Flu₂SiR₂)Nd(BH₄)(THF) [R₂ = Me₂, Et₂, (CH₂)₃, Me(Ph), Flu = C₁₃H₈] that were better characterised and more soluble than the previously described anionic [(Flu₂SiMe ₂)Nd(BH₄)₂]^-. The X-ray structures of three of these complexes were determined, and their solid-state geometrical parameters are very similar, despite the ring strain introduced by the silacyclobutane bridge in [Flu₂Si(CH₂)₃] Nd(BH₄)(THF). The main geometrical features were satisfactorily reproduced by DFT calculations. The catalytic activity of the title complexes in ethylene-butadiene copolymerisation reactions was assessed and compared to that of the reported activity of [(Flu₂SiMe₂)Nd(BH ₄)₂]^- under similar conditions. From these results it can be concluded that the cyclo-copolymerisation of ethylene with butadiene is characteristic of a catalyst featuring silylene-bridged bis(fluorenyl) ligands around neodymium, and appears to be independent of the substituents at the silicon atom.

Full text of DOI:10.1039/c0nj00274g

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/njc | New Journal of Chemistry
Neutral ansa-bis(fluorenyl)silane neodymium borohydrides: synthesis,
structural study and behaviour as catalysts in butadiene–ethylene
copolymerisationw
a
a
b
´
Guillaume Cortial, Xavier-Frederic Le Goff, Magali Bousquie,
a
Christophe Boisson,b Pascal Le Floch,z Franc
¸
ois Nief*a and Julien Thuilliezc
Received (in Montpellier, France) 14th April 2010, Accepted 26th May 2010
DOI: 10.1039/c0nj00274g
The reaction of silylene-bridged bis(fluorenyl)dipotassium salts with neodymium
tris(borohydride) afforded new neutral ansa-bis(fluorenyl)silane neodymium borohydrides:
(Flu2SiR2)Nd(BH4)(THF) [R2 = Me2, Et2, (CH2)3, Me(Ph), Flu = C13H8] that
were better characterised and more soluble than the previously described anionic
[(Flu2SiMe2)Nd(BH4)2]ꢀ. The X-ray structures of three of these complexes were determined,
and their solid-state geometrical parameters are very similar, despite the ring strain
introduced by the silacyclobutane bridge in [Flu2Si(CH2)3]Nd(BH4)(THF). The main
geometrical features were satisfactorily reproduced by DFT calculations. The catalytic
activity of the title complexes in ethylene–butadiene copolymerisation reactions was
assessed and compared to that of the reported activity of [(Flu2SiMe2)Nd(BH4)2]ꢀ
under similar conditions. From these results it can be concluded that the cyclo-
copolymerisation of ethylene with butadiene is characteristic of a catalyst featuring
silylene-bridged bis(fluorenyl) ligands around neodymium, and appears to be independent
of the substituents at the silicon atom.
Introduction
(Flu = C13H8) as well as anionic complexes of type A:
[(Flu2SiMe2)Nd(BH4)2]7ꢀ
which has been characterised
The development of a-olefin–diene copolymers is a very
challenging and thus very active field of research.1 Several of
these polymers have shown promising properties as new
vulcanisable elastomers for the tyre industry because of the
advantageous partial substitution of expensive butadiene by
cheaper ethylene. However, copolymerisation of these two
classes of monomers is not an obvious problem. There are
not so many metal-based systems that are able to catalyse both
the polymerisation of olefins and that of dienes. Among these,
lanthanide-based organometallic complexes are probably the
most promising systems, and over the past few years several
ansa-bis(substituted Cp) neodymium complex-based catalysts
have been developed.2–6
,
(Fig. 1). An interesting feature of the copolymer produced,
when a combination of these new complexes with an alkylating
agent as coactivator was used as a catalyst, was that cyclo-
hexane units were present in the microstructure observed.
Note that a mechanism rationalising the formation of these
cycles has been proposed.4,7
The presence of these rings in the microstructure imparts
beneficial properties to the copolymer in terms of relatively
high Tg values and absence of melt transitions. However, the
overall activity of the catalyst needs to be improved to
meet industrial requirements. Since structure A appeared
promising, we launched a research program aimed at studying
the effects of the substitution pattern of the silylene-substituted
bridge on catalytic activity and polymer microstructure.
Herein, we wish to report on the synthesis of new silylene-
bridged bis(fluorenyl)neodymium complexes as well as on
their respective catalytic activity in the copolymerisation of
ethylene with butadiene.
As a recent addition to this series, dimethylsilylene-bridged
ansa-bis(fluorenyl) neodymium complexes have been prepared,
like the structurally uncharacterised [(Flu2SiMe2)NdCl]
a
´ ´ ´ ´
Laboratoire Heteroelements et Coordination, Ecole Polytechnique,
CNRS, 91128 Palaiseau, France.
E-mail: francois.nief@polytechnique.edu; Fax: +33 (0)169334440
´
Universite de Lyon, CPE Lyon, CNRS UMR 5265 Laboratoire de
Chimie Catalyse Polymeres et Procedes (C2P2), LCPP team-Bat
b
`
´ ´
308F, 43 Bd du 11 novembre 1918, F-69616 Villeurbanne, France
c Manufacture Franc¸aise des Pneumatiques Michelin,
Baˆt. C1-Cataroux, 23 Place des Carmes,
63040 Clermont-Ferrand, France
w Electronic supplementary information (ESI) available: 1H NMR
spectra of 3b and 3d; computational details. CCDC reference numbers
772838 (1a), 772839 (1b), 772840 (1c), 772841 (1d), 772842 (3a), 772843
(3b) and 772844 (3c). For ESI and crystallographic data in CIF or
other electronic format see DOI: 10.1039/c0nj00274g
z Deceased.
Fig. 1 Anionic (type A) and neutral (type B) ansa-bis(fluorenyl)silane
neodymium borohydrides.
ꢁc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2010
2290 | New J. Chem., 2010, 34, 2290–2297

View Article Online
Results and discussion
Synthesis of the complexes
The first part of our study focused on the synthesis and the
structural characterisation of neutral ansa-bis(fluorenyl)-
neodymium complexes of type B (Flu2SiR2)Nd(BH4)(THF)
(Fig. 1). Indeed, these derivatives are expected to be more
easily accessible and also more soluble than their anionic
congeners of type A.
Scheme 2 Synthesis of neutral ansa-bis(fluorenyl)silane neodymium
borohydrides 3a–d.
The synthesis of silyl-substituted bis(fluorene) compounds is
relatively straightforward. In all cases, fluorenyllithium,
which can be obtained on a large scale, was reacted with
dichlorosilanes in THF at low temperature. After standard
workup, compounds 1a–d were obtained in fair to good yields.
Compounds 1a–c were further structurally characterised.
Deprotonation of the fluorene ligands was performed with a
non-nucleophilic base in order to avoid nucleophilic attack at
silicon.8 We found that KN(SiMe3)2 was well suited for this
purpose. Thus, reaction of two equivalents of KN(SiMe3)2
with 1a–d afforded silylene-bridged bis(fluorenyl)dipotassium
salts 2a–d which could be isolated as white powders
(Scheme 1).
complexity is very likely due to the absence of exchange of
the coordinated THF in this solvent.
X-Ray crystal structures and theoretical calculations
In order to look for possible geometrical differences in
complexes 3 due to the substituents on the silylene bridge,
we undertook X-ray crystal analyses of complexes 3a–c (no
suitable crystals of 3d could be grown). The main crystallo-
graphic data are presented in Table 1, whilst Fig. 2 shows
complex 3c as a representative example of compounds 3.
The structures of 3a–c are similar. The neodymium atom is
chelated by the ansa-bis(fluorenyl) ligand. The Nd–C(ring)
bonds are not equal, the shortest distances are Nd1–C1 and
Nd1–C17 (2.65–2.69 A), i.e. between Nd and the five-
membered ring carbon atoms directly bonded to the silylene
bridge, while the longest are between Nd and the other five-
membered ring atoms (C7, C8, C23 and C24: 2.87–2.93 A).
Nonetheless, the neodymium atom can be described as
Z5-bonded to the five-membered rings of the ansa-bis(fluorenyl)
ligand. The borohydride ligand displays the usual tridentate
coordination to neodymium, and the Nd–O bonds (2.44 A) are
also similar for the three complexes.
Neutral neodymium complexes can be synthesised by the
reaction of a suitable neodymium precursor with 2a–d. We
selected neodymium(III) borohydride because of its better
solubility than the chloride derivative. Furthermore, this
precursor is easily available from commercial NdCl3.9 Thus,
reaction of 2a–d with Nd(BH4)3(THF)3 in THF afforded
neutral complexes 3a–d in fair to good yield after toluene
workup (Scheme 2).
These complexes were characterised by proton NMR,
elemental analysis, and X-ray crystal structures (for 3a–c).
Despite the paramagnetism of neodymium(III), the proton
spectra could be fully assigned in most cases. In THF-d8
solution at room temperature, there is a configurational
instability at Nd due to rapid exchange of the coordinated
THF ligand. This is evidenced by the occurrence of only four
resonances for the fluorenyl protons in 3a–c, in which the
silylene bridge is symmetrical, and eight in complex 3d,
in which it is dissymmetrical. In toluene-d8 at room
temperature, no exchange of coordinated THF occurs and
the dissymmetrical configuration of the Nd(BH4)THF core is
retained. This is established by the appearance of the fluorenyl
signals in the spectrum of 3b in toluene-d8, which are now split
into eight resonances. Although we were not able to properly
assign the signals in the spectrum of 3d in toluene-d8, its
Fig. 2 An ORTEP plot (50% ellipsoids) of one molecule of 3c
together with the numbering scheme used. C15 is disordered over
two positions; only one is shown on the picture. Hydrogen atoms
(except those on the borohydride) have been omitted for clarity.
Selected distances (A) and angles (1): Nd1–O1 2.444(4), Nd1–B1
2.602(8), Nd1–C1 2.643(6), Nd1–C17 2.671(5), Nd1–C13 2.727(5),
Nd1–C18 2.743(5), Nd1–C2 2.756(6), Nd1–C29 2.822(5), Nd1–C8
2.855(6), Nd1–C7 2.860(6), Nd1–C23 2.906(5), Nd1–C24 2.931(6),
Nd1–H1B 2.36(8), Nd1–H2B 2.36(8), Nd1–H4B 2.51(8), Si1–C1
1.851(6), Si1–C17 1.859(6), Si1–C16 1.871(6), Si1–C14 1.871(6),
C1–Si1–C17 100.5(3), C16–Si1–C14 80.0(3)
Scheme 1 Synthesis of silylene-bridged bis(fluorenyl)dipotassium
salts 2a–d.
ꢁc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2010 New J. Chem., 2010, 34, 2290–2297 | 2291

View Article Online
Table 1 Main crystallographic and data collection parameters for 3a–c
Compound
3a
3b
3c
1
2
Molecular formula
Molecular weight
C32H34BNdOSi
617.73
C34H38BNdOSiꢂ (C7H8)
C33H34BNdOSi
629.74
691.85
Crystal habit
Crystal dimensions (mm)
Crystal system
Space group
a (A)
b (A)
c (A)
a (1)
Red blocks
0.18 ꢃ 0.18 ꢃ 0.06
Orthorhombic
P212121 (no. 19)
9.400(1)
11.887(1)
24.070(1)
90
90
90
2689.5(4)
4
1.526
1252
1.999
30.01
23520
7719
0.0616
6287
0.1367
Reddish orange needles
0.23 ꢃ 0.14 ꢃ 0.10
Triclinic
Red blocks
0.40 ꢃ 0.30 ꢃ 0.12
Monoclinic
P21/c (no. 14)
8.747(1)
34.629(1)
9.214(1)
90
92.708(1)
90
2787.8(4)
4
1.500
1276
1.930
27.48
22172
6121
0.0330
5494
0.1543
ꢀ
P1 (no. 2)
9.198(1)
11.563(1)
15.325(1)
97.919(1)
93.629(1)
93.853(1)
1606.4(2)
2
1.430
708
1.682
30.05
20537
9365
0.0300
8485
0.0792
0.0275
1.033
b (1)
g (1)
V (A3)
Z
d (g cmꢀ3
F(000)
)
m (cmꢀ1
)
Maximum y
Reflections measured
Unique data
Rint
Reflections used
wR2
R1
0.0380
1.084
ꢀ0.013(18)
0.0529
1.202
n.a.
GoF
Flack parameter
n.a.
The ‘‘bite angle’’ on neodymium can be evaluated by two
parameters: the dihedral angle a between the two mean planes
of the fluorenyl ligands, and the centroid–Nd–centroid angle
(d) around neodymium in the Z5 configuration (Fig. 3). These
two parameters are almost equal for the three complexes (73.81
to 75.61 for a and 1201–1211 for d). Therefore, there are no
significant geometrical differences between complexes 3a–c.
Fig. 3 Angles describing the geometry in complexes 3a–d: a, dihedral
angle between the two mean planes of the fluorenyl ligands; b, out-of-
plane bending angle of the silylene bridge; g, bond angle between the
alkyl or aryl substituents on the silylene bridge; d, centroid–Nd–
centroid angle.
Computational details
Since 4f orbitals are much smaller than 6s and 5d orbitals, it is
generally accepted that they are not involved in chemical
bonding in lanthanide metal complexes. Therefore we adopted
an effective core potential (ECP) for Nd, using the Stuttgart/
Dresden ECP MWB49 basis set. This ECP treats the 5s25p66s0
shells explicitly, whereas the [Kr]4d104f3 core of Nd(III) is fixed.
This type of ECP is often used in calculations of metallocene
derivatives of the lanthanides.10 Geometry optimisations were
performed with the Gaussian0311 set of programs, using the
B3LYP12 hybrid DFT functional with the 6-31G* basis set13,14
and the above-mentioned ECP for Nd.
complexes 3a and 3b and the theoretical model 3d. Indeed, in
3c, the values of a and b are very close to those of complexes
3a, b and d. This situation probably results from the hyper-
valent character of the silicon atom.
Polymerisation results
In the case of group 4 metallocenes, it has been shown that
introduction of substituents at different positions of ligands,
including the bridging unit, can greatly influence the
performance of the catalyst.15 Recently, we showed that the
tert-butyl substitution in positions 2 and 7 of the fluorenyl
ligands modified the cyclisation selectivity in cyclo-
copolymerisation of ethylene and butadiene.7 In the present
study, the effect of substituents of the silylene bridging unit on
catalyst performance was assessed.
Table 2 presents a comparison between experimental and
theoretical data. As can be seen, there is a good correlation
between these data, the only slight divergence being some
values of b (the bending angle between the Si–C1 bond and the
fluorenyl plane). However, this slight distortion (about 10%
on average) can be explained if we consider that the B3LYP
functional is known to slightly overestimate aromatic effects.
Therefore the sp2 character of the carbon atom C1 is more
important, leading to a smaller bending angle.
As previous studies showed that the ate-complex
{(Me2SiFlu2)Nd(m-BH4)[(m-BH4)Li(THF)]}2 (4) in combination
with a dialkylmagnesium compound exhibited a good catalytic
activity in the ethylene–butadiene copolymerisation process,
complexes 3a–d were also evaluated under similar experimental
From these data we conclude that modifying the substituents
at silicon only leads to a weak variation in the whole geometry
of complexes. The most significant example is given by
complex 3c, which features a smaller g angle (79.51) than
ꢁc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2010
2292 | New J. Chem., 2010, 34, 2290–2297

View Article Online
Table 2 Neodymium complexes geometries: X-ray and computational data
Complexes 3a–d: (R1)(R2)SiFlu2Nd(BH4)(THF)
Dihedral angle: a
Bending angle: b
R1–Si–R2 bond angle: g
Si–C1 bond length
R1 = R2 = Me (3a)
Experimental
Calculated
75.61
78.91
13.01
11.61
104.11
102.41
1.87 A
1.89 A
R1 = R2 = Et (3b)
Experimental
Calculated
75.11
78.71
13.11
11.61
105.61
104.21
1.87 A
1.89 A
R1 = R2 = (CH2)3 (3c)
Experimental
Calculated
73.81
79.11
13.61
11.81
79.51
78.81
1.84 A
1.87 A
R1 = Me, R2 = Ph (3d)
Calculated
79.71
11.51
102.81
1.88 A
conditions (see experimental section). n-Butyl-n-octyl-magnesium
(BoMag) was used in a tenfold excess ([Mg]/[Nd] = 10) as
activator. Indeed, dialkylmagnesiums have proven to be efficient
alkylating and transfer agents for lanthanide metallocene
catalysts.6,16,17
C. Boisson et al.4,7 proposed a four-step mechanism for the
cyclisation process involving one molecule of 1,3-butadiene
and two molecules of ethylene (Scheme 3). The first step
involves a 2,1-insertion of one molecule of butadiene to form
an Z3-allylneodymium complex (A). The second and the third
steps, which are similar, rely on the insertion of one molecule
of ethylene to yield an alkyl complex (B), in which the resulting
vinyl unit is coordinated to neodymium. Note that at this stage
an intramolecular insertion of this vinyl unit into the
neodymium–carbon bond would result in the formation of a
cyclobutane unit, which is highly disfavoured for evident
thermodynamic reasons related to the ring strain. Therefore,
the insertion of a second ethylene unit is favoured, yielding an
intermediate (C) in which the vinyl unit is still coordinated to
the neodymium. At this stage the intramolecular insertion is
then kinetically favoured to form an alkyl neodymium
complex featuring a 1,2-cyclohexane ring in the polymer chain.
According to this mechanism, the partial pressure of
butadiene is a critical factor for the formation of cyclic
structure; if a butadiene unit is inserted instead of an ethylene
during the ring-forming process, it would lead to a classical
1,2-vinyl pattern. Run 6 shows that with a lower butadiene
feed, the 1,2 vs. 1,4-addition ratio does not vary, as the trans-
1,4 molar ratio does not change. In contrast, the ring vs. vinyl
ratio increases with lower butadiene weight fraction.
The measured activities of complexes 3a–d, which are
reported in Table 3, are similar to those of the reference
complex 4. All catalysts provided a polymer of high molecular
weight with narrow molecular weight distribution except the
catalyst based on 3c. In that case, the strained 4-membered
ring may not remain intact in the reaction conditions, and this
would lead to several active species. However, the general
bis(flurorenyl) structure of the catalytic species is maintained
since the microstructure of polymers was not modified (run 4,
Table 3 and Table 4).
Microstructure and physical properties
The composition and the microstructures of the obtained
1
copolymers were investigated by 13C and H NMR spectro-
scopy, and the results are reported in Table 4. NMR studies
showed that the copolymers displayed similar microstructures:
they exhibit a high proportion of rings and vinyl groups
resulting from the 2,1-insertion of butadiene, together with
trans-1,4 butadiene and linear ethylene units. A slightly higher
trans-1,4 selectivity was observed for the ate-complex 4 in
comparison with 3a. This kind of microstructure was
previously observed on copolymers obtained with the [4/BoMag]
system.7 The cyclo-copolymerisation of ethylene with butadiene
is therefore characteristic of a catalyst featuring silylene-
bridged bis(fluorenyl) ligands around neodymium and appears
to be independent of the substituents at the silicon atom.
This kind of microstructure has an important impact on
the macroscopic properties. Indeed, the presence of rings
modifies chain flexibility, as inserted rings reduce the number
of degrees of freedom of the polymer chain. DSC analyses
of the obtained copolymers were performed, and the
measured glass-transition temperatures (Tg) are reported in
Table 4. The obtained copolymers present high Tg values
Table 3 Activity in ethylene/butadiene copolymerisation with 3a–d/BoMaga
Run
Compd.
Butadiene in feed (mol%)
n(Nd) (mmol)
Yield (g)
Time (min)
Mn (PDI)c (g molꢀ1
)
Activity (kg molꢀ1 hꢀ1
)
1b
2
3
4
5
6
7
8
9
4
30
30
30
30
30
15
15
15
60
20.0
20.0
19.5
18.7
21.0
16.8
22.0
19.6
19.5
7.5
180
180
173
171
180
180
114
180
180
33 300 (1.68)
32 000 (1.8)
24 200 (1.86)
21 400 (3.44)
33 100 (1.82)
25 900 (1.65)
36 700 (4.10)
20 100 (1.73)
11 400 (3.88)
123.0
119.8
114.4
106.4
100.5
175.4
173.9
133.8
41.4
3a
3b
3c
3d
3b
3c
3d
3d
7.19
6.43
5.67
6.33
8.84
7.27
7.87
2.42
a
b
c
Polymerisation conditions : toluene (200 mL), BoMag 0.2 mmol ([Mg]/[Nd] = 10), 80 1C, 4 bar. According to ref. 7. Mn: number-average
molecular weight; PDI: polydispersity index.
ꢁc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2010 New J. Chem., 2010, 34, 2290–2297 | 2293

View Article Online
Table 4 Microstructure of the copolymers
Run
Compd.
Butadiene in feed (mol%)
n(Nd) (mmol)
Yield (g)
Time (min)
Mn (PDI) (g molꢀ1
)
Activity (kg molꢀ1 hꢀ1
)
1b
2
3
4
5
6
7
8
9
4
30
30
30
30
30
15
15
15
60
20.0
20.0
19.5
18.7
21.0
16.8
22.0
19.6
19.5
7.5
180
180
173
171
180
180
114
180
180
33 300 (1.68)
32 000 (1.8)
24 200 (1.86)
21 400 (3.44)
33 100 (1.82)
25 900 (1.65)
36 700 (4.10)
20 100 (1.73)
11 400 (3.88)
123.0
119.8
114.4
106.4
100.5
175.4
173.9
133.8
41.4
3a
3b
3c
3d
3b
3c
3d
3d
7.19
6.43
5.67
6.33
8.84
7.27
7.87
2.42
a
b
Polymerisation conditions: same as in Table 3. Rings (according to ref. 7):
for 13C. Chemical shifts are expressed in parts per million
(ppm) relative to tetramethylsilane. The microstructure of
polymers was determined by Nuclear Magnetic Resonance
(NMR) using a Bruker DRX 400 spectrometer. Spectra were
measured with a 5mm QNP probe at 363 K. A 2 : 1 volume
mixture of tetrachloroethylene and hexadeuterobenzene was
used as the solvent. Gel Permeation Chromatography (GPC)
samples were prepared in 1,2,4-trichlorobenzene. Solutions
were injected at 150 1C into a Waters Alliance GPCV 2000
instrument equipped with a Waters viscosimeter and 3
Styragel columns (HT 6E and HT 2). The system was
calibrated with polystyrene standards using universal calibration.
Thermal studies were performed by Differential Scanning
Calorimetry (DSC) with a Setaram 131 Calorimeter. Samples
were heated from ꢀ130 1C to 120 1C in order to erase their
thermal past, then cooled to ꢀ130 1C, and heated again at a
rate of 10 1C minꢀ1. Microanalytical data were obtained
Scheme 3 Mechanism of the formation of 1,2-cyclohexane structures
in ethylene–butadiene copolymerisation with 3a–d.
compared to linear polybutadiene (ꢀ106 1C) or polyethylene
(ꢀ125 1C).18
It is known that Tg increases with 1,2-vinyl weight fraction19
in polybutadiene. Based on theoretical and experimental
studies, a linear relation has been proposed:20 Tg = ꢀ106 +
91 ꢃ (vinyl weight fraction). According to this, a Tg of ꢀ45 1C,
similar to that of the obtained copolymers (run 2–5), can be
achieved with a polybutadiene including 67% 1,2-vinyl units.
In comparison, the obtained copolymer for run 3 exhibits
weight fractions of 20% for 1,2-vinyl units. So the presence of
cyclic microstructures and ethylene units in the copolymer also
contributes to the increase of Tg.
through the Centre de Microanalyse de l’Universite
´
de Dijon.
Preparation of the compounds
Synthesis of fluorenyllithium. (FluLi):21 Fluorene (16.6 g,
100 mmol) was dissolved in 80 mL of toluene, and 1.05 eq.
BuLi (2.5 M in hexanes, 42 mL) was added dropwise at 0 1C.
The reaction mixture was warmed to room temperature and
stirred at 80 1C for 2 h, whereupon a yellow solid precipitated.
After cooling down to room temperature, the precipitate was
filtered, washed with toluene and hexane, and dried under
vacuum to afford 13.7 g of FluLi as a yellow powder
Conclusions
A series of neutral silylene bridged ansa-bis(fluorenyl) neodymium
borohydride complexes were prepared and characterised.
In combination with a dialkylmagnesium compound, all
complexes showed high activity in cyclo-copolymerisation of
ethylene with butadiene. Various substituents were introduced
on the silicon atom of the bridging unit, but few influences
were observed either on the crystal structure of organo-
metallic complexes or on the performance of catalysts in
polymerisation.
1
(80 mmol, 80%). H NMR (THF-d8): d = 5.95 (s, 1H), 6.41
(t, 2H, JH–H = 7 Hz), 6.78 (t, 2H, JH–H = 7 Hz), 7.28 (d, 2H,
JH–H=7 Hz), 7.89 (d, 2H, JH–H = 7 Hz).
General procedure (GP1) for the synthesis of the
Experimental section
ansa-bis(fluorenyl)silanes 1a–d
General
Fluorenyllithium was dissolved in THF, and 0.5 eq. of
dichlorosilane was added with stirring at ꢀ80 1C. The resulting
solution was warmed to room temperature. After 18 h, NH4Cl
was added to quench any residual fluorenyllithium. The
solution was evaporated, and the orange solid was redissolved
in CH2Cl2, dried on MgSO4 and purified by filtration on silica
gel. The clear solution was evaporated to a light yellow solid
that was further purified by recrystallisation from petroleum
ether. The product was dried at 60 1C under vacuum
overnight, affording 1a–d as white solids. 1a–d were further
All reactions involving air-sensitive compounds were performed
under a protective atmosphere of dry nitrogen or argon with
dry, oxygen-free solvents. Dichlorodiethylsilane was distilled
before use, and all other commercially available reagents were
obtained from the suppliers and used without further purification.
Nd(BH4)3(THF)3 was prepared as previously described.9
NMR spectra of all synthesised compounds were measured
at ambient temperature (20 1C) with a Bruker Avance
1
300 spectrometer operating at 300 MHz for H and 75.5 MHz
ꢁc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2010
2294 | New J. Chem., 2010, 34, 2290–2297

View Article Online
characterised by X-ray structures; single crystals were
obtained by slow evaporation of CHCl3 solutions at room
temperature.
Si(CH3)2), 6.5 (t, 4H, JH–H = 7 Hz). 6.9 (t, 4H, JH–H = 7 Hz),
7.9 (d, 4H, JH–H = 7 Hz), 8.0 (d, 4H, JH–H = 7 Hz). 13C NMR
(THF-d8): d = 2.6 (CH3–Si), 37.4 (C9), 119.4, 120.3, 125.7,
127.4 (C1–C8), 142.7, 144.1 (C10–C13).
Difluorenyldimethylsilane (1a). 1.63 g of 1a (4.2 mmol, 81%)
was prepared according to GP1 from 1.79 g FluLi (10.4 mmol),
and 0.63 mL Me2SiCl2 (0.67 g, 5.2 mmol) in 50 mL THF.
1H NMR (benzene-d6): d = ꢀ0.56 (s, 6H, Si(CH3)2), 4.18
Dipotassium (diethylbis(9-fluorenyl)silanediide) (2b). 1.2 g of
2b (2.4 mmol, 92%) was prepared according to GP2 from 1.11 g
1b (2.67 mmol) and 1.10 g KN(SiMe3)2 (5.5 mmol) in 20 mL of
toluene. 1H NMR (THF-d8): d = 1.31 (t, 6H, CH3), 1.97
(s, 2H, Si–CH), 7.15 (t, 4H, JH–H = 8 Hz), 7.27 (t, 4H, JH–H
=
8 Hz), 7.38 (d, 4H, JH–H = 8 Hz), 7.80 (d, 4H, JH–H = 8 Hz).
13C NMR (CDCl3): d = ꢀ6.9 (CH3), 40.5 (C9), 120.3, 124.4,
125.8, 126.40 (C1–C8), 140.9, 145.1 (C10–C13). MS (IC of
NH3): 389 [M + 1], 100%.
(q, 4H, CH2), 6.66 (t, 4H, JH–H = 7 Hz), 7.01 (t, 4H, JH–H =
7 Hz), 8.04 (d, 4H, JH–H = 7 Hz), 8.09 (d, 4H, JH–H = 7 Hz).
13C NMR (THF-d8): d = 0.1 (CH2–Si), 5.9 (CH3), 41.7 (C9),
118.1, 122.0, 123.3, 124.8 (C1–C8), 142.6, 146.0 (C10–C13).
Difluorenyldiethylsilane (1b). 1.19 g of 1b (2.8 mmol, 66%)
was prepared according to GP1 from 1.5 g FluLi (8.7 mmol),
and 0.65 mL Et2SiCl2 (0.68 g, 4.3 mmol) in 45 mL THF.
1H NMR (CDCl3): d = 0.38 (m, 10H, CH3–CH2–Si), 3.91
(s, 2H, Si–CH), 7.17 (t, 4H, JH–H = 8 Hz), 7.25 (d, 4H,
Dipotassium (bis(9-fluorenyl)-1,10-silacyclobutanediide) (2c).
410 mg of 2c (0.86 mmol, 85%) was prepared according to
GP2 from 401 mg of 1c (1.0 mmol) and 430 mg of KN(SiMe3)2
(2.16 mmol) in 10 mL of toluene;.1H NMR (THF-d8): d =
7.97 (d, 4H, JH–H = 8 Hz), 7.94 (d, 4H, JH–H = 8 Hz), 6.94
(t, 4H), 6.57 (t, 4H), 2.41 (m, 2H, CH2–CH2–CH2), 1.85 (t, 4H,
CH2–Si) 13C NMR (THF-d8) d = 17.9 (–CH2–Si), 23.0
(CH2–CH2–CH2), 45.3 (C9), 111.5, 119.4, 120.1, 121.5
(C1–C8), 145.7, 140.3 (C10–C13).
J
H–H = 7 Hz), 7.26 (t, 4H, JH–H = 7 Hz), 7.76 (d, 4H, JH–H =
8 Hz). 13C NMR (CDCl3): d = 3.1 (2 Si–CH2), 7.3 (2 –CH3),
38.7 (C9), 120.5, 124.8, 125.9, 126.6 (C1–C8), 141.2, 145.5
(C10–C13). MS (IC, NH3, 70 eV): 417 [M + 1], 100%.
1,1-Difluorenylsilacyclobutane (1c). 1.75 g (4.4 mmol, 57%)
of 1c was prepared according to GP1 from 2.65 g FluLi,
(15.4 mmol) and 0.91 mL of 1,1-dichlorosilacyclobutane
Dipotassium (methylphenylbis(9-fluorenyl)silanediide) (2d).
470 mg of 2c (0.89 mmol, 90%) was prepared according to
GP2 from 450 mg of 1d (1.0 mmol) and 438 mg of KN(SiMe3)2
1
1
(1.08 g, 7.7 mmol) in 60 mL THF. H NMR (CDCl3): d =
(2.2 mmol) in 10 mL toluene. H NMR (THF-d8) d = 1.25
(s, 3H, Si–CH3), 6.55 (d, 4H), 6.75 (t, 4H), 7.22 (m, 3H), 7.68
0.96 (t, 4H, –CH2-Si, JH–H = 8.2 Hz), 1.68 (m, 2H,
CH2–CH2–CH2, JH–H = 8.2 Hz), 3.72 (s, 2H, C9), 7.13 (d,
4H, JH–H = 7 Hz), 7.27 (t, 4H, JH–H = 7 Hz), 7.35 (t, 4H,
(d, 4H), 7.90 (m, 2H), 7.95 (d, 4H).
General procedure (GP3) for the synthesis of the
J
H–H = 7 Hz), 7.83 (d, 4H, JH–H = 7 Hz). 13C NMR (CDCl3):
ansa-bis(fluorenyl)silane neodymium borohydrides 3a–d.
d = 15.0 (–CH2–Si), 16.6 (CH2–CH2–CH2), 40.3 (C9), 120.8,
124.9, 126.4, 127.0 (C1–C8), 141.2, 144.8 (C10–C13). MS (IC,
NH3, 70 eV): 401 [M + 1], 100%.
An equimolecular mixture of dianions 2a–d and
Nd(BH4)3(THF)3 in THF was stirred 1 h at room temperature,
whereupon a white precipitate (KBH4) appeared. It was then
eliminated by centrifugation, and 20 mL of toluene was added
to the dark red solution. After stirring overnight, the solution
was evaporated to a brown solid, which was redissolved in
toluene and centrifuged to eliminate more KBH4. The clear
solution was evaporated to dryness, the residue was washed
with petroleum ether and dried under vacuum to give
complexes 3a–d. 3a–c were further characterised by X-ray
structures; pale-red single crystals were obtained by recrystall-
isation from a toluene–petroleum ether solution at ꢀ20 1C
Difluorenyl(methyl)phenylsilane (1d). 2.67 g of 1d (5.9 mmol,
82%) was prepared according to GP1 from 2.5 g of FluLi
(14.5 mmol) and 1.17 mL dichloro(methyl)phenylsilane (1.38 g,
7.2 mmol) in 70 mL THF. 1H NMR (CDCl3) d = ꢀ0.27
(s, 3H, Si–CH3)), 4.70 (s, 2H, Si–CH), 6.55 (d, 2H, J = 7,4 Hz),
6.92 (t, 2H, J = 7.5 Hz, Flu), 7.16 (t, 1H, J = 7.5 Hz), 7.22
(t, 2H, J = 7.4 Hz), 7.37 (8H), 7.80 (m, 6H). 13C NMR (CDCl3):
d = ꢀ10.07 (CH3–Si), 39.57 (C9), 120.10, 124.39, 125.68,
126.23 (C1–C8), 126.72 (CH), 131.6 (Cipso), 134.03 (CH),
140.99, 141.10, 144.35, 144.39 (C10–C13). MS (IC, NH3,
70 eV) 451 [M + 1], 100%.
[Dimethylbis(9-fluorenyl)silanediyl](borohydrido)(tetrahydro-
furano)neodymium (3a). According to GP3, starting from 464
mg 2a (1.0 mmol) and 404 mg of Nd(BH4)3(THF)3 (1.0 mmol)
in 16 mL of THF, 3a (380 mg) was obtained as a red-brown
General procedure (GP2) for the synthesis of
ansa-bis(fluorenylsilane) dianions 2a–d
1
A solution of 1a–d in toluene was added to a solution of
KN(TMS)2 (2.1 eq.) in toluene at room temperature. The
reaction mixture was heated at 50 1C for 16 h, whereupon
dianions 2a–d precipitated as yellow-orange powders that were
isolated, washed with toluene to remove any remaining
KN(SiMe3)2, and dried under vacuum.
powder. Yield: 0.55 mmol (55%). H NMR (THF-d8): d =
ꢀ2.4 (b, 8H), 3.4 (b, 4H), 5.0 (b, 4H), 13.1 (b, 6H, –CH3). Anal
(%), found: C, 61.07; H: 5.70; calcd. for C32H34BNdOSi =
617.75, C: 62.22, H: 5.55.
[Diethylbis(9-fluorenyl)silanediyl](borohydrido)(tetrahydro-
furano)neodymium (3b). According to GP3, starting from
308 mg of 2b (0.63 mmol) and 250 mg of Nd(BH4)3(THF)3
(0.63 mmol) in 10 mL of THF, 3b (270 mg) was obtained as a
red-brown powder that contained toluene. Anal(%): found, C:
64.95, H: 6.37; calcd. for C34H38BNdOSi = 645.80, C: 63.23,
Dipotassium (dimethylbis(9-fluorenyl)silanediide) (2a). 1.0 g
of 2a (2.2 mmol, 91%) was prepared according to GP2 from
0.92 g of 1a (2.35 mmol) and 0.98 g of KN(SiMe3)2 (5.2 mmol)
in 16 mL of toluene. 1H NMR (THF-d8) d = 0.9 (s, 6H,
ꢁc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2010 New J. Chem., 2010, 34, 2290–2297 | 2295

View Article Online
H: 5.95; calcd. for C34H38BNdOSiꢂ0.5(C7H8) = 696.87, C:
65.10; H, 6.12. Yield (as hemitoluene solvate): 0.39 mmol
(62%). 1H NMR (toluene-d8) d = ꢀ11.8 (b, 4H, THF),
ꢀ10.9 (b, 2H), ꢀ9.9 (b, 2H), ꢀ5.2 (b, 4H, THF), ꢀ5.2
(b, 2H, Flu), ꢀ4.7 (b, 2H), ꢀ4.1 (b, 2H), ꢀ3.8 (b, 2H), 9.8
(b, 3H, CH3), 9.9 (b, 2H), 10.8 (b, 2H, CH2), 11.9 (b, 2H), 13.1
(b, 3H, CH3), 20.6 (b, 2H, CH2). 1H NMR (THF-d8): d = ꢀ6.9
(b, 4H), ꢀ4.3 (b, 4H), 0.9 (b, 4H), 8.4 (b, 4H), 11.4 (b, 6H,
–CH3), 16.2 (b, 4H, –CH2–).
1d, chemical formula: C33H26Si, M: 450.63, crystal system:
triclinic, a (A): 9.3890(1), b (A): 10.038(1), c (A): 12.922(1),
a (1): 92.983(1), b (1): 90.812(1), g (1): 99.215(1), V (A3):
ꢀ
1200.2(2), space group: P1, Z: 2, T: 150.0(1) K, reflections
measured: 13 437, Rint: 0.0221, reflections used: 5451, final R1:
0.0423, wR2 (all data): 0.1176.
Polymerisation procedure
Polymerisation reactions were performed in a 250 mL glass
reactor, stirred with a stainless steel blade. A solution of the
alkylating agent (butyloctylmagnesium; BoMag) in toluene
(200 mL) was transferred via cannula over the Nd complexes
3a–d.The concentration of 3a–d was 100 mmol Lꢀ1 and that of
BoMag was 1 mmol Lꢀ1 (ratio [Mg]/[Nd] = 10). The mixture
was stirred for 1 min, and transferred to the reactor under a
stream of argon. The argon was then pumped out before
introducing the monomers. For low concentrations of
butadiene in the feed (r30 mol%), the reactor was charged
with the gaseous monomers (ethylene–butadiene mixture).
Temperature and pressure were then progressively increased
up to 80 1C and 4 bar respectively and kept constant. For
higher concentrations of butadiene, the latter was dissolved
in the toluene solution and then heated up to 80 1C.
Pure ethylene was finally added in order to obtain the
required composition. At the end of polymerisation,
the mixture was quenched with 1 mL of methanol, then the
polymer was stabilised with 2,6-di-tert-butyl-4-methylphenol,
precipitated by adding 600 mL EtOH and dried under vacuum
at 60 1C. Activities were measured via the mass of polymer
obtained.
[Bis(9-fluorenyl)-1,10-silacyclobutanediyl)](borohydrido)(tetra-
hydrofurano)neodymium (3c). According to GP3, starting from
465 mg (0.85 mmol) of 2c and 345 mg (0.85 mmol) of
Nd(BH4)3(THF)3 in 15 mL of THF, 3c (357 mg) was obtained
as a red-brown powder that contained toluene. Anal (%):
found, C: 64.03, H: 5.84; calcd. for C33H34BNdOSi = 629.76,
C: 62.94, H: 5.44; calcd. for C33H34BNdOSiꢂ0.5(C7H8) =
675.83, C: 64.87; H, 5.67. Yield (as hemitoluene solvate):
0.54 mmol (63%). 1H NMR (THF-d8): d = ꢀ3.0 (b, 4H),
ꢀ2.2 (b, 4H), 2.4 (s, 1.5H, toluene), 7.1 (b, 4H), 7.2 (m, 2.5H,
toluene), 9.9 (b, 2H), 10.8 (b, 2H, CH2–CH2–CH2), 12.0
(b, 4H, Flu), 14.8 (b, 4H, Si–CH2), 28.9 (b, 2H, –CH2–),
99.1 (vb, 4H, Nd–BH4),
[Methylphenylbis(9-fluorenyl)silanediyl](borohydrido)(tetra-
hydrofurano)neodymium (3d). According to GP3, starting from
348 mg (0.66 mmol) of 2d and 266 mg (0.66 mmol) of
Nd(BH4)3THF3 in 10 mL THF, 3d (380 mg) was obtained
as a dark-red powder that contained toluene. Yield (as
1
hemitoluene solvate): 0.51 mmol (78%). H NMR (THF-d8)
d = ꢀ7.9 (b, 2H), ꢀ5.9 (b, 2H), ꢀ5.2 (b, 2H), ꢀ4.8 (b, 2H),
ꢀ3.5 (b, 2H), ꢀ1.4 (b, 2H), 2.3 (s, 1.5H, toluene), 7.3–7.2 (m,
2.5H, toluene), 7.8 (b, 2H), 8.3 (b, 2H), 10.1 (b, 1H, Ph–Si),
11.0 (b, 2H), 16.1 (b, 3H, Si–CH3), 20.7 (b, 2H).
Acknowledgements
This work was financed by CNRS, Ecole Polytechnique, CPE
Lyon, and Manufacture Franc¸ aise des Pneumatiques
X-ray experimental section
MICHELIN.
Crystals for X-ray analysis were immersed in Paratone-N oil in
a drybox and mounted on a fiberglass needle in a cold stream
of dinitrogen. Data collection was performed at 150 K
on a Nonius KappaCCD diffractometer, using graphite-
monochromated Mo-Ka radiation (l = 0.71069 A).
1a, chemical formula: C28H24Si, M: 388.56, crystal system:
orthorhombic, a (A): 18.538(1), b (A): 35.874(1), c (A):
12.751(1), V (A3): 8479.8(8), space group: Aba2, Z: 16, T:
150.0(1) K, reflections measured: 24 164, Rint: 0.0512, reflections
used: 5763, final R1: 0.0404, wR2 (all data): 0.1191, Flack
parameter: 0.04(12).
Notes and references
1 C. Boisson, V. Monteil, J. Thuilliez, R. Spitz, C. Monnet,
M.-F. Llauro, F. Barbotin and P. Robert, Macromol. Symp.,
2005, 226, 17–23.
2 F. Barbotin, V. Monteil, M.-F. Llauro, C. Boisson and R. Spitz,
Macromolecules, 2000, 33, 8521–8523.
3 C. Boisson, V. Monteil, D. Ribour, R. Spitz and F. Barbotin,
Macromol. Chem. Phys., 2003, 204, 1747–1754.
4 V. Monteil, R. Spitz, F. Barbotin and C. Boisson, Macromol.
Chem. Phys., 2004, 205, 737–742.
5 J. Thuilliez, V. Monteil, R. Spitz and C. Boisson, Angew. Chem.,
Int. Ed., 2005, 44, 2593–2596.
1b, chemical formula: C30H28Si, M: 416.61, crystal system:
orthorhombic, a (A): 17.026(1), b (A): 9.952(1), c (A):
13.230(1), V (A3): 2241.7(3), space group: Pbcn, Z: 4, T:
150.0(1) K, Reflections measured: 14 437, Rint: 0.0473, reflections
used: 1797, final R1: 0.0892, wR2 (all data): 0.2523.
6 J. Thuilliez, R. Spitz and C. Boisson, Macromol. Chem. Phys.,
2006, 207, 1727–1731.
7 J. Thuilliez, L. Ricard, F. Nief, F. Boisson and C. Boisson,
Macromolecules, 2009, 42, 3774–3779.
8 L. Silaghi-Dumitrescu, I. Haiduc, J. Escudie
J. Satge, Synth. React. Inorg., Met.-Org., Nano-Met. Chem.,
1995, 25, 575–590.
´
, C. Couret and
´
1c, chemical formula: C29H24Si, M: 400.57, crystal system:
triclinic, a (A): 8.740(1), b (A): 9.839(1), c (A): 13.230(1), a (1):
104.023(1), b (1): 105.022(1), g (1): 102.433(1), V (A3):
9 S. Cendrowski-Guillaume, M. Nierlich, M. Lance and
M. Ephritikhine, Organometallics, 1998, 17, 786–788.
10 L. Maron and O. Eisenstein, J. Am. Chem. Soc., 2001, 123,
1036–1039.
11 GAUSSIAN 03 (Revision C.02), Gaussian, Inc., Wallingford, CT,
2004.
ꢀ
1046.8(2), space group: P1, Z: 2, T: 150.0(1) K, reflections
measured: 7940, Rint: 0.0195, reflections used: 4019, final R1:
0.0485, wR2 (all data): 0.1304.
ꢁc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2010
2296 | New J. Chem., 2010, 34, 2290–2297

View Article Online
12 R. C. Binning, Jr. and L. A. Curtiss, J. Comput. Chem., 1990, 11,
1206–1216.
17 S. Bogaert, T. Chenal, A. Mortreux and J.-F. Carpentier, J. Mol.
Catal. A: Chem., 2002, 190, 207–214.
13 G. A. Petersson, A. Bennett, T. G. Tensfeldt, M. A. Al-Laham,
W. A. Shirley and J. Mantzaris, J. Chem. Phys., 1988, 89,
2193–2218.
18 U. Gaur and B. Wunderlich, Macromolecules, 1980, 13, 445–446.
19 P. Dreyfus, Polymeric Material Encyclopedia, ed. J. C. Salamone,
CRC Press, New York, 1996, vol. 8, p. 5657.
14 G. A. Petersson and M. A. Al-Laham, J. Chem. Phys., 1991, 94,
6081–6090.
20 A. K. Sircar, J. Therm. Anal., 1997, 49, 293–301; W. S. Bahary,
D. I. Sapper and J. H. Lane, Rubber Chem. Technol., 1967, 40,
1529–1543.
15 H. G. Alt and A. Koppl, Chem. Rev., 2000, 100, 1205–1221.
¨
16 J.-F. Pelletier, A. Mortreux, X. Olonde and K. Bujadoux, Angew.
Chem., Int. Ed. Engl., 1996, 35, 1854–1856.
21 C. Fritze, G. Erker and R. Fruhlich, J. Organomet. Chem., 1995,
501, 41–45.
¨
ꢁc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2010 New J. Chem., 2010, 34, 2290–2297 | 2297