Zirconium(IV) benzyl complexes that contain chelating diamido ligands: Synthesis, fluxionality and ethylene polymerization activity

Article abstract of DOI:10.1016/S1381-1169(01)00462-9

The synthesis of chelating diamidodibenzyl complexes of the type Zr(ABAⁿ)(CH₂Ph)₂ [ABA¹ = N, N′-bis(trimethylsilyl)-2-amidobenzylamido (1), ABA² = N, N′-(methyldiphenylsilyl)(trimethylsilyl)-2-amidobenzylamido (2)] is described. The fluxionality of these complexes in solution has been studied by means of variable-temperature NMR. In the presence of one equivalent of a suitable alkyl abstractor, these compounds proved to behave as moderately active ethylene polymerization catalysts.

Full text of DOI:10.1016/S1381-1169(01)00462-9

Journal of Molecular Catalysis A: Chemical 182–183 (2002) 411–417
Zirconium(IV) benzyl complexes that contain chelating
diamido ligands: synthesis, fluxionality and ethylene
polymerization activity
Régis M. Gauvin1, Jacky Kress∗
Laboratoire de Chimie des Métaux de Transition et de Catalyse, UMR 7513 CNRS,
Institut Le Bel, 4 Rue Blaise Pascal, 67000 Strasbourg, France
Received 28 August 2001; accepted 23 October 2001
Abstract
ThesynthesisofchelatingdiamidodibenzylcomplexesofthetypeZr(ABAn )(CH2Ph)2 [ABA1 = N, Nꢀ-bis(trimethylsilyl)-
2-amidobenzylamido (1), ABA2 = N, Nꢀ-(methyldiphenylsilyl)(trimethylsilyl)-2-amidobenzylamido (2)] is described. The
fluxionality of these complexes in solution has been studied by means of variable-temperature NMR. In the presence of one
equivalent of a suitable alkyl abstractor, these compounds proved to behave as moderately active ethylene polymerization
catalysts. © 2002 Elsevier Science B.V. All rights reserved.
Keywords: Dibenzyl complexes; Diamido ligands; Zirconium; Fluxionality; Olefin polymerization
1. Introduction
[5,6]. The synthesis and the fluxionality of a series
of Zr(IV) dichlorides bearing disymmetric diamido
ligands based on the 2-aminobenzylamine framework
have been recently reported [7,8]. In the presence of
excess of MAO, these compounds display moderate
olefin polymerization activity. We now wish to re-
port the synthesis and the fluxional behaviour of two
corresponding dibenzyl derivatives, as well as their
catalytic polymerization properties in the presence
of one equivalent of the classical alkyl abstractors
mentioned above.
In the field of the transition metal-catalysed poly-
merization of olefins, recent attention has focused on
non-metallocene systems [1], and more particularly
on group 4 amido complexes [2–4]. Metal chloride
derivatives have been commonly used as catalyst
precursors, in combination with methylaluminoxane
(MAO) in large excess, but metal-alkyls have also
attracted much attention. These indeed can be ac-
tivated with stoichiometric amounts of co-catalysts
such as B(C6F5)3 or [CPh3][B(C6F5)4], which has
led in some instances to improved performances
2. Experimental
∗
Corresponding author. Tel.: +49-3-90-24-13-32;
fax: +49-3-90-24-13-29.
2.1. General procedures
E-mail address: jkress@chimie.u-strasbg.fr (J. Kress).
1
Present address: Laboratoire de Chimie Organome´tallique de
All experiments were carried out under an inert
atmosphere in a Vacuum Atmosphere drybox or by
Surface, CPE, UMR 9986 CNRS, 69616 Villeurbanne Cedex,
France.
1381-1169/02/$ – see front matter © 2002 Elsevier Science B.V. All rights reserved.
PII: S1381-1169(01)00462-9

412
R.M. Gauvin, J. Kress / Journal of Molecular Catalysis A: Chemical 182–183 (2002) 411–417
Schlenk techniques. Prior to use, solvents were re-
fluxed over an appropriate dehydrating agent [9],
distilled under argon and stored under argon over acti-
vated 4 Å molecular sieves. Deuterated solvents were
dried over activated 4 Å molecular sieves. Polymer-
ization studies were carried out in classical glassware
connected to an ethylene- or propylene-line and the
monomer pressure was controlled and monitored with
use of a mercury column. Polymerization runs were
quenched by addition of methanol. The polymers were
filtered off, washed with acetone and dried under vac-
uum. NMR spectra were recorded on Bruker AC-300
or AM-400 spectrometers. Chemical shifts are given
in ppm, with tetramethylsilane as the reference. Cou-
pling constants are given in Hz. Energy barriers were
calculated using the Eyring equation [10].
128.83 (Ar ABA), 127.92 (Co/m ZrCH2Ph), 123.75
(Cp ZrCH2Ph), 120.82 (Ar ABA), 120.32 (Ar ABA),
63.22 (ZrCH2Ph), 46.90 (CH2 ABA), 1.64 (SiMe3),
1.21 (SiMe3).
Compound 2: 1H NMR (C6D6, 300 MHz, 298 K): δ
7.58 (m, 4H, Ho SiPh), 7.15 (m, 7H, Hm+Hp SiPh+H3
or H6 ABA), 7.07 (t, 4H, Hm ZrCH2Ph), 6.94 (d, 1H,
H3 or H6 ABA), 6.89 (t, 2H, Hp ZrCH2Ph), 6.81 (dt,
1H, H4 or H5 ABA), 6.70 (dt, 1H, H4 or H5 ABA), 6.63
(d, 4H, Ho ZrCH2Ph), 4.68 (s, 2H, CH2 ABA), 2.13 (d,
2JH–H = 10.7 Hz, 2H, ZrCHHꢀPh), 1.46 (d, 2JH–H
=
10.7 Hz, 2H, ZrCHHꢀPh), 0.60 (s, 3H, SiMePh2), 0.01
(s, 9H, SiMe3). 13C NMR (C6D6, 75 MHz, 298 K): δ
151.27 (C2 ABA), 144.34 (Cipso ZrCH2Ph), 136.25
(Co SiPh), 135.62 (Cipso SiPh), 130.63 (Cp SiPh),
129.95 (Co/m ZrCH2Ph), 129.30 (C1 ABA), 128.78
(Cm SiPh), 128.59 (Ar ABA), 128.46 (Ar ABA),
127.33 (Cm/o ZrCH2Ph), 122.93 (Cp ZrCH2Ph +
Ar ABA), 121.63 (Ar ABA), 65.54 (ZrCH2Ph), 47.85
(CH2 ABA), 0.71 (SiMe3), −1.47 (SiMePh2).
2.2. Complexes 1 and 2
To a suspension of 300 mg of Zr(ABA1)Cl2
(7.03 × 10−4 mol) [7] or 500 mg of Zr(ABA2)Cl2
(9.08×10−4 mol) [7] in 15 ml of pentane at −78 ◦C (or
−55 ◦C) were added 1.45 ml (or 1.86 ml) of a 1.0 M
solution of Mg(CH2Ph)Br in diethylether (1.45 ×
3. Results and discussion
10−3 or 1.86 × 10−3 mol, respectively) [ABA1
=
3.1. Synthesis
N, Nꢀ-bis(trimethylsilyl)-2-amidobenzylamido (1),
ABA2 = N, Nꢀ-(methyldiphenylsilyl)(trimethylsilyl)-
2-amidobenzylamido (2)]. The reaction mixture was
let to warm up to room temperature. After overnight
stirring, the white precipitate was centrifuged and the
yellow solution was evaporated to dryness. The turbid
residual oil was then re-dissolved in pentane in order
to separate the remaining magnesium salts. Filtra-
tion and evaporation to dryness afforded 302 mg, or
599 mg, of an orange oil which solidified in the case
of 2 upon prolonged exposure to high vacuum (yield:
ca. 80% for both products).
n
The dibenzyl complexes Zr(ABA )(CH2Ph)2, that
differ in the nature of the anilinic substituent of the
n
ABA ligand, were synthesized by reaction of the cor-
responding dichlorides [7] with two equivalents of the
benzyl Grignard reagent (Fig. 1).
After filtration of magnesium chloride and evapo-
ration of the filtrate to dryness, 1 and 2 were obtained
in good yield as thick orange oils that solidify upon
prolonged exposure to vacuum. Both compounds are
highly soluble in hydrocarbons, and crystallization at-
tempts were unsuccessful. Protonolysis of tetraben-
Compound 1: 1H NMR (CD2Cl2, 300 MHz,
298 K): δ 7.17 (d, 1H, Ar ABA), 7.15 (t, 4H, Hm
ZrCH2Ph), 7.08 (dt, 1H, Ar ABA), 6.97 (t, 2H, Hp
ZrCH2Ph), 6.76 (dt, 1H, Ar ABA), 6.73 (d, 4H, Ho
ZrCH2Ph), 6.71 (d, 1H, Ar ABA), 4.73 (s, 2H, CH2
n
zylzirconium [11] by the diamines (ABA )H2 (n =
2
ABA), 1.88 (d, JH–H = 10.4 Hz, 2H, ZrCHHꢀPh),
2
1.82 (d, JH–H = 10.4 Hz, 2H, ZrCHHꢀPh), 0.07
(s, 9H, ArNSiMe3), 0.03 (s, 9H, CH2NSiMe3).
13C NMR (C7D8, 75 MHz, 298 K): δ 153.09 (C2
ABA), 144.68 (Cipso ZrCH2Ph), 131.09 (Co/m
ZrCH2Ph), 129.98 (C1 ABA), 128.91 (Ar ABA),
Fig. 1. Synthesis of 1 and 2.

R.M. Gauvin, J. Kress / Journal of Molecular Catalysis A: Chemical 182–183 (2002) 411–417
413
Fig. 2. 1H NMR spectrum of 1 at 185 K (CD2Cl2, 300 MHz, Bn = CH2Ph, S = CDHCl2; = organic impurities).
∗
1, 2) [7] also afforded complexes 1 and 2, but only
interacting with the metal centre [12–14]. The
as unseparable mixtures with the known bis-diamido
gate-decoupled 13C NMR spectrum of 2 in CD2Cl2
at 180 K confirms these conclusions. The ␩1- and
␩2-coordination modes lead for the CH2Ph car-
n
species Zr(ABA )2 [7,8]. Obviously, reaction of the
diamines with Zr(CH2Ph)4 first and 1 or 2 in a second
stage occurs at comparable rates.
1
bons of the two benzyl ligands to JC–H coupling
Low-temperature 1H NMR spectra of the two
compounds in CD2Cl2 reveal similar features. The
spectrum of 1 at 185 K is presented in Fig. 2. For both
1 and 2, relative integrations are in agreement with
the expected stoichiometry of one diamido ligand for
two benzyl groups. The latter are non-equivalent. The
three methylenic groups give rise to three AB systems,
the characteristics of which are reported in Table 1. In
both complexes, geminal coupling constants within
the benzylic methylene groups are indicative of the
presence of one ␩1-coordinated (2JH–H = ca. 11 Hz)
and one ␩2-coordinated (2JH–H = ca. 8.5 Hz) benzyl
ligand. A value lower than 10 Hz is indeed commonly
considered as characteristic of a dihapto benzyl lig-
and in which the ipso-carbon of the phenyl ring is
constants of 115 Hz (at 65.43 ppm) and 137 Hz (at
58.10 ppm), respectively, whereas the Cipso signal of
the ␩2-benzyl group appears at higher field than that
of the ␩1-benzyl one (137.4 ppm vs. 148.1 ppm).
It may be noticed further that two types of benzyl
ortho-protons could be localized at 6.77 ppm (poorly
Table 1
1H NMR characteristics of the different AB spin systems shown
2
by 1 and 2 in CD2Cl2 at 185 K (δ in ppm; JH–H in Hz)
1
2
n
CH2 ABA
5.03 + 4.35 (16.7)
2.59 + 2.19 (8.2)
1.18 + 0.42 (11.4)
5.29 + 4.41 (16.4)
1.34 + 0.88 (8.7)
2.56 + 1.28 (11.0)
CH2 ␩2-Bn
CH2 ␩1-Bn

414
R.M. Gauvin, J. Kress / Journal of Molecular Catalysis A: Chemical 182–183 (2002) 411–417
resolved doublet, 2H) and 5.82 ppm (broad peak, 2H)
1
on the low-temperature H NMR spectrum of 2, on
the basis of the variable-temperature studies described
below. The high field shift of the latter is not formally
assignable to the ␩2-benzyl ligand since no such ten-
dency has been established in the literature. It most
probably results from a shielding anisotropy effect of
the methyldiphenylsilyl group. The chemical shift of
the corresponding protons in 1, in which this group is
absent, is indeed unremarkable.
Fig. 3. Possible molecular structures for 1 and 2.
3.2. Fluxionality
These spectroscopic features, together with the
solid-state structures found for related Zr and Ti
complexes [7,8], are consistent with the two diastere-
omeric molecular structures shown in Fig. 3. The
relative position of the ␩1- and ␩2-benzyl ligands
with respect to the puckered 6-membered ring formed
1
Variable-temperature H NMR studies have been
carried out on 1 (Fig. 4) and 2 in CD2Cl2. On grad-
ual rising of the sample temperature, the three AB
systems of both compounds can be seen to coalesce
n
n
into new signals. That due to the ABA methylenic
by the ABA ligands and Zr remains unfortunately
protons merges into a singlet at 4.73 and 4.78 ppm
at room temperature for 1 and 2, respectively. Those
due to the CH2 protons of the two benzyl groups
unknown. The presence of a dihapto benzyl group
emphasizes the unsaturation of the metal centre in
these compounds.
Fig. 4. Variable-temperature 1H NMR spectra of 1 (CD2Cl2, 300 MHz).

R.M. Gauvin, J. Kress / Journal of Molecular Catalysis A: Chemical 182–183 (2002) 411–417
415
Table 2
Coalescence temperature (Tc), chemical shift difference between the coalescing signals (ꢀν) and calculated activation energies (ꢀG=) for
the coalescing systems in 1 and 2
1
2
Tc (K)
ꢀν (Hz)
ꢀG= (kJ mol−1
)
Tc (K)
ꢀν (Hz)
ꢀG= (kJ mol−1
)
CH2 (ABA)
Pair 1 (Bn)
Pair 2 (Bn)
Hortho (Bn)
213
1
204
651
303
–
40.7 0.2
–
40.4 0.3
–
205
211
198
205
2
2
5
2
352
488
160
380
38.2 0.4
38.8 0.4
230 > T c > 215a
215
–
2
38.1
1
38.1 0.4
a The coalescence plateau is masked by the resonances of pair 2.
n
of each compound coalesce pairwise to form a new
doublet of doublets centred at 1.88 and 1.82 ppm for
1 (2JH–H = 10.4 Hz) and 1.96 and 1.25 ppm for 2
(2JH–H = 10.6 Hz). The coalescing pairs were shown
to be 2.59/0.42 ppm (pair 1) and 2.19/1.18 ppm (pair
2) for 1, and 2.56/1.34 ppm (pair 1) and 1.28/0.88 ppm
(pair 2) for 2. Furthermore, raising the temperature
of 2 from 180 to 193 K leads to slight but significant
sharpening of the broad signal at 5.82 ppm, that we
assign to the ortho-protons of one of the benzyl lig-
ands. Further temperature increase causes this peak to
broaden again, before it eventually coalesces with the
ortho-protons signal of the other benzyl group to yield
a sharp doublet centred at 6.48 ppm at room tempera-
ture. The final room-temperature spectra of 1 and 2 in
CD2Cl2 as well as in C6D6 correspond to molecules
with apparent Cs symmetry, i.e. with two equivalent
rocking process of the ABA framework about the
ZrN2 plane already investigated earlier [7] takes also
place in the present benzyl complexes. In this case, it
is combined with the exchange between the ␩1- and
the ␩2-benzyl ligands, as illustrated in Fig. 5, prob-
ably through a bis(␩1-benzyl) transition state (vide
infra). This reversibly inverses the chirality of com-
plexes 1 and 2. The activation energies are close to
those found for the preceding examples [7], in which
an interaction between the zirconium centre and the
n
aryl ipso-carbon of the ABA ligands was shown
to be involved. Its strength is probably similar in 1
and 2. Further, the activation energy is again lower
for the ABA2 than for the ABA1 complex, which
could tentatively be explained by the difference in
steric hindrance between the SiMePh2 and the SiMe3
substituents of the anilinic nitrogen. The more volu-
minous SiMePh2 group would reduce the ꢀG= value
by raising the ground-state energy level.
The second dynamic process that is suspected
for 2 on the basis of the behaviour of the signal at
5.82 ppm between 180 and 193 K may correspond to
the reversible dissociation of the Cipso atom of the
␩2-benzyl ligand from the zirconium centre (Fig. 6).
This would exchange the two ortho-protons of this
n
benzyl groups and a planar ABA ligand.
The activation energies calculated from these data
with use of the Eyring equation [10], along with rele-
vant parameters are presented in Table 2. The narrow
range of values for a given compound shows that only
one dynamic process is operative in each case. The
average energy barriers are 40.7 0.4 kJ mol−1 for
1 and 38.2 0.4 kJ mol−1 for 2. Hence, the dynamic
Fig. 5. Dynamic rocking process in Zr(ABAn )(CH2Ph)2.

416
R.M. Gauvin, J. Kress / Journal of Molecular Catalysis A: Chemical 182–183 (2002) 411–417
Fig. 6. Proposal for the low-energy exchange phenomenon observed for 2.
benzyl group through rotation of the phenyl ring about
the CH2–Ph bond axis, and would be of lower activa-
tion energy than the ABA rocking process.
little modified over 60 min at −80 ◦C (entries 5 and 9),
where the catalytically active cationic species seems
to be fairly stable. Simultaneously, the polymerization
rate also slows down at lower temperature so that the
activity reaches a maximum of 12.5 kg (mol bar h)−1
at −30 ◦C (entries 1–5). Entries 6 and 7 show that
use of a less dissociating solvent such as chloroben-
zene has an inhibiting effect on the catalytic system.
The resulting polymers were not soluble enough for
SEC analysis (dichlorobenzene, 145 ◦C). However,
1H NMR spectra could be recorded that reveal the
quasi-absence of methyl groups in the polymer chains,
confirming the linearity and the large molecular
masses hinted by their poor solubility. Further exper-
iments have shown that these catalyst systems were
inactive, under similar conditions, for the polymer-
ization of ␣-olefins such as propene and hex-1-ene.
n
3.3. Ethylene polymerization activity
The reactivity of complexes 1 and 2 toward olefins
in the presence of one equivalent of an alkyl abstrac-
tor has been studied. Addition of B(C6F5)3 [15] to a
dichloromethane solution of 1 or 2 at −30 ◦C leads
to an immediate colour change from yellow to in-
tense orange. When this reaction was performed on
1 (36.9 ␮mol) under 1 bar of ethylene pressure, solid
polyethylene was produced as a fibrous polymer.
Formation of 20 mg of polymer within 25 min corre-
sponds to the modest activity of 1.3 kg (mol bar h)−1
.
The catalyst system 2/B(C6F5)3 is even less pro-
ductive and yields only traces of polyethylene under
comparable conditions.
Table 3
On the other hand, addition of 1 equiv. of
[CPh3][B(C6F5)4] [16,17] to dichloromethane so-
lutions of 1 has also been shown to afford a cata-
lyst system that is active in the polymerization of
ethylene (Table 3). At 21 ◦C, for instance, exposure
of this mixture (37.2 ␮mol of each) to an atmosphere
of ethylene (1 bar) for 5 min leads to the formation of
12 mg of solid polyethylene (entry 1). The activity of
3.9 kg (mol bar h)−1 is close to the one observed for
the MAO activated complex Zr(ABA1)Cl2 [7], but
remains modest compared with metallocene-based
systems. Following-up monomer consumption shows
moreover that the system is rapidly deactivated.
Slower deactivation is obtained at lower temperature,
as shown by the observation that the activity is halved
in 15 min at −30 ◦C (entries 3 and 6), but is only
Ethylene polymerization in the presence of the catalyst system
1/[CPh3][B(C6F5)4] (conditions: 20 mg (37.2 ␮mol) of 1 in 7 ml
of CH2Cl2; 1 equiv. of [CPh3][B(C6F5)4] in 3 ml of CH2Cl2; 1 bar
ethylene pressure)
Run
number
T (◦C)
t (min)
Yield
(mg)
A (kg
(mol bar h)−1
)
1
2
3
4
5
6
7b
8
9
21
0
5a
5
5
5
5
20
20
60
60
12
31
39
22
4
65
49
151
38
3.9
10.0
12.5
7.1
1.3
5.2
4.0
4.1
1.0
−30
−50
−80
−30
−30
0
−80
a Monomer consumption had stopped after a few minutes.
b Solvent C6H5Cl.

R.M. Gauvin, J. Kress / Journal of Molecular Catalysis A: Chemical 182–183 (2002) 411–417
417
4. Conclusion
[2] C.H. Lee, Y.-H. La, J.W. Park, Organometallics 19 (2000)
344.
[3] R.R. Schrock, A.L. Casado, J.T. Goodman, L.-C. Liang, P.J.
Bonitatebus Jr., W.M. Davis, Organometallics 19 (2000)
5325.
[4] S. Danièle, P.B. Hitchcock, M.F. Lappert, P.G. Merle, J.
Chem. Soc., Dalton Trans. (2001) 13.
[5] J.D. Scollard, D. McConville, N.C. Payne, J.J. Vittal,
Macromolecules 29 (1996) 5241.
[6] J.D. Scollard, D. McConville, J. Am. Chem. Soc. 118 (1996)
10008.
[7] R.M. Gauvin, C. Lorber, R. Choukroun, B. Donnadieu, J.
Kress, Eur. J. Inorg. Chem. 9 (2001) 2337.
[8] Y.-M. Jeon, J. Heo, W.M. Lee, T. Chang, K. Kim,
Organometallics 18 (1999) 4107.
[9] A.J. Gordon, R.A. Ford, The Chemist’s Companion, Wiley,
New York, 1972.
New dibenzyl-zirconium compounds that contain
chelating diamido ligands have been synthesized and
their fluxional behaviour analysed by low-temperature
NMR. In the presence of 1 equiv. of an alkyl abstrac-
tor, these compounds behave as moderately active
catalysts for the polymerization of ethylene. The cat-
alytically active cationic species probably generated
by the reaction of 1 with [CPh3][B(C6F5)4] is able
to polymerize ethylene at −80 ◦C, but is not stable
enough at higher temperatures to produce substan-
tial amounts of polymer. The polyethylenes seem to
consist of linear high molecular weight chains.
[10] H. Günther, La Spectroscopie de RMN, 5th Edition, Masson,
Paris, 1993.
[11] U. Zucchini, E. Albizzati, U. Giannini, J. Organomet. Chem.
26 (1971) 357.
Acknowledgements
[12] S.L. Latesky, A.K. McMullen, G.P. Nicollai, I.P. Rothwell,
Organometallics 4 (1985) 902.
[13] M. Bochmann, S.J. Lancaster, M.B. Hursthouse, K.M. Abdul
Malik, Organometallics 13 (1994) 2235.
[14] S.J. Lancaster, M. Bochmann, Organometallics 12 (1993)
633.
We thank the CNRS and the Ministère de
l’Education Nationale, de la Recherche et de la Tech-
nologie, for funding this work, and late Prof. J.A.
Osborn for his support.
[15] A.G. Massey, A.J. Park, J. Organomet. Chem. 2 (1964)
245.
[16] J.A. Ewen, H.J. Elder, European Patent Application
0-426-637A2 (1990).
References
[1] G.J.P. Britovsek, V.C. Gibson, D.F. Wass, Angew. Chem. Int.
Ed. Engl. 38 (1999) 428 and references therein.
[17] J.C.W. Chien, W.M. Tsai, M.D. Rausch, J. Am. Chem. Soc.
113 (1991) 8570.