Grafting of the zirconium complexes [Zr(η5-C 5H5){NC-amidine}Cl2] and [Zr(η5- C5H5)(NC-NacNac)Cl2] and the study of their behavior in ethylene polymerization

Article abstract of DOI:10.1016/j.molcata.2014.04.017

The synthesis of the new zirconium complex [Zr(η⁵-C ₅H₅){NC-amidine}Cl₂] (1) has been achieved by reaction of the amidine (E)-N-(4-cyanophenyl)-N′-(2,6-diisopropylphenyl) acetimidamide with [Zr(η⁵-C₅H₅)Cl ₃] and the new complex has been characterized by spectroscopic methods. The reaction of one equivalent of B(C₆F₅) ₃ with compound 1 produces the adduct [Zr(η⁵-C ₅H₅){(C₆F₅)₃B-NC-amidine} ]Cl₂ (2). Complexes 1 and 2 have been tested as homogeneous catalysts in olefin polymerization using MAO as cocatalyst. The grafting of 1 and the previously reported complex [Zr(η⁵-C₅H ₅)(NC-NacNac)Cl₂] (3) on to several inorganic solid supports, namely dehydroxylated silica (SiO₂(TEMP)) (S₁), dehydroxylated silica modified with O(SiMe₃)₂ (S ₂) or MAO (S₃), and MgCl₂ (S₄) has been carried out by addition of a freshly prepared solution of 1 or 3 to the corresponding inorganic solid. FT-IR studies of the new materials showed that the links between complexes 1 or 3 and S₁ and S₂ occur through the zirconium center, whereas with S₃ and S₄ the links involve the CN functional group of the ligand. These new materials have been studied as catalysts in olefin polymerization reactions in the presence of MAO. A UV-vis study was carried out in order to identify the minimum amount of MAO necessary to produce the active species. A preliminary study into the activity in olefin polymerizations was also carried out.

Full text of DOI:10.1016/j.molcata.2014.04.017

Journal of Molecular Catalysis A: Chemical 391 (2014) 130–138
Contents lists available at ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
Grafting of the zirconium complexes [Zr(␩5-C5H5){NC-amidine}Cl2]
and [Zr(␩5-C5H5)(NC-NacNac)Cl2] and the study of their behavior in
ethylene polymerization
Alan R. Cabreraa, Elena Villasen˜orb, Francisca Werlingera, Rene S. Rojasa,∗
,
Mauricio Valderramaa, Antonio Antin˜olob,∗, Fernando Carrillo-Hermosillab,
Rafael Férnadez-Galanb
a Departamento de Química Inorgánica, Facultad de Química, Pontificia Universidad Católica de Chile, Casilla 306, Santiago 22, Chile
b Departamento de Química Inorgánica, Orgánica y Bioquímica, Facultad de Ciencias y Tecnologías Químicas, Universidad de Castilla-La Mancha, Campus
Universitario de Ciudad Real, 13071 Ciudad Real, Espa˜na
a r t i c l e i n f o
a b s t r a c t
The synthesis of the new zirconium complex [Zr(␩5-C5H5){NC-amidine}Cl2] (1) has been achieved by
reaction of the amidine (E)-N-(4-cyanophenyl)-Nꢀ-(2,6-diisopropylphenyl)acetimidamide with [Zr(␩5-
C5H5)Cl3] and the new complex has been characterized by spectroscopic methods. The reaction of one
equivalent of B(C6F5)3 with compound 1 produces the adduct [Zr(␩5-C5H5){(C6F5)3B-NC-amidine}]Cl2
(2). Complexes 1 and 2 have been tested as homogeneous catalysts in olefin polymerization using MAO
as cocatalyst.
Article history:
Received 9 December 2013
Received in revised form 13 March 2014
Accepted 11 April 2014
Available online 21 April 2014
Keywords:
The grafting of 1 and the previously reported complex [Zr(␩5-C5H5)(NC-NacNac)Cl2] (3) on to several
inorganic solid supports, namely dehydroxylated silica (SiO2(TEMP)) (S1), dehydroxylated silica modified
with O(SiMe3)2 (S2) or MAO (S3), and MgCl2 (S4) has been carried out by addition of a freshly prepared
solution of 1 or 3 to the corresponding inorganic solid. FT-IR studies of the new materials showed that
the links between complexes 1 or 3 and S1 and S2 occur through the zirconium center, whereas with S3
and S4 the links involve the CN functional group of the ligand. These new materials have been studied
as catalysts in olefin polymerization reactions in the presence of MAO. A UV–vis study was carried out
in order to identify the minimum amount of MAO necessary to produce the active species. A preliminary
study into the activity in olefin polymerizations was also carried out.
Zirconium complexes
Ethylene polymerization
Cyanoamidinate complexes
© 2014 Elsevier B.V. All rights reserved.
1. Introduction
resulting polymers. Compared with conventional Ziegler–Natta
systems, metallocene-based catalysts offer greater versatility
Olefin polymerization is still attracting considerable interest
because the volume of polyolefin produced is increasing contin-
uously due to its versatility for the production of new and very
important materials in society. Control of the electronic and steric
environments of polymerization catalysts has been shown to
be very important in their activity and for the properties of the
and flexibility for both the synthesis and control of polyolefin
structures. The influence of the nature of the metallocene on
catalytic performance and the properties of the final polymer
are the subject of considerable research in the field of applied
organometallic chemistry [1]. In most cases modifications tend to
affect the stereochemistry of the resulting materials (for instance
syndio- vs. isotactic polypropylene) rather than the molecular
weight of the polymer [2]. In contrast, the molecular weights of
polymers obtained on using late transition catalysts are largely
dependent on the steric volume of the ligands, where the presence
of bulky groups in the catalyst leads to the production of high
molecular weight materials [3] in what would otherwise be an
oligomerization process. At the same time the design and synthesis
of half-zirconocene complexes containing amidinate ligands as
olefin polymerization catalysts is a growing area of interest [4].
∗
Corresponding authors at: Departamento de Química Inorgánica, Facultad de
Química, Pontificia Universidad Católica de Chile, Casilla 306, Santiago-22, Chile (R.
Rojas); Departamento de Química Inorgánica, Orgánica y Bioquímica, Facultad de
Ciencias y Tecnologías Químicas, Universidad de Castilla-La Mancha, Avda. Camilo
Jose Cela N 10, 13071 Ciudad Real, Ciudad Real, Spain (A. Antin˜olo);
Tel.: +0034 926295413; fax: +0034 926 295318.
E-mail address: Antonio.Antinolo@uclm.es (A. Antin˜olo).
http://dx.doi.org/10.1016/j.molcata.2014.04.017
1381-1169/© 2014 Elsevier B.V. All rights reserved.

A.R. Cabrera et al. / Journal of Molecular Catalysis A: Chemical 391 (2014) 130–138
131
Recently, some of us reported the design and synthesis of new
2.2. Synthesis and characterization of
[Zr(ꢀ5-C5H5){NC-amidine}Cl2] (1) and
[Zr(ꢀ5-C5H5){(C6F5)3B-NC-amidine}Cl2] (2)
zirconium systems bearing a cyclopentadienyl ring as well as a
bidentate diimine ligand with a functional group located remotely
on the complex backbone [5]. It was shown that the polymerization
activity of these systems could be increased by the action of a
Lewis acid on a functional group located remotely on the complex
backbone.
2.2.1. (1) [Zr(ꢀ5-C5H5){NC-amidine}Cl2]
Nꢀ-(2,6-diisopropylphenyl)-N-(4-cyanophenyl) acetimidamide
[12] (100 mg, 0.32 mmol) and KH (16 mg, 0.39 mmol) were stirred
in THF for 4 h. Gas was evolved during the reaction. The slightly
cloudy mixture was filtered through Celite and added dropwise to
[Zr(␩5-C5H5)Cl3] (84 mg, 0.32 mmol) in THF. The reaction mixture
was stirred for 12 h. The resulting light brown solution was filtered
through Celite. The solvent was extracted and the yellow solid was
washed three times with ether. A pale yellow powder correspond-
The main routes reported in the literature for metallocene
immobilization [6] have been classified by Ribeiro et al. [7]. The
first method involves direct impregnation of the metallocene onto
the support, either with or without prior modification of the
support. Soga et al. [8] prepared a number of supported cata-
lysts in which silica was initially modified by treatment with
Cl2SiMe2, followed by treatment with MAO and final grafting
of [Zr(␩5-C5H5)2Cl2]. Such systems were shown to be active in
the presence of common trialkylaluminium cocatalysts. Moroz
et al. [9] prepared a supported metallocene catalyst by treat-
ing a silylated silica, obtained by the reaction of dehydroxylated
silica with ClSiMe3, with [Zr(␩5-C5H5)2Cl2]. A few years ago
Otero et al. [10] presented results concerning a grafting link
between substituted and unsubstituted metallocenes on the sur-
faces of several modified silicas through the Zr Cl bond or
through a pendant substituent on the cyclopentadienyl unit.
One of the most promising methods for catalyst heterogeniza-
tion involves the reaction between a functional group of the
zirconium-containing molecule and a reactive group on the car-
rier surface. Recently, the synthesis and characterization of a
series of indenyl zirconocenes bearing pendant chains with sili-
con ethers or heterocyclic moieties as donor groups was described
[11].
The work described here is a continuation of our study of
zirconium chemistry with bidentate ligands bearing a cyclopenta-
dienyl ring and a bidentate ligand with a functional group located
remotely on the complex backbone. The presence of the nitrile
group as a functional group increase the catalytic activity of zir-
conium complexes interacting with Lewis acids and can also allow
support it easily onto inorganic solids having such centers, improv-
ing at same time its catalytic activity in olefin polymerization
processes.
The investigation concerns the synthesis and characterization
of a new amidinate zirconium bearing nitrile terminal groups,
complex 1, and its adduct with tris(pentafluorophenyl)borane (2).
We also present new results on the grafting of the half sand-
wich zirconium complexes 1 and the previously reported complex
[Zr(␩5-C5H5)(NC-NacNac)Cl2] (3) [5], both of which are supported
by a ligand bearing nitrile terminal groups, onto commercial par-
tially dehydroxylated silica and silica modified with MAO or silicon
ethers and MgCl2. The influence of the nature of these materi-
als after MAO activation with the lowest possible Al/Zr ratio was
also studied. The surface reactions were monitored by transmission
infrared spectroscopy (FT-IR).
ing to complex 1 was isolated in 55% yield (96 mg, 0.18 mmol). 1
H
NMR (400 MHz, CD2Cl2, 298 K): ␦/ppm = 7.74 (d, J = 8.1 Hz, 2H, NH),
7.06 (t, J = 7.7 Hz, 1H, m-Ph(CN)), 7.19 (d, J = 8.1 Hz, 2H, o-Ph(CN)),
7.13 (m, 2H, m-Ar), 7.12 (m, 1H, p-Ar), 6.20 (s, 5H, Cp), 3.32 (hept,
J = 6.9 Hz, 2H, HCiPr), 1.55 (s, 3H, MeC), 1.35 (d, J = 6.9 Hz, 6H, MeiPr),
1
1.21 (d, J = 6.9 Hz, 6H, MeiPr’). 13C{ H} NMR (100 MHz, CD2Cl2,
298 K): ␦/ppm = 172.7 (Ar–NCN–Ph(F)), 153.9 (i-Ph(CN)), 144.3 (o-Ar),
142.1 (i-Ar), 134.1 (m-Ph(CN)), 126.7 (o-Ph(CN)), 125.2 (p-Ar), 123.9
(m-Ar), 119.5 (C N), 113.8 (Cp), 107.5 (p-Ph(CN)), 28.1 (HCiPr), 25.3
(MeiPr), 24.6 (MeiPr’), 17.1 (MeC). FT-IR (KBr) = 2222 cm−1 (v˜ (C N),
s). Elemental analysis (%): C26H29Cl2N3Zr (M = 545.66 g/mol): cal-
culated C 57.23, H 5.36, N 7.70; found C 57.13, H 5.30, N 7.61%. For
additional 2D NMR data see Supporting information.
2.2.2. (2) [Zr(ꢀ5-C5H5){(C6F5)3B-NC-amidine}Cl2]
1 eq Of tris(pentafluorophenyl)borane (50 mg, 0.098 mmol in
1 mL of toluene) was added to a toluene suspension of 1 (53 mg,
0.098 mmol). The reaction mixture was stirred for 1 h at room tem-
perature, filtered through Celite, and the volatiles were removed
under vacuum and the residue was washed with pentane. Adduct
2 was isolated as a pale yellow solid in 95% yield (98.5 mg,
0.093 mmol). 1H NMR (400 MHz, CD2Cl2, 298 K): ␦/ppm = 7.96 (d,
J = 8.46 Hz, 2H, m-Ph(CN)), 7.25 (d, J = 8.46 Hz, 2H, o-Ph(CN)), 7.15
(m, 3H, m,p-Ar), 6.21 (s, 5H, Cp), 3.24 (hept, J = 6.86 Hz, 2H, HCiPr),
1.69 (s, 3H, MeC), 1.36 (d, J = 6.86 Hz, 6H, MeiPr), 1.22 (d, J = 6.86 Hz,
ꢀ
1
6H, MeiPr ). 13C{ H} NMR (100 MHz, CD2Cl2, 298 K): ␦/ppm = 173.8
1
(
Ar–NCN–Ph(CN)), 159.2 (i-Ph(CN)), 148.7 (dm, JFC ≈ 246 Hz, C6F5),
1
143.8 (o-Ar), 141.5 (i-Ar), 141.1 (dm, JFC ≈ 241 Hz, C6F5), 137.9
1
(dm, JFC ≈ 254 Hz, C6F5), 135.9 (m-Ph(CN)), 126.6 (p-Ar), 126.5
(o-Ph(CN)), 125.9 (br., i-C6F5), 124.2 (m-Ar), 116.3 (C N), 114.0
ꢀ
(Cp), 97.54 (p-Ph(CN)), 28.2 (HCiPr), 25.3 (MeiPr), 24.7 (MeiPr ),
18.0 (MeC). 11B{ H} NMR (192 MHz, CD2Cl2, 298 K): ␦/ppm = –3.3
1
(␯1/2 ∼ 500 Hz). FT-IR (KBr) = 2310 cm−1 (v˜ (C N), s). Elemental
analysis (%): C44H29BCl2F15N3Zr (M = 1057.64 g/mol): calculated C
49.97, H 2.76, N 3.97; found C 50.11, H 2.69, N 3.99%. For additional
2D NMR data see Supporting information.
2.3. Grafting of complexes 1 and 3
The supported catalysts were prepared under an inert atmo-
sphere using Schlenk techniques and a glove-box. A solution of the
zirconium complexes 1 or 3 in toluene (30 mL), to give a theoreti-
cal level of 1% Zr/SiO2, was added to partially dehydroxylated silica,
SiO2(773), S1, (1 g), and the mixture was stirred at 298 K for 16 h.
The slurry was filtered through fritted discs and washed three times
with toluene (20 mL). The resultant solids were dried under vacuum
at 343 K for 16 h. A similar procedure was followed for modified sil-
icas S2 and S3. In the case of S2, SiO2(773) (1 g) was treated with an
excess (5 mL) of the bis(trimethylsilyl)ether and the mixture was
stirred at 333 K for 16 h. The solids were carefully washed with
toluene, dried and the resulting materials were impregnated with
a solution of 1 or 3 as described above. In the case of MAO-modified
silica S3, SiO2(773) (1 g) was treated with 5 mL (to give a theoretical
2. Experimental
2.1. Material
Degussa silica (200 m2/g, according to the supplier) was
dehydroxylated under vacuum (10−2 mm Hg) for 16 h at the
desired temperature [SiO2(763)] and the sample was cooled
and stored under dry nitrogen. [Zr(␩5-C5H5)Cl3], AliBu3 (TIBA),
methylalumoxane (MAO) and B(C6F5)3 were supplied by Aldrich
and were used without further purification. Ethylene was pur-
chased from Matheson Tri-Gas (research grade, 99.99% pure).
Toluene, THF, ether and pentane were distilled from benzophenone
ketyl.

132
A.R. Cabrera et al. / Journal of Molecular Catalysis A: Chemical 391 (2014) 130–138
level of 15% Al/SiO2) of the toluene solution of MAO. The mixture
was stirred at 298 K for 1 h and dried in vacuo. The resulting mate-
rials were impregnated with the zirconium complexes 1 or 3 in
toluene solutions, and the mixtures were stirred for 16 h at 298 K.
The slurries were filtered through fritted discs and washed three
times with toluene (20 mL). The resultant solids were dried under
vacuum at 343 K for 16 h. Finally, in the case of S4, a solution of the
zirconium complexes 1 or 3 in toluene (30 mL) (to give a theoretical
level of 1% Zr/MgCl2) was added to activated MgCl2(523) [9] (1 g)
and the mixture was stirred at 333 K for 16 h. The slurries were fil-
tered through fritted discs and washed three times with toluene
(20 mL). The resultant solids were dried under vacuum at 343 K for
16 h.
Fig. 1. N,N ligands used to stabilize metal centers.
3. Results and discussion
2.4. Characterization
3.1. Synthesis and characterization of compounds 1 and 2
Infrared spectra were recorded on a Bruker Tensor-27 spec-
trophotometer using an infrared cell with CaF2 or KBr windows;
this set-up allowed in situ studies. A total of 32 scans were typically
accumulated for each spectrum (resolution 2 cm−1). The samples
consisted of ca. 20 mg of silica pressed into a 1-cm diameter self-
supported disc. The samples were dehydroxylated at the desired
temperature for 16 h. MAO-modified samples were treated with
a 10% solution of MAO in toluene and dried under vacuum. Zir-
conium complex grafting was performed using a 10−2 M solution
of the zirconocene in toluene. The samples were heated at 333 K
for 1 h, washed with toluene and dried under vacuum at 343 K
until further changes were not observed (sublimation of a slight
excess of complex was observed in some cases). The FT-IR analyses
were repeated on three different samples in order to ensure repro-
ducibility. NMR analyses were performed in CDCl3 or CD2Cl2 using
a Bruker Advance 400 NMR spectrometer. Most NMR assignments
were supported by additional 2D experiments. UV-Vis analyses
were performed using a Shimadzu UV-2501PC spectrophotometer.
The soluble complexes were dissolved in dry toluene and placed
in special cells under dry nitrogen (1.0 cm path length). The solid
samples were suspended in Nujol to form a slurry. The absorption
spectra were recorded between 190 and 900 nm and toluene or
Nujol were used as a reference.
The nacnac class of ligands (see Fig. 1a) has been extensively
used in organometallic chemistry as bidentate N,N ligands for the
stabilization of both early and late transition metal complexes [13].
In this paper we will focus our interest on a related NC-amidine
ligand family (see Fig. 1b) that allows the use of a common ligand
framework, while at the same time introducing groups that can
affect the steric and electronic properties of the ligand [14].
Compound 1 was synthesized by the reaction of the deep-
red ligand solution obtained by deprotonation of the NC-amidine
ligand, previously reported by our group [12], with KH and with
Zr(␩5-C5H5)Cl3 (THF, rt, 4 h) (see Scheme 1).
The 1H NMR characterization of complex 1 is summarized in the
Supporting Information file. The NMR spectrum contains a singlet
at 1.55 ppm (3H-Me) and a singlet at 6.20 (5H-Cp ring). In the 13
C
NMR spectrum, the Cp signal appears at 113.8 ppm and the cyano
carbon (CN) signal at 119.5 ppm. The signal at 107.5 ppm can be
assigned to the carbon that contains the CN group and the imine
carbon signal appears at 172.7 ppm. All of these resonances, espe-
cially the last one, are strongly shielded compared to those of the
free ligand. Similar behavior was previously observed for another
series of zirconium complexes including the (NC-nacnac) ligand
[14].
The polymers were characterized by gel permeation chromatog-
raphy (GPC) on a Waters Alliance GPC 2000 system equipped with
a refractive index detector. Samples were run at 150 ◦C in spec-
trophotometric grade 1,2,4-trichlorobenzene (TCB) stabilized with
BHT (0.5 g BHT/4 L solvent). Molecular weights were calculated
using a universal calibration from narrow polystyrene standards
in the molecular weight range from 580 to 7.5 million g/mol.
Mark–Houwink parameters of a = 0.7 and k = 47.7 were used to cor-
rect for polyethylene.
The reaction between one equivalent of B(C6F5)3 and compound
1 produced the adduct [Zr{(C6F5)3B-NC-amidine}(␩5-C5H5)Cl2] (2)
in 95% yield (see Scheme 1). 1H NMR characterization of the adduct
revealed the characteristic singlets associated with the Cp ring
(6.21 ppm) and the methyl group (1.69 ppm), both of which were
shifted slightly downfield compared to the signals for complex 1.
The 13C NMR spectrum contained the expected signal for the imine
carbon (C N) at 173.8 ppm, the Cp ring at 114.0 ppm, and the CCN
at 97.5 ppm. The CCN resonance is strongly shielded compared to
that for complex 1 (107.5 ppm). We attribute this change to the
inductive effects resulting from the intermolecular coordination of
the Lewis acid B(C6F5)3 with the C N group.
Polymer melting points were measured on a TA Instruments
differential scanning calorimeter (model DSC 2920) at a rate of
10 ◦C/min for two cycles in the temperature range 25–200 ◦C.
The IR spectrum of the borane adduct 2 shows a sharp C
N
stretching band at v˜ = 2310 cm−1 and this is shifted by ꢁv˜ = 88 cm−1
to higher wavenumbers compared to the corresponding band of
complex 1. This shift in wavenumber is expected from the simple
addition of a nitrile to a strong Lewis acid: sharing the nitrogen lone
pair with the boron atom effectively reduces the electronic repul-
sion between nitrogen and carbon to make the C N triple bond in
2 slightly stronger [5].
2.5. Polymerization tests
Polymerizations were carried out in a Parr autoclave reactor
(100 mL), loaded inside a glove-box with the appropriate amount
of the pre-catalyst (1, 2, 3 and 3-S4) and the corresponding MAO
with toluene, such that the final volume of the toluene solution
was 30 mL. The reactor was sealed inside the glove-box and was
attached to an ethylene line. Ethylene was fed continuously into
the reactor to a pressure of 12.5 bar. The pressurized reaction mix-
ture was stirred at the appropriate temperatures. After 10 min, the
ethylene was vented and acetone was added to quench the poly-
merization. The precipitated polymer was collected by filtration
and dried overnight.
3.2. Grafting of complexes 1 and 3
In this study we used four modified inorganic supports, namely
tSiO2 (S1), tSiO2/O(SiMe3)2 (S2), tSiO2/MAO (S3), and MgCl2 (S4), for
the grafting complexes 1 and 3 (Fig. 2, previously reported by our

A.R. Cabrera et al. / Journal of Molecular Catalysis A: Chemical 391 (2014) 130–138
133
Scheme 1. Synthesis of compounds 1 and 2.
Scheme 2. Grafting of complex 3 onto the prepared supports.
Table 1
FT-IR bands assigned to the v˜ (CN) group of complexes of 1-Sx and 3-Sx.
v˜ (CN)
MAO addition
1*
2222
2222
2222
2234
2243
2209
2209
2209
2216
2229
2310**
2230
2230
2234
2243
2261**
2216
2216
2216
2229
1-S1
1-S2
1-S3
1-S4
3*
3-S1
3-S2
3-S3
3-S4
Fig. 2. Complex 3.
group [5]). The grafting of complex 3 is summarized in Scheme 2,
being analogous to grafting of complex 1.
The resulting systems were characterized by FT-IR spectroscopy.
The v˜ (CN) band was used to verify the grafting and identify the
interaction mode of the complex on the surface.
Values in cm−1
.
*
Complexes 1 and 3 as references.
**
Adduct formed by the reaction of 1 or 3 and B(C6F5)3.
The wavenumbers corresponding to the nitrile stretching bands
for 3-Sx and 1-Sx are shown in Table 1 and the bands assigned to
the v˜ (CN) group of complexes 1 and 3 are included as references.
The compounds supported on S1 and S3 were treated with MAO
according to a reported procedure [10] (for more details see Sup-
porting Information). The wavenumbers of the stretching bands of
these systems are also shown in Table 1.
Comparison of the displacement of the corresponding C N band
for the isolated complexes and the S1-supported complexes did
not show any changes, indicating that the coordination does not
occur through this group. Based on results obtained with the same
support and metallocene compounds [10], it can be stated that the
coordination of complexes to S1 may be taking place through a Zr–O
bond after reaction with terminal OH groups and elimination of HCl.

134
A.R. Cabrera et al. / Journal of Molecular Catalysis A: Chemical 391 (2014) 130–138
Al/Zr Ratio
0
100
1,6
1,4
1,2
1,0
0,8
0,6
0,4
500
600
1000
1500
Scheme 3. Formation of the 3-S1-MAO system.
0,2
0,0
Similar behavior was observed after impregnation of complexes 1
and 3 on S2, with the cyanide band appearing at 2209 cm−1, i.e.,
in a similar position to those of the S1 systems. The ClSiMe3 elim-
ination reaction is thermodynamically favoured and leads to the
formation of strong Zr –O bonds. The coordination mode leaves
the C N group free in the complexes and this is available to coordi-
nate with another incoming Lewis acid group. Addition of MAO to
the complexes supported on S1 led to a shift of the cyanide bands
to higher wavenumbers (from 2209 to 2216 cm−1), thus confirming
the hypothesis outlined above (Scheme 3).
After the addition of complexes to the MAO-modified support
(S3, Scheme 2) the v˜ (CN) frequency appeared at 2216 cm−1, which
is quite different to that in the S1 systems but identical to that
found for the S1-MAO systems. This finding indicates that the com-
plexes are coordinated through the nitrile to acidic sites derived
from MAO. The addition of further quantities of MAO did not lead
to changes in the frequency or intensity of this characteristic band.
(see Supporting information Fig. 3).
275
300
325
350
375
400
425
450
475
500
Wavelenght (nm)
Fig. 3. UV–vis absorption spectrum of 3 at room temperature in toluene.
(homogeneous phase) and for supported systems 3-S3 and 3-S4
(heterogeneous phase). The presence of chromophoric aromatic
ligands directly attached to the metal enables the study of these
systems by UV–vis spectroscopy. The formation of various species
upon the incremental addition of MAO was identified by this tech-
nique.
The most interesting absorptions were ligand-to-metal charge
transfer transitions (LMCT), which occur because the ligands are
electron rich and the metal is electron poor. For the same reason,
neither metal-to-ligand charge transfer (MLCT) nor d–d transitions
occur, meaning that the interpretation of the spectra is relatively
straightforward.
In an effort to make this approach feasible for a larger-scale
industrial process by reducing the required amounts of MAO or
other molecular Lewis acids (B(C6F5)3), MgCl2 was used as a sup-
port (S4) [10]. MgCl2 is readily available, has a low-cost and contains
intrinsic Lewis acid sites on its surface to which compounds with a
Lewis base functionality could coordinate [15].
As indicated above, the addition of a solution of complexes 1 or
3 to a suspension of MgCl2 in toluene led to the formation of the
supported systems 1-S4 and 3-S4.
These experiments were carried out in a special cell under
dry nitrogen (1.0 cm path length) using dry toluene for the
homogeneous experiment and Nujol as the dispersant for the het-
erogeneous experiments (see Supporting information).
The UV–vis spectra of complex 3 in toluene (Al/Zr = 0 ratio used
as reference with [3]i = 1.2 × 10−6 M) and those resulting from the
successive additions of MAO in the studied range (100–1500) are
shown in Fig. 3.
The results of this study suggest that the addition of a small
amount of MAO (Al/Zr = 100) leads to the appearance of a new
species with an absorption maximum shifted to lower wavelength
(ꢂmax = 297 nm) in comparison to the band of the free complex
(ꢂmax = 321 nm). A further increase in the Al/Zr ratio brings about
the appearance of an additional band (ꢂmax = 315 nm). Based on the
study of metallocenes [14], the absorption band at 297 nm would
correspond to the dimethylated Zr compound (Scheme 4) and the
band at 315 nm is due to the catalytically active cationic species.
The combined form in which both bands appear accounts for the
balance between the two species (see Scheme 4). Finally, there is
an increase in the intensity of the band at 315 nm and this reaches
a maximum at Al/Zr = 600. This ratio corresponds to the minimum
MAO concentration required to achieve the maximum concentra-
tion of active species in solution.
FT-IR characterization of these systems (Table 1) showed that
the v˜ (CN)bands were shifted by about 20 cm−1 to higher wavenum-
bers in relation to the bands in the free complexes, thus confirming
the coordination of the complex through the nitrile fragment to the
magnesium cation centers. The more marked displacement of the
IR band of the v˜ (CN) fragment with respect to that observed for
S3 systems may be due to the more effective coordination of the
complexes, which is aided by the absence of bulky groups around
the magnesium centers on the surface and in the MAO.
3.3. UV–vis spectrophotometric study
Qualitative ethylene polymerization reactions were carried out
using 3-Sx, 1-Sx and ethylene at a pressure of 1 bar. The results of
these experiments indicated that the most active heterogeneous
systems in the presence of MAO are 3-S3 and 3-S4 and a UV–vis
spectrophotometric study of these complexes was therefore carried
out.
As mentioned above, one of the main reasons for carrying out
research into homogeneous and supported catalytic systems is the
reduction or elimination of the use of MAO as cocatalyst.
UV–vis spectroscopy was employed, in conjunction with our
previous results on metallocenes [11b,c], to identify the minimum
Al/Zr ratio needed to generate the active species for complex 3
In the UV-Vis studies of the supported systems 3-S3 and 3-S4
a
variable range of Al/Zr ratios were investigated with Nujol as the
dispersant. The results are shown in Figs. 4 and 5, respectively.
It can be seen that the generation of cationic species in hetero-
geneous systems does not include the dimethylated intermediates
(absence of an absorption at lower wavelengths) (Scheme 5).
The UV–vis spectra of the 3-S3 and 3-S4 systems show that the
addition of a very small amount of MAO (Al/Zr = 50) does not lead
to the formation of the cationic species. However, at Al/Zr = 100
a shoulder appears at higher wavelengths and this increases in

A.R. Cabrera et al. / Journal of Molecular Catalysis A: Chemical 391 (2014) 130–138
135
Scheme 4. Generation of cationic species for complex 3.
Scheme 5. Generation of cationic species for supported systems.
3-S
Al / Zr Ratio
3-S
3
Al / Zr Ratio
4
0
0
4
3
2
1
0
4
3
2
1
100
300
500
100
300
500
250
300
350
400
450
500
550
250
300
350
400
450
500
550
Wavelenght (nm)
Wavelenght (nm)
Fig. 5. UV–vis absorption spectrum of 3-S4 at room temperature in Nujol.
Fig. 4. UV–vis absorption spectrum of 3-S3 at room temperature in Nujol.
Table 2
Selected homogeneous ethylene polymerization reactions.a
intensity before reaching its maximum absorption at Al/Zr = 300.
This absorption band at ∼403 nm is associated with the active
cationic species.
d
Entry
Precatalyst
[Al]/[Zr] ratio
Tb
Ac
Mw
PDI
1
2
3
4
5
6
7
1
1
1
1
1
2
2
250
250
250
500
125
250
125
40
60
75
75
75
75
75
5210
8005
9782
10047
26
183
108
88
48
186
67
2.3
2.1
2.1
3.5
1.7
2.9
1.9
Comparison of the UV–vis spectra for the homogeneous and het-
erogeneous phases shows that the MAO concentration required to
generate the active species in the heterogeneous phase is half that
required for the homogeneous phase. These results are consistent
with the situation generally observed in heterogeneous catalyst
systems [8].
10112
6819
76
a
Ethylene polymerizations test were carried out in a 100 mL autoclave reac-
tor. The conditions are as follows: P = 12.5 bar; 6 × 10−6 mol of complex in 30 mL
of toluene .The cocatalyst was MAO. Reaction time 10 min.
3.4. Homogeneous ethylene polymerization using compounds 1
and 2
b
Reaction temperature in ◦C.
c
Activity (kgPE[mol(Zr)h]−1).
d
×103 g/mol.
A series of homogeneous ethylene reactions were carried out
using compounds 1 and 2 under various reaction conditions with
MAO as the cocatalyst. The results are summarized in Table 2.
The effect of reaction temperature on complex 1 is shown by
the results in entries 1–3. It can be seen that a significant increase
in activity occurs and this is consistent with an increase in the
propagation rate. Furthermore, the molecular weight decrease
observed on increasing the reaction temperature is due to an
increase in the chain termination process. The effect of MAO con-
centration is shown by the results in entries 3–5 and it can be seen

136
A.R. Cabrera et al. / Journal of Molecular Catalysis A: Chemical 391 (2014) 130–138
Table 3
that the molecular weight decreases and the PDI increases as the
MAO concentration is increased. These changes can be explained
by an increase in the cocatalyst transfer rate. MAO also appears to
have a small effect on the polymerization activity, which increases
slightly on increasing the MAO concentration. It is worth noting
that complex 1 has a very low activity at low MAO concentrations
(Al/Zr = 125, entry 5), probably due to the inability of the coactivator
to activate the metal center.
Selected ethylene polymerization reactions.a
c
Entry
Precatalyst
Al/Zr ratio
Ab
Mw
PDI
1
2
3
4
5
6
3
3
500
250
250
250
250
125
1910
1689
1944
4830
3240
1227
147
147
190
72
189
194
1.8
1.8
5.6*
4
1.8
2.1
3 + B(C6F5)3
3-S4
1-S4
1-S4
Adduct 2 was used with lower amounts of MAO (entries 6–7,
Al/Zr = 250 and 125, respectively) in order to confirm the previ-
ously reported observation that the activity remains constant as
the amount of MAO is decreased. On lowering the Al/Zr ratio to
125 it was found that the adduct had a very high activity (entry
7)—a situation in contrast to that found for complex 1, which is
practically inactive (entry 5) under the same reaction conditions,
probably because now the coactivator useful activates the metal
centre of the catalyst once the CN group of the ligand is blocked by
the B(C6F5)3 group.
a
Ethylene polymerizations test were carried out in a 100 mL autoclave reactor.
The conditions are as follows: P = 12.5 bar; 1.86 × 10−6 mol of complex or 100 mg
of supported catalyst in 30 mL of toluene .The cocatalyst was MAO. Reaction time
10 min.
b
Activity (kgPE[mol(Zr)h]−1).
c
×103 g/mol.
*
Bimodal distribution.
The heterogeneous system with MgCl2 as support was used to
optimize the olefin polymerization process as this material has
intrinsic Lewis acid sites that are capable of interacting with the
nitrile group of the ligand and enhancing the activity, in a simi-
lar manner similar to that seen in entry 3, where complex 3 was
coordinated with the borane derivative. However, in this case the
support is cheaper and is easy to prepare.
In all cases a linear polyethylene was obtained with a small
decrease in melting point (137 2 ◦C).
3.5. Heterogeneous ethylene polymerization
Entries 1 and 2 correspond to the behavior of complex 3 as cat-
alyst in the ethylene polymerization in the homogeneous phase
on varying the concentration of MAO. In fact, on decreasing the
amount of MAO to 50%, the activity decreased by 13% but the
properties of the polymer obtained in both cases are the same
(Mw, PDI and Tf). Entry 3 corresponds to the behavior as cata-
lyst in the homogeneous phase of the interaction product between
3 + B(C6F5)3. In this case the use of Al/Zr = 250 led to the same activ-
ity as previously found for complex 3, but in that case a higher
The heterogenized systems 3-Sx were tested as catalysts for
ethylene polymerization under a pressure of 1 bar. The most effi-
cient system was 3-S4 and this was subsequently studied in greater
depth. In this sense 3 and 3-S4 were studied under the same condi-
tions as previously reported [5] using MAO as cocatalyst. The new
results are presented in Table 3. The behavior of complex 3 and its
adducts 3 + B(C6F5)3 as catalysts has been reported previously by
our group [5].
Fig. 6. SEM images for systems 3-S4 and 3-S4-Polymer.

A.R. Cabrera et al. / Journal of Molecular Catalysis A: Chemical 391 (2014) 130–138
137
Al/Zr of 500 was required–a phenomenon that has been observed
in a previous study [5]. This system also led to a polymer with
higher molecular weight, probably due to the more marked steric
influence around the metallic center caused by the coordination of
tris(pentafluorophenyl)borane.
while for the heterogeneous system it was 200, i.e., a 66% decrease
in the required Al/Zr ratio.
Finally, it was shown that the activity of 3-S4 is twice as high
in the polymerization of ethylene and it requires only half the
cocatalyst in comparison to complex 3. This is the most significant
result because the use of MgCl2 as support not only activates the
system better than the Lewis acid B(C6F5)3, but leads to a reduction
in cost per gram in the synthesis of around 200 times.
The results for the heterogeneous system 3-S4 (entry 4) show
a significant increase in activity from 1689 for complex 3 to 1944
for [3 + B(C6F5)3] and 4830 for the heterogeneous system, i.e. more
than twice the activity of compound 3, but in both cases(entries 3 an
4) the polymers have large PDI values in agreement with the results
previously reported [5e]. However, on using the system 1-S4 (entry
5) the opposite effect to that shown by 3-S4 was observed, with the
activity being practically reduced to 30% of the activity of adduct
2 (entry 6, Table 2) under the same conditions. This difference is
attributed to the structures of the complexes used. Complex 3 has
a more hindered steric environment than complex 1 and this helps
to protect the metal center, whereas in the case of system 3-S4
possible interactions with the acid sites on the support may lead
to catalyst activation. It should be noted that although the activity
at a low Al/Zr ratio (125, entry 6, Table 3) is lower than that shown
by compound 2 under the same conditions, it is still higher than
that shown by compound 1, which is almost inactive. The most
interesting result is the difference in the molecular weights of the
polymers obtained [see entries 6 (Table 2) and 5 (Table 3) or entries
3 and 4(Table 3)]. In these cases a marked difference is observed on
using a boron compound or an inorganic support as the remote
activator for complex 1 (67 kg/mol and 189 kg/mol, respectively)
and for complex 3(190 kg/mol and 72 kg/mol, respectively), a result
that highlights the importance of the support and its effect on the
metal center.
SEM studies were carried out in order to determine the mor-
phology of the 3-S4 system as well as that of the polymer formed
on using this system. It can be seen in Fig. 6 that at 100 ␮m mag-
nification the 3-S4 system corresponds to a solid with irregular
size and shape but at a magnification of 20 ␮m the system con-
sists of aggregates of smaller particles with a regular morphology
that give rise to a compact solid. After the polymerization process
(3-S4-Polymer), the 100 ␮m magnification shows that these parti-
cles have a more regular size than in the case of 3-S4, which is close
to 30 ␮m. Finally, at the 20 ␮m magnification it is evident that these
particles correspond to a spongy material that is consistent with the
formation of the polymer on the agglomerated particles of the 3-S4
system.
Acknowledgments
We gratefully acknowledge the financial support of FONDECYT,
Chile (projects 1100286 and 1130077) and from the Ministerio
de Economía y Competitividad, Spain (Grants Consolider-Ingenio
2010 ORFEO CSD2007-00006 and CTQ2009-09214). A.R. Cabrera
thanks CONICYT for a PhD Fellowship and Doctoral Thesis Support
Grant.
Appendix A. Supplementary data
Supplementary data associated with this article can be
found, in the online version, at http://dx.doi.org/10.1016/j.molcata.
2014.04.017.
References
[1] (a) G.G. Hlatky, Chem. Rev. 100 (2000) 1347;
(b) T.J. Marks, Acc. Chem. Res. 25 (1992) 57;
(c) M.O. Kristen, Top. Catal. 7 (1999) 89;
(d) M. Jezequel, V. Dufaud, M.J. Ruiz-García, F. Carrillo-Hermosilla, U. Neuge-
bauer, G.P. Niccolai, F. Lefebvre, F. Bayard, J. Corker, S. Fiddy, J. Evans, J.-P. Broyer,
J. Malinge, J.-M. Basset, J. Am. Chem. Soc. 123 (2001) 3520;
(e) R. Muelhaupt, Macromol. Chem. Phys. 204 (2003) 289.
[2] (a) W. Kaminsky, A. Funck, H. Hahnsen, Dalton Trans. (2009) 8803;
(b) G.W. Coates, Chem. Rev. 100 (2000) 1223;
(c) W. Kaminsky, J. Chem. Soc., Dalton Trans. (1998) 1413.
[3] (a) S. Mecking, Angew. Chem. Int. Ed. 40 (2001) 534;
(b) L.K. Johnson, C.M. Killian, M. Brookhart, J. Am. Chem. Soc. 117 (1995) 6414;
(c) P. Crewdson, S. Gambarotta, M.C. Djoman, I. Korobkov, R. Duchateau,
Organometallics 24 (2005) 5214;
(d) W. Zhang, W.H. Sun, S. Zhang, J. Hou, K. Wedeking, S. Schultz, R. Fröhlich,
H. Song, Organometallics 25 (2006) 1961;
(e) S.D. Ittel, L.K. Johnson, M. Brookhart, Chem. Rev. 100 (2000) 1169.
[4] (a) C. Averbuj, M.S. Eisen, J. Am. Chem. Soc. 121 (1999) 8755;
(b) C. Averbuj, E. Tish, M.S. Eisen, J. Am. Chem. Soc. 120 (1998) 8640;
(c) R.J. Keaton, L. Koterwas, J.C. Fettinger, L.R. Sita, J. Am. Chem. Soc. 124 (2002)
5932;
(d) D. Kissounko, A. Epshteyn, J.C. Fettinger, L.R. Sita, Organometallics 25 (2006)
531;
(e) M. Sharma, H.S. Yameen, B. Tumanskii, S.-A. Filimon, M. Tamm, M.S. Eisen,
J. Am. Chem. Soc. 134 (2012) 17234;
(f) V. Volkis, E. Nelkenbaum, A. Lisovskii, G. Hasson, R. Semiat, M. Kapon, J. Am.
Chem. Soc. 125 (2003) 2179;
4. Conclusion
(g) R. Vollmerhaus, P. Shao, N.J. Taylor, S. Collins, Organometallics 18 (1999)
2731.
[5] (a) A.R. Cabrera, Y. Schneider, M. Valderrama, R. Fröhlich, G. Kehr, G. Erker, R.S.
Rojas, Organometallics 29 (2010) 6104;
The synthesis and characterization of two new zirconium com-
pounds (complex 1 and adduct 2) is reported. Both of these
compounds are active in the polymerization of ethylene and yield
linear, semicrystalline polyethylene of moderate molecular weight.
It is interesting to note that at low MAO concentrations (Al/Zr = 125
ratio) the adduct is at least two orders of magnitude more active
than the complex. Moreover, comparison of complexes 1 and 3
shows that a four-membered metallocycle (1) is five times more
active in olefin polymerization than the six-membered metallocy-
cle (3).
Complexes 1 and 3 were successfully heterogenized on differ-
ent inorganic supports and, due to the presence of the nitrile group,
there were two different interaction modes with the support. In
cases where the support has Lewis acid sites on its surface, the
complex participates in a Lewis acid-base interaction. However, if
the support does not have such sites, the complex is linked through
(b) S.L. Scott, B.C. Peoples, C. Yung, R.S. Rojas, V. Khanna, H. Sano, T. Suzuki, F.
Shimizu, Chem. Commun. (2008) 4186;
(c) B.M. Boardman, J.M. Valderrama, F. Mun˜oz, G. Wu, G.C. Bazan, R. Rojas,
Organometallics 27 (2008) 1671;
(d) J.D. Azoulay, R.S. Rojas, A.V. Serrano, H. Ohtaki, G.B. Galland, G. Wu, G.C.
Bazan, Angew. Chem. Int. Ed. 48 (2009) 1089;
(e) R.S. Rojas, A.R. Cabrera, B.C. Peoples, K. Spannhoff, M. Valderrama, R. Fröh-
lich, G. Kehr, G. Erker, Dalton Trans. 41 (2012) 1243.
[6] (a) Y. Choi, J.B.P. Soares, Can. J. Chem. Eng. 90 (2012) 646;
(b) J.C. Hicks, B. Mullis, C.W. Jones, J. Am. Chem. Soc. 129 (2007) 8426;
(c) R. Huang, R. Duchateau, C.E. Koning, J.C. Chadwick, Macromolecules 41
(2008) 579;
(d) B.Y. Lee, J.S. Oh, Macromolecules 33 (2000) 3194;
(e) N. Millot, S. Soignier, C.C. Santini, A. Baudouin, J.-M. Basset, J. Am. Chem. Soc.
128 (2006) 9361;
(f) C.P. Nicholas, T.J. Marks, Nano Lett. 7 (2004) 1557.
[7] M.R. Ribeiro, A.D. Deffieux, M.F. Portela, Ind. Eng. Chem. Res. 36 (1997) 1224.
[8] K. Soga, T. Ari, B.H. Hoang, T. Uozumi, Macromol. Rapid Commun. 16 (1995)
905.
[9] B.L. Moroz, N.V. Semikolenova, A.V. Nosov, A. Zakharov, S. Nagy, N.J. O’Reilly, J.
Mol. Catal. A: Chem. 130 (1998) 121.
a Si
O Zr bond. The results of a UV–vis study carried out to deter-
mine the smallest amount of MAO needed to activate the complex
showed an optimum Al/Zr = 600 ratio for the homogeneous system,

138
A.R. Cabrera et al. / Journal of Molecular Catalysis A: Chemical 391 (2014) 130–138
[10] (a) C. Alonso, A. Antin˜olo, F. Carrillo-Hermosilla, P. Carrión, A. Otero, J. Sancho,
E. Villasen˜or, J. Mol. Catal. A: Chem. 220 (2004) 285;
(c) M. Rahim, N.J. Taylor, S. Xin, S. Collins, Organometallics 17 (1998) 1315;
(d) D.P. Song, W.-P. Ye, Y.-X. Wang, J.-Y. Liu, Y.-S. Li, Organometallics 28 (2009)
5697;
(e) R. Vollmerhaus, M. Rahim, R. Tomaszewski, S. Xin, N.J. Taylor, S. Collins,
Organometallics 19 (2000) 2161.
(b) P. Carrión, F. Carrillo-Hermosilla, C. Alonso-Moreno, A. Otero, A. Antin˜olo, J.
Sancho, E. Villasen˜or, J. Mol. Catal. A: Chem. 258 (2006) 236.
[11] (a) C. Alonso-Moreno, A. Antin˜olo, F. Carrillo-Hermosilla, P. Carrion, I. López-
Solera, A. Otero, S. Prashar, J. Sancho, Eur. J. Inorg. Chem. (2005) 2924;
(b) P.J. Aragón, F. Carrillo-Hermosilla, E. Villasen˜or, A. Otero, A. Antin˜olo, A.M.
Rodríguez, Eur. J. Inorg. Chem. (2008) 330.
[14] (a) L. Bourget-Merle, M.F. Lappert, J.R. Severn, Chem. Rev. 102 (2002) 3031;
(b) C. Alonso-Moreno, P. Carrión, F. Carrillo-Hermosilla, A. Antin˜olo, A. Otero,
Appl. Organometal. Chem. 23 (2009) 241;
[12] F. Werlinger, Sintesis y caracterizacion de nuevos complejos de circonio con lig-
andos amidinicos para la polimerizacion de olefinas, in: Undergraduate Thesis,
Pontificia Universidad Católica, Chile, 2013.
[13] (a) I. El-Zoghbi, S. Latreche, F. Schaper, Organometallics 29 (2010) 1551;
(b) S. Latreche, F. Schaper, Organometallics 29 (2010) 2180;
(c) C. Alonso-Moreno, A. Antin˜olo, F. Carrillo-Hermosilla, P. Carrión, A.M.
Rodríguez, A. Otero, J. Sancho, J. Mol. Cat. A: Chem. 261 (2007) 53.
[15] V. Busico, M. Guardasole, R. Cipullo, L. Resconi, G. Morini, Chem. Abstr. 141
(2004) 278050.