Tailoring 2D and 3D molecular sieves structures for polyolefin composites: Do all roads lead to remarkable performances?

Article abstract of DOI:10.1039/c7dt03734a

Multiple synthetic strategies were performed in order to tether a zirconium-based catalyst to the 2D and 3D molecular sieves for olefin polymerizations. The anchoring of fluorene silane to the mesoporous MCM-41 was performed in order to obtain a stable catalyst for olefin polymerization (1@MCM-41). Using spectroscopic methods, this system was shown to have the metal center locked on a face down conformation with the surface. Also, immobilized zirconium complexes have been prepared on three different types of aminopropyl-modified supports (2@magadiite, 2@MCM-41 and 3@MCM-48). The advantage of this latter method of immobilization would be the reduction of the steric effect caused by the support: the catalyst, distant from the surface, is more exposed to the monomer and this situation may lead to an increase in the catalytic activity compared to 1@MCM-41. However, a medium size chain as a spacer between the support and the metallocene is still flexible enough to bend and predisposes the metal center to interact with the support surface; this effect is more evident when the nature of the support is of fixed pore dimensions. These supported catalysts exhibited activity for ethylene polymerization, resulting in linear PEs with high melting temperatures. In order to retain a metallocene assembled as in a homogeneous environment, a multi-step reaction was investigated (4@magadiite) but it led to the leaching of the organic moieties from the surface during catalyst preparation. The best catalytic performance was achieved when homogeneous Oct-amido catalyst (5) was reacted with the surface of magadiite and n-alkyl-AlPO-kan.

Full text of DOI:10.1039/c7dt03734a

Dalton
Transactions
View Article Online
View Journal
PAPER
Tailoring 2D and 3D molecular sieves structures
for polyolefin composites: do all roads lead to
remarkable performances?†
a
Cite this: DOI: 10.1039/c7dt03734a
H. M. Moura,a N. L. Gibbons,‡b S. A. Miller b and H. O. Pastore
*
Multiple synthetic strategies were performed in order to tether a zirconium-based catalyst to the 2D and
3D molecular sieves for olefin polymerizations. The anchoring of fluorene silane to the mesoporous
MCM-41 was performed in order to obtain a stable catalyst for olefin polymerization (1@MCM-41). Using
spectroscopic methods, this system was shown to have the metal center locked on a face down confor-
mation with the surface. Also, immobilized zirconium complexes have been prepared on three different
types of aminopropyl-modified supports (2@magadiite, 2@MCM-41 and 3@MCM-48). The advantage of
this latter method of immobilization would be the reduction of the steric effect caused by the support:
the catalyst, distant from the surface, is more exposed to the monomer and this situation may lead to an
increase in the catalytic activity compared to 1@MCM-41. However, a medium size chain as a spacer
between the support and the metallocene is still flexible enough to bend and predisposes the metal
center to interact with the support surface; this effect is more evident when the nature of the support is
of fixed pore dimensions. These supported catalysts exhibited activity for ethylene polymerization, result-
ing in linear PEs with high melting temperatures. In order to retain a metallocene assembled as in a
homogeneous environment, a multi-step reaction was investigated (4@magadiite) but it led to the leach-
ing of the organic moieties from the surface during catalyst preparation. The best catalytic performance
was achieved when homogeneous Oct-amido catalyst (5) was reacted with the surface of magadiite and
n-alkyl-AlPO-kan.
Received 4th October 2017,
Accepted 19th December 2017
DOI: 10.1039/c7dt03734a
rsc.li/dalton
ported catalysts are employed, they can act as a reinforcement
phase and control the polyolefin molecular architectures.5,6
1. Introduction
Polyolefins such as polypropylene (PP) and polyethylene
The most widely applied method to immobilize homo-
(PE) are the largest contributors to the world-scale polymer geneous olefin polymerization catalysts consists of anchoring
production.1,2 Such polyolefins are currently produced indus- metallocenes onto silica supports.7 Examples of silica-sup-
trially by using metallocene catalysts. Some problems still ported metallocene catalysts include systems prepared by direct
persist such as the control of the polymer morphology with deposition, MAO-pre-treated supports and covalent tethering.
homogeneous systems and the large amounts of cocatalysts However, the direct anchoring of Cp2ZrCl2, for example, onto
needed to achieve the best catalytic activity.3 A breakthrough silica is a method where leaching occurs to a large extent. The
in this field was the use of heterogeneous catalytic systems. second most leachable method is the MAO-pre-treatment prior
Although the activity of the supported systems is generally to the anchoring of the metallocene.8 The leaching of materials
lower compared to the homogeneous systems,3,4 if in situ sup- prepared by direct anchoring or pre-treatment of the support
with MAO can reach values between 37–80%.8–10
A reasonable method to overpass this complicated issue is
the covalent binding of organometallic compounds to the in-
organic support,11 which are known in the literature for their
low percentage of leaching (about 3–7%) by MAO.12 Plenty of
strategies for anchoring of the organometallic complexes via
covalent linkages between the support surface and one of the
ligands of the complex have been described in the literature,
including zirconocenes, ruthenium complexes and transition
metal compounds.11,13–21
aMicro and Mesoporous Molecular Sieves Group, Institute of Chemistry, University of
Campinas, Campinas, Sao Paulo, 13083-861, Brazil.
E-mail: gpmmm@iqm.unicamp.br
bThe George and Josephine Butler Laboratory for Polymer Research, Department of
Chemistry, University of Florida, Gainesville, Florida, 32611-7200, USA
†Electronic supplementary information (ESI) available. See DOI: 10.1039/
c7dt03734a
‡Current address: Intel Corporation, 2501 NW 229th Ave, Hillsboro, OR 97214,
USA.
This journal is © The Royal Society of Chemistry 2018
Dalton Trans.

View Article Online
Paper
Dalton Transactions
With respect to the zirconium complexes, many issues
regarding the tethering of a metallocene to the silica surface,
which can limit the use of supported metallocenes in commer-
cial olefin polymerization plants, have not yet been reasonably
evaluated and explained: (i) What would be the role of the
support in deactivation of the metallocenes during the immobil-
ization reaction? (ii) What is the effect of metallocene interaction
with the support? (iii) Does the morphology of the support affect
the diffusion of the co-catalyst and monomers to the immobi-
lized metallocenes? (iv) What are the consequences of this on
the catalytic activity and properties of the polymers obtained?
Answers to these questions require a systematic and com-
prehensive study allowing a deep investigation of the metallo-
cene, the support and the interaction between them. A com-
prehensive investigation of the influences of 2D and 3D in-
organic structures on the catalytic performance, as well as on
the properties of the metallocene complex, has not been
reported previously from a structural point of view. This is
because the tethering of organometallic compounds to a silica
support is one of the most challenging routes to immobilize
and characterize a catalyst system. There are two general
routes that are employed: (i) building-up the ligand and sub-
sequently the metal complex at the surface of the support or
(ii) tethering a pre-synthesized catalyst. In the present work,
our efforts focus on the understanding of both strategies in
relation to the 2D and 3D inorganic structures as the support
as well as the organic moieties in the catalytic system.
Chart 1 Synthesis of 1@MCM-41. RvOH, OEt, O–Siu.
room temperature under N2. After, the solvent was removed by
vacuum and zirconium tetrachloride was added, followed by
the condensation of DEE at −78 °C. The reaction was per-
formed under N2 for another 12 h at room temperature. The
solid was collected after filtration and rinsed with DEE. After
drying under vacuum, a yellowish powder was obtained and
collected inside of a glovebox and named 1@MCM-41.
2.2.2. Synthesis of 2 and 3. The synthesis of 12-bromo-octa-
methyloctahydrodibenzofluorene (C29H37Br) was performed by
first reacting 2,5-dimethyl-2,5-hexanediol with concentrated
hydrochloric acid to isolate 2,5-dichloro-2,5-dimethylhexane. It
was then added to 9-bromofluorene via a double cyclo Friedel–
Crafts alkylation (Chart 2). The final solid was recrystallized
with hexane.
Then, the initial step for preparing 2 and 3 over the sup-
ports was the grafting of aminopropyl groups on the surface of
the supports. A suspension of the dry support in toluene
(10 wt%) was stirred under N2 for 30 min.
Aminopropyltriethoxysilane (Si(OCH2CH3)3(CH2)3NH2) was
dispersed in dried toluene and the solution was transferred to
the support suspension. The reaction was stirred for 24 h
under reflux at 65–70 °C. The solid was recovered by filtration
and washed several times with toluene and dried under
vacuum overnight at 60 °C. A solution containing 9-bromo-
2. Experimental
2.1. Synthesis of the inorganic supports
The synthesis of MCM-41,22 Na-magadiite23 and n-alkyl-AlPO-
kanemite (AlPO3(OH)2[CH3(CH2)11(NH2)]0.43[CH3(CH2)3NH2]0.57
=
43dod,57but-AlPO-kan)24 was performed by following pro-
cedures previously developed by our group. For MCM-48
synthesis, the procedure reported by Castruita et al.25 was
followed. All the solids were pre-treated at 150–200 °C under
vacuum (∼10−5 mbar) for 12
h
prior to the organic
fluorene
(from
Aldrich)
or
12-bromo-octamethyl-
octahydrodibenzofluorene (Chart 2) in tetrahydrofuran (THF)
was added to the NH2-modified support and the system was
kept under stirring for 24 h. Then N,N-diisopropylethylamine
(DIPEA) was added to remove the HBr byproduct. After, the
solid was recovered by centrifugation and rinsed several times
with THF and dried under vacuum overnight.
modifications.
2.2. Synthesis of the supported catalysts
All the air sensitive procedures were performed under a N2
atmosphere using Schlenk lines and glovebox techniques.
2.2.1. Synthesis of 1. The catalyst was prepared by grafting
of triethoxy(fluoren-9-yl)silane on the surface of the inorganic
support (Chart 1). The silane was prepared by the Grignard
reaction. In parallel, the dry inorganic support, MCM-41, was
suspended in toluene and stirred for 30 min under N2. In par-
allel, a solution of the silane in toluene was prepared and
added to the support suspension and stirred under reflux for
another 24 h. The final white solid was rinsed with toluene
several times and dried under vacuum overnight at 60 °C. In
the next step, diethyl ether (DEE) was condensed at −78 °C
and n-butyllithium (1.6 M) was slowly syringed in it. The
mixture was maintained under magnetic stirring for 12 h at Chart 2 Synthesis of (C29H37Br).
Dalton Trans.
This journal is © The Royal Society of Chemistry 2018

View Article Online
Dalton Transactions
Paper
The addition of Zr was performed using a Swivel Frit appar-
atus. The Fl-NH2-solids were previously treated at 70 °C under
vacuum overnight to remove gases and solvent molecules poss-
ibly adsorbed. DEE was condensed at −78 °C and n-butyl-
lithium (1.6 M) was slowly syringed in it. The mixture was
maintained under magnetic stirring for 12 h at room tempera-
ture under N2. After, the solvent was removed by vacuum and
zirconium tetrachloride was added, followed by the conden-
sation of DEE at −78 °C. The reaction was performed under N2
for another 12 h at room temperature. The solid was collected
after filtration and rinsed with DEE. After drying under
vacuum, a yellowish powder was obtained and collected inside
Chart 5 Anchoring of catalyst 5 to the inorganic supports.
nitrogen. The solid was rinsed and the Friedel–Crafts acylation
reaction with benzoyl chloride in the presence of nitromethane
for 16 h at room temperature was performed. The benzophe-
none-magadiite was rinsed with nitromethane and then
reacted with cyclopentadienyl in the presence of sodium meth-
oxide and ethanol under nitrogen for 48 h at r.t. The rinsed
solid was reacted with fluorenyllithium for another 48 h at r.t.
For the incorporation of Zr, the solid was treated with n-butyl-
lithium in diethyl ether (DEE) for 12 h under N2 at r.t. The
solvent was removed by vacuum and ZrCl4 was added to the
system and DEE was condensed at −78 °C. The mixture was
warmed up to r.t and stirred overnight under N2. The solid was
rinsed with DEE and dried under vacuum.
of
a glovebox. The solids were named 2@magadiite,
2@MCM-48 and 3@MCM-41 (Chart 3).
2.2.3. Synthesis of 4. Catalyst 4 was synthesized by a multi-
step procedure as described below (Chart 4). CTA-magadiite
(magadiite swollen with cetyltrimethylammonium bromide)26
was heated at 100 °C under vacuum prior to the organic modi-
fications. Phenyltriethoxysilane was then grafted to the CTA-
magadiite in the presence of refluxing toluene for 72 h under
2.2.4. Synthesis of 5. The Oct-amido catalyst (5), Me2Si
(C29H36)(N-tBu)ZrCl2, was prepared as reported in the literature
by Miller et al.27 For the anchoring of 5 to the inorganic sup-
ports, the catalyst was placed in an air-free vessel with the
support, DEE was added at −78 °C and this mixture was
allowed to warm up to room temperature and stirred for 12 h
(Chart 5). After rinsing the solid with DEE, the material was
vacuum dried and the solid named 5@support was stored in
the glovebox.
2.3. Polymerization procedures
All the preparations were performed inside of a nitrogen-filled
glovebox and the polymerization reactions were performed
behind a blast shield. Methylaluminoxane (MAO) was pur-
chased as a 30 wt% toluene solution and used as a dry powder
obtained by vacuum removal of all volatiles. Toluene was dried
over sodium, distilled and stored in a Straus flask.
Chart 3 Synthesis of 2 (solid line) and 3 (dashed line) over the supports.
RvOH, OEt, O–Siu.
A 85 mL glass Lab-Crest (Andrews Glass Co.) cylindrical
polymerization vessel equipped with a stir bar was charged
with a 1000 equiv. of solid MAO (ca. 0.193 g) and 25 mL of
toluene. The reactor was then pressurized with ethylene,
heated to 60–62 °C and allowed to equilibrate while stirring.
After 10 min of equilibration, 1.0 mL of the catalyst suspension
in toluene was injected (3.33 μmol of Zr per mL) into the
vessel using a 2.5 mL Hamilton syringe. At the end, the reac-
tion was vented and quenched by adding approximately
200 mL of 10 wt% aqueous HCl in MeOH to the polymer solu-
tion. The polymer was collected by filtration, rinsed with fresh
methanol, and dried under high vacuum.
2.4. Characterization
X-ray diffraction (XRD) was performed in a XRD7000 Shimadzu
apparatus with Cu radiation (Kα = 0.15406 nm). Slits of 5 mm
Chart 4 Synthesis of 4@magadiite. RvOH, OEt, O–Siu.
This journal is © The Royal Society of Chemistry 2018
Dalton Trans.

View Article Online
Paper
Dalton Transactions
were used for dispersion and convergence radiation. High systems were studied by a combination of physico-chemical
power decoupling (HPDEC) 29Si NMR spectra of solid samples methods, such as XRD diffraction, solid state UV-Vis and NMR
were recorded on a Bruker Avance III+ 300 MHz. The samples spectroscopy.
were spun at 10 kHz in a zirconia rotor. More than 1024 scans
were obtained for proton-decoupled 29Si NMR spectra, with 60 s
3.1 The catalysts’ structures
delay time. For 13C CP/MAS NMR spectra, >1000 scans were
3.1.1 Catalyst 1@MCM-41. One of the attempts to tether a
accumulated, using a 3 s delay between pulses, an acquisition
fluorene derivative catalyst on the surface of the inorganic
time of 45 ms and a contact time of 4 s. The elemental ana-
support has been made by anchoring of the fluorene silane to
lyses of carbon, hydrogen and nitrogen were performed using
the mesoporous solid MCM-41. The X-ray diffractogram of
MCM-41 (Fig. 1, curve a) shows four main peaks below 7° 2θ,
assigned to the (100), (110), (200) and (210) Miller’s index
related to hexagonal pore organization. After functionalization
a CHNS/O Analyzer 2400 series II from PerkinElmer. The
UV-Vis spectra of the solids were collected using a Varian Cary
5000 using a diffuse reflectance apparatus without dilution.
The non-modified inorganic supports were used as a blank.
of the surface by fluorene silane, curve b, and the formation of
For the liquid samples, a quartz cuvette was employed and
catalyst 1@MCM-41 (curve c), there is a decrease in the inten-
sity of these signals due to the presence of higher organic
content in the solid.
In order to confirm the nature of porous MCM-41 before
and after the organic modifications, N2 adsorption/desorption
diethyl ether was used as a solvent. To determine the Zr con-
centration in solution, the Lambert–Beer Law was employed.
For that, a concentration versus absorbance curve was drawn to
determine the molar absorptivity coefficient (ε). The Zr con-
centration in the solid was calculated by the difference
isotherms were obtained and are seen in the ESI (SI_01†).
between the amount of Zr added initially to the reaction and
Table 1 shows the physical parameters obtained from N2 iso-
the amount remaining in the solution after the reaction. The
therms, such as the BET surface area, pore diameter and wall
content of zirconium was confirmed by X-ray fluorescence
thickness.
(XRF) as well as the amount of chlorine in the final catalysts.
The pores of 1@MCM-41 become slightly smaller as a
result of immobilization of the zirconium complex. The
The XRF data were collected using a Shimadzu XRF-1800
equipment. The N2 adsorption/desorption isotherms were
increase in the wall thickness indicates the incorporation of
obtained at cryogenic temperature using a NOVA 4200e equip-
the molecules inside the pores of MCM-41. The longest dis-
ment (Quantachrome Instruments). The samples were heated
tance of the organic moiety is approximately 9.1 Å.34 The pores
at 100 °C under vacuum for 24 h prior to the measurement.
of MCM-41 which possess an average diameter of 31 Å would
Surface areas were determined using the Brunauer–Emmett–
provide more than enough space for the organic molecule to
Teller (BET) method and pore size distribution data were col-
bind to the internal surface of MCM-41 cavities.
lected by the NLDFT method using cylindrical pores on silica
The 29Si HPDEC/MAS NMR spectrum of MCM-41 (Fig. 2,
for MCM-48 and BJH for MCM-41.
curve a) shows a peak in the region of Q4 [Si(4OSi)] at
Polymers were characterized by thermal analyses in a TA
−108 ppm and a peak of the Q3 [Si(3OSi)(OH)] site is observed
Instruments Q5000 from room temperature up to 600 °C at a
at −100 ppm. The immobilization of silane on the MCM-41
surface is confirmed by the presence of T-site signals in the
rate of 10 °C min−1 under nitrogen. Differential scanning
calorimetry (DSC) was performed in a TA Instruments Q600
under helium at a rate of 10 °C min−1 from 0 °C to 200 °C. 13
C
NMR samples were prepared by dissolving ∼50 mg of polymer
in 1.0 mL of 1,1,2,2-tetrachloroethane-d2. Spectra were
obtained using a Varian INOVA 500, 125 MHz at 100 °C. A
pulse of 70° was applied, the delay time was 5.5 s and 1500
scans were accumulated. The peak assignments and branching
calculations were performed as described previously.28–30 The
polymer molecular weights (Mw) were estimated by spin
diffusion using the 1H DOSY NMR technique in a Bruker
Avance 500 MHz spectrometer as described by Grubbs et al.31
3. Results and discussion
The molecular structure of the supported catalysts is a key
point for the understanding of the rules governing the stability
and reactivity of the system during polymerization reactions.
In this sense, detailed studies on inorganic materials as well
as the organic part of the supported catalysts are the subject of
Fig. 1 XRD profile of (a) MCM-41, (b) fluorene grafted-MCM-41 and (c)
great interest in this study. The molecular properties of these assembled catalyst 1@MCM-41.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2018

View Article Online
Dalton Transactions
Paper
Table 1 Physical parameters of the tethered catalysts on MCM-41 and MCM-48 obtained by N2 isotherms
Catalyst
d (nm)
Surface area BET (m2 g−1
)
Total pore volume (cm3 g−1
)
Pore diameter D (nm)
Wall thicknessa (nm)
MCM-41
d100 = 4.011
d100 = 4.085
d100 = 4.283
d211 = 3.247
d211 = 3.247
1033
953
498
1047
443
0.863
0.745
0.447
0.835
0.299
3.143
3.122
3.110
3.775
2.027
1.488
1.595
1.886
0.699
1.130
1@MCM-41
3@MCM-41
MCM-48
2@MCM-48
2 ꢀ d100
a
ξ0
D
2
a For MCM-41: Wall thickness ¼
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ꢁ D, where
D
=
BJH average pore diameter.32 For MCM-48: Wall thickness ¼
ꢁ
, where
pffiffiffi
3
a ¼ dhkl
ꢀ
h2 þ k2 þ l2, ξ0 = 3.0919 and D = NLDFT pore diameter.33
.
Fig. 3 (A) 13C CP/MAS NMR spectrum of 1@MCM-41 catalyst. (B) Solid
state UV-Vis spectra of (a) fluorene grafted-MCM-41 and (b) assembled
catalyst 1@MCM-41.
Fig. 2 29Si HPDEC MAS NMR spectra of (a) MCM-41, (b) fluorene
grafted-MCM-41 and (c) assembled catalyst 1@MCM-41. RvOH, OEt,
O–Siu.
This band is observed if the catalyst (or part of it) exhibits a
face-down type configuration37 with the surface as shown in
Scheme 1.
Table 2 reports the loading of the grafted fluorenyl and zir-
conium contents on the supported systems. In the case of
1@MCM-41, the fluorenyl/Zr molar ratio of 2.9/1.0 indicates
that for every two fluorenyl group, there is one zirconium
center, forming the complex, and about 30% of the fluorenyl
groups are not a zirconium complex, remaining as a free
ligand on the surface of the support. The Zr/Cl molar ratio of
1.0/1.1 is in agreement with the proposed face down Zr
complex with the support surface, indicated by the UV-Vis
spectra and shown in Scheme 1. The face down and face up
effects would greatly impact the catalyst activity since the
monomer and cocatalyst access to the metal center can be sig-
−50 and −75 ppm range in curve b, which accounts for 16% of
Si sites. After the reaction with ZrCl4, curve c, there is a
decrease in the SiT/SiQ ratio from 16% (curve b) to 7% in
1@MCM-41 (curve c), indicating partial leaching of silane in
the presence of n-butyllithium (used to generate the fluorenyl
anion in the formation of 1@MCM-41). The organic content in
the final catalyst 1@MCM-41 is confirmed by the 13C CP/MAS
NMR spectrum as shown in Fig. 3A. The presence of the aro-
matic carbons in the region of 100–150 ppm as well as the Si–C
bond at 30 ppm can be observed, proving that the organic part
of the catalyst is still attached to the surface of MCM-41.
The UV-Vis spectra of the fluorene-modified MCM-41
before and after the reaction with ZrCl4 are shown in Fig. 3B. A
band with a maximum of 288 nm is observed, assigned to the
electronic transitions of the fluorene ring (curve a). This same
band, when in the presence of zirconium (curve b), is dis-
placed in the region between 300 and 450 nm (with a
maximum at 320 nm) and it is broadened due to the fluorene-
Zr LMCT transitions, indicative of the formation of
1@MCM-41 as desired. In this same spectrum, curve b, a well-
defined band at 224 nm is observed, typical of the O2− → Zr4+
Scheme 1 Face-down and face-up conformation of catalyst
1@MCM-41. Only the cyclopentadienyl metal centre is shown for clarity.
charge transfer from Zr in a tetrahedral conformation.35,36 RvOH, OEt, O–Siu.
This journal is © The Royal Society of Chemistry 2018
Dalton Trans.

View Article Online
Paper
Dalton Transactions
Table 2 Elemental analysis of the tethered catalysts
Fluorenyl contenta
N contenta
Zr contentb
Cl contentb
Catalyst
(mmol g−1 solid)
(mmol g−1 solid)
(mmol g−1 solid)
(mmol g−1 solid)
Cl : Zr : N
1@MCM-41
2@Magadiite
2@MCM-48
3@MCM-41
0.455
3.048
3.641
1.395
—
0.157
1.023
1.618
0.420
0.171
2.647
1.855
0.699
1.1 : 1.0 : 0
3.254
3.852
1.433
3.2 : 1.0 : 2.6
1.2 : 1.0 : 2.4
1.6 : 1.0 : 3.4
a Elemental analysis (CHNO). b XRF analysis.
nificantly affected by these face-down and face-up X-ray diffraction (curve b) of the solid shows that the peaks in
conformations. the region of 22–30° 2θ, which already indicated the turbostra-
3.1.2 Catalysts 2@magadiite, 2@MCM-48 and 3@MCM-41. tic effect due to the presence of CTA+, suggest a significant
Due to the simplicity of syntheses of amino-modified supports, decrease in the organization of the lamella, one in relation to
they have been studied for many different applications. One of another. This effect is pronounced after the reaction with
their uses is for organometallic olefin polymerization catalysts. fluorene and the incorporation of Zr into the complex built on
In this application, tethering can prevent catalyst leaching in the magadiite’s surface, curves c and d, respectively. After the
slurry phase processes, for example. Here, primary-amine- reaction with fluorene (curve c), there is a decrease in the inter-
functionalized materials of varying 2D and 3D morphologies layer space to 2.67 nm (3.30° 2θ), probably caused by the π–π
have been prepared and used as supports for tethered cata- interactions of the aromatic structure of fluorene as well with
lysts. In this method, the metal centre would be distant from the magadiite surface. When the catalyst 2@magadiite is
the surface by a three-carbon alkyl linker.
finally assembled, curve d, the interlamellar plane displace-
3.1.2.1 Catalyst 2@magadiite. Fig. 4 shows the X-ray diffrac- ment is observed at smaller angles (2.28° 2θ/3.87 nm), indicat-
togram of magadiite during the preparation of 2 on its surface. ing an increase of 0.7 nm of the interlamellar space compared
Curve a presents the diffractogram of CTA-magadiite whose to the starting CTA-magadiite, reflecting the accommodation
basal spacing is 3.17 nm (2.78° 2θ).
of the organic on catalyst 2@magadiite.
The presence of a long chain organic cation like CTA+
Fig. 5a presents the typical magadiite 29Si resonances
(C16H33(CH3)3N+) is important in the process of organo- assigned to Q3 [Si(3OSi)(OH)] at −99 ppm and two peaks in
functionalization because it allows the expansion of the inter- the Q4 [Si(4OSi)] region at −112 and −109 ppm. The
lamellar space and makes it more organophilic, facilitating the functionalization of the CTA-magadiite surface can be seen in
access of the organic molecules to the space between the curve b of Fig. 5. One can observe the appearance of T2 (RSi
lamella of the silicate. After the aminopropyl grafting step, the (OSi)2OH) and T3 (RSi(OSi)3) signals at −59 and −67 ppm,
respectively, corresponding to 19% of the sites (SiT/SiQ ratio),
Fig. 4 XRD diffractograms of (a) CTA-magadiite, (b) NH2-modified
magadiite, (c) fluorene-NH2-magadiite and (d) 2@magadiite. RvOH,
OEt, O–Siu.
Fig. 5 29Si HPDEC/MAS NMR spectra of (a) CTA-magadiite, (b) NH2-
modified magadiite, (c) fluorene-NH2-magadiite and (d) catalyst
2@magadiite. RvOH, OEt, O–Siu.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2018

View Article Online
Dalton Transactions
Paper
which is kept after addition of fluorene (curve c). The final
catalyst 2@magadiite exhibits a lower amount of T sites (curve
d) compared to the functionalized precursor (curve b). The SiT/
SiQ ratio decreases to 11%, indicating that there was partial
loss of the surface aminopropyl groups in the presence of
n-butyllithium employed for the addition of zirconium.
The organic moieties of catalyst 2@magadiite were also
evaluated by 13C CP/MAS NMR (Fig. 6). The 13C NMR spectrum
of CTA-magadiite (curve a) shows the characteristic peaks of
the all-trans configuration of the organic CTA+ cation, which
keeps the magadiite’s lamella separated. The aminopropyl
group can be observed in the spectrum of curve b along with
CTA+. After the reaction of the aminopropyl groups with fluor-
ene, curve c, the appearance of a peak at 61.4 ppm was
observed (dotted square in curves c and d of Fig. 6) assigned to
the bond formed between the –NH2 group and the fluorene,
leading to the immobilized fluorene on the solid. In addition,
the aromatic carbons from fluorene moieties appear in the
region of 90 to 160 ppm as expected. After the reaction with
ZrCl4, curve d, the catalyst remains bonded to the solid as indi-
cated by the presence of the N–fluorene bond (peak in
61 ppm) and the fluorene aromatic carbons between 90 and
160 ppm.
Scheme 2 Possible species on 2@magadiite catalyst. RvOH, OEt, O–
Siu.
The elemental analysis of 2@magadiite in Table 2 shows
the Cl/Zr/N molar ratio of 3.2/1.0/2.6. This result indicates
that, beyond the proposed system which should be ideally Cl/
Zr/N molar ratio of 2/1/2 (species A in Scheme 2), there are
some possible species that cannot be neglected like the tri-
chloride zirconium complexes on the surface (species B in
Scheme 2) as well as a few free aminopropyl groups in the
solid (species C in Scheme 2).
Fig. 7 (A) XRD profile and (B) 29Si HPDEC NMR of (a) MCM-48, (b) NH2-
modified MCM-48, (c) fluorene-NH2-MCM-48 and (d) catalyst
2@MCM-48.
polymerization. MCM-48 is formed in a symmetric space
group of the Ia3d type with cubic organization of pores. The
typical MCM-48 diffractogram (Fig. 7A, curve a) contains the
3.1.2.2 Catalyst 2@MCM-48. Solids in which pore sizes
have fixed diameters, like MCM-41 and MCM-48, were also
investigated in the preparation of tethered catalysts for olefin
Fig. 6 13C CP/MAS NMR spectra of (a) CTA-magadiite, (b) NH2-modified magadiite, (c) fluorene-NH2-magadiite and (d) catalyst 2@magadiite. * CH2
and CH3 from DEE solvent (66 and 16 ppm).
This journal is © The Royal Society of Chemistry 2018
Dalton Trans.

View Article Online
Paper
Dalton Transactions
peaks assigned to the planes (211), (220), (321), (400), (420) assigned to the bonding of the –NH group to fluorene, again
and (332). confirming the immobilization of fluorene to the mesoporous
Here, the same procedure as that of aminopropyl-grafting, silica surface. After assembling of catalyst 2@MCM-48, curve c,
followed by the fluorene reaction with the –NH2 groups and the 13C NMR profile is maintained.
zirconium addition, reported above for magadiite, was per-
The Cl/Zr/N molar ratio of 2@MCM-48 is 1.2/1.0/2.4. This
formed. One can observe that after the organic modifications, system is slightly different from 2@magadiite: although both
curves b–d in Fig. 7, the main diffraction peaks (211) and (220) catalysts can exhibit free aminopropyl groups like species C in
are broadened. This is probably due to the increase in the Scheme 2, while 2@magadiite presents trichloride zirconium
organic content (that does not diffract) as well as the hetero- complexes and 2@MCM-48 presents monochloride zirconium
geneous distribution of the organic groups in the MCM-48 species.
channels.
3.1.2.3 Catalyst 3@MCM-41. Sterically expanded fluorene
The pore volume and diameter of 2@MCM-48 are signifi- was also synthesized over the surface of MCM-41 in order to
cantly smaller than the non-modified MCM-48 (see Table 1) as investigate its performance on olefin polymerization. Catalyst
well as the surface area. The approximate 50% decrease in the 3@MCM-41 was synthesized as reported for 2@magadiite and
surface area can be assigned to the successful anchoring of 2@MCM-48, using an Oct-fluorene derivative instead of fluor-
the catalyst on the internal surfaces of the mesoporous ene. The powder XRD diffractograms of the solids after each
material. This is confirmed by the decrease in the adsorption reaction step are reported in the ESI (Fig. SI_02†). MCM-41 pre-
pore volume from 0.8 to 0.3 cm3 g−1 and the wall thickening sented sharp, well-defined peaks which, in the final assembled
from 0.7 nm to 1.1 nm with the addition of the organic com- 3@MCM-41 catalyst, are broadened as a result of a slightly dis-
ponent similar to 1@MCM-41.
turbed scattering contrast.38 As the effect might be inhomo-
Surface functionalization of these materials is confirmed by geneous or affect pores differently, some will show thicker
29Si NMR spectra (Fig. 7B), where T sites are observed between walls than others, changing a0 and creating a set of peaks,
−40 and −70 ppm (curve b). After insertion of fluorene fol- slightly spaced one from the other compared to the non-
lowed by the reaction with ZrCl4, there is a decrease in the SiT/ organic modified support. The net result is broadening of the
SiQ ratio from 28% (curve b) to 8% in the final catalyst signals.
2@MCM-48 (curve d), the effect of the partial hydrolysis of
The grafting of the surface by the silane is confirmed by
some aminopropyl groups as mentioned before for the other 29Si HPDEC/MAS NMR, Fig. 9. After aminopropyl grafting,
systems. Regarding the organic content on 2@MCM-48, Fig. 8 curve b, the evident appearance of signals referring to the T3
(curve a) displays the 13C CP/MAS NMR spectrum of the ami- sites (RSi(OSi)3) at −67 ppm is observed, confirming the incor-
nopropyl functionalized MCM-48 with characteristic peaks of poration of the aminopropyl pendent to the inorganic struc-
the immobilized aminopropyltriethoxysilane moiety. After the
reaction with fluorene, curve b, the solid presents peaks in the
region of 100 to 160 ppm due to the 13C atoms of the fluorene
aromatic system and a broad and intense peak at 61 ppm
Fig. 8 13C CP/MAS NMR of (a) NH2-modified MCM-48, (b) fluorene-
NH2-MCM-48 and (c) catalyst 2@MCM-48. RvOH, OEt, O–Siu. * CH2
and CH3 from DEE solvent (66 and 16 ppm).
Fig. 9 29Si HPDEC NMR spectra of (a) MCM-41, (b) NH2-modified
MCM-41, (c) Oct-fluorene-NH2-MCM-41 and (d) catalyst 3@MCM-41.
RvOH, OEt, O–Siu.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2018

View Article Online
Dalton Transactions
Paper
ture. When reacting with Oct-fluorene (curve c) and ZrCl4
(curve d) there is the appearance of T2 sites (RSi(OSi)2OH) at
−58 ppm, resulting from the partial loss of some aminopropyl
groups. The organic moieties were investigated by 13C CP/MAS
NMR spectroscopy and present a similar behaviour described
for 2@magadiite or 2@MCM-48 catalysts and can be seen in
the ESI (SI_03†). The catalyst is attached to the MCM-41
surface as indicated by the peak at 60 ppm in the 13C NMR
spectrum.
The Oct-fluorenyl content was determined as 1.395 mmol
g−1 of solid (Table 2). Also, catalyst 3@MCM-41 presents a Cl/
Zr/N molar ratio of 1.6/1.0/3.4, which means that, as observed
for 2@MCM-48, free aminopropyl groups and monochloride
zirconium species as well can be present in the solid. All of the
possible species tethered on the surface of MCM-41 might be
at the internal cavities of the support. The significant decrease
in the surface area from 1033 m2 g−1 to 498 m2 g−1 can be
assigned to the anchoring of the catalyst on the surface of the
MCM-41. The wall thickening from ca. 1.5 nm to ca. 1.9 nm
indicates that the catalyst was assembled in the inner part of
the support. Moreover, there is a plenty of room for the catalyst
to be built inside of the pore: the maximum length that cata-
lyst 3 can have is ca. 13.5 Å (ref. 34) while MCM-41 possesses
31.4 Å of pore diameter.
Fig. 10 UV-Vis spectra of the solid catalysts 2@magadiite, 2@MCM-48
and 3@MCM-41.
Catalysts 2@magadiite, 2@MCM-48 and 3@MCM-41 were
prepared using the aminopropyl group as a bridge for their
immobilization onto the surface of the supports. The 13C
peaks of the aminopropyl group exhibited narrow resonances
reflecting the high flexibility or freedom of the molecular
moiety. After insertion of fluorene or Oct-fluorene, these
propyl chain peaks are broader especially when MCM-41 and
MCM-48 are the supports, indicating that the degree of flexi-
bility of the aminopropyl chain decreased. This effect is less
evident in magadiite due to its expandable interlayer space
and the presence of interlamellar CTA+. Also, in all cases the
peak at 60 ppm, assigned to the –NH–Fluorene bond is quite
broad, indicating the participation of different conformations
and possible interactions of the organic part with the surface
of the support. These surface interactions directly affect the
catalytic activity of the metal centre during the polymerization
of olefins and can be understood by UV-Vis spectroscopy.
3.1.2.4 Catalysts 2@magadiite, 2@MCM-48 and 3@MCM-41
Scheme 3 Schematic representation of the interaction of the organic
part of the catalyst with the surface of the support. Only the cyclopenta-
dienyl metal centre is shown for clarity. RvOH, OEt, O–Siu.
than 2/1 as seen for the set of catalysts shown here. This inter-
action with the surface occurs more intensely in mesoporous
solids, whose pore arrangements are rigid and confine the
catalyst within a more restricted space. In these cases, the Zr
center interacts with the supports’ surface at the expense of
the organic chain flexibility. When the catalyst is built on the
surface of magadiite, which has expandable interlayer space,
this effect was not observed. This confinement has an impor-
tant effect on the production of polyolefins and is discussed
next.
3.2 The polymerization of olefins by tethered catalysts
from the point of view of UV-Vis spectroscopy. Looking at the Catalysts 1@MCM-41, 2@magadiite, 2@MCM-48 and
UV-Vis spectra of the catalysts shown in Fig. 10, a band can be 3@MCM-41 were employed for the polymerization of ethylene
observed between 270 and 320 nm, present in all solids, as well as propylene and are summarized in Table 3. Catalyst
characteristic of the π interactions of the fluorene aromatic 1@MCM-41 did not work under any of the conditions
ring with the pair of electrons of the NH group to which it is employed for both ethylene and propylene polymerizations. As
attached,39 besides a band in the region of 431 nm due to the seen by UV-Vis spectra shown in Fig. 3B, the metal centre
interaction of the fluorenyl anion with the zirconium, assigned becomes irreversibly blocked in a face-down configuration, in
to ligand–metal charge transfer (LMCT) fluorene → Zr4+.
addition to limiting the conformational freedom of the catalyst
In 2@MCM-48 and 3@MCM-41 an intense band in the on the surface, preventing the capture of the monomer to poly-
region of 220 nm is observed and it seems possibly due to the merize. If the catalyst is built too close to the inorganic surface
charge transfer O2− → Zr4+ in a tetrahedral configuration. For of the support, the interaction of the metal center with the
this charge transfer to occur, the metal atom of the catalyst surface can lead to some drawbacks, including the possible
should interact with the surface of the carrier as shown in formation of deactivated catalysts through their interactions
Scheme 3, and the Cl/Zr molar ratio is expected to be lower with the support material. As illustrated in Scheme 1,
This journal is © The Royal Society of Chemistry 2018
Dalton Trans.

View Article Online
Paper
Dalton Transactions
Table 3 Polymerization results using tethered catalysts 1@MCM-41, 2@magadiite, 2@MCM-48 and 3@MCM-41
Activity (kg per mol
Zr per h)
Tmax of
decomposition (°C)
Entry
Catalysts
Monomer
TM (°C)
TC (°C)
1
2
3
4
5
6
7
8
9
1@MCM-41
1@MCM-41
2@Magadiite
2@Magadiite
2@Magadiite
2@MCM-48
2@MCM-48
2@MCM-48
3@MCM-41
3@MCM-41
Propylene
Ethylene
Propylene
Ethylene
Ethylenea
Propylene
Ethylene
Ethylenea
Propylene
Ethylene
0
0
14
656
595
0
353
301
0
—
—
—
135
125
—
130
131
—
—
—
—
120
118
—
117
113
—
—
—
—
394
452
—
398
451
—
10
30
133
118
396
Zr = 3.33 μmol. PE: 5.5 bar, 60 °C, 10 min. PP: 30 mL of propylene, r.t., 30 min. MAO/Zr = 1000. a 2.5 mmol of 1-octene was added to the reactor.
1@MCM-41 depicts a metallocene molecule bonded to the sites are in a similar situation as before and the polymeriz-
support by two short tether points. One would expect not only ation catalyst cannot reach its best performance.
to confine movement (translational and rotational) of the
3.3 Catalyst 4@magadiite
molecule, but also to make the metal center in the face-up con-
figuration, i.e., facing away from the silica surface. The face-up
configuration ensures sufficient accessibility of the metal
center by monomers during polymerization. However, what
was found in the present work, following the strategy pro-
posed, is that the movement restriction by two tether points to
the surface is not strong enough, so the bonded molecule may
still take a face-down configuration. When an alkyl chain is
added between the support and the catalyst metal centre as in
catalysts 2@magadiite, 2@MCM-48 and 3@MCM-41, the scen-
ario changes a little. These three catalysts showed low activity
in the polymerization of ethylene and were completely inactive
in the preparation of polypropylene. Other conditions were
tested and there was no formation of polypropylene.
The interesting results come from the ethylene polymeriz-
ation reaction (Table 3, entries 4, 7 and 10): a single peak in
DSC curves between 130–135 °C was observed in all samples,
suggesting the formation of HDPE (High Density
PolyEthylene). Ethylene polymerization produced linear PE, as
observed by the single peak at 30 ppm in the 13C NMR spec-
trum (ESI SI_04†), showing the role of the carrier as a limiting
factor in the insertion of branching even as small as a methyl
group as in the case of propylene. Moreover, in order to induce
branching formation on the PE chain, 1-octene was added to
the polyethylene synthesis using 2@magadiite and
2@MCM-48 as catalysts (entries 5 and 8 of Table 3, respect-
ively). However, there was no increase in the degree of branch-
ing of PE as observed by the presence of a single peak at
30 ppm in the 13C NMR spectrum (ESI SI_04†).
Keeping these results in mind, another strategy for the immo-
bilization of olefin polymerization catalysts was performed
using a multi-step methodology in order to construct an
assembled catalyst, similar to the homogeneous one, on the
surface of magadiite. So, first CTA-magadiite (Fig. 11, curve a)
was functionalized with triethoxyphenylsilane (phenyl-CTA-
magadiite), curve b. Then, through a Friedel Crafts acylation,
the benzophenone molecule was bonded to the support (ben-
zophenone-CTA-magadiite), curve c. After condensation of the
aromatic cyclopentadienyl to the system, the diphenylfulvene
group was anchored to the support (diphenylfulvene-CTA-
magadiite), curve d. Next, the reaction with fluorene led to the
formation of a Ph2C(C5H5)(C13H9) ligand bound to magadiite,
curve e. The final magadiite catalyst is obtained after the reac-
tion of ZrCl4 with the solid, curve f, leading to the catalyst
4@magadiite. It is observed that during diphenylfulvene-CTA-
magadiite formation (curve d) the characteristic interlamellar
distance increases from 3.17 nm to 3.82 nm, accommodating
Using the aminopropyl chain, the covalent tethering
method was expected to avoid the disadvantage of the direct
grafting of the fluorene method by putting a covalent spacer
between an active site and the support surface. In fact, it
helped the system and some polyethylene was synthesized. But
the flexibility of a long chain also provides the active site a
pseudo-solution environment, i.e., there is still some degree of
flexibility of the chain that provides a route for the direct inter-
Fig. 11 XRD diffractograms of (a) CTA-magadiite, (b) phenyl-CTA-
magadiite, (c) benzophenone-CTA-magadiite, (d) diphenylfulvene-CTA-
magadiite,
(e)
Ph2C(C5H5)(C13H9)-magadiite
and
(f)
catalyst
action of the active site with the surface, so some of the active 4@magadiite.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2018

View Article Online
Dalton Transactions
Paper
Fig. 13 UV-Vis spectra of (a) Ph2C(C5H5)(C13H9)-magadiite and (b) cata-
lyst 4@magadiite (before and after ZrCl4 addition, respectively). RvOH,
OEt, O–Siu.
Fig. 12 29Si HPDEC/MAS NMR spectra of (a) CTA-magadiite, (b) phenyl-
CTA-magadiite, (c) benzophenone-CTA-magadiite, (d) diphenylfulvene-
CTA-magadiite, (e) Ph2C(C5H5)(C13H9)-magadiite and (f) catalyst
4@magadiite. RvOH, OEt, O–Siu.
ligand, it reflects in steric and electronic effects of smaller or
higher order; at the end, it influences the reaction mechanism
and the production of the polymer. The catalysts built farther
away from the surface had low activity in the production of PE
due to the formation of a more stable species coming from the
interaction with the surface. Catalysts built close to the surface
had no activity in the production of olefins. Two factors are
related: (i) the system can be locked on a face-down configur-
ation with the surface or (ii) the system is sensitive to the reac-
tive conditions and ends up being leached from the structure.
The Zr–O interactions are difficult to eliminate in these
systems and seem to be certain using the current strategies. A
good supporting technique should not only strongly immobi-
lize a metallocene catalyst, i.e., no leaching of active species
from the support, but it also preserves the most useful features
of the catalyst.
Thus, in order to prepare an efficient heterogeneous metallo-
cene catalyst, the direct deposition approach was taken into
consideration. This method can be very convenient for metal-
locene immobilization. In a typical process, the pre-treated
support is stirred with a solution of a metallocene compound
under an inert atmosphere at room temperature or elevated
temperatures for a period of time. Then, the slurry is filtered,
and the remaining solid product is washed with solvent
several times to remove weakly adsorbed metallocene mole-
cules. The washed product is dried under vacuum to remove
the solvent. The metallocene compound reacts with hydroxyl
groups on the silica surface and bond to the surface though a
M–O–Si bond (Chart 5). Good catalyst performance can be
reached with this strategy.40,41
the ligand. By adding fluorene and Zr, curves e–f, the inter-
actions between the aromatic rings force the system to
approach the lamella and the interlayer space reduces to
2.68 nm.
The 29Si NMR spectra (Fig. 12) show that there was a
decrease in the concentration of T sites from 22 to 10% when
catalyst 4@magadiite is assembled after addition of ZrCl4,
curve f, resulting in the partial loss of the silane in the pres-
ence of n-butyllithium employed for the addition of zirconium.
In fact, if we consider that the molar absorptivity is not
changed from the free ligand to the coordinated situation, the
UV-Vis spectrum (Fig. 13) shows that after the reaction with Zr,
there is a significant decrease of the aromatic organic absorp-
tion in the region of UV-Vis, indicating that there was, in fact,
leaching of organic part from the support catalyst. When
tested for the polymerization of ethylene and propylene, the
catalyst 4@magadiite had no activity in the reactions. Several
reaction conditions were tested, but no polymer was obtained.
The final solution, after addition of acidified methanol, was
rotoevaporated in an attempt to precipitate some polymeric
materials, but without success, proving that the metallocene
complex is sensitive to the reaction conditions employed for
its systems and was removed from the structure when the
multi-step methodology was chosen to tether the metallocene
to the support surface.
3.4. Catalyst 5@layered supports
In this sense, the Oct-amido catalyst (catalyst 5) was reacted
with different layered supports and the final solids represent a
survey of the types of supports that could be used with the Oct-
amido catalyst. Fig. 14A and B display the XRD diffractograms
The results shown here point to the intrinsic influence of the
support on the polymerization processes catalysed by the
metallocene center: because the presence of the support in the
surroundings of the metallocene plays the role of a bulky
This journal is © The Royal Society of Chemistry 2018
Dalton Trans.

View Article Online
Paper
Dalton Transactions
lyst. No exfoliation can be observed regarding the stacking of
the lamella peak below 5° 2θ. Solid state UV-Vis spectra was
used to certify that the catalyst was attached to the support
surface (Fig. 15). The presence of a peak at 220 nm shows that
the Zr–O bond was formed and indicates that the catalyst is
attached to the surface. The increase in absorbance around
345–415 nm also indicates the presence of the support in the
sample.
Once 5 was adsorbed onto the supports, polymerizations of
ethylene were performed. The layered supports were chosen
because of their special ability to accommodate guest com-
pounds in between the layers and they can also exfoliate to
better accommodate bulky molecules or polymers. The results
for polymers made from homogeneous 5, 5@n-alkyl-AlPO-kan
and 5@magadiite are shown in Table 4.
The homogeneous catalyst polymerizations, under the con-
ditions employed, have an activity of about 2000 kg of PE per
mol Zr per h with a melting temperature of 112 °C. Branching
is one of the most important features of catalyst 5: it is able to
Fig. 14 XRD diffractograms of (A) magadiite (a) and 5@magadiite (b)
and (B) n-alkyl-AlPO-kan (a) and 5@n-alkyl-AlPO-kan (b).
produce highly branched PE and low Mw (about 1000 g mol−1
)
from a source of pure ethylene, without the addition of co-
monomers. Overall, the supported catalysts had lower activities
compared to the homogeneous catalyst possibly due to the
steric hindrance of the layers, making it difficult to prepare
large amounts of the polymer or for monomer to diffuse into
the layers as well as some bimolecular deactivation from the
interaction of the catalyst and support.
of the layered supports magadiite and n-alkyl-AlPO-kan,
respectively, before and after anchoring of the Oct-amido cata-
The melting temperatures for all of the supported systems
increased to 134–135 °C. This is due to an absence of any
branching. This result was quite surprising. By supporting this
catalyst on inorganic supports, either the active site becomes
sterically hindered and will no longer accept the large α-olefins
made in situ to produce branching or it affects the mechanism
by which the catalyst affects the β-hydride elimination reac-
tion. What is more interesting is that this catalyst, once acti-
vated by MAO, is possibly detached from the support but is
contained within the layers and this still drastically changes
the mechanism and forces the catalyst to only produce linear
PE (see the 13C NMR spectrum in ESI SI_05†).
To help elucidate whether the support is affecting the
β-hydride elimination portion of the mechanism or whether it
is hindering the active site and preventing the α-olefins from
inserting into the polymer backbone, copolymerization with
Fig. 15 UV-Vis spectra of homogeneous Oct-amido catalyst 5 in CHCl3
(a), 5@magadiite (b) and 5@n-alkyl-AlPO-kan (c).
Table 4 Ethylene polymerizations with supported 5
Branchingc
Zr loading on the
Activity
(kg of PE per mol Zr per h)
Catalyst
supporta (mmol Zr per g of solid)
TM (°C)
Mw b (g mol−1
)
NE
NL
NT
5 Homogeneous
5@Magadiite
5@n-alkyl-AlPO-kan
—
0.0455
0.0555
2120
599
750
112
135
134
1048
20 036
—
6.3
0
0
39.5
0
0
45.8
0
0
Polymerizations were performed at 5.5 bar of ethylene, 60 °C for 10 min, with 3.3 μmol of Zr and 1000 equivalents of MAO. a Calculated by the
difference between the amount of catalyst added and the amount of catalyst remained in solution after reaction, using UV-Vis spectroscopy.
b Estimated by 1H DOSY NMR. c Number of ethyl (NE), long (n-hexyl or greater, NL) and total branches (NT) per 1000 carbons.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2018

View Article Online
Dalton Transactions
Paper
Table 5 Ethylene/1-octene copolymerizations with 5@magadiite
catalysts, not only for zirconocenes but also for metal com-
plexes as Grubbs’ ruthenium catalyst,16 RuCl3 anchored onto
functionalized silica for oxidation of benzylic alcohols,44 W,
Mo, Re, and Ru organometallic complexes45 and so on. The
formation of covalent bonds between the support surface and
metal complexes is a challenge. For zirconocene complexes as
well as for other organometallic complexes, immobilization
can be reached by the direct grafting of the complex onto the
support and anchoring of the complex via a covalent linkage
between the support surface and one of the ligands of the
complex.11,13–21
Branching
Activity (kg of PE
1-Octene (mmol)
per mol Zr per h)
TM (°C)
NE
NL
NT
3.2
1.6
40
340
125
122
0
0
0
0
0
0
Polymerizations were performed at 5.5 bar of ethylene, 60 °C for
30 min, with 3.3 μmol of Zr and 1000 equivalents of MAO.
ethylene and 1-octene was performed. If the co-monomer
inserts as a hexyl branch, then the support is influencing how
the catalyst makes the macromonomers. However, if only
linear polyethylene is produced then the active site is sterically
hindered by the support. The polymerization trials with
1-octene are shown in Table 5. No branching occurred when
1-octene is added to the reaction. This indicates that the
support is hindering the active site, preventing any α-olefins—
whether made in situ or manually added—into the backbone
of the polymers.
Immobilizations via non-covalent interactions like ionic
interactions or physisorption, for example, are less common
due to the leaching aspects. In the case of zirconocenes, when
applied for polymerization, regardless the type of anchoring
employed, the complexes as well as the zirconium metal are
retained in the final polymer chain. Thus, a judicious choice
of the supports allows the formation of interesting composites.
Moreover, whether the catalyst does not deactivate and is
sufficiently stable under the reaction conditions, a good immo-
bilization methodology would be reached.
Microscopy images of these polymers were obtained and
are shown in Fig. 16. PE made from the homogeneous 1
(Fig. 16A and B) show that flakey layers of polymers are formed
during polymerization. As these polymers are not melted
before microscopy, the images show how the PE forms during
polymerization. These polymers form quite differently when
compared to the PE produced from the supported catalyst. The
ones made by 5@magadiite (C and D) are not flakey at all, the
flat surface of the polymer displays several defects. A closer
view of the image shows the silica support in white (Fig. 16D).
PE from 5@AlPO-kanemite shows a more defective surface but
there are some indications of flakey layers within the material
(E and F). These images show that the supports drastically
influence the fashion in which the polymer forms during
polymerization.
These results on polymerization using supported Oct-
amido (catalyst 5) shed some light on the way to control the
degree of branching on the PE chain using heterogeneous cat-
alysts. Recent results42 on the preparation of polyethylene
nanocomposites by constrained geometry catalysts (CGCs)
have been obtained by reacting the cocatalyst MAO and catalyst
5 with different 2D supports. Polyethylene nanocomposites
were produced with good activities and a plenty of control on
the branching content of PE.
The presence of the support in the surroundings of the
metallocene plays the role of a bulky binder, reflecting in
steric and electronic effects of smaller or higher order, that
end up influencing the monomer access to the active center
and the properties of the final polymer composites.
The catalyst built in close proximity to the surface of a
mesoporous solid, like 1@MCM-41 employing the silane strat-
egy shown in the present work, had no activity in the pro-
duction of olefins. One of the factors related is that the system
can be assembled in face-up or face-down configurations due
to the proximity to the support surface. An endeavour to
reduce such interaction was binding the metal center far from
the surface using an alkyl group as a spacer. Thereupon, cata-
lyst 2 was assembled on the surface of a mesoporous material
(MCM-48) and the system was active for ethylene polymeriz-
ation (350 kg of PE per mol Zr per h). However, a bulky organic
complex like catalyst 3@MCM-41 can hinder the access of the
monomers into the cavity of the support where the active sites
are lodged. Besides the hindrance effect, the restricted region
forced the zirconium complex to interact with the OH group
on the surface of the support. The Cl/Zr molar ratios of
2@MCM-41 and 3@MCM-41 samples are 1.2/1.0 and 1.6/1.0,
respectively, which is evidence of the confining effect and
possible binding of the zirconium with the surface. In an ideal
situation, where no interaction with the surface occurs, the
ratio should be 2/1. The monochloride species found for these
catalysts forces the metal center to interact with the surface
3.5. The nature of 2D and 3D structures and tethering
catalyst strategies
What sets zirconocenes apart from other catalysts is that the groups of the support. This is not unexpected since the
zirconocene unit is able to perform selective carbon–carbon complex has been assembled into the cavity of these meso-
bond-forming reactions due to its ability to stabilize alkynes porous solids as the data from N2 isotherms showed by the
and alkenes.43 Turning these systems heterogeneous adds the thickening of their walls after the catalyst construction.
benefits of heterogeneous catalysis like recyclability with the
If the inner region of the support where the catalyst is
controlled transformations that occur at the single site metal housed can be expanded, as in the case of the layered maga-
center, provided by the homogeneous system. There are many diite, the bending of the organic chain is not extensive and the
strategies proposed to generate well-defined heterogeneous catalyst performance could be improved. This is what happens
This journal is © The Royal Society of Chemistry 2018
Dalton Trans.

View Article Online
Paper
Dalton Transactions
Fig. 16 SEM images of PE. (A) Image and (B) close up image of homogeneous 5. (C) Image and (D) close up image of PE made by 5@magadiite.
(E) Image and (F) close up image of PE from 5@AlPO-kanemite.
when 2 is built on the surface of magadiite. The performance porous solids MCM-41 and MCM-48 showed evidence of
of 2@magadiite on the synthesis of polyethylene almost limited molecular movement within the pores and, in most of
doubled compared to the same zirconium complex the cases, leading to the trapping of the zirconium complexes
2@MCM-48. This clearly is evidence of the role of the bidimen- on the surface. The less constrained environments within a
sional structure in comparison with the tridimensional one.
solid structure, like layered supports, open up the possibilities
For polymerization to occur in the heterogeneous catalyst, of designing systems and their final products, e.g., when cata-
the monomers must be able to penetrate into the structure lyst 5 was supported on two different types of layered supports
and access the metal sites. Thus, the porosity of the solid (magadiite and n-alkyl-AlPO-kanemite), the ability to fine tune
structure should aid the polymerization reaction if there is the guest species in these supports can lead to bimodal PE or
enough room to do so. In the cases studied here, the meso- to a purely linear HDPE.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2018

View Article Online
Dalton Transactions
Paper
4. Conclusions
References
Primary-amine-functionalized materials have been prepared
and used as supports for tethered catalysts. Other strategies
tested were the grafting of fluorene and the direct construction
on the inorganic surfaces. Direct anchoring of homogeneous
catalysts was also performed. From the ethylene polymeriz-
ation results using methylaluminoxane as the co-catalyst, two
main features have to be highlighted when these catalysts are
tethered: (i) the site for covalent attachment should be remote
from the catalytic site for olefin coordination and insertion.
Due to the presence of a flexible –(CH2)3NH– chain, the metal-
lic center interacts with –OH on the inorganic surface (con-
firmed by UV-Vis spectra of the final solids) and access of the
ethylene monomers and/or short chains to the Zr4+ center is
difficult, leading to lower activities and poor-branched PE pro-
duction. Actually, the activity of 2@magadiite (656 kg of PE
per mol Zr per h) whose layers are expandable are expressively
higher than the activity of 2@MCM-48 (353 kg of PE per mol
Zr per h) and catalyst 3@MCM-41 (30 kg of PE per mol Zr per
h) where the catalysts are hindered into the pores. Both cata-
1 R. Geyer, J. R. Jambeck and K. L. Law, Sci. Adv., 2017, 3, 25–29.
2 M. Stürzel, S. Mihan and R. Mülhaupt, Chem. Rev., 2016,
116, 1398–1433.
3 M. Jezequel, V. Dufaud, M. J. Ruiz-Garcia, F. Carrillo-
Hermosilla, U. Neugebauer, G. P. Niccolai, F. Lefebvre,
F. Bayard, J. Corker, S. Fiddy, J. Evans, J. P. Broyer,
J. Malinge and J. M. Basset, J. Am. Chem. Soc., 2001, 123,
3520–3540.
4 J. C. W. Chien, Top. Catal., 1999, 7, 23–36.
5 P. C. Lebaron, Z. Wang and T. J. Pinnavaia, Appl. Clay Sci.,
1999, 15, 11–29.
6 J. Liu, W.-J. Boo, A. Clearfield and H.-J. Sue, Mater. Manuf.
Processes, 2006, 21, 143–151.
7 K. Tangjituabun, B. Jongsomjit and P. Praserthdam, Catal.
Commun., 2009, 10, 1319–1323.
8 W. Kaminsky and C. Strübel, J. Mol. Catal. A: Chem., 1998,
128, 191–200.
9 N. V. Semikolenova and V. A. Zakharov, Macromol. Chem.
Phys., 1997, 198, 2889–2897.
lysts produced HDPE with melting temperatures in the range 10 M. C. Sacchi, D. Zucchi, I. Tritto, P. Locatelli and
of 130–135 °C. T. Dall’Occo, Macromol. Rapid Commun., 1995, 16, 581–590.
Apart from these results, the direct grafting of the fluorene 11 J. R. Severn, J. C. Chadwick, R. Duchateau and
on the support surface led to catalyst 1@MCM-41 which N. Friederichs, Chem. Rev., 2005, 105, 4073–4147.
assembles in an undesired face-down conformation with the 12 B. Y. Lee and J. S. Oh, Macromolecules, 2000, 33, 3194–3195.
surface. Once the metal center becomes hindered, it is limited 13 V. Varga, M. Večeřa, R. Gyepes, J. Pinkas, M. Horáček,
in its conformational freedom on the surface, preventing the
monomer coordination and no polymer can be obtained in
J. Merna and M. Lamač, ChemCatChem, 2017, 9, 3160–
3172.
this case. (ii) The catalyst must be chemically stable and not be 14 J. Pinkas and M. Lamač, Coord. Chem. Rev., 2015, 296, 45–90.
cleaved during the catalytic cycle as in the case of catalyst 4, as 15 V. Varga, M. Horáček, Z. Bastl, J. Merna, I. Císařová,
seen by NMR and UV-Vis right after the Zr addition, and if
detached from the surface, it needs to be kept assembled.
Supporting the homogeneous Oct-amido catalyst (5) on in-
J. Sýkora and J. Pinkas, Catal. Today, 2012, 179, 130–139.
16 K. Melis, V. De D., P. Jacobs and F. Verpoort, J. Mol. Catal.
A: Chem., 2001, 169, 47–56.
organic supports affects the active site and changes how the 17 M. Kwanten, B. a. M. Carrière, P. J. Grobet and P. a. Jacobs,
catalyst polymerized ethylene and propylene. Once the catalyst Chem. Commun., 2003, 1508.
was supported on the layered supports magadiite and AlPO- 18 M. De Bruyn, M. Limbourg, J. Denayer, G. V. Baron,
kanemite, the active site became sterically hindered and the
V. Parvulescu, P. J. Grobet, D. E. De Vos and P. A. Jacobs,
catalyst could no longer incorporate the α-olefins produced
Appl. Catal., A, 2003, 254, 189–201.
in situ into the polymer back bone. This was shown by both an 19 H. G. Alt, J. Chem. Soc., Dalton Trans., 1999, 1703–1709.
increase in melting temperatures and a lack of branching 20 E. I. Iiskola, S. Timonen, T. T. Pakkanen, O. Härkki,
peaks in carbon NMR.
P. Lehmus and J. V. Seppälä, Macromolecules, 1997, 30,
2853–2859.
21 M. Benaglia, A. Puglisi and F. Cozzi, Chem. Rev., 2003, 103,
3401–3429.
22 H. O. Pastore, M. Munsignatti, D. R. S. Bittencourt and
M. M. Rippel, Microporous Mesoporous Mater., 1999, 32,
211–228.
Conflicts of interest
There are no conflicts to declare.
23 G. B. Superti, E. C. Oliveira, H. O. Pastore, A. Bordo,
C. Bisio and L. Marchese, Chem. Mater., 2007, 19, 4300–
4315.
Acknowledgements
The authors acknowledge the Funding agencies that have sup- 24 D. L. Felix, M. Strauss, L. C. Ducati and H. O. Pastore,
ported this work: Conselho Nacional de Desenvolvimento Microporous Mesoporous Mater., 2009, 120, 187–194.
Científico (CNPq, grant number 140737/2013-8) and 25 H. I. Meléndez-Ortiz, Y. A. Perera-Mercado, L. A. García-
Coordenação de Aperfeiçoamento de Pessoal de Nível Superior
(CAPES, grant number 7266/15-4).
Cerda, J. A. Mercado-Silva and G. Castruita, Ceram. Int.,
2014, 40, 4155–4161.
This journal is © The Royal Society of Chemistry 2018
Dalton Trans.

View Article Online
Paper
Dalton Transactions
26 H. O. Pastore, M. Munsignatti and A. J. S. Mascarenhas, 36 E. Fernández López, V. Sánchez Escribano, M. Panizza,
Clays Clay Miner., 2000, 48, 224–229.
27 L. J. Irwin, J. H. Reibenspies and S. A. Miller, J. Am. Chem.
Soc., 2004, 126, 16716–16717.
28 S. A. Miller and J. E. Bercaw, Organometallics, 2004, 23,
1777–1789.
29 G. B. Galland, R. F. de Souza, R. S. Mauler and F. F. Nunes,
Macromolecules, 1999, 32, 1620–1625.
30 L. J. Irwin, J. H. Reibenspies and S. A. Miller, J. Am. Chem.
Soc., 2004, 126, 16716–16717.
31 W. Li, H. Chung, C. Daeffler, J. A. Johnson and
R. H. Grubbs, Macromolecules, 2012, 45, 9595–9603.
M. M. Carnasciali and G. Busca, J. Mater. Chem., 2001, 11,
1891–1897.
37 X. Cheng, O. W. Lofthus and P. A. Deck, J. Mol. Catal. A:
Chem., 2004, 212, 121–126.
38 B. Marler, U. Oberhagemann, S. Vortmann and H. Gies,
Microporous Mater., 1996, 6, 375–383.
39 M. Yadavi, A. Badiei, G. Ziarani and A. Abbasi, Chem. Pap.,
2013, 67, 751–758.
40 A. C. dos Ouros, C. Favero, M. O. de Souza, R. F. de Souza
and H. O. Pastore, J. Mol. Catal. A: Chem., 2016, 422,
59–68.
32 C. Alonso-Moreno, D. Pérez-Quintanilla, D. Polo-Cerón, 41 S. Zakrzewska, A. C. Dos Ouros, N. Schramm, D. Jehnichen,
S. Prashar, I. Sierra, I. del Hierro and M. Fajardo, J. Mol.
Catal. A: Chem., 2009, 304, 107–116.
L. Häussler, D. Pospiech, B. Voit, U. Schulze and
H. O. Pastore, Eur. Polym. J., 2015, 65, 238–251.
33 P. I. Ravikovitch and a. V. Neimark, Langmuir, 2000, 16, 42 H. M. Moura, N. L. Gibbons, S. A. Miller and H. O. Pastore,
2419–2423.
2D-Aluminum-Modified Solids as Simultaneous Support
and Cocatalyst for In Situ Polymerizations of Olefins.
Submitted to Journal of Catalysis (Ref: JCAT-17-1511).
34 The dimensions of the organic structures were simulated
using the WebMO freeware developed by: J. R. Schmidt and
W. F. Polik, WebMO Enterprise, Version 18.0.002e; WebMO 43 R. D. Broene and S. L. Buchwald, Science, 1993, 261, 1696–
LLC: Holland, MI, USA, 2018, available at http://www.
webmo.net (accessed in 01/08/2018).
1701.
44 S. Sahi and S. Paul, Green Chem. Lett. Rev., 2012, 5, 343–
35 A. Ramanathan, B. Subramaniam, R. Maheswari and
351.
U. Hanefeld, Microporous Mesoporous Mater., 2013, 167, 45 H. Balcar and J. Čejka, Coord. Chem. Rev., 2013, 257, 3107–
207–212.
3124.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2018