- Email: [email protected]
Facile high-temperature synthesis of weakly entangled polyethylene using a highly activated Ziegler-Natta catalyst
Journal of Catalysis 360 (2018) 145–151
Contents lists available at ScienceDirect
Journal of Catalysis journal homepage: www.elsevier.com/locate/jcat
Facile high-temperature synthesis of weakly entangled polyethylene using a highly activated Ziegler-Natta catalyst Wei Li a,⇑, Lei Hui a, Bing Xue a,b, Chuanding Dong c, Yuming Chen d, Linxi Hou b, Binbo Jiang d, Jingdai Wang d, Yongrong Yang d a
Department of Polymer Science and Engineering, School of Material Science and Chemical Engineering, Ningbo University, Ningbo 315211, Zhejiang, PR China School of Chemical Engineering, Fuzhou University, Fuzhou 350100, Fujian, PR China Institut für Physik and Institut für Mikro- und Nanotechnologie, Technische Universität Ilmenau, 98693 Ilmenau, Germany d School of Chemical and Biochemical Engineering, Zhejiang University, Hangzhou 310027, Zhejiang, PR China b c
a r t i c l e
i n f o
Article history: Received 16 November 2017 Revised 17 January 2018 Accepted 24 January 2018
Keywords: Heterogenous catalysts Chain entanglements Density functional theory Ziegler-Natta catalyst
a b s t r a c t A Polyhedral oligomeric silsesquioxane (POSS) modified Ziegler-Natta catalyst is reported to synthesize the weakly entangled ultra-high molecular weight polyethylene at a high temperature and with a high activity. Incorporation of POSS can form numbers of aggregates with a mean size of 48 nm adsorbing on the surface of SiO2. The structure of nanoaggregates is evidenced by the experiments and theoretical calculations where one POSS can coordinate with multiple MgCl2 molecules serving as an electric donor to the further immobilized TiCl4. The catalyst is thus featured by two distinct active regions positioned in the POSS/MgCl2 nanoaggregates and d-MgCl2, respectively. The former regions present extremely low activity on ethylene polymerization, which function as horizontal separators for isolating the active TiCl4 sites and the growing chains. This catalyst exhibits an exceptional activity (i.e. 1.3 106 gPE 1 bar1) for the synthesis of weakly entangled polyethylene at 60 °C. mol1 Ti h Ó 2018 Elsevier Inc. All rights reserved.
1. Introduction The balance of mechanical and processable property is one of the key issue for developing high performance polyolefin. To this end, numerous studies have focused on the precise control of microstructures such as molecular weight and distribution, branch degree and distribution [1]. Recently, the chain entanglement of nascent polymers, caused by the intertwined chains during the polymerization, is highlighted and is proved to show significantly influence on the mechanical, thermal and processable properties of the final products [2]. Some approaches have shown validity to synthesize the nascent polyethylene (PE) with a weakly entangled state, where the growing chains are separated from each other by diluted, compartmented or isolated active species [2]. For instance, our group employed a polyhedral oligomeric silsesquioxane (POSS) modified silica for immobilizing fluorinated bis(phenoxyimine)Ti complexes. The POSSs were chemically bonded on the silica surface, serving as horizontal spacers between the active sites to reduce the chain overlap in polymerization [3]. By this modification, the ⇑ Corresponding author. E-mail address: [email protected] (W. Li). https://doi.org/10.1016/j.jcat.2018.01.024 0021-9517/Ó 2018 Elsevier Inc. All rights reserved.
weakly entangled polyethylene (PE) can be synthesized by the heterogeneous catalyst with a considerable activity. To date, all the strategies for synthesizing the weakly entangled polyethylene are conducted below 30 °C at which a fast crystallization rate (even faster than the chain growth rate) can be achieved [2]. This fast crystallization rate can make the growing chains to be crystallized as soon as they are growing out, controlling a single chain to form a single crystal [2a,2b]. However, this low temperature will either lower the catalytic activity or harden the removal of heat transfer, making the process highly energy-consuming and difficult to be scaled-up [4]. Importantly, the growth rate of the polyethylene crystal presents exponential decrement when the temperature is increased from 30 to 90 °C [2b,5]. This fact takes a big challenge for synthesizing the weakly entangled polyethylene above 60 °C which is the economically critical temperature for industrial polymerization process. Heterogeneous catalysts play a very important role in the synthesis of polyolefin and show great potential of being transferred to industrial continuous processes thanks to its efficiency of directing the polymers morphology and preventing reactor fouling [4,6]. The active species are dispersed randomly on the porous supports depending on the location of nucleophilic groups [2a]. It was reported that the onset of chain folding was from 65 to 150
146
W. Li et al. / Journal of Catalysis 360 (2018) 145–151
carbon atoms for solution-crystallized linear n-alkanes [5]. Thus, the close distance between active species will result in a higher probability of chain overlap, especially before the formation of crystals [2,3]. We report here a facile method of modifying the traditional Ziegler-Natta (ZN) catalyst by the POSS, which has been proven effective in producing the weakly entangled PE at a high temperature with an exceptional activity (Scheme 1). The POSS molecule has a large molecular weight (i.e. 891.72 g/mol) and contains only two AOH groups (See Fig. S1 for the structure of POSS which is an incompletely condensed silsesquioxane). The incorporation of POSS contributes a large weight fraction but only a small increment of the mole fraction of nucleophilic groups (i.e. hydroxyl). We show that the POSS can capture the MgCl2 and form POSS/MgCl2 nanoaggregates which can isolate the active sites and hinder the chain overlap. Considering the critical length scale of polymer chains for crystallization (ranging from 8.5 nm to 19.3 nm) [5], the size and the density of the obstructer must reach a certain level to be able to effectively prevent the formation of entanglement. A macropores SiO2 with an open-framework structure (Macro-SiO2) is chosen as the support [7], since its open structure is helpful for (i) observing the assembly of POSS/MgCl2 nanoaggregates, and (ii) imposing less confinement to the chain crystallization and propagation [7], thus amplifying the influence of obstructers on the entanglement formation (See Fig. S2 for the morphology of Macro-SiO2).
2. Experiments 2.1. Materials All manipulations of air-sensitive and moisture-sensitive compounds were conducted under an inert nitrogen atmosphere using standard Schlenk techniques or in a glovebox. The macroporessilica (average pore diameter = 0.9 lm, porosity = 90.8%), was synthesized following our previous work and was marked as MacroSiO2 [7]. 955 silica was purchased from Grace Davison Company (China). Disilanolisobutyl POSS was purchased from Hybrid Company (USA) and dried for 24 h before use. The Macro-SiO2 and 955 silica were activated at 600 °C for 5 h with nitrogen flow before use. The titanium (IV) chloride (TiCl4, 99.9 wt%) was supplied by the Acros Organics. The anhydrous magnesium chloride (MgCl2) was purchased from the Aladdin Chemical Reagent. The 1,4-butanediol (BD) was obtained from the Aladdin Chemical Reagent, and was dried over molecular sieves for 2 days before use. The triethylaluminium (TEA) (1 mol/l in hexane) was purchased from the J&K Chemical Corp. The tetrahydrofuran (THF), n-heptane and n-hexane were distilled over the sodium/diphenyl ketone before use.
2.2. Catalyst preparation 0.5 g of MgCl2 and POSS were stirred with 30 ml of THF at 60 °C for 2 h, achieving MgCl2/POSS/THF solution. The 0.5 g of MacroSiO2 was mixed with 0.25 ml of BD and 20 ml of THF in another Schlenk flask at 40 °C for 2 h, obtaining Macro-SiO2/BD/THF. The weight fraction of POSS to the Macro-SiO2 was varied as 0, 10, 20, 30, 50 wt%. The corresponding mole ratio of [POSS]/[MgCl2] was 0, 0.01, 0.02, 0.03 and 0.05, respectively. The prepared MgCl2/ POSS/THF solution was then added into the Macro-SiO2/BD solution. The mixture was stirred for another 2 h. The obtained solids were filtered and washed three times with 30 ml of n-hexane, and dried under vacuum for 3 h, achieving the supports (SupPOSS-X, where X (=0, 10, 20, 30 and 50) represented the nominal loading of POSS). Subsequently, 30 ml of n-hexane and 5 ml of TEA were mixed with 1.0 g of Sup-POSS-X to remove the coordinated BD. The slurry was stirred at 60 °C for 2 h and washed with 30 ml of n-hexane three times. Finally, 30 ml of n-hexane and 1 ml of TiCl4 were added and stirred for another 2 h at 60 °C in order to synthesize the catalyst. The catalyst, labelled as Cat-POSS-X, was washed three times with 30 ml of n-hexane to remove the extra TiCl4 and then dried under vacuum for 3 h. For the synthesis of 955 silica supported catalyst, all the procedure was the same with Cat-POSS-10, only changing the Macro-SiO2 to this commercial 955 silica. In order to study the reaction between the POSS, MgCl2 and TiCl4, the POSS/MgCl2 adducts and POSS/MgCl2/TiCl4 adducts were synthesized with equimolar quantity of these molecules. For synthesizing the POSS/MgCl2 and the POSS/MgCl2/TiCl4, 1 mmol of MgCl2 was dissolved into 30 ml of THF and was then added droplets into the POSS solution (1 mmol of POSS solved into 30 ml of THF). The mixture was further stirred for 2 h at 60 °C. The solution was dried under vacuum for another 3 h at 60 °C achieving solids. The solids were dissolved into 30 ml of n-hexane again and then filling out the unsolvable solids in order to remove the uncoordinated MgCl2. The solution was dried achieving a solid named as POSS/MgCl2. The POSS/MgCl2 was dissolved into 30 ml of nhexane achieving a transparent solution. 1 mmol of TiCl4 was added into the solution. The solution was transferred to the yellowish supernatant and some solids was seed out after 4 h. The solid was washed by the n-hexane 3 times to remove the POSS/ TiCl4 (formed by the chemical bonding between hydroxyl and TiACl bonds). This solid was further dried under vacuum for 3 h at 50 °C, achieving the POSS/MgCl2/TiCl4 catalysts (Cat-1). An excessive amount of TiCl4 is incorporated (i.e., 8 mmol) for synthesizing the Cat-2. The POSS/TiCl4 catalysts were synthesized as following and used for a benchmark. 1 mmol of POSS and 2 mmol of TiCl4 are mixed with 30 ml of n-hexane, achieving a homogenous solution. This solution was directly used as a homogenous catalyst for ethylene polymerization. The catalyst (i.e., TiCl4 immobilized on the POSS/MgCl2 modified Macro-SiO2) was synthesized following the preparation method of Cat-POSS-X. 0.2 g of the POSS/MgCl2 was dissolved into 30 ml of THF, instead of the excessive MgCl2 in the Sup-POSS-X. 0.2 g of Macro-SiO2 and 0.1 ml of BD were also used. 2.3. Ethylene polymerization
Scheme 1. POSS modified ZN catalysts for synthesizing weakly entangled PE.
Ethylene polymerization was carried out in a 1.0 L Buchi stainless steel autoclave reactor, equipped with a mechanical stirrer and a temperature control equipment. The reactor was heated above 130 °C under vacuum for more than 3 h and repeatedly purged with nitrogen before polymerization. Then, the reactor temperature was reduced to 60 °C. 350 ml of n-heptane was added to the reactor. After the introduction of TEA as the cocatalyst, 12 ± 2 mg of the catalyst (0.025 mmol of [Ti]) was injected into the reactor.
W. Li et al. / Journal of Catalysis 360 (2018) 145–151
The polymerization was carried out under a continuous ethylene flow to meet 3 bars at a stirring rate of 450 rpm. The polymerization proceeded for 60 min. At the end of the polymerization, the autoclave was quickly vented. The synthesized product was precipitated and washed with acidified ethanol (5 wt% hydrochloric acid), and dried at 60 °C under vacuum for 12 h. 2.4. Density functional theory (DFT) simulation The structural relaxations for the systems 1 POSS + 1MgCl2 and 1 POSS + 1MgCl2 + 1TiCl4 were carried out by using DFT calculations. To find the energetically preferred configurations, various relative positions and orientations between the molecules are considered for each system. The calculations used 6-31G⁄ basis set [8] and B3LYP exchange-correlation functional [2]. All the DFT calculations were performed with GAUSSIAN09 package. 2.5. Characterization The particle size distribution (PSD) of POSS was determined by dynamic light scattering (Zetasizer 3000HSA, Malvern, UK) at room temperature. THF was used as the dispersion. The sample were sonicated for 30 min before the measurement. The average particle size of POSS was 9.5 nm (See the PSD curve of POSS in Fig. S3). The morphology of the supports and catalysts were observed using a transmission electron microscopy (TEM, Tecnai F20, USA) and scanning electron microscopy (SEM, Hitachi S-4700, Japan). The catalyst was first mixed with n-hexane and then was put into ultrasound for 15 min. One droplet of the mixture are added into the copper screen and was dried in the glove box. The dried copper screen was used for TEM observing. The Energy dispersive X-ray detector (EDX) was used for analysing the elemental composition. The red points in the Fig. 1c are the detected points. The samples were sputter-coated by Pt before the measurement of SEM. The
147
distribution of nanoaggregates was calculated by Image J software. The specific surface area of Macro-SiO2 was measured by an Automatic Mercury Porosimeter (PoreMaster-60, Quantachrome, USA) at a pressure range of 0.3–30,000 psi. The contact angle and surface tensions were 130° and 0.480 N/m, respectively. Powder X-ray diffraction (XRD) measurements were carried out on a Bruker GADDS diffract meter with the Cu Ka radiation at 40 kV and 40 mA (k = 0.154 nm, 5–60°). The step size was 0.02°. X-ray photoelectron spectroscopy (XPS) measurements were carried out on Escalab 250Xi using Al Ka radiation as the X-ray source (300 W, 1486.3 eV). The vacuum chamber was about 5 109 Torr. The binding energy was calibrated to C1s peak at 284.6 eV. 1H NMR spectra of the POSS, POSS/MgCl2 and POSS/MgCl2/TiCl4 were obtained on a Bruker AC-80 400 MHz spectrometer at ambient temperature. Deuterated chloroform (CDCl3) was used as the solvent for all the samples in the NMR measurement. Tetramethylsilane (TMS) was used as the internal standard. The samples for NMR measurements were set in the sealed NMR tube for isolating the water and oxygen. The titanium content of the prepared catalysts was measured by a ultraviolet–visible (UV–vis) measurement through the hydrogen peroxide colorimetric method. The supported catalysts were dissolved in 5 ml of H2SO4 solution (1 mol l1), and 2 ml of H2O2, and then the solution was diluted with an H2SO4 solution to 25 ml. The UV measurement was performed by a TU-1901 spectrophotometer (PERSEA Corp, China). The intensity of the peak at 410 nm was used to quantify the titanium content. The weight-average molar mass (Mw) and the molecular weight distribution (MWD) were measured by a gel permeation chromatography (GPC) at 150 °C with a PL-GPC-220 instrument (Polymer Laboratories, Shropshire, U.K.). The molecular weight of standard polystyrene ranges from 1000 to 14,000,000 g/mol. 1,2,4-trichlorobenzene was used as the solvent. X-ray diffraction (XRD) measurements of nascent PE were carried out on a Bruker GADDS diffract meter with an area detector operating under
Fig. 1. (a) SEM morphology, (b) the particle size distribution of the nanoaggregates, (c) TEM morphology and (d) specific surface area of the POSS modified ZN catalysts. The catalysts are labelled as Cat-POSS-X, where X (=0, 10, 20 and 50) represents the nominal loading of POSS. (a) and (c) are from Cat-POSS-50, serving as examples.
148
W. Li et al. / Journal of Catalysis 360 (2018) 145–151
40 kV and 40 mA, using Cu Ka radiation (k = 0.154 nm) and a step of 0.02° min1. The XRD curves were decomposed into three components using Origin 8.5 in order to calculate the linear crystallinity of the nascent polymers. Lorentz function was used for fitting the curves [3]. Differential scanning calorimeter (DSC) measurements were performed with a DSC-7 instrument (PerkinElmer Corp., USA) to measure the melting point and crystallinity of products. Samples (ca. 5 mg) were first heated from 50 to 160 °C at 10 °C min1 and then held for 3 min to release thermal history. Subsequently, they were cooled to 50 °C at 10 °C min1 and then held for 3 min. The second heating cycle was the same as the first cycle. Melting temperature was taken at the peak of the endotherm. The crystallinity was calculated by comparison with the heat of fusion for polyethylene with a perfectly crystalline, 289 J/g [2a,3]. Rheological studies were performed on a straincontrolled rheometer (HAAKE III instrument) for analysing the entanglements of nascent polymers. A disk of 20 mm diameter was compressed under 20 tons at 120 °C for 30 min and used in all rheological studies [3]. The disk between the parallel plates of the rheometer was heated to 160 °C under a nitrogen environment to prevent thermo-oxidative degradation. After thermal stabilization at 160 °C (5 min), the rheology experiments were started. The dynamic amplitude sweep test was performed at a fixed frequency of 1 Hz to determine the linear viscoelastic regime. The dynamic time sweep test was performed to follow the entanglement formation at a fixed frequency of 1 rad/s and strain in the linear viscoelastic regime.
3. Results and discussion Typical surface morphologies of catalysts are shown in the Fig. 1 (See detailed SEM morphologies in the Fig. S4). The incorporation of POSS can form numerous nanoaggregates with a mean size of 48.8 nm, uniformly distributing on the surface of Macro-SiO2 supports. This indicates the POSS can adsorb effectively on the silica surface (See Fig. 1a–c). A low POSS-loading (i.e. 10 wt%) can reduce the mean size of nanoaggregates, which increases the tortuosity of the pores and leads to a higher specific surface area (SSA, See Fig. 1d) [3,9]. Further increasing the POSS loading (i.e. larger than 20 wt%) changes little on the size of the nanoaggregates but accumulated in the pores (See Fig. 1c). This accumulation leads to a linear decrement of the SSA [9]. The chemical composition of the nanoaggregates is further characterized by the TEM-EDX, where the concentration of Ti, Mg and Si atoms are 0.95, 7.1 and 25.0 wt%, respectively (see the red points in Fig. 1c). This indicates that the formed nanoaggregates may contain POSS, MgCl2, and can also immobilize the Ti complex. Fig. 2a, b shows the XRD curves of POSS modified Macro-SiO2 supports and their immobilized ZN catalysts. The XRD curves of pure POSS and a-MgCl2 are shown in the Fig. S5. The peaks at 2h = 10.8° (0 0 1), 15.0° (0 0 3), 22.5° (0 0 2) and 35° (0 0 4) represent the patterns of MgCl2/1,4-butanediol (BD) adducts [7,10]. The distinctly high intensity of these planes indicates that the structure of a-MgCl2, with a distorted cubic close packing of Cl atoms, is destroyed [10]. Moreover, the typical peaks of POSS (2h 10°) disappear, which indicates that the major aggregation of pure POSS are destroyed [3]. After immobilizing the TiCl4, the characteristic peaks for MgCl2/BD adducts disappear, while the peaks around 15°, 29–32° and 50.5° emerge (corresponding to the resonance of d-MgCl2). This d-MgCl2 prefers to immobilize the active titanium, which will present a high activity on the ethylene polymerization [10]. The peaks around 40.8° are shown both in the POSS modified supports and catalysts. The intensities of these peaks are dependent on the POSS loading in the supports, which gives another evidence of the formation of nanoaggregates [3,10]. More
importantly, the high intensities of these peaks are maintained for all the Cat-POSS-X systems, suggesting that the incorporation of TEA and TiCl4 cannot destroy the structure of nanoaggregates [10,11]. Surface element analysis of Si atoms of the support is shown in Fig. 2c, d. The Si 2p curves can be separated into two peaks at 101.3 and 103.1 eV (see Fig. 2c), which correspond to the surface Si atoms and bulk Si atoms, respectively [3]. interestingly, (see Fig. 2d), the concentration of surficial Si reduces linearly with the increment of POSS loading. This may be attributed to the reason that the MgCl2 molecules migrate to the surfaces of the nanoaggregates and cover on them, thus decreasing the amount of surficial Si atoms. To explore the interactions between POSS, MgCl2 and TiCl4 during the catalyst immobilization, model systems with equimolar quantity of these molecules are favorable. To this end, we performed both NMR measurements for POSS/MgCl2/TiCl4 mixtures with equal mole ratio, and density functional theory (DFT) calculations [8] for systems containing one POSS and one MgCl2 molecule (1 POSS + 1 MgCl2), as well as those containing one additional TiCl4 molecule (1 POSS + 1MgCl2 + 1 TiCl4) (See Fig. 3). The simulated results show that the O atoms of SiAOASi and the hydroxyl of POSS can be coordinated with the Mg atoms, forming conformation A, B or C (See Fig. 3a). The conformation C (i.e. 1 MgCl2 coordinated with the double hydroxyl of POSS) presents the lowest total energy (TE) and highest binding energy (BE), representing the most thermally stable conformation [12]. Conformation B, where one Mg atom binds to one hydroxyl, presents a BE of 34.1 kcal/mol. The coordination between the Mg atoms and the O atoms of SiAOASi bonds (conformation A) also shows a considerable BE (i.e. 15.9 kcal/mol). Considering the excess amount of the MgCl2 in the Cat-POSS-X (i.e. the mole ratio of [POSS]/[MgCl2] is 0.01–0.05), the coordination of Mg atoms with both O atoms in the hydroxyl (A, B) and those in the SiAOASi (conformation C) may exist [12]. The coordination between Mg and O is further confirmed by the NMR measurements. The chemical shift at 3.9 ppm (halo peak) reflects the resonance of hydroxyl groups of POSS (See Fig. 3b, POSS) [3b]. This resonance becomes narrow and shifts to 4.1 ppm with the incorporation of MgCl2 (See Fig. 3b, POSS/MgCl2 is synthesized by the nominally equimolar quantity), since the coordination of ClAMgAOH enhance the Lewis acid of the proton in the hydroxyl groups [10]. The CAH in the i-butyl of POSS, with the chemical shift of 1.8 ppm (See Fig. 3c), is far away from the O atom of hydroxyl, serving as an internal standard. The mole ratio of [H]C–H/[OH] presents independence on the incorporation of MgCl2, indicating that no H is detached after the incorporation of MgCl2. The multiple peaks around 0.6 ppm correspond to the resonance of CH2 which is bonded with Si atoms (See Fig. 3c). Compared with the pure POSS, a new peak presents at a lower chemical shift and the mole ratio of [POSS]/[Mg] is measured to be 5 (measured by SEM-EDX [1c]). in the POSS/MgCl2. This demonstrates that only partial POSSs are coordinated with the MgCl2 through MgAOASi, which increase the electron density of the CH2. The uncoordinated MgCl2 will give a chance to form the d-MgCl2 in the Cat-POSS-X systems. The reaction between the POSS, MgCl2 and TiCl4 are further investigated by experiments with the equimolar quantity of the reactants as well as DFT calculations. Importantly, the pure POSS and coordinated POSS/MgCl2 adducts is n-hexane soluble. After reacting with the TiCl4, the POSS/MgCl2 adducts becomes nhexane insoluble (forming POSS/MgCl2/TiCl4) thanks to the increased polarity, whereas the POSS, uncoordinated with the MgCl2, will form POSS/TiCl4 upon the reaction with hydroxyl (forming SiAOATi). The POSS/MgCl2/TiCl4 solid (i.e. Cat-1) can be separated, and the POSS/TiCl4 can be leaching-out owing to its excellent solubility in the n-hexane. Excitingly, the mole ratio of [Mg]/[POSS] (measured by SEM-EDX [1c]) is 7:1 in the Cat-1
W. Li et al. / Journal of Catalysis 360 (2018) 145–151
149
Fig. 2. (a) XRD curves of the POSS modified Macro-SiO2, (b) XRD curves of the POSS modified catalysts. (c), (d) XPS analysis of the catalysts. The supports and catalysts are labelled as Sup-POSS-X and Cat-POSS-X, respectively.
Fig. 3. (a) DFT calculations for the systems of POSS and MgCl2 with the equimolar quantity, (b), (c) NMR studies of POSS, POSS/MgCl2 and POSS/MgCl2-TiCl4 (Cat-1 and Cat-2). The initial mole ratio of TiCl4/POSS is 1.0 in the Cat-1. The initial dosage of TiCl4 is excessive in the Cat-2.
150
W. Li et al. / Journal of Catalysis 360 (2018) 145–151
although their nominal loading is 1:1, demonstrating that the O atoms of SiAOASi must participate in the coordination with the Mg atoms. This gives the evidence that the MgCl2 can cover the surface of POSS, decreasing the surficial Si concentration in the XPS results (See Fig. 2d). Here, we must mention that the composition of the nanoaggregates in the Cat-POSS-X measured by TEM-EDX. Supposing all the Si atoms in the red points (Fig. 1c) is attributable to the POSS, with which give the highest mole ratio of [POSS]/[Mg] equal to 2.7. This value indicates the nanoaggregates of Cat-POSS-X systems contain the coordinated POSS/MgCl2 and the uncoordinated POSS. Based on the stoichiometry, incorporating the TiCl4 with the equimolar quantity of POSS (namely half quantity of the hydroxyl) results in a disappearance of the –OH (3.9 ppm) and a split of CAH resonance in the Cat-1 (see Fig. 3d). Excessively incorporated TiCl4 takes no variance on the NMR spectrum (See Cat-2 in Fig. 3b, d). These demonstrate that the TiCl4 is able to react with the hydroxyl of the POSS/MgCl2 (forming the reacted conformation, see Fig. S6). The coordinated MgCl2 can also coordinate/immobilize the TiCl4 (forming coordinated conformation, see Fig. S6) because the molar ratio of [Ti]/[POSS] is measured to be 4 in the Cat-1. This fact is further evidenced by the DFT calculations where the reacted and coordinated conformation show similar BE and TE (See Fig. S6). Therefore, the loading amount of TiCl4 may be higher in the CatPOSS-X compared with that of Cat-POSS-0. The catalytic activity of POSS/MgCl2/TiCl4 with the excessive dosage of TiCl4 (Cat-2) is studied since the amount of TiCl4 is excessive for synthesizing the Cat-POSS-X. The Cat-2 presents a low activity on the ethylene polymerization (i.e. 2.0 104 gPE mol1 Ti h1 bar1). The POSS/TiCl4 catalyst shows a much lower activity 1 (i.e. 2.2 102 gPE mol1 bar1). The Immobilization of TiCl4 on Ti h the POSS/MgCl2 modified Macro-SiO2 shows a considerable Ti loading (i.e. 1.0 wt%) but no catalytic activity. Thus, POSS/MgCl2 adducts are suitable to serve as horizontal spacers to isolate the active species loading on the d-MgCl2 [3]. Table 1 shows the immobilization results of TiCl4 on the MacroSiO2/POSS support. The Incorporation of POSS indeed shows a higher TiCl4 loading than the Cat-POSS-0 as we mentioned above.
The decrement of Ti loading with the increased POSS amount may be due to the reduced SSA. The BE of Mg 2p are decreased in the Cat-POSS-X compared with the case of Cat-POSS-0, suggesting that the electron density of Mg atoms is increased by the coordination between MgCl2 and POSS [3,11]. The increased BE of Ti2p1/2 suggests the presence of electron deficient Ti atoms owing to the electron donor effects of POSS [3]. The multiple chemical environments of TiCl4 broaden the Ti2p1/2 peak in the POSS modified catalysts. The results of ethylene polymerization are shown in Table 2. The Cat-POSS-0 exhibits a lower activity than the POSS modified 1 one. A super high activity (1.3 106 gPE mol1 bar1) is Ti h achieved with Cat-POSS-20, because the MgCl2/POSS nanoaggregates serve as horizontal spacers to enlarge the distance between the active sites and therefore inhibiting bimetallic deactivation [3]. The decreased activity of Cat-POSS-30 and Cat-POSS-50 can be explained by the MgCl2/POSS aggregators immobilizing more deactivated TiCl4, and also by the decreased SSA of the supports [3,10b]. The synthesized polyethylene presents a high molecular weight (i.e. larger than 151 104 g/mol), indicating the production of ultra-high molecular weight polyethylene (UHMWPE). Interestingly, the synthesized polymers by the POSS modified catalysts present high crystallinities (i.e. Xc,XRD 81.9% and Xc,DSC 62.5%), indicating a highly ordered structure of the nascent PE. The nascent UHMWPE shows a high melting point (larger than 144.0 °C), which is usually considered as a reflection of chain-extended crystal with the thickness more than 1 lm (See Fig. S7 for the DSC curves) [2a]. The high crystallinity and high melting point are lost during the second heating scan, indicating the weakly entangled state of the nascent polyethylene [2a,3]. This weakly entangled sate of nascent PE is further evidenced by the rheology. Using the nascent PE-POSS-50 one can achieve the most weakly entangled state where the modulus build-up curves shows the lowest starting modulus and takes the longest time to reach the thermo-dynamically stable state (see Fig. 4) [2a,3]. To the best of our knowledge, this is the first time that the weakly entangled UHMWPE can be synthesized by the ZN catalyst at 60 °C with a high activity, showing a great potential to realize the ‘drop-in’
Table 1 Immobilization of TiCl4 on the Macro-SiO2/POSS support.
a b c
Catalyst
[POSS]/[MgCl2]a
Ti loading (wt%)b
Mg 2p BE (eV)c
Cat-POSS-0 Cat-POSS-10 Cat-POSS-20 Cat-POSS-30 Cat-POSS-50
0 0.01 0.02 0.03 0.05
6.9 8.0 7.7 7.5 7.3
51.6 51.0 51.0 51.0 51.0
Ti 2p1/2c BE (eV)
FWHM (eV)
464.4 464.9 464.9 464.8 464.8
2.5 2.8 2.9 2.8 2.8
Initial mole ratio of POSS to MgCl2. Measured by ICP. Measured by XPS.
Table 2 Ethylene polymerization results of the catalyst.a
a b c d e
Catalyst
PE yield (g)
Activityb
Mwc
MWD
Xc,XRDd %
Xc1,DSCe %
Tm1 °C
Xc2,DSCe %
Tm2 °C
Cat-POSS-0 Cat-POSS-10 Cat-POSS-20 Cat-POSS-30 Cat-POSS-50
36.0 81.5 97.9 47.9 42.7
480 1168 1305 638 570
181.0 159.0 205.3 151.0 158.0
6.0 8.4 9.5 8.9 9.1
78.5 82.1 81.9 81.9 84.6
62.5 64.2 65.5 66.9 66.4
143.7 144.0 144.2 144.3 144.7
46.5 47.5 53.3 52.9 54.3
136.5 137.0 137.2 137.5 137.8
Polymerizations. catalysts: 0.025 mmol of [Ti], 60 min of reaction, 500 ml of toluene, 3 bar of C2H4, 60 °C and 100 M ratio of [Al]/[Ti]. 1 103 gPE mol1 bar1. Ti h 104 g/mol. Measured by XRD. Measured by DSC.
W. Li et al. / Journal of Catalysis 360 (2018) 145–151
151
Science Foundation of China No. 21376054, Prof. Yang thanks the support of Natural Science Foundation of China No. U1663222, the Project of Natural Science Foundation of Ningbo (2016A610048), The National Science Fund for Distinguished Young, 21525627, and sponsorship by the K. C. Wong Magna Fund in Ningbo University are gratefully acknowledged. C. D would like to thank the computer center of Technische Universität Ilmenau for the support of all computations. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at https://doi.org/10.1016/j.jcat.2018.01.024. References Fig. 4. Dynamic time sweep test (at 160 °C) for the nascent PE with storage modulus (normalized by maximum plateau modulus G0 max). The synthesized polymer is labelled as PE-POSS-X, where X (=0, 10, 20 and 50) represents the nominal loading of POSS/MgCl2.
techniques in the industry. The reduced entanglements can be attributed to that the steric hindrance, generated by the MgCl2/ POSS nanoaggregates, can enlarge the distance between the growing chains and decrease accordingly the probability of chain overlap [2,3]. Finally, when change the Macro-SiO2 to the commercial silica with confined pores (i.e. Grace-955 SiO2), the isolators (POSS/MgCl2 nanoaggregates) can still work in the separation of active sites, where the weakly entangled UHMWPE with even higher molecular weight (i.e. 2.9 106 g/mol) can be synthesized 1 with a super high activity (i.e. 4.5 106 g PE mol1 ) at 60 °C Ti h (See Fig. S8). 4. Conclusions In summary, we have shown that the POSSs can capture the MgCl2 molecules through the coordination, serving as electron donors with a cage structure. The POSS/MgCl2 adducts can be effectively adsorbed and uniformly distributed on the surface of MacroSiO2 forming nanoaggregates with a mean size of 48 nm. The formed POSS/MgCl2 nanoaggregates can immobilize the TiCl4 and present extremely low activity on ethylene polymerization. These nanoaggregates are thus served as horizontal separators for isolating the active TiCl4 and the growing polyethylene chains. The weakly entangled UHMWPE can be synthesized by the facile modification of ZN catalyst at 60 °C with a super high activity, demonstrating a great potential to realize the ‘drop-in’ techniques. The reported results provide a new strategy of controlling the distribution of active sites in the heterogeneous catalyst and opening the perspective of preparing highly active catalyst and highperformance polyolefin. Acknowledgements The research was supported by the Natural Science Foundation of China No. 21776141, Prof. Hou thanks the support of Natural
[1] (a) D.J. Read, D. Auhl, C. Das, J.D. Doelder, M. Kapnistos, I. Vittorias, T.C.B. Meleish, Science 333 (2011) 1871–1874; (b) M. Stürzel, S. Mihan, R. Mülhaupt, Chem. Rev. 116 (2016) 1398–1433; (c) D. Cicmil, J. Meeuwissen, A. Vantomme, J. Wang, I.K. van Ravenhorst, H.E. van der Bij, A.M. Murillo, B.M. Weckhuysen, Angew. Chem. Int. Ed. 54 (2015) 13073–13079. [2] (a) S. Rastogi S, D.R. Lippits, G.W.M. Peters, R. Graf, Y. Yao, H.W. Spiess, Nat. Mater. 4 (2005) 635–641; (b) A. Osichow A, C. Rabe, K. Vogtt, T. Narayanan, L. Harnau, M. Drechsler, M. Ballauff, S. Mecking, J. Am. Chem. Soc. 135 (2013) 11645–11650; (c) M. Santoro1, F.A. Gorelli, R. Bini, J. Haines, A. Lee, Nat. Commun. 4 (2013) 1557–1563; (d) L. Kolb, V. Monteil, R. Thomann, S. Mecking, Angew. Chem. Int. Ed. 44 (2005) 429–432; (e) K. Kageyama, J. Tamazawa, T. Aida, Science 285 (1999) 2113–2115. [3] (a) W. Li, H.Q. Yang, J.J. Zhang, J.S. Mu, D.R. Gong, X.D. Wang, Chem. Commun. 52 (2016) 11092–11095; (b) W. Li, C. Guan, J. Xu, J. Mu, D. Gong, Z. Chen, Polymer 55 (2014) 1792–1798. [4] (a) M. Klapper, D. Joe, S. Nietzel, J.K. Krumpfer, K. Müllen, Chem. Mater. 26 (2014) 802–819; (b) W. Kaminsky, Adv. Polym. Sci. 258 (2013) 1–363. [5] (a) G. Ungar, J. Stejny, A. Keller, I. Bidd, M.C. Whiting, Science 229 (1985) 386– 389; (b) N. Waheed, G.C. Rutledge, J. Polym. Sci. Polym. Phys. 43 (2005) 2468–2473. [6] (a) T.F.L. Mckenna, A.D. Martino, G. Weickert, J.B.P. Soares, Macrom. React. Eng. 4 (2010) 40–64; (b) M.P. McDaniel, ACS Catal. 1 (2011) 1394–1407. [7] B. Xue, L. Hui, H.Q. Yang, Y.L. Zhao, L.X. Hou, W. Li, Ind. Eng. Chem. Res. 56 (2017) 135–142. [8] V.A. Rassolov, M.A. Ratner, J.A. Pople, P.C. Redfern, L.A. Curtiss, J. Comput. Chem. 22 (2001) 976–984. [9] (a) M. Choi, F. Kleitz, D. Liu, H.Y. Lee, W. Ahn, R. Ryoo, J. Am. Chem. Soc. 127 (2005) 1924–1932; (b) H. Li, B. Xu, X. Liu, W. Gao, Y. Mu, Chem. Comm. 51 (2015) 16703–16706. [10] (a) P. Sozzani, S. Bracco, A. Comotti, R. Simonutti, I. Camurati, J. Am. Chem. Soc. 125 (2003) 12881–12893; (b) R. Credendino, D. Liguori, Z.Q. Fan, G. Morini, L. Cavallo, ACS Catal. 5 (2015) 5431; (c) T. Taniike, M. Ternao, J. Catal. 293 (2012) 39–50. [11] (a) M.M.C. Forte, F.V. Cunha, J.H.Z. dos Santos, J. Mol. Cat. A Chem. 175 (2001) 91–103; (b) A. Andoni, J. Chadwick, H. Niemantsverdriet, P. Thune, J. Catal. 257 (2008) 81–86; (c) T. Taniike, T. Funako, M. Terano, J. Catal. 311 (2014) 33–40. [12] (a) D. Hossain, C.U. Pittman, S. Saebo, J. Phys. Chem. C 111 (2007) 619–6206; (b) D. Hossain, C.U. Pittman, F. Hagelberg, S. Saebo, J. Phys. Chem. C 112 (2008) 16070–16077; (c) A.R. Bassindale, M. Pourny, P.G. Taylor, M.B. Hursthouse, M.E. Light, Angew. Chem. Int. Ed. 42 (2003) 3488–3490.
Contents lists available at ScienceDirect
Journal of Catalysis journal homepage: www.elsevier.com/locate/jcat
Facile high-temperature synthesis of weakly entangled polyethylene using a highly activated Ziegler-Natta catalyst Wei Li a,⇑, Lei Hui a, Bing Xue a,b, Chuanding Dong c, Yuming Chen d, Linxi Hou b, Binbo Jiang d, Jingdai Wang d, Yongrong Yang d a
Department of Polymer Science and Engineering, School of Material Science and Chemical Engineering, Ningbo University, Ningbo 315211, Zhejiang, PR China School of Chemical Engineering, Fuzhou University, Fuzhou 350100, Fujian, PR China Institut für Physik and Institut für Mikro- und Nanotechnologie, Technische Universität Ilmenau, 98693 Ilmenau, Germany d School of Chemical and Biochemical Engineering, Zhejiang University, Hangzhou 310027, Zhejiang, PR China b c
a r t i c l e
i n f o
Article history: Received 16 November 2017 Revised 17 January 2018 Accepted 24 January 2018
Keywords: Heterogenous catalysts Chain entanglements Density functional theory Ziegler-Natta catalyst
a b s t r a c t A Polyhedral oligomeric silsesquioxane (POSS) modified Ziegler-Natta catalyst is reported to synthesize the weakly entangled ultra-high molecular weight polyethylene at a high temperature and with a high activity. Incorporation of POSS can form numbers of aggregates with a mean size of 48 nm adsorbing on the surface of SiO2. The structure of nanoaggregates is evidenced by the experiments and theoretical calculations where one POSS can coordinate with multiple MgCl2 molecules serving as an electric donor to the further immobilized TiCl4. The catalyst is thus featured by two distinct active regions positioned in the POSS/MgCl2 nanoaggregates and d-MgCl2, respectively. The former regions present extremely low activity on ethylene polymerization, which function as horizontal separators for isolating the active TiCl4 sites and the growing chains. This catalyst exhibits an exceptional activity (i.e. 1.3 106 gPE 1 bar1) for the synthesis of weakly entangled polyethylene at 60 °C. mol1 Ti h Ó 2018 Elsevier Inc. All rights reserved.
1. Introduction The balance of mechanical and processable property is one of the key issue for developing high performance polyolefin. To this end, numerous studies have focused on the precise control of microstructures such as molecular weight and distribution, branch degree and distribution [1]. Recently, the chain entanglement of nascent polymers, caused by the intertwined chains during the polymerization, is highlighted and is proved to show significantly influence on the mechanical, thermal and processable properties of the final products [2]. Some approaches have shown validity to synthesize the nascent polyethylene (PE) with a weakly entangled state, where the growing chains are separated from each other by diluted, compartmented or isolated active species [2]. For instance, our group employed a polyhedral oligomeric silsesquioxane (POSS) modified silica for immobilizing fluorinated bis(phenoxyimine)Ti complexes. The POSSs were chemically bonded on the silica surface, serving as horizontal spacers between the active sites to reduce the chain overlap in polymerization [3]. By this modification, the ⇑ Corresponding author. E-mail address: [email protected] (W. Li). https://doi.org/10.1016/j.jcat.2018.01.024 0021-9517/Ó 2018 Elsevier Inc. All rights reserved.
weakly entangled polyethylene (PE) can be synthesized by the heterogeneous catalyst with a considerable activity. To date, all the strategies for synthesizing the weakly entangled polyethylene are conducted below 30 °C at which a fast crystallization rate (even faster than the chain growth rate) can be achieved [2]. This fast crystallization rate can make the growing chains to be crystallized as soon as they are growing out, controlling a single chain to form a single crystal [2a,2b]. However, this low temperature will either lower the catalytic activity or harden the removal of heat transfer, making the process highly energy-consuming and difficult to be scaled-up [4]. Importantly, the growth rate of the polyethylene crystal presents exponential decrement when the temperature is increased from 30 to 90 °C [2b,5]. This fact takes a big challenge for synthesizing the weakly entangled polyethylene above 60 °C which is the economically critical temperature for industrial polymerization process. Heterogeneous catalysts play a very important role in the synthesis of polyolefin and show great potential of being transferred to industrial continuous processes thanks to its efficiency of directing the polymers morphology and preventing reactor fouling [4,6]. The active species are dispersed randomly on the porous supports depending on the location of nucleophilic groups [2a]. It was reported that the onset of chain folding was from 65 to 150
146
W. Li et al. / Journal of Catalysis 360 (2018) 145–151
carbon atoms for solution-crystallized linear n-alkanes [5]. Thus, the close distance between active species will result in a higher probability of chain overlap, especially before the formation of crystals [2,3]. We report here a facile method of modifying the traditional Ziegler-Natta (ZN) catalyst by the POSS, which has been proven effective in producing the weakly entangled PE at a high temperature with an exceptional activity (Scheme 1). The POSS molecule has a large molecular weight (i.e. 891.72 g/mol) and contains only two AOH groups (See Fig. S1 for the structure of POSS which is an incompletely condensed silsesquioxane). The incorporation of POSS contributes a large weight fraction but only a small increment of the mole fraction of nucleophilic groups (i.e. hydroxyl). We show that the POSS can capture the MgCl2 and form POSS/MgCl2 nanoaggregates which can isolate the active sites and hinder the chain overlap. Considering the critical length scale of polymer chains for crystallization (ranging from 8.5 nm to 19.3 nm) [5], the size and the density of the obstructer must reach a certain level to be able to effectively prevent the formation of entanglement. A macropores SiO2 with an open-framework structure (Macro-SiO2) is chosen as the support [7], since its open structure is helpful for (i) observing the assembly of POSS/MgCl2 nanoaggregates, and (ii) imposing less confinement to the chain crystallization and propagation [7], thus amplifying the influence of obstructers on the entanglement formation (See Fig. S2 for the morphology of Macro-SiO2).
2. Experiments 2.1. Materials All manipulations of air-sensitive and moisture-sensitive compounds were conducted under an inert nitrogen atmosphere using standard Schlenk techniques or in a glovebox. The macroporessilica (average pore diameter = 0.9 lm, porosity = 90.8%), was synthesized following our previous work and was marked as MacroSiO2 [7]. 955 silica was purchased from Grace Davison Company (China). Disilanolisobutyl POSS was purchased from Hybrid Company (USA) and dried for 24 h before use. The Macro-SiO2 and 955 silica were activated at 600 °C for 5 h with nitrogen flow before use. The titanium (IV) chloride (TiCl4, 99.9 wt%) was supplied by the Acros Organics. The anhydrous magnesium chloride (MgCl2) was purchased from the Aladdin Chemical Reagent. The 1,4-butanediol (BD) was obtained from the Aladdin Chemical Reagent, and was dried over molecular sieves for 2 days before use. The triethylaluminium (TEA) (1 mol/l in hexane) was purchased from the J&K Chemical Corp. The tetrahydrofuran (THF), n-heptane and n-hexane were distilled over the sodium/diphenyl ketone before use.
2.2. Catalyst preparation 0.5 g of MgCl2 and POSS were stirred with 30 ml of THF at 60 °C for 2 h, achieving MgCl2/POSS/THF solution. The 0.5 g of MacroSiO2 was mixed with 0.25 ml of BD and 20 ml of THF in another Schlenk flask at 40 °C for 2 h, obtaining Macro-SiO2/BD/THF. The weight fraction of POSS to the Macro-SiO2 was varied as 0, 10, 20, 30, 50 wt%. The corresponding mole ratio of [POSS]/[MgCl2] was 0, 0.01, 0.02, 0.03 and 0.05, respectively. The prepared MgCl2/ POSS/THF solution was then added into the Macro-SiO2/BD solution. The mixture was stirred for another 2 h. The obtained solids were filtered and washed three times with 30 ml of n-hexane, and dried under vacuum for 3 h, achieving the supports (SupPOSS-X, where X (=0, 10, 20, 30 and 50) represented the nominal loading of POSS). Subsequently, 30 ml of n-hexane and 5 ml of TEA were mixed with 1.0 g of Sup-POSS-X to remove the coordinated BD. The slurry was stirred at 60 °C for 2 h and washed with 30 ml of n-hexane three times. Finally, 30 ml of n-hexane and 1 ml of TiCl4 were added and stirred for another 2 h at 60 °C in order to synthesize the catalyst. The catalyst, labelled as Cat-POSS-X, was washed three times with 30 ml of n-hexane to remove the extra TiCl4 and then dried under vacuum for 3 h. For the synthesis of 955 silica supported catalyst, all the procedure was the same with Cat-POSS-10, only changing the Macro-SiO2 to this commercial 955 silica. In order to study the reaction between the POSS, MgCl2 and TiCl4, the POSS/MgCl2 adducts and POSS/MgCl2/TiCl4 adducts were synthesized with equimolar quantity of these molecules. For synthesizing the POSS/MgCl2 and the POSS/MgCl2/TiCl4, 1 mmol of MgCl2 was dissolved into 30 ml of THF and was then added droplets into the POSS solution (1 mmol of POSS solved into 30 ml of THF). The mixture was further stirred for 2 h at 60 °C. The solution was dried under vacuum for another 3 h at 60 °C achieving solids. The solids were dissolved into 30 ml of n-hexane again and then filling out the unsolvable solids in order to remove the uncoordinated MgCl2. The solution was dried achieving a solid named as POSS/MgCl2. The POSS/MgCl2 was dissolved into 30 ml of nhexane achieving a transparent solution. 1 mmol of TiCl4 was added into the solution. The solution was transferred to the yellowish supernatant and some solids was seed out after 4 h. The solid was washed by the n-hexane 3 times to remove the POSS/ TiCl4 (formed by the chemical bonding between hydroxyl and TiACl bonds). This solid was further dried under vacuum for 3 h at 50 °C, achieving the POSS/MgCl2/TiCl4 catalysts (Cat-1). An excessive amount of TiCl4 is incorporated (i.e., 8 mmol) for synthesizing the Cat-2. The POSS/TiCl4 catalysts were synthesized as following and used for a benchmark. 1 mmol of POSS and 2 mmol of TiCl4 are mixed with 30 ml of n-hexane, achieving a homogenous solution. This solution was directly used as a homogenous catalyst for ethylene polymerization. The catalyst (i.e., TiCl4 immobilized on the POSS/MgCl2 modified Macro-SiO2) was synthesized following the preparation method of Cat-POSS-X. 0.2 g of the POSS/MgCl2 was dissolved into 30 ml of THF, instead of the excessive MgCl2 in the Sup-POSS-X. 0.2 g of Macro-SiO2 and 0.1 ml of BD were also used. 2.3. Ethylene polymerization
Scheme 1. POSS modified ZN catalysts for synthesizing weakly entangled PE.
Ethylene polymerization was carried out in a 1.0 L Buchi stainless steel autoclave reactor, equipped with a mechanical stirrer and a temperature control equipment. The reactor was heated above 130 °C under vacuum for more than 3 h and repeatedly purged with nitrogen before polymerization. Then, the reactor temperature was reduced to 60 °C. 350 ml of n-heptane was added to the reactor. After the introduction of TEA as the cocatalyst, 12 ± 2 mg of the catalyst (0.025 mmol of [Ti]) was injected into the reactor.
W. Li et al. / Journal of Catalysis 360 (2018) 145–151
The polymerization was carried out under a continuous ethylene flow to meet 3 bars at a stirring rate of 450 rpm. The polymerization proceeded for 60 min. At the end of the polymerization, the autoclave was quickly vented. The synthesized product was precipitated and washed with acidified ethanol (5 wt% hydrochloric acid), and dried at 60 °C under vacuum for 12 h. 2.4. Density functional theory (DFT) simulation The structural relaxations for the systems 1 POSS + 1MgCl2 and 1 POSS + 1MgCl2 + 1TiCl4 were carried out by using DFT calculations. To find the energetically preferred configurations, various relative positions and orientations between the molecules are considered for each system. The calculations used 6-31G⁄ basis set [8] and B3LYP exchange-correlation functional [2]. All the DFT calculations were performed with GAUSSIAN09 package. 2.5. Characterization The particle size distribution (PSD) of POSS was determined by dynamic light scattering (Zetasizer 3000HSA, Malvern, UK) at room temperature. THF was used as the dispersion. The sample were sonicated for 30 min before the measurement. The average particle size of POSS was 9.5 nm (See the PSD curve of POSS in Fig. S3). The morphology of the supports and catalysts were observed using a transmission electron microscopy (TEM, Tecnai F20, USA) and scanning electron microscopy (SEM, Hitachi S-4700, Japan). The catalyst was first mixed with n-hexane and then was put into ultrasound for 15 min. One droplet of the mixture are added into the copper screen and was dried in the glove box. The dried copper screen was used for TEM observing. The Energy dispersive X-ray detector (EDX) was used for analysing the elemental composition. The red points in the Fig. 1c are the detected points. The samples were sputter-coated by Pt before the measurement of SEM. The
147
distribution of nanoaggregates was calculated by Image J software. The specific surface area of Macro-SiO2 was measured by an Automatic Mercury Porosimeter (PoreMaster-60, Quantachrome, USA) at a pressure range of 0.3–30,000 psi. The contact angle and surface tensions were 130° and 0.480 N/m, respectively. Powder X-ray diffraction (XRD) measurements were carried out on a Bruker GADDS diffract meter with the Cu Ka radiation at 40 kV and 40 mA (k = 0.154 nm, 5–60°). The step size was 0.02°. X-ray photoelectron spectroscopy (XPS) measurements were carried out on Escalab 250Xi using Al Ka radiation as the X-ray source (300 W, 1486.3 eV). The vacuum chamber was about 5 109 Torr. The binding energy was calibrated to C1s peak at 284.6 eV. 1H NMR spectra of the POSS, POSS/MgCl2 and POSS/MgCl2/TiCl4 were obtained on a Bruker AC-80 400 MHz spectrometer at ambient temperature. Deuterated chloroform (CDCl3) was used as the solvent for all the samples in the NMR measurement. Tetramethylsilane (TMS) was used as the internal standard. The samples for NMR measurements were set in the sealed NMR tube for isolating the water and oxygen. The titanium content of the prepared catalysts was measured by a ultraviolet–visible (UV–vis) measurement through the hydrogen peroxide colorimetric method. The supported catalysts were dissolved in 5 ml of H2SO4 solution (1 mol l1), and 2 ml of H2O2, and then the solution was diluted with an H2SO4 solution to 25 ml. The UV measurement was performed by a TU-1901 spectrophotometer (PERSEA Corp, China). The intensity of the peak at 410 nm was used to quantify the titanium content. The weight-average molar mass (Mw) and the molecular weight distribution (MWD) were measured by a gel permeation chromatography (GPC) at 150 °C with a PL-GPC-220 instrument (Polymer Laboratories, Shropshire, U.K.). The molecular weight of standard polystyrene ranges from 1000 to 14,000,000 g/mol. 1,2,4-trichlorobenzene was used as the solvent. X-ray diffraction (XRD) measurements of nascent PE were carried out on a Bruker GADDS diffract meter with an area detector operating under
Fig. 1. (a) SEM morphology, (b) the particle size distribution of the nanoaggregates, (c) TEM morphology and (d) specific surface area of the POSS modified ZN catalysts. The catalysts are labelled as Cat-POSS-X, where X (=0, 10, 20 and 50) represents the nominal loading of POSS. (a) and (c) are from Cat-POSS-50, serving as examples.
148
W. Li et al. / Journal of Catalysis 360 (2018) 145–151
40 kV and 40 mA, using Cu Ka radiation (k = 0.154 nm) and a step of 0.02° min1. The XRD curves were decomposed into three components using Origin 8.5 in order to calculate the linear crystallinity of the nascent polymers. Lorentz function was used for fitting the curves [3]. Differential scanning calorimeter (DSC) measurements were performed with a DSC-7 instrument (PerkinElmer Corp., USA) to measure the melting point and crystallinity of products. Samples (ca. 5 mg) were first heated from 50 to 160 °C at 10 °C min1 and then held for 3 min to release thermal history. Subsequently, they were cooled to 50 °C at 10 °C min1 and then held for 3 min. The second heating cycle was the same as the first cycle. Melting temperature was taken at the peak of the endotherm. The crystallinity was calculated by comparison with the heat of fusion for polyethylene with a perfectly crystalline, 289 J/g [2a,3]. Rheological studies were performed on a straincontrolled rheometer (HAAKE III instrument) for analysing the entanglements of nascent polymers. A disk of 20 mm diameter was compressed under 20 tons at 120 °C for 30 min and used in all rheological studies [3]. The disk between the parallel plates of the rheometer was heated to 160 °C under a nitrogen environment to prevent thermo-oxidative degradation. After thermal stabilization at 160 °C (5 min), the rheology experiments were started. The dynamic amplitude sweep test was performed at a fixed frequency of 1 Hz to determine the linear viscoelastic regime. The dynamic time sweep test was performed to follow the entanglement formation at a fixed frequency of 1 rad/s and strain in the linear viscoelastic regime.
3. Results and discussion Typical surface morphologies of catalysts are shown in the Fig. 1 (See detailed SEM morphologies in the Fig. S4). The incorporation of POSS can form numerous nanoaggregates with a mean size of 48.8 nm, uniformly distributing on the surface of Macro-SiO2 supports. This indicates the POSS can adsorb effectively on the silica surface (See Fig. 1a–c). A low POSS-loading (i.e. 10 wt%) can reduce the mean size of nanoaggregates, which increases the tortuosity of the pores and leads to a higher specific surface area (SSA, See Fig. 1d) [3,9]. Further increasing the POSS loading (i.e. larger than 20 wt%) changes little on the size of the nanoaggregates but accumulated in the pores (See Fig. 1c). This accumulation leads to a linear decrement of the SSA [9]. The chemical composition of the nanoaggregates is further characterized by the TEM-EDX, where the concentration of Ti, Mg and Si atoms are 0.95, 7.1 and 25.0 wt%, respectively (see the red points in Fig. 1c). This indicates that the formed nanoaggregates may contain POSS, MgCl2, and can also immobilize the Ti complex. Fig. 2a, b shows the XRD curves of POSS modified Macro-SiO2 supports and their immobilized ZN catalysts. The XRD curves of pure POSS and a-MgCl2 are shown in the Fig. S5. The peaks at 2h = 10.8° (0 0 1), 15.0° (0 0 3), 22.5° (0 0 2) and 35° (0 0 4) represent the patterns of MgCl2/1,4-butanediol (BD) adducts [7,10]. The distinctly high intensity of these planes indicates that the structure of a-MgCl2, with a distorted cubic close packing of Cl atoms, is destroyed [10]. Moreover, the typical peaks of POSS (2h 10°) disappear, which indicates that the major aggregation of pure POSS are destroyed [3]. After immobilizing the TiCl4, the characteristic peaks for MgCl2/BD adducts disappear, while the peaks around 15°, 29–32° and 50.5° emerge (corresponding to the resonance of d-MgCl2). This d-MgCl2 prefers to immobilize the active titanium, which will present a high activity on the ethylene polymerization [10]. The peaks around 40.8° are shown both in the POSS modified supports and catalysts. The intensities of these peaks are dependent on the POSS loading in the supports, which gives another evidence of the formation of nanoaggregates [3,10]. More
importantly, the high intensities of these peaks are maintained for all the Cat-POSS-X systems, suggesting that the incorporation of TEA and TiCl4 cannot destroy the structure of nanoaggregates [10,11]. Surface element analysis of Si atoms of the support is shown in Fig. 2c, d. The Si 2p curves can be separated into two peaks at 101.3 and 103.1 eV (see Fig. 2c), which correspond to the surface Si atoms and bulk Si atoms, respectively [3]. interestingly, (see Fig. 2d), the concentration of surficial Si reduces linearly with the increment of POSS loading. This may be attributed to the reason that the MgCl2 molecules migrate to the surfaces of the nanoaggregates and cover on them, thus decreasing the amount of surficial Si atoms. To explore the interactions between POSS, MgCl2 and TiCl4 during the catalyst immobilization, model systems with equimolar quantity of these molecules are favorable. To this end, we performed both NMR measurements for POSS/MgCl2/TiCl4 mixtures with equal mole ratio, and density functional theory (DFT) calculations [8] for systems containing one POSS and one MgCl2 molecule (1 POSS + 1 MgCl2), as well as those containing one additional TiCl4 molecule (1 POSS + 1MgCl2 + 1 TiCl4) (See Fig. 3). The simulated results show that the O atoms of SiAOASi and the hydroxyl of POSS can be coordinated with the Mg atoms, forming conformation A, B or C (See Fig. 3a). The conformation C (i.e. 1 MgCl2 coordinated with the double hydroxyl of POSS) presents the lowest total energy (TE) and highest binding energy (BE), representing the most thermally stable conformation [12]. Conformation B, where one Mg atom binds to one hydroxyl, presents a BE of 34.1 kcal/mol. The coordination between the Mg atoms and the O atoms of SiAOASi bonds (conformation A) also shows a considerable BE (i.e. 15.9 kcal/mol). Considering the excess amount of the MgCl2 in the Cat-POSS-X (i.e. the mole ratio of [POSS]/[MgCl2] is 0.01–0.05), the coordination of Mg atoms with both O atoms in the hydroxyl (A, B) and those in the SiAOASi (conformation C) may exist [12]. The coordination between Mg and O is further confirmed by the NMR measurements. The chemical shift at 3.9 ppm (halo peak) reflects the resonance of hydroxyl groups of POSS (See Fig. 3b, POSS) [3b]. This resonance becomes narrow and shifts to 4.1 ppm with the incorporation of MgCl2 (See Fig. 3b, POSS/MgCl2 is synthesized by the nominally equimolar quantity), since the coordination of ClAMgAOH enhance the Lewis acid of the proton in the hydroxyl groups [10]. The CAH in the i-butyl of POSS, with the chemical shift of 1.8 ppm (See Fig. 3c), is far away from the O atom of hydroxyl, serving as an internal standard. The mole ratio of [H]C–H/[OH] presents independence on the incorporation of MgCl2, indicating that no H is detached after the incorporation of MgCl2. The multiple peaks around 0.6 ppm correspond to the resonance of CH2 which is bonded with Si atoms (See Fig. 3c). Compared with the pure POSS, a new peak presents at a lower chemical shift and the mole ratio of [POSS]/[Mg] is measured to be 5 (measured by SEM-EDX [1c]). in the POSS/MgCl2. This demonstrates that only partial POSSs are coordinated with the MgCl2 through MgAOASi, which increase the electron density of the CH2. The uncoordinated MgCl2 will give a chance to form the d-MgCl2 in the Cat-POSS-X systems. The reaction between the POSS, MgCl2 and TiCl4 are further investigated by experiments with the equimolar quantity of the reactants as well as DFT calculations. Importantly, the pure POSS and coordinated POSS/MgCl2 adducts is n-hexane soluble. After reacting with the TiCl4, the POSS/MgCl2 adducts becomes nhexane insoluble (forming POSS/MgCl2/TiCl4) thanks to the increased polarity, whereas the POSS, uncoordinated with the MgCl2, will form POSS/TiCl4 upon the reaction with hydroxyl (forming SiAOATi). The POSS/MgCl2/TiCl4 solid (i.e. Cat-1) can be separated, and the POSS/TiCl4 can be leaching-out owing to its excellent solubility in the n-hexane. Excitingly, the mole ratio of [Mg]/[POSS] (measured by SEM-EDX [1c]) is 7:1 in the Cat-1
W. Li et al. / Journal of Catalysis 360 (2018) 145–151
149
Fig. 2. (a) XRD curves of the POSS modified Macro-SiO2, (b) XRD curves of the POSS modified catalysts. (c), (d) XPS analysis of the catalysts. The supports and catalysts are labelled as Sup-POSS-X and Cat-POSS-X, respectively.
Fig. 3. (a) DFT calculations for the systems of POSS and MgCl2 with the equimolar quantity, (b), (c) NMR studies of POSS, POSS/MgCl2 and POSS/MgCl2-TiCl4 (Cat-1 and Cat-2). The initial mole ratio of TiCl4/POSS is 1.0 in the Cat-1. The initial dosage of TiCl4 is excessive in the Cat-2.
150
W. Li et al. / Journal of Catalysis 360 (2018) 145–151
although their nominal loading is 1:1, demonstrating that the O atoms of SiAOASi must participate in the coordination with the Mg atoms. This gives the evidence that the MgCl2 can cover the surface of POSS, decreasing the surficial Si concentration in the XPS results (See Fig. 2d). Here, we must mention that the composition of the nanoaggregates in the Cat-POSS-X measured by TEM-EDX. Supposing all the Si atoms in the red points (Fig. 1c) is attributable to the POSS, with which give the highest mole ratio of [POSS]/[Mg] equal to 2.7. This value indicates the nanoaggregates of Cat-POSS-X systems contain the coordinated POSS/MgCl2 and the uncoordinated POSS. Based on the stoichiometry, incorporating the TiCl4 with the equimolar quantity of POSS (namely half quantity of the hydroxyl) results in a disappearance of the –OH (3.9 ppm) and a split of CAH resonance in the Cat-1 (see Fig. 3d). Excessively incorporated TiCl4 takes no variance on the NMR spectrum (See Cat-2 in Fig. 3b, d). These demonstrate that the TiCl4 is able to react with the hydroxyl of the POSS/MgCl2 (forming the reacted conformation, see Fig. S6). The coordinated MgCl2 can also coordinate/immobilize the TiCl4 (forming coordinated conformation, see Fig. S6) because the molar ratio of [Ti]/[POSS] is measured to be 4 in the Cat-1. This fact is further evidenced by the DFT calculations where the reacted and coordinated conformation show similar BE and TE (See Fig. S6). Therefore, the loading amount of TiCl4 may be higher in the CatPOSS-X compared with that of Cat-POSS-0. The catalytic activity of POSS/MgCl2/TiCl4 with the excessive dosage of TiCl4 (Cat-2) is studied since the amount of TiCl4 is excessive for synthesizing the Cat-POSS-X. The Cat-2 presents a low activity on the ethylene polymerization (i.e. 2.0 104 gPE mol1 Ti h1 bar1). The POSS/TiCl4 catalyst shows a much lower activity 1 (i.e. 2.2 102 gPE mol1 bar1). The Immobilization of TiCl4 on Ti h the POSS/MgCl2 modified Macro-SiO2 shows a considerable Ti loading (i.e. 1.0 wt%) but no catalytic activity. Thus, POSS/MgCl2 adducts are suitable to serve as horizontal spacers to isolate the active species loading on the d-MgCl2 [3]. Table 1 shows the immobilization results of TiCl4 on the MacroSiO2/POSS support. The Incorporation of POSS indeed shows a higher TiCl4 loading than the Cat-POSS-0 as we mentioned above.
The decrement of Ti loading with the increased POSS amount may be due to the reduced SSA. The BE of Mg 2p are decreased in the Cat-POSS-X compared with the case of Cat-POSS-0, suggesting that the electron density of Mg atoms is increased by the coordination between MgCl2 and POSS [3,11]. The increased BE of Ti2p1/2 suggests the presence of electron deficient Ti atoms owing to the electron donor effects of POSS [3]. The multiple chemical environments of TiCl4 broaden the Ti2p1/2 peak in the POSS modified catalysts. The results of ethylene polymerization are shown in Table 2. The Cat-POSS-0 exhibits a lower activity than the POSS modified 1 one. A super high activity (1.3 106 gPE mol1 bar1) is Ti h achieved with Cat-POSS-20, because the MgCl2/POSS nanoaggregates serve as horizontal spacers to enlarge the distance between the active sites and therefore inhibiting bimetallic deactivation [3]. The decreased activity of Cat-POSS-30 and Cat-POSS-50 can be explained by the MgCl2/POSS aggregators immobilizing more deactivated TiCl4, and also by the decreased SSA of the supports [3,10b]. The synthesized polyethylene presents a high molecular weight (i.e. larger than 151 104 g/mol), indicating the production of ultra-high molecular weight polyethylene (UHMWPE). Interestingly, the synthesized polymers by the POSS modified catalysts present high crystallinities (i.e. Xc,XRD 81.9% and Xc,DSC 62.5%), indicating a highly ordered structure of the nascent PE. The nascent UHMWPE shows a high melting point (larger than 144.0 °C), which is usually considered as a reflection of chain-extended crystal with the thickness more than 1 lm (See Fig. S7 for the DSC curves) [2a]. The high crystallinity and high melting point are lost during the second heating scan, indicating the weakly entangled state of the nascent polyethylene [2a,3]. This weakly entangled sate of nascent PE is further evidenced by the rheology. Using the nascent PE-POSS-50 one can achieve the most weakly entangled state where the modulus build-up curves shows the lowest starting modulus and takes the longest time to reach the thermo-dynamically stable state (see Fig. 4) [2a,3]. To the best of our knowledge, this is the first time that the weakly entangled UHMWPE can be synthesized by the ZN catalyst at 60 °C with a high activity, showing a great potential to realize the ‘drop-in’
Table 1 Immobilization of TiCl4 on the Macro-SiO2/POSS support.
a b c
Catalyst
[POSS]/[MgCl2]a
Ti loading (wt%)b
Mg 2p BE (eV)c
Cat-POSS-0 Cat-POSS-10 Cat-POSS-20 Cat-POSS-30 Cat-POSS-50
0 0.01 0.02 0.03 0.05
6.9 8.0 7.7 7.5 7.3
51.6 51.0 51.0 51.0 51.0
Ti 2p1/2c BE (eV)
FWHM (eV)
464.4 464.9 464.9 464.8 464.8
2.5 2.8 2.9 2.8 2.8
Initial mole ratio of POSS to MgCl2. Measured by ICP. Measured by XPS.
Table 2 Ethylene polymerization results of the catalyst.a
a b c d e
Catalyst
PE yield (g)
Activityb
Mwc
MWD
Xc,XRDd %
Xc1,DSCe %
Tm1 °C
Xc2,DSCe %
Tm2 °C
Cat-POSS-0 Cat-POSS-10 Cat-POSS-20 Cat-POSS-30 Cat-POSS-50
36.0 81.5 97.9 47.9 42.7
480 1168 1305 638 570
181.0 159.0 205.3 151.0 158.0
6.0 8.4 9.5 8.9 9.1
78.5 82.1 81.9 81.9 84.6
62.5 64.2 65.5 66.9 66.4
143.7 144.0 144.2 144.3 144.7
46.5 47.5 53.3 52.9 54.3
136.5 137.0 137.2 137.5 137.8
Polymerizations. catalysts: 0.025 mmol of [Ti], 60 min of reaction, 500 ml of toluene, 3 bar of C2H4, 60 °C and 100 M ratio of [Al]/[Ti]. 1 103 gPE mol1 bar1. Ti h 104 g/mol. Measured by XRD. Measured by DSC.
W. Li et al. / Journal of Catalysis 360 (2018) 145–151
151
Science Foundation of China No. 21376054, Prof. Yang thanks the support of Natural Science Foundation of China No. U1663222, the Project of Natural Science Foundation of Ningbo (2016A610048), The National Science Fund for Distinguished Young, 21525627, and sponsorship by the K. C. Wong Magna Fund in Ningbo University are gratefully acknowledged. C. D would like to thank the computer center of Technische Universität Ilmenau for the support of all computations. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at https://doi.org/10.1016/j.jcat.2018.01.024. References Fig. 4. Dynamic time sweep test (at 160 °C) for the nascent PE with storage modulus (normalized by maximum plateau modulus G0 max). The synthesized polymer is labelled as PE-POSS-X, where X (=0, 10, 20 and 50) represents the nominal loading of POSS/MgCl2.
techniques in the industry. The reduced entanglements can be attributed to that the steric hindrance, generated by the MgCl2/ POSS nanoaggregates, can enlarge the distance between the growing chains and decrease accordingly the probability of chain overlap [2,3]. Finally, when change the Macro-SiO2 to the commercial silica with confined pores (i.e. Grace-955 SiO2), the isolators (POSS/MgCl2 nanoaggregates) can still work in the separation of active sites, where the weakly entangled UHMWPE with even higher molecular weight (i.e. 2.9 106 g/mol) can be synthesized 1 with a super high activity (i.e. 4.5 106 g PE mol1 ) at 60 °C Ti h (See Fig. S8). 4. Conclusions In summary, we have shown that the POSSs can capture the MgCl2 molecules through the coordination, serving as electron donors with a cage structure. The POSS/MgCl2 adducts can be effectively adsorbed and uniformly distributed on the surface of MacroSiO2 forming nanoaggregates with a mean size of 48 nm. The formed POSS/MgCl2 nanoaggregates can immobilize the TiCl4 and present extremely low activity on ethylene polymerization. These nanoaggregates are thus served as horizontal separators for isolating the active TiCl4 and the growing polyethylene chains. The weakly entangled UHMWPE can be synthesized by the facile modification of ZN catalyst at 60 °C with a super high activity, demonstrating a great potential to realize the ‘drop-in’ techniques. The reported results provide a new strategy of controlling the distribution of active sites in the heterogeneous catalyst and opening the perspective of preparing highly active catalyst and highperformance polyolefin. Acknowledgements The research was supported by the Natural Science Foundation of China No. 21776141, Prof. Hou thanks the support of Natural
[1] (a) D.J. Read, D. Auhl, C. Das, J.D. Doelder, M. Kapnistos, I. Vittorias, T.C.B. Meleish, Science 333 (2011) 1871–1874; (b) M. Stürzel, S. Mihan, R. Mülhaupt, Chem. Rev. 116 (2016) 1398–1433; (c) D. Cicmil, J. Meeuwissen, A. Vantomme, J. Wang, I.K. van Ravenhorst, H.E. van der Bij, A.M. Murillo, B.M. Weckhuysen, Angew. Chem. Int. Ed. 54 (2015) 13073–13079. [2] (a) S. Rastogi S, D.R. Lippits, G.W.M. Peters, R. Graf, Y. Yao, H.W. Spiess, Nat. Mater. 4 (2005) 635–641; (b) A. Osichow A, C. Rabe, K. Vogtt, T. Narayanan, L. Harnau, M. Drechsler, M. Ballauff, S. Mecking, J. Am. Chem. Soc. 135 (2013) 11645–11650; (c) M. Santoro1, F.A. Gorelli, R. Bini, J. Haines, A. Lee, Nat. Commun. 4 (2013) 1557–1563; (d) L. Kolb, V. Monteil, R. Thomann, S. Mecking, Angew. Chem. Int. Ed. 44 (2005) 429–432; (e) K. Kageyama, J. Tamazawa, T. Aida, Science 285 (1999) 2113–2115. [3] (a) W. Li, H.Q. Yang, J.J. Zhang, J.S. Mu, D.R. Gong, X.D. Wang, Chem. Commun. 52 (2016) 11092–11095; (b) W. Li, C. Guan, J. Xu, J. Mu, D. Gong, Z. Chen, Polymer 55 (2014) 1792–1798. [4] (a) M. Klapper, D. Joe, S. Nietzel, J.K. Krumpfer, K. Müllen, Chem. Mater. 26 (2014) 802–819; (b) W. Kaminsky, Adv. Polym. Sci. 258 (2013) 1–363. [5] (a) G. Ungar, J. Stejny, A. Keller, I. Bidd, M.C. Whiting, Science 229 (1985) 386– 389; (b) N. Waheed, G.C. Rutledge, J. Polym. Sci. Polym. Phys. 43 (2005) 2468–2473. [6] (a) T.F.L. Mckenna, A.D. Martino, G. Weickert, J.B.P. Soares, Macrom. React. Eng. 4 (2010) 40–64; (b) M.P. McDaniel, ACS Catal. 1 (2011) 1394–1407. [7] B. Xue, L. Hui, H.Q. Yang, Y.L. Zhao, L.X. Hou, W. Li, Ind. Eng. Chem. Res. 56 (2017) 135–142. [8] V.A. Rassolov, M.A. Ratner, J.A. Pople, P.C. Redfern, L.A. Curtiss, J. Comput. Chem. 22 (2001) 976–984. [9] (a) M. Choi, F. Kleitz, D. Liu, H.Y. Lee, W. Ahn, R. Ryoo, J. Am. Chem. Soc. 127 (2005) 1924–1932; (b) H. Li, B. Xu, X. Liu, W. Gao, Y. Mu, Chem. Comm. 51 (2015) 16703–16706. [10] (a) P. Sozzani, S. Bracco, A. Comotti, R. Simonutti, I. Camurati, J. Am. Chem. Soc. 125 (2003) 12881–12893; (b) R. Credendino, D. Liguori, Z.Q. Fan, G. Morini, L. Cavallo, ACS Catal. 5 (2015) 5431; (c) T. Taniike, M. Ternao, J. Catal. 293 (2012) 39–50. [11] (a) M.M.C. Forte, F.V. Cunha, J.H.Z. dos Santos, J. Mol. Cat. A Chem. 175 (2001) 91–103; (b) A. Andoni, J. Chadwick, H. Niemantsverdriet, P. Thune, J. Catal. 257 (2008) 81–86; (c) T. Taniike, T. Funako, M. Terano, J. Catal. 311 (2014) 33–40. [12] (a) D. Hossain, C.U. Pittman, S. Saebo, J. Phys. Chem. C 111 (2007) 619–6206; (b) D. Hossain, C.U. Pittman, F. Hagelberg, S. Saebo, J. Phys. Chem. C 112 (2008) 16070–16077; (c) A.R. Bassindale, M. Pourny, P.G. Taylor, M.B. Hursthouse, M.E. Light, Angew. Chem. Int. Ed. 42 (2003) 3488–3490.