Fabrication of CNT/MgCl₂-Supported Ti-based Ziegler-Natta Catalysts for Trans-selective Polymerization of Isoprene

Hydroxylated carbon nanotubes (CNTOH) (outer diameter (OD) around 8 nm, inner diameter (ID): 2-5 nm, length: 10-30 μm) and large inner diameters carbon nanotubes (CNTL) were provided by Chengdu Organic Chemistry Co., Ltd., China. Anhydrous magnesium chloride (purity, 98%), titanium tetrachloride (purity, 99%), and isoprene (purity, 99%) were procured from Sigma-Aldrich, China. Triisobutylaluminum (TIBA) (purity, 98%) was obtained from Akzonobel, USA.

This section was designed to utilize ball milling to prepare the CNTOH/MgCl₂/TiCl₄ catalyst/CNTL/MgCl₂/TiCl₄ catalyst with well dispersed CNTOH/CNTL sheets. 30 g MgCl₂ and 10 g CNTL (CNTOH) were added into the ball-milling agate jar under nitrogen atmosphere. Keep ball milling at 500 rpm for 1 hour, the mixture turned out to be more uniform. Then, 1.18 mL TiCl₄ were directly added into the ball-milling agate jar under nitrogen atmosphere respectively. Hours later, the CNTL/MgCl₂/TiCl₄ (CNTOH/MgCl₂/TiCl₄) catalyst was obtained. Then, these two kinds of catalysts were applied in the isoprene polymerization.

The polymerization was implemented in a 250 mL glass reactor filled with nitrogen to remove the moisture oxygen. 180 mL isoprene and the co-catalyst TIBA was added to the reactor then the glass reactor was placed in an ice-water mixture for the pre-polymerization (in the absence of pre-polymerization experiments, this process was abolished). After half an hour the catalyst was poured into the reactor, then pre-polymerization in the ice-water mixture for 10 minutes. After reacting at 20°C for 10 hours the polymerization was terminated by adding 10% HCl-ethanol solution, and then dried under vacuum oven at 40°C until a constant weight was achieved. The polymers obtained by the catalyst of CNTOH/MgCl₂/TiCl₄ named TPI/CNTOH and obtained by the catalyst of CNTL/MgCl₂/TiCl₄ named TPICNTL, respectively.

Fourier transform infrared spectroscopy (FTIR) of polyisoprene composites were recorded using a VERTEX 70 infrared spectrometer in the range of 400-4000 cm⁻¹ . Morphologies of the catalyst and composites were examined using a scanning electron microscopy (SEM, JEM-2100F). X-ray diffraction (XRD) patterns were obtained using a Rigaku D/MAX-2500 X-ray diffractometer (Japan). Thermogravimetric analysis (TGA) was conducted using a Q800 thermal analyser (TA, America) under nitrogen atmosphere at a heating rate of 20°C/min. The weight loss was measured from 30 to 800°C.

The preparation of CNTL/MgCl₂ and CNTOH/MgCl₂ supported Ti-based Ziegler–Natta catalyst is illustrated in Scheme 1. In the process of preparation of the catalyst, the CNTL/MgCl₂ and CNTOH/MgCl₂-supported Ziegler-Natta catalysts were synthesized via high-energy ball milling. During this step, the surfaces of the CNTL and CNTOH were covered with MgCl₂-TiCl₄ catalyst, which prevented possible aggregation of CNT/CNT layers. The -OH group on the surfaces of CNT can interact with the surfaces of MgCl₂ in two distinct ways: by interaction of hydrogen atoms with surface chlorine atoms or via interaction of electron pairs of the oxygen with magnesium.²⁸ To be specific, there were three major factors preventing the aggregation of CNT/CNTOH. Firstly, MgCl₂ attached the surface of the CNTL and CNTOH so it can be prevented the Van der Waals force and π-π interaction. Besides, during the high-energy ball milling process, the planetary ball mill possessed greater impact, compression and shearing intensity.^29,30 The crashing effect became more intensive and the powders were mixed more uniform, which could turn MgCl₂ attached to the surface of CNT will have more active sites exposed to the surface, and this will be beneficial for the catalyst activity. Hence, here we have formulated CNT/MgCl₂ supported TiCl₄ based Ziegler-Natta catalyst using a solvent-free, ball milling based facile method. Subsequently, the resultant two kinds of catalysts were used during isoprene polymerization to prepare the TPI/CNTL and TPI/CNTOH nanocomposite.

1.1. Morphology Characterization of catalyst by SEM

Morphologies of the two catalysts have been compared using SEM analysis and shown in Figure 2. Figure 2(a) shows the dispersed state of CNTOH in CNTOH/MgCl₂/TiCl₄ catalyst; Figure 2(b) shows the dispersed state of CNTL in TiCl₄/CNT/MgCl₂ catalyst. As shown in these two figures, CNTOH and CNTL disentangled and dispersed uniformly in the catalysts. In the ball milling process lamella, MgCl₂ inserts between CNT or partially cover on the surface of CNT,^19,20 which weakens the π-π interaction between CNT and thus makes CNT disentangle and disperse uniformly in the catalyst.

Figure 1. Schematic representation for synthesis of CNTL/MgCl₂/TiCl₄ and CNTOH/MgCl₂/TiCl₄ catalyst.

Download Original Figure

Figure 2. SEM images of (a) CNTOH in CNTOH/MgCl₂/TiCl₄ catalyst (b) CNTL in CNTL/MgCl₂/TiCl₄ catalyst.

Download Original Figure

1.2. Influence of ball-milling time on catalytic efficiency

Figure 3 shows the catalytic efficiency of CNTL/MgCl₂/TiCl₄ and CNTOH/MgCl₂/TiCl₄ Ziegler-Natta catalysts at different milling time . After the milling time reaches 4 hours with 500 rpm, prolonging the milling time brings minor effects on the catalytic efficiency. Due to the fact that the faster revolving speed and stronger impacting, shearing, milling and compressing effects than traditional mill,³¹ planetary ball mill presents obvious advantages in breaking and homogenizing powder. After a specific period, the planetary ball mill achieved a kind of balance between smashing and dispersing systems. In consideration of the energy consumption and catalytic activity, the optimum ball-milling time is 4 hours. Hence, the present research sets 4 hours as the milling time for the preparation of catalysts. At the same time, the CNTOH/MgCl₂/TiCl₄ Ziegler-Natta catalysts shows higher activity than CNTL/MgCl₂/TiCl₄ Ziegler-Natta catalysts, maybe the -OH group on the surfaces of CNT can interact with the surfaces of MgCl₂, so the MgCl₂ have higher specific surface area which may cause more titanium tetrachloride active sites to be anchored to the catalyst surface and eventually lead to an increase in catalytic efficiency.

Figure 3. Rate of polymerization (Rp) using CNTL/MgCl₂/TiCl₄ and CNTOH/MgCl₂/TiCl₄ at different ball milling time (Rp: rate of polymerization, $\text{Rp=}\left(\frac{\text{weight of TPI (g)}}{\text{mmol of Ti}}\right)/\text{h}\text{.)}$

Download Original Figure

1.3. Structure Characterization of catalyst by WAXD

To measure the activity of catalyst and the dispersed state of CNTL and CNTOH, fabrication of more active catalyst has been predicted by comparative XRD analyses of pure MgCl₂ carrier, pure CNT carrier, MgCl₂/TiCl₄ catalyst, CNTL/MgCl₂/TiCl₄ and CNTOH/MgCl₂/TiCl₄ catalyst. From Figure 4, in the pure MgCl₂ carrier curve it can be seen that high-intensity diffraction peaks are observed when 2θ is 15, 30, and 35. It means MgCl₂ carrier is α crystal form. When 2θ is 15, 32 and 50.5, the profile shapes of both MgCl₂/TiCl₄ catalyst and CNTOH/MgCl₂/TiCl₄ catalyst are wider at the bottom and lower at the height than the diffraction peaks of α-MgCl₂. The WAXD spectrogram of MgCl₂/TiCl₄ catalyst is basically same with that of CNTOH/MgCl₂/TiCl₄. It indicates that the magnesium chloride in both the prepared CNTOH/MgCl₂/TiCl₄ catalyst and TiCl₄/MgCl₂ catalyst is transferred to δ crystal of activity from the primitive α crystal. The high and steep diffraction peaks in α-MgCl₂ become wide and weak in δ-MgCl₂ of disordered structure. The reason is that the increase of δ-MgCl₂ of disordered structure results in the increase of content of non-crystalline parts, which increases the specific surface area of the above-mentioned Zeiger-Natta carbon nanotube catalyst; besides, the non-crystalline part of the disordered-structure MgCl₂ is in favor of the permeation of TiCl₄ into the depth. Both the increase of specific surface area and easier permeation of active component effectively increase the Ti content of carriers.

Figure 4. XRD patterns of pure MgCl₂ carrier, pure CNTOH carrier, pure CNTL carrier, MgCl₂/TiCl₄ catalyst, CNTL/MgCl₂/TiCl₄ and CNTOH/MgCl₂/TiCl₄ catalyst.

Download Original Figure

In Figure 5(a), the polymerization reaction starts at 30°C without pre-polymerization. Both the two kinds of catalytic activity rapidly reaches the top and then gradually declines as the reaction goes on. The possible reason is that the power-like catalyst contact monomers at a large-scale at the beginning. However, as the reaction goes on, the catalyst pieces are covered by polymer and monomers become poorer fluidity. The decrease in the catalytic activity can be attributed to the polymer of early formation functions as a physical barrier, retarding the diffusion of the isoprene into the catalyst core. With the increasement of reaction time, the rate of polymerization continuously declines till the isoprene is completely depleted.

Figure 5. Rate of polymerization (Rp) using CNTL/MgCl₂/TiCl₄ and CNTOH/MgCl₂/TiCl₄ with reaction time (a) without pre-polymerization; (b) with pre-polymerization

Download Original Figure

In Figure 5(b), the polymerization reaction includes pre-polymerization in icy water for 0.5 h and then conducting polymerization reactions at 30°C. It can be seen that the catalytic activity rapidly reaches to the peak and then gradually declines during the pre-polymerization process, however, the decreasing trend was much lower than in the direct polymerization process. The possible reason is that during the beginning stage of polymerization reaction, triisobutyl aluminum deoxygenates TiCl₄ to TiCl₃ and Ti³⁺ triggers the isoprene polymerization. At the stage, the chain initiation rate and chain conversion rate are both high, so that the polymerization action presents relatively high activity. Without the pre-polymerization at low temperature, the initial polymerization rate of polyisoprene is higher than its crystallization rate (when the trans-1,4-structure is over 95%). The polymerization products possess poor formability and adhere together. While, pre-polymerization at low temperature can stabilize the active centers of catalysts, prolong catalysts' active time and effectively lower the viscidity of the system, which facilitates better morphology of polymers. During the process of polymerization, triisobutyl aluminum further deoxidizes Ti³⁺ to Ti²⁺, which has no activity to initiate the isoprene polymerization. The foregoing mentioned reduction reaction consumes the triisobutyl aluminum. Besides, such excessive reduction reaction reduces the concentration of triisobutyl aluminum, which eventually causes the concentration reduction of catalysts active centers initiating isoprene polymerization. The catalysts present declining polymerization activity. Due to the reducing concentration of triisobutyl aluminum, the probability of polymerization active centers transferring to triisobutyl aluminum also reduces, which further reduces the chain transfer rate. As the polymerization reaction goes on for 30 minutes, chain initiation rate and chain transfer rate reach to a balanced state and catalytic active centers tend to stablilize.

Thus, the relationship of the TiCl₄/isoprene monomer mole ratio (shorted as n(Ti)/n(IP)) with the monomer conversion rate was studied. As shown in Figure 6, when the n(Ti)/n(IP) is within the range of 1~5×10⁻⁵ , the monomer conversion rate increases with the increase of n(Ti)/n(IP). The reason is that when the concentration of the active component TiCl₄ is high, the probability that isoprene attacks TiCl₄ is also high. As a result, the monomer conversion rate in unit time is high. The catalytic efficiency of catalyst firstly increases and then declines with the increase of n(Ti)/n(IP) as well as reaches to the peak when the n(Ti)/n(IP) is 3.5×10⁻⁵ . The reason is that when the catalyst concentration is low and n(Al)/n(Ti) mole ratio reminds unchanged, the relative concentration of the cocatalyst triisobutyl aluminum in monomer is relatively low; besides, due to the active chemical characters, triisobutyl aluminum may react with some residual impurities like water and oxygen in the system, which consumes a certain amount of triisobutyl aluminum. In actual reactions, the n(Al)/n(Ti) declines and the cocatalyst triisobutyl aluminum fails to fully cause complex reaction with active centers, which lead to the declining of polymerization rate (Rp). However, after the n(Ti)/n(IP) reaches to a specific value (3.5×10⁻⁵), the numbers that monomer attacks the active center relatively declines as the n(Ti)/n(IP) increases. Then the catalyst's CE calculated with unit mass of catalyst will declines. The result is consistent with relationship of CE and conversion rate gained through theoretical calculation.

Figure 6. Rate of polymerization (Rp) and conversion using CNTL/MgCl₂/TiCl₄ and CNTOH/MgCl₂/TiCl₄ with different n(Ti)/n(IP)

Download Original Figure

As can be seen from Figure 7, the polymerization efficiency increases with the increase of Al-Ti molar ratio of Al-Ti, and the conversion rate of polymerization decreases with the increase of molar ratio of Al-Ti, which shows that the active component of the main catalyst, tetrachloro Titanium and monomer and cocatalyst triisobutyl aluminum there is the best ratio between. When the molar ratio of Al and Ti is too low, the concentration of triisobutylaluminium is too low, the triisobutylaluminum can not completely remove the impurities in the system, resulting in the loss of activity of some of the active components of the catalyst and the impurities; meanwhile, the co-catalyst The amount of triisobutylaluminium is too small, the main active component of titanium tetrachloride can not be fully coordinated with the reaction and can not form a sufficient number of active reaction centers; and when the the molar ratio of aluminum-titanium molar ratio is too high, the cocatalyst may be the main catalyst Over-reduction of titanium tetrachloride to inactive titanium dichloride will also lead to a decrease in the conversion of isoprene in the polymerized monomer. Meanwhile, an excessive amount of triisobutylaluminium will also be present in the subsequent reaction of the catalytic system Chain transfer, which is equivalent to reducing the number of active centers during the actual reaction, may also result in a decrease in the conversion of the isoprene monomer. In the catalyst system, when the ratio of aluminum to titanium is 80, the catalytic efficiency and the conversion rate achieve the optimal ratio.

Figure 7. Rate of polymerization (Rp) and conversion using CNTL/MgCl₂/TiCl₄ and CNTOH/MgCl₂/TiCl₄ with different n(Al)/n(Ti)

Download Original Figure

2.1. The structural characteristics of TPI/CNT nanocomposites by FTIR

As shown in Figure 8(a) and (b), when n(Al)/n(Ti) increases from 30 to 80, in fingerprint region, no obvious changes is ob-served to the peak shape of polyisoprene carbon nanotube composite. Thus, it can be confirmed that the changes in both n(Al)/n(Ti) and n(Ti)/n(IP) will not lead to structural changes of polyisoprene carbon nanotube composite under catalytic system. At the same time, the structure of these TPI/CNTOH and TPI/CNTL nanocomposites is close to that of natural eucommia rubber, and therefore all of them belong to trans-1,4-polyisoprene, as shown in Figure 8(c). According to the reference (Table 1),³² the signals at 910 cm⁻¹ due to the 1,2-units and the signals at 890 cm⁻¹ due to the 3,4-units, which is not appear in the FTIR spectra. So here there is only contains trans 3,4-units. Then we speculated on the possible mechanism of this reaction, as shown in Figure 9.

Figure 8. FTIR spectra of (a) TPI/CNTOH nanocomposites; (b) TPI/CNTL nanocomposites with different n(Al)/n(Ti); (c) TPI/CNTOH, TPI/CNTL nanocomposites and natural eucommia rubber.

Download Original Figure

Figure 9. Mechanism of isoprene TiCl₄/MgCl₂-Al(i-Bu)₃ system.

Download Original Figure

Table 1. Characteristic Bands of Characteristic Groups of NR and EU Gum.

Characteristic groups	Characteristic bands
	NR (Natural rubber)	EU gum (Eucommia rubber)
-C=C-	1645 cm−1	1645 cm−1
-CH3, -CH2-	1450 cm−1 , 1380 cm−1	1450 cm−1 , 1380 cm−1
-CH2-C(CH3)=CH-CH2-	835 cm−1	845 cm−1
Crystalline peaks		875 cm−1 , 798 cm−1 , 758 cm−1 , 595 cm−1 , 465 cm−1
Configuation	cis	trans

Download Excel Table

2.3. The thermal stability characteristics of TPI/CNT nanocomposites by TGA

As shown in Figure 10, Compared with pure polyisoprene, as the nanotube content increases, the highest temperature at thermal weight loss is increased from 380°C to 402°C. The obvious improvement in thermal stability is related to the network structure of the composite and seldom relates to the types of CNT.

Figure 10. SEM images of (a)TPI/CNTOH nanocomposites and (b) TPI/CNTL nanocomposites.

Download Original Figure

Figure 11. The TGA 1st Derivative curves of TPI/CNTOH nanocomposites and TPI/CNTL nanocomposites with Mw ≈ 420~460 k. Insert: magnified TGA curves from 25°C to 800°C.

Download Original Figure