Analysis of solution polybutadiene polymerizations performed with a neodymium catalyst.
Zirgler-Natta catalysts based on Nd are used for 1,3-butadiene polymerization in industrial scale. These catalysts allow for production of polymer materials with excellent properties with high yields, which are suitable for use as tires and elastic materials [1, 2]. However, the molecular weight distributions (MWDs) of the resulting polymers are not controlled well because of the presence of multiple active sites and the heterogeneity of the catalyst system [1-10]. In addition, cocatalysts are believed to exert a significant effect on the shape of the final MWD of obtained polymer materials [1-9].
In the past years, distinct catalyst systems (Ni [11, 12], Co [13, 14], Nd [1-10, 15-17], Ti [18, 19], and Sm [20-22]) have been developed for 1,3-diene polymerizations using methylaluminoxane (MAO) [2-4], borate , and alkylaluminum/halides [1-10] as cocatalysts. Titanocene/MAO [18-19] and samarocene/[AlR.sub.3]/borane [20-22] catalysts present single-active catalyst sites and are able to produce high cis-1,4 polybutadienes (PB) with the simultaneous control of the MWD and stereoregularity of the obtained polymer material; however, these catalysts require large amounts of MAO to scavenge impurities and activate the catalyst. Therefore, there are huge incentives for development of alternative catalyst systems that do not require the use of MAO as cocatalyst, given its high costs and difficulty in handling.
The cobalt- and nickel-based catalysts lead to production of high cis-1,4 PBs with long-chain branches because of reincorporation of double bonds [23-25]. As a consequence, the final polymer material usually presents higher molecular weights than the materials obtained with the other catalysts. Particularly, PBs produced with cobalt catalysts present higher gel contents compared with other PBs because of cross-linking and reincorporation of internal unsaturations. The gel formation leads to reactor fouling and reduces the mechanical and rheological properties of the final polymer [26, 27].
Nd-based catalysts allow for production of PBs with high cis-1,4 content (>98%), which leads to superior performance in strain-induced crystallization, allows for vulcanization without crack growth, and leads to improved fatigue resistance [24, 25, 28, 29]. Furthermore, the gel content of Nd-based PBs is close to zero, which also contributes to the improvement of the mechanical and rheological properties of the obtained materials. An additional advantage is related to the fact that various cocatalysts are able to activate neodymium catalysts. When MAO is used as cocatalyst, PBs with broad MWDs are produced because of the heterogeneous nature of the active sites [2, 3]. Allyl-neodymium catalysts are able to produce PBs with narrow MWDs (polydispersity [PDI] around 2.0), but the metal complex is difficult to prepare and to handle [30-32]. The activation of neodymium catalysts with a combination of cocatalysts, such as the mixture of alkylaluminum ([AlR.sub.3]) and alkylaluminum halides ([AlR.sub.2]Cl), constitutes an effective method to obtain high polymer yields and high cis content simultaneously [1-9, 15-17]. However, to obtain long-chain branched polymers, it is necessary to activate the complex with diethylaluminum chloride (DEAC) at more severe polymerization conditions (higher temperatures and higher cocatalyst concentrations - Al/Nd = 15) during multiple polymerization steps [23, 29].
Nd-based catalysts generally present good activity and produce PBs with high cis-1,4 content. However, the MWD cannot be controlled effectively, which is generally attributed to the heterogeneity of the catalyst system and to the existence of multiple active sites [1-10]. Thus, the catalyst and cocatalyst feed sequence, the catalyst aging temperature, and the polymerization conditions (including cocatalyst and monomer concentrations) can lead to formation of different catalysts with different active sites distribution [33, 10]. The numerical deconvolution of obtained MWDs also indicated the existence of multiple active sites in Nd[Cl.sub.3] heterogeneous catalyst systems [34, 35].
This work investigates the effect of the ratio between the DEAC and neodymium molar concentrations, polymerization temperatures, catalyst concentrations, and butadiene concentrations on the final MWD of obtained polymers and on the catalyst performance in butadiene polymerizations performed with a neodymium versatate catalyst. DEAC is used as cocatalyst to evaluate the formation of long-chain branches in the obtained PB samples. It is shown that the cocatalyst/catalyst ratio (DEAC/Nd) and the reaction temperature exert the most important effects on the catalyst performance and on the final polymer properties. Particularly, the analysis of reaction conditions on the formation of long-chain branches is presented here for the first time, based on triple detector size exclusion chromatography (Tri-SEC) analyses. It is shown that PBs produced with neodymium versatate catalyst are not necessarily linear and that polymerization variables can affect chain branching frequencies significantly. In addition, detailed deconvolution of the MWDs of PBs produced with neodymium versatate catalysts are reported here for the first time, indicating that three or more catalyst sites are required to explain the final MWD of the polymer samples.
Neodymium versatate was purchased from Rhodia Rare Earths, and gallium as 50% w/v solution in hexane. Hexane was provided by Petroflex Ind. & Com. S.A. as a mixture of hexane and cyclohexane (20/80 to 30/70 wt%). This solvent was purified by drying in alumina bed columns before use. The blend of butadiene and hexane was provided as 35% w/v solution without inhibitor. The solution was also purified with alumina columns and maintained at 5[degrees]C. Nitrogen was provided by White Martins and purified by flowing through molecular sieves before use. DEAC was provided by Akzo Nobel with purity of 99% w/v. Diisobutylaluminum-hydride (DIBAL-H) was provided by Akzo Nobel with purity of 97% w/v. Cocatalysts were used as received.
The catalyst was synthesized in a glass reactor under nitrogen atmosphere. Reagents were added in the following order: the hexane solution containing known amounts of DIBAL-H, the neodymium versatate catalyst precursor, and finally DEAC. After addition of the reagents, the mixture was mixed with magnetic stirrer for 1 hr and aged for 12 hr at 5[degrees]C.
Polymerization was performed in a pilot plant equipped with a 1-L PARR stainless steel reactor. Alumina beds were used to dry the butadiene/hexane blend and the hexane. The reagents were added into the reactor in the following order: half of the prescribed hexane amount, the hexane/butadiene blend, the catalyst, and the remaining amount of hexane, to remove catalyst residues from the manifold. Polymerization was performed for 1 hr, and the produced polymer was transferred into a 2-L glass reactor, where the hexane solution was added. The hexane solution contained 3,5-di-terc-butil-4-hydroxitoluene (50% w/v), trinitrophenyl-phosphine, and octadecyl 3,5-di-(tert)-butyl-4-hydroxyhydrocinnamate (Irganox) (10% w/v). The resulting polymer solution was coagulated and dried in an oven with air recirculation. All manipulation was performed under nitrogen atmosphere.
The molecular weight averages and the MWDs were measured by gel permeation chromatography (GPC) at 40[degrees]C in an instrument equipped with a refraction index detector and a series of four columns. Tetrahydrofuran was used as eluent at a flow rate of 1 mL/min. Polystyrene standards were used for GPC calibration. Sample solutions (1 mg/mL) were filtered through a 0.45-[micro]m micro-filter before injection. Chain branching frequencies were determined through Tri-SEC with the help of a Viscotek - 303-040 Triple Detector Array equipped with differential refractive index, four capillary differential viscometers, and a low-angle light scattering detector. Tetrahydrofuran was used as solvent at a flow rate of 1 mL/min. Sample solutions (1 mg/mL) were filtered through a 0.45-[micro]m microfilter before injection. The Zimm method was used to calculate the molar masses [M.sub.i] and the mean square radius of gyration for each eluted volume, as reported in the literature .
RESULTS AND DISCUSSIONS
Main Effect Analysis
Experiments were performed in accordance with the experimental design ([2.sup.4-1] factorial) presented in Table 1. Replicates were performed at the central point to evaluate the experimental reproducibility. The following reaction variables were evaluated: DEAC/Nd ratio, reaction temperature, catalyst concentration, and 1,3-butadiene concentration. Variable ranges were selected in accordance with previous experience and published material. Particularly, as described in the previous section, it is known that the analyzed variables can indeed affect the performance of the catalyst system and the final properties of produced PBs. It is important to say that the catalyst characteristics depend on many other variables, such as the feeding sequence, aging temperature, and synthesis temperature, in addition to the polymerizations conditions [1-10, 33]. Despite that, the analysis of these variables is not performed here because it is intended to focus on the effects of polymerization variables, as the catalyst synthesis had already been optimized previously in respect to distinct objectives, such as the polymer productivity and catalyst stability, among others . For this reason, catalyst preparation conditions are kept constant throughout this article.
TABLE 1. Experimental results. Run DEAC/Nd Temperature [Nd] [BD] Activity codes ([degrees]C) (mmol/L) (g/100 mL) (kgPB/mmolNd/min) R1 10 70 0.2 7.5 55 R2 10 70 0.2 7.5 67 R3 10 70 0.2 7.5 55 LB2 5 60 0.3 10.0 41 LB3 5 80 0.1 10.0 102 LB4 5 80 0.3 5.0 23 LB5 5 60 0.1 5.0 Traces LB6 15 80 0.3 10.0 34 LB7 15 60 0.1 10.0 75 LB8 15 60 0.3 5.0 26 LB9 15 80 0.3 5.0 35 Run [M.sub.w] [M.sub.n] PDI ([M.sub.w]/[M.sub.n]) codes ([10.sup.3] Da) ([10.sup.3] Da) R1 204 76 2.7 R2 204 75 2.8 R3 205 76 2.7 LB2 656 202 3.2 LB3 552 172 3.2 LB4 319 144 2.2 LB5 n.a. n.a. n.a. LB6 167 55 3.1 LB7 671 171 3.9 LB8 73 17 4.4 LB9 197 60 3.3 The Cl/Nd molar ratio is equal to the DEAC/Nd molar ratio. n.a., not available; LB, long branching; R, center replicates.
Table 1 also shows the obtained experimental results for catalyst activity, average molecular weights, and PDT of the obtained PBs (Table 1). Experimental results were used to evaluate the effects of the analyzed polymerization variables with the help of the following linear empirical model (Eq. 1):
[y.sub.k] = [a.sub.0] + [n.summation of(i=1)][a.sub.i][x.sub.i] + [n.summation of(i=1)][n.summation of(j=1)][a.sub.ij][x.sub.i][x.sub.j] (1)
In Eq. 1, [a.sub.0] is the independent term, [x.sub.i] represents the independent polymerization variables (DEAC/Nd ratio, reaction temperature, catalyst concentration, and 1,3-butadiene concentration), [y.sub.k] represents model responses (catalyst activity, molecular weight averages, and PDI), and [a.sub.i] and [a.sub.ij] are model parameters (variable effects). The parameters (main and interaction effects) of the linear model were estimated using the experimental data presented in Table 1.
Figures 1-3 show the predicted and the experimental values for catalyst activity, weight average molecular weight ([M.sub.w]), and PDI, when Eq. 1 is used for main effect analyses (Figs. 1-3). Experiments performed at the central point indicate fair reproducibility (one must observe that variation of catalyst activities and molecular weight averages are negligible compared with the full range of variation of response variables). Figs. 1-3 show that there is a good agreement between the linear empirical model and the experimental values, which makes main effect analysis possible and meaningful (although an outlier is seemingly present in Fig. 1; Figs. 1-3). Obtained numerical results are presented in Table 2. In Table 2, one can find the linear coefficients ([a.sub.i]) of Eq. 1 for each analyzed response (Table 2). The significant effects (as computed with 95% of confidence) are presented in bold. The interaction effects were not found to be significant and are not shown in Table 2 (interaction effects may be assumed to be equal to zero). In Table 3, one can find the predicted model values, used to build Figs. 1-3.
[FIGURE 1 OMITTED]
TABLE 2. Main effect analysis of polymerization parameters. Yield (g) [M.sub.w] (Da) PDI ([M.sub.w]/[M.sub.n]) Al/Nd 1.19 -58042 0.51 Temp 7.19 -25865 -0.59 [Nd] -23.77 -5527 -0.16 [BD] 5.96 29331 0.05 Significant effects are presented in bold - 95%.
It can be noted in Figs. 1-3 and Table 2 that the polymer yield and the average molecular weights are sensitive to variation of polymerization parameters (Figs. 1-3; Table 2). All response variables, including polydispersities, change considerably in the investigated experimental range. Much attention has been dedicated to polydispersities because of the likely existence of multiple active sites and of the difficulty to reproduce the catalyst synthesis. Polydispersities vary between 2 and 4, indicating that narrow and broad MWDs can be produced when the polymerization conditions are changed. Similarly, catalyst activities vary between 0 and 100 kgPB/mmolNd/min and weight average molecular weights vary between 70 and 600 kDa, indicating that response variables are affected vary strongly by the varying polymerization conditions.
It is shown in Table 2 and Fig. 1 that the polymerization yield is affected mainly by the neodymium concentration, although the increase of the catalyst concentration leads to a decrease of the polymer yield. This unusual result can be explained by the formation of bimolecular species between propagating chains and the neodymium complex, which causes the decrease of the number of active sites and consequently of the polymer yield. This effect had been observed previously for other similar catalysts [13, 37, 38]. In addition, it is believed that the neodymium catalyst can form oligomers (Fig. 4), so that only terminal neodymium complexes would be able to produce PB [13, 37, 38] (Fig. 4). The synthesis of monomeric neodymium complexes was performed by Kwag , who showed that the catalyst activity increases with the neodymium content when only monomeric catalyst species are present in the reaction medium. From a practical point of view, it may be said that small catalyst concentrations should be used in the industrial environment for attainment of proper (and economical) catalyst activities.
[FIGURE 4 OMITTED]
Table 2 and Fig. 2 show that the weight average molecular weight is influenced by the neodymium concentration, the A1(DEAC)/Nd molar ratio, and the butadiene concentration. The increase of the DEAC concentration leads to a decrease in the molecular weight, which may be explained by chain transfer to DEAC . Another possible explanation is related to sensitivity of the distinct active sites to the increase of DEAC concentrations. In this case, it may be wondered that active sites that produce polymers of high molecular weights may be more sensitive to deactivation by DEAC, although this assumption is difficult to prove at the current time. Table 2 also shows that the polymerization temperature also exerts an important effect on the molecular weight averages. The increase of reaction temperature leads to a decrease in the weight average molecular weight. This can be explained by occurrence of spontaneous chain transfer, which is influenced by the polymerization temperature. The existence of spontaneous chain transfer reactions has already been observed in the literature for similar catalysts systems [2-4, 39]. In addition, the polymerization temperature can also influence the nature of the active sites in the reaction. As described later, the low-molecular weight polymers can be produced by active sites that are less sensitive to polymerization temperature . The increase of the butadiene concentration causes the increase of the molecular weight averages, as observed in Table 2. The higher butadiene concentrations lead to higher propagation rates and consequently to higher molecular weights.
[FIGURE 2 OMITTED]
The PDI of polymer samples is influenced mostly by the DEAC concentration and the polymerization temperature, as shown in Table 2 and Fig. 3. The influence of DEAC concentration can probably be explained by the formation of different active sites, which present distinct sensitivity to modifications of the DEAC concentration. The production of PB fractions with different molecular weights at distinct active sites can cause broadening of the MWD. It must be emphasized that the increase of chain transfer concentrations usually causes the decrease of PDI, as the chain growth becomes limited by a single reaction factor. As the opposite effect is observed here, the effect of multiple active sites on the shape of the MWD becomes more plausible. In addition, DEAC can affect the kinetic equilibrium between two or more active sites, as discussed in the literature for similar catalysts . Reaction temperature exerts the inverse effect on polydispersity, as shown in Table 2. The increase of the polymerization temperature leads to a decrease of the PDI. This can indicate that kinetic constants for chain growth and chain transfer may present distinct activation energies at the distinct active sites.
[FIGURE 3 OMITTED]
Analysis of MWDs
Figure 5 shows that the MWDs of the polymer samples can be broad (Fig. 5a) and present bimodal behavior (Fig. 5b). This gives additional support to the idea that the distinct kinetic behavior of the many existing catalyst sites controls the kinetics of the polymerization and the final characteristics of the polymer material. Significantly different MWDs can be obtained by varying the polymerization conditions. It can be observed, for instance, that the MWDs are broader at lower temperatures (LB6, LB9, R1, and R2).
[FIGURE 5 OMITTED]
Deconvolution of the MWDs of the PB samples was performed to investigate the relationship among polymerization conditions and the formation of the different active sites. Deconvolution was performed as described in the literature [40, 41], by assuming that each individual catalyst site produces polymer chains that follow the Flory distribution, as
[P.sub.i] = [N.summation over (j=1)] [[alpha].sub.j][(1 - [q.sub.j]).sup.2][iq.sub.j.sup.(i-1)] (2)
where [P.sub.i] represents the concentration of polymer chains of length i, [[alpha].sub.j], is the molar fraction of active catalyst site j, and [q.sub.j] is the propagation probability of active catalyst site j. When the polymer is essentially linear and average chain size is controlled mainly by chain transfer reactions, Eq. 2 describes closely the MWD of the polymer material produced at distinct catalyst sites. It is important to emphasize that previous deconvolution studies [34, 35] were performed for [NdCl.sub.3] catalysts, showing that multiple active sites were required to explain the MWD of the final polymer samples.
Figure 6 shows the experimental and deconvoluted distributions for some PB samples (LB4, LB3, and LB7). Figure 6 also shows the MWD of the polymer produced by each deconvoluted active site. It can be observed that the MWD deconvolution requires the definition of at least three distinct active sites, which is an indirect confirmation of the existence of many distinct catalyst sites in the reaction environment. As macromonomer reincorporation may occur, the polymer fraction of high molecular weight can be in principle produced by this mechanism , although this cannot be explained by observed branching frequencies (which are low and discussed in the following section) and by the dynamic evolution of molecular weight averages (which decrease with time). Therefore, the distinct nature of the different catalyst sites is the probable explanation for the distinct polymer fractions obtained through deconvolution. Deconvolution results show that the fractions of high molecular weights tend to decrease with time, which indicate that the sites that produce fractions of high molecular weights lose activity along the time. (Deconvoluted curves are usually different from experimental MWDs at low chain sizes because of the many sample preparation procedures, which tend to wash out the shorter polymer chains.)
[FIGURE 6 OMITTED]
Analysis of Branching Frequencies
As polymer chains present unsaturated double bonds, macromonomer reincorporation and long-chain branching may occur, following different reaction steps. Vinyl-terminated macromonomer reincorporation is common in olefin polymerizations performed with metallocene catalysts. However, the existence of internal unsaturations in polydienes allows for reincorporation of such internal unsaturations, leading to formation of distinct branching patterns. In addition, reincorporation of internal unsaturations of growing polymer chains is also possible, leading to formation of cyclic structures. These different mechanisms of macromonomer reincorporation are shown in Fig. 7.
[FIGURE 7 OMITTED]
The effects of polymerization conditions on long-chain branching frequencies are shown in Table 4 (experimental values are averages obtained for triplicates, and reproducibility may be regarded as very good) (Table 4). The neodymium catalyst produces predominantly linear PB. However, although long-chain branching frequencies are low in all cases, branches can be observed in almost all PB samples obtained in this work. It is important to notice that DEAC was used as cocatalyst and that some polymerizations were performed at relatively high temperatures. As mentioned by Nakajima and Yamaguchi , these conditions favor the production of branched PBs. However, available data indicate for the first time that all reaction variables can affect chain branching, which probably means that chain branching is not controlled by a single experimental factor. This clearly indicates that comprehensive reaction models are required for quantitative interpretation of chain branching in these systems.
TABLE 4. The influence of polymerization parameters on long-chain branching frequencies. Runs (PLB) DEAC/Nd Temperature [Nd] [BD] [B.sub.n] ([degrees]C) (mmol/L) (% w/v) (ramif/1000C) 2 5 60 0.3 10.0 0.00 (linear) 3 5 80 0.1 10.0 1.75 3 5 80 0.1 10.0 2.01 4 5 HO 0.3 5.0 3.73 4 5 80 0.3 5.0 3.82 6 15 80 0.3 10.0 0.22 6 15 80 0.3 10.0 0.20 7 15 60 0.1 10.0 2.95 7 15 60 0.1 10.0 3.10 The Cl/Nd molar ratio is equal to the DEAC/Nd molar ratio.
Chain branching probably becomes more important in the end of the batch, when the butadiene concentration becomes much smaller and macromonomer reincorporation is favored because of the higher PB concentrations. This observation can be supported by the comparative analysis of results obtained for runs LB3 and LB4, as it is possible to observe that the increase of polymer concentration in respect to monomer (the increase of catalyst concentration and reduction of monomer concentration causes the relative increase of polymer concentration) leads to increase of the chain branching frequency. This might already be expected, as it may be said that monomer insertion and double bond reincorporation are competitive kinetic steps. In addition, chain branching is also affected by the MWD (and therefore by temperature and concentrations of butadiene and DEAC), as short chains can probably be incorporated at higher rates, given the higher mobility and availability of the internal unsaturations of these chains. This may explain why chain branching frequencies are so sensitive to variations of the DEAC concentration and reactor temperature, as shown in Table 4. However, it is important to note in run LB7 that significant chain branching frequencies can be obtained even at mild reaction conditions, when the DEAC concentration is sufficiently high.
The intrinsic viscosity, [[eta]], and the mean square radius of gyration, [[left angle bracket][rg.sup.2][right angle bracket].sup.1/2], of some PB samples are presented in Fig. 8. As one might already expect, branched samples present lower [[left angle bracket][rg.sup.2][right angle bracket].sup.1/2] and [[eta]] values than those obtained for the linear sample. It is interesting to notice that the observed plot behavior changes significantly in the range from 5 X [10.sup.4] to 10 X [10.sup.4] Da, as this is the range where most polymer chains are expected to present at least one branch (according to data presented in Table 3). This can be observed more clearly when contraction factors g are computed as the ratio between the intrinsic viscosity of the branched samples and the intrinsic viscosity of the linear sample.
[FIGURE 8 OMITTED]
TABLE 3. Experimental vs. predicted data obtained with empirical model. Yield (g) DEAC/Nd Temperature [Nd] [BD] (% w/v) Predicted Experimental ([degrees]C) (mmol/L) data 10 70 0.2 7.5 22 22 10 70 0.2 7.5 22 27 10 70 0.2 7.5 22 22 5 60 0.3 10.0 26 25 5 80 0.1 10.0 20 20 5 80 0.3 5.0 15 14 15 80 0.3 10.0 21 20 15 60 0.1 10.0 17 15 15 60 0.3 5.0 14 16 15 80 0.3 5.0 24 21 [M.sub.w] (Da) PDI ([M.sub.w]/[M.sub.n]) DEAC/Nd Predicted Experimental data Predicted Experimental data 10 293507 203966 2.8 2.7 10 293507 208474 2.8 2.8 10 293507 205000 2.8 2.7 5 683540 656000 3.2 3.2 5 579916 551533 2.9 3.2 5 194063 319044 2.3 2.2 15 152310 167130 3.1 3.1 15 503645 671196 4.2 3.9 15 61135 73122 4.2 4.4 15 192057 197062 3.3 3.3 The Cl/Nd molar ratio is equal to the DEAC/Nd molar ratio.
Previous works [1-10] showed that neodymium carboxylates produce predominantly linear PBs. For this reason, detailed analyses of branching frequencies of PBs obtained with neodymium catalyst have not been reported. Despite that, it has been shown that small variations of the chain branching frequencies can have a strong impact on the PB performance and processability [28, 29]. As shown in this work, branched PBs can be indeed produced with neodymium versatate/DEAC catalyst systems at relatively mild polymerization conditions. Furthermore, as already said, modification of the chain branching frequencies can be achieved through modification of the operation conditions when the DEAC concentration is sufficiently high. This may be important for development of actual reactor operation policies at plant site and production of PB grades with improved performances.
The effect of the polymerization conditions on the catalyst activity and polymer properties during butadiene polymerizations performed with a neodymium versatate catalyst was analyzed. It was observed that polymer yields and catalyst activities depend essentially on the neodymium catalyst concentration. In addition, results indicate that bimolecular catalyst interaction occurs at significant extents. The MWDs are influenced by the butadiene concentration, reaction temperatures, and DEAC concentration. In addition, deconvolution of MWDs indicates the existence of multiple catalyst sites. In contrast, PDI is influenced mainly by the A1/Nd ratio and polymerization temperature. Obtained results indicate that the concentration and kinetic rate constants of the distinct catalyst sites respond differently to modifications of the reaction conditions, as PDI decreases with temperature and increases with the DEAC concentration, probably suggesting the modification of the relative concentration of active sites. Finally, it was shown that branched PBs can be produced with neodymium versatate/DEAC/DIBA-H catalyst systems at relatively mild polymerization conditions and that modification of the chain branching frequencies can be achieved through modification of the operation conditions when the DEAC concentration is sufficiently high. This may be important for development of actual reactor operation policies at plant site and production of PB grades with improved performances.
(1.) N.M.T. Pires, Sintese de Polibutadieno com alto teor do isomero 1,4-cis utilizando catalisador Ziegler-Natta, PhD Thesis, Universidade Federal do Rio de Janeiro, Rio de Janeiro, Brasil (2004) (in Portuguese).
(2.) W. Dong and T. Masuda, Polymer, 44, 1561 (2003).
(3.) W. Dong and T. Masuda, J. Polym. Sci. Part A: Polym. Chem., 40, 1838 (2002).
(4.) W. Dong, K. Endo, and T. Masuda, Macromol. Chem. Phys., 204, 104 (2003).
(5.) N.M.T. Pires, F.M.B. Coutinho, and M.S. Costa, Eur. Polym. J., 2599 (2004).
Luis Carlos Ferreira Jr., (1) Luiz Claudio de Santa Maria, (2) Marcos A.S. Costa, (2) Neusa Maria Tochetto Pires, (3) Marcio Nele, (4) Jose Carlos Pinto (1)
(1) Programa de Engenharia Quimica/COPPE, Universidade Federal do Rio de Janeiro, Cidade Universitaria CP: 68502, Rio de Janeiro CEP: 21941-914 RJ Brasil
(2) Instituto de Quimica, Universidade do Estado do Rio de Janeiro, Rua Sao Francisco Xavier 524 Maracana, Rio de Janeiro, Brasil, CEP: 20550-900
(3) Petroflex Industria & Comercio - Rua Marumbi, 600, Campos Eliseos, Duque de Caxias, RJ, Brasil, CEP: 25221-000
(4) Escola de Quimica, Universidade Federal do Rio de Janeiro, Av. Brigadeiro Trompowski, s/n, Ilha do Fundao, Rio de Janeiro, Brasil, CEP: 21941-909
(6.) R.P. Quirk, A.M. Kellsa, K. Yunlub, and J.P. Cuifb, Polymer, 41, 5903 (2000).
(7.) L. Friebe, O. Nuyken, H. Windisch, and W. Obrecht, Macromol. Chem. Phys., 203, 1055 (2002).
(8.) A. Oehme, U. Gebauer, K. Gehrke, and M.D. Lechner, Angew Makromol Chem, 235, 121 (1996).
(9.) L. Friebe, H. Windisch, O. Nuyken, and W. Obrecht, J. Macromol. Sci., Part A: Pure Appl. Chem., 41,3, 245 (2004).
(10.) G. Ricci, S. Italia, F. Cabassi, and L. Porri, Polym. Commun., 28, 223 (1987).
(11.) Y. Jang, P. Kim, Y. Ho, and H.L. Jeong, J. Mol. Catal. A: Chem., 206, 29 (2003).
(12.) Y. Jang, D.Y. Choi, and S. Han, J. Polym. Sci. Part A: Polym. Chem., 42, 1164 (2004).
(13.) P. Cass, K. Pratt, B. Laslett, and E. Rizzardo, J. Appl. Polym. Sci., 39, 2256 (2001).
(14.) D.C.D. Nath, T. Shiono, and T. Ikeda, J. Polym. Sci. Part A: Polym. Chem., 40, 3086 (2002).
(15.) E. Kobayashi, N. Hayashi, S. Aoshima, and J. Furukawa, J. Polym. Sci. Part A: Polym. Chem., 36, 1707 (1998).
(16.) F. Bonnet, M. Visseaux, A. Pereira, and D.B. Baudry, Macromolecules, 38, 3162 (2005).
(17.) D.J. Wilson, J. Polym. Sci. Part A: Polym. Chem., 33, 2505 (1995).
(18.) N. Naga and Y. Imanishi, J. Polym. Sci. Part A: Polym. Chem., 41, 939 (2003).
(19.) A. Miyazawa, T. Kase, and K. Soga, Macromolecules, 33, 2796 (2000).
(20.) W.J. Evans, T.M. Champagne, D.G. Giarikos, and J.W. Ziller, Organometallics, 24, 570 (2005).
(21.) W.J. Evans, D.G. Giarikos, and N.T. Allen, Macromolecules, 36, 4256 (2003).
(22.) S. Kaita, Y. Takeguchi, Z. Hou, M. Nishiura, Y. Doi, and Y. Wakatsuki, Macromolecules, 36, 7923 (2003).
(23.) K.L. Makovetskii, V.A. Yakolev, T.G. Golenko, and G.N. Bondarenko, Polym. Sci. Ser. B, 48, 61 (2006).
(24.) N.M.T. Pires, F.M.C. Coutinho, M.A.S. Costa, L.C. Santa Maria, and I.L. Mello, Revista de Quimica Industrial, 719, 15 (2002).
(25.) C. Wang, Mater. Chem. Phys., 89, 116 (2005).
(26.) A. Osmam, F. Barsan, PCT WO 00/14130, March 16, (2000).
(27.) M.N. Dedecker, H.L. Wright, and J.M. Pyle, U.S. Patent, US 6.727.330, April, 27, (2004).
(28.) N.M.T. Pires, A.A. Ferreira, C.H. Lira, P.L.A. Coutinho, L.F. Nicolini, B.G. Soares, and F.M.B. Coutinho, J. Appl. Polym. Sci., 99, 88 (2006).
(29.) N. Nakajima and Y. Yamaguchi, J. Appl. Polym. Sci., 61, 1525 (1996).
(30.) S. Maiwald, M.G. Sommer, and R. Taube, Macromol. Chem. Phys., 203(7), 1029 (2002).
(31.) R. Taube, S. Maiwald, and J. Sieler, J. Organomet. Chem., 621, 327 (2001).
(32.) S. Maiwald, R. Taube, H. Helming, and H. Schumann, J. Organomet. Chem., 552, 195 (1998).
(33.) J. Wang, Y.X. Wu, X. Xu, H. Zhu, and G.Y. Wu, Polym. Int., 54, 1320 (2005).
(34.) N.N. Sigaeva, T.S. Usmanov, V.P. Budtov, S.I. Spivak, G.E. Zaikov, and Y.B. Monakov, J. Appl. Polym. Sci., 87, 358 (2003).
(35.) N.N. Sigaeva, T.S. Usmanov, V.P. Budtov, S.I. Spivak, G.E. Zaikov, and Y.B. Monakov, J. Appl. Polym. Sci., 89, 674 (2003).
(36.) D.B. Trowbridge, Characterization of Long-Chain Branching in Polybutadiene, PhD Thesis, Graduated Faculty of the University of Akron, USA (1996).
(37.) G. Kwag, Macromolecules, 35, 4875 (2002).
(38.) V.G. Koslov, N.N. Sigaeva, K.V. Nefedjev, I.G. Saveleva, N.G. Marina, and Y.B. Monakov, J. Polym. Sci. Part A: Polym. Chem., 32(7), 1237 (1994).
(39.) W.J. Wang, S. Kharchlenko, K. Migler, and S. Zhu, Polymer, 45, 6495 (2004).
(40.) V. Matos, A.G.M. Neto, and J.C. Pinto, J. Appl. Polym. Sci., 79(11), 2076 (2001).
(41.) M. Nele and J.C. Pinto, Macromol. Theory Simul., 11, 293 (2002).
Correspondence to: Jose Carlos Pinto; e-mail: email@example.com
Contract grant sponsors: Conselho Nacional de Desenvolvimento Cientifico e Tecnologico, Fundacao Carlos Chagas Filho de Apoio a Pesquisa do Estado do Rio de Janeiro, PETROFLEX Industria e Comercio.
Published online in Wiley Online Library (wileyonlinelibrary.com).
[C] 2010 Society of Plastics Engineers
|Printer friendly Cite/link Email Feedback|
|Author:||Ferreira, Luis Carlos, Jr.; Maria, Luiz Claudio de Santa; Costa, Marcos A.S.; Pires, Neusa Maria Toc|
|Publication:||Polymer Engineering and Science|
|Date:||Apr 1, 2011|
|Previous Article:||Shaping conjugated hollow fibers using a four-segmented arc spinneret.|
|Next Article:||Visualization analysis of side-edge flow phenomena in different thickness/width rectangular cavities using a rotary runner exchange system.|