Natta catalyst

Kinetics of short-duration ethylene polymerization with MgCl2-supported Ziegler–Natta catalyst: Two-stage initiation evidenced by changes in active center concentration Akbar Khan, Yintian Guo, Zhisheng Fu, Zhiqiang Fan MOE Key Laboratory of Macromolecular Synthesis and Functionalization, Department of Polymer Science and Engineering, Zhejiang University, Hangzhou 310027, China Correspondence to: Z. Fan (E - mail: [email protected])

The kinetics of ethylene polymerization with a TiCl4/MgCl2-type Ziegler–Natta catalyst was studied. Changes in polymerization activity and concentration of active centers ([C*]) in the first 5 min were determined. Initiation of the active centers was found to proceed in two stages. In the first stage, [C*]/[Ti] quickly rose to about 1% in less than 30 s and then remained stable in the subsequent 60 s. Then the [C*]/[Ti] value started to increase again, forming the second buildup stage. The polymerization activity was found to change roughly in parallel with the change in [C*]/[Ti]. Changes in the polymer/catalyst particle morphology and polymer molecular weight distribution with polymerization time were studied. A mechanistic model was proposed to explain the two-stage kinetics: initiation of active sites on the outer surface of catalyst particles takes place in the first stage, and initiation of active sites buried inside the particles takes place in the second stage. These buried sites are released when the catalyst particles are fragmented C 2017 Wiley Periodicals, Inc. J. Appl. Polym. Sci. 000: 000–000, 2017 by the expanding polymer phase. V

ABSTRACT:

KEYWORDS: catalysts; kinetics; polyolefins

Received 19 December 2016; accepted 13 March 2017 DOI: 10.1002/app.45187 INTRODUCTION

MgCl2-supported Ziegler–Natta (ZN) catalysts play a dominant role in polyolefin production. However, even after more than three decades of fundamental studies, there are still many unsolved problems related to the mechanism of the catalysis process. Among these problems, the formation or initiation of catalytic centers in the first few minutes of olefin polymerization is especially important, as it will profoundly influence the kinetics, particle morphology, and polymer chain structure of the whole polymerization process. Many researchers have reported the kinetics and particle morphology of ZN-catalyzed olefin polymerization in the initial stage.1–12 In a recent work, Busico et al.12 found that in propylene polymerization with a preactivated supported ZN catalyst, the initiation was completed within a few seconds. Skoumal et al.3 also observed a quick initiation in propylene polymerization. In propylene polymerization with a Mg(OEt)2-based ZN catalyst, the polymerization rate starts to speed up again after a very short stationary stage, which was correlated to catalyst fragmentation in the process.8 This means that the whole initiation process is accompanied by the exposure of hidden catalyst surfaces containing a large number of potential active sites. In ethylene polymerization for a longer time, there are usually sharp transitions of the rate curve

from buildup type to decay type in the first few minutes.13 Therefore, a time span of 0–10 min is necessary for completion of the initiation process. Because of the lack of convenient methods to trace changes of active center concentration ([C*]) in the first 10 min, knowledge of the kinetics and mechanism of the whole initiation process is rather limited. In recent years, we have developed a new method of active center counting based on selective labeling of the propagation chains by quenching the catalyst with 2-thiophenecarbonyl chloride (TPCC). This method enabled us to trace the changes of [C*] with various polymerization conditions.14–19 In the initial stage of propylene polymerization with a MgCl2-supported ZN catalyst, [C*] was found to increase in the first 3 min and then level out in the remaining 10 min, and meanwhile the polymerization rate kept decreasing in the first 3 min.14 These results mean that the microscopic scenario of the initiation process is rather complicated. In this work, the initiation process of ethylene polymerization with a TiCl4/MgCl2-type ZN catalyst was studied by tracing the change in [C*] with time in the first 5 min. For the first time it was found that initiation of the active centers proceeds in two stages, which proved a gradual release of active centers through catalyst fragmentation in the polymerization process.

C 2017 Wiley Periodicals, Inc. V

WWW.MATERIALSVIEWS.COM

45187 (1 of 6)

J. APPL. POLYM. SCI. 2017, DOI: 10.1002/APP.45187

ARTICLE

WILEYONLINELIBRARY.COM/APP

EXPERIMENTAL

Reagents A commercial MgCl2-supported ZN catalyst (Ti content 5 5.8 wt %, provided by SINOPEC, Beijing, China) was used for polymerization. The catalyst particles have sizes of 228 lm and are nearly spherical in shape. Triethylaluminum (TEA) was purchased from Albemarle (Charlotte, NC) and used as a 2 M solution in n-heptane. Ethylene and propylene (polymerization grade, supplied by Minxing Gas Co., Hangzhou, China) were purified by passing through molecular sieves and a manganese-based deoxygenation agent. The 2-thiophenecarbonyl chloride was purchased from Alfa Aesar Co. (Ward Hill, MA) and diluted to a 2 M solution in n-heptane before use. n-Heptane was dried over 4 A molecular sieves under dry nitrogen and refluxed over Na before use. Polymerization and Quenching Reaction All operations were carried out under dry nitrogen atmosphere using a standard Schlenk line or glove-box techniques. Polymerizations were performed in 300 mL or 150 mL glass reactors with magnetic stirrer and gas inlet. A designated amount of n-heptane was added to the reactor, followed by the catalyst. After stirring for 5 min, TEA (Al/Ti 5 20) was added and stirred for 10 min. After this, an ethylene flow of 1 atm pressure and 1.5 SLM (standard liter per minute) was continuously bubbled through the liquid phase to initiate the polymerization for a designated time tp, and TPCC (TPCC/Al 5 2) was then introduced to quench the polymerization for 5 min. After the quenching, acidified ethanol was added to decompose the catalyst and quencher, and the polymer was precipitated with an excess of ethanol. When the effects of prepolymerization were studied, the catalyst–cocatalyst precontact time was reduced to 9 min, and 1 atm propylene gas was bubbled through the liquid phase for 1 min as the prepolymerization step. After this step, the ethylene polymerization and quenching reactions were conducted in the same procedures. The polymer samples were purified by the procedure described in the literature.14

Analysis The sulfur content of the quenched polymer was measured with a YHTS-2000 fluorescence UV sulfur analyzer (Jiangyan Yinhe Instrument Co., Jiangyan, China, detection limit 5 0.05 ppm). Three parallel measurements were made for each sample (2–4 mg solid powder, weighed to 60.01 mg). The molecular weight and molecular weight distribution (MWD) of the polymer were measured by gel permeation chromatography (GPC) in a PL 220 GPC instrument (Polymer Laboratories, Shropshire, UK). The analysis was performed at 150 8C using 1,2,4-trichlorobenzene as the solvent at a flow rate of 1.0 mL/min. Universal calibration against polystyrene standards was adopted. Differential scanning calorimetry (DSC) analysis of the polymers was done with a TA Q200 DSC instrument (TA Instruments, New Castle, DE) under N2 atmosphere. About 2–3 mg of sample was sealed in aluminum sample pan, and the sample was heated from 40 to 180 8C at a heating rate of 10 8C/min and kept melting for 5 min, and then the sample was cooled down to 40 8C. Finally, the melting endotherm of the sample was recorded at a heating rate of 10 8C/min from 40 to 180 8C. Scanning electron microscope observations of the PE particles were made with a Hitachi-4800 SEM (Hitachi High-Technologies Corp., Tokyo, Japan). Micrographs was taken at an acceleration voltage of 3 kV. The samples were coated with a thin layer of gold before SEM observation. RESULTS AND DISCUSSION

Kinetics of Polymerization Ethylene slurry polymerization with a MgCl2-supported ZN catalyst was conducted for short durations from 30 to 300 s. Before introducing ethylene into the reactor, the catalyst was stirred with TEA for 10 min (precontact step) in order to activate the potential active sites beforehand. After this precontact step, the initiation reaction, namely insertion of the first monomer in the TiAC bond of the active sites can immediately start when the monomer is introduced. The changes of polymer yield and catalytic activity with time are shown in Table I. The

Table I. Changes of Polymer Yield, Activity, and Polymer Properties with Time Sample

tpra (s)

tpb (s)

PE/catc (g/g)

Activity (kg/g(Ti) h)

Mn (kg/mol)

PDId

P1

0

30

0.553

1.17

9.3

14.9

130

190

P2

0

60

0.722

0.80

10.4

13.5

127

161

P3

0

90

1.432

0.98

6.1

37.2

133

218

P4

0

120

4.455

2.34

8.1

28.7

133

207

P5

0

300

10.18

2.16

9.1

28.8

134

191

P6

60

30

0.614

1.30

8.3

12.0

127

131

P7

60

60

1.061

1.20

9.2

13.7

128

133

P8

60

90

1.509

1.07

5.8

23.8

128

136

P9

60

120

5.070

2.70

8.0

25.3

131

157

P10

60

300

11.72

2.45

7.5

28.8

133

183

Tme (8C)

DHme (J/g)

Conditions: n-heptane 5 250 mL (P1, P2, P3, P6, P7, P8) or 125 mL (P4, P5, P9, P10); [Ti] 5 2.5 mmol/L; TEA/Ti 5 20 (mol/mol); ethylene and propylene pressure 5 1 atm; TPCC/Al 5 2 (mol/mol); quenching time 5 5 min; temperature 5 50 8C. a Time of prepolymerization. b Time of ethylene polymerization. c Polymer yield. d Polydispersity index (5 Mw/Mn). e Melting temperature (Tm) and melting enthalpy (DHm) determined by DSC.


45187 (2 of 6)


ARTICLE


Table II. Changes in Number of Active Centers and Propagation Rate Constant with Time Sample

tp (s)

[C*]/[Ti] (%)

Rp (mol/L s)

kp (L/mol s)

P1

30

0.98

0.0014

760

P2

60

1.08

0.0010

464

P3

90

0.84

0.0012

745

P4

120

2.14

0.0028

692

P5

300

5.49

0.0026

250

P6

30

0.92

0.0016

898

P7

60

1.43

0.0014

534

P8

90

0.89

0.0013

768

P9

120

2.57

0.0032

667

P10

300

6.14

0.0029

253

Figure 1. Changes in active concentration and polymerization rate with time.

The polymerization conditions are the same as in Table I.

activity versus time profile can be divided into two stages. Before 90 s, the activity was low and stable in the period of 30– 90 s. After 120 s, the activity rapidly rose to a much higher level. Such peculiar kinetic behavior is similar to that reported by Terano et al.,8 though the second stage took much longer to appear in the present work. The effects of prepolymerization on the initiation kinetics have also been studied. Propylene was introduced for a short (1 min) prepolymerization before introducing ethylene into the reactor. The prepolymerization caused a slight increase in activity, but the two-stage polymerization kinetics was unchanged. Prepolymerization with propylene has been found to improve the polymer particles’ morphology and activity.20–22 The rate enhancement effect shown in Table I is rather weak compared to that in the literature, possibly because of the short prepolymerization time adopted in this work. It is worth mentioning that the melting temperature and melting enthalpy of the polymer samples changed very slightly during the polymerization process without prepolymerization. By quenching the polymerization with TPCC and measuring the sulfur content of the polymer samples, the number of active centers ([C*]/[Ti]) in each polymer sample was determined. According to the rate equation Rp 5 kp[C*][M] that has been well established for most catalyzed olefin polymerizations, the chain propagation rate constant kp was also calculated. The polymerization rate Rp has been determined from differentiation of the curve of polymer yield versus time. The results are shown in Table II and Figure 1. The value of [C*] also increased in two stages. In 0–30 s, [C*]/ [Ti] quickly rose to about 1%, and then it changed little from 30–90 s. This stage may be explained by the quick initiation of active sites located on the exposed surfaces of catalyst particles. Because these sites are readily accessible to the cocatalyst and monomer, their activation can be completed in the precontact step, and their initiation can be completed very quickly. The first steady stage in the period of 30–90 s can be reasonably ascribed to the first induction-steady stage of [C*]. Emergence of the second induction stage in 90–300 s strongly suggests that many more active sites were exposed and initiated after the first


stage. The polymerization rate also experienced a quick increase from 90 to 120 s. A reasonable explanation for this phenomenon is the exposure of more active sites by disintegration of the catalyst particles in this period. The increasing weight of polymer should be the main reason for catalyst disintegration after 2 min of polymerization. Morphology of the Polymer Particles To search for further evidence for the two-stage initiation process, the particle morphology of the polymer was observed by SEM, as shown in Figure 2. The whole particles had sizes ranging from 5 to 100 lm and were aggregates of many subparticles [see Figure 2(g)]. Subparticles of 1–2 lm can be seen in Figure 2(a,c,e), but their morphologies evidently changed with the polymerization time. At a tp of 60 s, the surfaces of the subparticles were rather smooth, and there were only a limited number of cracks and holes. At a tp of 120 s, the subparticles were cracked into many interconnected smaller particles, and their surfaces became rather coarse, meaning that they began to disintegrate under the expansion force of the growing polymer phase. At a tp of 300 s, the morphology of the subparticles was not markedly changed; only their cracks became deeper and broader. Therefore, the morphology change in the period 60– 120 s is stronger than that in the later stage. This means that emergence of the second buildup stage can be correlated with the change in particle morphology. Many researchers have studied the structure and morphogenesis of polyolefin catalyst particles. Three levels of structure were found in particles of MgCl2-supported ZN catalysts: primary particles or MgCl2 crystallites bearing supported Ti species, subparticles formed by aggregation of many primary particles, and the whole catalyst particle composed of multiple subparticles.23–25 The present work implies that there are Ti species on most primary particles, but only those on the external surfaces of the subparticles are accessible to the cocatalyst and monomer at the beginning of polymerization. They can be initiated at the beginning, but the rest of the Ti species are still buried inside the subparticles. Because the volume of the polymer phase is not large enough to fragment the subparticles before 90 s, a stable [C*] level can be observed in 30–90 s. When the polymer volume increases to a threshold value (yield/catalyst ratio > 1.5), the distance between the tightly packed primary particles is enlarged to

45187 (3 of 6)


ARTICLE


Figure 2. SEM pictures of polymer particles at different polymerization times: (a,b) 60 s; (c,d) 120 s; (e–g) 300 s (samples P2, P4, and P5 in Table I).

allow for penetration of TEA and ethylene, making activation and initiation of more active sites possible. The acceleration of polymerization by [C*] increase will further promote particle disintegration, forming a new buildup stage. A simplified model is shown in Figure 3 to depict this two-stage initiation process. The present results suggest that the catalyst contains a high proportion of buried active species, as the [C*]/[Ti] value at tp 5 300 s is about five times the value at 60 s. More buried sites could be released when the yield/catalyst ratio is raised further. Molecular Weight Distribution of the Polymer The change in polymer MWD with time can provide indirect evidence for the formation of more active centers in the second buildup stage. As shown in Figure 4 and Table I, the MWD was markedly broadened when the polymerization proceeded from 60 s to 90 s. Such MWD broadening could be caused by a markedly increased diffusion limitation in this period, as the yield/catalyst ratio increased a lot in this period. Liu et al.26 have found that diffusion limitation on the microscale can cause severe MWD broadening in olefin polymerization. It is interesting to see a slight decrease of polydispersity index at a tp of 120 and 300 s, when the MWD curves looked more like that at tp 5 30 s. This means that the diffusion limitation is eased in this period, because more


active centers are exposed in this period, which can reduce the average diffusion distance for a monomer molecule to reach a certain active center. On the other hand, the increasing number of cracks in the polymer/catalyst particle in the second stage (see Figure 2) can also reduce the degree of diffusion limitation. The change of kp with time is quite interesting (see Table II). The value of kp experienced a rapid decrease in 0–60 s, then partly recovered in the period of 90–120 s, and sharply decreased again until tp 5 300 s. Two factors may have caused the changes in kp. One is the diffusion barrier that leads to underestimation of kp when calculating it based on the equilibrium monomer concentration. On the other hand, among the multiple active centers of heterogeneous ZN catalysts, those with lower intrinsic reactivity (smaller kp) are more unstable than those with a higher kp value.27 Fast deactivation of the former will cause an increase in the observed kp value. As shown in Figure 1, there was about a 20–30% decrease in [C*] in the period of 30–60 s. When these two factors play dominant roles in different periods of polymerization, complex changes in kp may be observed. The introduction of the prepolymerization step did not change the kinetic behaviors in the first 5 min and only slightly enhanced the kp value at tp 5 30 s. The prepolymerization

45187 (4 of 6)


ARTICLE


Figure 3. Schematic model of the two-stage initiation process. The gray platelets represent MgCl2 crystallites with adsorbed TiCl4 as the precursor of active sites, and the zigzag lines represent polyethylene chains.

treatment also reduced the melting temperature and melting enthalpy of the polymer (see Table I), as insertion of a small amount of propylene in the polyethylene chain can reduce the crystallinity. CONCLUSIONS

By tracing the change of [C*]/[Ti] with time, the initiation of active sites in ethylene polymerization with a TiCl4/MgCl2-type ZN catalyst was studied. The initiation proceeded in two stages. In the first stage, [C*]/[Ti] quickly rose to about 1% in less

than 30 s and then remained stable in the subsequent 60 s. After 90 s of polymerization, the [C*]/[Ti] value started to increase again, forming the second buildup stage. The polymerization activity changes basically in parallel with the change in [C*]/[Ti]. The two-stage kinetics can be explained by first initiation of active sites on the outer surface of catalyst particles and subsequent initiation of active sites that are buried inside the particles. These buried sites are released in catalyst fragmentation by the expanding polymer phase. ACKNOWLEDGMENTS

Supports from the National Natural Science Foundation of China (Grant Nos. 21374094 and U1462114) are gratefully acknowledged. REFERENCES

1. McKenna, T. F. L.; Tioni, E.; Ranieri, M. M.; Alizadeh, A.; Boisson, C.; Monteil, V. Can. J. Chem. Eng. 2013, 91, 669. 2. Pimplapure, M. S.; Zheng, X. J.; Loos, J.; Weickert, G. Macromol. Rapid Commun. 2005, 26, 1155. 3. Skoumal, M.; Cejpek, I.; Cheng, C. P. Macromol. Rapid Commun. 2005, 26, 357. 4. Zheng, X. J.; Pimplapure, M. S.; Weickert, G.; Loos, J. e-Polymers 2006, no. 028. 5. Heuvelsland, A.; Wichmann, S.; Schellenberg, J. J. Appl. Polym. Sci. 2007, 106, 354. Figure 4. MWD curves of polymer samples collected at different polymerization times (with no prepolymerization).


6. Machado, F.; Lima, E. L.; Pinto, J. C.; McKenna, T. F. Macromol. React. Eng. 2009, 3, 47.

45187 (5 of 6)


ARTICLE


7. Taniike, T.; Thang, V. Q.; Binh, N. T.; Hiraoka, Y.; Uozumi, T.; Terano, M. Macromol. Chem. Phys. 2011, 212, 723.

17. Xu, T.; Yang, H. R.; Fu, Z. S.; Fan, Z. Q. J. Organomet. Chem. 2015, 798, 328.

8. Dwivedi, S.; Taniike, T.; Terano, M. Macromol. Chem. Phys. 2014, 215, 1698.

18. Zhang, B.; Zhang, L. T.; Fu, Z. S.; Fan, Z. Q. Catal. Commun. 2015, 69, 147.

9. Silva, F. M.; Broyer, J. P.; Novat, C.; Lima, E. L.; Pinto, J. C.; McKenna, T. F. Macromol. Rapid Commun. 2005, 26, 1846.

19. Yang, H. R.; Zhang, L. T.; Zang, D. D.; Fu, Z. S.; Fan, Z. Q. Catal. Commun. 2015, 62, 104.

10. Di Martino, A.; Broyer, J. P.; Spitz, R.; Weickert, G.; McKenna, T. F. Macromol. Rapid Commun. 2005, 26, 215.

20. Pater, J. T. M.; Weickert, G.; van Swaaij, W. P. M. AIChE J. 2003, 49, 180.

11. R€ onkk€ o, H.-L.; Korpela, T.; Knuuttila, H.; Pakkanen, T. T.; Denifl, P.; Leinonen, T.; Kemell, M.; Leskel€a, M. J. Mol. Catal. A: Chem. 2009, 309, 40.

21. Qi, M. Z.; Zhang, B.; Fu, Z. S.; Xu, J. T.; Fan, Z. Q. J. Appl. Polym. Sci. 2016, 133, 43207.

12. Yu, Y.; Busico, V.; Budzelaar, P. H. M.; Vittoria, A.; Cipullo, R. Angew. Chem. Int. Ed. 2016, 55, 8590. 13. Han-Adebekun, G. C.; Hamba, M.; Ray, W. H. J. Polym. Sci. Part A: Polym. Chem. 1997, 35, 2063. 14. Shen, X. R.; Hu, J.; Fu, Z. S.; Lou, J. Q.; Fan, Z. Q. Catal. Commun. 2013, 30, 66.

22. Tan, N.; Yu, L. Q.; Tan, Z.; Mao, B. Q. J. Appl. Polym. Sci. 2015, 132, 41816. 23. Chang, M.; Liu, X. S.; Nelson, P. J.; Munzing, G. R.; Gegan, T. A.; Kissin, Y. V. J. Catal. 2006, 239, 347. 24. McKenna, T. F. L.; Mattiolli, V. Macromol. Symp. 2001, 173, 149. 25. McKenna, T. F. L.; Di Martino, A.; Weickert, G.; Soares, J. B. P. Macromol. React. Eng. 2010, 4, 40.

15. Shen, X. R.; Fu, Z. S.; Hu, J.; Wang, Q.; Fan, Z. Q. J. Phys. Chem. C 2013, 117, 15174.

26. Chen, Y.; Liu, X. G. Polymer 2005, 46, 9434.

16. Yang, H. R.; Zhang, L. T.; Fu, Z. S.; Fan, Z. Q. J. Appl. Polym. Sci. 2015, 132, 41264.

27. Kissin, Y. V.; Mink, R. I.; Nowlin, T. E. J. Polym. Sci. Part A: Polym. Chem. 1999, 37, 4255.


45187 (6 of 6)
