Evaluation of the Initial Stages of Gas-Phase Ethylene Polymerizations with a SiO 2 -Supported Ziegler-Natta Catalyst

Full Paper

Evaluation of the Initial Stages of Gas-PhaseEthylene Polymerizations with a SiO2-Supported Ziegler–Natta Catalyst

Fabricio Machado,* Enrique Luis Lima, Jose Carlos Pinto,Timothy F. McKenna

The very early stages of gas-phase ethylene polymerization on an SiO2-supported Ziegler–Natta catalyst were studied with the help of a short-stop reactor. The short-stop-reactor-basedtechnique was useful in studying nascent polymerization, providing insights at very short,controlled times into important phenomenaregarding catalyst fragmentation and the acti-vation and deactivation of catalyst sites that takeplace during the very early stages of the reaction.Experimental results indicate that the growth ofthe polymer chains occurs at unsteady conditionsduring the initial stages of the polymerization.Hydrogen has a strong influence on the initialkinetics, leading to a significant decrease ofpolymerization activity. Polymer crystallinityincreases with the reaction time, probably dueto the formation of long chains with a high degreeof crystallinity.

F. MachadoQUATTOR, Rua Hidrogenio, 1404, Polo Petroquımico,CEP: 42810-010, Camacari, BA, BrazilFax: þ55 (71) 3632-2206E-mail: [email protected]. Machado, T. F. McKennaLCPP–CNRS/ESCPE-Lyon, 43 Blvd du 11 Novembre 1918, Bat 308F,BP 2077, 69616 Villeurbanne Cedex, FranceT. F. McKennaDepartment of Chemical Engineering, Queen’s University, 19Division Street, Kingston, ON, K7L 3N6, CanadaF. Machado, E. L. Lima, J. C. PintoPrograma de Engenharia Quımica/COPPE, Universidade Federaldo Rio de Janeiro, Cidade Universitaria, CP 68502, Rio de Janeiro,21945-970, RJ, Brazil

Macromol. React. Eng. 2009, 3, 47–57

� 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Introduction

The continuous advances in organometallic chemistry

over the course of recent decades have given us catalysts

capable of polymerizing at ever higher polymerization

rates. These increases in the intrinsic rates of reaction

have, in turn, significant implications in terms of reactor

operation, productivity, monomer incorporation, fines

generation, polymer particle morphology, etc. The mor-

phological control of polymer particles is desired because

of problems associated with process viability and reduc-

tion of plant operation costs. Generally, it is desired that

polymer particles present regular shape because the

absence of fines prevents reactor fouling problems

and undesirable fluidization effects.[1] In addition, the

DOI: 10.1002/mren.200800037 47

F. Machado, E. L. Lima, J. C. Pinto, T. F. McKenna

48

morphology of the resulting polymer particle will have a

strong influence on the rate at which monomers arrive

from the bulk phase of the reactor to the active sites. For

instance, in a polymer particle, the more it is necessary for

monomer to diffuse through layers of polymer to get to the

active sites, the more likely it is that the rate of reaction

and chain microstructure will be limited by mass

transfer.[2] It is therefore clear that a better understanding

of issues that control particle morphology would be

valuable.

Initial observations on the catalyst particle fragmenta-

tion date from 1950s. Natta and Pasquon,[3] based on

kinetic considerations, proposed the existence of a

progressive fragmentation front due to the development

of mechanical stresses caused by the polymer production

within the narrow channels of the porous particle. Other

works subsequently contributed to the establishment of a

basis of the fragmentation and growth mechanisms,

including the following concepts: that there is a multi-

granular structure and that polymer growth takes place on

the primary grains of the catalyst; that final polymer

particles replicate the shape of the catalyst particles; that

the polymerization rates may be reduced due to diffusion

limitations; and, that the support breaks up due to the

generation of mechanical stresses in the pores.[4,5]

The nascent phase of these polymerizations has been

widely studied, but more from the point of view of

exploring the relation between the crystallization, forma-

tion of the polymer chains and the impact of monomer

transport upon the polymer morphology.[6–13] In spite of

the importance of studies regarding the development of

polymer particle morphology, the real impact of the

reaction conditions on the morphogenesis of the polymer

particles has not been completely explained. The mor-

phology of the final polymer particle depends on the

mechanical and structural properties of both the catalyst

support and the polymer formed in the critical very early

stage of the polymerization.[14,15] However, very little is

currently known about the details of the fragmentation

process of heterogeneous Ziegler–Natta catalysts (distri-

bution of active sites on the catalyst carrier, fragmentation

mechanism of carrier, fragment distribution of carrier in

the polymer chain matrix, etc.). In part, this stems from a

lack of tools adapted to its study in the very early stages of

the olefin polymerizations.

More recently, studies have begun to focus on the

development of the morphology of the polymer particles

obtained with heterogeneous Ziegler–Natta catalyst sys-

tems at very short times of polymerization[16–18] and

under industrial conditions.[19–21] In particular, the studies

of Di Martino et al.[22–24] provided a detailed investigation

of the kinetics, morphology and polymer properties; these

were conducted in slurry polymerizations of ethylene on a

MgCl2-supported Ziegler–Natta catalyst at very short

Macromol. React. Eng. 2009, 3, 47–57

� 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

times (t � 40 ms) performed with a quenched-flow

apparatus. More recently, Olalla et al.[25] improved the

original gas-phase short-stop technique[5,26] in order to

study heat transfer in gas-phase olefin polymerization on

supported catalyst; they found that the rate of reaction has

a very strong influence on the rate of fragmentation and

also showed that the shape of the rate curve is dependent

on the initial conditions of the polymerization, suggesting

that the mechanism of fragmentation of the catalyst

carrier depends on the start-up of the polymerization. The

latter observation confirmed an earlier result of Knocke

et al.[27] These and other articles in a similar vein all point

to the fact that fragmentation influences the morphology,

but that the rate of polymerization during the fragmenta-

tion step and the physical properties of the polymer

present during this stage all interact in a complex manner.

Tools for the study of these aspects of olefin polymeriza-

tion on supported catalysts are therefore extremely useful.

Other techniques have been developed for the purpose

of elucidating the mechanism of fragmentation of

different carriers employed in the synthesis of hetero-

geneous catalysts used in olefin polymerizations. Among

them, we should mention the work of Eberstein et al.,[28]

Zollner and Reichert[29] and Knocke et al.,[30] based on

the concept of videomicroscopy (combination of a micro-

reactor and a microscope) to observe polymerizations

performed with heterogeneous catalysts. Pater et al.[31,32]

described a newmethod for online observation of polymer

particle growth, based on optical microscopy and infrared

images. With the aid of infrared images, it was possible to

determine the temperature of the particles surface in situ

during the polymerization. Microscopy has been used by

several research groups for in situ monitoring of homo-

polymerizations and copolymerizations reactions in the

gas phase, mainly because of the non-destructive feature

of this technique.[33] Recently, Abboud et al.[34–36]

employed videomicroscopy to follow the polymer particles

melting obtained from heterogeneous catalysts. Laser

scanning confocal fluorescence microscopy is also avail-

able for online monitoring of olefin polymerization.[37,38]

Although these techniques are useful for describing

certain features of olefin polymerization, one should bear

in mind that their use is frequently limited by the

resolution of videomicroscopy and infrared techniques.

Fluorescent microscopy is difficult to use under normal

reaction conditions, and techniques like calorimetry do not

always give fine-enough resolution of the reaction rate

curves during the critical initial period.

A detailed description of an experimental short-stop

reactor (SSR) developed to study fragmentation of catalyst

carrier, nascent particle morphology and reaction kinetics

in gas-phase polymerization of olefins on supported

catalysts has presented by Machado et al.[5,26] and

the reader is referred to these publications for

DOI: 10.1002/mren.200800037

Evaluation of the Initial Stages of Gas-Phase Ethylene Polymerizations . . .

detailed information. In this work, the SSR-based techni-

que is employed to evaluate the main features of the

heterogeneous catalyst fragmentation in the very early

stages of the gas-phase ethylene polymerizations carried

out using a Ziegler–Natta catalyst system supported on

silica.

Experimental Part

Pentane (SDS, France) with a minimum purity of 99% after pre-

treatment on 3 A molecular sieves was used for preparation of

cocatalyst solution and catalyst system slurry. Acid-washed glass

beads (Supelco, USA) with an average particle size of 75 mm were

used as seedbed. The beads were treated with triethylaluminum

(TEA), vacuum dried at 150 8C and kept under an argon

atmosphere at room temperature. Other chemicals were used

as received; argon (Air Liquide, France), with a minimum purity of

99.5%, was used to keep the reaction environment free of oxygen;

xylene (mixed isomers), with a minimum purity of 99% (Laurylab,

France), was used to extract the polymer from catalyst/polymer

particles; 2,6-di-butyl-4-methylphenol (BHT) (AcrosOrganics, USA)

with a purity of 99% was used as antioxidant. The TEA cocatalyst

was obtained from Witco (Germany). Ethylene with a minimum

purity of 99.5% and hydrogen with a minimum purity of 99.9%

were obtained from Air Liquide (France). Carbon monoxide with a

minimum of purity of 99.5% was used as reaction poison. A

proprietary SiO2-supported Ziegler–Natta catalyst with a titanium

content of 2.0 wt.-% was used for ethylene polymerization. This

catalyst system is similar to the one used by Aboudd et al.[35,36]

and the reader is referred to those works for more information

about the catalysts.

Gas-phase ethylene polymerizationswere carried out in a 1 cm3

fixed-bed microreactor at different temperatures and partial

pressures of ethylene and hydrogen. The stainless-steel micro-

reactor employed for the polymerizations was equipped with a

replaceable cartridge made of a sintered porous metal with an

average pore diameter of 7mm.A lid connected to a spring element

was used to maintain the powder bed, formed by the active

catalyst and the glass beads (used to assure dispersion of catalyst

and to disperse the heat generated by the polymerization

reaction), fixed inside the cartridge.

Gas feed lines were equipped with two miniature solenoid

valves (Asco Joucomatic, France) with a minimum time between

subsequent actions of 0.01 s. A pressure transmitter (Scaime ATM,

France) was used to monitor the reactor pressure. A heating bath

(Lauda E100, USA) maintained the temperature of the reaction

medium at the desired value. A T-type thermocouple (TC, France)

was used to monitor the inlet gas temperature, and a heating tape

(Aldrich, France) was used to maintain the gas temperature

constant and equal to the reactor temperature. The temperature of

the gas wasmeasuredwith a CHY 506 thermometer (CHY, France).

A logic controller (Crouzet, Millenium IIþ, France) was used to

control the time of opening of the miniature solenoid valves.

Initially the reactor was swept with argon until the reaction

temperaturewas reached, followed by evacuation of the system to

remove the inert gas. Then, the first solenoidwas opened, allowing

the preheated monomer to enter the reactor. When the pre-

Macromol. React. Eng. 2009, 3, 47–57

� 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

determined reaction time (depending on the reaction condition,

the time ranged from 0.1 to 6 s) was reached, the first valve was

closed and then immediately the second valve opened a CO2 line,

allowing the flow of CO2 and rapidly poisoning the catalyst and

stopping the reaction.

The following timescales were respected when the logical

controller acted on the opening and closing of the miniature

solenoid valves: first and second valve (response time 5–10 ms);

for each procedure the CO2 valve was maintained opened for 10 s.

The system was kept under isoperibolic conditions and the

polymerizations were performed with 20–50 mg of commercial

Ziegler–Natta catalyst and 0.8–1.0 g of pre-treated glass beads. The

polymer yieldswere determined byweighting the cartridge before

and after the polymerizations. The experimental setup used to

carry out the polymerization reactions was similar to the one

described byMachado et al.,[5,26] and the reader is referred to these

publications for a more detailed description of the protocol.

Polymer samples were separated from the catalyst/polymer

particles via a Soxhlet extraction technique with boiling xylene

(stabilized with BHT to avoid oxidative degradation) for 4 h and

vacuum dried at 100 8C. Samples so-extracted were analyzed by

differential scanning calorimetry (DSC) and gel permeation

chromatography (GPC). The melting temperature was determined

by DSC measurements in a Pyris 1 calorimeter (Perkin Elmer,

California, USA) at heating rates of 5 8C �min�1. The weight-

average molecular weight (MW) and the molecular-weight

distribution (MWD) of the polymers were measured by GPC

(Waters, Alliance GPCV 2000). The system was equipped with a

refractometer, a viscometer andWaters Styragel HT2 andHT6E gel

columns. Analyses were performed at 150 8C using trichloroben-

zene as the solvent. The weight-average molecular weight was

calculated by the standard procedure based on the universal

calibration of polystyrene.

Results and Discussion

Different experimental conditions were analyzed with the

purpose of studying the gas-phase ethylene polymeriza-

tion in the very early stages of the reaction, as shown in

Table 1.

Polymerization Kinetics

Figure 1–3 show the activity and the polymer yield profiles

obtained from polymerizations performed with an SiO2-

supported Ziegler–Natta catalyst at different tempera-

tures. Similar behaviors were observed at all three

pressures: the polymerization activity increases very

quickly, then decays exponentially with time, reaching a

pseudo-steady state rate around 0.5 s. The peak rates are

proportional to the pressure as onewould expect. As can be

seen in Figure 4, the deactivation continues at longer

times, but at much lower rates. Regarding the polymer

yields, it a significant increase was observed under all

experimental conditions analyzed during the first

www.mre-journal.de 49

F. Machado, E. L. Lima, J. C. Pinto, T. F. McKenna

Table 1. Gas-phase ethylene polymerizations using an SiO2-supported Ziegler–Natta catalyst.

Experiment T P TR }} << Mn Mw PI Tm DH xC

-C bar s gPE � gCATS1 gPE � gCATS1 �hS1 g �molS1 g �molS1 -C J � gS1 %

01 60 1.5 0.70 0.1978 1017.26 35229 94 419 2.68 127.93 140.29 51.95

02 60 4.0 0.70 0.4386 2255.82 30926 115732 3.74 128.52 132.61 49.11

03 60 4.0 0.70 0.5037 2590.52 35537 127977 3.60 123.83 100.58 37.25

04 60 8.0 0.10 0.2815 10 133.45 37026 98 276 2.65 122.24 62.42 23.11

05 60 8.0 0.30 0.5526 6631.58 27270 85 939 3.15 120.35 41.52 15.38

06 60 8.0 0.50 0.8368 6024.68 34844 167012 4.79 130.92 123.57 45.76

07 60 8.0 0.70 1.2002 6172.31 36921 214813 5.82 129.57 135.4 50.14

08 75 1.5 0.10 0.1816 1307.21 17616 39 264 2.23 119.45 33.98 12.58

09 75 1.5 0.30 0.2558 1315.54 13370 43 764 3.27 124.15 107.09 39.66

10 75 4.0 0.10 0.1780 6406.61 24278 47 937 1.97 124.64 76.61 28.37

11 75 4.0 0.30 0.2918 3502.09 33495 68 217 2.04 127.68 129.86 48.09

12 75 4.0 0.50 0.3732 2687.22 21279 77 668 3.65 128.18 156.02 57.77

13 75 4.0 0.70 0.4766 2451.19 19236 94 405 4.91 129.33 160.91 59.59

14 75 4.0 1.00 0.6042 2175.17 21296 86 354 4.05 124.71 112.89 41.80

15 75 4.0 2.00 1.0333 1859.97 20597 100420 4.88 129.48 164.51 60.92

16 75 4.0 4.00 1.5204 1368.39 23578 105719 4.48 129.41 176.66 65.42

17 75 4.0 6.00 2.0010 1200.59 19095 113199 5.93 128.16 161.65 59.86

18 75 8.0 0.10 0.3110 11 196.00 19493 78 226 4.01 127.44 154.87 57.35

19 75 8.0 0.30 0.5620 6743.78 16416 78 080 4.76 125.14 129.75 48.05

20 75 8.0 0.50 0.8925 6425.84 15728 80 069 5.09 126.20 145.29 53.80

21 75 8.0 0.70 1.1699 6016.47 13681 76 713 5.61 128.42 177.83 65.85

22 75 7.0 0.50 0.5831 4198.18 57394 308839 5.38 132.95 162.41 60.14

23 75 7.0 0.70 0.7645 3931.89 56750 308025 5.43 133.03 173.11 64.10

24 75 7.0 0.50 0.5911 4256.15 59444 328046 5.52 132.45 168.86 62.53

25 75 7.0 0.70 0.9022 4639.81 40598 205555 5.06 133.28 158.03 58.52

50

moments of the polymerization. Pronounced differences in

the polymerization activity profiles were observed only in

the experiments presented in Figure 3. It is very important

to emphasize that the work pressure range of the valve

used to control the gas line pressure was large (0–40 bar). It

was therefore very difficult to keep the ethylene pressure

constant in the experiments performed at 1.5 bar.

As shown in Figure 1 to 3, the reaction temperature does

not seem to significantly affect the polymerization rate or

yield (although both quantities are slightly higher at 75 8Cthan at 60 8C). Figure 5 illustrates the temperature effect on

the molar concentration of ethylene over an appropriate

range of pressure. The ethylene concentration profiles

were determined using the revised Soave–Redlich–Kwong

equation of state (SRK-2 EOS).[39] According to Figure 5, the

temperature does not have a significant influence on the

concentration of ethylene in the range of pressures used in

the experiments. Based on this observation, it is believed

that the difference in activity is not an effect of

Macromol. React. Eng. 2009, 3, 47–57

� 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

temperature, but that probably is linked to the variations

of pressure in the ethylene line or another effect not

detected during the experiments. On the other hand, it

should be considered that the rate constant for propaga-

tion may be slightly higher at 75 than at 60 8C; the two

effects may balance out to some extent, so that it looks like

the temperature has no affect.

The kinetic profiles of ethylene polymerizations carried

out in the short-stop polymerizations using SiO2-

supported Ziegler–Natta catalysts can be represented by

a relatively simple kinetic mechanism. Consider that a

kinetic mechanism represented only by propagation and

catalyst site deactivation steps is able to describe the

ethylene polymerization in the early stages of the reaction.

We will consider the presence of two effective families of

catalytic sites: one of them deactivates and the other

remains active during the initial stage of the polymerization.

This is, of course, a gross simplification, but it can be used

to account for the rate behavior observed in these

DOI: 10.1002/mren.200800037

Evaluation of the Initial Stages of Gas-Phase Ethylene Polymerizations . . .

Figure 1. Polymerization profiles with an SiO2/TiCl4 catalyst sys-tem at 8 bar: a) catalyst activity, and, b) polymer yield.

Figure 2. Polymerization profiles with an SiO2/TiCl4 catalyst sys-tem at 4 bar: a) catalyst activity, and, b) polymer yield.

experiments. Similarly, neglecting the contribution of

chain transfer and reinitiation steps on the reaction

behavior, it is possible to write a polymerization kinetic

mechanism, as shown in Table 2. Based on this proposed

mechanism, it is possible to obtain the polymerization rate

(<}) as follows:

Macrom

� 2009

<} ¼ kP

X1i¼1

Si

!M (1)

The equation for each individual catalyst site can be

represented as:

dS1dt

¼ �k1S1 ; S1 ¼ S01e�k1t (2)

dS2dt

¼ 0 ; S2 ¼ S02 (3)

ol. React. Eng. 2009, 3, 47–57

WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

After mathematical manipulation and combination of

Equation (1) to (3), the polymerization rate can be

expressed as:

<} ¼ A1 � ExpA2

T

� �� P

� fþ 1� fð Þ � Exp �A3 � tð Þ½ � (4)

Where T is the reaction temperature (K), P is the reaction

pressure (bar), f is themolar fraction of the active site and t

is the reaction time (s). Estimated parameters values are

shown in Table 3.

The adjusted parameters Ai can be expressed as follows:

A1 is related to the pre-exponential factor for polymer

propagation and expressed in gPE � g�1CAT �h

�1; A2 is related

to the activation energy (cal �mol�1) and the ideal gas

constant (cal �mol�1 �K�1), EA/R, and expressed in K; A3 is

related to the deactivation constant of the catalyst and

expressed in s�1.

www.mre-journal.de 51

F. Machado, E. L. Lima, J. C. Pinto, T. F. McKenna

Figure 3. Polymerization profiles with an SiO2/TiCl4 catalyst sys-tem at 1.5 bar: a) catalyst activity, and, b) polymer yield.

Figure 4. Polymerization profiles with an SiO2-supported catalyst:a) catalyst activity, and, b) polymer yield.

Figure 5. The effect of temperature on the ethylene concen-tration.

52

Figure 6 shows that the experimental data can be

predicted using the proposed model. As can be noted, the

proposed model (see Equation (4)) is capable of satisfacto-

rily describing the profiles of activity and polymer yield of

the reactions performed on this catalyst. However, it

seems reasonable to admit that a certain fraction of

catalyst sites are rapidly deactivated in the very early

stages of the polymerization.

Regarding the polymerization-rate behavior, similar

results were obtained by Di Martino et al.[21] in slurry

polymerizations of ethylene on a MgCl2-supported

Ziegler–Natta catalyst at very short times (t � 40 ms)

performed with a stopped-flow reactor, indicating that the

proposed model can be useful to describe the reaction

behavior in polymerization performed with heteroge-

neous Ziegler–Natta catalyst system.

Macromol. React. Eng. 2009, 3, 47–57

� 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim DOI: 10.1002/mren.200800037

Evaluation of the Initial Stages of Gas-Phase Ethylene Polymerizations . . .

Table 2. Simplified kinetic mechanism.

Step Equation

PropagationSi þ M �!kP

Siþ1

Catalyst Site DeactivationSi �!

k1S�0

Si �!k2

Si

Table 3. Estimated model parameters.

Parameter Value Unit

A1 720.77� 15.36 gPE � gCATS1 �hS1

A2 381.39� 63.49 K

A3 7.94� 2.83 sS1

f 0.33� 0.05 –

Figure 6. Polymerizations model predictions: a) correlationbetween the predicted and observed activities, and, b) catalystactivity.

Figure 7. The effect of hydrogen on the polymerization at 75 8C:a) catalyst activity, and, b) polymer yield.

Macromol. React. Eng. 2009, 3, 47–57

� 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Hydrogen Effect on the Polymerization Kinetics

Figure 7 and Table 4 show the effect of hydrogen on the

gas-phase ethylene polymerization kinetics in reactions

carried out at different conditions. It appears that

hydrogen significantly impacts the beginning of the

reaction, leading to a significant decrease of polymeriza-

tion activity in the initial 0.1 s when compared to reaction

activity obtained from polymerization performed in the

absence of hydrogen.

Table 4 also shows a comparison between the results of

catalyst activity and polymer yield obtained from experi-

ments performed under similar reaction conditions.

Apparently, hydrogen has an inductive effect upon the

catalyst sites at the beginning of the polymerization.

However, this effect seems to become less pronounced as

the reaction time increases.

www.mre-journal.de 53

F. Machado, E. L. Lima, J. C. Pinto, T. F. McKenna

Table 4. Hydrogen effect on the polymerization.Time Polymer Yield

P¼ 4 bar P¼ 8 bar

xH2¼ 0.00

mol � LS1

xH2¼ 0.09

mol � LS1

xH2¼ 0.00

mol � LS1

xH2¼ 0.09

mol � LS1

xH2¼ 0.09

mol � LS1

xH2¼ 0.17

mol � LS1

s gPE � gCATS1 gPE � gCATS1 gPE � gCATS1 gPE � gCATS1 gPE � gCATS1 gPE � gCATS1

0.10 0.1780 0.1069 0.3110 0.1327 0.1490 0.2205

0.30 0.2918 0.2958 0.5620 0.4955 0.5327 0.4019

0.50 0.3732 0.3560 0.8925 0.7032 0.7462 0.6575

0.70 0.4766 0.4321 1.1699 0.8711 0.9173 0.9200

Time Activity

P¼ 4 bar P¼ 8 bar

xH2¼ 0.00

mol � LS1

xH2¼ 0.09

mol � LS1

xH2¼ 0.00

mol � LS1

xH2¼ 0.09

mol � LS1

xH2¼ 0.09

mol � LS1

xH2¼ 0.17

mol � LS1

s gPE � gCATS1 �hS1 gPE � gCATS1 �hS1 gPE � gCATS1 �hS1 gPE � gCATS1 �hS1 gPE � gCATS1 �hS1 gPE � gCATS1 �hS1

0.10 6 406.61 3 849.75 11 196.00 4 778.45 5 362.72 7 939.06

0.30 3 502.09 3 549.18 6 743.78 5 946.03 6 392.70 4 822.73

0.50 2 687.22 2 563.52 6 425.84 5 063.26 5 372.96 4 733.79

0.70 2 451.19 2 222.28 6 016.47 4 479.87 4 717.61 4 731.35

54

Molecular Weight and Molecular Weight Distribution

Figure 8 and 9 show the evolution of both average

molecular weight and polydispersity indexes obtained

Figure 8. Average molecular weight, polydispersity indexes and crystallinity in polymer-izations using an SiO2-supported catalyst system at 75 8C and 4 bar.

from GPC analyses of resins produced

with the SiO2-supported catalyst. The

polydispersity indexes increase as a

function of the polymerization time,

indicating that the growth of the poly-

mer chains probably occurs at a non-

steady state during the very early stages

of the gas-phase ethylene reactions.

Given the timescales involved, it is

believed that the reaction stops before

the growth of the larger chains is

interrupted due to a chain transfer

mechanism. It is possible that the reac-

tion is poisoned before all of the chains

formed in the first population have

transferred off of the active sites.

Figure 8 and 9 exemplify the influence

of the polymerization conditions on the

weight- and number-average molecular

weights. According to Figure 8, consider-

ing the polymerizations performed at

75 8C and 4 bar, the weight-average

Macromol. React. Eng. 2009, 3, 47–57

� 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

molecular weight (Mw) increases while the number-

average molecular weight (Mn) remains constant as the

polymerization time increases (see Table 1). Figure 9

illustrates the temperature effect on the averagemolecular

DOI: 10.1002/mren.200800037

Evaluation of the Initial Stages of Gas-Phase Ethylene Polymerizations . . .

Figure 9. Average molecular weight, polydispersity indexes and crystallinity in polymer-izations using an SiO2-supported catalyst system: a) T¼60 8C, P¼8 bar; and,b) T¼ 75 8C, P¼ 8 bar.

Figure 10. Polymer molecular weight distributions: a) T¼ 75 8C, P¼ 7 bar; b) T¼ 75 8C, P¼ 8 bbar.

Macromol. React. Eng. 2009, 3, 47–57

� 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

weights. For example, Figure 9a shows a

significant increase in the Mw values,

while the Mn remains essentially con-

stant. Regarding the profiles of average

molecular weights presented in

Figure 9b, a different behavior was

observed: Mw remains constant while

Mn decreases slightly with the polymer-

ization time (see Table 1).

It is possible that the average mole-

cular weight is affected by the local

concentration of monomer at the cata-

lyst active site, contributing to the

formation of both large and small

polymeric chains, depending on the

monomer concentration accessible to

the catalyst site. In particular, the

experimental data presented in

Figure 8 and 9 show that different

profiles of the evolution of the molecular

weight can be obtained.

ar; c) T¼60 8C, P¼ 8 bar; and, d) T¼ 75 8C, P¼4

www.mre-journal.de 55

F. Machado, E. L. Lima, J. C. Pinto, T. F. McKenna

56

Modifications in the Mn and Mw profiles during the

polymerization indicate that the overall propagation and

chain transfer rates change during the early stages of the

polymerization. This may be caused by different factors,

which include the modification of catalyst sites, monomer

concentration, other species acting as effective chain

transfer agent, changes in the intraparticle temperature,

etc. Hence, important dynamic effects take place during

the initial stages of the gas-phase ethylene polymerization.

Figure 10 shows the MWDs of the polymeric resins

produced under different experimental conditions.

Depending on the reaction conditions, the shapes of the

MWDs change significantly in the initial stage of the

polymerization, as shown in Figure 10c. However, in some

cases, the shape of the MWD is, apparently, promptly

defined at the beginning of the polymerization. In these

cases, there were no pronounced changes in the shape of

the MWD after 0.5 s of polymerization, as shown in

Figure 10a and b. The different shapes of the MWDs

Figure 11. Typical DSC curves of the polymer samples: a) T¼ 75 8C,P¼ 1.5 bar; and, b) T¼ 75 8C, P¼4 bar.

Macromol. React. Eng. 2009, 3, 47–57

� 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

probably reflect the dynamic effects of activation and

deactivation of the catalyst sites, which distinctly respond

to the reactions conditions.

Crystallinity

The crystallinity of the polymer samples was determined

by means of the following relation:

xCð%Þ ¼ DHDSC

DH�f

100 (5)

WhereDHDSC is the heat of fusion of polyethylene obtained

from DSC analyzes and DH�f is heat fusion of polyethylene

100% crystalline (64.8 cal � g�1).[40]

Figure 8–9 and Table 1 show the dynamic evolution of

the polymer crystallinity. As can be observed, the crystal-

linity increases with the polymerization time, probably

due to the formation of the longer polymer chains with

higher degrees of crystallinity. Figure 11 shows typical DSC

curves for the polymer samples. Depending on the

polymerization time, the melting temperatures of the

polymer are significantly different (see also Table 1).

The crystallinity of the resins increases with the poly-

merization time, probably due to the formation of the

longer polymer chains that lead to different crystalline

formations in the very early stages of polymerization.

Conclusion

The initial stages of the gas-phase ethylene polymerization

were analyzed in reactions performed using a high activity

SiO2-Supported Ziegler–Natta catalyst system. Different

aspects of the polymerization were studied, including:

catalyst activities, polymer crystallinity, molecular weight

distributions, hydrogen effect, etc. The kinetics of the gas-

phase olefin polymerization on supported catalysts,

molecular weight distributions and polymer crystallinity

were successfully evaluated with an SSR-based technique

during the very early stages of the reaction.

Experimental results showed that the growth of the

polymer chains occurred at non-steady state conditions

during the initial stages of the polymerization. A

significant increase of the polymer yields and decrease

of the catalyst activity occurred simultaneously during the

first moments of the polymerization. It was also shown

that hydrogen significantly affected the beginning of

reaction, leading to a significant decrease of polymeriza-

tion activity, when compared to the reaction activity

obtained from polymerization performed in the absence of

hydrogen at polymerization times of the order of 0.1 s. It

was observed that the crystallinity of the resins increased

DOI: 10.1002/mren.200800037

Evaluation of the Initial Stages of Gas-Phase Ethylene Polymerizations . . .

with polymerization time, possibly due to the formation of

longer polymer chains that form crystals more readily, and

possibly due to crystals of a different form being found at

lower molecular weights

The kinetic model proposed was able to describe very

well the experimental data regarding catalyst activity and

polymer yield. The kinetic profiles of ethylene polymer-

izations carried out in the short-stop polymerizations

using SiO2-supported Ziegler–Natta catalysts can be

represented by a relatively simple kinetic mechanism,

considering only two types of catalyst active site.

Acknowledgements: The authors would like to thank Coorde-nacao de Aperfeicoamento de Pessoal de Nıvel Superior (CAPES,Brazilian Agency, Project N8 BEX 2813/03-3) and ConselhoNacional de Desenvolvimento Cientıfico e Tecnologico (CNPq) forproviding scholarships. This workwas also partially funded by theEC FP5 project ‘‘Polyprop: Polyolefins: improved property controland reactor operability’’, contract no. GSRD-CT-20012-0597 andproject no. GRD2-2000-30189.

Received: September 19, 2008; Revised: December 3, 2008;Accepted: December 4, 2008; DOI: 10.1002/mren.200800037

Keywords: catalyst fragmentation; gas-phase polymerization;polyethylene (PE); SiO2-supported Ziegler–Natta catalyst; Ziegler–Natta polymerization

[1] F. Machado, E. L. Lima, J. C. Pinto, T. F. McKenna, Eur. Polym. J.2008, 44, 1130.

[2] T. F. L. McKenna, Macromol. Symp. 2008, 260, 65.[3] G. Natta, I. Pasquon, Adv. Catal. 1959, 11, 1.[4] G. Cecchin, E. Marchetti, G. Baruzzi, Macromol. Chem. Phys.

2001, 202, 1987.[5] F. Machado, J. P. Broyer, C. Novat, E. L. Lima, J. C. Pinto, T. F.

McKenna, Macromol. Rapid Commun. 2005, 26, 1846.[6] B. Wunderlich, ‘‘Macromolecular Physics: Crystal Nucleation,

Growth, Annealing’’, Academic Press, New York 1976, Vol. 2.[7] H. D. Chanzy, E. Bonjour, R. H.Marchessault, Colloid Polym. Sci.

1974, 252, 8.[8] H. D. Chanzy, J. F. Revol, R. H. Marchessault, A. Lamande,

Kolloid Z. Z. Polym. 1973, 251, 563.[9] J. Loos, M. Arndt-Rosenau, U. Weingarten, W. Kaminsky, P. J.

Lemstra, Polym. Bull. 2002, 48, 191.[10] J. Loos, P. J. Lemstra, E. M. E. V. Kimmenade, J. W. Niemants-

verdriet, G. W. H. Hohne, P. C. Thune, Polym. Int. 2004, 53, 824.[11] E. M. Ivan’kova, L. Pmyasnikova, V. A. Marikhin, A. A. Baulin,

B. Z. Volchek, J. Macromol. Sci., Phys. 2001, B40, 813.[12] V.M. Egorov, E.M. Ivan’kova, V. A.Marikhin, L. P. Myasnikova,

A. Drews, J. Macromol. Sci., Phys. 2002, B41, 939.

Macromol. React. Eng. 2009, 3, 47–57

� 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

[13] S. W. Webb, W. C. Conner, R. L. Laurence, Macromolecules1989, 22, 2885.

[14] M. A. Ferrero, R. Sommer, P. Spanne, K. W. Jones, W. C. Conner,J. Polym. Sci., Part A: Polym. Chem. 1993, 31, 2507.

[15] D. A. Estenoz, M. G. Chiovetta, J. Appl. Polym. Sci. 2001, 81,285.

[16] H. Kono, T. Ichiki, H. Mori, H. Nakatani, M. Terano, Polym. Int.2001, 50, 568.

[17] H. Kono, H.Mori,M. Terano, Macromol. Chem. Phys. 2001, 202,1319.

[18] M. Yamahiro, H. Mori, K. H. Nitta, M. Terano, Macromol.Chem. Phys. 1999, 200, 134.

[19] H. Mori, M. Yamahiro, V. V. Prokhorov, K. H. Nitta, M. Terano,Macromolecules 1999, 32, 6008.

[20] H. Mori, M. Yamahiro, M. Terano, M. Takahashi, T. Matsu-kawa, Die Angew. Makromol. Chem. 1999, 273, 40.

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

[22] A. Di Martino, J.-P. Broyer, D. Schweich, C. D. Bellefon,G.Weickert, T. F. L. McKenna, Macromol. React. Eng. 2007, 1, 284.

[23] A. Di Martino, G. Weickert, T. F. L. McKenna, Macromol. React.Eng. 2007, 1, 229.

[24] A. Di Martino, G. Weickert, T. F. L. McKenna, Macromol. React.Eng. 2007, 1, 165.

[25] B. Olalla, J.-P. Broyer, T. F. L. McKenna, Macromol. Symp. 2008,271, 1.

[26] F. Machado, ‘‘Study on the Production of Polyolefin Materials:Prepolymerization and Synthesis of Propylene/1-Butene Copo-lymers’’, D.Sc. Thesis, COPPE/UFRJ, Rio de Janeiro, RJ, Brazil (inPortuguese) 2006.

[27] S. Knoke, F. Korber, G. Fink, B. Tesche, Macromol. Chem. Phys.2003, 204, 607.

[28] C. Eberstein, B. Garmatter, K.-H. Reichert, G. Sylvester, Chem.Ing. Tech. 1996, 68, 820.

[29] K. Zollner, K. H. Reichert, Chem. Eng. Technol. 2002, 25, 707.[30] S. Knoke, D. Ferrari, B. Tesche, G. Fink, Angew. Chem. Int. Ed.

2003, 42, 5090.[31] J. T. M. Pater, G. Weickert, W. P. M. Swaaij, Chimia 2001, 55,

231.[32] J. T. M. Pater, G. Weickert, W. P. M. Van Swaaij, AIChE J. 2003,

49, 450.[33] D. Ferrari, G. Fink, Macromol. Mater. Eng. 2005, 290, 1125.[34] M. Abboud, P. Denifl, K.-H. Reichert, Macromol. Mater. Eng.

2005, 290, 1220.[35] M. Abboud, P. Denifl, K.-H. Reichert, J. Appl. Polym. Sci. 2005,

98, 2191.[36] M. Abboud, P. Denifl, K.-H. Reichert, Macromol. Mater. Eng.

2005, 290, 558.[37] Y.-J. Jang, C. Naundorf, M. Klapper, K. Mullen, Macromol.

Chem. Phys. 2005, 206, 2027.[38] Y.-J. Jang, K. Bieber, C. N. N. Nenov, M. Klapper, K. Mullen,

D. Ferrari, S. Knoke, G. Fink, e-Polymers 2005, 13, 1.[39] L. S. Wang, J. Gmehling, AIChE J. 1999, 45, 1125.[40] M. Inoue, J. Polym. Sci., Part A: Gen. Papers 1963, 2697.

www.mre-journal.de 57